James E. Haber1, Moreshwar B. Vaze1, Sang Eun Lee1,
Ayelet Arbel-Eden1, André Walther1, Achille
Pellicioli2, Chiara Lucca2 and Marco Foiani2
1Rosenstiel Center and Department of Biology, Brandeis University
and Department of Biology, Waltham, MA 02254-9110 and 2 Istituto
F.I.R.C. di Oncologia Molecolare, 20122, Milan and Dipartimento di Genetica
e di Biologia dei Microrganismi, Universita' degli Studi di Milano, Italy.
As a model system for studying the effects of low-dose LET radiation, we have developed several strains of budding yeast in which there is a single, double-strand break (DSB) created by the galactose-inducible HO endonuclease. In some strains the lesion can be readily repaired by homologous recombination; in others, repair can only occur by nonhomologous end-joining (NHEJ). When HO expression is continuous, efficient NHEJ is futile because the re-joined site is cleaved continuously; consequently most cells fail to repair the DSB and will eventually die. Cells suffering one DSB arrest at the G2/M point in the cell cycle (prior to anaphase) by virtue of the DNA damage checkpoint. After 10-15 hours these DNA damaged cells adapt, and resume cell cycle progression, even though they still have a broken chromosome. Cells can divide several more times before they die because of loss of essential genetic information as the broken chromosome is slowly degraded (4 kb/hr). Similar results are obtained in disomic strains where recombination is prevented and where the broken chromosome is, after many generations, simply lost.
In these experiments we therefore model low-dose radiation by inflicting a single DSB per cell. Because it is at a single location in all cells, as opposed to random, stochastic damage caused by radiation and other agents, we can analyze in detail the metabolism of the broken DNA ends, their repair and the consequences of failure to repair the damage.
Adaptation
DNA damage arrest is accompanied by the hyperphosphorylation of Rad53p and Chk1p protein kinases by Mec1p. Other kinases such as Dun1p and Cdc5p are also activated in a RAD9, RAD17-dependent fashion. Cells with one DSB adapt; those with two DSBs do not, nor do yku70 cells with one DSB that is resected more rapidly to produce more single-stranded DNA (ssDNA). Thus the cell appears to monitor the amount of damage via the extent of ssDNA present.
Adaptation involves the loss of hyperphosphorylated forms of Rad53p and Chk1p. In yku70, tid1 and cdc5-ad mutants, cells cannot adapt and Rad53p kinase remains active for > 24 hr. A mutation in the ssDNA binding complex, RPA (rfa1-t11) suppresses this defect in yku70 and tid1 cells but not cdc5-ad.
We recently showed that maintenance of arrest depends on continued activity of Mec1p. We now report that Chk1p is not needed to establish the checkpoint but only to maintain it, whereas Rad9p, Rad17p and Mec1p are needed to establish the arrest. A chk1 cell arrests after a single DSB is created and resumes cell cycle progression about the time as wild type cells. But whereas a cdc5-ad strain never adapts, chk1 cdc5-ad strains do adapt.
Recovery
We constructed a new strain in which an HO-induced DSB at the LEU2 locus can be repaired by the process of single-strand annealing, but only after 6 hr. This was accomplished by placing a segment of the 5' end of the LEU2 gene 25 kb distal on the same chromosome, so that - at 4 kb/hr - resection of the DSB end would take 6 hr to reach this point and allow annealing to take place. During this time, cells all arrest at G2/M and, after repair, resume cell cycle progression about 2 hr later. More than 80% of cells survive. Resumption of cell cycle progression is faster than during the adaptation process and Rad53p becomes dephosphorylated much more synchronously and rapidly.
We report that the absence of the Srs2p helicase prevents cells both from recovering (after the DSB is repaired) as well as from adapting (when the damage persists). In contrast, other adaptation mutations (yku70, tid1, cdc5-ad) are all proficient in recovery.
Damage inducible genes caused by a single DSB.
In collaboration with A.P. Gasch, D. Botstein and P. Brown (Stanford University) we have found about 150 genes whose mRNA levels change significantly in response to a single HO-induced, unrepaired DSB, even in cells whose progression through the cell cycle was already prevented by pre-arresting them with nocodazole. We will now determine which of these genes are regulated by different genes in the DNA damage checkpoint.
In collaboration with J. Demeter and T. Stearns (Stanford University) we also found that the DNA damage signal that actually causes arrest is nuclear autonomous and thus is unlikely to involve new mRNA synthesis.
Use of the "recovery" strain to study the adaptive response.
In the adaptive response, if cells are treated with a low dose of radiation and then exposed to higher doses, they prove more resistant to the killing effects of the high dose. This could be due to increased expression of DNA damage inducible genes or to post-translational regulation of recombination proteins or possibly to arrest in the cell cycle. To study the adaptive response, we have used strain YMV2 so that ever cell receives one DSB that - 6 hr later - will be repaired. We find that if such cells are challenged with 0.1 to 0.3% MMS that cells experiencing this DSB they are1.5 to 2.5 times more resistant to these doses than are cells that were not first challenged with a single DSB. We believe this will be an excellent model system to study how a single DSB causes such a change in DNA damage resistance.