E2F Mediates Sustained G2 Arrest and Down-regulation of Stathmin and AIM-1 Expression in Response to Genotoxic Stress*

Exposure of cells to genotoxic agents results in activation of checkpoint pathways leading to cell cycle arrest. These arrest pathways allow repair of damaged DNA before its replication and segregation, thus preventing accumulation of mutations. The tumor suppressor retinoblastoma (RB) is required for the G1/S checkpoint function. In addition, regulation of the G2 checkpoint by the tumor suppressor p53 is RB-dependent. However, the molecular mechanism underlying the involvement of RB and its related proteins p107 and p130 in the G2 checkpoint is not fully understood. We show here that sustained G2/M arrest induced by the genotoxic agent doxorubicin is E2F-dependent and involves a decrease in expression of two mitotic regulators, Stathmin and AIM-1. Abrogation of E2F function by dominant negative E2F abolishes the doxorubicin-induced down-regulation of Stathmin and AIM-1 and leads to premature exit from G2. Expression of the E7 papilloma virus protein, which dissociates complexes containing E2F and RB family members, also prevents the down-regulation of these mitotic genes and leads to premature exit from G2 after genotoxic stress. Furthermore, genotoxic stress increases the levels of nuclear E2F-4 and p130 as well as their in vivo binding to the Stathmin promoter. Thus, functional complexes containing E2F and RB family members appear to be essential for repressing expression of critical mitotic regulators and maintaining the G2/M checkpoint.

Cell cycle arrest in response to DNA damage is an important mechanism for maintaining genomic integrity. This cell cycle arrest provides time for DNA repair to prevent replication or segregation of damaged DNA. Induction of growth arrest by DNA damage occurs mainly through the activation of checkpoint pathways that delay cell cycle progression at G 1 , S, and G 2 (1,2).
Growth arrest at both G 1 and G 2 is believed to occur in two steps, resulting in a rapid arrest that is followed by a more sustained arrest. Initial arrest at G 1 involves phosphorylation and degradation of both the protein phosphatase cdc25A and cyclin D1, resulting in inhibition of G 1 cyclin-dependent kinases (3)(4)(5). Initial arrest at G 2 involves phosphorylation and inhibition of the protein phosphatase cdc25C, leading to inhibition of cdc2 activity (6,7).
Through transactivation of p21, the tumor suppressor p53 is one of the essential mediators of sustained arrest at both G 1 and G 2 in response to DNA damage (8 -11). p21 binds to cyclincyclin-dependent kinase complexes and inhibits their ability to phosphorylate the retinoblastoma tumor suppressor, RB, 1 and its related proteins, p107 and p130 (12). Cells deficient of RB fail to arrest at G 1 after DNA damage, indicating that RB plays an essential role in this arrest (8,13). In addition, p53 regulation of DNA damage-induced G 2 arrest was shown to be RB-dependent (14).
RB, p107, and p130 play a key role in negative regulation of cell cycle progression, and their growth inhibitory activity is largely attributed to their association with members of the E2F family of transcription factors (15,16). The E2F family is composed of six members, E2F-1-E2F-6, which heterodimerize with the DP proteins, DP-1 or DP-2, to form the DNA-binding, active transcription factor (15,16). E2F plays a crucial and well established role in the control of cell cycle progression mainly by up-regulating expression of genes required for the G 1 /S transition as well as for DNA replication (15,16).
This transcriptional activity of E2F is inhibited by its interaction with RB, p107, and p130 (15,16). In addition, the complex containing E2F and RB family members also actively represses transcription. Assembly of such repressive complexes, containing E2F and RB family members (referred to herein as E2F-RB, although they may contain p107 or p130), on promoters that have E2F-binding sites is critical for growth suppression by RB family members (17,18). The combination of cessation of repression of some E2F-regulated genes by the E2F-RB complex and the activation of others by activated E2F constitutes a major step in promoting G 1 exit.
Recent studies indicate that E2F regulates the expression of genes involved not only in cell cycle progression but also in other biological processes (33)(34)(35)(36). Many of these novel E2F targets function in various cellular responses to DNA damage, including activation of checkpoints, DNA repair, and apoptosis, thus implicating E2F in the DNA damage response. Further support to this notion comes from recent reports demonstrating that the E2F-1 protein is stabilized and its levels are increased following DNA damage (37)(38)(39)(40). This stabilization is due to phosphorylation of E2F-1 by the protein kinase ATM, one of the master controllers of the response to DNA damage (41).
We and others have shown that E2F up-regulates expression of a number of genes involved in entry to and progression through mitosis (33,35). However, the mode of regulation of these mitotic genes by E2F and the biological consequences of this regulation are not fully understood. We show here that expression of two of these mitotic genes, AIM-1 and Stathmin, is also elevated by transcriptionally inactive E2F-1, indicating that they are subjected to E2F-dependent repression. Furthermore, we show that E2F-containing complexes are required for DNA damage-induced down-regulation of AIM-1 and Stathmin. This repression of gene expression is correlated with E2F-dependent maintenance of DNA damage-induced growth arrest at G 2 .

EXPERIMENTAL PROCEDURES
Cell Culture-NIH3T3 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum.
Rat-1a-MT-wtE2F-1 and Rat-1a-MT-E2F-1dlTA, which are the Rat1 cell lines transfected with an inducible plasmid expressing either wild type of a dominant negative mutant of E2F-1, were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and G418 (500 g/ml). For zinc induction, cells were maintained for 48 h in medium with 0.1% serum, and then 100 M ZnCl 2 were added to the medium. 293 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum.
All cells were maintained at 37°C in a humidified 8% CO 2containing atmosphere.
Activation of the E2F-1 fused to estrogen receptor ligand-binding domain (ER-E2F-1) was induced by addition of 4-hydroxytamoxifen to a final concentration of 30 nM.
Cycloheximide was added to a final concentration of 10 g/ml. Doxorubicin was added to a final concentration of 0.2 g/ml.
Retroviral Infection-Cells of the packaging cell line 293 were cotransfected with 10 g of ecotropic packaging plasmid, pSV--E-MLV, and 10 g of the relevant plasmid using the calcium phosphate method in the presence of chloroquin (25 M final concentration, Sigma C6628). After 8 h, the transfection medium was replaced with fresh medium, and 5 ml of retroviral-containing cell supernatant was collected at 6-h intervals. Five collections were pooled together and frozen in aliquots. For infection, NIH3T3 cells were incubated for 5 h at 37°C in 3 ml of retroviral supernatant, supplemented with 8 g/ml polybrene (Sigma H9268). Then, 7 ml of medium was added, and after 24 h the medium was replaced with fresh medium containing 10% serum and 2 g/ml puromycin (Sigma P7130).
Fractionation and Western Blotting-For nuclear and cytoplasmic fractions, cell pellets were resuspended in four packed cell volumes of hypotonic buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 20 g/ml aprotinin, and 10 g/ml leupeptin) and incubated for 15 min on ice. Cells were then lysed by adding Nonidet P-40 to a final concentration of 0.6% and vortexing. After a short centrifugation, the cytoplasmic supernatant was taken out, and the nuclear pellet was lysed in two packed cell volumes of lysis buffer (20 mM HEPES, pH 7.9, 400 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 20 g/ml aprotinin, and 10 g/ml leupeptin). Protein concentrations were determined by Bradford assay. The ratio of loaded volumes of cytoplasmic and nuclear extracts from a given cell population was equivalent to the ratio of volumes of hypotonic and lysis buffers, respectively, and therefore it is considered as the per cell ratio of cytoplasmic and nuclear proteins. For whole cell extrcats, cells were lysed in lysis buffer (20 mM HEPES, pH 7.8, 450 mM NaCl, 25% glycerol, 0.2 mM EDTA, 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 20 g/ml aprotinin, and 1 g/ml leupeptin). Equal amounts of protein from each lysate, as determined by Bradford assay, were resolved by electrophoresis through an SDS 7.5-12.5% polyacrylamide gel and transferred to a filter (Protran BA 85, Schleicher & Schü ll). Filters were incubated with a primary antibody for 2 h, after 1 h blocking in PBS with 0.05% Tween 20 and 5% dry milk. Primary antibodies used were as follows: anti E2F-1 (sc-251, Santa Cruz), anti-RB (14001A, PharMingen), anti-p130 (sc-317, Santa Cruz), anti-AIM-1 (BD Biosciences), anti-Stathmin (STC, gift of Andre Sobel), anti-E2F-4 (sc-866 Santa Cruz), and anti-B23 (sc-6013, Santa Cruz). Binding of the primary antibody was detected using an enhanced chemiluminescence kit (ECL, Amersham Biosciences).
Cell Cycle Flow Cytometry Assays-Cells were trypsinized and fixed with methanol (Ϫ20°C). After fixation, cells were centrifuged for 5 min at 1200 rpm, resuspended in PBS and incubated for 30 min at 4°C. After recentrifugation cells were resuspended in PBS containing 5 g/ml propidium iodide and 50 g/ml RNase A and incubated for 30 min at room temperature. Fluorescence intensity was analyzed using a BD Biosciences flow cytometer.

RESULTS
Our initial studies aimed at understanding the regulation of mitotic genes by E2F focused on one of these E2F-regulated mitotic genes, Stathmin (also known as oncoprotein 18), which encodes a protein involved in microtubule dynamics and spindle assembly (44). To determine whether Stathmin is a direct target of E2F, we infected NIH3T3 cells with a retrovirus carrying E2F-1 fused to the estrogen receptor ligand-binding domain (ER-E2F-1). The ER-E2F-1 is expressed as an inactive fusion protein, which is activated upon addition of the ligand 4-hydroxytamoxifen (24). As was previously reported by Ishida et al. (33) induction of E2F-1 led to an increase in Stathmin mRNA levels (Fig. 1A). Interestingly, a similar E2F1-induced increase in Stathmin mRNA levels was detected in the pres-ence of the protein synthesis inhibitor, cycloheximide (CHX) (Fig. 1A). These data indicate that de novo protein synthesis is not required for E2F1-induced up-regulation of the Stathmin mRNA, suggesting that Stathmin is a direct target of E2F.
The observed E2F1-induced up-regulation of Stathmin may be due to either activation or derepression. To distinguish between these two possibilities we analyzed the expression of Stathmin in cells containing an inducible wild type E2F-1, or mutant E2F-1, E2F-1dlTA. This truncated E2F-1dlTA lacks both the transactivation and the RB-binding domains, and when overexpressed it can negate both activation of gene expression by E2F and repression by E2F-RB complexes. Once again, induction of wtE2F-1 resulted in an increase in Stathmin mRNA levels (Fig. 1B). Importantly, induction of E2F-1dlTA led to a similar increase in Stathmin mRNA levels (Fig.  1B). wt and mutant E2F-1 were expressed at comparable levels ( Fig. 1C). Expression of other E2F-regulated mitotic genes, including Cdc2, BUB1b, EB1, and SAK-a, was similarly up-regulated by both wtE2F-1 (35) and E2F-1dlTA (data not shown). These data strongly suggest that Stathmin, as well as other mitotic genes, is under E2F-dependent negative regulation.
A physiological setting in which such negative regulation may be important is during growth arrest in response to DNA damage.
Low doses of DNA-damaging agents induce cellular growth arrest, and we analyzed the effects of such a treatment on expression of both Stathmin and another E2F-regulated mitotic gene, AIM-1. AIM-1(aurora and Ipl-1-like midbody-associated protein kinase, also called aurora1 or STK12) is a serine/ threonine kinase required for cytokinesis (reviewed in Refs. 45 and 46). Following addition of the chemotherapeutic agent doxorubicin, NIH3T3 cells underwent growth arrest at the G 1 and G 2 /M phases of the cell cycle ( Fig. 2A). This DNA damageinduced growth arrest was already apparent 24 h after doxorubicin addition and persisted for at least 72 h. The growth arrest was accompanied by accumulation of hypophosphorylated RB and p130 (Fig. 2B). In addition, treatment with doxorubicin led to a significant decrease in mRNA levels of Stathmin and AIM-1 (Fig. 2C). Similar results were obtained using HCT116 human colorectal carcinoma cells (data not shown). The decrease in mRNA levels of the studied genes could be detected 24 h after treatment and was most evident at 48 h (Fig. 2C). Thus, the increase in hypophosphorylated RB and p130, which are the growth repressive forms, coincided with or preceded the down-regulation of Stathmin and AIM-1. This result raises the possibility that complexes containing either RB or p130 mediate the repression of Stathmin and AIM-1 in response to DNA damage.
Next we tested whether E2F mediates the decrease in expression of these mitotic genes in response to DNA damage. To this end, NIH3T3 cells were infected with retroviruses containing either E2F-1dlTA or an empty vector and then treated with doxorubicin.
As shown earlier for uninfected cells (Fig. 2), in vector-infected NIH3T3 cells, the doxorubicin-induced growth arrest was accompanied by a significant increase in levels of hypophosphorylated RB and p130 (Fig. 3B). In addition, as shown earlier for uninfected cells, treatment of empty vector-infected NIH3T3 cells with doxorubicin led to a decrease in mRNA levels of the E2F-regulated genes, AIM-1 and Stathmin (Fig.  3A). A similar doxorubicin-induced decrease was detected in protein levels of AIM-1 and Stathmin (Fig. 3B). Expression of E2F1-dlTA did not significantly affect the doxorubicin-induced accumulation of hypophosphorylated RB and p130; however, it abolished the decrease in mRNA and protein levels of AIM-1 and Stathmin (Fig. 3, A and B). Thus, in cells expressing E2F1-dlTA, levels of AIM-1 and Stathmin remained unchanged throughout the experiment.
Following doxorubicin addition, cells infected with an empty vector, similarly to uninfected cells, arrested at the G 1 and G 2 /M phases of the cell cycle. This arrest at G 1 and G 2 /M persisted for at least 96 h (Fig. 3C). Expression of E2F-1dlTA did not result in noticeable changes in cell cycle distribution of untreated cells; however, it had profound effects on cell cycle distribution after treatment with doxorubicin. Cells expressing E2F-1dlTA failed to arrest at G 1 and accumulated at G 2 /M. Interestingly, growth arrest at G 2 /M was not maintained, the percentage of cells at G 2 /M gradually decreased, and a concomitant increase in cells with Ͻ2 N and Ͼ4 N DNA content was detected (Fig. 3C). These data suggest that while E2F is not essential for the initiation of G 2 /M arrest, it is required for its maintenance.
To study more directly the role of E2F-RB complexes in the response to DNA damage, we tested the effect of their dissociation by the papilloma virus E7 protein. To this end, NIH3T3 cells were infected with a retrovirus containing either an empty vector, the wtE7 gene of HPV16, or a mutated E7, E7⌬21-35, which does not bind RB family members. The wt and mutated E7 were expressed at similar levels (data not shown); however, expression of wtE7, but not E7⌬21-35, inhibited the doxorubicin-induced decrease in protein levels of AIM-1 and Stathmin (Fig. 4A). In addition, expression of wtE7 led to profound changes in cell cycle distribution after doxorubicin treatment (Fig. 4B). The effects of wtE7 expression were highly similar to those of E2F-1dlTA and included: 1) failure of cells to arrest at G 1 and their accumulation at G 2 /M; 2) inability to maintain growth arrest at G 2 /M, with a gradual decrease in the percentage of cells at G 2 /M and an increase in cells with Ͻ2 N and Ͼ4 N DNA content (Fig. 4B). Expression of E7⌬21-35 had no apparent effect on cell cycle distribution after DNA damage (Fig. 4B). These findings suggest that endogenous E2F-RB complexes mediate repression of mitotic genes and sustained G 2 /M arrest in response to DNA damage.
To identify the distinct members of the E2F and RB families that mediate the repression of E2F-regulated genes in response   FIG. 1. Stathmin is a direct E2F target, and its expression is  up-regulated by transcriptionaly inactive E2F-1. A, NIH3T3 cells were infected with a retrovirus containing E2F-1 fused to the estrogen receptor ligand-binding domain (ER-E2F-1). After selection, the cells were kept in medium containing 0.5% bovine calf serum for 48 h. Then the fusion protein was activated by the addition of 4-hydroxytamoxifen (OHT) for the indicated times (in hours), in the absence or presence of cycloheximide (ϩCHX). Total RNA was extracted and used for RT-PCR performed with specific primers for the Stathmin and GAPDH genes. B, rat fibroblasts containing inducible wtE2F-1 (E2F-1) or an inducible E2F-1 mutant that lacks the transactivation and RB-binding domains (E2F-1dlTA) were kept in medium with 0.1% fetal calf serum for 48 h. Then E2F expression was induced with 100 M ZnCl 2 for 12 h (ϩ) or not induced (Ϫ) prior to extraction of total RNA. RT-PCR was performed on the total RNA using specific primers for the Stathmin and ARPP-PO genes. C, cells were treated as described in B and then lysed. Protein extracts were subjected to Western blot analysis using an anti-E2F-1 antibody.
to DNA damage we performed a ChIP using antibodies directed against specific family members. The human Stathmin promoter contains three putative E2F-binding sites at positions Ϫ28, Ϫ577, and Ϫ701 upstream to the transcription start site (47), and our analysis of the murine Stathmin promoter indicates that it too contains three putative E2F-binding sites. Taken together with our observation that E2F-induced up-regulation of Stathmin does not require de novo protein synthesis (Fig. 1A), this sequence information suggests that E2Fs may interact with the Stathmin promoter. Therefore, we analyzed this promoter for occupancy by E2F and RB family members. We observed a significant enrichment of the Stathmin promoter when using E2F-4 or p130 antibodies (Fig. 5A). We did not detect any enrichment after amplification of an unrelated genomic DNA fragment (Fig.  5A). The enrichment of the Stathmin promoter when using p130 and E2F-4 antibodies was more prominent in cells treated with   E2F-1dlTA). After selection infected cells were treated with doxorubicin (DOX) for the indicated times (in hours). Then cells were collected for RNA and protein extraction (A and B, respectively) and for FACS analysis (C). A, total RNA was used for RT-PCR analysis using specific primers for the AIM-1, Stathmin, and GAPDH genes. B, protein extracts were used for Western blot analysis with anti-AIM-1, Stathmin, RB, and p130 antibodies. doxorubicin (Fig. 5B), indicating an increase in the occupancy of this promoter by E2F-4 and p130 after genotoxic stress. In agreement with this observation, we detected an increase in levels of E2F-4 and hypophosphorylated p130 in the nuclei of cells treated with doxorubicin (Fig. 5C). Taken together, our data indicate that endogenous E2F-RB complexes downregulate expression of the mitotic genes AIM-1 and Stathmin in response to DNA damage. Furthermore, our findings strongly suggest that E2F-RB complexes play a role in the maintenance of DNA damage-induced G 2 /M arrest. DISCUSSION Recent screens aimed at identifying novel genes regulated by E2F suggest that E2F modulates the expression of a number of genes involved in entry to and progression through mitosis as well as genes that affect the G 2 /M checkpoint (33,35,36). We show here that two pivotal mitotic regulators, Stathmin and AIM-1, are subjected to E2F-mediated negative regulation, and their expression is elevated by a dominant negative mutant of E2F-1 that lacks the transactivation and RB-binding domains.  5. E2F-4 and p130 bind to the Stathmin promoter in response to DNA damage. A, NIH3T3 cells were treated with doxorubicin (DOX) for 48 h, and then ChIP assay was performed. Cross-linked chromatin was incubated with antibodies against E2F-4 (␣E2F-4), p130 (␣p130), or without antibody (no Ab). Immunoprecipitates from each sample were analyzed by PCR using primers specific for the Stathmin promoter (Stathmin) and for ␤-actin coding region (␤-actin). As a control, a sample representing 0.2% of the total input chromatin (input) was included. B, NIH3T3 cells were treated with doxorubicin (DOX) for 48 h or not treated, and ChIP assay was performed as in A. Relative occupancy represents binding of E2F-4 and p130 to the Stathmin promoter relative to ␤-actin control. C, NIH3T3 cells were treated with doxorubicin (DOX) for 48 h or not treated. Nuclear and cytoplsamic protein extracts were prepared and used for Western blot analysis with anti E2F-4, p130, ␣-tubulin, and B23 antibodies. ␣-Tubulin and B23 served as cytoplasmic and nuclear controls, respectively. Stathmin/oncoprotein 18 is a conserved cytoplasmic phosphoprotein that physically interacts with tubulin dimers (reviewed in Ref. 44). It is a critical regulator of microtubule dynamics during cell cycle progression and governs preferential microtubule growth around chromosomes during spindle assembly (48). Stathmin is expressed at elevated levels in various human tumors (49,50), and it is expressed at higher levels in proliferating cells as compared with non-proliferating cells (51). The Stathmin promoter contains three E2F-binding sites (47), and a recent study indicates that Stathmin levels are regulated by E2F (33). Importantly, antisense inhibition of Stathmin expression results in growth arrest and accumulation of cells in G 2 /M (52,53). Conversely, overexpression of Stathmin abrogates irradiation-induced G 2 /M arrest. Thus, Stathmin appears to have an essential role in the G 2 /M arrest mechanism (54).
The second E2F-regulated gene studied here, AIM-1, is an Aurora/Ipl1p-related serine/threonine kinase that is required for cytokinesis (reviewed in Refs. 45 and 46). Regulation of AIM-1 expression is not well characterized, and its promoter has not been studied extensively. Analysis of genomic sequences upstream to human AIM-1 coding sequence reveals an E2F-binding site at Ϫ61 upstream to the putative transcription start site, which is conserved in mouse AIM-1. Furthermore, E2F-4 was shown to interact with the human AIM-1 promoter (36).
AIM-1 expression is cell cycle-regulated, and its mRNA and protein accumulate at G 2 /M (55-57). The kinase activity of AIM-1 is also cell cycle-regulated with peak activity at the M phase (58). The intracellular localization of AIM-1 is consistent with a role in the later stages of mitosis as it is found in the central spindles of anaphase cells and at the midbody of telophase cells (56,58). Both expression of a kinase-inactive form of AIM-1 and overexpression of wt AIM-1 block cytokinesis and lead to polyploidity, indicating that proper AIM-1 expression and activity is critical for cytokinesis (55,56).
The results presented here demonstrate that expression of AIM-1 and Stathmin is repressed in response to DNA damage. We show that this repression is abrogated by dominant negative E2F-1 as well as by dissociation of E2F-RB complexes, indicating that such complexes play a critical role in the regulation of these genes. In support of this notion, we detected in vivo binding of E2F-4 and p130 to the Stathmin promoter. Importantly, DNA damage leads to an increase in levels of nuclear E2F-4 and p130 and enhancement of their binding to the Stathmin promoter. Of note, overexpression of p53 was shown to increase in vitro binding of E2F-4 and p130 to the promoter of cdc2, another E2F-regulated mitotic gene (59).
Our data also implicate E2F in maintenance of induced G 2 /M arrest and indicate that functional E2F-RB complexes are required to sustain G 2 /M arrest following genotoxic stress. Treatment of unmanipulated cells and cells expressing mutated E7 (that does not dissociate E2F-RB complexes) with doxorubicin resulted in a sustained arrest at G 1 and G 2 /M. In sharp contrast, doxorubicin treatment of cells expressing either E7 (which dissociates E2F-RB complexes) or E2F-1dlTA (which competes with E2F-RB complexes for DNA binding) led to a transient G 2 /M arrest. Furthermore, in cells expressing E7 or E2F-1dlTA, the premature exit from G 2 after genotoxic stress was accompanied by appearance of cells with Ͼ4 N DNA content, indicating that loss of functional E2F and E2F-RB complexes may result in endoreduplication. This is in agreement with the suggested role of AIM-1 in maintaining cell diploidity and with previous studies demonstrating that lack of RB is correlated with increased endoreduplication (60). The data presented here further support a role for endogenous E2F in pre-venting endoreduplication after DNA damage.
Overall, the data presented here demonstrate that E2F-RB complexes play an important role both in maintaining G 2 /M arrest and in repression of two mitotic genes after DNA damage. The complete panel of genes that are repressed by E2F-RB complexes and their relative contribution to sustained G 2 /M arrest remain to be determined. Cdc2 and cyclin B1 are two E2F-regulated genes (32, 33) whose promoter activity was shown to be down-regulated upon sustained G 2 /M arrest or p53 activation (59,61,62). Protein levels of cdc2 and cyclin B1 as well as cyclin B1/cdc2 kinase activity also decrease in association with sustained G 2 /M arrest after DNA damage (14,59,61), and this decrease is abrogated by E7, implicating the RB family in their negative regulation. These data strongly suggest that upon DNA damage, expression of both cyclin B1 and cdc2 is repressed by E2F-RB complexes. However, there is conflicting data regarding the dependence of sustained G 2 /M arrest on reduction of cyclin B1/cdc2 activity. While one study demonstrated that constitutive activation of cyclin B1/cdc2 kinase activity overrides p53-mediated G 2 /M arrest (63), others detected prolonged G 2 /M arrest after DNA damage even in cells with high cyclin B1/cdc2 activity (61). Thus, additional genes repressed by E2F-RB complexes probably play a pivotal role in sustaining G 2 /M arrest. A number of additional E2F-regulated genes, including cyclin A, thymidine kinase, topoisomerase II␣, and RAD51 are repressed upon DNA damage (64). We show here that the E2F-dependent decrease in expression of two genes involved in mitosis, AIM-1 and Stathmin, is associated with sustained G 2 /M arrest after DNA damage. It is tempting to speculate that sustaining G 2 arrest after DNA damage involves the concerted, E2F-mediated, repression of a large panel of genes.