Definition of the p53 functional domains necessary for inducing apoptosis.

The p53 protein contains several functional domains necessary for inducing cell cycle arrest and apoptosis. The C-terminal basic domain within residues 364-393 and the proline-rich domain within residues 64-91 are required for apoptotic activity. In addition, activation domain 2 within residues 43-63 is necessary for apoptotic activity when the N-terminal activation domain 1 within residues 1-42 is deleted (DeltaAD1) or mutated (AD1(-)). Here we have discovered that an activation domain 2 mutation at residues 53-54 (AD2(-)) abrogates the apoptotic activity but has no significant effect on cell cycle arrest. We have also found that p53-(DeltaAD2), which lacks activation domain 2, is inert in inducing apoptosis. p53-(AD2(-)DeltaBD), which is defective in activation domain 2 and lacks the C-terminal basic domain, p53-(DeltaAD2DeltaBD), which lacks both activation domain 2 and the C-terminal basic domain, and p53-(DeltaPRDDeltaBD), which lacks both the proline-rich domain and the C-terminal basic domain, are also inert in inducing apoptosis. All four mutants are still capable of inducing cell cycle arrest, albeit to a lesser extent than wild-type p53. Interestingly, we have found that deletion of the N-terminal activation domain 1 alleviates the requirement of the C-terminal basic domain for apoptotic activity. Thus, we have generated a small but potent p53-(DeltaAD1DeltaBD) molecule. Furthermore, we have determined that at least two of the three domains (activation domain 1, activation domain 2, and the proline-rich domain), are required for inducing cell cycle arrest. Taken together, our results suggest that activation domain 2 and the proline-rich domain form an activation domain for inducing pro-apoptotic genes or inhibiting anti-apoptotic genes. The C-terminal basic domain is required for maintaining this activation domain competent for transactivation or transrepression.

p53 is frequently mutated in cancers. Mutations in the p53 DNA binding domain or certain mutations in the nuclear localization signal and tetramerization domain that indirectly affect DNA binding abrogate or diminish p53 activity in cell cycle arrest and apoptosis (1,5). The proline-rich domain has been shown to be required for efficient growth suppression (12). Recent experiments indicate that the proline-rich domain is necessary for apoptosis but not cell cycle arrest (16 -18). In addition, the proline-rich domain plays an important role in the induction of several endogenous target genes, but is not required for activation of the exogenously introduced promoters of these target genes (17). These results suggest that the proline-rich domain may participate in the induction of cellular target gene(s) responsible for mediating apoptosis. However, the role of other p53 functional domains (especially the Nterminal activation domain 1 and the C-terminal basic domain) in apoptosis is still not certain. Earlier reports have shown that in some experimental protocols (19 -21) including our own (22), p53 transactivation activity is dispensable for apoptosis. It should be noted that this conclusion is based at least in part on the observation that an activation domain 1-deficient mutant (a double point mutation at residues 22-23, AD1 Ϫ ) 1 is capable of inducing apoptosis (21,22). Recently, we and others have shown that p53-(AD1 Ϫ ) contains an intact activation domain 2 (9 -11), and therefore, p53-(AD1 Ϫ ) is still competent in transactivation (10). Furthermore, when both activation domain 1 and activation domain 2 are mutated (a quadruple point mutation at residues 22-23 and 53-54, AD1 Ϫ AD2 Ϫ ), the resulting protein is inert in transactivation and in inducing cell cycle arrest and apoptosis (9 -11).
The C-terminal basic domain has been subjected to extensive analysis, and all evidence suggests that the basic domain is a regulatory domain. This basic domain can regulate the DNA binding activity when it is phosphorylated (1,5), acetylated (23)(24)(25), deleted (26), or associated with anti-p53 antibody (26,27) or peptides derived from the C terminus of p53 (28,29). Interestingly, the mechanism by which these latter peptides enhance p53 DNA binding activity is the ability of the peptides to interact with three separate domains in p53, that is, the proline-rich domain (30), the DNA binding domain (31), and the C-terminal basic domain (30,31). The C-terminal basic domain also interacts with several cellular proteins, such as TFIIH subunits XPB and XPD (32,33), and Werner syndrome protein (WRN) (34,35), which all lead to efficient induction of p53-mediated apoptosis. These results support a hypothesis that the C-terminal basic domain is a negative regulatory do-main whose effect on the DNA binding activity can be alleviated by interacting with other cellular proteins, peptides derived from the p53 C terminus, or other modifications. However, several groups have shown that p53-(⌬BD), which lacks the C-terminal basic domain, has a reduced ability to induce several cellular target genes and becomes incapable of inducing apoptosis (22,32,36). These results suggest that the C-terminal basic domain can regulate p53 activity both positively and negatively.
In this study, we show that activation domain 2 and the proline-rich domain form an activation domain for inducing pro-apoptotic genes or inhibiting anti-apoptotic genes. The Cterminal basic domain is required for maintaining this activation domain competent for transactivation or transrepression. We also found that an activation domain capable of inducing at least partial cell cycle arrest can be formed by activation domain 1 plus activation domain 2, activation domain 1 plus the proline-rich domain, or activation domain 2 plus the prolinerich domain. The ability of these activation domains to induce cell cycle arrest can be enhanced by the presence of the Cterminal basic domain.
The above mutant p53 cDNAs were cloned separately into a tetracycline-regulated expression vector, pUHD10-3, at its EcoRI site (37), and the resulting plasmids were used to generate cell lines that inducibly express p53.
Growth Rate Analysis, Trypan Blue Dye Exclusion Assay, DNA Histogram Analysis, and Annexin V Staining-Growth rate analysis, trypan blue dye exclusion assay, and DNA histogram analysis were performed as described previously (10,17,22). Propidium iodide and RNase A were purchased from Sigma. Fluorescein isothiocyanate-labeled annexin V was purchased from Roche Molecular Biochemicals, and staining was performed as described by the manufacturer.
RNA Isolation and Northern Blot Analysis-Total RNA was isolated using Trizol reagents (Life Technologies, Inc.). Northern blot analysis was performed as described (10). The p21, BAX, and glyceraldehyde-3phosphate dehydrogenase probes were prepared as described previously (10).

RESULTS
The Activity of Activation Domain 2 Is Necessary for Inducing Apoptosis-Previously, we have shown that the activity of activation domain 2 is required for inducing apoptosis when a double point mutation at residues 22-23 or deletion of the N-terminal 42 amino acid residues renders activation domain 1 dysfunctional (10). To further determine the function of activation domain 2 in apoptosis, we generated an activation domain 2-deficient mutant, p53-(AD2 Ϫ ), which contains a double point mutation at residues 53-54. We then established several cell lines that inducibly express this mutant in p53-null H1299 lung carcinoma cells. Western blots from two representative cell lines, p53-(AD2 Ϫ )-6 and -8, are shown in Fig. 1A. After normalization to the levels of actin protein expressed, we found that the levels of p53 protein in p53-(AD2 Ϫ )-6 and -8 cells were comparable with that in p53-3 and HA⅐p53-15 cells, which express wild-type p53 and HA-tagged wild-type p53, respectively (Fig. 1A, upper two panels, compare lanes 5-8 with lanes [1][2][3][4]. To determine the transcriptional activity of p53-(AD2 Ϫ ), we measured the level of p21 protein induced by p53-(AD2 Ϫ ). Surprisingly, we found that the ability of p53-(AD2 Ϫ ) to induce p21 was severely diminished (Fig. 1A, p21 panel, lanes [5][6][7][8]. These results are similar to that observed for the activation domain 1-deficient mutant (6,10,22). In contrast, p21 was strongly induced by wild-type p53 and HA-tagged wild-type p53 (Fig. 1A, p21 panel, lanes 1-4).
To determine the activity of p53-(AD2 Ϫ ) in H1299 cells, the growth rate of p53-(AD2 Ϫ )-6 cells was determined over a 5-day period. When induced to express p53-(AD2 Ϫ ), cells failed to multiply (Fig. 1B), but visible microscopic cell death was not significantly increased (data not shown).
To determine whether the growth suppression by p53-(AD2Ϫ) is because of cell cycle arrest, apoptosis, and/or both, we performed a DNA histogram analysis and an annexin V staining assay. When induced to express the mutant p53-(AD2 Ϫ ) for three days, we found that the percentage of cells in S phase decreased from 35 to 8% whereas cells in G 1 increased from 49 to 75%, suggesting that p53-(AD2 Ϫ ) arrested cells primarily in G 1 (Fig. 1, D-E). However, no apparent apoptosis was detected by either DNA histogram analysis (Fig. 1, D-E) or annexin V staining (Fig. 1, F-G). Thus, the activity in activation domain 2 is necessary for inducing apoptosis. As a positive control, we analyzed p53-3 and HA⅐p53-15 cells. When induced to express wild-type or HA-tagged p53 for three days, we found that both p53-producing cells were arrested primarily in G 1 and underwent apoptosis, consistent with previous reports (10,22). We also analyzed p53-(AD2 Ϫ ⌬BD)-9 cells. We found that no significant apoptosis was observed, and cells primarily arrested in G 1 when induced to express p53-(AD2 Ϫ ⌬BD) (data not shown).
To determine the activity of the entire activation domain 2 (residues 43-62), we generated p53-(⌬AD2), which lacks the entire activation domain 2 and p53-(⌬AD2⌬BD), which in turn lacks activation domain 2 and the C-terminal basic domain. We then established several cell lines that inducibly express p53-(⌬AD2) and p53-(⌬AD2⌬BD), respectively (Fig. 2, A and C). We found that p53-(⌬AD2) and p53-(⌬AD2⌬BD) suppressed cell proliferation (Fig. 2, B and D), albeit to a lesser extent than p53-(AD2 Ϫ ) and p53-(AD2 Ϫ ⌬BD) (Fig. 1, B and C). Furthermore, we found that cells were arrested primarily in G 1 but did not undergo apoptosis when induced to express these p53 mutants (data not shown, Table I). However, we found that p21 was not significantly induced (Fig. 2, A and C), suggesting that p53-dependent cell cycle arrest in G 1 can be mediated by a gene(s) other than p21.
The Proline-rich Domain Contributes to the Ability of p53 to Induce Cell Cycle Arrest-Previously, we and others have shown that the proline-rich domain (16 -18) and the C-terminal basic domain (22,32) are necessary for inducing apoptosis but not cell cycle arrest. To determine whether both domains are dispensable for inducing cell cycle arrest, we generated p53-(⌬PRD⌬BD), which lacks both the proline-rich domain and the C-terminal basic domain. We then established several cell lines that inducibly express this mutant. Western blots from three representative cell lines, p53-(⌬PRD⌬BD)-2, -6, and -7, are shown in Fig. 2E. We found that the level of p53 expressed in p53-(⌬PRD⌬BD)-2 cells was comparable with that in p53-3, HA⅐p53-15, and p53-(⌬BD)-1 but slightly lower than that in p53-(⌬PRD)-5, which inducibly expresses a p53 mutant lacking FIG. 1. The activity of activation domain 2 is necessary for inducing apoptosis. A, levels of p53, p21, and actin were assayed by Western blot analysis in cell lines as shown above the blots. Cell extracts were prepared from non-induced cells (Ϫ) and cells induced to express p53 for 24 h (ϩ). HA-tagged p53 was detected with 12CA5 antibody. p53 was detected with anti-p53 monoclonal antibody Pab240. p21 was detected with anti-p21 monoclonal antibody (Ab-1). Actin was detected with anti-actin polyclonal antibody. B and C, growth rates of p53-(AD2 Ϫ )-6 and p53-(AD2 Ϫ ⌬BD)-9 cells in the absence (छ) or presence (Ⅺ) of p53 over a 5-day period. D and E, DNA content was quantified by propidium iodide staining of fixed cells that were non-induced (Ϫ p53) or induced (ϩ p53) to express p53-(AD2 Ϫ ) for 3 days. F and G, apoptotic cells were quantified by propidium iodide-annexin V staining of cells that were non-induced (Ϫ p53) or induced (ϩ p53) to express p53-(AD2 Ϫ ) for 3 days.
p53 Functional Domains for Apoptosis the proline-rich domain (Fig. 2E, p53 panel). To determine whether p21 can be induced, we found that p53-(⌬PRD⌬BD) was much less potent in inducing p21 than wild-type p53, HA-tagged p53, p53-(⌬BD), or p53-(⌬PRD) (Fig. 2E, p21 panel). However, when the DNA binding activity was determined in vitro, we found that p53-(⌬PRD⌬BD) was as potent as wildtype p53 in binding to the ribosomal gene cluster p53 response element (data not shown). This suggests that deletion of both the proline-rich domain and the C-terminal basic domain does not affect the activity of the p53 DNA binding domain. Growth rate analysis showed that p53-(⌬PRD⌬BD) had a much reduced ability to suppress cell proliferation (Fig. 2F). In addition, DNA histogram analysis and annexin V staining assay showed that a partial arrest in G 1 , but no apoptosis, was detected in p53-(⌬PRD⌬BD)-2 cells (data not shown).
p53-(⌬AD1⌬BD) Is Small but Potent in Inducing Cell Cycle Arrest and Apoptosis-We and others have shown that p53-(⌬BD), which lacks the C-terminal basic domain, is inactive in inducing apoptosis (22,32,36) whereas p53-(⌬AD1), which lacks activation domain 1 (residues 1-42), is very active (10). To determine whether the C-terminal basic domain is necessary for p53-(⌬AD1) to induce apoptosis, we generated p53-(⌬AD1⌬BD), which lacks activation domain 1 and the C-terminal basic domain. We then established several cell lines that inducibly express p53-(⌬AD1⌬BD). Western blots from three representative cell lines, p53-(⌬AD1⌬BD)-3, -6, and -7, are shown in Fig. 3A. We found that the level of p53 expressed in these cells was comparable with that in p53-3, HA⅐p53-15, and p53-(⌬BD)-1 cells, but lower than that in p53-(⌬AD1)-2 cells (Fig. 3A, p53 panel). p53-(⌬AD1)-2 cells are derived from H1299 cells that inducibly express p53-(⌬AD1), which lacks activation domain 1 (10). We found that p53-(⌬AD1⌬BD) retained the ability to induce p21. Induction of p21 by p53-(⌬AD1⌬BD) was greater than induction by p53-(⌬AD1) but less than induction by wild-type p53 and p53-(⌬BD) (Fig. 3A, p21  panel). Growth rate analysis showed that cells failed to multiply when induced to express p53-(⌬AD1⌬BD) (Fig. 3, B and C). Microscopic examination showed that the p53-expressing cells detached from plates and shrank to form apoptotic bodies (data not shown). DNA histogram analysis showed that the percentage of cells in S phase decreased from 35 to 11% but the percentage of cells in G 1 increased from 55 to 75%, suggesting that these cells arrested primarily in G 1 (Fig. 3, D-E). We also found that the number of cells with a sub-G 1 DNA content was not significantly increased. However, when stained for annexin V, we found that the percentage of stained cells increased from 7 to 31%, suggesting that these cells also underwent apoptosis (Fig. 3, F-G).
Growth rate analysis showed that cells failed to multiply when induced to express wild-type p53 or p53-(⌬AD1⌬BD) (Fig. 4, B-C). Microscopic examination showed that the p53-expressing cells detached from plates and shrank to form apoptotic bodies (data not shown). DNA histogram analysis showed that the percentage of cells that had a sub-G 1 DNA content was increased from 3 to 37% by wild-type p53 (Fig. 4, D-E) and from 4 to 49% by p53-(⌬AD1⌬BD) (Fig. 4, H-I). In addition, annexin V staining assay showed that the percentage of the annexin V-stained cells was increased from 7 to 28% by wild-type p53 and from 9 to 29% by p53-(⌬AD1⌬BD). These data indicate that p53-(⌬AD1⌬BD) is a potent apoptotic inducer. At Least Two of the Three Domains, i.e. Activation Domain 1, Activation Domain 2, and the Proline-rich Domain Are Required for Inducing Cell Cycle Arrest-To further define the role of activation domain 1, activation domain 2, the prolinerich domain, and the C-terminal basic domain in inducing cell cycle arrest and apoptosis, we generated six p53 mutants that are dysfunctional in two or three of the four functional domains (Fig. 5). We then established several cell lines that inducibly express these p53 mutants individually (Fig. 5). These are p53-(⌬AD1AD2 Ϫ ⌬BD),  5E. The level of p53 expressed in some of these mutant p53producing cells was comparable with or higher than that in p53-3 cells (Fig. 5, A-E, p53 panel). However, none of these mutants were capable of inducing p21 (Fig. 5, A-E, p21 panel). In addition, cell cycle arrest and apoptosis were not detected by growth rate and DNA histogram analyses and annexin V staining assay (data not shown). These data suggest that at least two of the three domains (activation domain 1, activation domain 2, and the proline-rich domain) are required for p53 activity.
Regulation of p21 and BAX by p53 Mutants-To determine the ability of various p53 mutants that lack activation domain 1, activation domain 2, and/or the C-terminal basic domain in inducing p21 and BAX, we performed a Northern blot analysis (Fig. 6). We found that wild-type p53 was very active (lanes 1-2). p53-(R249S), a tumor-derived mutant that is defective in the DNA binding domain, was nearly inert (lanes 3-4). Although deletion of the C-terminal basic domain renders p53 constitutively active in binding to DNA in vitro (26), the ability of p53-(⌬BD) to induce p21 and BAX was approximately 2-fold less efficient than that of wild-type p53 (compare lanes 1-2 and   FIG. 3. The C-terminal basic domain is not necessary for apoptosis when activation domain 1 is absent. A, levels of p53, p21, and actin were assayed by Western blot analysis in cell lines as shown in the absence (Ϫ) or presence (ϩ) of p53 for 24 h. Antibodies used were as described in the legend to Fig. 1. B and C, growth rates of p53-(⌬AD1⌬BD)-6 and p53-(⌬AD1⌬BD)-7 cells in the absence (Ⅺ) or presence (छ) of p53 over a 5-day period. D and E, DNA content was quantified by propidium iodide staining of fixed cells that were non-induced (Ϫ p53) or induced (ϩ p53) to express p53-(⌬AD1⌬BD) for 3 days. F and G, apoptotic cells were quantified by propidium iodideannexin V staining of cells that were noninduced (Ϫ p53) or induced (ϩ p53) to express p53-(⌬AD1⌬BD) for 3 days.
p53 Functional Domains for Apoptosis [7][8]. p53-(AD1 Ϫ ) (lanes 5-6), p53-(⌬AD1) (lanes 9 -10), p53-(AD2 Ϫ ) (lanes [13][14], and p53-(AD2Ϫ⌬BD) (lanes [15][16] were extremely weak in inducing p21 and BAX (2-fold or less). It should be mentioned that p53-(AD2 Ϫ ) is extremely potent in inducing G 1 arrest (see Fig. 1, D-E), suggesting that a gene(s) other than p21 is responsible for this. Furthermore, when activation domain 1 and the basic domain were deleted, the ability of p53-(⌬AD1⌬BD) to induce p21 and BAX was partially restored (lanes [11][12], consistent with the result detected by Western blot analysis (Fig. 3A). DISCUSSION p53 induces apoptosis but the underlying mechanism remains unclear. To determine this mechanism, two major questions need to be addressed. What domains in p53 are required and is p53 transcriptional activity necessary for inducing apoptosis? Previous attempts to answer these questions have been inconclusive, because different experimental systems have been used (1,2). These include various types of cell lines and methods to express p53 (transient versus stable, ectopic versus inducible) and different types of p53 mutants (temperature-sensitive mutant versus wild-type p53; point mutations versus deletion mutations). To avoid these problems, we have applied the tetracycline inducible expression system to stably express various p53 mutants in p53-null H1299 cells. On the basis of the results obtained in this study (Table I) and several previous studies (11, 12, 16, 32, 36, 38 -40), including our own (10,17,22), we propose the following model for p53 functional domains in apoptosis (Fig. 7). First, p53 DNA binding activity is necessary for apoptosis because mutants that are defective in the DNA binding and tetramerization domains are inert. Second, activation domain 2 and the proline-rich domain can form an activation domain for transactivating pro-apoptotic genes or transrepressing anti-apoptotic genes, because mutation or deletion in either one of the domains abrogates the apoptotic activity. Third, activation domain 1 is not required because deletion of or mutation in activation domain 1 (p53-(⌬AD1), p53-(AD1 Ϫ )) has little effect on apoptosis. Fourth, the C-terminal basic domain is necessary for maintaining p53 competent in inducing apoptosis, probably by relieving the inhibitory activity of activation domain 1, because p53-(⌬AD1⌬BD), but not p53-(⌬BD), is capable of inducing apoptosis.
Several p53 inducible genes, such as BAX (41), KILLER/ DR5 (42), and several PIGs (43), may participate in the apoptotic process. These genes can be induced by either p53-(⌬PRD) (17) or p53-(AD2 Ϫ ) (data not shown), both of which are active in inducing cell cycle arrest but not apoptosis, suggesting that these genes are not required or insufficient for inducing apoptosis. Recent evidence has shown that p53 can repress specific genes, such as MAP4 (44). It is possible that transrepression of anti-apoptotic genes plays an important role in p53-mediated  H and I). F, G, J, and K, apoptotic cells were quantified by propidium iodide-annexin V staining of cells that were non-induced (Ϫ p53) or induced (ϩ p53) for 3 days to express p53 (F and G) or p53-(⌬AD1⌬BD) (J and K). p53 Functional Domains for Apoptosis apoptosis. Therefore, the cell lines that inducibly express the p53 mutants described in this study, especially p53-(⌬AD1⌬BD), can be used to identify and determine whether a cellular gene is necessary for mediating p53-dependent apoptosis.
p53 transcriptional activity has been shown to be necessary for inducing cell cycle arrest (1,2,4,45). In this study, we extend this observation. We found that an activation domain capable of inducing at least partial cell cycle arrest can be formed by activation domain 1 plus activation domain 2, activation domain 1 plus the proline-rich domain, or activation domain 2 plus the proline-rich domain (Table I). When two of the three domains, i.e. activation domain 1, activation domain 2, and the proline-rich domain, become dysfunctional, the activity in cell cycle arrest is abrogated (Table I). It should be mentioned that p53-(AD1 Ϫ ) is defective in inducing cell cycle arrest although two functional domains, i.e. activation domain 2 and the proline-rich domain are still intact (22). However, when part or all of the residues for activation domain 1 are deleted, as in p53-(⌬1-23) and p53-(⌬AD1), the ability to induce cell cycle arrest is retained. This suggests that the presence of the mutated activation domain 1 may mask the activity of, or inhibit the interaction of, a potential co-activator (or an adaptor) with the activation domain formed by activation do-main 2 and the proline-rich domain necessary for transactivation or transrepression.
The search for mediators of p53-dependent cell cycle arrest has identified many cellular p53 target genes (1,4,46). p21 cip1/waf1 , a well characterized cyclin-dependent kinase inhibitor, can mediate cell cycle arrest in G 1 when overexpressed (22,(47)(48)(49)(50)(51). Previous studies have shown that p53-(AD1 Ϫ ), which is deficient in inducing p21, is incapable of inducing arrest in G 1 , consistent with the hypothesis that p21 plays an important role in mediating p53-dependent arrest in G 1 (22,40). In this study, we found that p53-(AD2 Ϫ ) is extremely active in inducing arrest in G 1 , suggesting that activation domain 1, but not activation domain 2, plays an important role in inducing cell cycle arrest. However, p21 is only slightly induced by p53-(AD2 Ϫ ) (Fig. 1A). Because p53-(AD1 Ϫ AD2 Ϫ ), which is deficient in both activation domain 1 and activation domain 2, is inert in inducing cell cycle arrest (9 -11), this suggests that a gene(s) responsible for arrest by p53-(AD2 Ϫ ) must be induced. This is not surprising because DNA damage-induced G 1 arrest is delayed but not abolished in p21-null fibroblasts from p21deficient mice (52,53). Therefore, the cell line that inducibly expresses p53-(AD2 Ϫ ) can be used to identify other novel gene(s) responsible for G 1 arrest.
Previously, several studies have shown that the p53 protein can be cleaved by cellular proteases in cells treated with DNA damaging agents, which leads to formation of several smaller polypeptides with molecular masses ranging from 35-50 kDa (54 -58). In addition, the cleavage of p53 is concomitant with the  [11][12][13][14][15][16]. Antibodies used were as described in the legend to Fig. 1.   FIG. 6. Regulation of p21 and BAX by p53 mutants. A Northern blot was prepared using total RNAs isolated from non-induced cells (Ϫ) or cells induced for 24 h to express wild-type p53 or various p53 mutants as shown above the blot (ϩ). The blot was probed with cDNAs derived from the p21, BAX, and glyceraldehyde-3-phosphate dehydrogenase genes, respectively. After normalization to the amount of glyceraldehyde-3-phosphate dehydrogenase transcripts, the levels of induction by wild-type p53 or various p53 mutants were quantified by phosphorimager and are shown below the blot. p53 Functional Domains for Apoptosis 39933 onset of apoptosis in cells treated with DNA damaging agents, suggesting that the cleaved p53 polypeptides are potent in p53 activity and may participate in the apoptotic process (58). Interestingly, one of the cleaved p53 polypeptides, p50, is p53-(⌬N23), which lacks the N-terminal 23 residues (58). We have shown previously that p53-(⌬N23) is active in inducing cell cycle arrest and apoptosis (10). Thus, the cellular machinery can generate an active but smaller p53 polypeptide that would not be subject to negative regulation by MDM2 (59 -63). It is not clear whether p53-(⌬AD1⌬BD) is an in vivo cleavage product of p53. However, because p53-(⌬AD1⌬BD) lacks the MDM2 binding site, it would not be subjected to the negative regulation by MDM2. Thus, p53-(⌬AD1⌬BD) represents a small but potent, apoptosis-inducing form of p53. Recent clinical trials have shown that adenoviruses expressing p53 are effective in treating some advanced forms of human cancers (64,65). We suggest that p53-(⌬AD1⌬BD) is a good candidate to replace the larger, unwieldy wild-type p53 in cancer gene therapy.