Abstract
Aim: Thymidine deprivation is a common cancer treatment. This study examines the role of replication arrest and uracil DNA repair in response to thymidine deprivation. Materials and Methods: Strains of S. cerevisiae deficient in various replication and DNA repair functions were tested for sensitivity to thymidine deprivation induced by the antifolate aminopterin. Cell survival and DNA content were assayed following drug treatment. Results: Most arrest mutants were more sensitive to aminopterin than was the parental strain. Inactivation of uracil glycosylase in arrest mutants led to a partial reduction in toxicity for some double-mutants. DNA content during exposure to aminopterin was similar in parental and single mutants. However, cells deficient in both arrest and uracil glycosylase functions exhibited continued DNA synthesis, suggesting that uracil glycosylase activity contributes to replication arrest during thymidine deprivation. Conclusion: Replication arrest and uracil DNA repair are important and overlapping determinants of cellular response to thymidine deprivation.
Nucleotide pool limitations incite a number of responses in Saccharomyces cerevisiae. Depletion of deoxyribonucleotides (dNTPs), either by treatment with hydroxyurea or ribonucleotide reductase inactivation, disrupts replication resulting in S-phase arrest (1). DNA is structurally sensitive during replication arrest. It is unwound, partially single-stranded and significantly more vulnerable to damaging agents during S-phase arrest (2, 3).
A number of proteins are thought to be active in sensing and responding to replication stress caused by nucleotide pool depletion. These proteins serve to recognize replication difficulties or DNA damage, transmit appropriate signals and stabilize DNA. The roles of several proteins have been described during nucleotide pool depletion caused by decreased ribonucleotide reductase activity, which reduces the abundance of all four dNTPs. However, the role of these proteins in responding to selective depletion of thymidine has not been fully-elucidated.
Thymidine deprivation poses unique challenges. Thymidine monophosphate (dTMP) is produced by thymidylate synthase in a reaction that transfers a one-carbon group from reduced folate to dUMP to produce dTMP. Anti-folates block this reaction, leading to elevated levels of the precursor dUMP and subsequently of dUTP (4, 5). dUTP and dTTP are chemically similar and most DNA polymerases have poor discrimination between dUTP and dTTP. Therefore, when thymidylate synthase is inhibited, the dUTP level rises, the dTTP level drops and these conditions favor incorporation of dUTP into DNA (5, 6). Once incorporated into DNA, uracil becomes a target for uracil glycosylase DNA base repair, which involves base excision and strand incision. These activities could be harmful if performed on DNA already in vulnerable conformations due to stalled replication.
How do cells balance these potentially conflicting processes of cell-cycle arrest and uracil base excision repair? Understanding how arrest and repair activities are coordinated during thymidine deprivation may provide with unique insights into how these pathways interact. This study examines the role of replication arrest during thymidine deprivation. Interactions between arrest and uracil base excision repair are also examined.
Materials and Methods
Strains. Commercially available single-gene deletion strains of S. cerevisiae were used for this study. The parental strain BY4741 and deletion derivatives were obtained from Open Biosystems (Huntsville, AL, USA). Uracil DNA glycosylase mutant (ung1) strains were constructed by replacing the entire reading frame for UNG1 with the selectable histidine biosynthetic gene HIS3.
Media. Cells were grown in minimal media supplemented with charcoal-adsorbed casamino acids, uracil and tryptophan and adenine, as described previously (7). Thymidine deprivation was induced by adding sulfanilamide at 6 mg/ml to the media and resterilizing. Aminopterin was added at a final concentration of 200 μg per ml (7).
Thymidine deprivation. Cells were grown in supplemented minimal media overnight, diluted 1:100 and grown for 4-6 h. These log-phase cells were then diluted to 105 cells per ml in media containing sulfanilamide and aminopterin. Viable cell density (colony-forming units) was determined just after cells were added to the aminopterin culture and again after 24 hours. Cultures were diluted in water prior to plating to solid YPD media.
Survival calculations. The percentage survival was calculated by dividing the number of colony-forming units (viable cells) per ml at 24 hours by the colony-forming units per ml at time zero. Percentages shown in the tables are averages of a minimum of three independent cultures. Standard deviations were determined and Student's t-test performed to compare survival data.
Flow cytology analysis. Cultures were prepared for thymidine deprivation as described above with the exception of the use of a cell density of 106 cells per ml. Cultures were harvested at time 0 and 12 hours. Cells were lysed, stained for DNA content using propidium iodide and analyzed as described previously (7).
Results
The role of S-phase and replication arrest in mediating the toxicity of thymidine deprivation was assessed by examining aminopterin sensitivity in mutants deficient in these processes. Mutants chosen were viable as a haploid deletion strain and had a defined role in S-phase or replication arrest (8). Parental and mutant strains were treated with the antifolate aminopterin, as described in Materials and Methods. Multiple cultures were tested for each strain and average survival values are presented in Table I. Most mutants display a significant sensitivity to aminopterin. Mutants lacking functions important for replication fork maintenance during replication stalling, including tof1 and mrc1, are particularly sensitive to aminopterin. Mutants lacking tel1 and rad9 are exceptions and apparently not sensitive to aminopterin.
The rad6 mutant was significantly more sensitive than the parental strain in this screening and by Game et al. in a screening for mutants sensitive to thymidine deprivation produced by the antifolate trimethoprim (9). Here, the aminopterin sensitivity of strains harboring mutations in various activities within the RAD6 repair and ubiquitination pathways was determined. As shown in Table II, some of these mutant strains were sensitive to aminopterin. UBI4, coding for a poly-ubiquitin transcript, appears to play a critical role in responding to thymidine deprivation since ubi4 mutants are exquisitely sensitive to aminopterin. Indeed, ubi4 and doa1 mutants were more sensitive than strains deficient in other RAD6 repair pathway activities. This implies a role for ubiquitin outside the RAD6 pathway. Def1 functions to ubiquitinate RNA polymerase. def1 mutants are very sensitive to aminopterin, suggesting multiple roles for ubiquitin in response to thymidine deprivation.
Uracil base excision repair plays a role in response to thymidine deprivation (5, 7, 10). Strains deficient for both arrest and uracil glycosylase functions were treated with aminopterin. Survival data for the double mutants in comparison to the uracil glycosylase proficient arrest single mutant are shown in Table III. The increase in aminopterin sensitivity of rad17, rad24 and tof1 single-mutants was partially relieved by the co-deletion of ung1. rad50 ung1 double-mutant was marginally more sensitive to aminopterin than either single-mutant. ung1 co-deletion did not uniformly reduce sensitivity to aminopterin, suggesting some arrest functions have a more direct interaction with uracil repair than others.
DNA content in various strains was examined using propidium iodine staining and flow cytometry during aminopterin treatment. The parental strain arrested in S-phase after 12 h of drug treatment. The patterns for rad6 and ung1 single-mutants both resembled the parental pattern. Interestingly, the rad6 ung1 double-mutant appears to be able to continue DNA synthesis more than either single-mutant, as seen by an increase in the number of cells with greater than mid-S phase content of DNA (Figure 1). This suggests that the presence or activity of uracil glycosylase contributes to S-phase arrest during thymidine deprivation. Uracil glycosylase may contribute to S-phase arrest in rad6 mutants, and Rad6 may participate in arrest in uracil glycosylase mutants, but the double mutant lacks a normal arrest function. Similar findings were found for rad17, mrc1 and tof1 mutations in combination with ung1 mutations (data not shown).
Discussion
Nucleotide pool imbalances pose a substantial threat to growing and dividing cells. dNTP depletion leads to stalled replication forks and activation of several S-phase checkpoint proteins (8, 11). Replication stress induced by hydroxyurea, which depletes all four dNTPs, has been a model system for studying the consequences of nucleotide pool depletion (2). Proteins participating in replication checkpoints, including those for DNA damage recognition, signaling, DNA repair and replication fork stabilization, are all important in managing the consequences of hydroxyurea treatment (12).
Proteins that play an important role in responding to hydroxyurea may also play an important role in response to selective dTTP depletion. Aminopterin is an anti-folate that inhibits the conversion of dUMP to dTMP leading to thymidine depletion. Many of the replication checkpoint mutants tested were found to be sensitive to aminopterin. rad50 and mre11 mutants, deficient in early stages of DNA damage recognition are sensitive to aminopterin. Mutants lacking various functions in replication fork stabilization (tof1, mrc1, rad17, ddc1 and rad24) are also sensitive to aminopterin. chk1 mutants are not sensitive to aminopterin despite Chk1 having a role in replication fork protection (GO annotation, SGD database, http://www.yeastgenome.org/). chk1 mutants are not sensitive to hydroxyurea either (12), suggesting Chk1 is not essential in responding to the type of DNA damage caused by these drugs. Simon et al. also found rad50 and rad17 to be sensitive to a methotrexate analog (13).
Thymidylate synthase inhibition by raltitrexed in human cancer cell lines induced DNA damage foci containing RAD51, RPA, pChk1, pNBS1, and phosphorylated histone H2AX (14). Reducing the levels of ATM and ATR in colon cancer cell lines through siRNA leads to increased sensitivity to thymidine deprivation (15). These data in human cancer cell lines, and the data presented here confirm the importance of the replication checkpoint function in mediating the response to thymidine deprivation.
Inhibiting the conversion of dUMP to dTMP may increase levels of dUMP and therefore dUTP. Rising dUTP levels lead to the possibility of dUTP incorporation into DNA (5, 6). Uracil is removed from DNA by activities such as strand incision, which may create additional DNA damage and toxicity. Uracil glycosylase has been identified at the site of active replication forks (16). Uracil glycosylase activity contributes to aminopterin toxicity in rad17, rad24, and tof1 mutants because simultaneous deletion of ung1 from these replication checkpoint mutants alleviates some of their aminopterin sensitivity. While ung1 mutants have close to parental sensitivity to aminopterin (5, 17), mutants lacking the second step in uracil base excision repair, apn1, are exquisitely sensitive to thymidine deprivation (7, 10). Collura et al. have also described functional interactions between Rad17 and uracil glycosylase (18). Their data show cells lacking rad17 and apn1 to have high mutation rates that are dependent on a functional uracil glycosylase. High spontaneous mutation rates or increased aminopterin sensitivity both suggest that uracil glycosylase activity in rad17 mutants is potentially toxic, likely through difficulty in managing apyrimidinic sites.
Flow cytometric data suggest that uracil glycosylase contributes to cell-cycle arrest, at least in the absence of a fully functional replication checkpoint. DNA synthesis is halted in early S-phase following aminopterin treatment in parental, rad6 and ung1 cells, but progresses to late S or G2 content of DNA per cell in the rad6 ung1 double-mutant. This finding implies rad6 and uracil glycosylase may have redundant abilities to halt progression of DNA synthesis. Overexpression of uracil glycosylase in S. pombe also leads to cell-cycle arrest (19). Uracil glycosylase binds tightly to apurinic and apyrimidinic sites in DNA until displaced by Apn1 (20). Uracil glycosylase bound to DNA may serve as a replication block. Alternatively, apurinic and apyrmidinic sites generated by uracil glycosylase could block replication (21). Additional studies are required to further define how uracil glycosylase contributes to replication arrest.
An important role for the RAD6 pathway in mediating thymineless stress was identified over 35 years ago (9). The RAD6 pathway includes multiple activities involved with error prone and error-free post-replication DNA repair. Members of this group have helicase activity (SRS2 and RAD5), DNA polymerase activity (REV1, REV3) and ubiquitin ligase activities (MMS2, UBC13, RAD5, RAD18). The sensitivity of rad5 mutant suggests the key defect responsible for sensitivity for members of the RAD6 pathway lies in the ability of Rad5 to resolve stalled replication forks (22). Srs2 is also a helicase and mutants show aminopterin sensitivity, supporting the importance of helicase activity in coping with thymidine deprivation.
Mutants with lower ubiquitin-producing capacity (doa1 and ubi4) are more sensitive than mutants deficient in RAD6 pathway functions. The relative insensitivity of mutant RAD6 pathway members (e.g. bre1, mms2) suggests that inability to ubiquitinate targets such as PCNA or histone H2b is not a major impediment to survival during aminopterin treatment. Other targets of ubiquitination, such as RNA polymerase, may be even more important. Def1 is responsible for RNA polymerase ubiquitination in the presence of DNA damage and loss of def1 renders cells sensitive to aminopterin. ubi4 mutants are sensitive to oxidative stress (23) and thymidine deprivation may create oxidative stress due to preferential mitochondrial damage (24). Over 400 proteins are modified by ubiquitin (25), leaving many potential candidate proteins to explain the sensitivity of ubi4 and doa1.
Thymidineless death continues to be an intriguing puzzle. Continued efforts to understand thymidineless death will hopefully lead to a better understanding of how repair, arrest and recovery interact during replication stress. Activities beyond DNA replication and repair including transcription, mitochondrial function, amino-acid metabolism and others, may also play a significant role in response to thymidine deprivation (12, 25, 26, 27). Understanding thymidineless death will also hopefully lead to improved outcome for thousands of patients with cancer whose treatment is based on thymidine deprivation.
Acknowledgements
This work was supported by grants from the Holden Comprehensive Cancer Center, University of Iowa and the Duluth Medical Research Institute and the SMDC Foundation. Monika Johnson, Michael Lin and Greg Peduska provided technical assistance.
- Received January 18, 2013.
- Revision received April 9, 2013.
- Accepted April 10, 2013.
- Copyright© 2013 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved