Report

Restriction of DNA Replication to the Reductive Phase of the Metabolic Cycle Protects Genome Integrity

See allHide authors and affiliations

Science  29 Jun 2007:
Vol. 316, Issue 5833, pp. 1916-1919
DOI: 10.1126/science.1140958

Abstract

When prototrophic yeast cells are cultured under nutrient-limited conditions that mimic growth in the wild, rather than in the high-glucose solutions used in most laboratory studies, they exhibit a robustly periodic metabolic cycle. Over a cycle of 4 to 5 hours, yeast cells rhythmically alternate between glycolysis and respiration. The cell division cycle is tightly constrained to the reductive phase of this yeast metabolic cycle, with DNA replication taking place only during the glycolytic phase. We show that cell cycle mutants impeded in metabolic cycle–directed restriction of cell division exhibit substantial increases in spontaneous mutation rate. In addition, disruption of the gene encoding a DNA checkpoint kinase that couples the cell division cycle to the circadian cycle abolishes synchrony of the metabolic and cell cycles. Thus, circadian, metabolic, and cell division cycles may be coordinated similarly as an evolutionarily conserved means of preserving genome integrity.

Cyclic biological oscillators operate in numerous life processes over broad time scales (14). Two cardinal biological oscillators, the cell division cycle and the circadian rhythm cycle, are temporally coupled (2, 5), yet the biological significance of this coupling is poorly understood. The budding yeast, Saccharomyces cerevisiae, is not subject to a cycle of circadian dimensions. When grown in glucose-rich medium, S. cerevisiae preferentially ferment glucose in the absence of respiration to support rapid growth. In contrast, when grown under nutrient-limited conditions in continuous culture, they undergo oscillation between glycolytic and respiratory metabolism (69). Comprehensive studies of a 4- to 5-hour yeast metabolic cycle (YMC) demonstrated that this cycle facilitates temporal compartmentalization of cellular processes in a manner reminiscent of circadian rhythm (9). Cell division has been shown to be restricted to the reductive phase of the YMC when oxygen consumption is minimal (7, 9), which suggests that temporal segregation of DNA replication away from respiration might shield DNA from oxidative damage.

To explore the relation between the YMC and the cell division cycle (CDC), we first used a sensitive and temporally precise bromodeoxyuridine (BrdU)–labeling assay to monitor the progression of DNA replication throughout the YMC (10). Nuclear DNA replication started at the very beginning of the reductive building (RB) phase when respiration began to cease (time point T10), reached peak levels at T11 and T12, and diminished sharply thereafter (figs. S1 and S2). This time frame of DNA replication coincides precisely with a sharp increase in ethanol concentration, indicative of a highly glycolytic, nonrespiratory environment for replicating cells (9, 11).

We next asked whether the YMC-directed restriction of the CDC might be compromised in cell cycle mutants. To test this, we constructed 25 mutant strains, each bearing a disruption in a gene known to regulate the CDC (table S1). Mutants that failed to affect log growth rate in glucose-rich medium also failed to affect the length or amplitude of the YMC (Fig. 1, A and B). In contrast, mutants that slowed growth in glucose-rich medium shortened the temporal duration of the YMC by varying degrees. Although imperfect, a general correlation was observed; slower growth in the high-glucose state corresponded to more substantial truncation in metabolic cycle length (table S1). No correlation was observed between the distinct CDC phases prolonged by the mutation and the degree of YMC truncation. Consistent with morphological assays of budding as a function of the YMC (9), BrdU incorporation studies indicated that roughly half of the cells progress through the CDC per metabolic cycle for the wild-type (WT) CEN.PK strain, as well as for mutants exhibiting normal growth and YMC, e.g., cln1Δ (Fig. 1C). In contrast, we observed a reproducible decrease in the percentage of cells in the CDC per abbreviated YMC in growth-retarded mutants. When corrected for overall time, all of the strains tested allowed replication of 100% of the cell population in a period of 5 to 8 hours. For example, if 50% of the parental CEN.PK cells divide per 3.9-hour metabolic cycle, the strain can be predicted to require 7.8 hours for 100% of its cells to transit the CDC. Likewise, 100% replication estimates for bub1Δ, cdh1Δ, bem2Δ, swi6Δ, and sic1Δ corresponded to 7.4, 7.2, 6.7, 6.1, and 5.2 hours, respectively. We tentatively conclude that slower-growing mutants can maintain homeostatic cell density in the fermentor by abbreviating the temporal duration of the YMC. We next investigated whether the CDC might be equivalently restricted to the reductive phase of the YMC in strains exhibiting an abbreviated metabolic cycle. BrdU was fed to each strain for brief periods corresponding to the oxidative, RB, or reductive charging (RC) phase of the YMC (Fig. 2A). Mutant strains having a normally timed YMC, such as cln1Δ, revealed tight restriction of the CDC to the RB phase of the YMC. Strains exhibiting intermediate reductions in YMC cycle length, such as bub1Δ, cdh1Δ, bem2Δ, and swi6Δ, showed partially impaired temporal constraint of the CDC (Fig. 2A and fig. S3). sic1Δ, the strain with the shortest metabolic cycle (1.2 hours) had almost equivalent BrdU labeling of nuclear DNA in the three phases of the YMC. Consistent with this impaired restriction of the CDC, fluorescence-activated cell sorting (FACS) and quantitative real-time polymerase chain reaction (PCR) analyses also revealed increasingly diminished differences in the number of dividing cells (Fig. 2B) and cell cycle gene expression (fig. S4), at different YMC time points.

Fig. 1.

Growth-retarded cell cycle mutants exhibit shortened metabolic cycles. (A) Representative metabolic cycle profiles for the WT and six mutants each bearing a deletion in a cell cycle gene. For each profile, the y axis is dissolved O2 (dO2). (B) Median reductions in metabolic cycle length of 25 cell cycle mutants. Symbols (Δ, *, and X) denote reductions of various mutant strains in metabolic cycle length relative to the average WT cycle length of 3.9 hours. (C) Fewer cells undergo cell division during shortened metabolic cycles. BrdU was added to the culture during the oxidative phase, and cells collected at the same point of the next cycle were analyzed by fluorescence staining.

Fig. 2.

Mutants that allow DNA replication during the oxidative phase have increased spontaneous point-mutation rates. (A) Percentages of replicating cells during the three phases of the metabolic cycle; standard deviations are shown as error bars. Cells were labeled for ∼20 min with BrdU at three different time intervals corresponding to the oxidative (Ox), RB, and RC phases. (B) FACS analysis of DNA content at 12 time points over one full metabolic cycle. (C) Mutation rate comparisons normalized to those of WT; standard deviations are shown as error bars.

Having observed replication outside the RB phase, particularly during the oxidative phase of the abbreviated YMC, and knowing that replicating cells are vulnerable to DNA damage (1214), we evaluated the spontaneous mutation rate at the CAN1 locus. When the CAN1 gene is mutated, colonies can grow on selective plates containing the toxic arginine analog, canavanine (1517). All strains were grown under two conditions—either in glucose-rich medium, where yeast cells do not respire, or under nutrient-limiting conditions in continuous culture, where they cycle back and forth between glycolytic and respiratory metabolism. No statistically significant difference was observed in spontaneous mutation rates at the CAN1 locus among all strains when grown under nonrespiring conditions (Fig. 2C) (18). In contrast, mutants with an abbreviated metabolic cycle, especially those that permitted substantive DNA replication during the oxidative phase of the YMC, accumulated substantially higher levels of spontaneous mutations than the parental CEN.PK strain during continuous culture. We hypothesized that the enhanced mutation rate in uncoupled mutants results from DNA synthesis in the oxidative, respiratory phase of the YMC. It is formally possible, however, that enhanced mutation rates result from DNA replication in the RC phase of the YMC.

Progression of the YMC from the oxidative to the RB phase marks a sharp transition toward a more reductive metabolic environment (4, 9), which suggests that alteration in redox state might be crucial in the gating of the CDC/DNA replication to preserve genome integrity. To further explore the role of redox in the progression of these two cycles, we treated fermentor cultures with brief pulses of hydrogen peroxide (H2O2) at different times throughout the YMC (10). H2O2 failed to affect YMC progression when added during the oxidative phase and led to only modest phase delay after transition to the RB phase (Fig. 3, A and B). In contrast, H2O2 pulses elicited dramatic phase advancement of the YMC when administered during the RC phase. This oxidant-induced phase-response curve (Fig. 3B) is analogous to the light-induced phase-response curve of the circadian cycle (21, 22), wherein the zeitgeber can either advance or delay a phase in the cycle according to the time of administration.

Fig. 3.

The YMC exhibits a redox response curve and limits entry into the cell cycle. (A) H2O2 induces differential phase responses when pulsed at different times of the YMC. H2O2 was added at the indicated time points, which immediately caused a spike in the dO2 content (y axis) in the metabolic cycle traces. (B) The phase-response curve (PRC) of the YMC to H2O2. The phase shifts were calculated as the difference in cycle length between the cycle before and after the time of H2O2 pulses. Phase delay and phase advancement are denoted as negative and positive time changes, respectively (y axis). (C) Phase advancement of the metabolic cycle accelerates cell cycle entry. Fermentor cultures were treated with mock (H2O) (top), H2O2 (middle), and methionine (Met) (bottom) at T4, T4, and T6 (arrows), respectively. BrdU was then pulsed at the next time point (red dots) when H2O2 was fully depleted, and cells were subsequently collected and subjected to BrdU fluorescence staining. (Left) The metabolic cycle profiles. The onset and peak of DNA replication as evidenced by BrdU staining is denoted by light and dark green dots, respectively. (Right) Representative images of the BrdU staining: (tops) fluorescein isothiocyanate fluorescence staining; (bottoms) differential interference contrast (DIC).

To investigate whether phase advancement of the YMC might also advance the phase of the CDC, we performed BrdU labeling to examine the onset of DNA replication in either mock-treated or H2O2-treated fermentor cultures (10). DNA replication in cultures treated with H2O2 in the RC phase was first observed at T7 and peaked at T8, which indicated an H2O2-induced phase advancement of the CDC by three full time intervals (1 hour) relative to the mock-treated culture (Fig. 3C). To rule out a possible role of a mitogenic effect of H2O2 in this observed advancement, we applied methionine pulses at different intervals of the YMC and again observed clear evidence of advancement from the RC to the oxidative phase (10). A methionine pulse administered at T6 significantly advanced the phase of the YMC, and replicating cells were detected at T8 immediately after BrdU addition at T7, which indicated phase advancement of cell cycle entry by two full time intervals (40 min) (Fig. 3C). Taken together, these observations are consistent with the hypothesis that the metabolic cycle gates cell cycle entry. It is likely, however, that H2O2 (an oxidant) and methionine (a reductant) act through distinct mechanisms to advance the phase of the YMC (10).

Fungal geneticists have recently demonstrated that, in Neurospora crassa, the prd-4 gene, which is involved in the control of circadian rhythm, encodes a DNA checkpoint kinase that prevents cell cycle progression in response to damaged DNA, suggestive of a role in coupling the cell cycle to circadian rhythm (23). Disruption of the orthologous RAD53 gene in S. cerevisiae is lethal, but the organism can be rescued by concomitant disruption of the SML1 locus (23, 24). No difference was observed between parental cells and the sml1Δ single mutant in continuous culture (Fig. 4). By contrast, cells of the rad53Δ sml1Δ genotype sustained no more than three or four metabolic cycles before completely losing oscillatory behavior. FACS analysis of the double mutant gave evidence of partial CDC restriction during the initial metabolic cycles, but CDC synchrony to the YMC was fully abolished after cessation of metabolic oscillation (Fig. 4B). The rad53Δ sml1Δ double mutant also suffered the highest spontaneous point-mutation frequency of any strain tested to date (Fig. 4C). The mutation rates of these sml1Δ, rad53Δ sml1Δ, and parental strains were indistinguishable when tested under log-phase, high-glucose growth conditions.

Fig. 4.

A conserved DNA checkpoint kinase is required for synchrony of both the metabolic cycle and the cell cycle. (A) Metabolic cycle profiles of sml1Δ (top) and rad53Δ sml1Δ (bottom) mutants. (B) FACS analysis of DNA content at 12 time points over a typical metabolic cycle of the sml1Δ mutant (top), the initial full cycle of the rad53Δ sml1Δ double mutant (middle), and after the metabolic cycle of the double mutant was disrupted (bottom). (C) Mutation rate comparison. Mutation rates were normalized to those of the WT strain and are presented with standard deviation (error bars).

We hereby show that growth-retarded cell cycle mutants maintain continuous culture homeostasis by allowing more frequent “gate openings” for cell cycle entry, as well as DNA synthesis outside the reductive phase of the YMC, at the cost of increased spontaneous mutation rates. This trade-off between genome integrity and cell proliferation is reminiscent of cancer cells (25, 26). The YMC may temporally segregate cell division away from mutagenic consequences of DNA replication during periods of intense respiration. Analogous to the YMC, the circadian cycle gates cell division (2, 5, 27). Thegenethat couples the cell cycle and DNA damage response to the circadian cycle in Neurospora crassa (23) and mammalian cells (28) is also required for synchronization of the cell division and metabolic cycles of S. cerevisiae. In addition, the phase-response curve of the YMC to H2O2 pulses resembles that of the circadian cycle in response to light pulses, suggestive of a redox-driven regulatory apparatus that may control the YMC. These relations underscore the importance of regulatory systems that confine DNA synthesis to a properly protective reductive environment.

Supporting Online Material

www.sciencemag.org/cgi/content/full/316/5833/1916/DC1

Materials and Methods

Figs. S1 to S4

Table S1

References

References and Notes

View Abstract

Navigate This Article