When the excision repair process of eukaryote cells is arrested by inhibitors of repair synthesis including hydroxyurea (HU), l-β-D-arabinofuranosylcytosine (araC) or aphidicolin, major cellular changes follow the accumulation of repair-associated DNA breaks. These changes, each of which reflects more or less severe cellular stress, include cycle delay, chromosome behaviour, fall in NAD level, the development of double-stranded DNA breaks, rapid chromosome fragmentation and cell killing. Disruption of the repair process by agents such as araC after therapeutic DNA damage may, therefore, have some potential value in cancer treatment. The extreme cellular problems associated with the artificial arrest of repair may have their subtler counterparts elsewhere, and we discuss several systems where delays in the completion of excision repair in the absence of repair synthesis inhibitors have marked repercussions on cell viability. We also show that the average completion time of an excision repair patch varies according to the state of cell culture, and that completion time is extended after treatment with insulin or following trypsin detachment. Under certain growth conditions ultraviolet irradiation followed by mitogenic stimulation results in double-stranded DNA breakage and additional cell killing, and we discuss these data in the light of protocols that have been used successfully to transform human or rodent cells in vitro. Finally, we consider whether the rejoining of DNA breaks accumulated by repair synthesis inhibitors is a valid model system for studying ligation, and show that this protocol provides an extremely sensitive assay for most incision events and, thereby, a means for discriminating between normal human cells on the one hand, and Cockayne’s Syndrome cells and their heterozygotes on the other.

‡Present address: Genetics Institute, Biological Research Center, Hungarian Academy of Sciences, P.O.B. 501, Szeged, Hungary.

The response of cells to DNA damaging agents is complex. In addition to long-term genetic changes (ranging from point mutations to chromosome aberrations) there are many profound shifts in metabolism that occur much earlier. These include dramatic changes in the size and composition of deoxyribonucleotide triphosphate pools (Newman & Miller, 1984), inhibition of transcription (Hackett et al. 1978), nucleolar disruption and inhibition of protein synthesis (Collins et al. 1981), induction of proteolytic activity (Miskin & Reich, 1980), and arrest of DNA replication (Painter, 1977). In addition, DNA repair pathways are activated. Clearly, these must function most actively in eukaryote cells under conditions of serious cellular stress. Stress is a rather general term, though fairly widely used in a repair context (e.g. see Friedberg, 1985). In prokaryotes stress associated with DNA damage invokes the coordinated SOS response (Radman, 1974), one consequence being error-prone repair (Witkin, 1976). Though there is some evidence for an SOS response in eukaryote cells (e.g. see Sarasin & Benoit, 1986), the biochemical mechanism has not been identified. In this chapter the term stress is used in two ways. We examine the various cellular consequences that follow arrest of the excision repair process by inhibitors of repair synthesis. In general such inhibitors exacerbate and compound the cellular stress due to DNA damage. We also examine operation of the excision repair process itself when stress is applied to it by way of limiting the DNA precursor supply. One of the aims of this chapter is to chronicle some consequences of repair inhibition after ultraviolet (u.v.) damage and to consider how the stress induced by artificially blocking repair with inhibitors is a model for situations occurring normally in certain cells and conditions.

Complex enzymic mechanisms involving several steps should in theory be susceptible to a number of specific inhibitors whose application results in the accumulation of intermediate products. This is, as yet, only partly true in the case of the pathway for u.v.-induced excision repair. Fig. 1 shows a cartoon from 1984 (Collins & Johnson, 1984) depicting the points in the excision repair process at which different inhibitors were thought to act. For the best-characterized repair synthesis inhibitors, agents such as hydroxyurea, which reduces the supply of DNA precursors, and more or less specific inhibitors of DNA polymerase α such as aphidicolin or araC, the main application described in this chapter is that, in concert, such inhibitors can be used over short periods after u.v. irradiation to accumulate large numbers of DNA breaks created by the incision step of the repair process. The inhibition of repair synthesis by HU and polymerase inhibitors is thoroughly substantiated and has been reviewed recently (Collins & Johnson, 1984; Collins et al. 1984); it need concern us no further here except to note that it provides the most effective way of setting up abundant unligated excision repair sites with which to study subsequently the process of break sealing. This topic will be discussed later in the chapter.

Fig. 1.

A simplified scheme illustrating repair of DNA lesions and showing postulated sites of action of typical inhibitory agents. (From Collins & Johnson, 1984, with permission from Academic Press.)

Fig. 1.

A simplified scheme illustrating repair of DNA lesions and showing postulated sites of action of typical inhibitory agents. (From Collins & Johnson, 1984, with permission from Academic Press.)

Turning to novobiocin, a less-traditional inhibitor of replicative and repair DNA synthesis, the evidence in favour of its supposed mode of action via the inhibition of a topoisomerse-II-mediated preincision step must now be regarded as questionable. Novobiocin blocks the accumulation of u.v.-induced incision events by araC plus HU, monitored either by alkaline unwinding or nucleoid sedimentation (Collins & Johnson, 1979a; Mattern et al. 1982; Downes et al. 1985), and it also inhibits unscheduled DNA synthesis (Mattern & Scudiero, 1981). It is less effective when aphidicolin is used (Clarkson & Mitchell, 1983; Downes et al. 1985), probably because, unlike araC, aphidicolin does not require phosphorylation before it can act as an inhibitor. Despite an old observation (Frei et al. 1958) that novobiocin inhibits oxidative phosphorylation, it was only recently that its effect on mitochondria was revealed, when Downes et al. (1985) reported the massive (yet reversible) and rapid change in mitochondrial structure induced by the drug (Fig. 2). The shift in structure was associated with a fivefold reduction over 30 min in the ratio of ATP to ADP, an effect that presumably helps to explain the reduction in the amount of phosphorylated araC in the soluble pool during this period. Given the considerable energy dependence of excision repair, the inhibitory effect of novobiocin can at present be as easily attributed to constraint of energy supply as to inhibition of a putative topoisomerase-mediated pre-incision step.

Fig. 2.

Micrographs showing the effects of novobioicin on rhodamine 123 fluorescence in normal human fibroblasts. Cells were incubated for 10 min in medium with rhodamine, washed in normal medium and then incubated for another 30 min in: A, normal medium; or B, medium with 1 mM-novobiocin. A, ×450; B, ×1100. (From Downes et al. 1985, with permission from IRL Press.)

Fig. 2.

Micrographs showing the effects of novobioicin on rhodamine 123 fluorescence in normal human fibroblasts. Cells were incubated for 10 min in medium with rhodamine, washed in normal medium and then incubated for another 30 min in: A, normal medium; or B, medium with 1 mM-novobiocin. A, ×450; B, ×1100. (From Downes et al. 1985, with permission from IRL Press.)

Inhibition of repair synthesis by agents such as HU, araC and aphidicolin, acting after the incision step, results in an exaggeration of many cellular responses to u.v. Some of these are listed in Table 1. Not included in this table is the feedback inhibition of the repair process itself, due to immobilization of the repair polymerase, and presumably the associated endonuclease, at the inhibited site. Studies of feedback inhibition may thus yield information about the assembly and stability of excision repair components (Snyder et al. 1981; Squires et al. 1982). A detailed analysis of the process reveals considerable complexity in these interactions: the efficiency of feedback inhibition increases with time after irradiation, and the process culminates in a paradoxical state in which no further excision occurs, though many lesions remain unrepaired and the capacity to perform further incision in response to fresh lesions is only partly diminished (Downes, 1984). But we will not deal with this topic further in this chapter. An important general point to stress is that a brief treatment with inhibitors is usually sufficient to elicit the effects listed in Table 1.

Table 1.

Consequences of repair inhibition

Consequences of repair inhibition
Consequences of repair inhibition

Cell killing and the importance of DNA precursor pools

The most consistent feature of repair inhibition in u.v.-irradiated cells is additional cell killing, the extent of which depends on the stage in the cell cycle when u.v. was given and the type of inhibitor used. For many years it has been recognized that inhibitors of DNA synthesis (in the absence of DNA damage) selectively kill cells in S’ phase (Chu & Fischer, 1962; Sinclair, 1965, 1967; Pfeiffer & Tolmach, 1967; Burg et al. 1977), a not unexpected finding, though exactly how killing is achieved remains speculative; one possibility is that disruption of DNA replication results in the production of lethal chromosome aberrations (Yu & Sinclair, 1968). The effect of inhibitors after u.v. is best studied, therefore, in other stages of the cycle. For example, a 3h incubation of HeLa cells in 10−2M-HU, after u.v. irradiation in mitosis or early G1( resulted in greater killing than u.v. alone (Schor et al. 1975). The cycle-related pattern of killing of synchronized Chinese hamster ovary cells, u.v.-irradiated and incubated for 90 min with HU, first suggested that the size of the DNA precursor pool is an important parameter in the overall DNA repair capability (Burg et al. 1977). Mitotic and G2 CHO cells have large pools (Skoog et al. 1973; Walters et al. 1973) and show no, or only slight, additional u.v.- associated killing when a 90 min pulse of HU is given. In mid G1, at a time when precursor pools are much smaller, HU promotes u.v. killing: an effect that is nullified by the simultaneous provision of DNA precursors. Burg et al. (1977) concluded from these results that, by inhibiting the synthesis of DNA precursors, HU can disrupt repair DNA synthesis to such an extent that cell death occurs. Since the main effect of HU is to inhibit ribonucleotide reductase, a key enzyme in DNA precursor provision (Thelander & Reichard, 1979) and since the lethal effect of HU after u.v. damage apparently varies in relation to precursor pool size, this suggests that variations in cell sensitivity to u.v. radiation are, in turn, reflections of the size or availability of the pool: a point to which we will return shortly.

araC on its own also causes additional killing of u.v.-irradiated mitotic and G’i CHO cells; a 30 min exposure is not enough but 2 h is (Collins et al. 1980). As would be expected, in combination araC or aphidicolin plus hydroxyurea potentiate u.v. killing. Fig. 3 shows u.v. survival curves of a log phase repair-competent human cell, HD 1, given 90 min exposures to HU plus aphidicolin at two different times after u.v. In this case, treatment immediately after u.v. markedly reduces survival whereas treatment at about 4h after u.v., when there is much less repair activity, does not have much effect on survival. Given that inhibitors such as those used in the studies mentioned above result in the arrest of incision soon after u.v. (Snyder et al. 1981; Squires et al. 1982; Downes, 1984), with the consequence that continued dimer removal is blocked (Snyder et al. 1981), the prolonged presence of dimers in the DNA might be considered likely to account for the enhanced u.v. killing in the presence of inhibitor. However, calculations of the sort described by Collins et al. (1980), which take into account dimers that are not removed when inhibitors are used, strongly suggest that this is unlikely to be the case. For example, after 10 J m−2 HD1 cells incubated over a 30 min period with HU plus araC (a suitable substitute for aphidicolin) accumulate nine breaks per 109 daltons (Johnson et al. 1985), and so a 90 min incubation would temporarily prevent the removal of, at most, 27, dimers per 109 daltons, corresponding to a u.v. dose of about 2J m−2. It is, however, clear from data in Fig. 3 that the effect on survival of 90 min post-u.v. exposure to aphidicolin/HU is greater than would be produced by delivering an extra 2 J m−2 of u.v. The ‘extra’ lethality associated with repair inhibition strongly suggests that breaks accumulated during this period are intrinsically more lethal then the dimers that would otherwise have been replaced.

Fig. 3.

Survival of human repair-competent cell, HD1 (Johnson et al. 1985), after u.v. irradiation, in the presence or absence of 1·5-h pulses of aphidicolin (l0 μg ml−1) plus hydroxyurea (10−2M). One population was u.v.-irradiated with a graded series of fluences and incubated for the next 1·5 h with inhibitors (▵); another was incubated with inhibitors between 3·5 and 5 h after u.v. irradiation (▪); and a final population served as the u.v. control (▫). All populations were washed three times after exposure to inhibitors and the dishes (in triplicate) were incubated for 8 days to allow colonies to grow. The curves were fitted by least squares analysis. Standard errors for the D0 and Dq values with their standard errors are as follows: control u.v., D0 2·81 ± 0·2, Dq 3·44 ± 0·41; first 1·5- h exposure, D0 1·8 ± 0·3, Dq 0·7 ± 0·15; −5 h exposure, D0 2·17 ±0±1, Dq 2·31 ± 0·34. (I. Rasko & R. T. Johnson, unpublished results.)

Fig. 3.

Survival of human repair-competent cell, HD1 (Johnson et al. 1985), after u.v. irradiation, in the presence or absence of 1·5-h pulses of aphidicolin (l0 μg ml−1) plus hydroxyurea (10−2M). One population was u.v.-irradiated with a graded series of fluences and incubated for the next 1·5 h with inhibitors (▵); another was incubated with inhibitors between 3·5 and 5 h after u.v. irradiation (▪); and a final population served as the u.v. control (▫). All populations were washed three times after exposure to inhibitors and the dishes (in triplicate) were incubated for 8 days to allow colonies to grow. The curves were fitted by least squares analysis. Standard errors for the D0 and Dq values with their standard errors are as follows: control u.v., D0 2·81 ± 0·2, Dq 3·44 ± 0·41; first 1·5- h exposure, D0 1·8 ± 0·3, Dq 0·7 ± 0·15; −5 h exposure, D0 2·17 ±0±1, Dq 2·31 ± 0·34. (I. Rasko & R. T. Johnson, unpublished results.)

Turning to the relationship between DNA precursor availability and survival after u.v., it is clear that supplementing medium with DNA precursors enhances survival after u.v., most spectacularly for unstimulated B and T lymphocytes (Yew & Johnson, 1979), but also very strongly for fibroblasts (Collins & Johnson, 1979b). The greatest improvements in survival promoted by precursor supplement are in those cells (such as quiescent lymphocytes) or cycle stages (Go, early Gi) where DNA precursor pools are known or expected to be very small (Munch-Petersen et al. 1973; Skooget al. 1973; Snyder, 1984). In these cases, therefore, we might conclude that precursor availability is likely to be a major rate-limiting factor in the repair of u.v. damage.

Cell killing and secondary consequences of inhibited repair

Survival curves of repair-inhibited cells, such as those in Fig. 3, should not be taken at face value. That is to say the cellular response to primary DNA damage may be modified by, among other things, changes in cycle progression. For example, as Mayne (1984) has shown, inhibitors greatly increase the u.v. arrest of transcription, a perturbation serious enough to result in cycle delay. Fig. 4 shows that u.v. irradiation of synchronized G1 cells, followed by a short pulse of aphidicolin/HU or especially of araC/HU results in considerable delay in the onset of DNA synthesis (I. Rasko & R. T. Johnson, unpublished data), providing, in effect, an extended pre-S phase period when potentially lethal lesions might be removed (Simons, 1979). This is one possibility, but inspection of the survival curves suggests that it must be outweighed by more damaging consequences of the inhibited repair process itself. As we discussed earlier, diminished rates of dimer removal in the presence of inhibitors cannot entirely account for the substantial potentiation of killing observed and one must, therefore, suspect the development of secondary and more hazardous consequences, presumably at the DNA level.

Fig. 4.

Arrest in cell cycle progression after u.v. irradiation associated with repair synthesis inhibitors. HD1A, an excision repair defective, XPD-like hybrid cell line (Johnson et al. 1986), was reversibly blocked tn mitosis by nitrous oxide arrest (Johnson et al. 1978), and released into G1 phase 2·5 h later; when more than 90% of cells had entered G1 medium was removed and the cells u.v. irradiated in situ with 2 J m−2 in warm phosphate-buffered saline. One dish was incubated with araC (10−4M) and HU (10−2M) for 1·5 h (○); one dish with aphidicolin (10 μgml−1) plus 10−2M-HU for 1·5 h (▴), and another dish served as the u.v. control (•). After washing each plate three times with warm medium the cultures were incubated with medium containing 0·5 μCi ml−1 [3H]thymidine (50 Ci mmol−1), and at the times indicated cells were removed and cytocentrifuge preparations made. These were processed for autoradiography and the frequency of labelled nuclei was scored in 100–200 cells. (I. Rasko & R. T. Johnson, unpublished results.)

Fig. 4.

Arrest in cell cycle progression after u.v. irradiation associated with repair synthesis inhibitors. HD1A, an excision repair defective, XPD-like hybrid cell line (Johnson et al. 1986), was reversibly blocked tn mitosis by nitrous oxide arrest (Johnson et al. 1978), and released into G1 phase 2·5 h later; when more than 90% of cells had entered G1 medium was removed and the cells u.v. irradiated in situ with 2 J m−2 in warm phosphate-buffered saline. One dish was incubated with araC (10−4M) and HU (10−2M) for 1·5 h (○); one dish with aphidicolin (10 μgml−1) plus 10−2M-HU for 1·5 h (▴), and another dish served as the u.v. control (•). After washing each plate three times with warm medium the cultures were incubated with medium containing 0·5 μCi ml−1 [3H]thymidine (50 Ci mmol−1), and at the times indicated cells were removed and cytocentrifuge preparations made. These were processed for autoradiography and the frequency of labelled nuclei was scored in 100–200 cells. (I. Rasko & R. T. Johnson, unpublished results.)

Cell killing, DNA and chromosome fragmentation

We can obtain more precise and primary information about the potentiation of cell killing by repair synthesis inhibitors from a study of chromosome aberrations. In general, incubation with inhibitors is associated with increased frequencies of chromosome gaps and breaks, a phenomenon that, as Taylor et al. (1962) correctly inferred, results from the deprivation of precursor supply necessary to sustain adequate DNA repair. Several recent reviews have dealt with this area (e.g. see Collins & Johnson, 1984; Kihlman & Natarajan, 1984; Bryant & Iliakis, 1984) and these should be consulted for further details. However, it is clear that many uncertainties remain in the aetiology of chromosome aberrations, mostly related to the long time scale over which it is necessary to allow them to develop before cells reach mitotis. We (and others) have therefore been interested in developing more immediate chromosome techniques for exploring the relationship between DNA damage, repair and its inhibition, and the production of chromosome aberrations. Two procedures have been used: one, the cell fusion technique of premature chromosome condensation, which allows one to look directly at interphase chromosomes without waiting for them to enter mitosis (Johnson et al. 1982; Hittelman, 1984); and the second, metaphase repair, makes use exclusively of the mitotic cell (with fully packed chromosomes) to generate aberrations during this phase of the cycle (Mullinger & Johnson, 1985).

u.v. irradiation of metaphase or interphase cells followed by repair inhibition results in massive chromosome decondensation dependent on both time and dose. Metaphase and prematurely condensed chromosomes (PCC) finally lose all apparent structure (Schor et al. 1975; Mullinger & Johnson, 1985) (Fig. 5). Decondensation continues long after DNA strand breaks, as measured by alkaline unwinding, have saturated (Fig. 6). This figure also shows that accumulated breaks mostly disappear when the inhibitors, in this case araC and HU, are removed. Inspection of chromosomes from such cells, after reversal of inhibition (Mullinger & Johnson, 1985) reveals that though the bulk of strand breaks are rejoined there is an abundance of chromosome aberrations, mostly in the form of chromatid breaks (Fig. 7). The degree of chromosome fragmentation depends on the duration of repair inhibition and the u.v. dose (Mullinger & Johnson, 1985). For G1 PCC the picture is similar (Hittelman, 1984; Hittelman & Pollard, 1984), with fragmentation revealed after a period of inhibition. Fig. 8 contrasts the appearance of normal G1 PCC from a human fibroblast with those from a G1 cell irradiated with 10 J m−2 and incubated for 90 min with aphidicolin and HU before inhibitors were removed and cells fused to produce PCC. Fragmentation is evident. Table 2 supplies quantitative information and a comparison of the effects of araC/HU and aphidicolin/HU inhibition on PCC fragmentation in normal and Cockayne’s Syndrome (CS) fibroblasts. Both cell types respond similarly at this u.v. dose, and it appears that araC has a stronger effect on fragmentation than aphidicolin, a point we will return to later.

Table 2.

Inhibition of excision repair and the production of chromosome fragments in G1PCC from u.v.-irradiated human fibroblasts

Inhibition of excision repair and the production of chromosome fragments in G1PCC from u.v.-irradiated human fibroblasts
Inhibition of excision repair and the production of chromosome fragments in G1PCC from u.v.-irradiated human fibroblasts
Fig. 5.

Light micrographs of Indian muntjac chromosome spreads illustrating progressive stages of condensation after u.v. irradiation of metaphase cells pre-incubated for 30 min and post-incubated for 90 min with araC (10−4M) and HU (10−2M). A, 0·4 J m−2; B, 2 J m−2; C, 2 J m−2; D, 5Jm−2. ×900. (From Mullinger & Johnson, 1985.)

Fig. 5.

Light micrographs of Indian muntjac chromosome spreads illustrating progressive stages of condensation after u.v. irradiation of metaphase cells pre-incubated for 30 min and post-incubated for 90 min with araC (10−4M) and HU (10−2M). A, 0·4 J m−2; B, 2 J m−2; C, 2 J m−2; D, 5Jm−2. ×900. (From Mullinger & Johnson, 1985.)

Fig. 6.

DNA break accumulation in u.v.-irradiated, metaphase-arrested HeLa cells during incubation with araC (10−4M) plus HU (10−2M). Data are shown for 0·5 (▴), 1 (▿) and 5 (▪) J m−2 u.v. At 90 min after u.v. inhibitors were removed from a parallel 5 J m−2 sample, which was incubated for a further 60 min in the presence of the four deoxyribonucleosides, each at 10−4 M. DNA breaks were assayed after alkaline unwinding and hydroxyapatite chromatography. (From Mullinger & Johnson, 1985.)

Fig. 6.

DNA break accumulation in u.v.-irradiated, metaphase-arrested HeLa cells during incubation with araC (10−4M) plus HU (10−2M). Data are shown for 0·5 (▴), 1 (▿) and 5 (▪) J m−2 u.v. At 90 min after u.v. inhibitors were removed from a parallel 5 J m−2 sample, which was incubated for a further 60 min in the presence of the four deoxyribonucleosides, each at 10−4 M. DNA breaks were assayed after alkaline unwinding and hydroxyapatite chromatography. (From Mullinger & Johnson, 1985.)

Fig. 7.

Light micrographs of HeLa chromosomes showing reversal of decondensation leading to fragmented chromosomes. Metaphase cells were u.v.-irradiated (20 J m−2), incubated with araC (10−4M) plus HU (10−2M) for 90min, followed by a further 60min either in the continued presence of inhibitors (A), or in their absence but presence of 10−4M-deoxyribonucleosides (B). ×750. (From Mullinger & Johnson, 1985.)

Fig. 7.

Light micrographs of HeLa chromosomes showing reversal of decondensation leading to fragmented chromosomes. Metaphase cells were u.v.-irradiated (20 J m−2), incubated with araC (10−4M) plus HU (10−2M) for 90min, followed by a further 60min either in the continued presence of inhibitors (A), or in their absence but presence of 10−4M-deoxyribonucleosides (B). ×750. (From Mullinger & Johnson, 1985.)

Fig. 8.

u.v. irradiation and repair synthesis inhibition results in fragmentation of prematurely condensed chromosomes. Normal human fibroblasts in quiescent culture were either: A, fused with mitotic HeLa cells to induce premature chromosome condensation; or B, u.v. irradiated (10 Jm−2) and incubated with araC (10−4M) plus HU (10−2M) for 60min, at which point inhibitors were removed and the cells incubated for a further 30 min with 10−4M-deoxyribonucleosides; this population was fused with mitotic HeLa cells to induce PCC. (A. M. Mullinger, S. Squires & R. T. Johnson, unpublished results.)

Fig. 8.

u.v. irradiation and repair synthesis inhibition results in fragmentation of prematurely condensed chromosomes. Normal human fibroblasts in quiescent culture were either: A, fused with mitotic HeLa cells to induce premature chromosome condensation; or B, u.v. irradiated (10 Jm−2) and incubated with araC (10−4M) plus HU (10−2M) for 60min, at which point inhibitors were removed and the cells incubated for a further 30 min with 10−4M-deoxyribonucleosides; this population was fused with mitotic HeLa cells to induce PCC. (A. M. Mullinger, S. Squires & R. T. Johnson, unpublished results.)

Since virtually all single strand DNA breaks are removed after inhibitor reversal, this being particularly true for aphidicolin (see below), chromosome breakage probably reflects the small proportion of residual, barely detectable double strand (DS) DNA breaks. Using neutral elution or viscoelastometry, several groups have shown that DS DNA breaks develop in a time- and dose-related manner as a consequence of repair inhibition (Bradley & Taylor, 1983; Filatov & Noskin, 1983), while Hittelman (1984) and Mullinger & Johnson (1985) have related the appearance of DS breaks to the development of chromosome fragmentation. DS breaks can develop rapidly in these cells as the data in Fig. 9 show. Here a normal human fibroblast and a permanent human cell line, HD1, irradiated with 10 J m−-2, were incubated with araC/HU for up to 90 min and at various points the presence of DS breaks was assessed. Clearly, after 20 min of inhibition, elution behaviour has changed, implying the early development of DS lesions. The figure also shows that aphidicolin is as effective as araC in promoting DS breaks. Removal of the inhibitors (Mullinger & Johnson, 1985, and Fig. 9), or removal plus addition of DNA precursors, results in some loss of DS breaks, a finding in agreement with the results of Bryant & Iliakis (1984) using araA in combination with X-irradiation. We assume that in our system remaining DS breaks are reflected in chromosome fragmentation, though we do not have a quantitative correlation between the two. A small proportion of the DS breaks may be accounted for by overlap of inhibited repair sites in opposite strands, an idea proposed by Filatov & Noskin (1983), and one that is marginally strengthened by data suggesting that the size of repair patch in the presence of inhibitors is greater than usual (Clarkson, 1978; Francis et al. 1979; Th’ng & Walker, 1986). It is worth pointing out, however, that the increased size of repair patch in the presence of inhibitors is likely to be trivial (from 20 to 50 bases) compared to the distance between patches (about 100 000 base-pairs after 10 J m−2). Moreover, DS breaks and chromosome fragmentation develop even after rather low levels of u.v. (e.g. 5Jm−2), when rates of dimer coincidence in opposite strands would be extremely low. It seems likely, therefore, that a proportion of DS breaks arise from (secondary) nuclease action opposite the gaps at inhibited repair sites: though, as Hittelman (1984) has shown by means of S1 nuclease modification of neutral elution behaviour of DNA from repair arrested cells, not all single strand gaps are converted to DS breaks in the cells.

Fig. 9.

Effect of repair inhibitors on the elution behaviour of u.v.-irradiated metaphase DNA under neutral (pH 9·6) conditions. Normal human fibroblasts (▪), or human cell line HD1 (▴), uniformly prelabelled with [3H]thymidine, were u.v. irradiated (10 J m−2) and incubated with araC (10−4M) plus HU (10−2M) (filled symbols) or aphidicolin (10μg ml−1) plus HU (10−2M) (fibroblasts, (○)), for up to 90min. At the indicated times samples were prepared for neutral elution as described by Mullinger & Johnson (1985). In this figure the resulting elution profiles are expressed in terms of the proportion of DNA radioactivity remaining on the filter after a 15 h elution period. Without inhibitors elution of the DNA from u.v.-irradiated HD1 cells is negligible (▫). The arrow indicates the time of inhibitor removal from a fibroblast sample, which was then incubated for a further 30 min before cell lysis. Where bars are present they represent the range in duplicate samples. (R. T. Johnson, unpublished results.)

Fig. 9.

Effect of repair inhibitors on the elution behaviour of u.v.-irradiated metaphase DNA under neutral (pH 9·6) conditions. Normal human fibroblasts (▪), or human cell line HD1 (▴), uniformly prelabelled with [3H]thymidine, were u.v. irradiated (10 J m−2) and incubated with araC (10−4M) plus HU (10−2M) (filled symbols) or aphidicolin (10μg ml−1) plus HU (10−2M) (fibroblasts, (○)), for up to 90min. At the indicated times samples were prepared for neutral elution as described by Mullinger & Johnson (1985). In this figure the resulting elution profiles are expressed in terms of the proportion of DNA radioactivity remaining on the filter after a 15 h elution period. Without inhibitors elution of the DNA from u.v.-irradiated HD1 cells is negligible (▫). The arrow indicates the time of inhibitor removal from a fibroblast sample, which was then incubated for a further 30 min before cell lysis. Where bars are present they represent the range in duplicate samples. (R. T. Johnson, unpublished results.)

Despite our rather general ignorance of detailed metabolic changes that follow u.v. irradiation and subsequent incubation with inhibitors of DNA repair synthesis, we are now in a stronger position to conclude that the substantial extra killing that is commonly found with HU, araC and aphidicolin is predominantly due to degradative changes occurring at many of the inhibited sites of repair. The most important change in terms of potential lethality is the DS DNA break. Some of these can be repaired; others will form the basis of chromosome fragmentation, probably the immediate cause of lethality. Metaphase repair and PCC techniques allow us to narrow the time-scale between DNA damage, repair events and their perturbation, and the development of chromosome aberrations. It remains possible, however, that brief distortions of the repair process by inhibitors have additional deleterious and, as yet, unknown consequences for excision repair and such disturbance may contribute to cell killing. The most likely interpretation of the effect of novobiocin on repair probably arises from the cellular stress it induces via its disturbance of ATP metabolism. This mechanism can account for its potency until other data are produced.

The conversion of repair sites to lethal strand-breaks by DNA repair inhibitors offers possibilities in the chemotherapy of cancer. Potentially the most useful agent in this respect is araC, since its effect is not easily reversible. araC, acting as an inhibitor of replicative synthesis, is already widely used in chemotherapy, being the most effective agent for the treatment of myeloid leukaemia (Frei et al. 1969). Other tumours are more refractory to treatment with araC, in some cases because their replicative polymerase is less sensitive to inhibition (Tanaka & Yoshida, 1982). But even without replicative incorporation, the combination of araC and a DNA damaging agent is, as we have seen, more lethal than the damage alone. Kufe et al. (1984) have shown that the increased lethality correlates well with the incorporation of araCTP into repair patches. araC, or possibly other repair inhibitors, might therefore effectively increase the effectiveness of DNA damaging agents used therapeutically. In this context it is worth noting that the efficiency of discrimination against araCTP, as a substrate for incorporation, may be less with repair polymerases than with replicative polymerases, and be less in transformed than in diploid cells (Elliott & Downes, 1986).

The gross effects associated with inhibited repair that have been chronicled may have their subtler counterparts elsewhere. For example, there are several cases where, without our interference, cells behave very much as though their postincision repair processes were inhibited or delayed. Here, though the effects are less extreme, they may be far more biologicaly significant. The best examples of delayed repair are seen in the Chinese hamster mutant EM9 (see Thompson et al. 1987, this volume) and in the human fibroblast designated 46 BR (Webster et al. 1982). In both these cases, in the absence of inhibitors, the disappearance of repair-induced breaks is poor compared with wild-type counterparts (Thompson et al. 1982; Squires & Johnson, 1983; Teo et al. 1983). A survey of human cells has revealed that CS fibroblasts also accumulate breaks in the absence of inhibitors, but only if they have recently undergone trypsinization (Squires & Johnson, 1983). CS cells irradiated up to 6h after trypsin treatment accumulated significant numbers of breaks by comparison with normal cells, though always to a lesser degree than 46 BR. Unlike 46 BR this disturbed CS state can be alleviated by providing DNA precursors, implicating a possible abnormality in CS cells of precursor production or of the polymerization process itself in CS cells.

Normally, human fibroblasts do not generate long-lived breaks (unless treated with repair inhibitors) but in special circumstances they can. Squires et al. (unpublished) have investigated conditions under which delayed repair occurs. In particular, we have examined the dependence of break accumulation on trypsin detachment of cells in different growth states, seeking to establish whether the protocols used to transform human or rodent cells also impede the repair process. One mitogen, insulin, has been employed with success in.transformation studies (Milo et al. 1981; Umeda et al. 1963) and its effect on excision repair has been examined. Fig. 10 illustrates the excision repair behaviour of cells in three growth states given 4 J m−2 of u.v. over a 60 min period. The three cell states are log cultures reseeded 3 h earlier (A), quiescent cultures (B); and cultures seeded 3 h earlier from quiescence (C). Each data set includes time courses of DNA break frequencies in the presence and absence of repair synthesis inhibitors, and both of these conditions have an additional variable, the presence or absence of insulin. Several points are clear. First, in the presence of inhibitors each culture shows a similar profile of break accumulation, and insulin has little or no effect. Second, in the absence of inhibitors there are few long-lived breaks in log reseeded or quiescent cultures; but there are many more in cultures recently reseeded from quiescence, amounting to about 103 per cell. Third, insulin increases the frequency of long-lived breaks in each culture, with the greatest net value in quiescent reseeded cells of about 5 ×103/cell.

Fig. 10.

The accumulation of u.v.-induced incomplete repair sites in human fibroblasts as a function of growth state and insulin treatment.

A. Logarithmic cultures of normal human fibroblasts, labelled with [3]thymidine were irradiated with 4 J m−-2 of u.v. light 3 h after seeding. The cells were incubated in growth medium either with insulin (1 unit ml−1) (▴, ▵), or without (•, ○) for various periods in the presence of 10−2M-HU and 5×10−5M-araC (▴, •), or without DNA synthesis inhibitors (▵, ○). When inhibitors were used the cells were preincubated for 30min before irradiation. The frequency of breaks in DNA was determined by alkaline lysis and hydroxyapatite chromatography.

B. Quiescent cell cultures were obtained 2–3 weeks after seeding, with at least two medium changes during the interim period. The contact-inhibited cells were irradiated with 4 J m-, 3 h after serum stimulation.

C. Cultures obtained by reseeding, in fresh medium, trypsinized quiescent cells. These were irradiated 3 h later. The results shown represent the mean of several experiments (A, at least 3; B, 6; C, over 10). At 15 min incubation standard errors of the mean break frequency per 109 daltons were: A. (•), ±1·4; (), ±2; (▵), ±0·9. B. (•), ±3·4; (), ±3; (○), ±0·6; (▵), ±2·8. C. (•), ±2; (), ±24; (○), ±1·8; (▵), ±2·4. (From Squires, Elliott & Johnson, unpublished results.)

Fig. 10.

The accumulation of u.v.-induced incomplete repair sites in human fibroblasts as a function of growth state and insulin treatment.

A. Logarithmic cultures of normal human fibroblasts, labelled with [3]thymidine were irradiated with 4 J m−-2 of u.v. light 3 h after seeding. The cells were incubated in growth medium either with insulin (1 unit ml−1) (▴, ▵), or without (•, ○) for various periods in the presence of 10−2M-HU and 5×10−5M-araC (▴, •), or without DNA synthesis inhibitors (▵, ○). When inhibitors were used the cells were preincubated for 30min before irradiation. The frequency of breaks in DNA was determined by alkaline lysis and hydroxyapatite chromatography.

B. Quiescent cell cultures were obtained 2–3 weeks after seeding, with at least two medium changes during the interim period. The contact-inhibited cells were irradiated with 4 J m-, 3 h after serum stimulation.

C. Cultures obtained by reseeding, in fresh medium, trypsinized quiescent cells. These were irradiated 3 h later. The results shown represent the mean of several experiments (A, at least 3; B, 6; C, over 10). At 15 min incubation standard errors of the mean break frequency per 109 daltons were: A. (•), ±1·4; (), ±2; (▵), ±0·9. B. (•), ±3·4; (), ±3; (○), ±0·6; (▵), ±2·8. C. (•), ±2; (), ±24; (○), ±1·8; (▵), ±2·4. (From Squires, Elliott & Johnson, unpublished results.)

Data such as these in Fig. 10 allow us to estimate the average time taken to complete a repair site in human fibroblasts (Squires et al. unpublished). We assume that for a given culture state all repair sites give rise to breaks in the presence of inhibitors, and that all sites take the same time to be completed in the absence of inhibitors. Clearly, not all repair sites are converted into breaks in the presence of inhibitors (Downes, 1984; Smith, 1984), but we note from the data in Fig. 10 that the frequency of inhibited repair sites in the three cultures is similar, and we assume for this argument that the required correction factor would not reduce completion time by more than 30 %. Finally, we assume that the rate of incision is constant over a 15 min period of break accumulation in the presence of inhibitors. Breaks detected in the absence of inhibitors must therefore arise during a period after u.v. when there was too little incubation time left for completion at the site. Thus the fraction of repair sites that are incomplete at the end of 15 min equals the fraction of the 15 min needed to complete an incised site. These maximum estimates, shown in Table 3, indicate that the time required to complete a site varies with the growth state and also that insulin delays completion in all. For example, quiescent cells require four times longer to complete a repair site than proliferating cells and quiescent cells stimulated to cycle by trypsin treatment, and replating require about 18 times longer than their proliferating counterparts. The most extreme delays do not, however, exceed the 12 min duration of patch completion in 46BR fibroblasts, 40 times longer than a normal fibroblast (data taken from Squires & Johnson, 1983).

Table 3.

Estimated time for repair at a single site

Estimated time for repair at a single site
Estimated time for repair at a single site

The intriguing question of how the mitogens insulin and trypsin act to exacerbate delays in rejoining of repair patches can be partly answered. We know, from the fact that added DNA precursors can alleviate their effect, that DNA precursor pools are likely to be a target (Squires et al. unpublished). Now we have preliminary, direct evidence that deoxyribonucleotide pools are reduced by both trypsin and insulin treatment (Collins, Squires & Johnson, unpublished data). While insulin specifically reduces purine pools by up to 50 %, trypsin reduces purine and pyrimidine pools by 30–50%. How commonly mitogens act in this way remains to be seen; and how reduction in DNA precursor pools fits into the overall pattern of growth stimulation is, as yet, obscure. Trypsin effects are likely to be complex, especially since the enzyme can readily gain access into cells (Hodges et al. 1973); and also, by cleaving surface receptors such as those of insulin (Czech, 1985), it may radically change the cellular response to other mitogens.

Pre-S-phase fibroblasts irradiated with u.v. after detachment by trypsin rapidly develop double strand DNA damage, especially when insulin is present, though the effect is less marked than with repair inhibitors. Cells reseeded from quiescence show the greatest DS breakage, and with these cultures, also, insulin increases u.v. toxicity, reducing the D0 by a factor of almost 3 (Squires et al. unpublished).

The stressed fibroblasts considered above were not the first example to be discovered, of cell hypersensitivity to u.v. related to limitations of DNA precursor supply. The extreme u.v. sensitivity of unstimulated lymphocytes can be attributed in part at least to the limited size of the DNA precursor pool (Munch-Petersen et al. 1973), since supplies of deoxyribonucleosides greatly improve both B and T cell survival, and also speed up the completion of the excision repair process (Yew & Johnson, 1979). We may now be in a position to resolve the paradox of why noncycling lymphocytes should be so u.v.-sensitive even though they are in an extended holding state, which should allow repair to occur, however slowly, before the cells have to replicate their DNA. They may represent a repair-stressed situation where small pools are associated with delayed completion of the repair sites resulting in secondary DS DNA breaks and subsequent cell death. Additionally, however, a small proportion of lymphocytes may have advanced towards transformation because of this damage. This is an intriguing possibility. DS DNA breaks or gaps undergoing repair in mammalian cells may, as in yeast (Resnick, 1976; Orr-Weaver & Szostak, 1983), represent an excellent substrate for mitotic recombination. Resulting rearrangements, perhaps in the form of translocations and deletions, could form the basis of subsequent transformation. The work of Zajac-Kaye & Ts’o (1984) has shown a clear relationship between the production of DNA breaks with subsequent chromosome rearrangements and increased rates of transformation. In human fibroblasts the relationship between the induction after u.v. irradiation of long-lived repair sites by trypsin and insulin, and the promotion of genetic instability is speculative. However, several studies have demonstrated the usefulness of trypsin detachment and insulin treatment in promoting human and rodent cell mutagenesis and transformation from quiescent cultures (Milo et al. 1981; Zimmerman & Little, 1983; Grosovsky & Little, 1984; Terasima et al. 1985). Stressing the excision repair capacity after damage by reducing precursor availability and thereby delaying the completion of repair may be important general consequences of cell exposure to trypsin or other proteolytic enzymes, and to mitogens such as insulin. After tissues have experienced DNA damage they are subject to significant protease activity (Miskin & Reich, 1980, 1981), and also undertake rapid cell renewal presumably under control of growth factors.

Conclusions drawn about delays in the excision repair process not associated with repair synthesis inhibitors

There is now evidence that the completion time of the excision repair process is (1) related to the state of cell growth, and (2) greatly delayed by certain mitogenic stimuli. These statements relate to cells in vitro and are not necessarily associated with any mutant phenotype. Underlying each delay in completion is evidence of immediate DNA precursor limitation, either related to natural oscillations of pool ! size associated with growth state, or to induced pool depletion by agents other than those recognized as inhibitors of precursor production or DNA polymerization (Nicander & Reichard, 1985). In cell cultures the consequences of such delayed repair are clearly seen, though they are usually less extreme than when repair synthesis inhibitors have been used. In the organism there is no direct evidence for delayed repair associated with pool constraint and we must, therefore, be careful in making any extrapolation. However, given the evidence we have from cell cultures about the production of DS DNA breaks, chromosome aberrations, and transformation and mutagenesis associated with mitogenic stimulation of specific growth stages following DNA damage, we feel that some speculation is reasonable. We conclude that delay in the completion of repair in vivo may be an important consequence of the cellular response to DNA damage, especially since such repair behaviour is likely to occur in tissues where cell removal and renewal are commonplace. DNA templates suffering from such delay could act as substrates for recombination or deletion and both of these long-term genetic consequences could play a prominent role in cell transformation.

Earlier in this chapter, we noted that many of the DNA breaks accumulated with inhibitors after u.v. can be chased away when inhibitors are removed. This allows us to examine the kinetics of the ligation step in excision repair, and in particular its sensitivity to the ADP-ribosyl transferase inhibitor, 3-aminobenzamide (3AB); we will deal with this latter aspect first. Many laboratories have shown that DNA breakage, such as is caused by alkylation damage, results in the activation of ADP- ribosyl transferase, which utilizes NAD to produce polymers of ADP-ribose associated with nuclear proteins (Shall, 1984). Arrest of its activity by 3AB results in the retention of the NAD pool and the lack of sealing of alkylation-related breaks. One possible model for the action of 3AB is the dependence of DNA ligase II on ADP-ribosylation for its activation (Creissen & Shall, 1982).

Unlike alkylation damage, u.v. irradiation does not usually lead to the accumulation of many DNA strand breaks, and it produces only a slow, moderate fall in NAD levels. However, if repair synthesis inhibitors are used to accumulate DNA breaks in u.v.-irradiated cells, NAD levels fall further and faster as much additional poly(ADP-ribose) synthesis is induced (Jacobson et al. 1983; Collins, 1985; Squires et al. unpublished). NAD levels also fall by 20% in cultures of human fibroblasts reseeded from quiescence, which accumulate incomplete repair sites after u.v., especially in the presence of insulin (Fig. 10C) (Squires et al. 1986). Though the fall in NAD can be prevented entirely by 3AB, break sealing after removal of inhibitors (though still in the presence of 3AB) is normal (Collins, 1985; Squires et al. unpublished). A typical experiment is shown in Fig. 11. Panel A follows the NAD pool in u.v.-irradiated cells incubated with aphidicolin plus HU, with and without 3 AB. In the absence of 3 AB the pool falls to 40 % of its starting level by 60 min; with 3AB, NAD levels remain constant. Panel B registers the accumulated DNA breaks in this material. With aphidicolin/HU about 10 breaks accumulate in 60 min whether or not 3AB is present. Restoring repair synthesis by removing inhibitors and providing DNA precursors at the time indicated results in rapid break sealing (and no further loss of NAD). Under these circumstances we conclude that whichever DNA ligase is involved in repair of u.v. damage in fibroblasts, ADP ribosylation is not crucial for this activity.

Fig. 11.

Effect of 3AB on NAD levels and on the rate of gap sealing of DNA breaks accumulated after u.v. in the presence of repair synthesis inhibitors. Proliferating cultures were irradiated with 4 J m−2 of u.v. and incubated at 37·C, with 10−2M-HU and 10μgml−1 aphidicolin in the presence (•) or absence (○) of 5mM-3AB. An hour after irradiation the medium containing inhibitors was replaced with medium plus 10−4M- dX ± 5 mM-3AB (A) or not replaced (B) and the cells were incubated further. At various times samples were removed and NAD (A) or DNA breaks (B) were assayed as described (Squireset al. 1986). (From Squireset al. unpublished results.)

Fig. 11.

Effect of 3AB on NAD levels and on the rate of gap sealing of DNA breaks accumulated after u.v. in the presence of repair synthesis inhibitors. Proliferating cultures were irradiated with 4 J m−2 of u.v. and incubated at 37·C, with 10−2M-HU and 10μgml−1 aphidicolin in the presence (•) or absence (○) of 5mM-3AB. An hour after irradiation the medium containing inhibitors was replaced with medium plus 10−4M- dX ± 5 mM-3AB (A) or not replaced (B) and the cells were incubated further. At various times samples were removed and NAD (A) or DNA breaks (B) were assayed as described (Squireset al. 1986). (From Squireset al. unpublished results.)

The reversal of DNA breaks accumulated after u.v. in the presence of inhibitors such as araC or aphidicolin provides a simple and quantitative measure of postincision events including ligation, an activity that is othewise difficult to study. Breaks accumulated by aphidicolin are rejoined much more rapidly than those associated with araC-arrested repair, a phenomenon attributed to the different modes of action of these agents (Cleaver, 1983; Collins, 1987). For aphidicolin reversal (Fig. 11) the rate of break rejoining can be extremely rapid (Squires et al. unpublished) and for human and Chinese hamster cells, Collins (1987) has computed rates of rejoining in excess of one or two breaks per min per 109 dalton over the first 10 min, surpassing by 15-fold the maximum rates of incision in these cells. Such data indicate that ligation can validly be studied in this artificial situation, especially when the period of break accumulation is short. Extending the length of inhibition results in slower rates of rejoining and less-complete reversal (Collins, 1987), symptoms typical of increasing cellular stress associated with inhibited repair; and it seems likely that the development of DS DNA breaks and chromosome fragmentation during more prolonged inhibition could slow down or disrupt repair. Other explanations for less-efficient rejoining, involving DNA precursor availability, are also possible given the complex disturbance to the deoxyribonucleotide pools caused by u.v. light on the one hand (Das et al. 1983; Newman & Miller, 1984) and by inhibitors such as HU and aphidicolin on the other (Nicander & Reichard, 1985).

We have already seen that break rejoining is retarded in u.v.-irradiated normal fibroblasts by trypsin and insulin treatment (Fig. 10) (Squires et al. unpublished), and in CS fibroblasts by trypsin (Squires & Johnson, 1983). In both situations addition of DNA precursors restores normal break-rejoining behaviour, implying that the effect of trypsin/insulin was to disturb either precursor supply or precursor presentation to the repair polymerase. Given the exaggerated difficulties experienced by CS cells, we examined their capacity to rejoin accumulated breaks in a standard ‘recoil’ situation (Squires & Johnson, unpublished). Fig. 12 shows the time courses of break sealing for normal and CS fibroblasts and also for a CS heterozygous strain. Each culture had been reseeded recently from quiescence, before u.v. irradiation and incubation for 30min with aphidicolin plus HU. One striking feature is the very marked difficulty experienced by CS cells in rejoining breaks compared with a normal cell strain. But the figure also reveals that the CS heterozygous strain is delayed, to an intermediate degree. Even when the reversal is carried out with DNA precursors present to speed up rejoining the behaviour of both CS and CS heterozygous cells still differs from wild type, suggesting that disruption of the repair mechanism in CS cells may not simply be explained in terms of precursor supply. Studies are in progress to establish the generality of this finding with several CS strains and their heterozygotes.

Fig. 12.

Delays in break rejoining in fibroblasts from an individual with Cockayne’s Syndrome (CS) and a Cockayne’s Syndrome heterozygote (CSH). Confluent cultures of: (○ or ◑) normal human fibroblasts (embryonic lung); (▵ or ◮) CS3BE (CS); and (▫ or ◨) CSHIBI (CSH), were u.v.-irradiated (4 J m−2) and incubated with aphidicolin (10 μg ml−1) plus HU (10−4M) for 30min to accumulate repair-related DNA breaks. At this point inhibitors were removed and incubation continued for about 2h either in the presence of 10−4M-deoxyribonucleosides (broken lines, open symbols) or in standard growth medium (continuous lines, half-filled symbols). At different times samples were lysed in alkali, and the frequency of DNA breaks assessed by means of hydroxyapatite chromatography. (S. Squires & R. T. Johnson, unpublished results.)

Fig. 12.

Delays in break rejoining in fibroblasts from an individual with Cockayne’s Syndrome (CS) and a Cockayne’s Syndrome heterozygote (CSH). Confluent cultures of: (○ or ◑) normal human fibroblasts (embryonic lung); (▵ or ◮) CS3BE (CS); and (▫ or ◨) CSHIBI (CSH), were u.v.-irradiated (4 J m−2) and incubated with aphidicolin (10 μg ml−1) plus HU (10−4M) for 30min to accumulate repair-related DNA breaks. At this point inhibitors were removed and incubation continued for about 2h either in the presence of 10−4M-deoxyribonucleosides (broken lines, open symbols) or in standard growth medium (continuous lines, half-filled symbols). At different times samples were lysed in alkali, and the frequency of DNA breaks assessed by means of hydroxyapatite chromatography. (S. Squires & R. T. Johnson, unpublished results.)

Finally, it is worth pointing out that the susceptibility of post-incision steps of DNA repair processes to disruption by agents such as trypsin may be rather widespread among heritable human diseases. For example, in recently detached progeria cells, X-irradiated in suspension, there is clear evidence of slow breakrejoining compared with a normal control (Epstein et al. 1973). But when attached monolayers are irradiated, progeria cells reseal DNA breaks as rapidly as their normal counterparts (Setlow & Regan, 1974; Bradley et al. 1976). XP variant cells, without apparent defects in excision repair, have also been reported to accumulate long-lived breaks in parental DNA in the absence of inhibitors (Fornace et al. 1976) and, using the accumulated break-rejoining assay, Squires & Johnson (unpublished) have now confirmed that the only XP variant strain examined to date (XP30R0) is slow to remove breaks. Whether the ease with which excision repair steps can be dissociated in CS, progeria and XP variant cells is attributable in each case to abnormal DNA precursor provision or to defects in DNA polymerization/ligation steps remains uncertain. What does seem clear is that under most circumstances demands made on the precursor supply by repair synthesis will be met by the cell. Certainly this is the case for CS and XP variant in log growth, and in these conditions excision repair is apparently normal. For these cells the much more serious possibility lies in abnormal DNA replication behaviour when they are stressed either by damaging agents or by exposure to mitogenic signals. DNA damage in CS cells caused by u.v. is associated with prolonged inhibition of both DNA and RNA synthesis, and with failure of these cells to recover normally from potentially lethal u.v. damage when held in a non-dividing state (Lehmann et al. 1979; Mayne & Lehmann, 1982). We should like to know, by direct measurement, whether u.v. causes hyperinstability of nucleotide metabolism in CS. The stress induced in nucleotide metabolism by mitogenic stimulation is poorly understood, though the direct measurement of deoxyribonucleotide pool reduction associated with insulin or trypsin treatment mentioned above is perhaps the first evidence (Collins et al. unpublished data). But trypsin is also known to cause severe arrest of entry into S phase in normal cells (Campisi & Medrano, 1983), and can disturb precursor supply during.S’ phase (S. Squires, unpublished). Extrapolating from the sensitivity of excision-related DNA synthesis in CS, these cells may also display an exaggerated sensitivity of replication to disturbance. This is currently being investigated. If, as seems likely, replication in CS is hypersensitive to the stresses involved in handling cells in culture, this would not be the first hereditary defect to display such behaviour. Bloom’s Syndrome fibroblasts, for example, show exactly this pattern, with extremely slow rates of DNA chain growth by comparison with normal fibroblasts in sparse or recently reseeded cultures (Ockey, 1979). Each of these hereditary conditions might be regarded perhaps as a family of conditional replication mutants, defective in some aspect of DNA synthesis or synthesis-coupled step that is displayed most readily when an appropriate stress is applied. Given the complexity of the replication process and its roots in the synthesis and interconversion of precursors, the analysis may be prolonged.

We are grateful to the Cancer Research Campaign, of which R.T.J. is a Research Fellow, for their continued support, and to Ann Hill, Roger Northfield, David Oates, Peggy Pawley and Jacquie Whybrow for their excellent help.

Bradley
,
M. O.
,
Erickson
,
L. C.
&
Kohn
,
K. W.
(
1976
).
Normal DNA strand rejoining and absence of DNA crosslinking in progeroid and aging human cells
.
Mutât. Res.
37
,
279
292
.
Bradley
,
M. O.
&
Taylor
,
V. I.
(
1983
).
Repair-induced DNA double strand breaks after ultraviolet light and either aphidicolin or l-jfi-D-arabinofuranosylcytosine/hydroxyurea
.
Carcinogenesis
4
,
1513
1517
.
Bryant
,
P. E.
&
Iliakis
,
G.
(
1984
).
Possible correlations between cell killing, chromosome damage and DNA repair after X-irradiation
. In
DNA Repair and its Inhibition, Nucl. Acids Symp., Series, no. 13
(ed.
A. R. S.
Collins
, C. S. Downes &
R. T.
Johnson
), pp.
291
-
308
. Oxford, Washington, D.C.: IRL Press.
Burg
,
K.
,
Collins
,
A. R. S.
&
Johnson
,
R. T.
(
1977
).
Effects of ultraviolet light on synchronized Chinese hamster ovary cells; potentiation by hydroxyurea
.
J. Cell Sci.
28
,
29
48
.
Campisi
,
J.
&
Medrano
,
E. E.
(
1983
).
Cell cycle perturbations in normal and transformed fibroblasts caused by detachment from the substratum
.
J. cell. Physiol.
114
,
53
60
.
Chu
,
M. Y.
&
Fischer
,
G. A.
(
1962
).
A proposed mechanism of action of l-/3-D-arabino- furanosylcytosine as an inhibitor of the growth of leukemic cells
.
Biochem. Pharmac.
11
,
423
430
.
Clarkson
,
J. M.
(
1978
).
Enhancement of repair replication in mammalian cells by hydroxyurea
.
Mutat. Res.
52
,
273
284
.
Clarkson
,
J. M.
&
Mitchell
,
D. L.
(
1983
).
The effects of various inhibitors of DNA synthesis on the repair of DNA photoproducts
.
Biochim. biophys. Acta
740
,
417
427
.
Cleaver
,
J. E.
(
1983
).
Structure of repaired sites in human DNA synthesised in the presence of inhibitors of DNA polymerases a and fi in human fibroblasts
.
Biochim. biophys. Acta
739
,
301
311
.
Collins
,
A. R. S.
(
1985
).
Poly(ADP-ribose) is not involved in the rejoining of DNA breaks accumulated to high levels in UV-irradiated HeLa cells
.
Carcinogenesis
6
,
1033
1036
.
Collins
,
A. R. S.
(
1987
).
Estimates of the rate of ligation during excision repair of ultraviolet damaged DNA in mammalian cells
.
Biochim. biophys. Acta (in press)
.
Collins
,
A. R. S.
,
Downes
,
C. S.
&
Johnson
,
R. T.
(
1980
).
Cell cycle-related variations in UV damage and repair capacity in Chinese hamster (CHO-K1) cells
.
J. cell. Physiol.
103
,
179
191
.
Collins
,
A. R. S.
,
Downes
,
C. S.
&
Johnson
,
R. T.
(
1984
).
DNA Repair and its Inhibition, Nucl. Acids Symp. Series, no. 13. Oxford, Washington, D.C.: IRL Press
.
Collins
,
A. R. S.
,
Ord
,
M. J.
&
Johnson
,
R. T.
(
1981
).
Correlations of DNA damage and repair with nuclear and chromosomal damage in HeLa cells caused by methylnitrosamides
.
Cancer Res.
41
,
5176
5187
.
Collins
,
A.
&
Johnson
,
R.
(
1979a
).
Novobiocin: an inhibitor of the repair of UV-ionduced but not X-ray-induced damage in mammalian cells
.
Nucl. Acids Res.
7
,
1311
1320
.
COLLINS
,
A. R. S.
&
Johnson
,
R. T.
(
1979A
).
Repair and survival after UV in quiescent and proliferating Microtus agrestis cells: Different rates of incision and different dependence on DNA precursor supply
.
J. cell. Physiol.
99
,
125
138
.
Collins
,
A. R. S.
&
Johnson
,
R. T.
(
1984
).
The inhibition of DNA repair
. In
Advances in Radiation Biology,
vol.
11
(ed.
J. T.
Lett
), pp.
71
-
129
. London, New York: Academic Press.
Creissen
,
D.
&
Shall
,
S.
(
1982
).
Regulation of DNA ligase activity by poly(ADP-ribose)
.
Nature, Lond.
296
,
271
272
.
Czech
,
M. P.
(
1985
).
The nature and regulation of the insulin receptor: structure and function. A
.
Rev. Physiol.
47
,
357
381
.
Das
,
S. K.
,
Benditt
,
E. P.
&
Loeb
,
L. I.
(
1983
).
Rapid changes in deoxynucleoside triphosphate pools in mammalian cells treated with mutagens
.
Biochem. biophys. Res. Commun.
114
,
458
464
.
Downes
,
C. S.
(
1984
).
Approaches to the quantitative analysis of repair through the use of inhibitors
. In
DNA Repair and its Inhibition, Nucl. Acids Symp. series, no. 13
(ed.
A.
Collins
, C. S. Downes &
R. T.
Johnson
), pp.
231
-
254
. Oxford, Washington, D.C.: IRL Press.
Downes
,
C. S.
,
Ord
,
M. J.
,
Mullinger
,
A. M.
,
Collins
,
A. R. S.
&
Johnson
,
R. T.
(
1985
).
Novobiocin inhibition of DNA excision repair may occur through effects on mitochondrial structure and ATP metabolism, not on repair topoisomerases
.
Carcinogenesis
6
,
1343
1352
.
Elliott
,
G. C.
&
Downes
,
C. S.
(
1986
).
Qualitative differences between replicative and repair synthesis in normal and transformed mouse cells as measured by precursor discrimination
.
Mutât. Res. (in press)
.
Epstein
,
J.
,
Williams
,
J. R.
&
Little
,
J. B.
(
1973
).
Deficient DNA repair in human progeroid cells
.
Proc. natn. Acad. Sci. U.S.A.
70
,
977
981
.
Filatov
,
M. V.
&
Noskin
,
L. A.
(
1983
).
Sensitization of human cells by inhibitors of DNA synthesis following the action of DNA damaging agents
.
Mutât. Res.
110
,
393
399
.
Fornace
,
A. J.
,
Kohn
,
K. W.
&
Kann
,
H. E.
(
1976
).
DNA single strand breaks during repair of UV damage in human fibroblasts and abnormalities of repair in xeroderma pigmentosum
.
Proc, natn. Acad. Sci. U.S.A.
73
,
39
43
.
Francis
,
A. A.
,
Blevins
,
R. D.
,
Carrier
,
W. L.
,
Smith
,
D. P.
&
Regan
,
J. D.
(
1979
).
Inhibition of DNA repair in ultraviolet-irradiated cells by hydroxyurea
.
Biochim. biophys. Acta
563
,
385
392
.
Frei
,
E.
,
Bickers
,
J.
,
Hewlett
,
J.
,
Lane
,
M.
,
Leary
,
W.
&
Tally
,
R.
(
1969
).
Dose schedule and antitumor studies of arabinosyl cytosine
.
Cancer Res.
29
,
1325
1332
.
Frei
,
J.
,
Canal
,
N.
&
Gori
,
E.
(
1958
).
Novobiocin as an uncoupling agent
.
Experimentia
14
,
377
378
.
Friedberg
,
E. C.
(
1985
).
DNA Repair. New York: W. H. Freeman and Co
.
Grosovsky
,
A. J.
&
Little
,
J. B.
(
1983
).
Influence of confluent holding time on UV light mutagenesis in human diploid fibroblasts. Mutât. Res.
110
,
401
412
.
Hackett
,
P. B.
,
Traub
,
P.
&
Gallwitz
,
D.
(
1978
).
The histone genes in HeLa cells are on individual transcriptional units
.
J’, molec. Biol.
126
,
619
635
.
Hittelman
,
W. N.
(
1984
).
Prematurely condensed chromosomes: a model system for visualizing effects of DNA damage, repair and inhibition at the level of chromosome structure
. In
DNA Repair and its Inhibition, Nucl. Acids Symp. Series, no. 13
(ed.
A. R. S.
Collins
, C. S. Downes &
R. T.
Johnson
), pp.
341
-
371
. Oxford, Washington, D.C.: IRL Press.
Hittelman
,
W. N.
&
Pollard
,
M.
(
1984
).
Visualization of chromatin events associated with the repair of ultraviolet light-induced damage by premature chromosome condensation
.
Carcinogenesis
5
,
1277
1285
.
Hodges
,
G.
,
Livingstone
,
D.
&
Franks
,
C.
(
1973
).
The localisation of trypsin in cultured mammalian cells
.
J. Cell Sci.
12
,
887
902
.
Jacobson
,
E. L.
,
Antol
,
K. M.
,
Juarez-Salinas
,
H.
&
Jacobson
,
M. K.
(
1983
).
Poly(ADP- ribose) metabolism in ultraviolet-irradiated human fibroblasts
.
J. biol. Chem.
258
,
103
107
.
Johnson
,
R. T.
,
Collins
,
A. R. S.
&
Waldren
,
C. A.
(
1982
). Prematurely condensed chromosomes and the analysis of DNA and chromosome lesions. In
Premature Chromosome Condensation
(ed.
P. N.
Rao
, R. T. Johnson &
K.
Sperling
), pp.
253
308
.
London, New York
:
Academic Press
.
Johnson
,
R. T.
,
Mullinger
,
A. M.
&
Downes
,
C. S.
(
1978
).
Human minisegregant cells
. In
Methods in Cell Biology, vol. 20
(ed.
D. M.
Prescott
), pp.
255
-
314
. London, New York: Academic Press.
Johnson
,
R. T.
&
Rao
,
P. N.
(
1970
).
Mammalian cell fusion: Induction of premature chromosome condensation in interphase nuclei
.
Nature, Lond.
226
,
717
722
.
Johnson
,
R. T.
,
Squires
,
S.
,
Elliott
,
G. C.
,
Koch
,
G. L. E.
&
Rainbow
,
A. J.
(
1985
).
Xeroderma pigmentosum D-HeLa hybrids with low and high ultraviolet sensitivity associated with normal and diminished DNA repair ability respectively
.
J. Cell Sci.
76
,
115
133
.
Johnson
,
R. T.
,
Squires
,
S.
,
Elliott
,
G. C.
,
Rainbow
,
A. J.
,
Koch
,
G. L. E.
&
Smith
,
M.
(
1986
).
Analysis of DNA repair in XP-HeLa hybrids; lack of correlation between excision repair of UV damage and adenovirus reactivation in an XP(D)-like cell line
.
Carcinogenesis (in press)
.
Kihlman
,
B. A.
&
Natarajan
,
A. T.
(
1984
).
Potentiation of chromosomal alterations by inhibitors of DNA repair
. In
DNA Repair and its Inhibition, Nucl. Acids Symp. Series, no. 13
(ed.
A. R. S.
Collins
, C. S. Downes &
R. T.
Johnson
), pp.
319
-
339
. Oxford, Washington, D.C.: IRL Press.
Kufe
,
D. W.
,
Weichselbaum
,
R.
,
Egan
,
M.
,
Dahlberg
,
W.
&
Fran
,
R. J.
(
1984
).
Lethal effects of l-β-D-arabinofuranosyl cytosine incorporation into deoxyribonucleic acid during ultraviolet repair
.
Molec. Pharmac.
25
,
322
326
.
Lehmann
,
A. R.
,
Kirk-Bell
,
S.
&
Mayne
,
L. V.
(
1979
).
Abnormal kinetics of DNA synthesis in ultraviolet-irradiated cells from patients with Cockayne’s Syndrome
.
Cancer Res.
39
,
4238
4241
.
Mattern
,
M. R.
,
Paone
,
R. F.
&
Day
,
R. S.
(
1982
).
Eukaryotic DNA repair is blocked at different steps by inhibitors of DNA topoisomerases and of DNA polymerases α and β. Biochim. biophys,.Acta
697
,
6
13
.
Mattern
,
M. R.
&
Scudiero
,
D. A.
(
1981
).
Dependence of mammalian DNA synthesis on DNA supercoiling. III. Characterization of the inhibition of replicative and repair-type DNA synthesis by novobiocin and nalidixic acid
.
Biochim. biophys. Acta
653
,
248
258
.
Mayne
,
L. V.
(
1984
).
Inhibitors of DNA synthesis (Aphidicolin and araC/HU) prevent the recovery of RNA synthesis after UV-irradiation
.
Mutât. Res.
131
,
187
191
.
Mayne
,
L. V.
&
Lehmann
,
A. R.
(
1982
).
Failure of RNA synthesis to recover after UV irradiation: an early defect in cells from individuals with Cockayne’s Syndrome and xeroderma pigmentosum
.
Cancer Res.
42
,
1473
1478
.
Milo
,
G. E.
,
Weisbrode
,
S. A.
,
Zimmerman
,
R.
&
McCloskey
,
J. A.
(
1981
).
Ultraviolet radiation-induced neoplastic transformation of normal human cells, in vitro.
Chem-Biol. Interact.
36
,
45
59
.
Miskin
,
R.
&
Reich
,
E.
(
1980
).
Plasminogen activator: Induction of synthesis by DNA damage
.
Cell
19
,
217
224
.
Miskin
,
R.
&
Reich
,
E.
(
1981
).
Induction of plasminogen activator by UV light in normal and xeroderma pigmentosum fibroblasts
.
Proc. natn. Acad. Sci. U.SA.
78
,
6236
6240
.
Mullinger
,
A. M.
&
Johnson
,
R. T.
(
1985
).
Manipulating chromosome structure and metaphase status with ultraviolet light and repair synthesis inhibitors
.
J. Cell Sci.
73
,
159
186
.
Munch-Petersen
,
B.
,
Tyrsted
,
G.
&
Dupont
,
B.
(
1973
).
The deoxyribonucleoside 5′- triphosphate (dATP and dTTP) pool in phytohaemagglutinin-stimulated and non-stimulated human lymphocytes
.
Expl Cell Res.
79
,
249
256
.
Newman
,
C. N.
&
Miller
,
J. H.
(
1983
).
Mutagen-induced changes in cellular deoxycytodine triphosphate and thymidine triphosphate in Chinese hamster ovary cells
.
Biochem. biophys. Res. Commun.
114
,
34
40
.
Nicander
,
B.
&
Reichard
,
P.
(
1985
).
Evidence for the involvement of substrate cycles in the regulation of deoxyribonucleoside triphosphate pools in 3T6 cells
.
J. biol. Chem.
260
,
9216
9222
.
Ockey
,
C. H.
(
1979
).
Quantitative replicon analysis of DNA synthesis in cancer-prone conditions and the defects in Bloom’s Syndrome.J
.
Cell Sci.
40
,
125
144
.
Orr-Weaver
,
T. L.
&
Szostak
,
J. W.
(
1983
).
Yeast recombination: The association between double-strand gap repair and crossing-over
.
Proc. natn. Acad. Sci. U.SA.
80
,
4417
4421
.
Painter
,
R. B.
(
1977
).
Rapid test to detect agents that damage human DNA
.
Nature, Lond.
265
,
650
651
.
Pfeiffer
,
S. E.
&
Tolmach
,
L. J.
(
1967
).
Inhibition of DNA synthesis in HeLa cells by hydroxyurea
.
Cancer Res.
27
,
124
129
.
Radman
,
M.
(
1974
). Phenomenology of an inducible mutagenic DNA repair pathway in Escherichia coli:SOS repair hypothesis. In
Molecular and Environmental Aspects of Mutagenesis
(ed.
L.
Prakesh
,
F.
Sherman
,
M.
Miller
, C. Lawrence & H. W. Tabor), p. 128
.
Springfield
:
Thomas
.
Resnick
,
M.
(
1976
).
The repair of double strand breaks in DNA: A model involving recombinations
.
J. theor. Biol.
59
,
97
106
.
Sarasin
,
A.
&
Benoit
,
A.
(
1986
).
Error-prone replication of ultraviolet-irradiated simian virus 40 occurs in mitomycin c-treated host cells only at low multiplicity of infection
.
Molec. cell. Biol.
6
,
1102
1107
.
Schor
,
S. L.
,
Johnson
,
R. T.
&
Waldren
,
C. A.
(
1975
).
Changes in the organization of chromosomes during the cell cycle: Response to ultraviolet light
.
J. Cell Sci.
17
,
539
565
.
Setlow
,
R. B.
&
Regan
,
J. D.
(
1974
).
DNA repair in human progeroid cells
.
Fedn Proc. Fedn Am. Socs exp. Biol.
33
,
1539
.
Shall
,
S.
(
1984
).
Inhibition of DNA repair by inhibitors of nuclear ADP-ribosyl transferase
. In
DNA Repair and its Inhibition, Nucl. Acids Symp. Series, no. 13
(ed.
A. R. S.
Collins
, C. S. Downes &
R. T.
Johnson
), pp.
143
-
191
. Oxford, Washington, D.C.: IRL Press.
Simons
,
J. W. I. M.
(
1979
).
Development of a liquid-holding technique for the study of DNA repair in human diploid fibroblasts
.
Mutât. Res.
59
,
273
279
.
Sinclair
,
W. K.
(
1965
).
Hydroxyurea: Differential lethal effects on cultured mammalian cells during the cell cycle
.
Science
150
,
1729
1731
.
Sinclair
,
W. K.
(
1967
).
Hydroxyurea: Effects on Chinese hamster cells grown in culture
.
Cancer Res.
27
,
297
308
.
Skoog
,
K. L.
,
Nordenskjold
,
B. A.
&
Bjursell
,
K. G.
(
1973
).
Deoxyribonucleosidetriphosphate pools and DNA synthesis in synchronized hamster cells
.
Eur. J. Biochem.
33
,
428
432
.
Smith
,
C. A.
(
1984
).
Analysis of repair synthesis in the presence of inhibitors
.
In DNA Repair and its Inhibition, Nucl. Acids Symp. Series, no. 13
(ed.
A. R. S.
Collins
, C. S. Downes &
R. T.
Johnson
), pp.
51
—71. Oxford, Washington, D.C.: IRL Press.
Snyder
,
R. D.
(
1984
).
The role of deoxyribonucleoside triphosphate pools in the inhibition of DNA excision repair and replication in human cells by hydroxyurea
.
Mutât. Res.
131
,
163
172
.
Snyder
,
R. D.
,
Carrier
,
W. L.
&
Regan
,
J. D.
(
1981
).
Application of arabinofuranosyl cytosine in the kinetic analysis and quantitation of DNA repair in human cells after ultraviolet irradiation
.
Biophys. J.
35
,
339
350
.
Squires
,
S.
&
Johnson
,
R. T.
(
1983
).
UV induces long lived DNA breaks in Cockayne’s Syndrome and cells from an immunodeficient individual (46BR): defects and disturbance in post incision steps of excision repair
.
Carcinogenesis
4
,
565
572
.
Squires
,
S.
,
Johnson
,
R. T.
&
Collins
,
A. R. S.
(
1982
).
Initial rates of DNA incision in UV- irradiated human cells. Differences between normal, xeroderma pigmentosum and tumour cells
.
Mutât. Res.
95
,
389
404
.
Tanaka
,
M.
&
Yoshida
,
S.
(
1982
).
Altered sensitivity to l-α-D-arabinofuranosylcytidine-5′- triphosphate of DNA polymerase alpha from leukaemic blasts of acute lymphoblastic leukaemia
.
Cancer Res.
42
,
649
665
.
Taylor
,
J. H.
,
Haut
,
W. F.
&
Tung
,
J.
(
1962
).
Effects of fluorodeoxyuridine on DNA replication, chromosome breakage, and reunion
.
Proc. natn. Acad. Sci. U.S.A.
48
,
190
198
.
Teo
,
I. A.
,
Broughton
,
B. C.
,
Day
,
R. S.
,
James
,
M. R.
,
Karran
,
P.
,
Mayne
,
L. V.
&
Lehmann
,
A. R.
(
1983
).
A biochemical defect in the repair of alkylated DNA in cells from an immunodeficient patient (46BR)
.
Carcinogenesis
4
,
559
564
.
Terasima
,
T.
,
Yasukawa
,
M.
&
Kimura
,
M.
(
1985
).
Neoplastic transformation of plateau phase mouse 10T [ineq] cells following single and fractionated doses of X rays
.
Radiat. Res.
102
,
367
377
.
Thelander
,
L.
&
Reichard
,
P.
(
1979
).
Reduction of ribonucleotides. A
.
Rev. Biochem.
48
,
133
158
.
Th’ng
,
J. P. H.
&
Walker
,
I. G.
(
1986
).
Excision repair in the presence of aphidicolin
.
Mutat. Res.
165
,
139
150
.
Thompson
,
L. H.
,
Brookman
,
D. W.
,
Dillehay
,
L. E.
,
Carrano
,
A. V.
,
Mazrimas
,
J. A.
,
Mooney
,
C. L.
&
Minkler
,
J. L.
(
1982
).
A CHO-cell strain having hypersensitivity to mutagens, a defect in DNA strand-break repair, and an extraordinary baseline frequency of sister chromatid exchange
.
Mutat. Res.
95
,
427
440
.
Thompson
,
L. H.
,
Salazar
,
E. P.
,
Brookman
,
K. W.
,
Collins
,
C. C.
,
Stewart
,
S. A.
,
Busch
,
D. B.
&
Weber
,
C. A.
(
1987
).
Recent progress with DNA repair mutants of Chinese hamster ovary cells
.
J. Cell Sci. Suppl.
6
,
97
110
.
Umeda
,
M.
,
Tanaka
,
K.
&
Ono
,
T.
(
1983
).
Effect of insulin on the transformation of BALB/3T3 cells by X-ray irradiation
.
Gann
74
,
864
869
.
Walters
,
R. A.
,
Tobey
,
R. A.
&
Ratliff
,
R. L.
(
1973
).
Cell cycle-dependent variations of deoxyribonucleoside triphosphate pools in Chinese hamster cells
.
Biochim. biophys. Acta
319
,
336
347
.
Webster
,
D.
,
Arlett
,
C. F.
,
Harcourt
,
S. A.
,
Teo
,
I. A.
&
Henderson
,
L.
(
1982
). A new syndrome of immunodeficiency and increased cellular sensitivity to DNA damaging agents. In
Ataxia-Telangiectasia — A Cellular and Molecular Link Between Cancer, Neuropathology and Immune Deficiency
(ed. B. A. Bridges &
D. G.
Harnden
), pp.
379
386
.
London
:
John Wiley and Sons
.
Witkin
,
E. M.
(
1976
).
Ultraviolet mutagenesis and inducible DNA repair in Escherichia coli.
Bact. Rev.
40
,
869
907
.
Yew
,
F. H.
&
Johnson
,
R. T.
(
1979
).
Ultraviolet-induced DNA excision repair in human B and T lymphocytes II. Effect of inhibitors and DNA precursors
.
Biochim. biophys. Acta
562
,
240
251
.
Yu
,
C. K.
&
Sinclair
,
W. K.
(
1968
).
Cytological effects on Chinese hamster cells of synchronizing concentrations of hydroxyurea
.
J. cell. Physiol.
72
,
39
42
.
Zajac-Kaye
,
M.
&
Ts’o
,
P. O. P.
(
1984
).
DNAse I encapsulated in liposomes can induce neoplastic transformation of Syrian hamster embryo cells in culture
.
Cell
39
,
427
437
.
Zimmerman
,
R. J.
&
Little
,
J. B.
(
1983
).
Characteristics of human diploid fibroblasts transformed in vitro by chemical carcinogens
.
Cancer Res.
43
,
2183
2189
.