The authors have declared that no competing interests exist.
Conceived and designed the experiments: KSL YZ NS ZS. Performed the experiments: YZ NS ZS KSL. Analyzed the data: YZ KSL. Contributed reagents/materials/analysis tools: YZ KSL NS ZS. Wrote the paper: YZ KSL NS.
Inverted repeats capable of forming hairpin and cruciform structures present a threat to chromosomal integrity. They induce double strand breaks, which lead to gross chromosomal rearrangements, the hallmarks of cancers and hereditary diseases. Secondary structure formation at this motif has been proposed to be the driving force for the instability, albeit the mechanisms leading to the fragility are not well-understood. We carried out a genome-wide screen to uncover the genetic players that govern fragility of homologous and homeologous
Inverted repeats are found in many eukaryotic genomes including humans. They have a potential to cause chromosomal breakage and rearrangements that contribute to genome polymorphism and the development of diseases. Instability of inverted repeats is accounted for by their propensity to adopt DNA secondary structures that is negatively affected by the distance between the repeats and level of sequence divergence. However, the genetic factors that promote the abnormal structure formation or affect the ability of the repeats to break are largely unknown. Here, using a genome-wide screen we identified 38 mutants that destabilize imperfect human inverted
Long palindromic sequences (inverted repeats ∼100 bp or more each without a spacer or with a short spacer) present a threat to both prokaryotic and eukaryotic genome stability. In
Palindromic sequences can form hairpin and cruciform structures due to their intrinsic symmetry
Although the formation of hairpin and cruciform structures is deemed as the key initiation event for fragility at inverted repeats, the pathways that predispose eukaryotic cells to or provide protection against chromosomal breaks are still not well defined. Previously, deficiencies in Pol1, Pol3 and Rad27 proteins responsible for synthesis of the lagging strand during DNA replication were found to augment instability at inverted repeats
In this study, we carried out an unbiased genome-wide screen aimed at identifying the genetic factors controlling fragility of homologous and divergent
We systematically analyzed the effect of more than 6000 mutations on
(
We verified the effect of the identified mutants by recreating the hyper-fragile alleles in strains with the
38 mutants that exhibit a hyper-fragility phenotype in strains containing either 100% or 94% homologous
Previously, it has been shown that downregulation of or mutation in the DNA polymerases α and δ causes increased instability of inverted repeats
Besides the replication checkpoint surveillance mutants, the screen also revealed that GCRs mildly increase (2- to 4-fold) in Δ
The third group of mutants that amplify
The screen revealed that deletion of
Genetic background | GCR rate (×10−6) | |||
100% homologous | 94% homologous | |||
wild-type | 41 (30–52) |
5 (4–6) | ||
Replication mutants | ||||
TET- |
250 (100–280) | 6 |
170 (80–180) | 34 |
TET- |
240 (210–270) | 6 | 130 (90–150) | 26 |
TET- |
470 (380–500) | 11 | 100 (80–110) | 20 |
TET- |
460 (390–640) | 11 | 82 (72–102) | 16 |
TET- |
370 (290–390) | 9 | 69 (60–73) | 14 |
TET- |
280 (170–380) | 7 | 34 (21–44) | 6 |
TET- |
140 (110–160) | 3 | 38 (23–47) | 8 |
Δ |
720 (370–820) | 18 | 72 (61–85) | 14 |
TET- |
340 (260–470) | 8 | 170 (130–200) | 34 |
TET- |
150 (140–240) | 4 | 41 (16–62) | 8 |
TET- |
110 (80–230) | 3 | 62 (31–76) | 12 |
TET- |
120 (80–200) | 3 | 9 (9–12) | 3 |
Δ |
600 (450–920) | 15 | 90 (60–240) | 18 |
Δ |
240 (190–300) | 6 | 32 (28–36) | 6 |
Checkpoint response genes | ||||
Δ |
250 (170–300) | 6 | 39 (31–48) | 8 |
Δ |
370 (270–530) | 9 | 27 (22–37) | 5 |
Δ |
180 (160–250) | 4 | 14 (13–16) | 3 |
Δ |
140 (130–190) | 3 | 12 (11–17) | 2 |
Δ |
200 (160–220) | 5 | ND |
ND |
Δ |
43 (40–49) | 1 | ND | ND |
Helicase | ||||
Δ |
410 (300–490) | 10 | 35 (26–44) | 7 |
Δ |
480 (410–560) | 12 | ND | ND |
Telomere protection genes | ||||
TET- |
140 (120–230) | 3 | 13 (9–16) | 3 |
TET- |
120 (70–140) | 3 | 14 (11–18) | 3 |
Double strand breaks repair genes | ||||
Δ |
420 (370–440) | 10 | 210 (170–230) | 42 |
Δ |
400 (370–430) | 10 | 220 (200–250) | 44 |
Numbers in the brackets are 95% confidence intervals.
Fold increase in GCR rates in mutants compared to wild-type strains.
Not determined.
The fifth group of hyper-fragile mutants consisted of TET-
Previously, we demonstrated that the Mre11-Rad50-Xrs2 complex and the Sae2 protein are required to open hairpins to initiate DSB repair at inverted repeats
In the wild-type strain, DSBs induced by
We compared the levels of chromosomal breaks in the wild-type strain containing 100%
Yeast genomic DNA embedded in agarose plugs was digested with either AflII (
No DSBs were observed in the presence of Sae2 in both wild-type and mutant strains, likely due to hairpin opening and robust resection of the breaks. However, DSBs were readily detected in Δ
To test the premise of hairpin-capped breaks in the mutants experimentally, the DSB fragments in TET-
Yeast DNA embedded in agarose plugs was digested with AflII (
The symmetry of the breaks and the presence of covalently-closed hairpins at the DSB termini suggest that the final steps in breakage in mutants and wild-type are the same and include cruciform formation and resolution.
The screen revealed that mutants deficient in the DNA replication pathway comprise the major group that augments fragility at
DSB detection was carried out as described in
Genetic background | GCR rate (×10−6) | Fold increase over wild-type |
WT | 41 (30–52) |
1 |
Δ |
37 (27–50) | 1 |
TET- |
460 (390–640) | 11 |
TET- |
88 (58–108) | 2 |
TET- |
240 (210–270) | 6 |
TET- |
63 (46–79) | 1 |
Numbers in brackets are the 95% confidence interval.
To gain better insight into the mechanism underlying
DNA samples from the wild-type, TET-
The red helixes, blue pacman and orange hexamer depict the inverted repeats, the putative nuclease and the DNA replication helicase, respectively. In the case of normal replication, cruciform structure might form outside of S-phase as a result of chromatin packing or remodeling. On the other hand, long single-stranded DNA exposed due to compromised replication would facilitate the formation of a hairpin, which could further be converted into cruciform structure via template switch by Rad51. The intermediates of template switching present a strong obstacle for the replication machinery that is manifested as replication block in the replication-deficient strains. The Sgs1-Top3-Rmi1 dissolvosome might participate in unwinding the hairpin. Once formed, the cruciform structure might be attacked by the putative nuclease, leading to DSBs at the IRs.
Overall, these data reveal an important role of homologous recombination in promoting DSB formation at inverted repeats, specifically in replication-deficient mutants.
Palindromic sequences are strong inducers of DSBs and rearrangements in both prokaryotes and eukaryotes. The two distinct events that trigger fragility are considered to be the formation of either hairpin or cruciform structures at the repeats. In this study we found that when replication is compromised, replication delay imposed by inverted repeats is channeled into cruciform resolution via the action of homologous recombination pathway. These data led us to propose that the transition from hairpin to cruciform formation through Rad51-mediated template switching is the mechanism for fragility operating in cells under replication stress.
Inverted repeat-induced GCRs can be augmented in mutants that either influence secondary structure metabolism or alter repair of the broken chromosome. Previous studies from our lab have demonstrated that
The mutants identified in the screen that increase DSB formation and GCRs at
The screen revealed that depletion of the major components of the replication fork responsible for synthesis of both leading and lagging strands increases
Deletion of
The deficiencies described above are expected to create optimal conditions for the formation of a hairpin that impedes replication progression. The hairpin might be formed at lower frequencies in replication-proficient cells as well. In both replication-proficient and -deficient strains, the secondary structure or the arrested fork might trigger the activation of checkpoint response required to recruit proteins to remove the hairpin and promote replication restart (
It seemed reasonable to suggest the existence of helicases recruited to remove hairpins at the arrested fork. Indeed, the screen identified the Sgs1-Top3-Rmi1dissolvasome. Although Δ
An interesting group of mutants that destabilize
In wild-type strains carrying inverted repeats, the deduced mechanism of breakage is cruciform-resolution by a putative nuclease that cuts symmetrically at the base of the two hairpins. This generates two hairpin-capped molecules that are present in equimolar ratios
Based on our finding that deletion of
It is important to note that the Rad51 effect is specific in situations where replication is compromised. In replication-proficient strains, breaks and GCRs are not affected by Rad51 status, indicating that another mechanism for cruciform-formation exists. It is possible that in wild-type strains a homologous recombination-independent template switching mechanism leading to fragility operates, or that the cruciform formation is unrelated to replication. The latter hypothesis is supported by our recent finding that hairpin-capped breaks in the wild-type strain preferentially occur in G2 phase of the cell cycle (Sheng
Based on this study, we propose that in the human population, the carriers of hypomorphic alleles for the BLM-hTOPOIIIα-hRMI1-hRMI2 dissolvasome and proteins involved in DNA replication, replication-pausing checkpoint surveillance, Fe-S cluster biogenesis, telomere maintenance and protection might be susceptible to inverted repeat-induced breaks and carcinogenic GCRs. Importantly, the status of these proteins determines the stability of imperfect repeats with a spacer and divergent arms that are present in the human genome
yTHC, DAmP and YKO collections were purchased from Open Biosystems. All other strains in this study are derivatives of BY4742 (Open Biosystems). The genotype of the query strains for the screen is: MATα, Δ
The effect of mutant alleles identified from the screen was verified in derivatives of YKL36 that carries the GCR assay and has the following genotype: MATa, Δ
In strains used for DSB analysis,
The screen was carried out as described in Zhang et al., 2012
Yeast cells were grown on YPD plates for 3 days. For each strain, a minimum of 14 independent colonies were taken to perform fluctuation test to estimate GCR rates. Appropriate dilutions of cells were plated on YPD and canavanine-containing plates to determine the GCR frequency. The GCR rates were calculated using the formula μ = f/ln(Nμ) as described in Drake, 1991
Yeast cells from overnight cultures were embedded into 0.8% low-melting agarose plugs at a concentration of 24×108 cells/ml. The plugs were treated with 1.5 mg/ml lyticase for 3 hr, followed by overnight 1 mg/ml proteinase K treatment. For restriction digestion of the DNA, the plugs were washed twice with 1 X TE buffer (10 mM Tris-Cl [pH 8.0], 0.1 mM EDTA) for 30 min, treated with 1 mM PMSF for 1 hr, washed with distilled water for 1 hr and equilibrated with restriction buffer for 20 min. Each plug (∼40 µl) was digested with 50 units of AflII or BglII for 16 hr. Digested plugs were loaded in a 1% (AflII digestion) or 0.7% (BglII digestion) agarose gel, respectively, and run in 1 X TBE for 18 hr. The gels were treated with 0.25 N HCl for 20 min, alkaline buffer (1.5 M NaCl, 0.5 M NaOH) for 30 min and neutralization buffer (1.5 M NaCl, 1 M Tris [pH 7.5]) for 30 min. The gels were then transferred in 10 X SSC to charged nylon membrane for 2 hr through a Posiblotter (Stratagene). Southern hybridization was carried out using P32-labeled
Yeast plugs were prepared and digested as described above. Neutral/neutral and neutral/alkaline gel analysis was performed as previously described with small modifications
2D gel analysis was carried out as previously described in Brewer and Fangman, 1987
The genome-wide screen scheme. In the query strains, the chromosomal arm containing the GCR assay was marked by the
(TIF)
Detection of breakage intermediates in a subset of hyper-GCR mutants. Genomic DNA embedded in agarose plugs were digested by AflII (
(TIF)
Detection of breakage intermediates in Δ
(TIF)
Rad51 dependent breakage formation in Δ
(TIF)
DSB accumulation in TET-
(TIF)
Hyper-GCR mutants identified in the genome-wide screen. a+ shows mutants identified as hyper-GCR alleles from the libraries indicated. b* shows mutant alleles whose effect on
(DOCX)
Effect of
(DOCX)
Primers used in the study.
(DOCX)
We thank Dr. Natasha Degtyareva, Dr. Matthew Torres, and Anastasiya Lobacheva for critical reading of the manuscript and helpful discussions.