Plasmids.

Plasmids pC53-SN3 (27) containing human p53 cDNA under control of the cytomegalovirus promoter and the empty vector pCMV-Neo-Bam were provided by Bert Vogelstein. Constructs carrying mutant p53s were constructed by replacing the SgraI-StuI p53 portion with the same fragment of pLS89 derivatives containing the p53 mutations (23, 24). p53 constructs were confirmed by sequencing. Tet-Off cell lines were created with the pBI-EGFP plasmid (BD Biosciences Clontech).

SaOS2 cell lines with inducible p53 expression.

p53 tetracycline-inducible SaOS2 cell lines were established with a Tet-Off gene expression system (BD Biosciences Clontech). Tetracycline-responsive p53 expression constructs were made by cloning restriction fragments of wild-type (wt) and mutant p53 cDNAs into the tetracycline (Tet)-responsive reporter plasmid pBI-EGFP using the pCMV-p53 constructs described above. Constructs were confirmed by sequencing. The plasmids containing the p53s were cotransfected into SaOS2 Tet-Off cells (Clontech), which express the tetracycline-controlled transactivator (17) along with p-puromycin in the presence of transfection reagent FuGENE 6. Single colonies were obtained by limiting dilution with 400 μg/ml G418 and 250 μg/ml puromycin in the presence of 1 μg/ml doxycycline (dox) for 2 to 3 weeks. To obtain the stable cell lines (at least five for each construct), potential clones were first selected with puromycin. After two rounds of selection, putative cell lines were identified. Clones that had low background green fluorescent protein and homogeneous green fluorescent protein induction after removal of dox were selected and analyzed for p53 expression by Western blot analysis. p53 expression was kept “off” by 2 μg/ml dox. To induce the p53s, cells were washed 3× with phosphate-buffered saline (PBS) and placed in medium lacking dox.

Growth suppression.

Growth suppression was determined as reduction in colony formation. Cells (10⁵) were plated on a 35-mm plate, and 24 h later were transfected with 1.5 μg of p53 expression plasmid or pCMV-Neo empty vector. Forty-eight hours later, the cells were trypsinized and 10² cells/well were plated on a six-well plate. After 3-week G418 selection (0.5 mg/ml), colonies with more than 50 cells were determined. Colonies were methanol fixed, Giemsa stained, and counted.

Impact of UV and γ irradiation on cell survival.

For UV and ionizing radiation (γ), SaOS2 Tet-Off-inducible cell lines were grown in six-well plates in the presence of dox. Six hours prior to treatment, the cells were washed 3× in PBS. For UV treatment, cells were irradiated in a thin layer of PBS using a germicidal lamp at 1 J m²/s. For γ treatment, cells were irradiated at 1.56 Gy/min from a Shepherd cesium irradiator in the culture medium at room temperature. A clonogenic assay was used to determine cell survival. After irradiation, cells were trypsinized, resuspended in fresh growth medium with no dox, and reseeded at 10² cells/well in a six-well plate. Clonal survival was determined as described for the growth suppression assay.

Apoptosis and cell cycle evaluation.

Apoptosis was detected by morphology, fluorescence-activated cell sorter (FACS), and annexin V assays. For morphology, cells were stained with 5 mg/ml propidium iodide and 50 mg/ml acridine orange as described by Qutob and Ng (41). The percent apoptosis (>300 cells counted) was determined with a fluorescence microscope. For cell cycle and apoptosis evaluation by FACS analysis, adherent and detached cells were combined and fixed overnight with 75% ethanol in PBS at 4°C. After rinsing 2× with PBS, cells were incubated for 30 min with 1 ml of PBS containing 1 mg boiled RNase at 37°C. Cells were stained in 1 ml of PBS containing 1 mg of propidium iodide, and ~10⁴ cells were analyzed in a flow cytometer (FACSort; Becton Dickinson). Annexin analysis was done with the V-PE/7-AAD apoptosis detection kit (BD Pharmingen) according to protocol on ~10⁴ stained cells using a flow cytometer, where single positive staining for phycoerythrin (PE) corresponds to early apoptotic events, whereas double staining for PE and 7-amino-actinomycin (7-AAD) represents both late-apoptosis and necrotic cells. Absence of staining for both fluorochromes corresponds to live cells.

Western blot analysis.

Equal amounts of whole-cell extracts were separated on BisTris NuPAGE (4 to 12%) and transferred to polyvinylidene difluoride membranes (Invitrogen). The blots were probed with primary antibodies and bands were detected using horseradish peroxide-conjugated secondary antibodies (Santa Cruz) and the enhanced chemiluminescence detection system (Amersham). Specific monoclonal antibodies were as follows: p53 (pAb1801 and DO-1), p-p53 (Ser15-R), p-p53 (Ser392), hMdm2 (C-18), Bax (N-20), p21 (C-19), Gadd45 (C-20), Msh2 (N-20), and actin (C-11).

Chromatin immunoprecipitation assay.

Chromatin immunoprecipitation (ChIP) assays were as described by Kaeser and Iggo (25) using 1 μg mouse monoclonal anti-p53 antibody DO7 (BD Pharmingen). Amplification and PCR conditions are described in the supplemental material.

TaqMan real-time PCR.

Relative gene expression levels and differences among transient and stable transfected p53 SaOS2 cells were validated by TaqMan real-time PCR (20). The approach is detailed in the Materials and Methods in the supplemental material. Reactions were carried out in 96-well plates using the ABI PRISM 770 sequence detection system platform (PE Applied Biosystems) according to the manufacturer's recommendations.

Transfected p53 mutants result in many different phenotypes.

Initially, we determined the ability of the various p53 mutants to influence growth suppression and apoptosis when transfected into a cancer cell line, SaOS2, which does not express p53. The pCMV-Neo plasmids carrying the wild type or a specific p53 mutant were transfected using FuGENE 6 transfection reagent (see Materials and Methods). At 24 h posttransfection, p53 protein levels were comparable (Fig. ​(Fig.1A).1A). To evaluate the ability of mutant p53s to affect overall clonogenic survival in human cells, colony forming ability was determined by selection for the p53 expression vector neomycin (G418) marker 2 weeks after transfection (see Materials and Methods). As shown in Fig. 1B and C, the mutants exhibited dramatic differences in the ability to reduce colony formation and induction of apoptosis 24 h after transfection.

Transfection with wt p53 induced apoptosis in nearly 30% of the cells and led to an 80% reduction in clonal survival. The G245S and G279E mutants which are inactive for transactivation and, in the case of G245S, unable to bind p53 target REs (8, 36, 61), had little impact on these biological end points. Cells expressing the supertrans S121F and N288K mutants were clearly more effective at inducing growth inhibition and apoptosis than the other altered alleles, resulting in ~60% apoptotic cells versus 30% with wt p53 (Fig. ​(Fig.1C).1C). The other two alleles that were supertrans in yeast, T123P and H178Y, were much closer to the wild type in their effects, as was G279R. The T125R mutant appeared to be intermediate between the wt and null mutants in its effects. Mutations in the same domain, and even amino acid changes at the same position (G279) in the second helix domain, can have different biological impacts. As summarized in Table ​Table1,1, six of the eight mutants clearly retained p53-related growth suppression and apoptotic functions when ectopically expressed in SaOS2 cells. The retention of biological activity for most of these mutants has also been observed for transfection into other cell lines that have endogenous wt p53 activity (data not shown). Based on the results with the transfected p53s, there are several distinct patterns of biological responses for the eight mutants examined.

Differences in biological consequences of stably expressing p53 mutants with altered transactivation activities.

To better characterize the biological consequences of the various p53 mutants and to address effects on additional end points, SaOS2 cell lines were created containing p53 mutants that are tightly regulated. The p53s were regulated by the Tet-Off (BD Biosciences Clontech) tetracycline-controlled system. In the absence of the tetracycline derivative dox, the transactivator binds to the tetracycline response element in the minimal cytomegalovirus promoter allowing it to drive expression of the p53 cDNA. Transcription can be turned off in a dox-dose-dependent manner (17).

Cell lines were isolated that expressed p53 at biologically meaningful levels. The levels were comparable to those induced by damage in wild-type cells that are capable of expressing endogenous p53 (see Fig. S1 in the supplemental material) or in tumor cell lines that exhibit continuous mutant p53 expression (5). As shown in Fig. ​Fig.2A,2A, the selected clones for the wild-type p53 (DM wild type 10) and the mutants S121F (DM S121F-8), T125R (DM T125R-78), G279E (DM G279E-7), and G279R (DM G279R-76) did not produce p53 when cells were grown in the presence of dox. Removal of dox led to comparable levels of induction of the p53 protein, similar to levels detected in UV-irradiated p53⁺ cell lines (RKO, HT1080, and U2OS cells) (see Fig. S1 in the supplemental material).

Open in a separate window

FIG. 2.

Inducible regulated expression of p53 mutants results in different levels of apoptosis induction. (A) Western blot analysis of p53 induction. Stable SaOS2 cell lines with the different p53 transgenes regulated by the Tet-Off promoter were cultured in medium containing dox. Lanes 1 to 5 correspond to p53⁻. Lanes 6 to 10 correspond to cells growing for 24 h in medium lacking dox (i.e., p53 ON). Immunodetection of actin was used as an internal control. The numbers following the p53 designations refer to the cell lines. (B) Induction of apoptosis by expressed p53 mutants. SaOS2 cells were harvested 24 h after dox removal from the medium (i.e., p53 ON). The cells were stained with propidium iodide and subjected to flow cytometry. (C) Evaluation of apoptosis by annexin V staining and flow cytometry. SaOS2 inducible cell lines were harvested and counted 24 h after dox removal (p53 ON). Cells (1 ×10⁵) from each sample were stained with PE-conjugated annexin V and then analyzed by FACS. Presented are the average percentages and SD from three independent experiments. WT, wild type.

To address the effects of the various mutants on apoptosis, the presence of sub-G₁ cell populations at 24 h after dox removal was determined using flow cytometry (summarized in Fig. ​Fig.2B).2B). Apoptosis was also determined by the appearance of annexin V-stained cells (Fig. ​(Fig.2C).2C). The patterns of apoptosis induction by the stably expressed mutants in relation to wt p53 were comparable to those found when the mutants were transiently expressed (summarized in Table ​Table11).

Using FACS analysis and assessment of DNA content, the effects of the p53 mutants on cell cycle progression in the nonapoptotic cells were also determined (see Fig. S2 in the supplemental material). There was no effect on cell cycle profile for the mutants T125R, H178Y, G245S, G279R, and G279E or for cells that do not express p53. However, expression of the wild type and the T123P p53 mutant resulted in a decrease in the fraction of cells in S phase and an increase in G₁ cells. Cells expressing the p53 mutants S121F and N288K showed an even more prominent G₁ arrest and almost no S-phase cells, in addition to the high levels of apoptosis induced.

At 24 h after induction of p53, the presence of dead cells, which corresponds to the sum of apoptotic plus necrotic cells (see Materials and Methods), was assessed (Fig. ​(Fig.3A).3A). The expression of the wild type, T123P, H178Y, and G279R resulted in similar frequencies of dead cells. The G245S and G279E proteins had little effect on cell death, while the T125R was intermediate between these and the wild-type protein.

Open in a separate window

FIG. 3.

Expression of p53 mutants affects resistance to UV and γ radiation. (A) Cell death. SaOS2 cells with p53 mutants under the Tet-inducible promoter were grown in the presence of dox (p53 OFF). Six hours before UV or γ irradiation, dox was removed (p53 ON). Cells were examined at 24 h after exposure to UV (10 J/m²) or γ radiation (4 Gy) or no radiation. Total cell death is the sum of early and late apoptosis as well as necrosis quantified by flow cytometry analysis of 7-AAD- and annexin V-PE-stained cells. (B) Clonal survival. SaOS2 cells containing p53 mutants under the Tet-inducible promoter were grown in the presence of dox (p53 OFF) or in the absence of dox (p53 ON), and percent clonal survival was determined 21 days later. (C) Changes in sensitivity to UV resulting from expressed p53 mutants. (D) Changes in sensitivity to γ radiation resulting from expressed p53 mutants. Clonal survival of radiation-treated cells containing induced p53 was compared to p53 ON cells that were not irradiated. Presented are results and SD from three independent experiments. WT, wild type.

p53 has broad biological effects on cellular responses to DNA damage. Its impact on survival following exposure to DNA-damaging agents reflects its influence on a variety of biological processes that include apoptosis, checkpoint arrest, and DNA repair. Surprisingly, although p53 is considered a “guardian of the genome,” a lack of p53 can result in resistance to DNA-damaging agents, which may reflect opposing effects of these various processes (7, 31, 32, 59, 64). We therefore evaluated the consequences of mutational changes in the p53 regulatory network on sensitivities to UV and γ radiation measured as both clonal survival and damage-induced cell death, with the view that some alterations in the network might be beneficial and others might not.

To ensure that p53 would be present at the time of irradiation (see Fig. S3 in the supplemental material), the dox medium was changed to medium lacking dox at 6 h prior to exposure. Following UV (0, 5, or 10 J/m²) or γ irradiation (0, 2, 4, or 8 Gy), cells were continued in the dox⁻ medium (p53 ON). In the absence of radiation, clonal survival varied from 20% for the S121F and N288K mutants to 30% for the wild type to ~100% for the transactivation-inactive mutants G245S and G279E (Fig. ​(Fig.3B),3B), similar to growth suppression results obtained with transfectants (Table ​(Table11).

To determine the specific effects of the expressed p53s in response to UV, the clonal survival rates of irradiated cells were compared to those of unirradiated cells (Fig. ​(Fig.3C3C and Table ​Table1).1). The presence of wt p53 greatly increased UV sensitivity compared to p53 null cells, while the mutant p53s had various impacts on UV tolerance. Cells expressing loss-of-function G245S and G279E mutant p53s, as well as untransfected SaOS2, were resistant to UV radiation. The presence of S121F and N288K mutant proteins increased sensitivity over that of wt p53, while cells expressing H178Y, G279R, or T125R were more resistant.

Loss of UV clonal survival due to the presence of wt or mutant p53s appears related to the ability of UV to induce cell death at 24 h via the p53 pathway (Fig. ​(Fig.3A3A and Table ​Table1).1). Cells that were null for p53 or expressing the G279E or G245S proteins exhibited high clonal survival and little UV-induced cell death (i.e., “+++++” for survival and “+” for induced cell death in Table ​Table1).1). Wild-type and mutant p53s that decreased clonal survival after UV also resulted in greater UV-induced cell death. The full impact of S121F and N288K proteins on UV-induced cell death could not be determined because of the high level of spontaneous cell death. Thus, wt p53 and some of the mutants reduce UV tolerance to various extents. The amount of early (24 h) UV-induced apoptosis/necrosis appears to be a good predictor of clonal survival.

The impact of expressed p53s in responses to γ radiation was quite different. Distinctly for UV, cells expressing wt p53 as well as T123P conveyed resistance, in terms of clonal survival, that was similar to that observed with p53⁺ U2OS cells, while null cells or cells expressing either G279E or G245S were highly sensitive (Fig. ​(Fig.3D;3D; summarized in Table ​Table1).1). Expression of H178Y and G279R resulted in intermediate resistance. While their impact on UV clonal survival differed, cells expressing S121F, T125R, and N288K had similar effects, causing γ sensitivity to increase to nearly that of p53 null mutations.

Unlike results for UV, p53-dependent cell death 24 h after γ irradiation does not appear to be a good predictor of clonal survival. Cell death at 24 h after γ irradiation among mutants was also different from the UV responses (Fig. ​(Fig.3A).3A). While cells expressing the wild type and T123P had high clonal survival after γ irradiation, they also exhibited substantial p53-dependent apoptosis/necrosis at 24 h. Although clonal survival was low, the expression of null mutants G279E and G245S did not result in high levels of cell death at 24 h. In spite of similar abilities of S121F, T125R, and N288K to decrease clonal survival, the T125R had much less of an effect on inducing cell death after γ irradiation.

p53 mutations in the DBD result in at least five patterns of biological responses.

Overall, we found that the individual mutants can have very different effects on clonal survival and apoptosis/necrosis. As summarized in Tables ​Tables11 and ​and2,2, there were at least five distinguishable patterns of biological responses for the eight mutants examined when expressed at biologically relevant levels in human cells. The pattern 1 biological responses correspond to wt p53 and include T123P. The S121F and N288K mutants, comprising pattern 2 responses, have an impact that is stronger than the wild type. In pattern 3, observed with T125R expression, there are varied effects relative to the wild type depending on the end point examined, while in pattern 4 (H178 and G279R) there appear to be modest differences compared to the wild type. Pattern 5 (G279E and G2345S) corresponds to complete deficiency in p53, similar to results with the null p53 cell line. One candidate from each group was examined for its ability to influence the activity of genes in the p53 regulatory network: wt, S121F, T125R, G279R, and G279E. Each of the mutations has been identified in human tumors.

Expressed p53 mutants result in a variety of p53 target gene activities.

Given the differences in biological responses, we addressed whether they are specifically attributable to altered capabilities for transactivation. Three facets of activity of p53 target genes were assessed following the induction of wt p53 and mutant S121F, T125R, G279R, and G279E proteins: transcription, as determined by real time PCR; chromosomal occupancy by p53; and induction of proteins encoded by the target genes. While a direct relationship might be assumed, there are many factors, such as coactivators, stress, and the mutants themselves, that might influence the relationship. Gene activity was determined for the following genes: P21, associated with S-phase cell cycle arrest; MDM2, which prevents accumulation of p53; BAX, associated with apoptosis; and MSH2, the DNA mismatch repair gene associated with damage-induced apoptosis. MSH2 was also chosen because p53 control of the endogenous gene had not previously been described. The wt and p53 mutants were expressed from the stably integrated, dox-regulatable cassettes described above, and gene activity measurements were determined 24 h after induction.

Transactivation patterns differ between the p53 mutants.

The spectra and levels of genes transactivated by each of the mutant p53s are distinct from the wild type, as described in Fig. ​Fig.44 (for endogenous p53 target genes) and Fig. ​Fig.55 (transfected plasmids with p53 target gene REs placed upstream of a luciferase reporter) and summarized in Table ​Table2.2. The S121F, T125R, and G279R mutants clearly retain p53 function based on their abilities to transactivate target genes. The S121F (supertrans in yeast) was greatly reduced for P21 and MDM2 (similar to results of Kaeser and Iggo [25]) but supertrans for BAX and MSH2, whereas T125R was supertrans for MDM2 and deficient for MSH2 and P21. The G279R was only able to transactivate the P21and MDM2 genes (G279R and T125R were change-in-spectrum in yeast with enhanced activity on the MDM2 RE). It is interesting that single amino acid changes G→ R or G → E at codon 279 could change the pattern of transactivation dramatically. Thus, the differences in biological response patterns reflect, in some way, specific differences in spectra and levels of gene induction. The results with the expressed transgene were comparable to those with transfected p53 plus RE::luciferase reporters for P21, MDM2 and MSH2 REs (Fig. ​(Fig.5)5) (23).

Open in a separate window

FIG. 4.

Transactivation patterns resulting from expression of p53 mutants correlate with p53 occupancies at endogenous promoters and protein levels from target genes. SaOS2 cells containing stable Tet-Off-inducible p53s were induced for p53 (dox off) and 6 h later treated or not treated with 10 J/m² UV. The cells were continued in the p53-inducing medium and analyzed at 24 h. Presented are the mean values and SD from three independent experiments. (A) Differential induction of p53 target genes. Transactivation activities were analyzed by real-time PCR. The levels of mRNA expression are presented as induction relative to cells lacking p53. (B) Expression of p53 results in differential promoter occupancies by p53 mutants. ChIP analysis was done 24 h posttreatment (+ or − UV) using a p53-specific antibody (see Materials and Methods). The bar graphs represent the percentages of immunoprecipitated DNA with respect to input samples. For gene expression and ChIP assays, GAPDH was used as negative control. (C) Inducible expression of p53 mutants changes the protein expression spectra of several p53 target genes. Cell lysates were subjected to Western blot analysis. (D) Comparable p53 stabilization and posttranslational modifications in wild-type and p53-mutant proteins after UV. At 24 h post-UV treatment, phosphorylation of p53 at Ser 15 and Ser 392 was analyzed by Western blot using specific monoclonal antibodies. Actin was the internal loading control. IP, immunoprecipitated; WT, wild type.

The transactivation activities of S121F, T125R, and G279E towards additional endogenous target genes involved in apoptosis, DNA repair, cell cycle, and stress pathways were also examined using transfected p53 (see PCR results in Fig. S4A in the supplemental material). The cell cycle and stress genes Cyclin-G, GADD45, MDM2, and P21 were all highly induced (approximately fivefold) by wt p53. The repair genes MSH2 and P48^DDB2 were induced three- to fivefold, but there was little induction of XPC. Induction of the apoptotic genes AIP1, APAF, BAX, PIG3, and PUMA varied from two- to fivefold. Transient transfection by the supertrans S121F resulted in increased transactivation relative to the wild type for the apoptotic genes AIP1, APAF, PIG3, and PUMA, but there was a change in spectrum for the cell cycle/stress genes and the DNA repair genes. The expression profile was very different for the T125R allele which only induced MDM2, BAX, and PIG3 transcription and not the repair genes MSH2, P48^DDB2, or GADD45. The DNA-binding deficient mutant G279E had no impact on expression from any of the genes. The differential effects on the expression of DNA repair genes suggest differences in repair that might be addressable with these mutants.

Transactivation patterns correlate with p53 occupancies at endogenous promoters and induced target protein levels.

While transactivation requires direct interaction of p53 with REs, many factors can contribute to the final levels of transcription. We investigated the relationship between occupancy and extent of transactivation from the endogenous target genes, as well as the consequences of DNA damage on these two features of gene expression. Occupancy by the various wild-type and mutant p53s at the P21, MDM2, BAX, and MSH2 response elements was determined 24 h after removing dox using ChIP. The results are presented in Fig. ​Fig.44 and summarized in Table ​Table22 as target gene activity. There was high occupancy by wild-type p53 at promoters of the strongly expressed P21 and MDM2 genes (0.8%) and low occupancy for the weakly transactivated genes BAX and MSH2 (0.1%). The patterns of occupancy by the mutants were comparable to those for transactivation, suggesting that occupancy is a good predictor of transcription for the regulatable wild-type and mutant p53s. The S121F mutant had markedly increased occupancy for MSH2 (0.3%), somewhat higher than the wild type for BAX, and much lower for P21 and MDM2 promoters. For cells expressing the T125R allele, only the MDM2 promoter had a large increase in occupancy (0.9%). No occupancy was observed at the BAX and MSH2 promoters for the G279R mutant, but there was substantial occupancy at the P21 (0.5%) and MDM2 (0.4%) promoters. There was no significant binding by G279E protein at any of the promoters. Nonspecific precipitation of GAPDH DNA, used as a negative control, was minimal (0.02%) for wild-type p53 and the altered p53 alleles. Thus, differences in transactivation capacity towards the various REs are best explained as being due to differences in the ability of the proteins to bind rather than in the capability for transcription once bound.

As might be expected, the appearance of the P21, MDM2, BAX, and MSH2 proteins was directly related to transcription of the endogenous genes (Fig. ​(Fig.4C)4C) (also found for transfected p53 mutants, as shown in Fig. S4B in the supplemental material). For example, whereas wild-type p53 increased the levels of the four proteins, the levels of MSH2 and BAX were low although somewhat higher than in the p53 null SaOS2 cells. For S121F, the levels of BAX and MSH2 were higher than the wild type but much lower for p21 and MDM2, while the level of MDM2 was higher than the wild type with the T125R mutant. Thus, the different patterns of biological responses relate to changes in levels of proteins induced by the various mutants.

Overview of impact of wt and p53 mutants on biological responses and transactivation.

The impacts of the mutants on biological activities affected by p53 in human cells were examined at biologically relevant levels of protein expression. As summarized in Tables ​Tables11 and ​and2,2, the eight DBD mutants tested had markedly different biological consequences, resulting in five patterns for the six end points examined: growth suppression, spontaneous apoptosis, and UV and γ effects on clonal survival and on cell death. Since these are only a subset of possible biological stress responses, sampling a larger set of mutants and more biological end points should lead to an even greater number of patterns and, therefore, phenotypes. Using a representative from each pattern and a subset of genes in the p53 network, we established that the patterns of biological responses reflect differences in the activity of p53-responsive genes, as summarized for four mutants in Table ​Table2.2. Additional information for more target gene activities comes from transient expression, described in Fig. S4 in the supplemental material. While over 50 genes are targeted by p53, expression changes in the target genes examined provide insights into how the altered spectra of gene activities might relate to the observed patterns of biological responses.

As expected, expression of wt p53 in the p53 null cell line causes growth suppression and apoptosis which are likely related to the induction of P21, BAX (Table ​(Table2),2), and other apoptotic genes (see Fig. S4 in the supplemental material). It also sensitizes cells to UV, probably through increased apoptosis. Interestingly, there is a reduction in survival even though p53 was able to induce genes involved with repair of UV damage (P48^DDB2 and XPC) (also found for other p53 mutants). Unlike the case for UV, p53 provides protection against γ exposure. For both UV and γ radiation the impact of p53 on UV sensitivity or γ resistance is related to its sequence-specific transactivation capabilities (compare the wt versus G279E and null).

While the reduction in UV resistance resulting from expression of mutant as well as wt p53s is related to induction of apoptosis, the role for apoptosis in γ radiation survival responses is not clear based on results with the various p53s. Cells expressing some mutants, including the null, are highly sensitive in terms of clonal survival but exhibit low levels of cell death (apoptosis plus necrosis). On the contrary, cells expressing the wild type and T123P show high levels of clonal survival but also yield high levels of cell death. Furthermore, survival is similar for cells with expressed S121F and T125R, yet S121F causes much more γ radiation-induced apoptosis. Thus, we conclude that whether γ irradiation induced early cell death is not a predictor of sensitivity. Other factors, such as the types of damage induced by UV and γ radiation, the activation of repair pathways, and expression of other genes not examined in this study, may also contribute to the differences in sensitivity resulting from the expression of the p53 change-of-spectrum mutants.

The mutants also provide the opportunity to address the role of specific p53-targeted genes in various responses. Overall, the results with the various mutants demonstrate that change-of-spectrum mutants provide opportunities to separate out a general role for the p53 protein in biological responses (i.e., interactions with other factors) versus its role in transcriptional control of specific genes.

Altered responses due to S121F mutant.

The S121F mutant was identified as supertrans in yeast, but change-of-spectrum in human cells. In SaOS2 cells, it generally had a more dramatic effect on biological end points than wild type p53, including high levels of spontaneous apoptosis as well as low clonal survival following exposure to UV and γ radiation. These responses were consistent with the reduced level of P21 and the up-regulation of many of the endogenous p53 target genes in the apoptosis pathway (Table ​(Table2)2) (see Fig. S4 in the supplemental material) including MSH2 (see below). Our results are consistent with the higher apoptosis reported in several other human cell lines (26, 45) through the combination of reduced MDM2 induction and increased or unbalanced expression of other apoptosis-inducing p53 target genes (1, 45). We note that, unlike results for the various endogenous genes in the present study, Kakudo et al. (26) found that the S121F mutant generally resulted in reduced transactivation among the RE reporters examined in a transient transfection system.

The T125R change-of-spectrum mutant.

A very different pattern of biological responses and spectrum of target gene activity relative to the wt and S121F were observed for the T125R mutant (Tables ​(Tables11 and ​and2).2). The p53-specific biological responses were generally reduced from those in the wild type. The T125R protein caused less spontaneous growth suppression and apoptosis and it was similar to the null in that there was little UV-associated cell death or loss of clonal survival and the cells remained γ sensitive. The increased BAX and PIG3 expression (Table ​(Table2)2) (see Fig. S4 in the supplemental material) argues against these genes having specific roles in spontaneous or DNA damage-induced apoptosis, at least when controlled by T125R.

The possibility that this mutant affects the expression of other genes that antagonize or attenuate apoptosis cannot be excluded. For example, Minsky and Oren (35) showed that enhanced MDM2 protein expression (a feature of T125R) (Table ​(Table2)2) causes silencing of some p53-targeted promoters through MDM2-mediated histone monoubiquitylation and other chromatin modifications.

G279E loss-of-function versus G279R change-of-spectrum mutants.

The two amino acid changes at position 279, glycine to the negatively charged glutamic acid or the positively charged arginine, had markedly different effects on p53-associated end points. The mutation G279E was essentially null, although its presence did cause a slight increase in spontaneous apoptosis. On the other hand, the G279R mutation, which exhibited change-of-spectrum gene activities, retained some p53-associated biological properties including near-wild-type levels of spontaneous apoptosis in spite of the absence of BAX induction and intermediate resistance to γ radiation.

Results with T125R and G279R provide an opportunity to address a possible role for MSH2 in apoptosis and lethality. In addition to being essential in mismatch repair, MSH2 links DNA damage to apoptosis and cell cycle regulation (12, 19, 21, 54, 62). We established that p53 can induce expression of the endogenous MSH2 gene, supporting previous observations with a transfected MSH2 reporter (47). Expression of the wt, S121F, and G279R, but not T125R, resulted in substantial spontaneous apoptosis (Fig. ​(Fig.3A3A and Table ​Table1).1). Yet neither G279R nor T125R induced MSH2. The induction of apoptosis in UV- or γ-irradiated cells containing wt p53 was not found for G279R- or T125R-expressing cells. Possibly MSH2 participation in the apoptotic pathway depends on the presence of DNA damage.

Transactivation and occupancy by p53.

The present experiments also address the relationship between the level of p53 and stress-induced transactivation. p53 was expressed constitutively from a regulatable promoter at levels corresponding to those found in p53⁺/p53⁺ cells that are stressed (see Fig. S1 in the supplemental material) or in tumors that have mutant p53 proteins (14, 25). In our system, the transcriptional responses to the induced p53 at the four target genes examined were directly proportional to the occupancy. UV stress and subsequent modification of p53 did not markedly change the level of occupancy, arguing against a latency model for p53-induced gene activation, in agreement with other reports (25). Increases in nuclear p53 protein, rather than allosteric modifications, are associated with transcriptional activation (4, 33). The additional occupancy following UV may be due to increased synthesis of p53, stabilization, and/or phosphorylation (Fig. ​(Fig.4D4D).

Espinosa et al. (14) demonstrated with endogenously expressed p53 in p53⁺/p53⁺ cells that stresses can differentially alter the expression of p53 target genes. They showed that while there is a requirement for p53, recruitment of additional transcriptional factors and chromatin modifications play important roles in the kinetics of target gene induction following stress. Cofactors such as ASPP1 and even other p53 family members p63/p73 can selectively enhance the p53 apoptotic activity by facilitating its binding to proapoptotic promoters at least in some cell types (15, 46, 50). It should be noted that for unstressed p53⁺/p53⁺cells, p53 is bound to some of its target promoters (3, 13, 14, 53), and the eventual p53 transactivation following stress may not be directly related to occupancy (25, 53). In our case, there was no dramatic increase in transcription resulting from the UV stress other than a change that could be ascribed to the modest increase in p53 protein. For the two weaker promoters (BAX and MSH2) there was no increase in occupancy nor associated transcription. The difference in results for occupancy and transcription between ours and those of Espinosa et al. (14) could rest in the relatively high level of expressed p53 in our system and/or the establishment of transactivation prior to stress. In our system the transcriptional complex may already be in place (i.e., primed) at the time of exposure.

We thank Bert Vogelstein and Moshe Oren for the generous gift of expression vectors, Alice A. Wright, Carl Bortner, and Joyce Snipe for their valuable technical assistance, and Doug Bell, Trevor Archer, and Alex Merrick for critical comments and suggestions on the manuscript.