Direct monitoring of the strand passage reaction of DNA topoisomerase II triggers checkpoint activation

Abstract

By necessity, the ancient activity of type II topoisomerases co-evolved with the double-helical structure of DNA, at least in organisms with circular genomes. In humans, the strand passage reaction of DNA topoisomerase II (Topo II) is the target of several major classes of cancer drugs which both poison Topo II and activate cell cycle checkpoint controls. It is important to know the cellular effects of molecules that target Topo II, but the mechanisms of checkpoint activation that respond to Topo II dysfunction are not well understood. Here, we provide evidence that a checkpoint mechanism monitors the strand passage reaction of Topo II. In contrast, cells do not become checkpoint arrested in the presence of the aberrant DNA topologies, such as hyper-catenation, that arise in the absence of Topo II activity. An overall reduction in Topo II activity (i.e. slow strand passage cycles) does not activate the checkpoint, but specific defects in the T-segment transit step of the strand passage reaction do induce a cell cycle delay. Furthermore, the cell cycle delay depends on the divergent and catalytically inert C-terminal region of Topo II, indicating that transmission of a checkpoint signal may occur via the C-terminus. Other, well characterized, mitotic checkpoints detect DNA lesions or monitor unattached kinetochores; these defects arise via failures in a variety of cell processes. In contrast, we have described the first example of a distinct category of checkpoint mechanism that monitors the catalytic cycle of a single specific enzyme in order to determine when chromosome segregation can proceed faithfully.

Figure 1. Topo II Strand Passage Reaction (SPR) and Mutants Analyzed in this Study.

Left column, Main features of the SPR of the wild type Top2 enzyme. The cartoons and description are based on Dong and Berger 2007 and Wang 2002 , . a, Top2 homodimer bound to G-segment DNA (see Key). N-Gate is open in the absence of bound nucleotide. b, Binding of one molecule of ATP to each monomer is required for N-Gate closure. If a T-segment is captured, then G-segment cleavage is thought to be coupled with N-Gate closure, in order for the T-segment to be accommodated. c–d, hydrolysis of one ATP promotes conformation changes that swivel the Transducer domain, opening the DNA-Gate and leading to T-transport. e, DNA-Gate closure leads to G-segment re-ligation. f, C-Gate opening allows T-segment release followed by hydrolysis of the second ATP and N-Gate opening after release of the hydrolysis products. Mutants: K651A has greatly reduced affinity for G-segment DNA and thus does not perform appreciable strand passage reactions in vivo and cannot support viability . In vitro, relaxation of supercoiled DNA can be detected at very low levels, indicating that Top2^K651A can, albeit with a very limited capacity, perform the SPR and therefore overall folding of the enzyme is not abolished. Y782F lacks the active site tyrosine and therefore cannot cut G-segment DNA . It can bind to the G-segment and undergo rounds of ATP binding/hydrolysis as well as N-gate opening and closure. It is predicted to lack the ability to capture a T-segment due to space constraints within the N-terminal orifice of the enzyme in the absence of G-segment cleavage. G144I cannot bind nucleotide and therefore cannot lock the N-Gate closed . For this reason it is unlikely to capture a T-segment and since T-segment capture stimulates G-segment cleavage, it has much reduced cleavage activity. In vitro, however, cleavage activity has been measured and given this event, the enzyme may sample conformations normally associated with T-transport even in the absence of ATP hydrolysis and T-segment capture. E66Q has 200-fold reduced ATP hydrolysis activity and therefore inefficiently performs conformation changes that promote T-transport, including DNA-Gate opening . In vitro studies indicate that inefficient T-transport is followed by SPR arrest after release of the T-segment. A second SPR cycle is not possible because the N-Gate cannot open without release of the hydrolysis products. Top2-B44 is predicted to be defective in T-transport. The mutated residue is positioned where the TOPRIM-fold lies adjacent to the DNA binding domain. The mutant has a reduced rate of ATP hydrolysis (this study). L475A/L480P has wild type ATP hydrolysis activity but T-transport occurs at a much reduced rate . G738D and P824S are positioned in the C-gate portion of the enzyme. These mutants have a much reduced rate of the SPR but are predicted to not affect the T-transport steps associated with DNA-gate opening, but rather would affect a later step of the SPR . Right column, Indicates if checkpoint activation occurs upon expression of each mutant at endogenous levels in yeast cells depleted of Top2^deg during G1 (this study).

Figure 2. Characterization of a Yeast top2 Degron Strain.

a, Schematic of the chromosomal MET3-top2 ^deg gene encoding thermo-labile Top2^deg protein and controlled transcriptionally via the presence or absence of methionine in the growth medium. b, Western blot of Top2^deg. Temperature and carbon-source shifts (Figure S2) promote efficient degradation (Tub1-GFP, loading-control). c, Micrographs (left) and quantification (right) of failed nucleus segregation (DAPI) in anaphase cells with elongated spindles (Tub1-GFP) after Top2^deg was degraded in G1 and the subsequent cell cycle analyzed. Shaded region on graphs indicates the fraction of anaphase nuclei that were not segregated and the top left insets show expanded budding curves. d, CHEF gel analysis of separated chromosomes after Southern blotting to detect catenated topoisomers of the endogenous 2-micron plasmid. Wild type (WT), top2-4 and top2^deg strains were initially arrested in G1 (alpha-factor) at 26°C (37°C in the case of top2^deg) or subsequently allowed 2 hours to reach G2/M at 37°C in the presence of nocodazole to prevent anaphase onset (G2/M, 37°C). For top2^deg, additional samples were taken at 45 min and 70 min following alpha factor release at 37°C with nocodazole (Noc. 45 min, Noc. 70 min).

Figure 3. Analysis of G2/M Cell Cycle Checkpoint Activation in top2 Mutants.

a–c, Population Assays: kinetics of cell cycle progression based on budding (DIC microscopy) and spindle morphology (Tub1-GFP). Cell populations were returned to growth after depletion of Top2^deg and synchronization in G1 (alpha-factor). The interval between spindle assembly and anaphase (a, arrow) is not significantly different with wild-type endogenously expressed TOP2 or after Top2^deg degradation in G1. Panel b shows overlaid repetitions of each experiment to assess variation (green/blue/purple, indicate three experimental repeats). Panel c, histogram plot of average G2/M duration; see Material and Methods for statistical analysis. d–f, Single-cell Assays: kinetics of spindle assembly and elongation in single cells based on digital time-lapse microscopy of strains expressing Tub1-GFP. Images in panel d show representative images of each morphological state (also see Movie S1). Panel e, plot of average spindle length (or SPB diameter; first five time points) versus time for single cells aligned on the x-axis at the time of SPB separation (i.e. at time point 12 min). Standard deviation of lengths shows that SPB diameter and G2 spindle length are relatively constant. Standard deviation of spindle length increases markedly as some cells enter anaphase. Panel f, plots average time interval between SPB separation and the initiation of spindle elongation in anaphase B (+/− s.e.).

Figure 4. DNA-Associated Top2-B44 Activates a Mad2-Dependent G2/M Checkpoint.

Analysis of the kinetics of cell cycle progression (see Figure 3) following depletion of Top2^deg and release from G1 synchrony in cells expressing endogenous levels of the indicated mutant Top2 proteins. a, Population Assays: Histogram plots show average G2/M duration; see Material and Methods for statistical analysis. Western blots show each Top2 mutant relative to Tub1 loading control at G1 and G2. a values are significantly different to b values in the histogram plots. Strains with the same letter are not significantly different. b–c, Single-cell assays: b, plots of average spindle length versus time for single cells aligned on the x-axis at the time of SPB separation (i.e. at time point 12 min). Error bars show standard deviation of lengths. c, Histogram plots of average time interval between SPB separation and the initiation of spindle elongation in anaphase B (+/− s.e.).

Figure 5. Checkpoint Activation via Top2-B44 Requires Initiation of the Strand Passage Reaction.

a, Cartoon describing the catalytic defect in Top2^Y782F which cannot cleave G-segment DNA and thus performs non-productive cycles of ATP hydrolysis and N-gate closure/opening (see Figure 1 for complete Strand Passage Reaction and Key). b–d, Analysis of the kinetics of cell cycle progression (see Figure 3) following depletion of Top2^deg and release from G1 synchrony in cells expressing endogenous levels of the indicated mutant Top2 proteins. b, Population Assays: Histogram plots show average G2/M duration; see Material and Methods for statistical analysis. Western blots show each Top2 mutant relative to Tub1 loading control at G1 and G2. a values are significantly different to b values in the histogram plots. Strains with the same letter are not significantly different. c–d, Single-cell assays: c, plots of average spindle length versus time for single cells aligned on the x-axis at the time of SPB separation (i.e. at time point 12 min). Error bars show standard deviation of lengths. d, Histogram plots of average time interval between SPB separation and the initiation of spindle elongation in anaphase B (+/− s.e.).

Figure 6. Defective Top2 ATP Binding and Hydrolysis Activate the Mad2-Depedent Checkpoint.

a, Cartoon describing catalytic defects in Top2^G144I which cannot bind ATP (green annotations) and Top2^E66Q which cannot hydrolyze ATP (blue annotations). (See Figure 1 for complete Strand Passage Reaction and Key). b–j, Cell cycle analyses showing that Top2^G144I and Top2^E66Q activate Mad2-dependent but Rad53-independent checkpoint signaling. Analysis of the kinetics of cell cycle progression (see Figure 3) following depletion of Top2^deg and release from G1 synchrony in cells expressing endogenous levels of the indicated mutant Top2 proteins. (b–h) Population Assays: Histogram plots show average G2/M duration; see Material and Methods for statistical analysis. Western blots show each Top2 mutant relative to Tub1 loading control at G1 and G2. a values are significantly different to b values in the histogram plots. Strains with the same letter are not significantly different. (i,j) Single-cell assays: i, plots of average spindle length versus time for single cells aligned on the x-axis at the time of SPB separation (i.e. at time point 12 min). Error bars show standard deviation of lengths. j, Histogram plots of average time interval between SPB separation and the initiation of spindle elongation in anaphase B (+/− s.e.).

Figure 7. Defective T-Segment Transit may Activate the Mad2-Dependent Checkpoint.

Cell cycle analyses showing that Top2^L475A,L480P activates Mad2-dependent but Rad53-independent checkpoint signaling. a, Cartoon describing catalytic defect in Top2^L475A,L480P which has a reduced rate of T-segment transport due to inefficient G-segment cleavage and DNA-gate opening. b–d, Population Assays: Histogram plots show average G2/M duration; see Material and Methods for statistical analysis. Western blots show each Top2 mutant relative to Tub1 loading control at G1 and G2. a values are significantly different to b values in the histogram plots. Strains with the same letter are not significantly different. e–f, Single-cell assays: e, plots of average spindle length versus time for single cells aligned on the x-axis at the time of SPB separation (i.e. at time point 12 min). Error bars show standard deviation of lengths. f, Histogram plots of average time interval between SPB separation and the initiation of spindle elongation in anaphase B (+/− s.e.).

Figure 8. Slow Strand Passage does not Activate the Mad2-Dependent Checkpoint in the Absence of a T-Segment Transit Defect.

Analysis of cell cycle kinetics in top2 ^deg strains expressing Top2^G738D and Top2^P824S which have overall reduced rates of the catalytic cycle do not activate checkpoint signaling. a, Population Assays: Histogram plots show average G2/M duration; see Material and Methods for statistical analysis. Western blots show each Top2 mutant relative to Tub1 loading control at G1 and G2. a values are significantly different to b values in the histogram plots. Strains with the same letter are not significantly different. b–c, Single-cell assays: b, plots of average spindle length versus time for single cells aligned on the x-axis at the time of SPB separation (i.e. at time point 12 min). Error bars show standard deviation of lengths. c, Histogram plots of average time interval between SPB separation and the initiation of spindle elongation in anaphase B (+/− s.e.).

Figure 9. Top2-B44 is Defective in ATP Hydrolysis.

a, Cleavage activity of Top2-B44. Agarose gel electrophoresis of supercoiled (SC) plasmid DNA either untreated (C) or after incubation with purified wild type Top2 or Top2-B44 enzymes in the presence of increasing concentrations of etoposide (Etop). N = nicked forms, L = linear form. The linear form indicates plasmid that was cut by Top2 and re-ligation blocked by binding of etoposide to the Top2-DNA complex. b, Relaxation activity of Top2-B44. Left panel, Agarose gel electrophoresis of supercoiled (SC) plasmid DNA either untreated (C) or after 0–15 min. incubated with purified wild type Top2 or Top2-B44 enzymes. L = Linear form, R = relaxed topoisomers. Right panel, Quantification of relaxation activity of purified wild type (WT) Top2 or Top2-B44 (B44) enzymes versus time at 37°C and 28°C. c, Rate of ATP hydrolysis by wild type (WT) Top2 and Top2-B44 (B44) enzymes at 37°C and 28°C (measured by the release of free phosphate).

Figure 10. The C-Terminal Region (CTR) is Required for Checkpoint Activation by Top2-B44.

Top2-B44^ΔCTR does not activate checkpoint signaling and over-production of the CTR fragment of Top2 abrogates checkpoint signaling in top2-B44 cells. a, Population Assays: Histogram plots show average G2/M duration; see Material and Methods for statistical analysis. Western blots show each Top2 mutant relative to Tub1 control at G1 and G2 (in part d, * indicates a non-specific band). a values are significantly different to b values. Strains with the same letter are not significantly different. b–c, Single-cell assays: b, plots of average spindle length versus time for single cells aligned on the x-axis at the time of SPB separation (i.e. at time point 12 min). Error bars show standard deviation of lengths. c, Histogram plots of average time interval between SPB separation and the initiation of spindle elongation in anaphase B (+/− s.e.).

Figure 11. Checkpoint Activation by Top2-B44 is Independent of Kinetochores.

a–b, Top2-B44 induces checkpoint activation in G2 under conditions where the kinetochore protein Ndc10 was inactivated in the preceding G1-phase. a, Analysis of re-budding (progression into a second cell cycle) with or without nocodazole after release from G1 synchrony (alpha factor). During synchrony, the strains were shifted to the restrictive temperature for the ndc10-1 allele, thus destroying kinetochores in G1-phase, as previously described . b, Population analysis of the kinetics of cell cycle progression based on budding (DIC microscopy) and spindle morphology (Tub1-GFP) in the indicated strains after release from G1 synchrony. During synchrony, the strains were shifted to the restrictive temperature for the ndc10-1 allele, exactly as in panel a. Error bars on the line graphs show standard deviation between three repeats of the experiment. Histogram plot: The interval between spindle assembly and anaphase based on three repetitions of the experiment; see Material and Methods for statistical analysis.

Figure 12. Mad2-GFP does not Re-Localize to Kinetochores in top2-B44 Cells.

a, Expression of Mad2 tagged with three tandem repeats of GFP localizes either diffusely or at the periphery of the nucleus in untreated wild type cells (left panel). Upon treatment with nocodazole Mad2-GFP re-localizes to a single discrete focus (right panel). b, Time-lapse analysis of Mad2-GFP from bud emergence (min 0:00) to cytokinesis (min 40:00) in untreated top2-B44 cells.