A re-analysis of the metabolic fate of ultraviolet light (u.v.)-induced cyclobutyl pyrimidine dimers in the DNA of dermal fibroblasts from patients with different genetic forms of xeroderma pigmentosum (XP), a rare cancer-prone skin disorder, has provided new insight into the mode of dimer repair in normal human cells. When DNA isolated from post-u.v. incubated cultures was subjected to enzymic photoreactivation (PR) to probe dimer authenticity, single-strand scissions were produced in the damaged DNA of incubated XP group A and D cells, but not in DNA from XP group C cells or normal controls. Since enzymic PR treatment ruptures only the cyclobutane ring, these results suggested that in dimer excision-defective XP group A and D strains, the intradimer phosphodiester bond may have been cleaved without site restoration. Such a cleavage event had not previously been detected; the possibility that this reaction may be an early step in the normal excision-repair process is supported by the observed release of free thymidine (dThd) and its monophosphate (TMP), but not of thymine, upon photochemical reversal of the dimercontaining excision fragments isolated from post-u.v. incubated normal cells. The combined number of dThd and TMP molecules released was equal to ≈ 80% of the number of dimers photoreversed; for such release to occur, the dimer must both be at one end of an excised fragment and contain an internal phosphodiester break. Taken together, these data lead us to propose a novel model for dimer repair in human cells in which hydrolysis of the intradimer phosphodiester linkage precedes the concerted action of a generalized ‘bulky lesion-repair complex’ involving conventional strand incision/lesion excision/repair resynthesis/strand ligation reactions.

A milestone in the study of enzymic DNA repair processes in Homo sapiens occurred in 1968 with the disclosure by Cleaver that defective repair of ultraviolet light (u.v.)-induced cyclobutyl pyrimidine dimers is an intrinsic property of cultured cells from patients with xeroderma pigmentosum (XP), an autosomal recessive disorder predisposing to sunlight-associated skin cancer (Kraemer, 1983). Aside from demonstrating the crucial role of DNA repair systems in protecting mankind against an otherwise intolerable incidence of solar u.v.-induced malignancies, this seminal discovery unmasked a ready (indeed to date the sole) repository of mutant strains for detailed investigation into repair pathways that are operative on pyrimidine dimers in human cells. In the intervening two decades considerable effort has been expended in numerous laboratories in an attempt to define the full range of DNA repair anomalies associated with the syndrome (reviewed by Friedberg et al. 1979; Paterson et al. 1984; Kraemer, 1983). As indicated in Table 1, cultured dermal fibroblasts from no fewer than 119 genetically unrelated XP donors have now been subjected to one or more assays designed to monitor different parameters of the DNA repair machinery in general and the excision-repair process in particular. The cardinal biochemical anomaly in 98 of these unrelated XP strains (i.e. in ≈ 80% of those examined) is a malfunction in the so-called nucleotide mode of excision repair; in this pathway a bulky lesion, such as a pyrimidine dimer, is excised as part of an oligonucleotide; this is followed by insertion of a normal nucleotide sequence using the intact complementary strand as template (Haseltine et al. 1980; Sancar & Rupp, 1983; Paterson et al. 1984). On the basis of a biochemical complementation test, somatic cell fusion studies have thus far assigned the excision repair-defective XP strains to eight mutually complementing and hence genetically distinct groups designated A-H (see Table 1). The remaining 21 XP strains display a pronounced deficiency in executing postreplication repair (also called replicative or daughterstrand repair), a poorly defined process that is believed to promote base-pairing fidelity when the de novo DNA synthesis machinery is called upon to replicate past dimers or other non-coding lesions in template DNA (Cleaver, 1980). Pending the development of a suitable complementation assay, all of these 21 XP strains have been lumped together to form the ninth complementation group, termed variant.

Table 1.

Pertinent laboratory hallmarks of xeroderma pigmentosum complementation groups*

Pertinent laboratory hallmarks of xeroderma pigmentosum complementation groups*
Pertinent laboratory hallmarks of xeroderma pigmentosum complementation groups*

Despite the numerous inquiries into the DNA repair abnormalities in excision repair-defective strains belonging to XP complementation groups A-H (Kraemer, 1983; Paterson et al. 1984), little progress has been made in identifying the precise biochemical defect underlying any one of these eight genetic forms of the disease. This poor understanding of the root causes of the different XP genotypes has prompted us to undertake a new line of experimentation, one that promises not only to provide new insight into the basic deficiency in certain XP complementation groups but may also lead to a clearer definition of early reactions in the nucleotide repair mode that are operative on dimers in normal cells. A brief account of our findings to date forms the subject of this chapter.

As a prelude to a description of our investigations into the molecular mechanism of dimer repair in cultured human cells, we shall first outline contemporary models, derived from the more comprehensive and sophisticated experimentation that is possible in simpler prokaryotic systems, of different mechanisms for the nucleotide excision-repair mode. This is followed by a concise account of the relevant hallmarks of excision repair-defective XP strains; this latter review will emphasize group D cells, for it was the bizarre and seemingly inconsistent excision-repair traits of these cells that moved us to initiate the studies reported here.

Models of dimer excision-repair mechanisms

Extensive research into the metabolic fate of pyrimidine dimers in bacteria and bacteriophages has led to the identification of two closely related but distinct mechanisms by which the nucleotide excision-repair mode can operate on these u.v. photoproducts (for particulars, see Haseltine, 1983; Sancar & Rupp, 1983; Friedberg, 1985). While details concerning the innermost workings of these complex multistep systems remain to be elucidated, the basic biochemical reactions in each pathway are reasonably well-defined and candidate enzymes capable of catalysing most of these reactions have been detected and characterized to varying degrees (Friedberg, 1985). The most widely accepted models of the two mechanisms for effecting dimer excision in prokaryotes are diagrammatically illustrated in Fig. 1. The first mechanism, which was discovered in Escherichia coli over two decades ago, is thought to consist of the following sequential reactions (see right-hand ‘loop’ of Fig. 1): (1) the phosphodiester backbone of the dimer-containing strand is incised upstream from (i.e. on the 5 ′ side of) the lesion by a damage-specific enzyme referred to as ‘UV endonuclease’; (2) a second scission is introduced downstream from (i.e. on the 3 ′ side of) the u.v. photoproduct by an exonuclease, thereby facilitating the release of the lesion as part of a short oligonucleotide (N.B. in E. coli, these two cleavage steps may be performed concurrently by the concerted action of the so- called UVRABC excinuclease complex, leading to the liberation of the dimer near the middle of a single-strand fragment 12 – 13 nucleotides in length; the gap thus formed may be subsequently enlarged by exonucleolytic degradation of the chain in the 5 ′ ₒ 3 ′ direction); (3) nucleotides complementary to those in the opposite intact strand are then inserted into the resultant gap by the action of a DNA polymerase in an operation termed DNA repair synthesis; and finally (4) the repair patch and the juxtaposed pre-existing material are covalently linked by a DNA ligase, thus completing restoration of the site to normal structure and function (Sancar & Rupp, 1983; Friedberg, 1985).

Fig. 1.

Popular model for the two distinctive mechanisms by which the nucleotide mode of the excision-repair process is believed to operate on u.v.-induced pyrimidine dimers in different prokaryotic systems. In this model a dimer in either chain of the duplex DNA is excised as part of a short oligonucleotide, as described in the text. The term nucleotide excision distinguishes this mode from a second excision-repair mode called base excision. (In the latter, a carcinogen-damaged or non-conventional base is released in its free form (rather than in the nucleotide, or actually oligonucleotide, form) by the action of a highly specific DNA glycosylase, and the resultant AP site is then restored by sequential strand incision/AP site excision/repair patch insertion/strand ligation reactions (for details, consult Friedberg, 1985).) (From Patersone/al. (1985) by permission of Plenum Press.)

Fig. 1.

Popular model for the two distinctive mechanisms by which the nucleotide mode of the excision-repair process is believed to operate on u.v.-induced pyrimidine dimers in different prokaryotic systems. In this model a dimer in either chain of the duplex DNA is excised as part of a short oligonucleotide, as described in the text. The term nucleotide excision distinguishes this mode from a second excision-repair mode called base excision. (In the latter, a carcinogen-damaged or non-conventional base is released in its free form (rather than in the nucleotide, or actually oligonucleotide, form) by the action of a highly specific DNA glycosylase, and the resultant AP site is then restored by sequential strand incision/AP site excision/repair patch insertion/strand ligation reactions (for details, consult Friedberg, 1985).) (From Patersone/al. (1985) by permission of Plenum Press.)

The second mechanism by which nucleotide excision repair is known to operate on pyrimidine dimers was detected recently in Micrococcus luteus and bacteriophage T4-infected E. coli. As depicted in the left-hand ‘loop’ of Fig. 1, the UV endonucleases encoded by the M. luteus and phage T4 genomes accomplish the initial strand incision event by performing two concerted reactions: a dimer-recognizing DNA glycosylase activity hydrolyses the N-glycosy 1 bond between the 5 ′ -pyrimidine member of the dimer and its corresponding deoxyribose, and the phosphodiester linkage 3 ′ to the newly formed ‘denuded’ sugar is then severed by an apyrimidinic/apurinic (AP) endonuclease activity (Haseltine et al. 1980; McMillan et al. 1981; Nakabeppu & Sekiguchi, 1981). In all likelihood, the repair pathway then proceeds to completion in a manner similar to the last three steps in the first mechanism; that is, exonucleolytic cleavage downstream from the dimer (thus releasing a singlestrand fragment containing a dimerized pyrimidine-pyrimidylate moiety at its 5 ′ end), followed by insertion of a repair patch mediated by a DNA polymerase, and restitution of strand continuity at the site by the action of a DNA ligase (Friedberg, 1985).

Properties of excision repair-defective XP strains

In a number of laboratories representative strains of XP complementation groups A-H have been subjected to various assays monitoring different steps in excision repair in an attempt to pinpoint the stage at which blockage occurs in each complementation group. On the basis of these analyses, it appears that all eight XP groups are impaired at the same (strand incision) step (Friedberg et al. 1979; Kraemer, 1983; Paterson et al. 1984). The simplest interpretation of this unexpected finding is that no fewer than eight loci in the human genome may be involved in effecting the initial step in the nucleotide mode of excision repair. Other, less-direct, explanations cannot be excluded, however; these include the possibility that the action of otherwise normal strand-incising enzymic machinery may be precluded in some XP complementation groups, due to faulty execution of an unidentified preincision event(s) (e.g. the likely requirement that at particular structural and/or functional domains in the genetic material of higher organisms, the basic excision-repair process cannot be initiated at a dimer-containing site unless some form of localized processing of chromatin takes place first, such as ridding the DNA of associated histones and other nucleoproteins in order for repair enzymes to gain access to their substrates (Mortelmans et al. 1976; Mansbridge & Hanawalt, 1983; Mullenders et al. 1984; Karentz & Cleaver, 1986)).

These aforementioned assays also provide a quantitative estimate of the residual repair capacity in a given XP strain relative to that found in normal controls from healthy volunteers (see, e.g., Table 1). As a rule, XP strains assigned to a particular excision repair-defective complementation group carry out a similar amount of residual dimer repair, the particular value being characteristic of that specific group. This generally holds irrespective of which assay is utilized. XP group A cells, for example, appear to be extremely deficient in handling dimers via the nucleotide excision-repair mode, as judged by several criteria, including the following two: (1) a negligible ability to effect the removal of dimer-containing sites in DNA during incubation of u.v.-treated cells (detected in retrospect as u.v.-induced sites in extracted cellular DNA which are susceptible to the strand-incising activity of a UV endonuclease present in a crude protein extract of M. luteus-, such sites are henceforth referred to as UV endonuclease-sensitive sites); and (2) a marked inadequacy in performing u.v.-induced DNA repair synthesis during postirradiation cell incubation (measured as unscheduled DNA synthesis or DNA repair replication) (see Table 1). (NB The first criterion of repair capability, UV endonuclease-sensitive site removal, is presumed to monitor the second (lesion excision) step in the excision-repair process whereas the second, u.v.-induced DNA repair synthesis, is customarily taken as a direct measure of the third (repair patch insertion) step in the pathway (Paterson, 1978).) In some other excision repair-defective XP groups, while a deficiency in dimer repair is evident, an appreciable residual level still remains, especially in groups E and F and, to a lesser extent, group C; this is the case for both repair hallmarks in Table 1.

A notable exception to this general inter-assay consistency when quantifying residual repair capacities of XP strains is readily apparent in group D. Strains belonging to this group are markedly, if not completely, deficient in acting upon UV endonuclease-sensitive sites but, paradoxically, still manage to accomplish a considerable amount of u.v.-induced DNA repair synthesis (see Table 1). (A second exception is evident in group F; in this case, however, the paradox is just the opposite of that seen in group D. That is to say, group F cells appear to be much more efficient at eliminating UV endonuclease-sensitive sites than would be predicted on the basis of their ability to carry out u.v.-stimulated DNA repair synthesis (Table 1).)

Several additional pieces of assorted evidence support the notion that the primary defect in XP group D is distinct from that arising in the other seven XP groups defective in excision repair. Aside from their uniqueness in lacking one of the two AP endonucleases normally present in human cells (Kuhnlein et al. 1978), XP group D cells exhibit the peculiar and unprecedented property of being more competent than normal cells in performing host-cell reactivation of partially depurinated SV40 DNA (Kudrna et al. 1979). Furthermore, in contradistinction to other XP strains harbouring a defect in excision repair, group D strains are much less deficient at excising angelicin adducts from DNA relative to their deficit in removing pyrimidine dimers (Cleaver & Gruenert, 1984).

In attempting to resolve the apparent discrepancy in the excision-repair properties of XP group D strains, we first considered a possible trivial explanation, namely that the reported disparity in residual levels of UV endonuclease-sensitive site removal and of u.v.-induced DNA repair synthesis in these mutant strains may be simply due to some unidentified confounding variable in experimental protocol or cell culture methodology in different laboratories. Adopting a common experimental approach for both assays, we therefore monitored normal and XP group D cultures for their relative abilities: (1) to eliminate UV endonuclease-sensitive sites; and (2) to carry out u.v.-stimulated DNA repair replication. Our findings, like those of others (Zelle & Lohman, 1979; Kraemer et al. 1975) indicated that XP group D cells do indeed perform a significant amount of u.v.-induced repair synthesis even though they are markedly impaired in recognizing dimer-containing sites in their DNA (for details, see Paterson, 1982).

Having ruled out inter-laboratory differences in experimental conditions as a viable explanation for the peculiar excision-repair hallmarks reported for XP group D strains, an alternative working hypothesis was constructed. This hypothesis was based on the supposition that, on exposure to u.v. radiation, XP group D cells may operate on only a portion of the dimer-containing sites in their DNA, and do so in a faulty manner, somehow replacing pre-existing nucleotides (presumably downstream from the dimer) with new ones, thus accounting for the residual level of DNA repair synthesis, while at the same time failing to excise the photoproduct and thereby explaining the continued presence of all UV endonuclease-sensitive sites initially introduced. We further postulated that the insertion of such a putative ‘pseudo-repair patch’ could conceivably be accompanied by some structural modification in the immediate vicinity of the dimer. As a test of this postulate, we reinvestigated the metabolic fate of dimer-containing sites in XP group D fibroblasts by measuring the photoreactivability (a well-documented diagnostic probe of dimer authenticity; Harm, 1976) of such sites in naked DNA isolated from normal and XP group D strains as a function of cell incubation time after exposure to germicidal light (far u.v. of chiefly 254nm wavelength). A flow chart of our protocol is shown in Fig. 2. In short, after co-incubation for appropriate periods, parallel 3H-labelled, u.v.-irradiated and 14C-labelled, unirradiated cultures of each strain were lysed; the cellular DNAs were then co-extracted, subjected to enzymic photoreactivation (PR) treatment (using extensively purified Streptomyces griseus photolyase, an enzyme that binds to a dimer-containing site and, upon absorption of fluorescent light energy, ruptures the cyclobutane ring in situ, thereby regenerating the two constitutive pyrimidines (Sutherland, 1978)), and incubated with and/, luteus protein extract containing UV endonuclease activity. It was reasoned that this succession of treatments might uncover dimer-containing sites that may have been metabolically altered in vivo so as to render them refractory to restoration by enzymic PR but still susceptible to the strand-incising action of UV endonuclease. Finally, the incidence of single-strand nicks arising in the u.v.-damaged DNA as a result of this course of treatments, and hence the frequency of modified dimer-containing sites present in this DNA at the time of cell lysis, was quantified by the classical technique of velocity sedimentation in alkaline sucrose gradients (see Paterson et al. 1981).

Fig. 2.

Experimental protocol developed to assay for aberrant repair of u.v.-induced pyrimidine dimers in XP complementation group D fibroblasts. The treatment schedule was, in effect, identical to that followed in a conventional enzymic assay for dimer measurement (Paterson et al. 1981), except for the introduction of an enzymic PR step (to convert any intact dimers that remain back to constitutive monomeric bases in situ) before incubation with an.lf luteus protein extract. The 14C-labelled DNA from sham- irradiated cultures served as an internal control, which permitted correction for nonspecific strand breakage arising from the various physical manipulations and enzymic treatments to which the isolated DNA is subjected in performing the assay. (From Paterson (1982) by permission of Elsevier Biomedical Press.)

Fig. 2.

Experimental protocol developed to assay for aberrant repair of u.v.-induced pyrimidine dimers in XP complementation group D fibroblasts. The treatment schedule was, in effect, identical to that followed in a conventional enzymic assay for dimer measurement (Paterson et al. 1981), except for the introduction of an enzymic PR step (to convert any intact dimers that remain back to constitutive monomeric bases in situ) before incubation with an.lf luteus protein extract. The 14C-labelled DNA from sham- irradiated cultures served as an internal control, which permitted correction for nonspecific strand breakage arising from the various physical manipulations and enzymic treatments to which the isolated DNA is subjected in performing the assay. (From Paterson (1982) by permission of Elsevier Biomedical Press.)

In keeping with our reasoning, the protocol disclosed the appearance of strand breaks in the DNA of u.v.-irradiated and incubated group D (XP2NE and XP3NE) strains that was not observed in an identically treated normal (GM38) strain (for a fuller description of these initial results, see Paterson, 1982). As revealed in Fig. 3, the incidence of these peculiar strand breaks in the DNA of XP2NE cells reached a plateau value of ≈ 8 per 108 daltons by 48 h post-u.v. irradiation. This maximal yield was equivalent to approximately 15 % of the number of dimers acted upon by normal cells in the same 48-h period and was similar in relative magnitude to the residual amount of u.v.-induced DNA repair replication (≈ 15 – 20 % of normal) occurring in this same XP group D strain. These comparable residual levels of modified dimers and of DNA repair synthesis in group D strains are consistent with the notion that the sites of presumed dimer modification and pseudo repair-patch insertion are one and the same.

Fig. 3.

Incidence of novel sites, as detected by subsequent enzymic PR and M. luteus extract treatments, accumulating in the DNA of normal and XP strains as a function of postirradiation time. The experimental protocol was as outlined in Fig. 2. Each datum point is the mean of multiple determinations (S.E. < 15 %). (From Paterson et al. (1985) by permission of Plenum Press.)

Fig. 3.

Incidence of novel sites, as detected by subsequent enzymic PR and M. luteus extract treatments, accumulating in the DNA of normal and XP strains as a function of postirradiation time. The experimental protocol was as outlined in Fig. 2. Each datum point is the mean of multiple determinations (S.E. < 15 %). (From Paterson et al. (1985) by permission of Plenum Press.)

Accumulation of such novel sites in cellular DNA in response to germicidal light was not restricted to XP strains belonging to complementation group D. These same sites also appeared at similar rates in XP group A (XP12BE) fibroblasts; on the other hand, group C (XP4RO), E or variant fibroblasts responded like normal controls (Fig. 3; Paterson, 1982; Paterson et al. 1985).

In a later, more extensive experimental series, we ran controls in which the M. luteus extract treatment was omitted from the protocol of Fig. 2, expecting to observe no strand breaks, even though the modified sites presumably still accumulated. Much to our surprise, however, essentially the same incidence of strand breaks was detected in XP group A and D cells (but not in normal or XP group C cells) when the purified DNA from postirradiation-incubated cultures were subjected to enzymic PR treatment alone. This unforeseen observation led us to modify our original postulate: we supposed that during incubation of u.v.-damaged XP group A or D fibroblasts, the intradimer phosphodiester linkage at certain dimer containing sites may be severed and that at such modified sites the physical continuity of individual polynucleotide chains is then maintained solely by the cyclobutane bridge joining the two pyrimidine bases. In group A fibroblasts, the excision-repair process very probably aborts at this stage, given that these cells carry out little, if any, DNA repair synthesis in response to u.v. treatment (Table 1). In contrast, in group D cells, cleavage of the intradimer phosphodiester bond is possibly accompanied by insertion of an abnormal repair patch judging from, as pointed out above, the excellent correlation between residual levels of altered dimer-containing sites and of u.v.-stimulated repair synthesis in such cells.

The relationship between the accumulation of modified dimer-containing sites in u.v.-treated group D strains and the other abnormalities peculiar to these XP strains upon sustaining other forms of genotoxic damage (see Background Overview) must await further experimentation.

In contemplating the significance of the modified dimer-containing sites discovered in XP group A and D strains, it struck us that the occurrence of such sites may not necessarily be solely confined to these mutant strains. Instead, their accumulation in XP group A and D fibroblasts may represent a piling up of an intermediate reaction product (because of a defect in a later reaction) that is usually seen only transiently in normal human fibroblasts. If so, then our inability to detect these modified dimer-containing sites in u.v.-damaged normal cells may have been simply because such ‘internally incised dimer’ structures, once formed, are immediately released from genomic DNA by subsequent reactions that rapidly follow in a fully functional excision-repair system. It was further reasoned that if: (1) hydrolysis of the intradimer phosphodiester linkage does indeed occur in normal cells before excision of an oligonucleotide containing a thymine (Thy)-thymine (TT) dimer, for example; and (2) the lesion is located at one end of the excision fragment, then exposure of isolated excision products to a photochemically reversing (i.e. cyclobutane ring-rupturing) fluence of germicidal light should lead to the release of either free thymidine (dThd) or thymidine monophosphate (TMP). (It was assumed that one member of the dimer pair was not present as Thy, since such a structure would have been detected by La Belle & Linn (1982) in an earlier study designed to determine whether the excision of dimers in human cells is initiated by rupture of an intradimer N-glycosyl bond.)

To test the validity of this reasoning, we adopted an experimental protocol that was very similar to that used by La Belle & Linn (1982). This approach takes advantage of the fact that the dimer-containing excision products, which are released from genomic DNA by the enzymic excision-repair machinery, are retained within human fibroblasts for prolonged periods (⩾ 24 h); consequently, these excision fragments can be easily recovered in the trichloroacetic acid (TCA)-soluble fraction of cell cultures that have been incubated for appropriate times after u.v. irradiation. In brief, our experimental approach went as follows: normal (GM38) fibroblast cultures, whose DNA had been prelabelled with [3H]dThd, were exposed to 40 J m − 2 of germicidal light, incubated for 24h, and then lysed in 5 % TCA; finally, the dimer-containing excision fragments were isolated from the acid-soluble fraction and subjected to a six-step procedure utilizing reverse-phase high performance liquid chromatography (HPLC) in order to determine the quantitative relationship between Thy, dThd and TMP released, on the one hand, and Thy-containing dimers monomerized by a photoreversing fluence (5 · 5 kJ m −, 2) of germicidal light, on the other. The general protocol and rationale behind inclusion of each of the six steps are outlined in Fig. 4. It should be noted that in our chromatographic system, TMP migrates in the same region as the excision products and hence its presence has to be measured indirectly as the increase in dThd following bacterial alkaline phosphatase digestion.

Fig. 4.

Six-step protocol developed to assay for dimer photoreversal-induced liberation of Thy, dThd and TMP from TCA-soluble excision fragments isolated from u.v.- irradiated (40 J m − 2) and incubated (24 h) normal human (GM38) fibroblasts. (From Weinfeld et al. (1986) by permission of the American Chemical Society.)

Fig. 4.

Six-step protocol developed to assay for dimer photoreversal-induced liberation of Thy, dThd and TMP from TCA-soluble excision fragments isolated from u.v.- irradiated (40 J m − 2) and incubated (24 h) normal human (GM38) fibroblasts. (From Weinfeld et al. (1986) by permission of the American Chemical Society.)

The results obtained for each of the six steps in Fig. 4 are summarized in Table 2. It can be seen that exposure of isolated excision fragments to photoreversing irradiation did not result in an elevated level of free Thy, thus confirming the earlier observation of La Belle & Linn (1982). This inability to demonstrate the liberation of Thy upon photochemical monomerization of Thy-containing dimers in excision products strongly implies that hydrolysis of N-glycosyl bonds on dimerized pyrimidines is not an intermediate reaction in the excision-repair pathway operating on these photoproducts in human cells.

Table 2.

HPLC analysis of TCA-soluble material from post-u.v. (40Jm−, 2)- incubated (24 h) normal human (GM38) fibroblasts for release of Thy, dThd and TMP upon photochemical reversal of pyrimidine dimers in isolated excision fragments*

HPLC analysis of TCA-soluble material from post-u.v. (40Jm−, 2)- incubated (24 h) normal human (GM38) fibroblasts for release of Thy, dThd and TMP upon photochemical reversal of pyrimidine dimers in isolated excision fragments*
HPLC analysis of TCA-soluble material from post-u.v. (40Jm−, 2)- incubated (24 h) normal human (GM38) fibroblasts for release of Thy, dThd and TMP upon photochemical reversal of pyrimidine dimers in isolated excision fragments*

In contrast to the findings for Thy, 3 · 0% (5 · 3 % – 2 · 3%) and 5 · 4% (5 · 9% – 0 · 5 %) of the total TCA-soluble radioactivity was released as dThd and TMP, respectively, by the dimer-photoreversing treatment. The same u.v. fluence reduced the fraction of the total acid-soluble radioactivity in Thy-containing (i.e. TT and cytosine-thymine (CT)) dimers from 49% (25 · 2%+ 23 · 8%) to 28% (14 · 6% + 13 · 4%), indicating that 21 % of the radioactivity was photochemically converted from dimers to monomers. Assuming that: (1) the genomic DNA was uniformly labelled with [3H]dThd; (2) dimers were situated at one end of the isolated excision fragments; and (3) the intradimer phosphodiester bond was hydrolysed, then each photochemically monomerized TT or CT dimer should have released, on average, half of its radioactivity as dThd and/or TMP. Hence it follows that a maximum of 50 % of the tritium in photomonomerized Thy-containing dimers, that is, 10 · 5 % of the total acid-soluble counts, would be liberated as dThd and/or TMP. The actual amount of labelled dThd plus TMP recovered (8 · 4%) was equal to ≈ 80 % of this theoretical maximum, and we therefore estimate that about 80 % of the Thy-containing dimers have their intradimer phosphodiester linkage ruptured and are also located at one extremity of the excised oligonucleotides. This shortfall in dThd/TMP molecules photoliberated compared to dimers photoreversed can be easily rationalized if, for example, the second aforementioned assumption (namely, that the dimer lies at one end of the excision fragment) does not hold for every fragment: photochemical monomerization of a dimer with at least one adjoining nucleotide on both sides would generate two shorter pieces, each at least dimeric in number. Whatever the reason, our assumptions nonetheless seem to fit the vast majority of the excision fragments.

It is not feasible at present to ascertain whether the dimer-containing excision products from postirradiation-incubated normal fibroblasts have remained unaltered in length from the time of their enzymic excision from genomic DNA until their isolation from cultures at 24 h after u.v. treatment. It is, of course, entirely possible that the excision pieces examined by us first undergo postexcision degradation by non-specific endogenous nucleases. In fact, exonuclease activities operative on single-stranded DNA have been detected in various human sources (Doniger & Grossman, 1976; Friedberg et al. 1977). These and related nucleases may routinely catalyse the release of nucleotides from one or both termini of the excision fragments, until presumably becoming arrested upon encountering the distorted conformation of the oligonucleotides in the neighbourhood of the dimer. The average size of the repair patch inserted in the gap left by the release of a dimer is estimated to be approximately 20 – 25 nucleotides (derived from fig. 5 of Smith & Okumoto, 1984); consequently, the average size of the dimer-containing oligonucleotides at the time of their release from genomic DNA may well be considerably larger than their estimated size (≈ 3 · 7 nucleotides; Weinfeld et al. 1986) at the time of their isolation. (Alternatively, of course, this repair patch size may represent subsequent gapwidening.) Along a similar vein, our findings do not rule out the possibility that the apparent scission of the intradimer phosphodiester linkage in excision products may also be a consequence of postexcision exonucleolytic activity. We regard this latter possibility to be unlikely, however, in view of our evidence that the excision-repair apparatus appears to malfunction in XP group A and D fibroblasts after this intradimer incision reaction has occurred.

Fig. 5.

Proposed model for the nucleotide mode of excision repair acting on cyclobutyl pyrimidine dimers induced in the DNA of cultured human fibroblasts by u.v. irradiation. For simplicity, only the dimer-containing strand of the duplex DNA macromolecule is shown. The notable difference between this model and that for prokaryotic systems in Fig. 1 is that here the first reaction in the multi-step repair pathway is mediated by a putative pyrimidine dimer-DNA phosphodiesterase instead of a dimer-DNA glycosylase, as in M. luteus and phage T4-infected E.coli, or an excinuclease complex, as in uninfected E. coli. In the scheme outlined here, the dimer is located at the 5 ′ end of the excised oligonucleotide and the scission of the intradimer phosphodiester bond yields 3 ′ - P and 5 ′ -OH end groups. Our current data are equally compatible with other schemes, and the precise location and nature of the breaks leading to dimer removal must await further experimentation. (From Paterson et al. (1985) by permission of Plenum Press.)

Fig. 5.

Proposed model for the nucleotide mode of excision repair acting on cyclobutyl pyrimidine dimers induced in the DNA of cultured human fibroblasts by u.v. irradiation. For simplicity, only the dimer-containing strand of the duplex DNA macromolecule is shown. The notable difference between this model and that for prokaryotic systems in Fig. 1 is that here the first reaction in the multi-step repair pathway is mediated by a putative pyrimidine dimer-DNA phosphodiesterase instead of a dimer-DNA glycosylase, as in M. luteus and phage T4-infected E.coli, or an excinuclease complex, as in uninfected E. coli. In the scheme outlined here, the dimer is located at the 5 ′ end of the excised oligonucleotide and the scission of the intradimer phosphodiester bond yields 3 ′ - P and 5 ′ -OH end groups. Our current data are equally compatible with other schemes, and the precise location and nature of the breaks leading to dimer removal must await further experimentation. (From Paterson et al. (1985) by permission of Plenum Press.)

Instead, our preferred interpretation of the data presented here is that hydrolysis of the phosphodiester linkage between dimerized pyrimidines may constitute the initial reaction in the nucleotide excision-repair process that acts on these u.v. photoproducts in human dermal fibroblasts. In this model, which is illustrated in Fig. 5, the postulated role of the putative pyrimidine dimer-DNA phosphodiesterase activity is to induce a conformational change at the dimer-containing site such that the site becomes a substrate for a generalized ‘bulky lesion-repair complex’, possibly similar to the UVRABC excinuclease complex, which appears to recognize a vast array of chemically diverse lesions in the DNA of A1, coli (Sancar & Rupp, 1983). Breakage of the intradimer phosphodiester bond may also serve to relieve conformational stress introduced by the rigid cyclobutane ring fusing the two pyrimidines, thus restoring hydrogen bonding of nearby base-pairs in the doublestranded helix and, by so doing, presumably increasing the fidelity of de novo DNA synthesis on a u.v.-damaged template. According to our model, cleavage of the intradimer phosphodiester linkage is then followed by classical strand incision/lesion excision/patch insertion/strand ligation steps. This scheme for dimer repair in human cells differs from that postulated for M. luteus and phage T4-infected E. coli in that the N-glycosyl bond on the 5 ′ -pyrimidine of the dimer remains intact. It should be noted, however, that in both models the same phosphodiester bond is cut; in human cells this is proposed as the initial step, whereas in the two bacterial systems depyrimidination of the sugar moiety on the 5’ side of this bond is believed to precede the intradimer backbone-incision reaction.

This work was supported primarily by Atomic Energy of Canada Limited and the US National Cancer Institute through contract NO1-CP-21029 (Basic) with the Clinical and Environmental Epidemiology Branches, NCI, Bethesda, MD, and in the final stages of the study by a Heritage Medical Scientist award to M.C.P. and a Postdoctoral Fellowship to M.W. from the Alberta Heritage Foundation for Medical Research. We are grateful to V. Bjerkelund for secretarial assistance.

Cleaver
,
J. E.
(
1980
).
DNA damage, repair systems and human hypersensitive diseases
.
J. environ. Path. Toxicol.
3
,
53
68
.
Cleaver
,
J. E.
&
Gruenert
,
D. C.
(
1984
).
Repair of psoralen adducts in human DNA: Differences among xeroderma pigmentosum complementation groups
.
J. invest. Derm.
82
,
311
315
.
Doniger
,
J.
&
Grossman
,
L.
(
1976
).
Human correxonuclease. Purification and properties of a DNA repair exonuclease from placenta..J, biol. Chem.
251
,
4579
4587
.
Friedberg
,
E. C.
(
1985
).
DNARepair,
pp.
141
322
.
New York
:
W. H. Freeman & Co
.
Friedberg
,
E. C.
,
Cook
,
K. H.
,
Duncan
,
J.
&
Mortelmans
,
K.
(
1977
).
DNA repair enzymes in mammalian cells
.
Photochem. Photobiol. Rev.
2
,
263
322
.
Friedberg
,
E. C.
,
Ehmann
,
U. K.
&
Williams
,
J. I.
(
1979
).
Human diseases associated with defective DNA repair
.
Adv. Radiat. Biol.
8
,
85
174
.
Harm
,
H.
(
1976
). Repair of UV-irradiated biological systems: Photoreactivation.
In Photochemistry and Photobiology of Nucleic Adds,
vol.
2
(ed.
S. Y.
Wang
), pp.
219
263
.
New York
:
Academic Press
.
Haseltine
,
W. A.
(
1983
).
Ultraviolet light repair and mutagenesis revisted
.
Cell
33
,
13
17
.
Haseltine
,
W. A.
,
Gordon
,
L. K.
,
Lindan
,
C. P.
,
Grafstrom
,
R. H.
,
Shaper
,
N. L.
&
Grossman
,
L.
(
1980
).
Cleavage of pyrimidine dimers in specific DNA sequences by a pyrimidine dimer DNA-gly cosylase of M. luteus
.
Nature, Lond.
285
,
634
641
.
Karentz
,
D.
&
Cleaver
,
J. E.
(
1986
).
Excision repair in xeroderma pigmentosum group C but not group D is clustered in a small fraction of the total genome. Mutat. Res.
165
,
165
174
.
Kraemer
,
K. H.
(
1983
). Heritable diseases with increased sensitivity to cellular injury. In
Update: Dermatology in General Medicine
(ed.
T. B.
Fitzpatrick
,
A. Z.
Eisen
,
K.
Wolff
, I. M. Freedberg &
K. F.
Austen
), pp.
113
142
.
New York
:
McGraw-Hill Book Co
.
Kraemer
,
K. H.
,
Coon
,
H. G.
,
Petinga
,
R. A.
,
Barrett
,
S. F.
,
Rahe
,
A. E.
&
Robbins
,
J. H.
(
1975
).
Genetic heterogeneity in xeroderma pigmentosum: Complementation groups and their relationship to DNA repair rates
.
Proc. natn. Acad. Sci. U.S.A.
72
,
59
63
.
Kudrna
,
R. D.
,
Smith
,
J.
,
Linn
,
S.
&
Penhoet
,
E. E.
(
1979
).
Survival of apurinic SV40 DNA in the D complementation group of xeroderma pigmentosum
.
Mutat. Res.
62
,
173
181
.
Kuhnlein
,
U.
,
Lee
,
E.
,
Penhoet
,
E. E.
&
Linn
,
S.
(
1978
).
Xeroderma pigmentosum fibroblasts of the D group lack an apurinic DNA endonuclease species with a low apparent Km. Nucl. Acids Res
.
5
,
951
960
.
La Belle
,
M.
&
Linn
,
S.
(
1982
).
In vivo excision of pyrimidine dimers is mediated by a DNA A-glycosylase in Micrococcus luteus but not in human fibroblasts
.
Photochem. Photobiol.
36
,
319
324
.
Mansbridge
,
J. N.
&
Hanawalt
,
P. C.
(
1983
). Domain-limited repair of DNA in ultraviolet irradiated fibroblasts from xeroderma pigmentosum complementation group C. In
Cellular Responses to DNA Damage
(ed. E. C. Friedberg &
B. N.
Bridges
), pp.
195
207
.
New York
:
Alan R. Liss
.
McMillan
,
S.
,
Edenberg
,
H. J.
,
Radany
,
E. H.
,
Friedberg
,
R. C.
&
Friedberg
,
E. C.
(
1981
).
denV gene of bacteriophage T4 codes for both pyrimidine dimer-DNA glycosylase and apyrimidinic endonuclease activities
.
J. Virol.
40
,
211
223
.
Mortelmans
,
K.
,
Friedberg
,
E. C.
,
Slor
,
H.
,
Thomas
,
G.
&
Cleaver
,
J. E.
(
1976
).
Defective thymine dimer excision by extracts of xeroderma pigmentosum cells
.
Proc. natn. Acad. Sci. U.S.A.
73
,
2757
2761
.
Mullenders
,
L. H. F.
,
van Kesteren
,
A. C.
,
Bussman
,
C. J. M.
, van
Zeeland
,
A. A.
&
Natarajan
,
A. T.
(
1984
).
Preferential repair of nuclear matrix associated DNA in xeroderma pigmentosum. complementation group C
.
Mutat. Res.
141
,
75
82
.
Nakabeppu
,
Y.
&
Sekiguchi
,
M.
(
1981
).
Physical association of pyrimidine dimer DNA glycosylase and apurinic/apyrimidinic DNA endonuclease essential for repair of UV-damaged DNA
.
Proc. natn. Acad. Sci. U.SA.
78
,
2742
2746
.
Paterson
,
M. C.
(
1978
).
Use of purified lesion-recognized enzymes to monitor DNA repair in vivo.
Adv. Radiat. Biol.
7
,
1
53
.
Paterson
,
M. C.
(
1982
).
Accumulation of non-photoreactivable sites in DNA during incubation of UV-damaged xeroderma pigmentosum group A and D cells
.
Prog. Mutat. Res.
4
,
183
192
.
Paterson
,
M. C.
,
Gentner
,
N. E.
,
Middlestadt
,
M. V.
,
Mirzayans
,
R.
&
Weinfeld
,
M.
(
1985
). Hereditary and familial disorders linking cancer proneness with abnormal carcinogen response and faulty DNA metabolism.
In Epidemiology and Quantitation of Environmental Risk in Humans from Radiation and Other Agents
(ed.
A.
Castellani
), pp.
235
267
.
New York
:
Plenum Press
.
Paterson
,
M. C.
,
Gentner
,
N. E.
,
Middlestadt
,
M. V.
&
Weinfeld
,
M.
(
1984
).
Cancer predisposition, carcinogen hypersensitivity, and aberrant DNA metabolism
.
J. cell. Physiol. Suppl.
3
,
45
62
.
Paterson
,
M. C.
,
Smith
,
B. P.
&
Smith
,
P. J.
(
1981
). Measurement of enzyme-sensitive sites in UV- or y-irradiated human cells using Micrococcus luteus extracts. In
DNA Repair: A Laboratory Manual of Research Procedures,
vol.
1
, part A (ed. E. C. Friedberg &
P. C.
Hanawalt
), pp.
99
111
.
New York
:
Marcel Dekker, Inc
.
Sancar
,
A.
&
Rupp
,
W. D.
(
1983
).
A novel repair enzyme: UVRABC excision nuclease of Escherichia coli cuts a DNA strand on both sides of the damaged region
.
Cell
33
,
249
260
.
Smith
,
C. A.
&
Okumoto
,
D. S.
(
1984
).
Nature of DNA repair synthesis resistant to inhibitors of polymerase a in human cells
.
Biochemistry
23
,
1383
1391
.
Sutherland
,
B. M.
(
1978
). Enzymatic photoreactivation of DNA. In
DNA Repair Mechanisms
(ed.
P. C.
Hanawalt
, E. C. Friedberg &
C. F.
Fox
), pp.
113
122
.
New York
:
Academic Press
.
Weinfeld
,
M.
,
Gentner
,
N. E.
,
Johnson
,
L. D.
&
Paterson
,
M. C.
(
1986
).
Photoreversaldependent release of thymidine and thymidine monophosphate from pyrimidine dimercontaining DNA excision fragments isolated from ultraviolet-damaged human fibroblasts
.
Biochemistry
25
,
2656
2664
.
Zelle
,
B.
&
Lohman
,
P. H. M.
(
1979
).
Repair of UV-endonuclease-susceptible sites in the 7 complementation groups of xeroderma pigmentosum A through G
.
Mutat. Res.
62
,
363
368
.