Identification of a Novel p53 Functional Domain That Is Necessary for Mediating Apoptosis*

The ability of p53 to induce apoptosis requires its sequence-specific DNA binding activity; however, the transactivation-deficient p53(Gln22-Ser23) can still induce apoptosis. Previously, we have shown that the region between residues 23 and 97 in p53 is necessary for such activity. In an effort to more precisely map a domain necessary for apoptosis within the N terminus, we found that deletion of the N-terminal 23 amino acids compromises, but does not abolish, p53 induction of apoptosis. Surprisingly, p53(Δ1–42), which lacks the N-terminal 42 amino acids and the previously defined activation domain, retains the ability to induce apoptosis to an even higher level than wild-type p53. A more extensive deletion, which eliminates the N-terminal 63 amino acids, renders p53 completely inert in mediating apoptosis. In addition, we found that both p53(Δ1–42) and p53(Gln22-Ser23) can activate a subset of cellular p53 targets. Furthermore, we showed that residues 53 and 54 are critical for the apoptotic and transcriptional activities of both p53(Δ1–42) and p53(Gln22-Ser23). Taken together, these data suggest that within residues 43–63 lie an apoptotic domain as well as another transcriptional activation domain. We therefore postulate that the apoptotic activity in p53(Gln22-Ser23) and p53(Δ1–42) is still transcription-dependent.

The ability of p53 to induce apoptosis requires its sequence-specific DNA binding activity; however, the transactivation-deficient p53(Gln 22 -Ser 23 ) can still induce apoptosis. Previously, we have shown that the region between residues 23 and 97 in p53 is necessary for such activity. In an effort to more precisely map a domain necessary for apoptosis within the N terminus, we found that deletion of the N-terminal 23 amino acids compromises, but does not abolish, p53 induction of apoptosis. Surprisingly, p53(⌬1-42), which lacks the N-terminal 42 amino acids and the previously defined activation domain, retains the ability to induce apoptosis to an even higher level than wild-type p53. A more extensive deletion, which eliminates the N-terminal 63 amino acids, renders p53 completely inert in mediating apoptosis. In addition, we found that both p53(⌬1-42) and p53(Gln 22 -Ser 23 ) can activate a subset of cellular p53 targets. Furthermore, we showed that residues 53 and 54 are critical for the apoptotic and transcriptional activities of both p53(⌬1-42) and p53(Gln 22 -Ser 23 ). Taken together, these data suggest that within residues 43-63 lie an apoptotic domain as well as another transcriptional activation domain. We therefore postulate that the apoptotic activity in p53(Gln 22 -Ser 23 ) and p53(⌬1-42) is still transcription-dependent.
The p53 tumor suppressor protein serves as a checkpoint in maintaining genome stability (1)(2)(3). Several different biological responses that could play a role in maintaining genome stability have been strongly correlated with wild-type p53 function (1,3). Following stress conditions such as in the presence of damaged DNA or insufficient growth and survival factors, the cellular levels of p53 increase. This leads to one of at least three well understood cellular responses as follows: cell cycle arrest, differentiation, or apoptosis. Several factors have been shown to determine how a cell responds to the accumulation of p53, e.g. cell type and the presence of several cellular and viral proteins (4 -8). In addition, the levels of p53 in a given cell can dictate the response of the cell such that lower levels of p53 result in cell cycle arrest (9) or differentiation (10), whereas higher levels result in apoptosis (9,10).
The functional domains of p53 have been subjected to extensive analysis (1,3,4). A transcriptional activation domain has been shown to lie within N-terminal residues 1-42 (11,12). Within this region there are a number of acidic and hydrophobic residues, characteristics of the acidic activator family of transcriptional factors (13). Indeed, a double point mutation of the two hydrophobic amino acids at residues 22 and 23 renders p53 transcriptionally inactive (14). These two residues presumably are required for the interaction of the activation domain with the TATA box binding protein and/or TATA box binding protein-associated factors (15)(16)(17)(18).
It is well established that as a transcriptional activator, p53 up-regulates p21, a cyclin-dependent kinase inhibitor (19 -21), which leads to p53-dependent G 1 arrest. However, it is not certain what function(s) of p53 is required for apoptosis. The transactivation function of p53 was shown to be required in some experimental protocols (22)(23)(24). There are several candidate genes that play roles in apoptosis that can be activated in response to p53 induction, such as BAX (25), IGFBP3 (26), PAG608 (27), KILLER/DR5 (28), and several redox-related PIGs genes (29). Several other studies, including our own observations, have provided evidence that p53 might have a transcription-independent function in apoptosis (9, 30 -32). Recently, the proline-rich region between residues 60 and 90, which comprises five "PXXP" motifs (where P represents proline and X any amino acid), was found to be necessary for efficient growth suppression (33) and apoptosis (34) and to serve as a docking site for transactivation-independent growth arrest induced by GAS1 (35).
Previously, we showed that the region between residues 23 and 97 is necessary for apoptosis (9). To more precisely map such a domain in the N terminus necessary for apoptosis, we have made several new mutants. Analyses of these mutants lead to identification of a novel domain between residues 43 and 63 that can mediate apoptosis and activate cellular p53 targets. We also found that a double point mutation at residues 53 and 54 completely abolishes both the transcriptional and apoptotic activities mediated by this novel domain. Thus, we hypothesize that a transcriptional activity located in this novel domain regulates a subset of cellular p53 targets that are responsible for apoptosis.

EXPERIMENTAL PROCEDURES
Plasmids and Mutagenesis-Mutant p53 cDNAs were generated by polymerase chain reaction using the full-length wild-type p53 cDNA as a template. To generate p53(⌬1-23), the pair of primers used were as follows: forward primer N24, GAT CGA ATT CAC CAT GGG CTA CCC ATA CGA TGT TCC AGA TTA CGC TAA ACT ACT TCC TGA A; and reverse primer C393, GAT CGA ATT CTC AGT CTG AGT CAG GCC CTT. To generate p53(⌬1-42), the pair of primers used were as follows: forward primer N43, GAT CGA ATT CAC CAT GGG CTA CCC ATA CGA TGT TCC AGA TTA CGC TTT GAT GCT GTC CCC G; and reverse primer C393. To generate p53(⌬1-63), the pair of primers used were: forward primer N64, GAT CGA ATT CAC CAT GGG CTA CCC ATA CGA TGT TCC AGA TTA CGC TCC CAG AAT GCC AGA GGC T; and reverse primer C393. To generate p53(Gln 22 -Ser 23 /Gln 53 -Ser 54 ), cDNA fragments encoding amino acids 1-59 and 60 -393 were amplified in-dependently and ligated together through an internal AvaII site. The p53(Gln 22 -Ser 23 ) cDNA was used as a template (14). The pair of primers for the cDNA fragment encoding amino acids 1-59 were as follows: forward primer N1, GAT CGA ATT CAC CAT GGG CTA CCC ATA CGA TGT TCC AGA TTA CGC TGA GGA GCC GCA GTC AGA TCC; and reverse primer C59, TTC ATC TGG ACC TGG GTC TTC AGT GCT CTG TTG TTC AAT ATC. The pair of primers for the cDNA fragment encoding amino acids 60 -393 were as follows: forward primer N60, ACT GAA GAC CCA GGT CCA; and reverse primer C393. To generate p53(⌬1-42/Gln 53 -Ser 54 ), the p53(Gln 22 -Ser 23 /Gln 53 -Ser 54 ) cDNA was amplified by forward primer N43 and reverse primer C393. Mutations were confirmed by DNA sequencing.
The above mutant p53 cDNAs were cloned separately into a tetracycline-regulated expression vector, 10-3, at its EcoRI site, and the resulting plasmids were used to generate cell lines that inducibly express p53.
Cell Lines, Transfection, and Selection Procedures-The H1299 cell line was purchased from the American Type Culture Collection and grown with Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37°C with 5% CO 2 . Transfections were performed using the calcium chloride method as described (36). Cell lines expressing inducible proteins of interest were generated as described previously (9). Individual clones were screened for inducible expression of the p53 protein by Western blot analysis using monoclonal antibodies against p53. The H1299 cell lines that inducibly express either wild-type p53 or p53(⌬364 -393) are p53-3 and p53(⌬364 -393)-1, respectively, as described previously (9). The H1299 cell line that inducibly expresses p53(⌬62-91) is p53(⌬62-91)-5. 1 Western Blot Analysis-Cells were collected from plates in phosphate-buffered saline, resuspended with 1ϫ sample buffer, and boiled for 5 min. Western blot analysis was performed as described previously (37). Monoclonal antibodies used to detect p53 were Pab240 and Pab421 (37). The affinity purified monoclonal antibody against p21 (Ab-1) was purchased from Oncogene Science (Uniondale, NY). Affinity purified anti-actin polyclonal antibodies were purchased from Sigma.
Growth Rate Analysis-To determine the rate of cell growth, cells were seeded at 5-10 ϫ 10 4 cells per 60-mm plate, with or without tetracycline (2 g per ml). The medium was replaced every 72 h. At times indicated, two plates were rinsed with phosphate-buffered saline twice to remove dead cells and debris. Live cells on the plates were trypsinized and collected separately. Cells from each plate were counted three times by Coulter cell counter. The average number of cells from at least two plates were used for growth rate determination.
FACS 2 Analysis-Cells were seeded at 2.0 ϫ 10 5 per 90-mm plate with or without tetracycline. Three days after plating, both floating dead cells in the medium and live cells on the plate were collected and fixed with 2 ml of 70% ethanol for at least 30 min. For FACS analysis, the fixed cells were centrifuged and resuspended in 1 ml of phosphatebuffered saline solution containing 50 g/ml each of RNase A (Sigma) and propidium iodide (Sigma). The stained cells were analyzed in a fluorescence-activated cell sorter (FACSCaliber, Becton Dickinson) within 4 h. The percentage of cells in sub-G 1 , G 0 -G 1 , S, and G 2 -M phases was determined using the ModFit program. The percentage of cells in sub-G 1 phase was used as an index for the degree of apoptosis.
Cell Viability Assay by Trypan Blue Exclusion-Cells were seeded at 2 ϫ 10 5 per 90-mm plate with or without tetracycline. Three days after plating, both floating cells in the medium and live cells on the plate were collected and concentrated by centrifugation. After stained with trypan blue (Sigma) for 15 min, both live (unstained) and dead (stained) cells were counted two times in a hemocytometer. The percentage of dead cells from control plates was subtracted from the percentage of dead cells from experimental plates, and the resulting value was used as an index for the degree of apoptosis.
RNA Isolation and Northern Blot Analysis-Total RNA was isolated using Trizol reagent (Life Technologies, Inc.). Northern blot analysis was performed as described (37). The p21 probe was made from an 1.0-kilobase pair EcoRI-EcoRI fragment (19); the MDM-2 probe was made from a 2.1-kilobase pair NotI-SmaI fragment (38); the BAX probe was made from a 290-base pair PstI-BglII fragment (39); the GADD45 probe was from a 400-base pair EcoRI-BamHI fragment (40); the GAPDH probe was made from an 1.25-kilobase pair PstI-PstI cDNA fragment (41); and the MCG14 cDNA probe was a 200-base pair polymerase chain reaction fragment identified by CLONTECH PCR-Select cDNA subtraction. 2

A Novel Domain within Residues 43-63 Is Necessary for
Mediating Apoptosis-Previously, we showed that p53(⌬1-22), which lacks the N-terminal 22 amino acids, can still induce apoptosis as well as cell cycle arrest (9). Since both residues 22 and 23 are critical for p53 transcriptional activity (14), we decided to determine whether p53(⌬1-23), which deletes the N-terminal 23 amino acids, would also be able to induce apoptosis and activate cellular p53 targets.
We have previously established a cell line that expresses high levels of wild-type p53 called p53-3 (9). This line was established using a tetracycline-regulated expression system as described previously (42). By using similar techniques, we established nine stable cell lines that express p53(⌬1-23). Three representative cell lines, p53(⌬1-23)-9, -10, and -23, are shown in Fig. 1A. Western blot analysis showed that these cell lines express p53(⌬1-23) at levels comparable to wild-type p53 in p53-3 cells (Fig. 1A). To characterize p53(⌬1-23), we looked at its transcriptional and apoptotic activities and the growth rate of the cell line p53(⌬1-23)-9. The transcriptional activity was determined by monitoring the expression of the endogenous gene, p21, a well defined transcriptional target of p53 (19). We found that p53(⌬1-23) is still capable of activating p21, albeit to a much less degree than wild-type p53 (Fig. 1A). Next, the growth rates of p53(⌬1-23)-9 cells under both uninduced and induced conditions were determined, and these cells failed to multiply following p53 expression (Fig. 1B). To exclude potential effects of the regulator tetracycline and/or the tet-vp16 transactivator (42) on cell growth, we analyzed the growth rate of the cell line H24-1, which was similarly established but did not express any protein. The results showed that the growth rates of H24-1 cells under both the uninduced (ϩtet) and induced (Ϫtet) conditions were nearly identical (Fig. 1C), indicating that both tetracycline and tet-vp16 transactivator have no effect on cell growth. It is well established that the percentage of cells containing a sub-G 1 DNA content reflects the extent to which cells are undergoing apoptosis (9,23,31). Since p53 can induce apoptosis in H1299 cells (9,31), FACS analysis was used to observe the extent of apoptosis by determining the distribution of cells in each phase of the cell cycle. The results showed that 18% of cells expressing p53(⌬1-23) had a sub-G 1 DNA content 3 days after induction of this mutant, compared with less than 5% of the same cells expressing no p53 (Fig. 1, D and E; Table I). Trypan blue exclusion assay showed that 15% of cells were dead, which is consistent with FACS analysis. In contrast, about 45 and 30% of cells had a sub-G 1 DNA content at day 3 following expression of either wild-type p53 or transactivation-deficient p53(Gln 22 -Ser 23 ), respectively ( Table I). The FACS results also showed that the number of cells in S phase was decreased from 38 to 22.3% following induction of p53(⌬1-23), and these cells primarily arrested in G 1 (Fig. 1, D and E). Similar results were obtained using another high p53(⌬1-23) producer, p53(⌬1-23)-10.
Since p53(⌬1-23) is still capable of inducing apoptosis and p53 activation domain lies within residues 1-42 (11,12), we determined whether the other half (residues 24 -42) of the previously defined activation domain is required for apoptosis. To this end, we established 16 individual stable cell lines that inducibly express p53(⌬1-42) that lacks the N-terminal 42 amino acids. Three representative cell lines, p53(⌬1-42)-2, -5, and -11, were shown in Fig. 2A. Consistent with previous results that p53(Gln 22 -Ser 23 ) cannot activate p21 (9,22,33,43), p53(⌬1-42) only minimally activated p21 as compared with wild-type p53 (Fig. 2A). We then determined the growth rate of a high producer, p53(⌬1-42)-2. Surprisingly, we found that a majority of cells died within 3 days following induction of p53(⌬1-42) (Fig. 2B). In addition, both trypan blue exclusion assay and FACS analysis showed that approximately 50 -68% of cells underwent apoptosis ( Fig. 2C; Table I). Similar results were obtained from several other cell lines. These results suggest that the entire previously defined activation domain within the N-terminal 42 amino acids is dispensable for apoptosis. In fact, deletion of this region enhanced the ability of p53 to induce apoptosis (Table I).

Within Residues 43-63 Lies Another Activation Domain That Overlaps with the Domain Necessary for Mediating Apoptosis-
The ability of transactivation-deficient p53(Gln 22 -Ser 23 ) to induce apoptosis leads to the hypothesis that p53 has transcription-independent apoptotic activity (9, 31, 33). Since p53(⌬1-42) lacks the previously defined activation domain and only Ϫ p53(⌬1-63) Ϫ a Arrest was assayed by the relative growth rate of cells and the number of cells in S phase.
b Apoptosis was assayed at day 3 by the percentage of cells staining with trypan blue and by determination of the sub-G 1 phase cells using the Modfit program. FIG. 1. The N-terminal 23 amino acids are dispensable for apoptosis. A, levels of p53, p21, and actin in p53-3, and p53(⌬1-23)-9, -10, and -23 cell lines were assayed by Western blot analysis. Cell extracts were prepared from uninduced cells (Ϫ) or cells induced to express (ϩ) wild-type p53 or p53(⌬1-23). The upper portion of the blot was probed with a mixture of p53 monoclonal antibodies Pab421 and Pab240 and actin polyclonal antibody. Mutant p53(⌬1-23) migrates faster than wild-type p53 because it is missing 23 amino acids. The lower portion of the blot was probed with p21 monoclonal antibody. B, growth rates of p53(⌬1-23)-9 cells in the presence (छ) or absence (Ⅺ) of p53 were measured as described under "Experimental Procedures." C, growth rates of H24--cells in the presence ([Ⅺ) or absence (‚) of tetracycline. D, DNA contents were quantitated by propidium iodide staining of fixed cells at day 3 following withdrawal of tetracycline as described under "Experimental Procedures." E, The percentages of p53(⌬1-23)-9 cells in sub-G 1 , G 0 -G 1 , S, and G 2 -M phases in the presence or absence of p53 for 3 days were quantitated using ModFit program as described under "Experimental Procedures." p53-dependent Apoptosis minimally activates p21 as determined by Western blot analysis ( Fig. 2A), it appears that it can induce apoptosis in a transcription-independent manner. To ascertain whether p53(⌬1-42) contains a transcriptional activity, the expression patterns of four well defined cellular p53 targets, p21, MDM2, GADD45, and BAX, were analyzed in cells expressing p53(⌬1-42) by Northern blot analysis (Fig. 4A). The expression levels of these genes in cells with or without p53 were quantitated by PhosphorImage scanner, and the fold increase of their relative mRNAs was calculated after normalization to glyceraldehyde-3-phosphate dehydrogenase mRNA levels (Table II). The results showed clearly that p53(⌬1-42) significantly activated MDM2 (8-fold), GADD45 (7.03-fold), and BAX (3.9-fold) but only minimally activated p21 (1.83-fold). As expected, wild-type p53 but not mutant p53(Gln 22 -Ser 23 ) activated these cellular p53 targets (Fig. 4A; Table II). As a control, p53(⌬64 -91), which lacks all of the five PXXP motifs, was examined. The proline-rich domain in p53 is dispensable for transactivation (33,34). As expected, p53(⌬64 -91) activated these p53 targets ( Fig. 4A and Table II). Since p53(⌬1-63) failed to activate any of these p53-regulated genes (data not shown), the results suggest that another activation domain lies within residues 43-63. For clarity, we designate the originally defined activation domain located within residues 1-42 as activation domain I and this novel domain as activation domain II.
Since a double point mutation at residues 22 and 23 abolishes the transcriptional activity of the activation domain I (14), we looked for analogous hydrophobic amino acids within the activation domain II. Two were found: tryptophan at residue 53 and phenylalanine at residue 54. We therefore made identical mutations in these two amino acids in p53(Gln 22 -Ser 23 ) or p53(⌬1-42), changing tryptophan 53 to glutamine and phenylalanine 54 to serine to generate p53(Gln 22 -Ser 23 / Gln 53 -Ser 54 ) and p53(⌬1-42/Gln 53 -Ser 54 ). We then established a number of cell lines that inducibly express these mutants, and their ability to induce apoptosis and activate cellular p53 targets were similarly analyzed as above. Three representative cell lines that express either p53(Gln 22 -Ser 23 /Gln 53 -Ser 54 ) or p53(⌬1-42/Gln 53 -Ser 54 ) are shown in Fig. 5, A and C, respectively. As expected, Western blot analysis showed that p21 was not activated by either of these mutants (Fig. 5, A and C,  bottom panel). In addition, these mutants were unable to induce apoptosis, as demonstrated by the rate of cell growth (Fig.  5, B and D), trypan blue exclusion assay, and FACS analysis (Table I). Furthermore, the putative cellular p53 target MCG14, which can be activated by p53(⌬1-42) and p53(Gln 22 -Ser 23 ) (Fig. 4B), failed to be activated in cells expressing either p53(Gln 22 -Ser 23 /Gln 53 -Ser 54 ) or p53(⌬1-42/Gln 53 -Ser 54 ) (Fig.  4C). These results indicate that residues 53 and 54 are critical for the novel domain within residues 43-63 to induce apoptosis and activate cellular p53 targets.
Here we found that within residues 43-63 lies another novel domain that is necessary for apoptosis on the basis of the following observations: (i) p53(⌬1-42), which lacks the N-terminal 42 amino acids and the previously defined activation domain, contains a strong apoptotic activity; (ii) p53(⌬1-63), which lacks the N-terminal 63 amino acids but contains intact PXXP motifs, has no apoptotic activity; (iii) a double point mutation at residues 53 and 54 renders both p53(⌬1-42/Gln 53 -Ser 54 ) and p53(Gln 22 -Ser 23 /Gln 53 -Ser 54 ) completely inert in inducing apoptosis; and (iv) codon 53 is one of the frequently mutated sites outside the DNA binding domain in the p53 gene in human tumors (45), which underscores the importance of the apoptotic function within residues 43-63 in p53 tumor suppression.
How does this novel domain mediate an apoptotic activity? Previously, it was shown that p53(Gln 22 -Ser 23 ), which cannot activate several cellular p53 targets (9,14,22,31,43), is still capable of inducing apoptosis (9,31,34), and a p53 mutant, which lacks the proline-rich region, is capable of activating several p53 targets (33, 34) but cannot induce apoptosis (33,34). These results lead to a hypothesis that p53 has both transcription-dependent and -independent functions in apoptosis. However, it is well established that p53 mutants that are defective in sequence-specific DNA binding activity are also inert in inducing apoptosis (1,3,4), suggesting that p53 sequence-specific DNA binding activity and possibly its sequencespecific transcriptional activity are required for inducing apoptosis. Here we found that p53(⌬1-42), which lacks the entire previously defined activation domain I, not only induces apoptosis, but also activates the MDM2, BAX, and GADD45 genes through its activation domain II located between residues 43 and 63 ( Fig. 4; Table II). Since p53(Gln 22 -Ser 23 ) contains an

FIG. 4. Within residues 43-63 lies another activation domain.
A, Northern blots were prepared using 10 g of total RNA isolated from uninduced cells (Ϫ) or cells induced to express (ϩ) wild-type p53, p53(⌬1-42), p53(⌬62-91), or p53(Gln 22 -Ser 23 ). The blots were probed with p21, MDM2, GADD45, BAX, and GAPDH cDNAs, respectively. B, a Northern blot was prepared using 10 g of total RNA isolated from uninduced cells (Ϫ) or cells induced to express (ϩ) wild-type p53, p53(⌬1-42), p53(⌬62-91), p53(⌬364 -393), or p53(Gln 22 -Ser 23 ). The blot was probed with MCG14 cDNA. C, a Northern blot was prepared using 10 g of total RNA isolated from uninduced cells (Ϫ) or cells induced to express (ϩ) wild-type p53, p53(Gln 22 -Ser 23 /Gln 53 -Ser 54 ), or p53(⌬1-42/ Gln 53 -Ser 54 ). The blot was probed with MCG14 cDNA.  (46) recently showed that within residues 40 -83 lies a sub-activation domain, which can activate a reporter gene under control of a promoter with a p53-responsive element when p53 is cotransfected, and a double point mutation at residues 53 and 54 also abolished the transcriptional activity of the sub-activation domain. These results suggest that p53 has two independent activation domains. A second activation domain within a transcription factor is not without precedent. Herpes simplex virus protein VP16 also contains two independent activation domains (47). Thus, it appears that in response to various stress conditions and their subsequent modifications, the two independent activation domains might serve as an intrinsic factor of p53 that determines whether a given p53 target is activated. Although BAX, MDM2, and GADD45 are the activation domain II-regulated gene products, these cellular p53 targets might not mediate the p53-dependent apoptosis on the basis of two observations: (i) these genes were not activated by p53(Gln 22 -Ser 23 ) which is competent in inducing apoptosis ( Fig. 4A; Table II); (ii) these genes were activated by p53(⌬62-91) which is defective in inducing apoptosis ( Fig. 4A; Table II). Since cell type has been shown to influence the cellular response (cell cycle arrest or apoptosis) to p53 (1,4,8), cellular genetic background might then determine the modification of the two activation domains. Therefore, the results obtained in H1299 cells need to be confirmed in other cell types. It is intriguing that although p53(Gln 22 -Ser 23 ) contains an intact activation domain II, it fails to activate BAX, GADD45, and MDM2 (Fig. 4A). Since both p53(⌬1-23) and p53(⌬1-42) can activate these p53 targets, it suggests that the presence of the first 23 amino acids may mask the ability of the activation domain II in p53(Gln 22 -Ser 23 ) to activate these cellular p53 targets. Alternatively, it is also possible that when the activa-tion domain I is inactivated by a double point mutation at residues 22 and 23, the N-terminal 42 residues might then inhibit or block interaction of a co-activator (or an adaptor) with the activation domain II that is required for activation of some p53 targets, such as MDM2, p21, BAX, and GADD45, but not for activation of other p53 targets, such as MCG14. It is important to note that although the activation domain I is primarily responsible for activation of p21, the level of p21 in cells expressing either p53(⌬1-23) and p53(⌬1-42) was slightly increased upon p53 induction (Figs. 1A and 2A), suggesting that the activation domain II can weakly activate p21. Furthermore, our preliminary studies showed that activation of p21 was compromised by a double point mutation at residues 53 and 54 when p53(Gln 53 -Ser 54 ) was expressed at a low to intermediate level, 3 consistent with the idea that the activation domain II contributes to the activation of p21. Since several clones that express various expression levels of the target genes are required for determining the function of the targets (48), these results remain to be confirmed.
Previously, it was shown that overexpression of p21 can protect human colorectal carcinoma RKO cells from prostaglandin A2-mediated apoptosis (49). Lack of p21 expression due to homologous deletion of the p21 gene also renders HCT116 colorectal cancer cells susceptible to apoptosis following treatment with either ␥-radiation or chemotherapeutic agents (50). In addition, a significant fraction of tumors in mice deriving from p21 Ϫ/Ϫ HCT116 cancer cells were completely cured, and all tumors deriving from p21 ϩ/ϩ cancer cells underwent regrowth after treatment with ␥-radiation (50). It is interesting to note that p53(Gln 22 -Ser 23 ) and p53(⌬1-42), both of which lack a functional activation domain I, cannot significantly activate p21 (Fig. 2 and 4; Table II) but can induce apoptosis ( Table I). The strong apoptotic activity conferred by p53(⌬1-42) might be due to its failure of activating p21. Thus, we have generated a mutant, p53(⌬1-42), that might be better than wild-type p53 in the elimination of cancer cells and therefore a potential candidate for gene therapy.