J Biol Chem, Vol. 275, Issue 3, 1715-1722, January 21, 2000

A Novel DNA Damage Checkpoint Involving Post-transcriptional Regulation of Cyclin A Expression^*

Ning Guo, Douglas V. Faller, and Cyrus Vaziri

From the Cancer Research Center, Boston University School of Medicine, Boston Massachusetts 02118

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The intracellular metabolism of many carcinogenic polycyclic aryl hydrocarbons (PAHs, typified by the ubiquitous pollutant benzo[a]pyrene or B[a]P) generates electrophilic products that react covalently with genomic DNA. Cells that acquire PAH-induced DNA damage undergo growth arrest in a p53-independent manner (Vaziri, C., and Faller, D. V. (1997) J. Biol. Chem. 272, 2762-2769). In this report we have investigated the molecular basis of PAH-induced cell cycle arrest. Mitogenic signaling events involving cyclins D and E, Rb phosphorylation, and transcriptional activation of E2F-responsive genes (including cyclin E and cyclin A) were unaffected in cells containing PAH-damaged DNA. However, PAH-induced growth arrest was associated with post-transcriptional decreases in cyclin A expression. Mitogen-induced expression of cyclin B, an event that is temporally distal to cyclin A expression, was also inhibited in PAH-treated cells. The PAH-induced cell cycle block was transient, and arrested cells resumed DNA synthesis after a prolonged (~20 h) delay. Resumption of DNA synthesis in PAH-treated cells occurred concomitant with elevated expression of cyclins A and B. PAH-induced cell cycle arrest was overcome by ectopically expressed cyclin A (encoded by a recombinant adenovirus in transiently infected cells). Overall, our results suggest the existence of a DNA damage checkpoint pathway that arrests cell cycle progression via post-transcriptional control of cyclin A expression.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Polycyclic aryl hydrocarbons (PAHs),¹ typified by the common pollutant benzo[a]pyrene (B[a]P), are widespread and persistent environmental contaminants (1). PAHs are potent mutagens and carcinogens in experimental animals and transform rodent and human cells to malignancy in culture. (2). Epidemiological studies have shown strong correlations between human exposure to PAHs and susceptibility to certain cancers (3). For example, PAH-rich tobacco smoke is considered to be a causative factor for lung cancer in smokers (4).

Many PAHs, including B[a]P, are able to bind to and activate a ligand-dependent transcription factor termed the aryl hydrocarbon receptor or AhR (5). When activated by appropriate ligands, AhR translocates to the nucleus and dimerizes with a related protein termed aryl hydrocarbon receptor nuclear translocator, or ARNT (5). Ligand-activated AhR/ARNT heterodimers bind to specific nucleotide sequences in the promoter regions of AhR-responsive genes (termed xenobiotic response elements or XREs). Xenobiotic response elements are present in the promoter regions of genes encoding cytochromes P 450 1A1 and 1A2 (5, 6). These cytochromes are said to be "substrate-inducible," since many of the PAHs that induce AhR-dependent transcription of these genes are also substrates for the oxidative reactions carried out by the cytochrome P-450 proteins. PAHs such as B[a]P are not themselves genotoxic (DNA-damaging). However, the hydroxyl and epoxide species resulting from B[a]P metabolism are highly electrophilic and bind covalently to cellular macromolecules, forming bulky hydrophobic adducts (7). Benzo[a]pyrene diol epoxides (BPDEs) are the major genotoxic (DNA-damaging) species resulting from B[a]P metabolism and are considered to be the "ultimate carcinogens" generated from B[a]P (7). BPDEs bind covalently to the exocyclic amino group of guanosine (N2 position) or react with phosphate to produce phosphotriesters. PAH-induced mutagenesis and carcinogenesis results from DNA damage elicited by "metabolically activated" aryl hydrocarbons (7-9). Potentially, PAH-induced mutations could arise during DNA mis-replication at damaged sites of electrophilic attack or during error-prone DNA repair. Indeed, substitution mutations as well as frameshifts can result from replication of BPDE-adducted DNA. Therefore, multistep chemical carcinogenesis appears to be due to carcinogen-induced mutations in proto-oncogenes and/or tumor suppressor genes (8, 9). High levels of PAH-adducted DNA have been found in neoplastic tissues from cancer patients (10, 11), suggesting that environmental PAHs may indeed contribute to neoplasia in humans.

A crucial requirement for PAH-induced carcinogenesis is that mutations become "fixed" by mis-replication of damaged genes. Since cells are exposed to intrinsic and extrinsic DNA-damaging agents, they have evolved mechanisms to prevent mutagenic mis-replication of damaged DNA. When the integrity of the genome is compromised, checkpoint mechanisms exert negative controls on cell cycle progression (resulting in growth arrest or apoptosis). These controls prevent potentially error-prone DNA synthesis and therefore protect against mutagenesis and genetic instability (12). Individuals with inherited defects in known checkpoint genes (such as ataxia telagiectasia patients with defects in the ATM gene) are prone to cancer, which highlights the importance of checkpoint control mechanisms. We unexpectedly identified a novel PAH-induced and DNA damage-dependent cell cycle checkpoint in experiments in which we tested the effects of aryl hydrocarbons on growth factor signal transduction (13). The PAH-induced checkpoint is p53-independent and therefore fundamentally different from many previously identified checkpoints. PAH-activated cell cycle arrest is likely to guard normal cells against the mutagenic and tumorigenic effects of ubiquitous environmental contaminants (such as B[a]P). Therefore, it is important to understand the molecular basis of this potentially important checkpoint.

In this report, we have extended our initial studies of PAH-induced DNA damage responses and have identified a new mechanism of checkpoint control. Many genotoxic agents perturb the G₁ events that mediate cell cycle progression, thereby resulting in cell cycle arrest prior to S-phase entry (14-16). We initially tested the hypothesis that PAH-induced cell cycle arrest might similarly result from perturbation of mitogenic signal transduction pathways. Accordingly, we analyzed the effects of PAH-induced DNA damage upon the key mitogenic signals that mediate progression of cells through G₁ and into S-phase. Surprisingly, our data show that mitogenic signals involving the G₁ cyclins and cyclin-dependent kinases (CDKs) and the retinoblastoma (Rb)/E2F pathway are unperturbed by PAH-induced DNA damage. This contrasts with the cellular effects of other widely studied genotoxic agents (such as UV and ionizing radiation) that elicit p53-dependent G₁ arrest in an Rb-mediated manner (14-16). Instead, we show that PAH-induced cell cycle arrest occurs after completion of G₁ and involves post-transcriptional changes in the expression of cyclin A, a protein that is known to be rate-limiting for progression through S-phase.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Chemicals-- Unlabeled and [³H]B[a]P were purchased from Sigma and Amersham Pharmacia Biotech, respectively. BPDE was purchased from the NCI Carcinogen Repository. All other materials were purchased from previously specified sources.

Cells and Culture-- Swiss 3T3 cells were purchased from ATCC. Cell lines stably expressing the DHFR promoter-luciferase, cyclin A promoter-luciferase, and cyclin E promoter-luciferase plasmids were generated by co-transfection of the promoter-reporter construct with a vector conferring neomycin resistance (pClneo), followed by selection of stably transfected cells in medium containing G418.

Adenovirus-- To generate AdCycA, the human cyclin A cDNA was subcloned into the pACCMV shuttle vector (17). The resulting vector (pACCycA) and the pJM17 adenovirus plasmid were co-transfected into 293T cells. Recombinant adenovirus clones were isolated by plaque purification (17) and verified by restriction analysis and Southern blotting. AdGFP was derived in the same manner using a shuttle vector (pACGFP) containing the GFP cDNA. AdCon was derived similarly but by co-transfection of "empty" pACCMV shuttle vector with pJM17. Adenovirus particles were purified from 293T cell lysates by polyethylene glycol precipitation, CsCl gradient centrifugation, and gel filtration column chromatography (17).

Mitogenic Assays and Carcinogen Treatment-- Cells were synchronized in G₀ by serum starvation for 48 h as described previously (13). Quiescent cells were stimulated to re-enter the cell cycle by treatment with serum or purified growth factors, as indicated under "Results." Cell cycle progression was monitored by FACScan analysis or using measurements of [³H]thymidine incorporation as described previously (13). To test the effects of PAHs on cell cycle events, the carcinogens were added to quiescent cultures 24 h before treatment with mitogens, unless otherwise noted.

Analysis of B[a]P-adducted DNA-- To quantify DNA adducts, cultures were treated with [¹⁴C]- or [³H]benzo[a]pyrene (Amersham Pharmacia Biotech) for the indicated times. After purification of DNA from the [¹⁴C]benzo[a]pyrene-treated cultures (18), covalently bound radioactivity was measured by scintillation counting.

Transfections and Reporter Gene Analysis-- Transfections were performed using standard calcium phosphate co-precipitation methods (13). Luciferase reporter gene assays were performed as described previously (13).

RNA and Protein Blot Analysis-- RNA and protein samples were prepared as and analyzed (by RNA blotting and protein immunoblotting) exactly as described previously (13).

Cyclin-dependent Kinase Assays-- After immunoprecipitation of cyclin-CDK complexes from cell lysates, washed immune complexes were incubated with histone H1 and [³²P]ATP. Phosphorylated histone was electrophoresed on SDS-polyacrylamide gels and visualized by autoradiography. These experiments were performed exactly as described by Rosenblatt et al. (19).

Reproducibility-- All data are representative of experiments that were repeated at least three times with essentially similar results.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of B[a]P-induced DNA Damage on Expression of Cyclin D, Rb Phosphorylation, and E2F Activity-- We have previously shown that B[a]P-induced DNA damage inhibits progression of synchronized mitogen-treated cells through S-phase (Ref. 13 and Fig. 1A). Growth arrest elicited by many genotoxic agents resulted from negative controls exerted upon G₁ mitogenic events (14-16). We hypothesized that PAH-induced DNA damage may similarly perturb the ordered sequence of mitogen-induced events that normally mediate cell cycle progression. Accordingly, we have analyzed the effect of B[a]P-induced DNA upon mitogen-dependent expression of G₁ cyclins (cyclins D1 and E), Rb phosphorylation, and E2F activation.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Effect of B[a]P on cell cycle progression cyclin D1 expression, Rb phosphorylation, and E2F-dependent reporter gene activity. A, 10-cm plates of near-confluent Swiss 3T3 cells were serum-starved for 72 h. One plate of cells was pretreated with 1 µM B[a]P during the last 24 h of serum starvation. The cells were then given no treatment (control), 7% serum (serum), or 7% serum + 1 µM B[a]P. All plates also received 0.4 mg/ml nocodazole (to prevent re-entry into a subsequent G1 following completion of S-phase). After 25 h, cells were harvested and processed for FACScan analysis as described under "Materials and Methods." B, near-confluent cultures of fibroblasts were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 2 µM B[a]P during the last 24 h of serum starvation. The resulting cultures received no treatment or 7% serum for 15 h. RNA or whole cell extracts were isolated from the cells. Samples were analyzed for cyclin D1 expression by RNA blot analysis with a radiolabeled murine cyclin D1 cDNA (top panel) and by immunoblotting of whole cell extracts with antisera to mouse cyclin D1 (middle panel). Additionally, protein samples were analyzed for Rb phosphorylation status by immunoblotting with anti-Rb antisera. RNA and protein blot analyses were performed as described under "Materials and Methods." C, near-confluent cultures of S-E2F-LUC cells (which contain a stably transfected E2F-responsive reporter construct) were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 2 µM B[a]P during the last 24 h of serum starvation. The cultures then received no treatment or 7% serum for 15 h. Whole cell extracts were prepared from the cells and analyzed for luciferase activity as described under "Materials and Methods."

Mitogen treatment of quiescent cells resulted in transcriptional induction of cyclin D1, which attained peak levels of expression during mid-G₁ (20). When expressed, cyclin D contributed to the activation of CDKs 4 and 6. In turn, these active CDKs were believed to phosphorylate the Rb tumor suppressor protein, enabling the cells to traverse a restriction point termed the "R-point" (21). To test whether B[a]P-induced DNA damage perturbed these G₁ events, we analyzed the expression of cyclin D1 RNA and protein levels and the phosphorylation status of Rb in control and B[a]P-treated cultures of fibroblasts. As expected, serum-treated fibroblasts expressed high levels of cyclin D1 transcripts relative to control (quiescent) cultures (Fig. 1). As shown in Fig. 1B, when B[a]P-treated cultures were stimulated with serum, cyclin D1 transcripts were induced to the same high levels attained in control cultures (which lacked DNA damage). In a parallel experiment, we tested the effects of B[a]P upon serum-induced cyclin D1 protein levels. As we found for cyclin D1 transcript levels, cyclin D1 protein expression was unaffected by concentrations of B[a]P that inhibited cell cycle progression (Fig. 1A). These data indicated that cyclin D1 expression was not rate-limiting for cell cycle progression in cells containing PAH-damaged DNA. Furthermore, this suggested to us that B[a]P might perturb the ordered mitogen-dependent regulation of G₁-signaling events at a stage distal to cyclin D expression.

Cyclin D1 is thought to drive G₁ progression largely by contributing to the activation of CDKs and the resulting phosphorylation of Rb. We considered the possibility that B[a]P (and related PAHs) may prevent mitogen-dependent cell cycle progression by perturbing these processes. To test the effect of PAH treatment upon Rb phosphorylation, we performed immunoblot analysis of Rb protein in whole cell lysates from quiescent, serum-stimulated, and serum plus B[a]P-treated fibroblasts. As expected, mitogenic stimulation resulted in an increase in the amount of slowly migrating, hyperphosphorylated Rb (indicated by the arrow in Fig. 1B, lower panel). Mitogen-treated cells that had received B[a]P showed no change in the amount of hyperphosphorylated Rb relative to cells that were stimulated with serum alone. In the same experiments, replicate cultures received identical treatments but were harvested at a later time for FACScan analysis (Fig. 1A). These control experiments confirmed that the B[a]P-treated cells indeed failed to progress through S-phase. Therefore, Rb phosphorylation was not rate-limiting for S-phase entry under conditions of B[a]P-induced growth arrest.

Rb phosphorylation during late G₁ is known to result in de-repression of the activity of E2F family transcription factors (22). Mitogen-dependent activation of E2F and DP (E2F-related factors, which form transcriptionally active dimers with E2F family members)-regulated genes occurs subsequent to Rb phosphorylation and is thought to be necessary for the G₁/S-phase transition. Indeed, forced expression of E2F is sufficient to drive quiescent cells into S-phase, supporting the idea that E2F plays an essential role in normal cell cycle progression. We performed experiments to test if PAH-induced DNA damage perturbed E2F activity. We stably transfected cells with a heterologous plasmid containing a minimal E2F-responsive promoter linked to a luciferase reporter cDNA (23). Mitogen-dependent transcription of this reporter construct is entirely dependent upon the E2F sites within the promoter (23). The resulting stably transfected Swiss 3T3 cells (designated S.E.2F-LUC) were synchronized in G₀ by serum starvation for 48 h. As in previous experiments, some quiescent cultures were treated with B[a]P (to elicit DNA damage) before serum stimulation. After 15 h of serum treatment (time points distal to Rb phosphorylation and E2F activation), we prepared cytosolic extracts from the cells and assayed these for luciferase activity. As shown in Fig. 1C, serum treatment resulted in a marked increase in E2F-dependent luciferase activity (relative to control quiescent cultures that received no mitogen). Extracts from cells that were treated with B[a]P did not contain decreased luciferase activity. In fact we consistently noticed slight increases in E2F-dependent luciferase activity in the extracts from B[a]P-treated cells. A possible mechanism for this modest enhancement of serum induction will be discussed in a later section below. Nevertheless, these experiments suggested that E2F activity was not limiting for S-phase entry when the cells acquired B[a]P-induced DNA damage.

Effect of B[a]P on Cyclin E Signaling-- The experiments described above showed that the cyclin D/Rb/E2F signaling axis was unaffected by DNA damage resulting from doses of B[a]P that elicited cell cycle arrest. We hypothesized therefore, that cell cycle events distal to Rb phosphorylation/E2F activation might be perturbed as a result of B[a]P treatment. Transcriptional induction of cyclin E occurs subsequent to Rb phosphorylation and E2F activation (24) and plays a key role in cell cycle progression. Therefore, we tested the possibility that B[a]P-induced DNA damage might affect signaling events involving cyclin E.

To test whether transcriptional induction of cyclin E was affected by B[a]P, we stably transfected Swiss 3T3 cells with a heterologous DNA construct containing the promoter region of the murine cyclin E gene linked to a luciferase cDNA. We tested the effects of B[a]P-induced DNA damage on mitogen-dependent activation of the cyclin E promoter essentially as described above for S.E.2F-LUC cells.

As shown in Fig. 2A, cyclin E promoter-driven luciferase expression was induced by mitogen (serum) treatment, as we had observed for E2F-dependent promoter activity. Cyclin E promoter-dependent luciferase expression was not inhibited as a result of B[a]P treatment (in fact, luciferase activity was consistently slightly elevated in extracts from serum and B[a]P-treated cells, relative to cultures that were treated with serum alone). These data suggested that any putative B[a]P-induced lesions in cell cycle signaling events were likely to occur at a stage distal to cyclin E transcription. To test this possibility, we determined the effects of B[a]P-induced DNA damage upon levels of cyclin E protein and cyclin E-dependent CDK activity, as described below.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Effect of B[a]P on cyclin E signaling. A, near-confluent cultures of S-CycE-LUC cells (which contain a stably transfected cyclin E promoter-luciferase construct) were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 2 µM B[a]P during the last 24 h of serum starvation. The cultures then received no treatment or 7% serum for 15 h. Whole cell extracts were prepared from the cells and analyzed for luciferase activity as described under "Materials and Methods." B, near-confluent cultures of fibroblasts were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 2 µM B[a]P during the last 24 h of serum starvation. The resulting cultures received no treatment or 7% serum for 15 h. Whole cell extracts isolated from the cells were analyzed for cyclin E expression by immunoblotting with anti-cyclin E antisera (upper panel) or were immunoprecipitated with anti-cyclin E antibodies and assayed for histone kinase activity in vitro (lower panel) as described under "Materials and Methods."

As is in previous experiments, quiescent cultures of Swiss 3T3 cells were treated with B[a]P (or Me₂SO for controls) before treatment with mitogen. 15 h after serum stimulation, we prepared whole cell extracts from the cultures. These extracts were analyzed for cyclin E protein levels by immunoblotting (Fig. 2B, upper panel) or for cyclin E-dependent kinase activity (Fig. 2B, lower panel). As expected, extracts from serum-treated cells contained high levels of cyclin E protein relative to extracts from control cultures. However, B[a]P treatment had no effect on the serum-induced expression of cyclin E protein (Fig. 2), suggesting that cyclin E levels were not limiting for cell cycle progression in B[a]P-treated cells.

To test the potential effects of B[a]P on cyclin E-associated kinase activity, we performed in vitro kinase assays using anti-cyclin E immunoprecipitates from quiescent, serum-stimulated, and serum + B[a]P-treated cells. The lower panel in Fig. 2B represents an autoradiogram of the phosphorylated histone used as an in vitro substrate for CDK in these assays. Relatively low levels of cyclin E-associated kinase activity were present in extracts from quiescent cells. Serum stimulation resulted in an 11-fold increase in the levels of cyclin E-associated kinase activity recovered from the cell extracts (as determined by Cerenkov counting of the phosphorylated histone protein bands excised from the dried gel). The induction of cyclin E-dependent kinase activity elicited by serum was unaffected in cells that had received additional treatment with B[a]P. These data showed that cyclin E/CDK2 signaling was unaffected by PAH-induced DNA damage. Therefore, changes in cyclin E signaling were unlikely to mediate B[a]P-induced growth arrest.

Effect of B[a]P on Cyclin A Signaling-- Rb phosphorylation, E2F activation, and increases in cyclin E-dependent kinase activity occur very late in G₁ and mediate the transition into S-phase. The results described above showed that B[a]P treatment and acquisition of DNA damage did not perturb these events. This suggested to us that PAH-induced cell cycle arrest might actually occur subsequent to completion of G₁ (i.e. during S-phase rather than G₁). Consistent with this hypothesis, in previous experiments in which we assayed mitogenesis in PAH-treated cells using measurements of [³H]thymidine incorporation, doses of PAH that were maximally effective for inhibition of cell cycle progression failed to completely abrogate thymidine incorporation (see for example Fig. 1 in Vaziri and Faller (13)). Moreover, our FACScan profiles from serum plus B[a]P-treated cells show broadening of the apparent G₁ peak and the presence of a significant S-phase population (Fig. 1A), indicative of cell cycle arrest after commencement of DNA replication.

These data suggested that cell cycle arrest in response to PAH damage might result from putative checkpoint mechanisms involving S-phase signaling events. In mammalian cells, responses to DNA damage during the S-phase of the cell cycle have not been as extensively studied as G₁ checkpoint mechanisms. However, cells are prone to acquisition of DNA damage during all phases of the cell cycle. Therefore it is highly likely that appropriate mechanisms have evolved to modify cell cycle progression in response to damage acquired during S-phase. Indeed, distinct signaling mechanisms that mediate cell cycle stage-specific checkpoint responses to DNA damage (including S-phase-specific responses) have been described in some eukaryotic organisms (25). In experiments described below, we investigated potential lesions in S-phase signaling events that might constitute an S-phase checkpoint mechanism.

During an unperturbed cell cycle, transcriptional induction of cyclin A expression occurs at the G₁/S-phase boundary, subsequent to Rb phosphorylation and E2F activation (26). Expression of cyclin A continues to rise throughout S-phase and peaks during late S/G₂. Cyclin A is considered to play an important and necessary role in regulating the transition from G₁ into S-phase (as well as progression through S-phase). Since we did not detect any PAH-induced lesions in critical G₁ signaling cascades and because the results of our mitogenic assays were consistent with an S-phase block in response to B[a]P, we tested the consequences of PAH-induced DNA damage upon induction of cyclin A expression and cyclin A-dependent kinase activity.

We stably transfected Swiss 3T3 cells with a DNA construct containing the promoter region of the murine cyclin A gene linked to a luciferase cDNA. We tested the effects of treatment with B[a]P upon subsequent mitogen-dependent activation of the cyclin A promoter in the resulting stable transfectants (designated S-CycA-LUC). As expected, serum elicited a strong induction of luciferase expression at time points corresponding to activation of the endogenous cyclin A gene by mitogenic stimuli (Fig. 3A). Luciferase activity in extracts from B[a]P-treated cells was not changed (relative to that present in cells given serum alone), suggesting that transcription of the cyclin A gene was unaffected by B[a]P-induced DNA damage. Consistent with this result, serum stimulation resulted in induction of cyclin A transcripts in control (PAH-untreated) as well as in B[a]P-treated cultures (Fig. 3B). These results suggested that neither transcription of the cyclin A gene nor levels of cyclin A transcripts were limiting for progression through S-phase in PAH-treated cells.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Effect of B[a]P on cyclin A signaling. A, near-confluent cultures of S-CycA-LUC cells (which contain a stably transfected cyclin A promoter-luciferase construct) were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 2 µM B[a]P (BP) during the last 24 h of serum starvation. The cultures then received no treatment or 7% serum for 22 h. Whole cell extracts were prepared from the cells and analyzed for luciferase activity as described under "Materials and Methods." B, near-confluent cultures of Swiss 3T3 cells were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 2 µM B[a]P during the last 24 h of serum starvation. The cultures then received no treatment or 7% serum for 22 h. RNA extracted from the cells was separated by electrophoresis, transferred to nitrocellulose, and probed with a radiolabeled cyclin A cDNA as described under "Materials and Methods." C, near-confluent cultures of Swiss 3T3 were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 1 µM B[a]P during the last 24 h of serum starvation. The cultures then received no treatment or 7% serum for 23 h. Extracts from the cells were then separated by SDS-PAGE, transferred to nitrocellulose filters, and analyzed for cyclin A (top panel) or cyclin B (middle panel) expression as described under "Materials and Methods." For histone kinase assays (lower panel) extracts from the cells were immunoprecipitated with anti-cyclin A antibodies. Immune complexes were recovered and assayed for histone kinase activity as described under "Materials and Methods."

Fig. 3C shows the result of an immunoblotting experiment (performed in parallel with the RNA blot analysis described above) in which we compared levels of cyclin A protein in control and PAH-treated cultures. In control cultures (that had not received PAH), mitogen treatment resulted in induction of cyclin A protein expression relative to quiescent cultures. Interestingly, however, cyclin A protein levels were markedly reduced in cells that had received B[a]P. Serum-induced expression of cyclin B (which ordinarily occurs distal to expression of cyclin A during a normal cell cycle) was also prevented following acquisition of B[a]P-induced DNA damage (Fig. 3C). We also performed histone H1 kinase assays in vitro using anti-cyclin A immunoprecipitates from control, mitogen-treated, and mitogen + B[a]P-stimulated cells. As predicted from our immunoblotting experiments, reduced levels of cyclin A-associated kinase activity were recovered from serum plus B[a]P-treated cells relative to cultures that received only serum (Fig. 3).

Ectopically Expressed Cyclin A Overcomes PAH-induced Cell Cycle Arrest-- Our data suggested that cyclin A expression may be rate-limiting for cell cycle progression in cells containing B[a]P-adducted DNA. To test this hypothesis, we determined the effects of ectopically overexpressed cyclin A upon PAH-induced cell cycle arrest. Because of the high efficiency of gene transfer with adenoviral vectors, we constructed and employed a recombinant adenovirus for delivery of the cyclin A cDNA (under transcriptional control of a strong CMV promoter) in these experiments. Since Swiss 3T3 fibroblasts were not infected efficiently by adenovirus (data not shown), we performed these studies in an alternative fibroblast line (the Rat1 cell line), which is susceptible to infection by adenovirus.

As expected, serum-starved populations of Rat1 fibroblasts comprised predominantly G₁ cells (Fig. 4A, top panels). Serum stimulation of these quiescent cultures resulted in near-complete passage through S-phase and into G₂/M (Fig. 4A, second panels from top). As in earlier experiments in Swiss 3T3 cells, PAH-treated Rat1 cells underwent cell cycle arrest and failed to complete S-phase following mitogen treatment (Fig. 4A, bottom two panels on left-hand side). Interestingly, however, when mitogen and PAH-treated cells were infected with a recombinant adenovirus encoding cyclin A (Fig. 4A, right-hand panels), the resulting cells overcame the DNA damage-induced cell cycle arrest and progressed into G₂/M. Serum plus PAH-treated cells infected with a control adenovirus (AdCon, which lacks the cyclin A cDNA), were susceptible to cell cycle arrest (Fig. 4A, left-hand panels). Therefore, adenovirus infection alone was not a responsible bypass of the DNA damage-induced checkpoint. Adenovirus constructs encoding other pro-mitogenic proteins (including v-ras, mdm2) did not overcome the DNA damage-induced cell cycle block (data not shown), showing that the effect of cyclin A was not a general consequence of overexpressed mitogenic signaling proteins. The ability of ectopically expressed cyclin A to overcome PAH-induced cell cycle arrest is consistent with our hypothesis that cyclin A expression is limiting for cell cycle progression in cells containing B[a]P-adducted DNA.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Effect of ectopically expressed cyclin A on PAH-induced cell cycle arrest. A, 10-cm plates of confluent Rat1 R12 cells were serum-starved for 48 h. Some plates were then treated with 0.5 or 1.0 µM BPDE for 1 h. Cultures were then infected with 1011 AdCon or AdCycA particles (this concentration of a similarly prepared recombinant virus encoding GFP was sufficient to express GFP in greater than 90% of the population, as determined by fluorescence microscopy). Two hours after infection, cells were stimulated with 7% serum (or were given no serum for controls) as indicated by the figure. All plates also received 0.4 mg/ml nocodazole (to arrest cells with a G2/M content). 26 h after the addition of serum, cells were harvested for FACScan analysis as described under "Materials and Methods." B, some of the profiles from A were superimposed on a three-dimensional histogram to enable direct comparison of the effects of serum, BPDE, and AdCycA on cell cycle progression. The three-dimensional histograms represent profiles from cells that received, left to right, AdCon, no serum, no BPDE; AdCon, 7% serum, no BPDE; AdCon, 7% serum, 0.5 µM BPDE; AdCycA, 7% serum, 0.5 µM BPDE.

Resumption of S-phase in PAH-arrested Cells Occurs Concomitant with Increases in Cyclin A and B Protein Expression-- PAH-induced cell cycle arrest was transient, as shown by the FACScan profiles in Fig. 5A. PAH-treated cells remained arrested in S-phase for approximately 20 h (during which time the control cells treated with mitogen alone returned to a G₀/G₁ state). However, approximately 40 h after mitogen treatment, the PAH-treated cultures recovered from cell cycle arrest and gradually progressed through S-phase and into G₂. Recovery from cell cycle arrest correlated temporally with expression of cyclin A protein (which was followed by cyclin B expression), as shown by the immunoblotting experiment in Fig. 5B. Cyclin A mRNA levels remained at high levels in the PAH-treated cells both during cell cycle arrest and resumption of S-phase (Fig. 5B). These data also indicated that the changes in cyclin A protein levels during S-phase arrest and recovery were likely to be accounted for, at least in large part, by post-transcriptional mechanisms.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Resumption of S-phase progression and cyclin expression in PAH-treated cells. a, near-confluent cultures of Swiss 3T3 cells were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 2 µM B[a]P during the last 24 h of serum starvation. The cultures then received 7% serum for 0, 22, 40, and 48 h. At each time point following serum addition, cells were trypsinized, fixed, and stained with PI. Cell cycle profiles were obtained from the resulting samples as described under "Materials and Methods." B, near-confluent cultures of Swiss 3T3 cells were serum-starved for 72 h. Some serum-starved cultures received pretreatment with 2 µM B[a]P during the last 24 h of serum starvation. The cultures then received 7% serum for 0, 17, 20, 24, 45, and 70 h. At each time point following serum addition whole cell extracts were prepared from the cultures. The resulting samples were analyzed for expression of cyclin A (upper panel) and cyclin B (middle panel) by SDS-PAGE and immunoblot analysis with appropriate antisera. In a similar experiment performed in paral- lel, RNA was extracted from the cells. The resulting samples were tested for expression of cyclin A transcripts by RNA blotting (lower panel) as described under "Materials and Methods."

Checkpoint mechanisms elicited by damaged DNA are hypothesized to delay cell cycle progression and prevent mis-replication of damaged genes, thereby providing growth-arrested cells additional time to repair DNA lesions before re-entry into the cell cycle (12, 25). However, in experiments in which we quantified levels of DNA damage during recovery from growth arrest we did not observe any decline in B[a]P adducts (data not shown). Interestingly, Saccharomyces cerevisiae have been shown to resume cell cycle progression following DNA damage-induced cell cycle delay, despite the persistence of unrepaired DNA lesions (25). Our data suggest the existence of a similar checkpoint adaptation phenomenon in higher eukaryotic organisms.

Overall, our data show that cell cycle arrest resulting from carcinogenic PAHs is associated with perturbation of S-phase events (expression of cyclin A protein, cyclin A-dependent kinase activation, cyclin B expression), but that earlier mitogenic signaling events during G₁ remain intact. Expression of cyclin A (and activation of cyclin A-dependent kinases) is necessary for entry into and progression through S-phase (26). Our results show that PAH-induced DNA damage negatively regulates cyclin A expression via post-transcriptional mechanisms. These data suggest that cyclin A protein levels are rate-limiting for cell cycle progression in cells containing PAH-adducted DNA. Our finding that cyclin A expression was reduced in PAH-treated cells provided a plausible explanation for the earlier observations that E2F-responsive promoters were moderately induced in serum + PAH-treated cells relative to serum-treated controls. Cyclin A/CDK is known to be a negative regulator of E2F/DP-dependent transcriptional activity (27). Potentially, loss of the cyclin A/CDK-mediated negative regulation of E2F/DP complexes in PAH-treated cells may account the increases in E2F-responsive promoter activity in our experiments. Additionally, PAH-induced expression of the mdm2 oncoprotein (13) may result in E2F activation (23), possibly contributing to elevated responses of E2F-dependent reporters following PAH-treatment.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We have shown that cells containing PAH-adducted DNA undergo cell cycle arrest. This phenomenon is superficially analogous to the growth arrest elicited by other genotoxic agents such as UV and ionizing radiation. DNA damage resulting from these latter agents is known to negatively regulate G₁ signal transduction events, causing growth arrest prior to entry into S-phase (14, 15). We hypothesized that PAH-induced DNA damage might similarly exert negative controls on G₁ events and cell cycle phase transitions. Therefore, we have analyzed the key mitogen-induced G₁ signaling events that are thought to mediate cell cycle progression through G₁ and into S-phase. Our results show that in contrast with well documented responses to other genotoxic agents, PAH-induced DNA damage does not perturb mitogenic signaling events involving the G₁ cyclins (and their partner CDKs), Rb phosphorylation, or E2F activation. p53-mediated G₁ arrest involves inhibition of the CDK/Rb/E2F signaling axis. Our finding that these events are not perturbed as a consequence of PAH-induced DNA damage is consistent with our previous experiments, which showed that p53 is not required for PAH-induced cell cycle arrest (13).

There exist several potential explanations for the lack of a demonstrable p53-mediated G₁ arrest in response to PAH-induced DNA damage. Potentially the amount of DNA damage elicited by PAHs under the experimental conditions we have described may be insufficient to trigger p53-mediated cell cycle responses. Although numerous experiments have shown p53-mediated G₁ arrest in response to genotoxic agents, those experiments have not generally quantified the numbers of DNA lesions present in the growth-arrested cells. In our experiments, relatively low levels of DNA adducts were formed following treatment with doses of B[a]P that were maximally effective for cell cycle arrest (0.4 adducts/10⁶ bases after 24 h of treatment with 2 µM B[a]P). Therefore, more severe DNA damage (than elicited under our experimental conditions in the studies described here) may be required to elicit p53-mediated transcriptional responses that result in G₁ arrest.

Alternatively, the nature of the DNA damage elicited by PAH treatment may be inadequate for stimulation of p53-mediated growth arrest. Genotoxic B[a]P metabolites generate bulky DNA adducts by covalent linkage with the exocyclic amino group of dG residues, yet cause little or no oxidative damage. In contrast with PAHs, IR and UV radiation (genotoxic agents that elicit a commonly studied p53-mediated cell cycle arrest in G₁) generate a spectrum of DNA lesions. These forms of damage can result from direct interaction of radiation with DNA. Additionally, UV- and ionizing radiation-induced DNA damage may result indirectly from reactive species formed by the excitation of water and other small molecules in the nuclear environment. Thus, IR generates single- and double-strand breaks, clustered base damage, and sugar damage as well as inter-strand cross-links. UV irradiation is known to result in the formation of cyclobutane pyrimidine dimers, pyrimidine-pyrimidone (6-4) photoproducts, and thymine glycol as well as other complex lesions. The different forms of DNA damage described above are recognized and repaired by distinct mechanisms (28). It is not unreasonable to postulate that cell cycle responses to different forms of DNA damage might also mediated via different checkpoint pathways (for example p53-dependent and p53-independent signaling cascades). Indeed p53 has been shown to be differentially phosphorylated in cells following exposure to different DNA damaging agents (14, 15). It is also possible that cellular stresses other than DNA damage that are elicited by radiation treatment contribute to p53 activation. For example, UV radiation can activate signal transduction events that are initiated at the plasma membrane, as shown by experiments using enucleated cells. Potentially, such DNA damage-independent events could affect the overall cellular response to DNA damage.

Although p53-regulated checkpoint mechanisms are not operational in response to genotoxic PAHs, cells that have acquired PAH-DNA adducts do undergo cell cycle arrest. Apparently PAH-induced growth arrest actually occurs in early S-phase and not in G₁ as we previously suggested. This is evidenced by the incomplete inhibition of mitogen-induced thymidine incorporation in mitogenic assays with PAH-treated cells and by the presence of S-phase cells in FACScan profiles of mitogen + PAH-treated cells. Our finding that G₁ mitogenic events (expression of cyclins D and E, Rb phosphorylation, E2F activation) are not limiting for cell cycle progression in PAH-treated cells also suggests that cell cycle arrest occurs after passage of the "R" restriction point and completion of G₁.

Cyclin A expression is the earliest mitogen-dependent cell cycle event that we have found to be perturbed in synchronized PAH-treated cells (concomitant with cell cycle arrest). Cyclin A expression is absolutely required for progression through S-phase. Therefore it is highly likely that cyclin A levels are limiting for S-phase progression in cells containing PAH-adducted DNA. Thus, decreases in cyclin A protein might play a causal role in aryl hydrocarbon-induced cell cycle arrest. Consistent with this hypothesis, we have shown that ectopic expression of cyclin A is able to abrogate a PAH-induced block to cell cycle progression. These data suggest that cyclin A expression is indeed an important rate-limiting factor for S-phase progression in cells containing PAH-adducted DNA.

Expression of cyclin A has been reported to be down- regulated in response to the DNA cross-linking agent cis-platin (29). In the latter study, down-regulation of cyclin A expression was attributed to reduced activation of the cyclin A promoter by E2F (and appeared to be a consequence of p53 activation and p21 induction). In contrast with the responses to cis-platin studied by Spitovsky et al. (29), in our experiments mitogen-dependent E2F activity, activation of the cyclin A promotor, and cyclin A mRNA transcript levels were unperturbed by PAHs. Therefore, the decreases in cyclin A expression in response to PAH treatment are likely to result from altered translation or stability of the cyclin A protein. It appears that multiple mechanisms may contribute to down-regulation of cyclin A expression in response to DNA-damaging agents.

Expression of cyclin proteins is known to be regulated in part by proteolysis (30). Our data raise the interesting possibility that the proteolytic machinery that determines cyclin A levels is an effector of a checkpoint pathway that detects DNA damage (or unreplicated DNA). In preliminary studies designed to test this hypothesis we have investigated the effects of proteasome inhibitors upon B[a]P-induced repression of cyclin A expression. Interestingly, in such studies, increased amounts of an electrophoretically retarded cyclin A accumulate in PAH-treated cells relative to control cultures following inhibition of proteasome activity. Experiments are currently under way to test the nature of the putative post-translational modification, which retards cyclin A mobility and may target the protein for degradation. Interestingly, an N-terminal-truncated (and proteolysis-resistant) cyclin A fusion protein has been identified as an oncoprotein in a human hepatocellular carcinoma (31). Oncoproteins frequently promote genomic instability and multi-step carcinogenesis by overcoming checkpoints that are otherwise imposed upon normal cells. Potentially, a proteolysis-resistant cyclin A could abrogate S-phase checkpoints, such as the one elicited by PAH-adducted DNA, thereby resulting in mutagenic mis-replication of the genome.

In conclusion, PAHs induce a form of cell cycle arrest that is fundamentally different from the checkpoints elicited by other well studied genotoxic agents. As noted previously, B[a]P and related PAHs are abundant and ubiquitous environmental pollutants with well documented carcinogenic properties. Therefore, it is important to understand the checkpoint mechanisms that prevent "fixation" of PAH-induced mutations. It is likely that the molecular components of the PAH-induced checkpoint machinery serve an important tumor suppressive function. Experiments are under way to identify the (putative) signaling pathways that mediate PAH-induced cell cycle arrest.

	ACKNOWLEDGEMENTS

We thank all our colleagues who provided reagents for these studies. Promoter-luciferase constructs for cell cycle genes were provided by Drs. Pidder Jansen-Durr (cyclin E-Luc), Jim Xiao (E2F-Luc), and J. M. Blanchard (Cyclin A-Luc). The cDNA probe for cyclin D was provided by Dr. Charles Sherr. We are grateful to Dr. Ed Loechler for his insightful comments and suggestions during the course of these studies.

	FOOTNOTES

^* This work was funded by American Cancer Society Grant IRG-72-001-24-IRG, a grant from the American Cancer Society, Massachusetts Division, Inc., and National Institutes of Health Grants ES09558 (to C. V.) and CA50454 (to D. V. F.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Cancer Research Center, Boston University School of Medicine, 80 E. Concord St., Boston MA 02118. Tel.: 617-638-4175; Fax: 617-638-5609; E-mail: cvaziri@acs.bu.edu.

	ABBREVIATIONS

The abbreviations used are: PAH, polycyclic aryl hydrocarbon; BP, benzo[a]pyrene; AhR, aryl hydrocarbon receptor; BPDE, benzo[a]pyrene diol epoxides; CDK, cyclin-dependent kinase; CMV, cytomegalovirus.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES