Role of Wild Type P53 in the G2 Phase: Regulation of the G-Irradiation- Induced Delay and DNA Repair

Role of wild type p53 in the G2 phase: regulation of the g-irradiation- induced delay and DNA repair

Dov Schwartz, Nava Almog, Amnon Peled, Naomi Gold®nger and Varda Rotter

Department of Molecular Cell Biology, Weizmann Institute of Science, Rehovot, Israel 76100

Upregulation of the p53 protein was shown to induce cell al., 1995; Wells, 1996). We have recently found that cycle arrest at the G1/S border and in some cases at the both the loss of wild type p53, or overexpression of

G2/M border. Furthermore, it was suggested that p53 is mutant p53 in myeloid cells, are associated with associated with the induction of the various DNA repair polyploidity (Peled et al., 1996). This supports the pathways. Previously, we demonstrated that cells co- conclusion that p53 may play a role at the spindle expressing endogenous wild type p53 protein, together checkpoint. with dominant negative mutant p53, exhibit deregulation p53 was shown to be involved in active cellular of apoptosis, G1 arrest and delay in G2 following g- repsonses that remove DNA damage, either by repair irradiation. In the present study, we investigated the role of the damaged DNA, or by an active killing of p53 protein in the DNA damage response at the G2 (apoptosis) of cells possessing extensive, unrepairable phase. Using p53-null, wild type p53 and mutant p53- DNA damage (reviewed in Enoch and Norbury, 1995). producer cell lines, we found that the two C-terminally p53-dependent apoptosis is triggered by various stress spliced p53 forms could prevent g-irradiation induced conditions, resulting in a clastrogenic eect on the mutagenesis prior to mitosis, at the G2/M checkpoint. DNA (Yonish-Rouach et al., 1991, 1993; Lowe and

We found that at the G2 phase, p53 may facilitate repair Ruley, 1993; Lowe et al., 1993; Lee et al., 1993; Clarke of DNA breaks giving rise to micronuclei, and regulate et al., 1993; Symonds et al., 1994). The activity of p53 the exit from the G2 checkpoint. At the G1 phase, only in DNA repair processes (Pan et al., 1995; Wang et al., the regularly spliced form of p53 caused growth arrest. 1995, 1996; Ford and Hanawalt, 1995; Bogue et al., In contrast, both the regularly and the alternatively 1996; McDonald III et al., 1996) was shown to be spliced p53 forms directed postmitotic micronucleated controlled by the C-terminus of the molecule, that cells towards apoptosis. These results provide a func- binds several DNA repair-related proteins, including tional explanation for the cell cycle-independent expres- the human Rad51 (Sturzbecher et al., 1996), RPA sion of p53 in normal cycling cells, as well as in cells (Dutta et al., 1993) and ERCC3 (Wang et al., 1994, where p53 is up-regulated, following DNA damage. 1995). In addition, GADD45 and the waf-1 gene, that are transcriptionally regulated by p53, were found to

Keywords: p53; G2 checkpoint; DNA damage; micro- bind PCNA, another DNA repair protein (Chen et al., nucleated cells 1995). Moreover, the last 100 amino acids of human p53 contain non-speci®c single strand (Bakalkin et al., 1994, 1995; Jayaraman and Drives, 1995) and mismatched (Lee, et al., 1995) DNA binding activity. Introduction The C-terminal domain of the regularly spliced p53 form (RSp53), also contains a DNA helicase activity It is well established that p53 is up regulated by various (Brain and Jenkins, 1994; Bakalkin et al., 1995; Wu et stress conditions, characterized by DNA damage al., 1995), characteristic of many DNA repair proteins. induction (Maltzman, Czyzyc, 1984; Fritsche et al., The eect of p53 on DNA repair is also evident 1993; Nelson and Kastan, 1994). One of the major from measuring residual DNA damage following proposed functions for p53 under these conditions, is genotoxic stress. One of the detectable manifestations the maintenance of genomic stability (Livingstone et of residual DNA damage in cells following g- al., 1992; Lane, 1992; Yin et al., 1992; Symonds et al., irradiation is the appearance of micronuclei (Heddle, 1994). Carrano, 1977). Micronuclei originate from acentric p53 was suggested to play a role in growth arrest at double strand breaks in chromosomes and from the G1/S border prior to DNA replication, and at the dicentric chromosomes, that give rise to chunks of

G2 phase prior to mitosis, by allowing the repair of chromatin that do not segregate with the rest of the faulty DNA that accumulates in cells (Kastan et al., genome after mitosis, but are often separately engulfed 1991, 1992; Di Lionardo et al., 1994). p53 was also by the nuclear membrane (Heddle and Carrano, 1977; shown to regulate the spindle checkpoint in mitotic Nusse et al., 1994). Thus, the presence of micronuclei embryonic ®broblasts. This checkpoint prevents the in cells is an indicator of either chromosomal formation of aneuploid cells that are bound to harbor aberrations that were not repaired until completion of an unstable genome after cytokinesis, and also prevents mitosis, or of an unstable karyotype. Interestingly, ectopic DNA endoreduplication in these cells (Cross et reticulocytes from p53-knockout mice contain elevated levels of micronuclei under normal growth conditions and in cells with DNA damage (Lee et al., 1994). Moreover, bone marrow cells from these mice exhibit a Correspondence: V Rotter high frequency of chromosomal abnormalities (Bouer Received 29 April 1997; revised 18 July 1997; accepted 21 July 1997 et al., 1995). Furthermore, wild type p53, in contrast to p53 and the G2 checkpoint D Schwartz et al 2598 mutant p53, was found to accumulate in micronuclei of cells that are p53-null lymphoid pre-B, stably g-irradiated hepatoma cell lines (Unger et al., 1994). expressing regularly (RSp53) or alternatively-spliced

Although it is clear that the onset of the G2 p53 (ASp53) forms and their mutant counterparts, checkpoint is p53-independent, and involves a rapid were analysed. Both alternatively spliced wild type p53 phosphorylation of the MPF cyclin B1-cdc2 complex, forms facilitated the exit from the g-irradiation-

p53 may play a role as a secondary regulator of the G2 dependent G2 delay. Analysis of micronuclei content

delay (Stewart et al., 1994; Paules et al., 1995). In following exit from the G2 delay indicated that cell several p53-null non-lymphoid cell lines, reconstitution lines expressing the two forms of wild type p53 of wild type p53 conformation (Stewart et al., 1994), or exhibited 20% of micronuclei detected in p53-null transcriptional induction of wild type p53 without any cells. Analysis of p53 distribution in post mitotic

DNA damage (Agarwal et al., 1995) resulted in a G2- diploid micronucleated cells indicated that wild type

as well as a G1- arrest. The role of p53 as a second p53 accumulates in the L12 cells expressing wild type

barrier at the G2 checkpoint, is further supported by p53, and promotes their apoptosis. This eect was not the ®nding that annihilation of the phosphorylation- related to the spindle checkpoint, since in these cells

dependent G2-delay by caeine, is reduced in wild type this checkpoint was p53-independent. While both p53 expressing cells (Powell et al., 1995). This agrees alternatively spliced p53 forms are involved in the

with our observation that functional inactivation of G2-related activities, the ability to arrest at the G1/S endogenous p53 in lymphoid cells, by stably expressing border following g-irradiation was evident only in L12 dominant negative mutant p53, induces premature exit cells reconstituted with wild type RSp53.

from the g-irradiation-dependent G2 arrest (Aloni- Grinstein et al., 1995). The goal of this study was to elucidate the role of Results p53 in the response to DNA damage at the G2 phase. We compared the time course of the cell cycle, the Wild type p53 controls the kinetics of exit from G delay repair of DNA damage following g-irradiation of cells 2 in g-irradiated L12 cells at the G2 phase and the selective apoptosis of micronucleated postmitotic cells. Since most of the Previous data indicated that wild type p53 plays a role

DNA repair associated activites were ascribed to the in the G2 checkpoint. To analyse this activity, regularly spliced C-terminal of p53 (Shaulian et al., logarithmically growing cells of p53 producer and 1994; Lee et al., 1995; Wu et al., 1995), we compared non-producer L12 derivatives were enriched for the eect of re-introduction into p53 null cells of the tetraploid DNA constant by centrifugal elutriation two alternatively spliced p53 forms that vary at their under viable conditions. The cells were collected, C-terminal (Wolf et al., 1985; Arai et al., 1986). L12 replated and g-irradiated (with 300 R), or kept under

Figure 1 Wild type p53 accelerates the exit from the G2 delay after irradiation: Cell cycle patterns. L12 clones expressing regular spliced wild type RSp53-10 (marked in gray in a, b, c, alternatively spliced wild type ASp53-22 (marked as empty in a, b, c)or empty pSVL plasmid, clone pSVL-3 (marked in black in a, b, c) were grown at logarithmic conditions and enriched for the G2 phase using centrifugal elutriation under viable conditions (a). Normal cell cycle distribution is depicted by dotted line in a. The cells were either directly replated, or g-irradiated (300 R) and then replated. (b) shows the cell cycle of pattern g-irradiated cells 12 h after g- irradiation. (c) shows the cell cycle pattern of non-irradiated clones, 6 h after replating. (d, e, f) parallel (a, b, c) except that the clones used in these experiments are mutant RSp53-M11-3 (marked in gray), mutant ASp53-M8-3A2 (marked as empty) or empty pSVL plasmid pSVL-3 (marked in black) p53andtheG2 checkpoint D Schwartz et al 2599 the same conditions as the g-irradiated cells, and as the G2-enriched fractions. Since less cells progressed replated under normal growth conditions. At various from the diploid peak of wild type RSp53 expressing times after replating, cells were ®xed in 70% ethanol, clones, we concluded that under normal growth and analysed by FACS. To measure the distribution of conditions, wild type RSp53 confers a prolonged G1 the various cell populations corresponding to the phase, compared to the p53-null control (Figure 2b). phases of the cell cycle, standard PI DNA-content Mutant p53 (Figure 2e) as well as wild type ASp53 analysis by FACS was employed (Figure 1). (Figure 2b) did not signi®cantly alter the length of the

The cell cycle pattern of either g-irradiated or non- G1 phase. Furthermore, following g-irradiation, only irradiated G2/M enriched fractions of p53-null clones wild type RSp53 (Figure 2c), but not mutant (Figure (pSVL-3, pSVL-1) was compared to regularly spliced 2f) or wild type ASp53 (Figure 2c) altered the length of wild type p53 expressors (RSp53-10 and RSp53-20), the G1 checkpoint. the alternative spliced wild type p53 producers (ASp53- Standard cell cycle analysis using PI incorporation is

20 and ASp53-22), and to mutant p53 expressing L12 not suitable for the detection of a G2 delay, that is clones (RSp53M11-3 and ASp53M8-3A2). Cell cycle manifested by the disappearance of the mitotic cells. distribution of L12 clones expressing the various p53 To determine the existence of such delay, we utilized types immediately following enrichment for G2 phase, the FACS assay based on Acridine Orange (AO) DNA is shown in Figure 1a and d. Analysis of the kinetics of denaturability. This assay is based on dierential DNA progression from G2 of non-irradiated L12 cell denaturability of interphase and mitotic nuclei at low fractions indicated that both wild type p53 expressors pH, thus enabling to calculate the percentage of mitotic

(Figure 1c) and mutant p53 expressors (Figure 1f) as well as G1, S and G2 cells in a given population. The exhibited the same length of the G2 phase compared to supercoiled DNA of mitotic cells has less re-annealing the p53-null L12 control cells. In contrast, following g- capacity upon DNA denaturation, resulting in more irradiation, L12 cells harboring wild type p53 exited single stranded DNA (ssDNA), that can be detected faster from the G2 delay as compared to the p53-null using the metachromatic DNA binding dye AO. AO parental cell line (Figure 1bb, or cell lines that harbor has dierent emission wavelengths upon binding to mutant p53 (Figure 1e). Furthermore, clones harboring double stranded versus single stranded DNA. The G2/ wild type RSp53 were more ecient than wild type M enriched fraction that was obtained from the

ASp53 expressors in the exit from the G2 delay (Figure various cell lines using centrifugal elutriation, was g- 1b). irradiated and subjected to AO analysis. Figure 3a

To evaluate the eect of p53 expression at the G1 shows the cell cycle distribution obtained by the AO phase in these L12 clones, G1-enriched fractions were DNA denaturability assay. The onset of the G2 arrest isolated by centrifugal elutriation (Figure 2a and d) is very fast in all L12 derivatives as demonstrated by and analysed under the same experimental conditions the complete disappearance of the mitotic populations,

Figure 2 Wild type RSp53 arrests cells at G1 phase following g-irradiation: Cell cycle pattern. L12 clones expressing regular spliced wild type RSp53-10 (marked in gray in a, b, c); alternatively spliced wild type ASp53-22 (marked as empty in a, b, c) or empty pSVL plasmid, clone pSVL-3 (marked in black in a, b, c) were grown as baove (see Figure 1) and enriched for the G1 phase using centrifugal elutriation at viable conditions (a). Normal cell cycle distribution is depicted by dotted line in a. The cells were either directly replacted or g-irradiated (300 R) and then replated. (c) shows the cell cycle pattern of G1 enriched fraction 6 h after irradiation. (b) shows the cell cycle pattern of the non-irradiated clones. (d, e, f) parallel (a, b, c) except that the clones used in these experiments are mutant RSp53 (marked in gray), mutant ASp53 (marked as empty) or empty pSVL plasmid (marked in black) p53 and the G2 checkpoint D Schwartz et al 2600 as early as 2 h post g-irradiation (Figure 3a, second We next analysed the same AO stained samples, by

row). The mitotic populations of wild type p53 gating on the cell populations of the G2 phase. Figure 4

expressing cells, (Figure 3a, arrows) are evident shows the amount of G2 cells at various times, as

already 6 h after irradiation (third row). However, percent of the number of G2 cells at time 0. Figure 4a mutant p53 expressors (Rsp53M11-3 and ASp53M8- shows the comparison between wild type p53 3A2), or p53-null control cells (pSVL-3), were still expressors and the p53-null control, while Figure 4b

arrested at G2 at this time point. compares pSVL p53 null cells to mutant p53 The overall plot of the percentage of mitotic cells at expressors. g-irradiation induced speci®c accumulation

various time points following g-irradiation shown in at the G2 phase in all L12 clones. The exit from the G2 Figure 3b, con®rms that wild type p53 expressors of arrest, manifested by the reducton in the percent of the

both RS and AS forms, have earlier exit from the G2 G2-arrested cells, is much faster in wild type p53 arrest compared to p53-null control cells. RSp53 expressing clones than in mutant p53 expressors, or in

expressors exit the G2 delay signi®cantly faster than the p53-null parental clone. Analysis of cell cycle ASp53 expressing cells. Mutant p53 expressors had progression of the mutant p53 expressing clones

similar kinetics of exit from G2 delay to that of p53- (Figure 4b), indicated that these p53 mutants have no null pSVL control cells (data not shown). dominant negative eect, neither on regulation of the

Figure 3 The eect of irradiation on the mitotic phase of the G2 enriched L12 clones. The clones shown in the upper part were enriched for G2 cells (the upper row), and irradiated (300 R). (a) shows the amount of mitotic and interphase cells (see markings in the pSVL column). The enriched fractions were analysed by FACS using the Acridine Orange DNA denaturability method enabling to distinguish between interphase and mitotic cells, evaluating the cell cycle stage in interphase cells. The second and the third rows show the cell cycle phase 2 h and 6 h after irradiation. (b) shows the percent of mitotic cells normalized to time 0 of control. -&- L12 pSVL; -&- wild type RSp53; -~- L12 wild type ASp53-22 p53andtheG2 checkpoint D Schwartz et al 2601

Figure 4 Kinetics of the G2 phase in normal growth and after g-irradiation of L12 clones. (a) shows the percent of G2 cells in G2- enriched fractions with or without irradiation: -*- L12 pSVL irradiated; -*- L12 pSVL non-irradiated; -&- L12 wild type RSp53- 10 irradiated; -&- L12 wild type RSp53-10 non-irradiated; -~- L12 wild type ASp53-22 irradiated; -~- L12 wild type ASp53-22 non irradiated. (b) shows the percent of G2 cells in G2-enriched fractions with or without irradiation: -&- L12 pSVL irradiated; -*- L12 pSVL non-irradiated; -&- L12 mutant RSp53-M11 irradiated; -~- L12 mutant RSp53-M11 non-irradiated; -~- L12 mutant ASp53-M8 irradiated; -*- L12 mutant ASp53-M8 non-irradiated

G2 delay upon DNA damage, nor on the length of the not undergo apoptosis, thus enabling their quantita-

G2 phase at logarithmic cell growth. tion.

These results suggest that both RSp53 and ASp53 Elutriation enriched G1 and G2/M fractions of forms facilitate the exit from the G2 delay with logarithmically growing L12 clones harboring wild dierent kinetics. Mutant p53 expression did not type or mutant p53 protein were replated, g- signi®cantly aect the cell cycle pattern, as compared irradiated (at 300 R) and the amount of micronuclei to p53 null parental cells. at various time points was determined by FACS. Cells were treated with hypotonic buer containing non-ionic detergent to obtain isolated nuclei, and p53 is involved in the reapir of micronuclei generated by micronuclei. The samples were stained with PI and DNA damage at the G phase 2 analysed by FACS for DNA content, size and

Accumulation of cells at the G2 phase results from granularity, that enables to distinguish micronuclei DNA damage. It was previously shown that the from intact nuclei, apoptotic blebs and cell debris amount of micronuclei in postmitotic cells induced by (Figure 5a). Figure 5b shows percent micronuclei DNA damage, is inversely correlated with the eciency normalized to the amount of divided cells in L12 of the double strand break (DSB) DNA reapir at the clones expressing either wild type or mutant p53 preceding G2 phase, and to the integrity of the G2 and protein. The ratio between micronuclei and intact spindle checkpoints. We did not detect alterations in nuclei indicated that all logarithmically growing L12 the arrest at the G2 checkpoint, since all clones lacked clones contained up to 1.5% of micronuclei (Figure mitotic cells 2 h after g-irradiation. Analysis of the 5b, sections G1 0 h and G2 0 h). After g-irradiation integrity of the spindle checkpoint, using prolonged and release of the g-irradiated cells from the G2 cultivation in presence of nocodazole, indicated that all delay, wild type RSp53 and ASp53 expressors L12 clones developed low super-tetraploid endoredu- contained 8 ± 20% micronuclei in the postmitotic cell plicated cell populations, characteristic of a disrupted population, while p53-null control cells and mutant spindle checkpont (data not shown). Thus, both G2 p53 expressors contained 38 ± 50% micronuclei and the spindle checkpoints were probably intact in all (Figure 5b, section G2 12 h IR). L12 clones, and could not account for the dierent The notion that the formation of micronuclei kinetics of exit from the G2 delay. Dierences in DNA following DNA damage is dependent on transition repair pro®ciency between wild type p53 expressors through mitosis is supported by the ®nding that in and non-expressors at the G2 phase, could possibly both G1 and G2 enriched cells, the time course of account for the dierential exit from the G2 delay that micronuclei induction after DNA damage paralleled we observed. the resumption of mitosis (Figure 5b sections, G1 12 h We have, therefore, examined the contribution of IR; G2 12 h IR). p53 to the elimination of DNA breaks at G2,by The reduced number cells with micronuclei in clones determining the amount of micronuclei in cells harboring wild type p53 suggests that wild type p53 immediately after exit from the G2 delay. At this may aect the repair of damaged DNA at the G2 point the micronucleated cells were still intact and did phase. p53 and the G2 checkpoint D Schwartz et al 2602 p53 protein is upregulated in micronucleated cells stained by p53 much more intensely than other cells (Figure 6b), indicating a `speci®c activation' of p53 in We next studied the distribution of wild type p53 in response to chromosomal aberrations or DNA breaks. micronucleated cells. In all L12 derived cell lines p53 Apoptotic blebs (Figure 6a upper row), appear to loose was expressed constitutively at low levels. To determine p53 expression. These results indicated that the p53 the subcellular distribution of p53 in micronucleated pathway is speci®cally activated in micronucleated cells, cells which overexpressed wild type p53 protein cells. Judged by the patterns of p53 expression and were chosen. We generated L12 derived clones the dierence in DNA staining, the micronucleated expressing high levels of the temperature sensitive cells could be distinguished from apoptotic blebs. Val135 RSp53 variant. Cells were grown at 378C, g- irradiated (500 R) and incubated for 24 h at 328Cto Wild type p53 induces apoptosis in micronucleated cells allow wild type p53 conformation (Gottlieb, et al., 1994). Cells were then ®xed and stained using the PAb- In order to examine whether wild type p53 is also able 421 or PAb-248 anti-p53 monoclonal antibodies, and to eliminate selectively cells containing micronuclei, the DNA was stained with Hoechst-33342. As seen in L12 p53 non-producer cell line (pSVL), wild type Figure 6, staining (upper row) revealed that micro- RSp53 (RSp53-10) and wild type ASp53 (ASp53-22) nuclei (marked by arrows) appeared as de®ned nuclear were g-irradiated (600 R) and grown for 28 h when extensions with the same DNA pattern as the nucleus. apoptosis had reached its peak (data not shown). Each The apoptotic blebs exhibited a typical picnotic DNA cell line was analysed by FACS for DNA content using pattern (Figure 6a). Micronucleated cells are often the vital dye HO and for cell viability, by PI exclusion. Figure 7Aa shows the typical HO-PI distributions obtained by this method (represented by the pSVL clone). Cells containing a sub-diploid population a (typical for apoptosis) also exhibit intense PI uptake, indicating high membrane permeability (typical for dead cells). Cellular populations containing dying cells possessed intermediate PI incorporation that retained normal DNA content, and living cells lacking PI. The living and dying fractions of each clone were sorted by the FACStar+ ¯ow cytometer using these parameters. Such analysis revealed that these dierent populations in each clone were more than 96% pure. The sorted dying cell fraction (represented in the pSVL clone) is shown in Figure 7Ab, and the sorted living cell fraction is shown in Figure 7Ac. The non-sorted cells (Figure

AB DNA staining b p35 Expression

Figure 6 p53 distribution in micronucleated and apoptotic cells following g-irradiation. Logarithmically growing L12 cells stably expressing the temperature sensitive p53 mutant were g-irradiated Figure 5 Analysis of micronuclei by FACS. Cells were enriched (500 R) and grown at 328C for 24 h. The cells were ®xed in for the various cell cycle fractions by centrifugal elutriation, g- methanol and stained for p53 protein using monoclonal anti-p53 irradiated (300 R) and analysed by FACS for the amount of antibodies PAb-421 and PAb-248. The chromatin structure was micronucleated cells. (a) shows a typical distribution of the visualized using the Hoechst-33342 DNA dye. The upper row micronucleated cells. (b) shows a chart with the amount of shows the DNA pattern, and the lower row shows p53 micronuclei divided by the amount of nuclei (percent) from the distribution. Column A shows the comparative p53 and DNA total population (in G1 and G2 0 h, and in G1 12 h IR patterns of apoptotic versus micronucleated cells and column B categories), or of the amount of post mitotic diploid cells represents micronucleated versus intact cells. The micronuclei are appearing after the release from the G2 (12 h) arrest marked by black arrows p53andtheG2 checkpoint D Schwartz et al 2603

Figure 7 Wild type p53 eliminates cells harboring damaged DNA. The p53-null clone containing the empty pSVL plasmid (pSVL), the wild type RSp53 (RSp53-10), or wild type ASp53 (ASp53-22) were irradiated (600 R), grown for 26 h, and incubated for 1 h with 5 mg/ml of Hoechst-33342 to label DNA. Cells were also stained with 50 mg/ml of PI to label dying and dead cells. Cells were analyzed and sorted by FACS using the FACStar+ ¯ow cytometer (A a). The sorted fractions of dying (A b) and living (A c) cells, as well as the unsorted total cell population were subjected to micronuclei assay by FACS on a FACSort ¯ow cytometer. The ratio of micronuclei in the dying cell fraction/micronuclei in the living fraction for each cell line is shown in section B

7Aa), as well as the two sorted fractions of each cell that p53 expression during embryogenesis is con®ned line, were further examined for the amount of to proliferating and newly dierentiating regions micronuclei using the FACsort ¯ow cytometer as (Rogel et al., 1985; Schmid et al., 1991). described above (Figure 5aa. Figure 7b depicts the p53 appears to be a key protein in the control of the relative abundance of micronuclei found in dying cells cell cycle. It is generally accepted that in addition to its as a fraction of micronuclei found in the living cell role in the G1 phase, p53 is also involved in S and G2 population, for each clone. Wild type RSp53 and the phases of the cell cycle (Remvikos et al., 1990; Danova ASp53 producer cell lines, retained less micronucleated et al., 1990). p53 was shown to be upregulated living cells than the parental p53-null clones. More- following DNA damage, at all stages of the cell cycle over, signi®cantly more micronucleated dying cells (Yamaizumi, Sugano, 1994). In addition to the arrest were detected in wild type p53 producer clones. at the G1/S border (Kastan et al., 1991, 1992; Kuerbitz These results suggest that p53 may selectively direct et al., 1992; Di Leonardo et al., 1994) elicited by p53 cells with DNA damage that result in micronuclei via the waf-1 pathway (El-Deiry et al., 1994), high formation, towards apoptosis. levels of wild type p53 can also induce a G2 arrest (Stewart et al., 1994; Agarwal et al., 1995). In Ataxia- Telangiectasia, where the p53 pathway is inactivated or

Discussion deregulated, the DNA damage-dependent G2 arrest is much longer than in normal cells. This is explained by The p53 knockout mouse model has proven unequi- the lower repair eciency, especially of double strand vocally that p53 functions as a tumor supressor gene. breaks that occur in this syndrome (Hong et al., 1994; p53 null mice developed a high incidence of tumors Barlow et al., 1996; Hawley and Friend, 1996). In during adult life (Donehower et al., 1992; Jacks et al., agreement with this notion, we have previously 1994). Moreover, the recent observations showing that observed that inactivation of wild type p53 function these mice exhibit a variety of developmental aberra- by dominant negative mutant p53, shortened the g- tions (Armstrong et al., 1995; Sah et al., 1995), support irradiation induced G2 cell arrest (Aloni-Grinstein et the conclusion that p53 is also essential for normal al., 1995). Furthermore, ®broblasts derived from Li- development. This agrees with early reports showing Fraumeni patients, lost the normal G2 checkpoint p53 and the G2 checkpoint D Schwartz et al 2604

response following homozygotisation of the mutant The integrity of the G2 checkpoint could be analyzed p53 allele (Paules et al., 1995; Kaufmann et al., 1995). accurately only by measuring the percent of mitotic p53 was also shown to prevent endoreduplication of cells at various times following DNA damage. We the tetraploid genome during prolonged mitotic arrest utilized the AO DNA denaturability assay by FACS,

by regulating the spindle checkpoint (Cross et al., 1995; enabling to de®ne mitotic, as well as G1, S and G2 cells Wells, 1996). In agreement with this notion, we found in a given population. In agreement with previous

that after induction of the mutant p53 conformation, results, all clones arrested at the G2 phase following g- M1/2 myelomonocytic cells expressing temperature irradiation, did contain mitotic cells ± 2 h after treat-

sensitive mutant p53 had more giant polyploid cells, ment. Accurate analysis of the G2 fraction using this and induction of the wild type p53 conformation assay con®rmed that clones expressing wild type p53 of reduced their number (Peled et al., 1996). both alternatively spliced p53 forms existed from the

The notion that a p53-dependent pathway is G2 checkpoint faster than the mutant p53 expressors involved in DNA repair in addition to the regulation and p53-null cells. p53, therefore, aects more

of cell cycle checkpoints and apoptosis of damaged signi®cantly the DNA repair at the G2 phase in these

cells by p53 is supported by several ®ndings: For cells, causing an accelerated exit from the G2 delay.

example, p53 was shown to accelerate excision repair The analysis of the eciency of DNA repair at the G2 (Pan et al., 1995; Wang et al., 1995; Bogue et al., 1996). phase in all clones, by determining the number of

Waf-1-null human cancer cells, possess mismatch micronuclei in cells that exited the G2 delay, indicated repair de®ciency (McDonald III et al., 1996). Li- that wild type p53 expressors exhibited a signi®cantly Fraumeni derived ®broblasts homozygous for p53 lower levels of micronuclei. Mutant p53 expressors had mutations, are de®cient in global DNA repair (Ford the same levels of micronuclei as those in p53 non- and Hanawalt, 1995). p53 was found to bind to several producers, suggesting that mutant p53 does not aect

genes directly involved in excision repair, such as RPA these G2-related activities. Non-irradiated cells, had very (Dutta et al., 1993), ERCC3 (Wang et al., 1994), and low quantities of micronuclei, and the time course of human Rad51 protein involved in double strand break micronuclei appearance followed that of cell division. To recombination repair pathway (Sturzbecher et al., prove that the number of micronuclei represents 1996). Moreover, p53 functions as a DNA helicase authentically the eciency of DNA repair at the

via its C-terminus, that has the ability to bind preceding G2 phase, it was essential to determine nonspeci®cally single stranded (Bakalkin et al., 1994, micronuclear number before the onset of apoptosis. 1995; Jayaraman and Prives, 1995) and mismatched, or The results indicated that p53 has a role in the regulation

nicked DNA (Lee et al., 1995). of double strand break repair of DNA in G2 phase, as well

This study addressed the role of p53 in DNA repair as in the exit from the DNA damage dependent G2-delay.

and cell cycle arrest at the G2 phase. The eect of p53 We also analysed the possibility that wild type p53 is

upon g-irradiation of G2 phase enriched cellular activated in micronucleated cells. Immunocytochemical

fractions was analyzed regarding the G2 delay and analysis of p53 overproducer L12 clones expressing the the eciency of double strand break repair in various temperature sensitive RSp53 after g-irradiation, indi- L12 clones. These cell lines were generated by cated that DNA damage induced the accumulation of introducing p53 expression regulated by a weak p53 in the micronucleated cells. Occasionally, only the promoter that codes for low levels of p53, thus micronucleus had intense p53 staining, while the mimicking physiological condition. We found that the regular nucleus showed only weak staining for p53. spindle checkpoint is inteact in all L12 clones, and This may point to the possibility that p53 has the therefore, we have focused on the function of p53 at ability to be targeted to damaged DNA in vivo.

the G2 checkpoint. We analyzed the cell cycle pattern Finally, we have shown that wild type p53 is also by standard FACS analysis of logarithmically growing able to direct micronucleated cells towards apoptosis.

L12 clones enriched for G1 or G2/M phases by Comparison of the incidence of micronuclei in living centrifugal elutriation, before and after g-irradiation. and dying fractions of wild type p53 producers versus We found that both alternatively spliced forms of wild p53-null cells revealed that the former had signi®cantly type p53 protein, but not the corresponding mutant more micronucleated dying cells than the latter. forms, accelerated the exit, but did not aect the entry, Moreover, there were less micronucleated living cells

into g-irradiation-induced G2 delay when compared to left in the p53 producer populations. These observa- the p53-null parental clone. This implies that the onset tions suggest that wild type p53 has the ability to

of the G2 checkpoint is p53-independent in L12 clones. speci®cally eliminate micronucleated cells. Non-g-irradiated wild type and mutant p53 expressing It appears therefore, that p53 seems to exhibit two cell lines exhibited similar kinetics of progression when opposite eects on DNA damage-dependent cell cycle

the G2 enriched fractions were analysed. checkpoints. Based on our previous results it appears

Examination of the g-irradiation reponse of G1 that p53 arrests damaged cells until their DNA is enriched populations showed that only the wild type repaired, or alternatively, promotes cells with unrepair-

RS p53 expressing clones displayed G1 arrest, that was able damage towards apoptosis (Aloni-Grinstein et al., evident at 6 h after g-irradiation. Constitutive expres- 1995). On the other hand, our present data support the sion of low levels of wild type p53 in these cells notion that p53 facilitates the repair of damaged DNA,

prolonged the G1 phase, also under normal growth thus accelerating the exit of cells from the DNA conditions (Shaulsky et al., 1991). The inability of wild damage-dependent cell cycle checkpoints. This implies

type ASp53 to regulate eciently the G1 arrest is in that the net eect of p53 on the cell cycle response to agreement with the notion that this form of p53 has DNA damage may depend on the cell type, the level of

functions that are speci®c for the G2 phase (Kulesz- p53 in the cell, and on the type of DNA damage, that Martin et al., 1994). is repaired by the dierent pathways. p53andtheG2 checkpoint D Schwartz et al 2605 Taken together, these results can provide a Measurement of cell cycle using the AO DNA denaturability functional explanation for the cell cycle-independent assay by FACS expression pattern of p53 during normal growth, and Cells were harvested and ®xed for 2 ± 24 h in 70% the upregulation and nuclear targeting of p53 upon ethanol ± 30% HBSS (Hanks Balanced Salt Solution) v/v, DNA damage, in diploid and tetraploid cell cycle at 7208Cat16106 cells/ml. The cells were washed once stages. with HBSS, and resuspended in HBSS with 0.5 mg/ml RNase A. After 1 h incubation at 378C, the cells were washed once and resuspended in HBSS at 26106 cells/ml. Cell suspensions (200 ml) were added to 0.5 ml of 0.1 M HCl. After 30 s incubation at RT, 2 ml of Acridine Orange Materials and methods (AO) (Molecular Probes) staining solution of pH 2.6, containing 90% v.v. 0.1 M Na-Citrate and 10% v/v 0.2 M Cell lines and culture medium Na2HPO4 and 6 mg/ml AO were added (Darzynkiewicz, The murine L12 cells were grown in suspension at 378Cin Gong, 1994). Cells were analysed by the FACS using RPMI-1640 medium containing 10% fetal calf serum and FACSort ¯ow cytometer and CellQuest list mode analysis 261075M b-mercaptoethanol. L12 clones carrying the software (Beckton-Dickinson) (Darzynkiewicz, 1994). regularly spliced wild type p53 cDNA, L12-RSp53-10; 20 (Shaulsky et al., 1991), and alternative spliced wild type Detection of micronuclei by ¯ow cytometry p53, L12 ASp53-22; 23, generated by transfection with alternatively spliced C-terminal p53 coding plasmid and the The protocol was described before (Nusse et al., 1992, pSV2-gpt plasmid containing the drug resistance gpt gene, 1994; Schrieber et al., 1992). Brie¯y, the cells were washed were grown in RPMI-1640 medium containing 2 mg/ml in PBS once. The pellet (0.1 ± 16106 cells) was resuspended mycophenolic acid, 150 mg/ml xantine and 15 mg/ml and vortexed for 20 s in 1 ml of Solution I containing hypoxantine. The L12 clones carrying mutant p53 were: 584 mg/L NaCl; 1 gr/L Na-citrate; 10 mg/L RNAse A L12-M11 (Regularly spliced p53 Glu?Gly mutation at from bovine pancreas, preheated to inactivate the residual codon 168) and L12-M8 (alternative spliced p53 Cys?Phe DNAses (Boeringher Mannheim); 0.3 ml/L NP-40 (Sigma) mutation at codon 133). Clone L12-109 (tsRSp53) was and 25 mg/ml PI (Sigma). The cell suspension was established by infection of the parental L12 cells by co- incubated for 1 h in the dark at RT, vortexed, and 1 ml cultivating with pLXSN-p53Va/135 gpe+ virus producing of Solution II containing 40 mg/ml PI, 15 mg/L Na-Citrate cells (Gottlieb et al., 1994) (kindly provided by Dr M Oren, and 0.25 M sucrose was added. After a short vortexing, the The Weizmann Institute of Science) for 48 h and the samples were analysed on a FACSort. selection of single cell clones was achieved by limited dilution and incubation with 2 mg/ml G-418 drug. p53 Incubation of cells with nocodazole expression in G-418-resistant cells was determined by Western blotting. Logarithmically growing L12 clones were incubated with 0.05, 0.1 and 0.5 Mg/ml of Nocodazole (Sigma) for 12, 24, 48 and 96 h. The cells were ®xed in 80% ethanol (BioLab) Antibodies at 7208C for 24 h and the cell cycle pattern was analysed Monoclonal antibodies to p53 PAb-421 (Harlow et al., at each time point using routine DNA content analysis by 1981), PAb-242, and PAb-248 (Yewdell et al., 1986; FACS. Cell clumps were excluded, by live gating on single Gannon et al., 1990) were obtained of ascitic ¯uids of cells, according to the area and the width of the PI (FL/2) the peritoneum of hybridoma-bearing syngeneic Balb/c signal. mice, that were spun to remove tissue debris, diluted in PBS and used without further puri®cation. Immunocytochemical staining for p53 and for chromatin Logarithmically growing L12 clones were washed once Cell cycle fractionation by centrifugal elutriation and g- with PBS and ®xed overnight in 80% methanol (BioLab) at irradiation 48C in suspension. The ®xed cells (1.56106) were washed 1±36108 cells were loaded into an elutriation rotor (J-6MI once in PBA (PBS containing 0.5% w/v BSA (fraction VII centrifuge equipped with a JE-5.0 elutriation system, Sigma) and 0.02% w/v sodium azide (Sigma), and including a Sanderson chamber (Beckmann Instruments incubated for 1 h on ice in 100 MI of ascitic ¯uid Inc.) and MasterFlex (Cole-Parmer Instruments) peristaltic containing either PAb-421 or PAb-248 monoclonal pump pre-sterilized with 5% Sodium hypochlorite. The antibodies diluted 1 : 40 in PBA. The cells were washed separation was performed at room temperature (20 ± 228C) twice with PBA and incubated on ice with Fluorolink- using RPMI medium supplemented at 2% heat inactivated Cy2TM labeled Goat anti-Mouse IgG (H&L) antibody fetal calf serum with a constant centrifuge speed of 2200 (Amersham Cat. PA42002) for another 1 h. After two 6 r.p.m. (rmax=*470). The cells were loaded with pump washes with PBA the cells were concentrated to 5610 speed of 12 ml/min, and diploid (G0-G1): (at 15 ± 21 pump cells/ml in PBA containing 5 mg/ml Hoechst 33342 speed) and tetraploid (at pump speed 29 ± 36) fractions (Molecular Probes Inc.). The cell suspension was either were collected. After the separation, cells were washed once directly applied on slides or centrifuged in a cytospin with DPBS, counted and replated in RPMI supplemented centrifuge and the ¯uorescent cells were visualized and with 8% fetal calf serum at 0.3 ± 0.66106 cells/ml. Part of photographed in an Axioplan ¯uorescent microscope the cells were g-irradiated (with 300 rads) after replating, (Zeiss) using 1600 ASA strong BW KODAK ®lm. and their fate was followed. Viability and DNA content analysed by HO-PI: sorting of living Standard cell cycle analysis by FACS and dying population Cells were ®xed in 70% ethanol (BioLab) at 7208C, for at Logarithmically growing L12 cells were incubated for 1 h least 2 h and stained with 50 mg/ml Propidium Iodide with 5 mg/ml of Hoechst-33342 (Molecular Probes) that (Sigma). The cells were analysed by a FACSort ¯ow intercalates into DNA, and for 3 min with 25 mg/ml PI cytometer (Beckton Dickinson), using the CellQuest (Sigma) that is incorporated in dying cells, having (Beckton Dickinson) software. ruptured membranes (Belloc et al., 1994). The cells were p53 and the G2 checkpoint D Schwartz et al 2606 analysed on a dual laser FACStar+ cell sorter (Beckton nonsorted cells were subjected to the micronucleus assay as Dickinson Inc). The u.v. laser excited Hoechst-33342 described. ¯uorescence was recorded on a FL/3 channel, and the PI ¯uorescence was recorded on the FL/2 channel. Using these two parameters, the cells were sorted into two Acknowledgements groups: Living-Low FL/2 ¯uorescence, Dying-Intermedi- The authors wish to thank Prof Avri Ben-Ze'ev and David ate FL/2 ¯uorescence. The sorting conditions were also Wisemann for fruitful discussion and criticism. This work selected for more than diploid DNA content according to was supported in part by grants from the Leo and Julia FL/3 ¯uorescence. Forchheimer Center for Molecular Genetics, DKFZ and 0.56106 cells were sorted from each fraction and the Minerva Foundation. VR is the incumbent of the analysed for purity. In all fractions, the enrichment was Norman and Helen Asher Professional Chair in Cancer more than 96%. Subsequently, the sorted fractions, and Research at the Weizmann Institute.

References

AgarwalML,AgarwalA,TaylorWRandStarkGR.(1995). Gottlieb E, Haner R, von Ruden T, Wagner EF and Oren Proc. Natl. Acad. Sci. USA, 92, 8493 ± 8497. M. (1994). EMBO J., 13, 1368 ± 1374. Aloni-Grinstein R, Schwartz D and Rotter V. (1995). EMBO Harlow E, Crawford LV, Pim DC and Williamson NM. J. 14, 1392 ± 1401. (1981), J. Virol., 39, 861 ± 869. Arai N, Nomura D, Yokota K, Wolf D, Brill E, Shohat O Hawley RS and Friend SH. (1996). Genes. and Dev., 10, and Rotter V. (1986). Mol. Cell. Biol. 6, 3232 ± 3239. 2382 ± 2388. Armstrong JF, Kaufman MH, Harrison DJ and Clarke AR. Heddle JA and Carrano AV. (1977). Mutat. Res., 44, 63 ± 69. (1995). Curr. Biol. 5, 931 ± 936. Hong JH, Gatti RA, Huo YK, Chiang CS and McBride WH. Bakalkin G, Selivanova G, Yakovleva T, Kiseleva E, (1994). Radia. Res., 140, 17 ± 23. Kashuba E, Magnusson KP, Szekeley L, Klein G, Jacks T, Remington L, Williams BO, Schmitt EM, Halachmi Terenius L and Wiman KG. (1995). Nucl. Acid. Res., 23, S, Bronson RT and Weinberg RA. (1994). Cur. Biol., 4, 1± 362 ± 369. 7. Bakalkin G, Yakovleva T, Selivanova G, Magnusson KP, Jayaraman L and Prives C. (1995). Cell., 81, 1021 ± 1029. Szekely L, Kiseleva E, Klein G, Terenius L and Wiman Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and KG. (1994). Proc. Natl. Acad. Sci. USA, 91, 413 ± 417. Craig RW. (1991). Cancer Res., 51, 6304 ± 6311. Barlow C, Hirotsune S, Paylor R, Liyanage M, Eckhaus M, Kastan MB, Zhan Q, El-Deiry WS, Carrier F, Jacks T, Collins F, Shiloh Y, Crawley NJ, Ried T, Tagle D and Walsh WV, Plunket BS, Vogelstein B and Fornace AJ. Wynshaw-Boris A. (1996). Cell, 86, 159 ± 171. (1992). Cell., 71, 587 ± 597. BellocF,DumainP,BoisseauMR,JalloustreC,ReiersJ, Kaufmann W, Levedakou E, Grady H, Paules R and Stein Bernard P and Lacombe F. (1994). Cytometry, 17, 59 ± 65. G. (1995). Cancer Res., 55, 7 ± 11. Bogue MA, Zhu C, Aguilar-Cordova E, Donehower LA and Kuerbitz SJ, Plunkett BS, Walsh WV and Kastan MB. Roth DB. (1996). Genes Dev., 10, 553 ± 565. (1992). Proc. Natl. Acad. Sci. USA, 89, 7491 ± 7495. Bouer SD, Kemp CJ, Balmain A and Cox R. (1995). Kulesz-Martin MF, Lisafeld B, Huang H, Kisiel ND and Lee Cancer Res., 55, 3883 ± 3889. L. (1994). Mol. Cell. Biol., 14, 1698 ± 1708. Brain R and Jenkins JR. (1994). Oncogene, 9, 1775 ± 1780. Lane DP. (1992). Nature, 358, 15 ± 16. Chen I-T, Smith ML, O'Connor PM and Fornace AJJ. Lee JM, Abramson JL, Kandel R, Donehower LA and (1995). Oncogene, 11, 1931 ± 1937. Bernstein A. (1994). Oncogene, 9, 3731 ± 3736. Clarke AR, Purdie CA, Harrison DJ, Morris RG, Bird CC, Lee JM and Bernstein A. (1993). Proc. Natl. Acad. Sci. USA, Hooper ML and Wyllie AH. (1993). Nature, 362, 849 ± 90, 5742 ± 5746. 852. Lee S, Elenbaas B, Levine A and Grith J. (1995). Cell., 81, Cross SM, Sanchez CA, Morgan CA, Schimke MK, Ramel 1013 ± 1020. S, Idzerda RL, Raskind WH and Reid BJ. (1995). Science, Livingstone LR, White A, Sprouse J, Livanos E, Jacks T and 267, 1353 ± 1356. Tisty TD. (1992). Cell., 70, 923 ± 935. Danova M, Giordano M, Mazzini G and Riccardi A. (1990). Lowe SW and Ruley HE. (1993). Genes. and Dev., 7, 535 ± Leukemia Res., 14, 417 ± 422. 545. Darzynkiewicz Z. (1994). Cell Biology Laboratory Hand- Lowe SW, Schmitt EM, Smith SW, Osborne BA and Jacks T. book, 1, 261 ± 271. (1993). Nature, 362, 847 ± 853. Darzynkiewicz Z, Li X and Gong J. (1994). Methods in Cell Maltzman W and Czyzyk L. (1984). Mol. Cell. Biol., 4, Biology, 41, 15 ± 38. 1689 ± 1694. Di Leonardo A, Linke SP, Clarkin K and Wahl GM. (1994). McDonald III ER, Wu GS, Waldman T and El-Deiry WS. Genes & Dev., 8, 2540 ± 2551. (1996). Cancer Res., 56, 2250 ± 2255. Donehower LA, Harvey M, Slagle BT, McArthur MJ, Nelson WG and Kastan MB. (1994). Mol. Cell. Biol., 14, Montgomery CA, Butel JS and Bradly A. (1992). 1815 ± 1823. Nature, 356, 215 ± 221. Nusse M, Beisker W, Kramer J, Miller BM, Schreiber GA, Dutta A, Ruppert JM, Aster JC and Wincherester E. (1993). Viaggi S, Weller EM and Wessels JM. (1994). Meth. Cell. Nature, 365, 79 ± 82. Biol., 42, 149 ± 158. El-Deiry WS, Harper JW, O'Connor PM, Velculescu VE, Nusse M, Kramer J and Miller BM. (1992). Int. J. Radiat. Canman CE, Jackman J, Pietenpol JA, Burrell M, Hill DE Biol., 62, 587 ± 602. and Wang Y. (1994). Cancer Res., 54, 1169 ± 1174. Pan ZQ, Reardons JT, Li L, Florez-Rozas H and Legerski R. Enoch T and Norbury C. (1995). TIBS, 20, 426 ± 434. (1995). J. Biol. Chem., 270, 22008 ± 22016. Ford JM and Hanawalt PC. (1995). Proc. Nat. Acad. of Sci. Paules R, Levedakou E, Wilson S, Innes C, Rhodes N, Tisty USA, 92, 8876 ± 8890. T, Galloway D, Donehower L, Tainsky M and Kaufmann Fritsche M, Haessler C and Brander G. (1993). 8, 307 ± 318. W. (1995). Cancer Res., 55, 1763 ± 1773. Gannon JV, Greaves R, Iggo R and Lane DP. (1990). EMBO Peled A, Schwartz D, Elkind NB, Wolkowicz R, Li R and J., 9, 1595 ± 1602. Rotter V. (1996). Oncogene, 13, 1677 ± 1685. p53andtheG2 checkpoint D Schwartz et al 2607 Powell S, Defrank J, Connell P, Eogan M, Preer F, Wang XW, Forrester K, Yeh H, Feitelson MA, Gu J and Dombkowski D, Tang W and Friend S. (1995). Cancer Harris C. (1994). Proc. Natl. Acad. Sci. USA, 91, 2230 ± Res., 55, 1643 ± 1648. 2234. Remvikos Y, Lauret-Puig P, Salmon RJ, Frelat G, Wang XW, Vermwulen W, Coursen JD, Gibson M, Lupold Dutrillaux B and Thomas G. (1990). Int. J. Cancer, 45, SE, Forrester K, Xu G, Elmore L, Yeh H, Hoeijmakers 450 ± 456. JHJ and Harris CC. (1996). Gen. and Devel., 10, 1219 ± Rogel A, Poplinker M, Webb SG and Oren M. (1985). Mol. 1232. Cell. Biol., 5, 2851 ± 2855. Wang XW, Yeh H, Schaeer L, Roy R, Moncoli V, Egly J- Sah VP, Attardi LD, Mulligan GJ, Williams BO, Bronson M, Wang Z, Friedberg EC, Evans MK, Tae BG, Bohr RT and Jacks T. (1995). Nat. Genetics, 10, 175 ± 180. VA, Weeda G, Hoeijmakers JHJ, Forrester K and Harris Schmid P, Lorenz A, Hameister H and Montenarh M. CC. (1995). Nat. Genetics, 10, 188 ± 195. (1991). Development, 113, 857 ± 865. Wells WAE. (1996). Trends in Cell Biol., 6, 228 ± 234. Schrieber GA, Beisker W, Braselmann H, Bauchinger M, Wolf D, Harris N, Gold®nger N and Rotter V. (1985). Mol. Bogl KW and Nusse M. (1992). Int. J. Radiat. Biol., 62, Cell. Biol., 5, 127 ± 132. 695 ± 709. Wu L, Bayle H, Elenbaas B, Pavletich NP and Levine AJ. Shaulian E, Haviv I, Shaul Y and Oren M. (1994). Oncogene, (1995). Mol. Cell Biol., 15, 497 ± 504. 10, 671 ± 680. Yamaizumi M and Sugano T. (1994). Oncogene, 9, 2775 ± Shaulsky G, Gold®nger N, Peled A and Rotter V. (1991). 2784. Proc. Natl. Acad. Sci. USA, 88, 8982 ± 8986. Yewdell JW, Gannon JV and Lane DP. (1986). J. Virol., 59, Stewart N, Hicks GG, Paraskevas F and Mowat M. (1994). 444 ± 452. Oncogene, 10, 109 ± 115. Yin Y, Tainsky MA, Bischo FZ, Strong LC and Wahl GM. Sturzbecher H-W, Donzelmann B, Henning W, Knippschild (1992). Cell, 70, 937 ± 948. U and Buchhop S. (1996). EMBO J., 15, 1992 ± 2002. Yonish-Rouach E, Grunwald D, Wilder S, Kimchi A, May Symonds H, Krall L, Remington L, Saenz-Robles M, Lowe E, Lawrence JJ, May P and Oren M. (1993). Mol. Cell. S, Jacks T and Van Dyke T. (1994). Cell, 78, 703 ± 711. Biol., 13, 1415 ± 1423. Unger C, Kress S, Buchmann A and Schwarz M. (1994). Yonish-Rouach E, Resnitzky K, Lotem J, Sachs L, Kimchi A Cancer Res., 54, 3651 ± 3655. and Oren M. (1991). Nature, 352, 345 ± 347.