Chromosomes were isolated from Chinese hamster ovary cells that had been pulse-labeled with 3H-thymidine and bromodeoxyuridine and allowed to proceed to metaphase in the presence of colcemid. The chromosomes were then irradiated with 313 nm light and the length of the pulse-labeled DNA segments was estimated by the distribution of 3H-DNA in alkaline sucrose gradients. The average rate of fork displacement, calculated from this measurement, was 0.9 μm/min, which was in excellent agreement with measurements from interphase cells. By the same method, hydroxyurea was shown to retard DNA chain growth by 50–60% in both isolated chromosomes and interphase cells.

The discovery that ataxia-telangiectasia (A-T) patients (Morgan et al., 1968; Gotoff et al., 1967; Cunliffe et al., 1975), as well as cells from these patients (Taylor et al., 1975), are hypersensitive to ionizing radiation (Fig. 1) has sparked an outpouring of research aimed at determining the basis of this radiosensitivity and its relationship to the pathology of the disease. The increased susceptibility of A-T cells to killing by ionizing radiation can largely be correlated with the increased frequency of radiation-induced chromosomal aberrations (Bender et al., 1985, 1988; Higurachi and Conen, 1973; Nagasawa et al., 1985; Natarajan and Meyers, 1979; Taylor, 1982; Zampetti-Bosseler and Scott, 1981); however, the molecular basis of this extraordinary radiosensitivity is still not understood. Many chromatid-type aberrations are observed after irradiation of A-T cells in the G1 phase of the cell cycle (Bender et al., 1985; Natarajan and Meyers, 1979; Taylor et al., 1976; Taylor, 1978), something almost never seen in normal cells or in cells from patients with other genetic diseases. This suggests that some kind of damage that is always repaired in normal cells before they move into S phase is not repaired in A-T cells. Unfortunately, this is a clue that no one has yet been able to exploit.

Transfection, with a human cosmid clone library, of an ataxia-telangiectasia (AT) cell line (AT5BIVA) from complementation group D previously resulted in the isolation of a cell line (1B3) with partially restored resistance to ionizing radiation. We rescued the integrated cosmid sequences within 1B3 and obtained two cosmid clones that contained overlapping DNA from chromosomal region 11q23, previously shown to be the region containing the AT gene(s) from three complementation groups. Isolation of an apparently full-length 3.0-kb cDNA from a HeLa cell library demonstrated a previously unidentified gene (ATDC) within these cosmid clones. The transfected copy of the ATDC gene in 1B3 is truncated at the 3' end but is a complete transcription unit, because of the presence of SV40 termination sequences within the adjacent cosmid DNA. After further screening of cosmid clones from a chromosome 11 library, we identified contiguous DNA that contained the missing portion of the gene. Southern blot analysis indicated that the ATDC gene is present in a single copy in the human genome; however, RNA blot analysis revealed mRNA of several sizes (1.8, 2.6, 3.0, 4.7, and 5.7 kb) that varied among different cell lines. Because no large rearrangements were detected in AT5BIVA cells by Southern or RNA blot analysis, any alteration in the ATDC gene in this cell line would involve a point mutation or a small rearrangement. Transfection of the AT5BIVA cell line with one of the cosmids partially restored radioresistance. Analysis of 100 X-radiation hybrid cell lines containing various fragments from the chromosomal region 11q23 showed that the ATDC gene is closely linked to THY1. The ATDC gene therefore lies outside the linkage region predicted to contain the AT gene(s) for complementation groups A and C, indicating a separate locus for the AT complementation group D gene.

A model for the formation of DNA rearrangements in somatic cells is presented. Two double-strand breaks at the junctions between unreplicated DNA and newly replicated DNA generate four double-stranded DNA molecules that can recombine to form tandem duplications, inversions, deletions and extrachromosomal DNA circles.

The release of DNA 5'-terminal deoxyribose-phosphate residues from enzymatically incised apurinic/apyrimidinic sites by human cell extracts has been under investigation. During the course of these studies, we observed that ataxia telangiectasia cell extracts modify deoxyribose-phosphate (dRp) residues by converting them to an altered form, dRp-X, which shows altered chromatographic properties on HPLC analysis. The chemical nature of the adduct is as yet unknown, but dRp-X is stable to both heat and acid. The modification requires an enzymatic activity and a low-molecular weight co-factor. Extracts of normal cells contain a dialyzable inhibitor that suppresses the reaction occurring with ataxia telangiectasia cell extracts. Formation of dRp-X has been observed in 7 out of 7 ataxia telangiectasia lymphoblastoid lines which represent at least 3 genetic complementation groups. Similar modification of dRp did not occur with extracts of cells of normal origin, nor those representing Fanconi's anaemia, xeroderma pigmentosum, Bloom's syndrome, Werner's syndrome or Friedreich's ataxia.

There is an old saying: "I'd rather be lucky than good, any day." I sometimes (read often) believe that I am living proof of the truth of that adage. In 1955, intermediary metabolism was in its last throes as the "glamour" subject of biology. I finished my Ph.D. work at Ohio State with a dissertation on how Pseudomonas utilized ethylene glycol as a sole carbon source. It was not a brilliant piece of work-it was never published-and I was left feeling very uncertain about my potential as an independent investigator. I had applied to many places for a job (postdoctoral appointments were relatively rare in those days), and one of them was to Brookhaven National Laboratory. I had earlier generated an interest in the killing effects of incorporated radioisotopes, when working on 32p uptake into bacteria, and in my letter to Brookhaven I mentioned this interest. The letter came to the attention of Dr. Walter L. (Pete) Hughes, then head of the Microbiology Division of the Medical Department, who was envisaging the use of tritium as a means of selectively killing growing cells. Thus the letter fortuitously hit the right target. I went to work with Pete in January 1956, just as he prepared the first batch of tritium-labeled thymidine. I had stumbled in on the development of one of the most important tools of modern biology, namely, the use of tritium as a biological tracer in general and, as [3H]thymidine, a specific labeled precursor for DNA. Early in the summer of 1956 Pete informed me that some fellow by the name of Herb Taylor was visiting Phil Woods in the Biology Department and wanted to use [3H]thymidine for a tracer experiment he had devised. We had already begun to grow cells in [3H]thymidine to see if it would kill them, so the [3H]thymidine was available, and Pete agreed

Sixty-eight human fibroblast cell strains were assayed for radioresistant DNA synthesis (RDS), which is defined here as the absence of a steep component of inhibition of DNA synthesis in a dose-response curve when rate of DNA synthesis is plotted against radation doses from 0 to 20 Gy or more. Twenty-seven strains from patients who were previously diagnosed to have ataxia-telangiectasia (AT) were positive for this feature. Among the cell strains that did not show RDS were two from AT obligate heterozygotes (i.e., the parents of AT patients), two from patients with Alzheimer disease, two from patients with Friedreich ataxia, one from a patient with Bloom syndrome, one from a patient with Down syndrome, and six from patients with various immunodeficiencies. Four strains demonstrated RDS that was less pronounced than in most AT cells: one was from a patient with Nijmegen breakage syndrome, one was from a patient without ataxia but with choreiform movement disorder, telangiectasia, and elevated concentrations of α-fetoprotein in the blood, and two were from AT patients. RDS therefore is not a necessary trait of human genetic diseases that involve radiosensitivity or immunodeficiency. Although recent reports suggest that some AT patients do not exhibit RDS, we found RDS in all the AT cell we tested.

SV40-transformed ataxia-telangiectasia (AT) cells were transfected with a cosmid that contains a normal human DNA library and a selectable marker, the neo gene, which endows successfully transformed mammalian cells with resistance to the antibiotic G418. After a three-part selection protocol for G418 resistance and radioresistance, a cell line stably resistant to ionizing radiation was recovered. Cells from this line were irradiated with 50 Gy of X-rays and fused with non-transfected AT cells. Among the G418-resistant colonies recovered was one that was stably resistant to radiation. Resistance to ionizing radiation of both the primary transfectant line and its fusion derivative was intermediate between that of AT cells and normal cells, as assayed by colony-forming ability and measurement of radiation-induced G2 chromatid aberrations; both cell lines retained AT-like radioresistant DNA synthesis. These results suggest that, because radioresistance in the transfected cells was not as great as that in normal human cells, the two hallmarks of AT, radiosensitivity and radioresistant DNA synthesis, may still be the result of a single defective AT gene.

Patients with ataxia-telangiectasia (AT), a human autosomal recessive genetic disease, are uniformly hypersensitive to ionizing radiation as measured by colony-forming ability and by chromosomal aberrations. Obligate heterozygotes, i.e., parents of AT patients, are slightly more radiosensitive than normal humans in terms of both colony-forming ability and chromosomal aberrations formed in G2. Thus, this system not only furnishes a model system to study factors that are responsible for radioresistance in normal human beings, but is also a unique tool for determining the role of gene dosage on radiation-induced cell killing. Because AT cells seem to be hypomutable to ionizing radiation, they also can be used to study the relationship between radiosensitivity and mutability and, therefore, carcinogenesis. Isolation of the defective gene that causes hypersensitivity in AT cells and its counterpart in normal cells should lead to a breakthrough in our understanding of radiation effects and how they can be prevented in human beings.

An X-ray-sensitive Chinese hamster ovary cell line was isolated by means of a semi-automated procedure in which mutagenized cells formed colonies on top of agar, were X-irradiated, and were photographed at two later times. We compared the photographs to identify colonies that displayed significant growth arrest. One of the colonies identified in this manner produced a stable line (irs1SF) that is hypersensitive to ionizing radiation. The X-ray dose at which 10% of the population survives (D10) is 2.25 Gy for irs1SF and 5.45 Gy for the parental line. The new mutant is also moderately sensitive to ethyl methanesulfonate. irs1SF performs only half as much X-ray-induced repair replication as the parental line, indicating a defect in excision repair. This defect is believed to be the primary cause of the line's radiosensitivity. Although irs1SF repairs DNA double-strand breaks at a normal rate, it repairs single-strand breaks more slowly than normal. irs1SF has an elevated number of spontaneous chromatid aberrations and produces significantly higher numbers of X-ray-induced chromatid aberrations after exposure during the G1 phase of the cell cycle. The line is hypomutable, with X-ray exposure inducing only one-third as many 6-thioguanine-resistant colonies as the parental line.

We assessed the role of acentric chromosome fragments in gene amplification by using cell fusion techniques to introduce the fragmented chromosomes of a donor Chinese hamster ovary (CHO) cell line that contained the dihydrofolate reductase (dhfr) gene(s) into a CHO cell line deficient for dhfr. Chromosome fragments were successfully integrated into cells at a frequency of approximately 3%. Methotrexate-resistant variants arose much more frequently in two cell lines derived from these successful cell fusions than in wild-type CHO cells. The hybrid cell lines also amplified their dhfr genes more readily than did the CHO cell line used as dhfr donor.

ABSTRACT One of the first responses observed in S phase mammalian cells that have suffered DNA damage is the inhibition of initiation of DNA replicons. In cells exposed to ionizing radiation, a singlestrand break appears to be the stimulus for this effect, whereby the initiation of many adjacent replicons (a replicon cluster) is blocked by a single-strand break in any one of them. In cells exposed to ultraviolet light (u.v.), replicon initiation is blocked at fluences that induce about one pyrimidine dimer per replicon. The inhibition of replicon initiation by u.v. in Chinese hamster cells that are incapable of excising pyrimidine dimers from their DNA is virtually the same as in cells that are proficient in dimer excision. Therefore, a single-strand break formed during excision repair of pyrimidine dimers is not the stimulus for inhibition of replicon initiation in u.v.-irradiated cells. Considering this fact, as well as the comparative insensitivity of human ataxia telangiectasia cells to u.v.-induced inhibition of replicon initiation, we propose that a relatively rare lesion is the stimulus for u.v.-induced inhibition of replicon initiation.

Two SV40-transformed human cell lines, GM637, derived from a normal human subject, and GM5849, derived from a patient with ataxia-telangiectasia (A-T), were grown in increasing concentrations of the cytotoxic agent methotrexate (MTX). The GM637 line was naturally more resistant to methotrexate than was GM5849 and, over a 5-month period, became resistant even to very high concentrations (up to 100 microM). The GM5849 line became resistant to 500 nM methotrexate during the same period. However, dot blot and Southern blot analyses showed that both cell lines had amplified their dihydrofolate reductase (dhfr) genes to about the same extent, approx. 50-fold. Using the GM5849 line with amplified dhfr, we attempted to determine if interruption of DNA synthesis by hydroxyurea would cause DNA to be replicated twice within a single cell cycle, as has been reported for Chinese hamster ovary cells. No evidence for such a phenomenon was obtained.

Previous reports have suggested that transient inhibition of DNA synthesis by chemicals or ultraviolet light causes some of the DNA to replicate more than once in one cell cycle, i.e., that it induces overreplication of DNA. The data that led to this suggestion were obtained from cesium chloride equilibrium density gradient analyses, in which cells were incubated with bromodeoxyuridine so that the DNA synthesized after incubation with the inhibitor could be densitometrically separated from the DNA that had been radioactively labeled before incubation with the inhibitor. An unresolved problem with these analyses was that the data also suggested that overreplication must have occurred in control cells, i.e., those not incubated with an inhibitor of DNA synthesis. We show here that the latter result is probably due to an artifact of cesium chloride equilibrium density gradient analysis, probably because of nonspecific trapping of DNA in regions of the gradients where there are large amounts of DNA. We also used another protocol that avoids this artifact; with this protocol any overreplicated DNA would be found where heavy-heavy DNA bands and nonspecific trapping cannot occur. When this protocol was used there was no evidence that transient inhibition of DNA synthesis induces overreplication of DNA.

Although the mechanisms leading to gene amplification are poorly understood, it has recently been proposed that the initial event of amplification is the rereplication of a variable, but relatively large, amount of the genome within a single cell cycle. We sought evidence for rereplication of DNA as a basis for gene amplification through two cytogenetic techniques: differential staining for sister-chromatid exchange analysis and premature chromosome condensation. Synchronized Chinese hamster ovary cells were incubated continuously with bromodeoxyuridine and treated with hydroxyurea (HU) when cells were approximately 2 h into the S phase. After 6 h exposure to HU, the drug was removed and at 3 h intervals thereafter metaphase cells were collected and the chromosomes were stained by the fluorescence-plus-Giemsa procedure. No staining patterns consistent with rereplication of DNA were observed. Since HU causes cytogenetic damage, the premature chromosome condensation technique was used to determine the kinetics of chromosome damage after removal of HU. Extensive G2 chromosome damage within 1 h after removal of HU from the medium was found, although cesium chloride gradient analysis showed that there was no rereplication of DNA during this time. Contrary to a previous report, these results provide no evidence that incubation of cells with HU during S phase induces rereplication of DNA within a single cell cycle. The results observed are consistent with the hypothesis that drug-induced aberrations and the subsequent abnormal segregation of chromosomal fragments are the first steps in the process that leads to gene amplification in drug-treated mammalian cells.