Molecular Cancer Research Genome w/Abstract AM No Date
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Molecular Cancer Research 5, 403-412, April 1, 2007. doi: 10.1158/1541-7786.MCR-06-0146
© 2007 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Schuster, K.
Right arrow Articles by Harris, L. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Schuster, K.
Right arrow Articles by Harris, L. C.

Signaling and Regulation

MDM2 Splice Variants Predominantly Localize to the Nucleoplasm Mediated by a COOH-Terminal Nuclear Localization Signal

Katja Schuster1, Liying Fan1 and Linda C. Harris1,2

1 Department of Molecular Pharmacology, St. Jude Children's Research Hospital, and 2 Department of Pathology and Laboratory Medicine, University of Tennessee Health Science Center, Memphis, Tennessee

Requests for reprints: Linda C. Harris, St. Jude Children's Research Hospital, 332 N Lauderdale St, Mail Stop 230, TN 38105. Phone: 901-495-3833; Fax: 901-495-4293. E-mail: Linda.Harris{at}stjude.org


   Abstract
Top
 Notes
 Abstract
Introduction
 Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 
Of the >40 alternative and aberrant splice variants of MDM2 that have been described to date, the majority has lost both the well-characterized nuclear localization signal (NLS1) and the nuclear export signal (NES) sequence. Because cellular localization of proteins provides insight regarding their potential function, we determined the localization of three different MDM2 splice variants. The splice variants chosen were the common variants MDM2-A and MDM2-B. In addition, MDM2-FB26 was chosen because it is one of the few variants described that contains the complete p53-binding site. All three splice variants predominantly localized to the nucleus. Nuclear localization of MDM2-A and MDM2-B was controlled by a previously uncharacterized nuclear localization signal (NLS2), whereas nucleoplasmic localization of MDM2-FB26 was mediated by NLS1. p53 and full-length MDM2 colocalized with the splice variants in the nucleus. MDM2-A and MDM2-B both contain a COOH-terminal RING finger domain, and interaction with full-length MDM2 through this domain was confirmed. MDM2-FB26 was the only splice variant evaluated that contained a p53-binding domain; however, interaction between MDM2-FB26 and p53 could not be shown. p14ARF did not colocalize with the splice variants and was predominantly expressed within the nucleoli. In summary, nuclear localization signals responsible for the nucleoplasmic distribution of MDM2 splice variants have been characterized. Colocalization and interaction of MDM2-A and MDM2-B with full-length MDM2 in the nucleus have important physiologic consequences, for example, deregulation of p53 activity. (Mol Cancer Res 2007;5(4):403–12)


   Introduction
Top
 Notes
 Abstract
 Introduction
Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 
To date, more than 40 splice variants of MDM2 have been detected in several different tumor types, including tumors of the ovary and bladder (1), brain (2), soft tissues (3-5), and the mammary gland (6). Expression of alternative MDM2 splice variants often correlates with a more advanced tumor stage (1, 2), and in a study of invasive breast cancers, the presence of aberrant splice forms was associated with decreased patient survival (6).

Splicing of the pre-mRNA of MDM2 occurs either through the use of alternative existing intron-exon borders or by an aberrant mechanism that results in splicing within introns and exons (6, 7). As a result, splice variants of the MDM2 gene encode proteins that lack important domains of the protein and that probably display altered biological activity.

The functions of MDM2 splice variants are unclear. Several studies have evaluated the cellular localization of various human (8-11) and murine (12) MDM2 proteins to predict potential activities. In a study by Evans et al. (9), the tumor splice variant MDM2-B was excluded from the nucleus and was visualized predominantly within the cytoplasm of MEFs. In another study, MDM2-{Delta}150-230 localized to the cytoplasm of U20S cells (11). Both of these MDM2 proteins lacked part of the NH2-terminal region that contains a nuclear localization signal (NLS), suggesting that loss of this signal prevented the nuclear entry of these two proteins. However, two other studies showed that the loss of this NLS did not prevent MDM2 entry into the nucleus of NIH3T3 and 293 cells, respectively (8, 10). Furthermore, an MDM2 splice variant (V3) lacking most of the p53-binding domain but containing the complete NLS was excluded from the nucleus and predominantly localized to the cytoplasm of MEF (12). These data show that cellular localization of MDM2 proteins is complex and seems to depend on factors other than the well-characterized NH2-terminal NLS (NLS1).

Because of the contradicting published information, our goal was to evaluate the cellular localization of three tumor MDM2 splice variants in vitro to (a) determine whether the lack of the p53 binding domain or the NH2-terminal NLS influenced the localization of MDM2; and (b) evaluate whether the MDM2 isoforms colocalized with known binding proteins; full-length MDM2, p53, or p14ARF (ARF).


   Results
Top
 Notes
 Abstract
 Introduction
 Results
Discussion
 Materials and Methods
 Acknowledgements
 References
 
Three splice variants were chosen for this project (Fig. 1 ): MDM2-A, which was the most common splice variant found in pediatric rhabdomyosarcoma; MDM2-B, which is the most common variant found in human cancer; and MDM2-FB26 (5), one of only two splice variants described to date that contains the complete p53 binding domain. MDM2-A and MDM2-B both lack most of the p53-binding domain, the well-characterized NLS1 and the NES. However, the MDM2-A protein contains an additional 78 amino acids compared with MDM2-B that includes the central acidic domain (Fig. 1). In contrast, splice variant MDM2-FB26 contains the complete p53-binding site and also NLS1. This variant lacks the COOH terminus due to an out-of-frame sequence and a premature stop codon.


Figure 1
View larger version (15K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 1. Schematic of full-length MDM2 protein and the splice variants MDM2-A, MDM2-B, and MDM2-FB26. NLS, nuclear localization signal; NES, nuclear export signal. Numbers, exons, nucleotides, or amino acids of MDM2 as indicated.

 
MDM2 Splice Variants and Full-Length MDM2 Localize to the Nucleus
To determine where the individual MDM2 proteins localize within a cell, MDM2-A, MDM2-B, MDM2-FB26, and full-length MDM2 proteins were expressed with fusion epitope tags in triple knock-out MEFs (TKO; null for p53, MDM2, and ARF). The splice variants MDM2-A and MDM2-B localized predominantly to the nucleus with faint expression in the cytoplasm (Fig. 2 ). MDM2-FB26 localized to the nucleoplasm and was excluded from the cytoplasm and the nucleoli. As predicted, full-length MDM2 seemed to be exclusively expressed into the nucleoplasm and excluded from the nucleoli and the cytoplasm (Fig. 2). Similar results were observed when each of the four MDM2 proteins expressed independently in double knock-out MEFs (DKO; null for p53 and MDM2; data not shown).


Figure 2
View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 2. Cellular localization of MDM2 splice variants and full-length MDM2 in MDM2/p53/ARF null murine fibroblasts (TKO). MDM2 proteins were visualized using anti-V5 or anti-Myc antibodies and an anti-mouse FITC-conjugated secondary antibody. The DNA in the nuclei was stained with Hoechst 33342. MDM2-A and MDM2-B localized to the nucleus with faint cytoplasmic expression. Full-length MDM2 and MDM2-FB26 expressed within the nucleus and were excluded from the nucleoli and the cytoplasm.

 
Alternate NLS Sequences Facilitate Nuclear Entry of MDM2 Splice Variants
The nuclear localization of MDM2 splice variants A and B, which lack the previously characterized NLS1 (8, 9, 11), suggested that alternate sequences might facilitate their nuclear import. By analysis of the MDM2 amino acid sequence, a potential second NLS (NLS2) was identified within the RING finger domain, which has previously been described as cryptic nucleolar localization signal (11). A minimal feature of the majority of NLSs is the motif K-X-X-K/R (lysine-X-X-lysine/arginine; ref. 13). NLS1 consists of amino acid numbers 181 to 185 with the motif RKRHK. The second potential transport signal (NLS2) consists of amino acid 466 to 473, resulting in three overlapping motifs within the same sequence KKLKKRNK.

The splice variant MDM2-A, which lacks NLS1, was mutated to introduce mutant sequence 1 (Fig. 3A ) into NLS2 to generate MDM2-ANLS2. The MDM2-ANLS2 protein was expressed predominantly within the nucleus with faint expression in the cytoplasm of TKOs (Fig. 3B) in an identical manner to the nonmutated protein. Similarly, splice variant MDM2-B, which also had mutant sequence 1 incorporated into its NLS2 and lacks the NLS1, localized predominantly in the nucleoplasm with visible expression in the cytoplasm of TKO cells (data not shown), as seen with the nonmutant protein. These results show that the introduction of mutant sequence 1 into the NLS2 motif could not inhibit protein transport to the nucleus and suggested that the NLS signal was not disrupted by the introduced mutations. Therefore, additional site-directed mutagenesis was undertaken to further mutate NLS2 in MDM2-A to determine whether additional changes could interrupt the cellular transport of the splice variant to the nucleus. In addition to the amino acid changes at position 469 and 470 in NLS2, lysine residues 466 and 467 were also mutated (Fig. 3A; mutant sequence 2). The MDM2-A protein sequence was changed as follows: K466N, K467Q, K469N, and K470E to generate MDM2-ANLS2+. These additional mutations changed the cellular localization of MDM2-A. In TKOs, MDM2-ANLS2+ was now expressed predominantly in the cytoplasm with faint nuclear staining (Fig. 3B), in contrast to the predominantly nuclear localization observed upon wild-type MDM2-A or mutant MDM2-ANLS2 expression. The NLS2 mutant protein was 54% to 58% nuclear, whereas following additional mutagenesis, only 20% to 24% of the NLS2+ protein was nuclear. These data show that additional mutations within NLS2 were necessary to reduce nuclear entry of MDM2-ANLS2+ and provide evidence for a second NLS in the COOH-terminal portion of the MDM2 protein. However, these data are in contrast to those obtained for full-length MDM2 protein. In full-length MDM2, only NLS1 functions as a NLS. When NLS1 was mutated, as shown in Fig. 3A, full-length MDM2 localized to the cytoplasm (Fig. 3C). In contrast, when NLS2 was mutagenized (NLS2+), full-length MDM2 remained nucleoplasmic (Fig. 3C). These data suggest that NLS2 only functions as a NLS in MDM2 splice variants. It seems that the change in conformation of the COOH-terminal RING finger in MDM2 splice variants lacking a central portion of the protein allows NLS2 to become unmasked.


Figure 3
View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 3. Mutagenesis of MDM2 NLSs (NLS1 and 2). A. Full-length MDM2 contains NLS1 and a putative NLS2. Mutations were introduced to eliminate each signal independently, as shown. MDM2-A lacks the first NLS (NLS1). Mutations were introduced into the second NLS (NLS2). B. MDM2-A containing NLS2 mutant sequence 1 (MDM2-ANLS2) was detected in the nucleus with visible expression in the cytoplasm. Additional mutations were introduced into NLS2 of MDM2-A (NLS2+). MDM2-A–containing NLS2 mutant sequence 2 (MDM2-ANLS2+) was excluded from the nucleus and localized predominantly within the cytoplasm. C. Full-length MDM2 with mutations in NLS1 resulted in cytoplasmic protein, whereas following the mutagenesis of NLS2 (NLS2+), the protein remained nuclear.

 
MDM2 Splice Variants Colocalize with Full-Length MDM2 in the Nucleus
To coexpress full-length MDM2 with its splice variants, the plasmids encoding the different proteins were transiently transfected into mutant TKO mouse embryonic fibroblasts. To distinguish between the different proteins when coexpressed, a secondary antibody conjugated to FITC (green fluorescence) was used to detect the primary antibody against the V5 tag of the splice variants, whereas a Texas Red (red fluorescence)–conjugated secondary antibody was used to detect the primary antibody against the MYC-tagged full-length MDM2 protein.

Dual staining to detect the three different MDM2 splice variants when coexpressed with full-length MDM2 revealed that as expected, all three colocalized with full-length MDM2 in the nucleoplasm (Fig. 4A ). Negative controls for antibody staining and staining of untransfected cells showed that the observed staining was specific.


Figure 4
View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 4. A. Coexpression of MDM2 splice variants with full-length MDM2. TKO cells were stained with anti-V5 antibody to detect the splice variants followed by a FITC-conjugated secondary antibody. Full-length MDM2 was visualized using an anti-Myc antibody and a Texas Red–conjugated secondary antibody. The DNA in the nuclei was stained with Hoechst 33342. The MDM2 splice variants colocalized with full-length MDM2 within the nucleoplasm. B. Immunoprecipitation of MDM2 splice variants following the expression of full-length MDM2 in TKO cells. Truncated MDM2 proteins were precipitated using a V5-specific antibody. Full-length MDM2 and splice variants were detected with an MDM2-specific antibody. MDM2-A and MDM2-B formed complexes with full-length MDM2. Splice variant MDM2-FB26 did not bind full-length protein. FL, full-length MDM2; A, MDM2-A; B, MDM2-B; FB26, MDM2-FB26; UT, untreated; and WCE, whole cell extract.

 
MDM2 Splice Variants Bind Full-Length MDM2 through Their RING Finger Domains
Because MDM2-A, MDM2-B, and MDM2-FB26 colocalized with full-length MDM2 in the nucleus, further experiments were undertaken to determine whether the MDM2 splice variants could bind full-length MDM2. The MDM2 splice variants and full-length MDM2 were coexpressed in TKOs and immunoprecipitated using an anti-V5 antibody (Fig. 4B).

As others have previously shown (9), MDM2-B was immunoprecipitated in complex with full-length MDM2 (Fig. 4B). In a similar manner, MDM2-A binding to full-length MDM2 was detected. The full-length MDM2 protein could not be detected in complex with FB26 (Fig. 4B). FB26 was the only variant evaluated that lacked the complete COOH-terminal region, including the RING finger domain necessary for MDM2/MDM2 interactions (Fig. 1), and therefore, our results are consistent with the idea that the interaction of MDM2-A and MDM2-B with full-length MDM2 occurred through the COOH-terminal RING domain.

Coexpression of MDM2 Splice Variants with p53 in the Nucleus
To determine the effect of p53 expression on the cellular localization of MDM2 splice variants, p53 was expressed in the mutant TKO mouse embryonic fibroblasts, together with the MDM2 splice variants. Expression of p53 was achieved by transduction with an adenoviral vector containing the p53 cDNA (as described in Materials and Methods). To visualize both proteins within the cell, a secondary antibody conjugated with FITC was used to detect the primary antibody against the V5-tagged splice variants; whereas a Texas Red–conjugated secondary antibody was used to detect the primary antibody against p53. As expected, p53 expression was primarily detected in the nucleoplasm and colocalized with the MDM2 splice variants MDM2-A, MDM2-B, and MDM2-FB26. These data are not shown but were identical to those in Fig. 4A for full-length MDM2.

MDM2 Splice Variants Do Not Interact with p53
MDM2-FB26 was the only splice variant evaluated that contained the complete p53-binding domain. MDM2-FB26 and the other splice variants colocalized with p53 in the nucleus, suggesting that they had the potential to form a complex with the p53 protein. Therefore, we evaluated whether MDM2-A, MDM2-B, and MDM2-FB26 could bind p53 protein by transiently transfecting the cDNAs of different MDM2 isoforms into TKOs. To express the p53 tumor suppressor, cells were transduced with adenoviral particles containing the p53 cDNA. Immunoprecipitation assays were carried out as described in Materials and Methods.

As previously shown, full-length MDM2 coprecipitated with p53 (Fig. 5 ). Splice variants MDM2-A and MDM2-B were unable to bind p53 (Fig. 5). These results were expected because both isoforms lack the p53-binding domain. However, surprisingly, FB26 was also unable to bind p53 protein, although it contains the complete p53-binding domain (Fig. 5). These data suggest that the lack of the COOH-terminal region in MDM2-FB26 might influence the structure of the p53-binding domain, preventing its interaction with p53.


Figure 5
View larger version (20K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 5. Immunoprecipitation of MDM2 splice variants following adenoviral transduction of p53 into TKO cells. A V5-specific antibody was used to immunoprecipitate truncated MDM2 proteins and an MDM2-specific antibody (SMP14) was used to immunoprecipitate full-length MDM2. MDM2 proteins were visualized with an MDM2-specific antibody (R&D Systems). Anti-p53 antibody was used to detect any p53 bound to the MDM2 proteins. As expected, full-length MDM2 bound p53. MDM2-A, MDM2-B, and FB26 did not complex with p53 protein.

 
Coexpression of MDM2 Splice Variants with ARF
To determine whether ARF expression influenced the localization of MDM2 splice variants, a retroviral vector containing the ARF cDNA was used to express ARF in the TKO fibroblasts. To distinguish between ARF and the MDM2 proteins, a secondary antibody conjugated with FITC was used to detect the primary antibody against the splice variant V5 tag, whereas a Texas Red–conjugated secondary antibody was used to detected the primary ARF antibody.

ARF localized predominantly within the nucleoli of the TKO (Fig. 6A ), and it did not influence the nuclear localization of the MDM2 splice variants.


Figure 6
View larger version (22K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 6. A. Coexpression of MDM2 splice variants and full-length MDM2 with ARF in TKOs. Cells were stained with anti-V5 antibody to detect the splice variants followed by an FITC-conjugated secondary antibody. ARF was visualized using an anti-ARF antibody and a Texas Red–conjugated secondary antibody. The DNA in the nuclei was stained with Hoechst 33342. ARF expressed predominantly within the nucleoli of the cell and was excluded from the nucleus and cytoplasm. MDM2-A, MDM2-B, and MDM2-FB26 did not colocalize with ARF. B. Immunoprecipitation of MDM2 following retroviral transduction of ARF into TKO cells. An ARF-specific antibody was used to immunoprecipitate ARF protein. An MDM2 antibody was used to visualize MDM2 proteins (R&D Systems). Full-length MDM2 forms a complex with ARF. ARF binding to the MDM2 splice variants could not be detected.

 
Immunoprecipitation of MDM2 Splice Variants with ARF
MDM2-A was the only splice variant evaluated that contained the acidic domain to which ARF has previously been shown to bind. Therefore, immunoprecipitation analysis was undertaken to determine whether MDM2-A and other MDM2 splice variants bound ARF protein. ARF was immunoprecipitated following retroviral transduction of TKOs as described in Materials and Methods. MDM2 proteins were detected using an MDM2-specific antibody (anti-MDM2 rabbit, R&D Systems, Minneapolis, MN).

Full-length MDM2 formed a complex with ARF, as shown in Fig. 6B. However, ARF binding to the MDM2 splice variants could not be detected. Unfortunately, due to the cross-reaction of the horseradish peroxidase–conjugated secondary antibody with the ARF antibody used for the immunoprecipitation, any ARF binding to FB26 would have been obscured. Expression levels of full-length MDM2 (90 kDa), MDM2-A (~53 kDa), MDM2-B (~45 kDa), and FB26 (~35 kDa) in the TKO cell extracts before the immunoprecipitation are shown in Fig. 6B.


   Discussion
Top
 Notes
 Abstract
 Introduction
 Results
 Discussion
Materials and Methods
 Acknowledgements
 References
 
In this study, the cellular localization of human MDM2 splice variants was investigated in the presence and absence of potential binding partners: p53, ARF, and full-length MDM2. The splice variants and full-length MDM2 were expressed in knock-out mouse embryonic fibroblasts (DKO, TKO). All of the MDM2 proteins, including those that lacked the previously characterized nuclear localization signal (NLS1), localized predominantly to the nucleus. Splice variants MDM2-A and MDM2-B both lack the well-characterized NLS1 and NES (Fig. 1), and yet they were detected predominantly in the nucleoplasm. After eliminating the possibility that the epitope tag may have played a role in the nuclear localization (data not shown), we suggested that MDM2-A and MDM2-B might contain an alternative signal that mediated the nuclear localization of these proteins.

The motif K-X-X-K/R (lysine-X-X-lysine/arginine) containing highly charged amino acids has been shown to mediate nuclear localization of several proteins, including Lamin A, human C-MYC, and SV40 large T (13). This motif is the minimal consensus sequence of the nuclear transport signals found in nuclear proteins. This motif was found thrice within the RING finger region as 466KKLKKRNK473 of MDM2 (1, 8, 11). This sequence has been previously described as a cryptic nucleolar localization signal that facilitates nucleolar localization of MDM2 upon ARF binding (11). However, the nucleolar localization motif is very similar to that of the NLS R/K-R/K-X-R/K (arginine/lysine, arginine/lysine X arginine/lysine; ref. 14) and the MDM2 RING finger sequence contains both motifs. It is unclear what other factors may be required to direct the localization of proteins containing motifs that could potentially mediate either nuclear or nucleolar localization.

To test the hypothesis that the cryptic nucleolar localization signal could also act as a NLS, multiple missense mutations were incorporated to disrupt all three possible transport motifs (NLS2; Fig. 3A). The amino acid substitutions chosen were similar to mutations that had previously been introduced into the NLS sequences of plant NPR1 (15) and viral SV40 large T (13) and which were confirmed to disrupt the ability of these proteins to localize to the nucleus. Initially, two mutations were generated in the NLS2 sequence (466KKLKKRNK473) of MDM2-A and MDM2-B changing lysine residues 469 and 470 into asparagine and glutamic acid, respectively (mutant sequence 1, 466KKLNERNK473, Fig. 3A). However, this first set of mutations within NLS2 did not interrupt the nucleoplasmic import of MDM2-A (Fig. 3B) and MDM2-B (data not shown). The mutants MDM2-ANLS2 and BNLS2 remained predominantly localized to the nucleoplasm identical to that observed for the wild-type MDM2-A and MDM2-B proteins (compare Figs. 2 and 3B). These data suggested that the amino acid changes in the three motifs 466KKLN469, 467KLNE470, or 470ERNK473 were not sufficient to disrupt the nuclear transport.

Because it has been previously shown that specific lysine residues can be very important in directing nuclear localization (13), we introduced two additional mutations into MDM2-ANLS2 that changed lysine residues 466 and 467 to asparagine and glutamine, respectively (466NQLNERNK473). These additional mutations changed the localization of MDM2-A to be predominantly cytoplasmic (Fig. 3B). These data show that Lys466 and Lys467 are critical for nuclear localization of MDM2-A. The COOH-terminal NLS signal (NLS2) has previously been described as a cryptic nucleolar localization signal that only becomes unmasked upon ARF binding to MDM2 (11). This sequence does not seem to facilitate nuclear localization of full-length MDM2 (Fig. 3C), but acts as a NLS only for MDM2 splice variants. The change in protein conformation that occurs when the central portion of MDM2 is deleted seems to unmask the NLS contained within the RING finger.

Our data are in contrast to two previously published studies that evaluated cellular localization of MDM2 splice variants and truncated MDM2 proteins (9, 11). Data from one of those studies (9) showed predominantly cytoplasmic expression of HDM2-ALT1 (human MDM2-B) in p53/MDM2 null mouse fibroblasts (DKO). In addition, truncated MDM2 isoforms, which lacked amino acids {Delta}150-230 including NLS1, were expressed only in the cytoplasm in U2OS cells (ARF–/–; ref. 11). However, our data are in agreement with two other studies that evaluated MDM2 splice variants and MDM2 deletion mutants (8, 10). In the first study, the deletion mutant MDM2-{Delta}491-154 that does not contain NLS1 expressed predominantly in the nucleus of NIH3T3 cells (8). The second study described an enhanced green fluorescent protein-MDM2-B fusion protein that localized to the nucleus in human HEK293 cells (10). A potential reason for the differences could be that expression of MDM2 proteins is influenced by the different cellular backgrounds in which these studies have been carried out. It is also possible that the different MDM2 proteins, which varied in their deleted domains, may display differences in their conformation that could be critical for unmasking the COOH-terminal NLS.

However, these reasons do not explain why we observed nuclear localization for MDM2-B compared with the cytoplasmic localization observed by Evans et al. (9). In both cases, the splice variant was expressed in p53/MDM2 null mouse embryonic fibroblasts (DKO), although it is likely that the cells used were of different passage number, which might have changed the characteristics of the cells in culture. However, this suggestion is speculative, and additional work would be required to determine the exact reason for the observed differences.

Previous studies have shown that both full-length MDM2 and p53 predominantly localize in the nucleoplasm (16, 17). Furthermore, p53, nucleolar ARF (18), and truncated MDM2 proteins have been shown to interact with full-length MDM2, resulting in an altered cellular localization. We therefore asked the question whether the splice variants evaluated in this study colocalized with full-length MDM2, p53, or ARF, and whether the MDM2 splice variant cellular localization was altered upon expression of these proteins.

Full-length MDM2, p53, and ARF were each coexpressed with individual MDM2 splice variants in mutant TKO fibroblasts. Full-length MDM2 and p53 were predominantly nuclear, whereas ARF was predominantly localized to the nucleoli (Figs. 4 and 6). Localization of the MDM2 splice variants did not change when coexpressed with these potential binding partners. In a previously published report, MDM2 splice variants with an intact COOH-terminal RING finger domain such as MDM2-B, bound full-length MDM2 and sequestered it in the cytoplasm (9). Our results are consistent with MDM2-A and MDM2-B interacting with full-length MDM2 through their RING finger domains (Fig. 4B) and suggest that full-length MDM2 and the splice variants can interact in the nucleus without being exported to the cytoplasm. This hypothesis is supported by a previous study that showed that the interaction of full-length MDM2 with its splice variants occurred independent of the cellular compartment (12). Therefore, data indicate that nuclear export is not necessary for the interaction of full-length MDM2 with MDM2-A and MDM2-B or for the activation of p53.

The p53 tumor suppressor protein was visualized in the nucleus as expected (19, 20) and colocalized with the MDM2 splice variants in the nucleoplasm (data not shown). Because the splice variant FB26 contains the p53-binding site, binding with p53 was predicted. However, no binding of the two proteins was observed (Fig. 5), suggesting either that the p53 binding domain of FB26 was not in the correct conformation, or that other domains within MDM2 might be needed to stabilize a p53/MDM2-FB26 complex. This result was surprising because a similar splice variant (MDM2-E) containing the p53-binding domain has previously been shown to bind p53 protein (1). With the exception of the NLS1 in FB26, the amino acid sequences that differ between MDM2-FB26 and MDM2-E contain no obvious motifs. Therefore, it is difficult to predict what other sequences may influence p53-MDM2 binding. MDM2-A and MDM2-B both lack the p53-binding domain, and as expected, these proteins did not bind p53 protein. Although expression of MDM2-A was very low in the Western blot analysis shown in Fig. 5, both MDM2-A and MDM2-B have previously been shown not to bind p53 (1). The interaction of MDM2 splice variants A and B with full-length MDM2 in the nucleus would likely inhibit the activity of full-length protein to appropriately regulate p53, thereby resulting in p53 activation.

To determine whether ARF influenced the cellular localization of MDM2 splice variants, ARF was expressed in the MEFs using a retroviral vector. ARF has been shown to localize to the nucleolus independent of either p53 or MDM2 expression (21, 22). However, once induced, for example, by oncogenic activation, ARF mediates the translocation of full-length MDM2 to the nucleolus, thereby releasing control of p53 (22). In this manner, cells are protected from oncogene-induced transformation by induction of p53-mediated apoptosis. Therefore, we asked the question whether elevated ARF expression could mediate the transfer MDM2 splice variants to the nucleolus.

Upon ARF expression, MDM2-A, MDM2-B, and MDM2-FB26 remained within the nucleoplasm (Fig. 6A). Of the three splice variants analyzed, MDM2-A was the only one that contained the central highly acidic domain to which ARF has been shown to bind (amino acids 210-304). Therefore, MDM2-A would be predicted to bind and colocalize with ARF in the nucleoli. However, immunoprecipitation analysis revealed that although MDM2-A contained the ARF binding site, it could not bind to the ARF protein (Fig. 6B), thus explaining the maintenance of its nuclear localization upon ARF expression. MDM2-B and MDM2-FB26 both lack the complete central acidic domain (amino acids 222-300 and 222-437, respectively), and these proteins also remained nucleoplasmic upon ARF expression, as expected.

In summary, all three MDM2 splice variants evaluated predominantly localized to the nucleus in mouse embryonic fibroblasts. This mechanism seems to involve two NLSs within the protein. Although the localization of the splice variants was not altered by the expression of full-length MDM2 or p53, their colocalization within the cell is probably important for their function. MDM2-A and MDM2-B colocalization with full-length MDM2 in the nucleoplasm would allow their binding and subsequent activation of p53. MDM2-FB26 does not contain the COOH-terminal RING finger domain but does contain the p53-binding site, although binding to p53 protein was not observed. Therefore, MDM2-FB26 would not be predicted to inactivate p53 protein in a similar manner to full-length MDM2 protein. MDM2-A that contained the ARF-binding site was not shown to bind ARF protein or localize to the nucleoli. Therefore, upon ARF activation, only full-length MDM2 would be predicted to move to the nucleoli with ARF. The expression of MDM2 splice variants within such a cell would also enhance p53 activation independent of ARF activity.


   Materials and Methods
Top
 Notes
 Abstract
 Introduction
 Results
 Discussion
 Materials and Methods
Acknowledgements
 References
 
Cell Culture
Mutant mouse embryonic fibroblasts were kindly provided by Dr. G. Lozano (MD Anderson, Houston, TX; p53/MDM2 null, DKO) and Dr. G. Zambetti (St. Jude Children's Research Hospital, Memphis, TN; p53/MDM2/ARF null, TKO). Cells were grown in DMEM supplemented with 10% fetal bovine serum and 2% L-glutamine in 10% CO2.

Generation of the MDM2 Splice Variant Constructs
The cDNAs of splice variants MDM2-A and MDM2-FB26 were amplified from total mRNA of primary rhabdomyosarcoma tumor samples as previously described (5) and subcloned into helper plasmid pCR 2.1-TOPO (Invitrogen, Carlsbad, CA). The cDNA of splice variant MDM2-B was obtained from Dr. John Lunec (University of Newcastle, United Kingdom). The full-length MDM2 cDNA was provided by Dr. Frank Bartel (Martin-Luther-University, Halle-Wittenberg, Germany). The variants were cloned into a mammalian expression vector containing the sequence for the V5 epitope tag (pcDNA4/V5-his, Invitrogen). Full-length MDM2 was cloned into a mammalian vector and expressed as a fusion protein with the MYC epitope tag (pcDNA6/MYC-his, Invitrogen). The plasmids were transiently transfected into mouse embryonic fibroblasts in which endogenous p53, MDM2, and ARF had been deleted (TKO). Some of the experiments as described in the text were carried out using MEFs in which only MDM2 and p53 were deleted (DKO).

Site-Directed Mutagenesis
The QuikChange Multi Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA) was used to generate mutations or correct PCR-introduced mutations within a DNA sequence. Oligonucleotides were generated with a melting temperature (Tm) at or above 75°C. The formula Tm = 81.5 + 0.41(%GC) – 675/N – % mismatch (N = number of nucleotides in primer) was used to calculate the correct Tm. The Hartwell Center (St. Jude Children's Research Hospital, Memphis, TN) synthesized the primers, which were phosphorylated at the 5' end. The primers used for mutagenesis of the NLSs are for NLS1 MutAS182/183 5'-CTGGTGAACGACAAAGAAATGGCCACAAATCTGATAG-3'; for NLS2 MutAS469/470 5'-GCAAAGAAGCTAAATGAAAGGAATAAGCCCTGCCC-3'; and MutAS466/467 5'-GGCCTGCTTTACATGTGCAAATCAGCTAAATGAAAGG-3'. Mutagenesis reactions were carried out as described in the manufacturer's manual.

Retroviral Transduction
To create ecotropic retroviral particles, the MSCV-IRES-GFP (23) or the pSR{alpha}MSVtkCD8 viral vector (Dr. Martine Roussel, St. Jude Children's Research Hospital, Memphis, TN) containing the cDNA of the gene of interest (MDM2 splice variants or ARF, respectively) was cotransfected with the helper plasmid pEQECO (Peter Houghton, St. Jude Children's Research Hospital, Memphis, TN) into 293T packaging cells (Mary-Ann Bjornsti, St. Jude Children's Research Hospital, Memphis, TN). Replication-incompetent retroviral particles were collected 48 h post-transfection. Fresh media were added to the 293T cells, and new retroviral particles were collected after another 24 h. The media were either used immediately for transduction or stored at –80°C.

Fibroblasts were plated at a subconfluent density and allowed to attach before removal of the growth media. The supernatant containing retroviral particles was filtered with a Steriflip filter (0.45 µm) to remove all cell debris. The supernatant was diluted 1:2 with complete media and Polybrene (1 mg/mL; Sigma, St. Louis, MO) was added to a final concentration of 1 µg/mL to enhance virus-host interaction. After 24 h, cells transduced with MSCV-IRES-GFP were analyzed for green fluorescent protein expression to confirm viral transduction and then incubated for another 24 h before Western blot analysis or immunofluorescence assay.

Adenoviral Transduction
Cells were plated at 70% confluency, and growth media were removed. Ad-p53 (Av1p53; Genetic Therapy Inc., Novartis Co., Gaithersburg, MD; ref. 24) containing the cDNA of p53 was diluted into 2 mL of adenoviral transduction media (2% fetal bovine serum) to a concentration of 100 plaque-forming units per cell before adding to the mouse embryonic fibroblast. After a 2-h incubation at 10% CO2 and 37°C, complete growth media were added. The culture media were changed after 24 h, and the cells were incubated for a further 24 h before Western blot analysis and immunofluorescence assays.

Immunofluorescence Assay
MEF cells (3 x 104 cells per well) were plated into chamber slides (Nalge Nunc Intern. Rochester, NY) and incubated at 37°C and 10% CO2 for 24 h before analysis. Cells in chamber slides were fixed with 1% paraformaldehyde (Sigma) in FA-buffer (Difco, Fisher, Fair Lawn, NJ) and permeabilized with 0.25% Triton (Surfact-Amps X-100, Pierce, Rockford, IL) in FA at room temperature. The slides were blocked with 10% swine serum (DAKO, Carpinteria, CA) in FA and stained with antibodies. MDM2 proteins were visualized using the following primary antibodies: anti-V5 mouse (Invitrogen), anti-MYC mouse (Invitrogen), anti-MDM2 rabbit (R&D Systems), and anti-Flag-M2 mouse (Sigma) followed by the corresponding secondary antibodies: anti-mouse FITC-conjugated, anti-mouse Texas Red–conjugated and anti-rabbit Texas Red–conjugated (all from Jackson Inc., West Grove, PA). Isotype-matched negative control antibodies anti-immunoglobulin G2a (IgG2a) mouse, anti-IgG1 mouse, anti-Ig rabbit (all from Jackson Inc.), and anti-IgG mouse (Santa Cruz Biotechnology, Santa Cruz, CA) were used to determine if the staining was specific. The nuclei were visualized using a 30 µmol/L Hoechst 33342 staining solution (Promega, Madison, WI). Cells were visualized with the fluorescence microscope Zeiss Axioplan 2 imaging at a magnification of 40x or 63x (Carl Zeiss Vision GmbH, Hallbermoos, Germany). Quantitations were done using Scanalytics Software (BD Biosciences, Rockville, MD).

To distinguish between ARF, p53, full-length MDM2, and the MDM2 splice variant proteins in the dual staining assays, a secondary antibody conjugated with FITC was used to detect the primary antibody against the MDM2 splice variant V5 tag, whereas a Texas Red–conjugated secondary antibody was used to detect the anti-p14ARF rabbit antibody (Novus Biologicals, Littleton, CO), the anti-p53 rabbit antibody (Santa Cruz Biotechnology), or the primary antibody against the full-length MDM2 Myc tag (anti–c-Myc-epitope rabbit, Affinity Bioreagents, Inc., Golden, CO).

Immunoprecipitation Assay
For immunoprecipitation, pelleted cells were resuspended in cooled immunoprecipitation-lysis buffer containing 20 mmol/L HEPES (Sigma; pH 7.5), 150 mmol/L sodium chloride, 0.1% Tween 20, and 10% glycerol. Samples were incubated on dry ice for 5 min and thawed for 5 min in a 37°C water bath (five times). A p14ARF-specific antibody, anti-p14 ARF (C-18) antibody (Santa Cruz Biotechnology), was used to precipitate ARF and its binding proteins from 200 µg of total protein for 1 to 2 h at room temperature. Immunoprecipitation complexes were washed thrice in ice-cold immunoprecipitation-lysis buffer. Proteins were separated by electrophoresis and transferred to polyvinylidene difluoride membranes.

Western Blotting
Samples containing 50 to 200 µg of total protein were electrophoresed using a Novex 4% to 20% gradient Tris-glycine gel (Invitrogen). Proteins were electroblotted to a polyvinylidene difluoride membrane. For immunodetection of proteins, membranes were washed thrice in 1x TBS-T [0.2 mol/L Tris (pH 7.6), 1.37 mol/L NaCl, and 10 mL of Tween] for 5 min each. Membranes were blocked with Blotto for 1 h at room temperature. Membranes were incubated with the primary antibodies diluted in Blotto (5% fat-free dry milk in 1x TBS-T [0.2 mol/L Tris (pH 7.6), 1.37 mol/L NaCl, and 10 mL of Tween]) between 1 h and overnight, and at room temperature or 4°C, respectively. The secondary antibodies conjugated with horseradish peroxidase were added, and the membranes were incubated at room temperature for 1 h. ECL Western blotting reagent kit (Amersham, Piscataway, NJ) was used for protein detection.


   Acknowledgements
Top
 Notes
 Abstract
 Introduction
 Results
 Discussion
 Materials and Methods
 Acknowledgements
References
 
We thank Dr. Pamela McKenzie for her assistance with the immunofluorescence staining procedure. We also thank Misty Cheney and Queen Rogers for technical assistance and Dr. Mary Danks for helpful discussions.


   Notes
Top
 Notes
Abstract
 Introduction
 Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 
Grant support: NIH grants CA92401, CA21765, and the American Lebanese Syrian Associated Charities.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Note: Current address for Katja Schuster: Simmons Comprehensive Cancer Center, University of Texas Southwestern Medical Center, Dallas, Texas.

Received 5/21/06; revised 1/ 5/07; accepted 2/ 6/07.


   References
Top
 Notes
 Abstract
 Introduction
 Results
 Discussion
 Materials and Methods
 Acknowledgements
 References
 

  1. Sigalas I, Calvert AH, Anderson JJ, Neal DE, Lunec J. Alternatively spliced mdm2 transcripts with loss of p53 binding domain sequences: transforming ability and frequent detection in human cancer. Nat Med 1996;2:912–7.[CrossRef][Medline]
  2. Matsumoto R, Tada M, Nozaki M, Zhang CL, Sawamura Y, Abe H. Short alternative splice transcripts of the mdm2 oncogene correlate to malignancy in human astrocytic neoplasms. Cancer Res 1998;58:609–13.[Abstract/Free Full Text]
  3. Bartel F, Meye A, Wurl P, et al. Amplification of the MDM2 gene, but not expression of splice variants of MDM2 mRNA is associated with prognosis in soft tissue sarcoma. Int J Cancer 2001;95:168–75.[CrossRef][Medline]
  4. Tamborini E, Della Torre G, Lavarino C, et al. Analysis of the molecular species generated by MDM2 gene amplification in liposarcomas. Int J Cancer 2001;92:790–6.[CrossRef][Medline]
  5. Bartel F, Taylor AC, Taubert H, Harris LC. Novel mdm2 splice variants identified in pediatric rhabdomyosarcoma tumors and cell lines. Oncol Res 2002;12:451–7.
  6. Lukas J, Gao DQ, Keshmeshian M, et al. Alternative and aberrant messenger RNA splicing of the mdm2 oncogene in invasive breast cancer. Cancer Res 2001;61:3212–9.[Abstract/Free Full Text]
  7. Bartel F, Taubert H, Harris LC. Alternative and aberrant splicing of MDM2 mRNA in human cancer. Cancer Cell 2002;2:9–15.[CrossRef][Medline]
  8. Brown DR, Thomas CA, Deb SP. The human oncoprotein MDM2 arrests the cell cycle: elimination of its cell-cycle-inhibitory function induces tumorigenesis. EMBO J 1998;17:2513–25.[CrossRef][Medline]
  9. Evans SC, Viswanathan M, Grier JD, Narayana M, El-Naggar AK, Lozano G. An alternatively spliced HDM2 product increases p53 activity by inhibiting HDM2. Oncogene 2001;20:4041–9.[CrossRef][Medline]
  10. Evdokiou A, Atkins GJ, Bouralexis S, et al. Expression of alternatively-spliced MDM2 transcripts in giant cell tumours of bone. Int J Oncol 2001;19:625–32.[Medline]
  11. Lohrum MA, Ashcroft M, Kubbutat MH, Vousden KH. Identification of a cryptic nucleolar-localization signal in MDM2. Nat Cell Biol 2000;2:179–81.[CrossRef][Medline]
  12. Dang J, Kuo ML, Eischen CM, Stepanova L, Sherr CJ, Roussel MF. The RING domain of Mdm2 can inhibit cell proliferation. Cancer Res 2002;62:1222–30.[Abstract/Free Full Text]
  13. Roberts B. Nuclear location signal-mediated protein transport. Biochim Biophys Acta 1989;1008:263–80.[Medline]
  14. Weber JD, Kuo ML, Bothner B, et al. Cooperative signals governing ARF-mdm2 interaction and nucleolar localization of the complex. Mol Cell Biol 2000;20:2517–28.[Abstract/Free Full Text]
  15. Kinkema M, Fan W, Dong X. Nuclear localization of NPR1 is required for activation of PR gene expression. Plant Cell 2000;12:2339–50.[Abstract/Free Full Text]
  16. Mayo LD, Donner DB. A phosphatidylinositol 3-kinase/Akt pathway promotes translocation of Mdm2 from the cytoplasm to the nucleus. Proc Natl Acad Sci U S A 2001;98:11598–603.[Abstract/Free Full Text]
  17. Rubbi CP, Milner J. Non-activated p53 co-localizes with sites of transcription within both the nucleoplasm and the nucleolus. Oncogene 2000;19:85–96.[CrossRef][Medline]
  18. Lindstrom MS, Klangby U, Inoue R, Pisa P, Wiman KG, Asker CE. Immunolocalization of human p14(ARF) to the granular component of the interphase nucleolus. Exp Cell Res 2000;256:400–10.[CrossRef][Medline]
  19. Shaulsky G, Goldfinger N, Peled A, Rotter V. Involvement of wild-type p53 protein in the cell cycle requires nuclear localization. Cell Growth Differ 1991;2:661–7.[Abstract]
  20. Shaulsky G, Goldfinger N, Tosky MS, Levine AJ, Rotter V. Nuclear localization is essential for the activity of p53 protein. Oncogene 1991;6:2055–65.[Medline]
  21. Tao W, Levine AJ. P19(ARF) stabilizes p53 by blocking nucleo-cytoplasmic shuttling of Mdm2. Proc Natl Acad Sci U S A 1999;96:6937–41.[Abstract/Free Full Text]
  22. Weber JD, Taylor LJ, Roussel MF, Sherr CJ, Bar-Sagi D. Nucleolar Arf sequesters Mdm2 and activates p53. Nat Cell Biol 1999;1:20–6.[CrossRef][Medline]
  23. Persons DA, Allay JA, Allay ER, et al. Retroviral-mediated transfer of the green fluorescent protein gene into murine hematopoietic cells facilitates scoring and selection of transduced progenitors in vitro and identification of genetically modified cells in vivo. Blood 1997;90:1777–86.[Abstract/Free Full Text]
  24. Pirollo KF, Hao Z, Rait A, et al. p53 mediated sensitization of squamous cell carcinoma of the head and neck to radiotherapy. Oncogene 1997;14:1735–46.[CrossRef][Medline]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Schuster, K.
Right arrow Articles by Harris, L. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Schuster, K.
Right arrow Articles by Harris, L. C.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online