We have utilized DNA transfer and recombinant DNA techniques to probe DNA double-strand break repair in the human ionizing radiation-sensitive genetic syndrome ataxia-telangiectasia (A-T). Using restriction enzyme-generated double-strand breaks in the coding sequence of a selectable gene we have detected a significantly greater frequency of mis-repair of such breaks in a permanent A-T cell line compared with cell lines of normal radiosensitivity. This mis-repair in A-T can plausibly explain many of the clinical features of the disease but was insufficiently detailed to address the broad problem of DNA repair mechanisms relevant to ionizing radiation-induced damage. To extend these observations of DNA double-strand break mis-repair we have now applied this type of repair assay to novel, de novo induced mammalian X-ray-sensitive cell lines and to appropriate Escherichia coli mutants. In both cellular systems we have now found some equivalence to the A-T repair defect. In particular, studies on one E. coli mutant have provided evidence suggesting an involvement of a topoisomerase activity in DNA double-strand break misrepair, which may be relevant to the biochemical defect in A-T.

Mechanisms of DNA repair may be resolved through detailed in vitro studies on cellular mutant phenotypes that, through specific DNA-repair deficiencies, exhibit increased sensitivity to DNA damaging agents. However, it is a paradox that while radiobiological studies were instrumental in the discovery of cellular repair phenomena we now have a much better understanding of the mechanisms of repair of ultraviolet light-induced and chemically induced DNA damage than for the repair of damage induced by ionizing radiation. However, the advent of recombinant DNA techniques and their application to radiobiological problems is now beginning to change this rather dismal picture and in this chapter we summarize recent studies in this laboratory on radiosensitive cellular mutant phenotypes that may provide new insights into mechanisms of DNA repair following exposure to ionizing radiation.

While the main focus of these studies has been the resolution of the putative DNA- repair defect in ionizing radiation-sensitive cells cultured from patients with the human genetic disorder, ataxia-telangiectasia, we also discuss recent unpublished data from studies on other mammalian and bacterial radiosensitive mutants that contribute to our understanding of the problem.

Ataxia-telangiectasia (A-T) is an autosomal recessive genetic disease presenting during childhood and associated with a number of progressive and variable clinical features. These include neuromotor dysfunction associated with neuronal loss (Sedgewick, 1982), variable immunodeficiency (Waldman, 1982), susceptibility to sinopulmonary infection, a high frequency of lymphoreticular neoplasia and other cancers (Spector et al. 1982), high levels of serum alpha fetoprotein (Waldman & McIntire, 1972), cytogenetic abnormalities of peripheral lymphocytes (Taylor, 1982) and the development of occulocutaneous telangiectases. Following adverse reaction of A-T patients to conventional radiotherapy (see Cox, 1983), cells cultured from a large number of patients have been shown to be hypersensitive to ionizing radiation both in terms of induction of chromosome aberration (Taylor, 1982) and loss of reproductive capacity (Taylor et al. 1975; Cox et al. 1978).

The observation that A-T cells are specifically sensitive to ionizing radiation has been the motivation for many cellular and biochemical studies, which have attempted to relate A-T radiosensitivity to specific defects in DNA metabolism. Many of these studies have been collated (Bridges & Harnden, 1982), forming an important resource for those interested in the broad aspects of the disorder. However, the published studies on A-T present a wide array of observed cellular phenomena with little consensus as to the underlying deficiency present. Thus whilst there is general agreement as to its clinical description, research into A-T has not provided a coherent biochemical or enzymic definition to explain the cellular radiosensitivity, let alone the clinical features of the disease. Of the various radiobiological characteristics described there is, though, broad agreement on a few cellular manifestations of the genetic defect in A-T. These include the lack of post-irradiation DNA synthesis delay that is observed in cells of normal radiosensitivity, a normal sensitivity to ultraviolet light (u.v.), relatively less hypersensitivity to ionizing radiations of increasing linear energy transfer (LET), and a hypersensitivity to strand breaking agents such as bleomycin (Lehmann, 1982; Cox, 1982). Most importantly, and central to an understanding of A-T, following exposure to ionizing radiation A-T cells rejoin the radiation-induced DNA double-strand breaks as efficiently as normal cells (Lehmann, 1982). This characteristic distinguishes A-T cells from other reported mutants hypersensitive to ionizing radiation, which are defective in the rejoining of ionizing radiation-induced DNA double-strand breaks, i.e. xrs mutants of hamster cells (Kemp et al. 1984), radt>2 series mutants in yeast (Resnick & Martin, 1976) and rorA and recN mutants in Escherichia coli (Glickman et al. 1971; Picksley et al. 1984). DNA double-strand breaks constitute a major genetic lesion with great potential to cause recombination, rearrangements and deletions. An inability to repair such lesions, which are directly induced in the cellular DNA by ionizing radiation, would logically explain a cellular sensitivity to ionizing radiation. Such a repair defect would not necessarily also impart increased sensitivity to u.v. as DNA double-strand breaks are not efficiently induced directly by u.v. With one possible exception (Coquerelle & Weibezahn, 1981) A-T cells are quantitatively proficient in DNA double-strand break repair and, overall, it would seem that a gross deficiency in such repair cannot account for A-T radiosensitivity.

Accepting that DNA double-strand breaks are one of the most potent DNA lesions induced by ionizing radiation it is important to recognize that the conventional assay techniques (see Lehmann, 1982) used to measure repair of such lesions are only quantitative and cannot comment on the fidelity of the whole process. We have developed an experimental approach that enables the fidelity of DNA double-strand break repair to be examined using DNA-mediated gene transfer techniques (Cox et al. 1984). This approach involves the introduction of a unique double-strand break in the coding sequence of a dominant, selectable gene (gpt) in a normally circular DNA vector by restriction enzyme cleavage. For more detailed studies an improved vector (Fig. 1) was subsequently developed and utilized (unpublished observations). This DNA vector contains a linked second dominant and selectable gene (G418R), which acts as a control for DNA transfer and gene expression. If the double-strand break introduced into the non-selected gene is not rejoined or is rejoined incorrectly within the recipient cell, then the gene will be inactive and this can be detected by subsequent lack of expression of its function. Because this technique involves gene transfer it is unfortunately not easily applied to primary diploid cell cultures and is most suitable for transformed cell lines (Debenham et al. 1984). However, using this approach we have shown that the one transformed A-T cell line available at present (AT5BIVA) rejoins DNA double-strand breaks with at least a 7- to 10-fold higher frequency of mis-repair than cell lines of normal ionizing radiation sensitivity (Cox et al. 1984, 1986).

Fig. 1.

A schematic representation of the experimental protocol to detect mis-repair of a restriction enzyme generated break in the gpt gene. (1) pPM17 or pPMHSVl6 DNA is linearized by cleavage with either KpnI or AcoRV. Both enzymes cut in the coding sequence of the gpt gene. (2) The linearized DNA is precipitated onto a monolayer of cells via a standard calcium phosphate procedure. (3) Cells that take up the DNA (often more than one copy) and express the undamaged G418R gene will grow in medium containing G418. The transferred DNA is eventually integrated into the cellular DNA, so that on division each daughter cell will contain the same configuration of this DNA. (4) If during the growth in G418 medium the cell has correctly joined one or more Kpn I (or EcoRV) sites the cell will have an active gpt gene and survive in XHATM medium (Mulligan & Berg, 1981). (S) If no correct joining of Kpnl (or AcoRV) sites occurs the recipient cell will not have an active gpt gene and this cell (colony) will die in XHATM medium.

Fig. 1.

A schematic representation of the experimental protocol to detect mis-repair of a restriction enzyme generated break in the gpt gene. (1) pPM17 or pPMHSVl6 DNA is linearized by cleavage with either KpnI or AcoRV. Both enzymes cut in the coding sequence of the gpt gene. (2) The linearized DNA is precipitated onto a monolayer of cells via a standard calcium phosphate procedure. (3) Cells that take up the DNA (often more than one copy) and express the undamaged G418R gene will grow in medium containing G418. The transferred DNA is eventually integrated into the cellular DNA, so that on division each daughter cell will contain the same configuration of this DNA. (4) If during the growth in G418 medium the cell has correctly joined one or more Kpn I (or EcoRV) sites the cell will have an active gpt gene and survive in XHATM medium (Mulligan & Berg, 1981). (S) If no correct joining of Kpnl (or AcoRV) sites occurs the recipient cell will not have an active gpt gene and this cell (colony) will die in XHATM medium.

The simplest explanation for such a mis-repair phenotype is that either exonuclease activities, which degrade DNA at double-strand breaks, are over-produced, or factors that normally protect DNA termini from such degradation are altered or lacking. Since no gross elevation of exonuclease activity was detected in crude extracts of A-T cells (data not shown) some form of deficiency of protection of DNA termini seems a more plausible explanation. In order to explain the observed misrepair in A-T it was postulated that an equilibrium exists, at DNA termini, between their correct rejoining and exonuclease degradation (Cox et al. 1986). In normal cells a protein factor/complex is present that concomitantly protects the termini from exonuclease activity and promotes their correct repair. In A-T cells the equilibrium is perturbed by the alteration or deficiency of this factor so that sequence information is more frequently lost at the termini of a DNA double-strand break than in the wild-type cell. Consequently, a variable, and unpredictable, degree of misrepair may occur at any DNA double-strand break.

Are the data on which this hypothesis is based sufficient to warrant further extrapolation and experimentation? Given the controversy surrounding the basis of the A-T disorder it is important to establish whether the results obtained are due to experimental artefacts or are essentially unique to the single transformed A-T cell line investigated to date. Since DNA transfer remains a poorly understood cellular process it may be that this component of the experimental approach is the most vulnerable to misinterpretation.

The most relevant aspect of the DNA transfer process that might artefactually produce an apparent mis-repair phenotype in AT5BIVA is the copy number of linearized vector taken up/maintained by the normal and A-T cell lines. A single functional copy of the gpt gene (the gene cleaved in the mis-repair assay) is sufficient to provide resistance to the selective medium (XHATM; Thacker et al. 1983). Other things being equal, a cell receiving or maintaining many copies of the linearized vector has a higher probability of producing at least one correctly repaired gene than does a cell receiving a small number of copies. Southern blot analysis in the initial study showed that for both the normal cell line used (MRC5CV) and AT5BIVA only a small number of copies of the vector (pPM17) were found per cell (Cox et al. 1986). We have subsequently found that the effective copy number can be further reduced to approximately one by using a vector (pPMHSVl6) with a more efficiently expressed G418R gene such that one copy of this gene is now sufficient to establish cell growth in the initial selection following DNA transfer (Fig. 1; and unpublished observations). The G418R gene in pPM17 is not sufficiently active to achieve this in single copy. Consequently, using pPMHSVl6 more transformant colonies are obtained for a given amount of transferred DNA reflecting successful growth of cells with only one or a few copies of the vector. With pPM17 such colonies were only poorly viable in the selective medium. Therefore, the use of the pPMHSVl6 vector should more precisely monitor the nature of repair of each restriction enzyme generated break.

It was found that monitoring the rejoin fidelity of the Kpn\ site using pPMHSVl6 as the vector, rather than pPM17, reduces the observed wild-type (MRC5CV) fidelity from ≈88% to 56% as scored by gpt expression in transformed colonies (Table 1). It has long been known that transferred DNA molecules are subject to DNase activity during DNA transfer and, for MRC5CV, the effect on DNA transfer of reducing the copy number of transferred DNA molecules it consistent with the activity of degradative cellular processes (Goebel & Schiess, 1975; Wigler et al. 1978). Although the same non-specific degradative events appear to be occurring in AT5BIVA, reducing the observed fidelity from ≈12% to ≈6%, the differential rejoin fidelity, thought to be a consequence of the genetic defect in A-T, is still clearly apparent. As far as is known MRC5CV and AT5BIVA were not derived from related donors, and numerous other genetic differences may also exist that influence the observed repair fidelity. Additional genetic differences may also derive from the establishment of these cell lines by simian virus 40 (SV40) infection. However, other unrelated SV40 established cell lines of normal sensitivity to ionizing radiation were examined in earlier studies and found to give results comparable to MRC5CV (Cox et al. 1984, 1986). Given the increased sensitivity of using the pPMHSVl6 vector the repair of the Kpnl site in XP12ROSV cells has been re-examined. These cells are extremely sensitive to u.v. (from a patient with xeroderma pigmentosum, group A; see Lehmann & Stevens, 1980) but have a normal response to ionizing radiation. In these studies it was found that although the fidelity of rejoining of double-strand breaks in XP12ROSV cells was not quite as high as in MRC5CV cells, it was more than five times higher than in AT5BIVA cells (Table 1). The difference between MRC5CV and XP12ROSV responses may merely reflect a natural range in repair fidelity that, on the basis of the data presented here and elsewhere, remains clearly distinct from the A-T cell response.

Table 1.

Rejoin fidelity of linearized pPMl 7 and pPMHSVl6 vectors (XHATMR/G418R)

Rejoin fidelity of linearized pPMl 7 and pPMHSVl6 vectors (XHATMR/G418R)
Rejoin fidelity of linearized pPMl 7 and pPMHSVl6 vectors (XHATMR/G418R)

Unfortunately, no other suitable A-T cell lines are available at present to extend and confirm the results obtained with AT5B1VA. However, it is still possible to ask whether the mis-repair defect observed in AT5BIVA is unique or whether parallels exist in other equivalent mutant cell lines. A-T cells clearly differ from other published examples of ionizing radiation-hypersensitive cell lines in being competent at rejoining ionizing radiation induced DNA double-strand breaks. Recently, however, four new X-ray-sensitive mammalian sub-lines derived from V79 hamster cells have been isolated in this laboratory (Jones, unpublished), which have been placed in different genetic complementation groups from previously characterized mutants such as the xrs series (Jeggo & Kemp, 1983) or EM 9.1/EM 7.2 (Thompson et al. 1980). The vector-mediated assay has recently been applied to two of these mutant lines that show the greatest sensitivity to ionizing radiation, at present termed A4 (D37 for X-rays of 134 rad) and Cll (D37 for X-rays of 141 rad), as well as their parental cell line V79 (D37 for X-rays of 420 rad (Jones, unpublished). Obviously these cell lines all have the same genetic background and allow for more direct comparison than is possible in the available human cell lines. It appears that in hamster cell lines (V79 and CHO; Table 2, and unpublished observations) there is a high degree of degradative action at the Kpnl site in the gpt gene such that only approximately 10% of transformed colonies have functional gpt gene activity. The Kpnl site is clearly subject to mis-repair in rodent cells, which may mask mis-repair in the ionizing radiation-sensitive cell lines. However, even with this possible masking effect A4 mutant cells have been reproducibly observed to show a fivefold reduction in the fidelity of rejoining at the Kpnl site compared with V79 (Table 2). This mis-repair phenomenon in A4 cells is also apparent if the gpt gene is linearized by £coRV, which cleaves DNA to leave flush termini as opposed to the 3′ single-strand ends generated by Kpnl. Thus A4 shows a mis-repair that appears independent of the nature of the DNA double-strand break and is therefore qualitatively equivalent to the defect in human A-T cells. Conversely, Cll and xrsl (data not shown), although hypersensitive to ionizing radiation, are both as proficient as their parental cell lines at correctly rejoining restriction-enzyme-generated DNA double-strand breaks at the Kpnl and EcoRN sites. This is particularly intriguing as xrs 1 has been shown to be quantitatively defective in the repair of ionizing radiation-induced DNA double-strand breaks (Kemp et al. 1984).

Table 2.

Rejoin fidelity of linearized pPMHSVl6 vector (XHATMR,/G418R,/SUP>)

Rejoin fidelity of linearized pPMHSVl6 vector (XHATMR,/G418R,/SUP>)
Rejoin fidelity of linearized pPMHSVl6 vector (XHATMR,/G418R,/SUP>)

Comparable studies have also been undertaken in three E. coli mutants specifically sensitive to ionizing radiation (rorA, rorB and recN). Mutants rorA and recN have been shown to be defective in the rejoining of ionizing radiation-induced doublestrand breaks (Glickman et al. 1971; Picksley et al. 1984). Mutant rorB is a newly isolated mutant that appears to be competent with respect to ionizing radiation- induced DNA double-strand break repair (unpublished observations). Unlike rorA and recN, rorB appears to rejoin erroneously restriction enzyme-generated doublestrand breaks in plasmid DNA molecules (unpublished observations). Thus rorB, like AT5BIVA appears to be an example of a cellular phenotype showing competent rejoining of ionizing radiation-induced DNA double-strand breaks, but which may involve low fidelity in the rejoining process. Conversely, rorA and recN competently rejoin the simple, restriction enzyme-generated double-strand breaks in plasmid DNA molecules (unpublished observations). Thus rorB, like AT5BIVA appears to be an example of a cellular phenotype showing competent rejoining of ionizing radiation-induced DNA double-strand breaks, but which may involve low fidelity in the rejoining process. Conversely, rorA and recN competently rejoin the simple, restriction enzyme-generated double-strand breaks with high fidelity but are deficient in processes rejoining a proportion of perhaps more complex ionizing radiation-induced strand breaks. It is possible that the hamster mutant xrsl falls into a repair class similar to that of rorA and recN.

Is there any clue as to the nature of the two distinct repair pathways (described above) acting on DNA double-strand breaks? The pathway deficient in E. coli mutants rorA and recN, and also yeast mutants rad52. etc., has for some time been argued to involve DNA recombination, which in turn had been thought to be the major process by which cells repair ionizing radiation-induced DNA double-strand breaks (Resnick, 1976).

It has been shown here that a second aspect of DNA double-strand break repair that influences the fidelity of the process is also of major importance. A possible clue to the enzymic basis of this mis-repair was the surprising finding (unpublished observations) that two independently isolated rorB mutants appear hypersensitive to coumermycin Al, a specific antagonist of the ATP binding subunit (gyrB) of E. coli gyrase (Sugino et al. 1978). Mutants rorA and recN have a normal sensitivity to coumermycin Al. rorB has a normal sensitivity to nalidixic acid, which inhibits the other subunit of E. coli gyrase (gyrA; Gellert et al. 1977), suggesting rorB has a very specific change in its gyrase activity. However, the gene complementing rorB is not gyrB and the cloned gyrB gene when introduced into rorB cells does not complement the radiation sensitivity of this mutation. Thus, it seems that the rorB gene codes for a factor that alters gyrase activity.

The similarities to rorB, with respect to a DNA mis-repair phenotype associated with a hypersensitivity to ionizing radiation, prompted the question whether A-T cells and A4 cells also showed alteration in a topoisomerase type II activity that is the equivalent of gyrase in E. coli. Novobiocin is the mammalian equivalent of coumermycin Al, inactivating topoisomerase type II in mammalian cells (Gellert, 1981), although it is thought to affect other cellular processes without acting -via topoisomerase (e.g. see Gottesfield, 1986). Fig. 2 shows the novobiocin sensitivities of MRC5CV, XP12ROSV and AT5BIVA cells and it is clear that AT5BIVA is considerably more novobiocin-sensitive than MRC5CV and XP12ROSV. These analyses of pooled data were made difficult by the form of the cellular response to novobiocin and there was some variation in individual sensitivity with different drug preparations, nevertheless there were consistent differences in each experiment between AT5BIVA and MRC5CV/XP12ROSV.

Fig. 2.

MRC5CV (○), XP12ROSV (▫) and AT5BIVA (•) cells were plated into 10 ml medium in 9-cm diam. dishes containing freshly prepared novobiocin at the indicated concentrations. The cells were incubated, without change, for 14 days at 37°C. Viable colonies were scored relative to the viability of the cells in medium without novobiocin (approx. 36%, 45% and 40%, respectively). Data presented are the means (±S.D., shown by bars, where appropriate) from three experiments for MRC5CV and AT5BIVA, and two experiments for XP12ROSV.

Fig. 2.

MRC5CV (○), XP12ROSV (▫) and AT5BIVA (•) cells were plated into 10 ml medium in 9-cm diam. dishes containing freshly prepared novobiocin at the indicated concentrations. The cells were incubated, without change, for 14 days at 37°C. Viable colonies were scored relative to the viability of the cells in medium without novobiocin (approx. 36%, 45% and 40%, respectively). Data presented are the means (±S.D., shown by bars, where appropriate) from three experiments for MRC5CV and AT5BIVA, and two experiments for XP12ROSV.

In a similar study the novobiocin sensitivities of V79, A4 and Cll cells were also studied. V79 hamster cells were also found to show rapidly decreasing survival with increasing novobiocin concentration. V79 and mutant Cl 1 did not differ markedly in their response to the drug but mutant A4 was clearly more novobiocin-sensitive (unpublished observations). Thus a consistent picture may be arising in which the repair of at least some DNA double-strand breaks involves or is associated with a topoisomerase type II activity and, overall, the studies summarized here imply that alteration in this process may cause the relevant breaks to be repaired with relatively low fidelity.

The hypothesis that DNA double-strand breaks are subjected to a high frequency of mis-repair in A-T cells appears from the data discussed here to receive plausible support from studies with other radiosensitive cellular phenotypes. How well does this model of repair specifically address the many aspects of the A-T phenotype? The hypothesis can explain, in broad terms, cellular radiosensitivity in A-T and some aspects of the more readily understood clinical features of this genetic disorder (Cox et al. 1986). Since the misrepair is focussed on DNA double-strand breaks it is also consistent with the lack of A-T hypersensitivity to u.v. exposure and the majority of base-damaging chemical agents. A-T cells are sensitive to bleomycin (see Lehmann, 1982), which directly induces single-strand breaks and thus could also produce double-strand DNA breaks from two closely adjacent scissions. While it is therefore possible to consider the repair of DNA double-strand breaks as the principal biochemical problem for A-T, some degree of deficiency in single-strand break repair fidelity is quite feasible and, indeed, cytogenetic observation of excessive chromatid type damage in irradiated GQ A-T cells lends some support to this possibility (Taylor, 1982).

A-T cells are relatively less hypersensitive to radiations of increasing linear energy transfer (high LET; Cox, 1982). High LET radiation tracks (e.g. those of a particles) will tend to cause greater deposition of energy in a greater number of bases on traversal of a DNA duplex. This increase in energy depositions must result in many more chemical interactions such that sites of high LET radiation-induced double-strand breaks will probably be associated with aberrant chemical structures such as intra- and inter-strand cross-linking, and cross-linking with neighbouring proteins. In a genetic context, many of these lesions in both normal and A-T cells may be irreparable through simple rejoining, and only recombination processes would be expected to reconstitute the sequence information lost from both DNA strands. Consequently, as the inherent ‘rejoinability’ of lesions decreases so the difference in sensitivity of A-T cells to such radiation, relative to normal cells would decrease. Thus the mis-repair hypothesis can accommodate the reduced relative sensitivity of A-T cells to radiations of increasing LET (Cox, 1982).

However, it is not so obvious why a perturbation of the fidelity of DNA doublestrand break repair should influence the progression of the cell through the DNA replicative cycle. Following exposure to ionizing radiation, normal cells do not continue their progression through S-phase but are delayed for several hours (Houldsworth & Lavin, 1980; Edwards & Taylor, 1980). In A-T cells S-phase progression is not greatly influenced by such exposure. It is thought that one DNA strand break is enough to stop initiation of DNA synthesis within a group of replicons (Povirk, 1977). However, as the majority of A-T cell strains are both kinetically and quantitatively competent at rejoining double-strand breaks the explanation does not seem to be related to the repair or mis-repair of DNA doublestrand breaks. An explanation may be found by examining possible subsidiary rôles of the cellular factor that is deficient in A-T. Is it possible, for example, that following irradiation normal DNA replication is retarded by the consumption of a factor normally required at sites of DNA double-strand break repair? To address this possibility we need to know what factors are associated with the process of correctly rejoining DNA double-strand breaks. The previous section of this chapter has indicated the possibility that a topoisomerase type II activity could have an important rôle in this process.

Topoisomerase type II function in cells is thought to regulate the conformation of the DNA by generating transient DNA double-strand breaks and passing the double helix through them, thus altering the degree of supercoiling of the DNA. During this process the DNA termini are protected (Morrison & Cozzarelli, 1981). It is thus tempting to speculate that a topoisomerase type II activity, particularly its termini protection activity, is in some way utilized to promote the correct rejoining of DNA double-strand breaks including those produced by ionizing radiation. In A-T cells it is this function that would be altered. However, the mutation in A-T may not reside in a gene structurally constituting topoisomerase type II activity and may act indirectly. On this point it may be relevant that the gene complementing the E. coli rorB mutation is neither gyr\ or gyrB (the only genes required to determine structurally DNA gyrase activity in A1, coli).

The increased sensitivity to novobiocin/coumermycin of the ionizing radiation-sensitive mammalian and bacterial cells characterized by a mis-repair of DNA double-strand breaks may reflect a defective and potentially lethal process in the cell normally counteracted or controlled by topoisomerase type II activity. Recently, and independently, it has been shown that AT5BI cells are hypersensitive to the drug VP16, which mediates its cellular toxicity via the topoisomerase type II enzyme (P. J. Smith, personal communication). These studies reveal that the intrinsic sensitivity of topoisomerase type II activity to VP 16 in A-T cells is normal but that accumulation of VP16-induced DNA damage is greater than normal. Thus a process controlling topoisomerase type II enzyme level or its sites of activity is proposed to be altered.

The A-T defect, as discussed earlier, causes a lack of control of DNA replication following exposure to ionizing radiation. In the normal cell, replication is curtailed for several hours following exposure but the mechanism involved is unknown. The unwinding of condensed (supercoiled) chromosome structures by topoisomerase II (and possibly also topoisomerase I) probably plays an essential role in replicative DNA synthesis. It may therefore be argued that in normal cells, following exposure to ionizing radiation, topoisomerase type II activity is temporarily consumed in protecting DNA double-strand breaks from degradation, thus hindering initiation of new replicons. In A-T cells the factor interacting with topoisomerase type II activity for DNA double-strand break repair is defective, allowing continued availability of topoisomerase type II activity for the de-condensation of replicons. On this basis it may be expected that, in A-T, DNA replication will continue without major perturbation in the presence of radiation-induced DNA strand breaks.

We present initial evidence and arguments to link three phenomena: a cellular hypersensitivity to ionizing radiation, an ability to rejoin DNA double-strand breaks but with a higher frequency of misrepair than comparable wild-type cells, and a sensitivity to novobiocin that may reflect altered topoisomerase type II function. The evidence is derived from cell lines hypersensitive to ionizing radiation from E. coli, hamsters and humans. Taken together, these data highlight DNA double-strand breaks as crucial lesions induced by ionizing radiation and link their correct repair with post-irradiation DNA synthesis through regulation of topoisomerase II enzyme activity. Further to this, we suggest that the primary genetic defect in radiosensitive cells from human ataxia-telangiectasia patients may reside in the regulation of topoisomerase II activity.

This study is supported in part by C.E.C. Contract 816-144-UK.

Bridges
,
B. A.
&
Harnden
,
G. D.
(
1982
).
Ataxia-telangiectasia. New York: Wiley
.
Coquerelle
,
T. M.
&
Weibezahn
,
K. F.
(
1981
).
Rejoining of DNA double strand breaks in human fibroblasts and its impairment in one ataxia-telangiectasia and two Fanconi strains
.
J. supramolec. Struct, cell. Biochem.
17
,
369
376
.
Cox
,
R.
(
1982
). A cellular description of the repair defect in ataxia-telangiectasia. In
Ataxiatelangiectasia
(ed. B. A. Bridges &
D. G.
Harnden
), pp.
141
153
.
New York
:
Wiley
.
Cox
,
R.
(
1983
). In vivo and in vitro radiosensitivity in ataxia-telangiectasia. In
The Biological Basis of Radiotherapy
(ed.
G. G.
Steel
, G. E. Adams &
M. J.
Peckham
), pp.
105
112
.
Oxford
:
Elsevier
.
Cox
,
R.
,
Debenham
,
P. G.
,
Masson
,
W. K.
&
Webb
,
M. B. T.
(
1986
).
Ataxia-telangiectasia: A human mutation giving high frequency misrepair of DNA double strand scissions
.
Molec. Biol. Med.
3
,
229
244
.
Cox
,
R.
,
Hosking
,
G. P.
&
Wilson
,
J.
(
1978
).
Ataxia-telangiectasia: Evaluation of radiosensitivity in cultured skin fibroblasts as a diagnostic test
.
Archs Dis. Childh.
53
,
386
—390.
Cox
,
R.
,
Masson
,
W. K.
,
Debenham
,
P. G.
&
Webb
,
M. B. T.
(
1984
).
The use of recombinant DNA plasmids for the determination of DNA repair and recombination in cultured mammalian cells
.
Br.J. Cancer
49
, Suppl. VI, 67-72.
Debenham
,
P. G.
,
Webb
,
M. B. T.
,
Masson
,
W. K.
&
Cox
,
R.
(
1984
).
DNA-mediated gene transfer into human diploid fibroblasts derived from normal and ataxia-telangiectasia donors: parameters for DNA transfer and properties of DNA transformants
.
Int. J. Radiat. Biol.
45
,
525
536
.
Edwards
,
M. J.
&
Taylor
,
A. M. R.
(
1980
).
Unusual levels of (ADP-ribose)n and DNA synthesis in ataxia-telangiectasia cells following y-irradiation
.
Nature, Lond.
287
,
745
747
.
Gellert
,
M.
(
1981
).
DNA topoisomerases. A
.
Rev. Biochem.
50
,
879
910
.
Gellert
,
M.
,
Mizuuchi
,
K.
,
O’Dea
,
M. H.
,
Itoh
,
T.
&
Tomizawa
,
J.-I.
(
1977
).
Nalidixic acid resistance: A second genetic character involved in DNA gyrase activity
.
Proc. natn. Acad. Sci. U.SA.
74
,
4772
4776
.
Glickman
,
B. W.
,
Zwenk
,
H.
, Van
Sluis
,
C. A.
&
Rorsch
,
A.
(
1971
).
The isolation and characterisation of an X-ray sensitive ultraviolet-resistant mutant of Escherichia coli
.
Biochim. biophys. Acta
254
,
144
154
.
Goebel
,
W.
&
Schiess
,
W.
(
1975
).
The fate of bacterial plasmid DNA in mammalian cells
.
Molec. gen. Genet.
138
,
213
223
.
Gottesfield
,
J. M.
(
1986
).
Novobiocin inhibits RNA polymerase III transcription in vitro by a mechanism distinct from DNA topoisomerase IL Nucl. Acids Res.
14
,
2075
2087
.
Houldsworth
,
J.
&
Lavin
,
M. F.
(
1980
).
Effect of ionising radiation on DNA synthesis in ataxiatelangiectasia
.
Nucl. Acids Res.
8
,
3709
3720
.
Jeggo
,
P. A.
&
Kemp
,
L. M.
(
1983
).
X-ray sensitive mutants of Chinese hamster ovary cell line. Isolation and cross-sensitivity to other DNA-damaging agents
.
Mutat. Res.
112
,
313
—327.
Kemp
,
L. M.
,
Sedgwick
,
S. G.
&
Jeggo
,
P. A.
(
1984
).
X-ray sensitive mutants of Chinese hamster ovary cells defective in double-strand break rejoining
.
Mutat. Res.
132
,
189
196
.
Lehmann
,
A. R.
(
1982
). The cellular and molecular responses of ataxia-telangiectasia cells to DNA damage. In
Ataxia-telangiectasia
(ed. B. A. Bridges &
D. G.
Harnden
), pp.
83
101
.
New York
:
Wiley
.
Lehmann
,
A. R.
&
Stevens
,
S.
(
1980
).
A rapid procedure for measurement of DNA repair in human fibroblasts and for complementation analysis of xeroderma pigmentosum cells
.
Mutat. Res.
69
,
177
190
.
Morrison
,
A.
&
Cozzarelli
,
N. R.
(
1981
).
Contacts between DNA gyrase and its binding site on DNA: Features of symmetry and asymmetry revealed by protection from nucleases
.
Proc. natn. Acad. Sci. U.SA.
78
,
1416
1420
.
Mulligan
,
R. C.
&
Berg
,
P.
(
1981
).
Selection for animal cells that express the Escherichia coligene coding for xanthine-guanine phosphoribosyltransferase
.
Proc. natn. Acad. Sri. U.S A..
78
,
2072
2076
.
PiCKSLEY
,
S. M.
,
Attfield
,
P. V.
&
Lloyd
,
R. G.
(
1984
).
Repair of DNA double strand breaks in Escherichia coli K12 requires a functional recN product
.
Molec. gen. Genet.
195
,
267
274
.
Povirk
,
L. F.
(
1977
).
Localisation of inhibition of replicon initiation to damaged regions of DNA
.
J. molec. Biol.
114
,
141
151
.
Resnick
,
M. A.
(
1976
).
The repair of double strand breaks in DNA: A model involving recombination., theor. Biol.
59
,
97
106
.
Resnick
,
M. A.
&
Martin
,
P.
(
1976
).
The repair of double strand breaks in the nuclear DNA of Saccharomyces cerevisiae and its genetic control
.
Molec. gen. Genet.
143
,
119
129
.
Sedgewick
,
R. P.
(
1982
). Neurological abnormalities in ataxia-telangiectasia. In
Ataxiatelangiectasia
(ed. B. A. Bridges &
D. G.
Harnden
), pp.
23
35
.
New York
:
Wiley
.
Spector
,
B. D.
,
Filipovitch
,
A. H.
,
Perry
,
G. S.
&
Kerser
,
J. H.
(
1982
). Epidemiology of cancer in ataxia-telangiectasia. In
Ataxia-telangiectasia
(ed. B. A. Bridges &
D. G.
Harnden
), pp.
103
138
.
New York
:
Wiley
.
Sugino
,
A.
,
Higgins
,
N. P.
,
Brown
,
P. O.
,
Peebles
,
C. L.
&
Cozzarelli
,
N. R.
(
1978
).
Energy coupling in DNA gyrase and the mechanism of action of novobiocin
.
Proc. natn. Acad. Sci. U.S.A.
75
,
4838
4842
.
Taylor
,
A. M. R.
(
1982
). Cytogenetics of ataxia-telangiectasia. In
Ataxia-telangiectasia
(ed. B. A. Bridges &
D. G.
Harnden
), pp.
53
81
.
New York
:
Wiley
.
Taylor
,
A. M. R.
,
Harnden
,
D. G.
,
Arlett
,
C. F.
,
Harcourt
,
S. A.
,
Lehmann
,
A. R.
,
Stevens
,
S.
&
Bridges
,
B. A.
(
1975
).
Ataxia-telangiectasia: a human mutation with abnormal radiation sensitivity
.
Nature, Land.
258
,
427
429
.
Thacker
,
J.
,
Debenham
,
P. G.
,
Stretch
,
A.
&
Webb
,
M. B. T.
(
1983
).
The use of a cloned bacterial gene to study mutation in mammalian cells
.
Mutat. Res. Ill, 9-23
.
Thompson
,
L. H.
,
Rubin
,
J. S.
,
Cleaver
,
J. E.
,
Whitmore
,
G. F.
&
Brookman
,
K.
(
1980
).
A screening method for isolating DNA repair-deficient mutants of CHO cells
.
Sornat. Cell Genet.
6
,
391
405
.
Waldman
,
T. A.
(
1982
). Immunological abnormalities in ataxia-telangiectasia. In
Ataxiatelangiectasia
(ed. B. A. Bridges &
D. G.
Harnden
), pp.
37
51
.
New York
:
Wiley
.
Waldman
,
T. A.
&
McIntire
,
K. R.
(
1972
).
Serum alpha-fetaprotein levels in patients with ataxia-telangiectasia
.
Lancet ii, 1112-1115
.
Wigler
,
M.
,
Pellicer
,
A.
,
Silverstein
,
S.
&
Axel
,
R.
(
1978
).
Biochemical transfer of single-copy eucaryotic genes using total cellular DNA as donor
.
Cell
14
,
725
731
.