DNA repair in mammalian cells: Direct DNA damage reversal: elegant solutions for nasty problems.

The genomic integrity of all living organisms is constantly jeopardized by physical [e.g. ultraviolet (UV) light, ionizing radiation] and chemical (e.g. environmental pollutants, endogenously produced reactive metabolites) agents that damage the DNA. To overcome the deleterious effects of DNA lesions, nature evolved a number of complex multi-protein repair processes with broad, partially overlapping substrate specificity. In marked contrast, cells may use very simple repair systems, referred to as direct DNA damage reversal, that rely on a single protein, remove lesions in a basically error-free manner, show high substrate specificity, and do not involve incision of the sugar-phosphate backbone or base excision. This concise review deals with two types of direct DNA damage reversal: (i) the repair of alkylating damage by alkyltransferases and dioxygenases, and (ii) the repair of UV-induced damage by spore photoproduct lyases and photolyases. (Part of a Multi-author Review).

Relationship of the xeroderma pigmentosum group E DNA repair defect to the chromatin and DNA binding proteins UV-DDB and replication protein A.

Cells from complementation groups A through G of the heritable sun-sensitive disorder xeroderma pigmentosum (XP) show defects in nucleotide excision repair of damaged DNA. Proteins representing groups A, B, C, D, F, and G are subunits of the core recognition and incision machinery of repair. XP group E (XP-E) is the mildest form of the disorder, and cells generally show about 50% of the normal repair level. We investigated two protein factors previously implicated in the XP-E defect, UV-damaged DNA binding protein (UV-DDB) and replication protein A (RPA). Three newly identified XP-E cell lines (XP23PV, XP25PV, and a line formerly classified as an XP variant) were defective in UV-DDB binding activity but had levels of RPA in the normal range. The XP-E cell extracts did not display a significant nucleotide excision repair defect in vitro, with either UV-irradiated DNA or a uniquely placed cisplatin lesion used as a substrate. Purified UV-DDB protein did not stimulate repair of naked DNA by DDB- XP-E cell extracts, but microinjection of the protein into DDB- XP-E cells could partially correct the repair defect. RPA stimulated repair in normal, XP-E, or complemented extracts from other XP groups, and so the effect of RPA was not specific for XP-E cell extracts. These data strengthen the connection between XP-E and UV-DDB. Coupled with previous results, the findings suggest that UV-DDB has a role in the repair of DNA in chromatin.

Two human homologs of Rad23 are functionally interchangeable in complex formation and stimulation of XPC repair activity.

XPC-hHR23B protein complex is specifically involved in nucleotide excision repair (NER) of DNA lesions on transcriptionally inactive sequences as well as the nontranscribed strand of active genes. Here we demonstrate that not only highly purified recombinant hHR23B (rhHR23B) but also a second human homolog of the Saccharomyces cerevisiae Rad23 repair protein, hHR23A, stimulates the in vitro repair activity of recombinant human XPC (rhXPC), revealing functional redundancy between these human Rad23 homologs. Coprecipitation experiments with His-tagged rhHR23 as well as sedimentation velocity analysis showed that both rhHR23 proteins in vitro reconstitute a physical complex with rhXPC. Both complexes were more active than free rhXPC, indicating that complex assembly is required for the stimulation. rhHR23B was shown to stimulate an early stage of NER at or prior to incision. Furthermore, both rhHR23 proteins function in a defined NER system reconstituted with purified proteins, indicating direct involvement of hHR23 proteins in the DNA repair reaction via interaction with XPC.

Cloning of a human homolog of the yeast nucleotide excision repair gene MMS19 and interaction with transcription repair factor TFIIH via the XPB and XPD helicases.

Nucleotide excision repair (NER) removes UV-induced photoproducts and numerous other DNA lesions in a highly conserved 'cut-and-paste' reaction that involves approximately 25 core components. In addition, several other proteins have been identified which are dispensable for NER in vitro but have an undefined role in vivo and may act at the interface of NER and other cellular processes. An intriguing example is the Saccharomyces cerevisiae Mms19 protein that has an unknown dual function in NER and RNA polymerase II transcription. Here we report the cloning and characterization of a human homolog, designated hMMS19, that encodes a 1030 amino acid protein with 26% identity and 51% similarity to S.cerevisiae Mms19p and with a strikingly similar size. The expression profile and nuclear location are consistent with a repair function. Co-immunoprecipitation experiments revealed that hMMS19 directly interacts with the XPB and XPD subunits of NER-transcription factor TFIIH. These findings extend the conservation of the NER apparatus and the link between NER and basal transcription and suggest that hMMS19 exerts its function in repair and transcription by interacting with the XPB and XPD helicases.

Different removal of ultraviolet photoproducts in genetically related xeroderma pigmentosum and trichothiodystrophy diseases.

To understand the heterogeneity in genetic predisposition to skin cancer in different nucleotide excision repair-deficient human syndromes, we studied repair of cyclobutane pyrimidine dimers (CPDs) and of pyrimidine(6-4)pyrimidone (6-4PP) photoproducts in cells from trichothiodystrophy (TTD) patients. TTD is not associated with increased incidence of skin cancer, although 50% of the patients are photosensitive and carry a defect in the nucleotide excision repair pathway, similar to Xeroderma pigmentosum patients. However, in striking contrast to TTD, Xeroderma pigmentosum is highly prone to cancer. To address this apparent paradox, two types of studies were conducted: (a) reactivation of UV-irradiated plasmids harboring actively transcribed reporter genes, with or without photolyase treatment before transfection of SV40-transformed fibroblasts; and (b) the kinetics of removal of UV-induced CPDs and 6-4PPs in genomic DNA by immunoblot analysis using lesion-specific mAbs in SV40-transformed and untransformed fibroblasts representative of all genetic TTD complementation groups. Results showed that all cell lines from photosensitive TTD patients efficiently express Cat or luciferase genes in transfected plasmids carrying non-CPD lesions, including 6-4PP, and display wild-type or near-wild-type (50-70% in 3 cell lines) 6-4PP repair in the overall genome after immunoblot analysis. However, CPD lesions (the repair of which is defective in the overall genome) also block the expression of the reporter gene in transfected plasmids. Two cell lines from nonphotosensitive TTD patients showed wild-type levels of repair for both photoproducts in overall genome. A model on the lesion-specific repair in the context of the molecular defect in TTD is proposed. The implication of the defective CPD repair and efficient 6-4PP repair subpathways in cancer prevention in TTD patients is discussed.

Effects of microinjected photoreactivating enzyme on thymine dimer removal and DNA repair synthesis in normal human and xeroderma pigmentosum fibroblasts.

UV-induced thymine dimers (10 J/m2 of UV-C) were assayed in normal human and xeroderma pigmentosum (XP) fibroblasts with a monoclonal antibody against these dimers and quantitative fluorescence microscopy. In repair-proficient cells dimer-specific immunofluorescence gradually decreased with time, reaching about 25% of the initial fluorescence after 27 h. Rapid disappearance of dimers was observed in cells which had been microinjected with yeast photoreactivating enzyme prior to UV irradiation. This photoreactivation (PHR) was light dependent and (virtually) complete within 15 min of PHR illumination. In general, PHR of dimers strongly reduces UV-induced unscheduled DNA synthesis (UDS). However, when PHR was applied immediately after UV irradiation, UDS remained unchanged initially; the decrease set in only after 30 min. When PHR was performed 2 h after UV exposure, UDS dropped without delay. An explanation for this difference is preferential removal of some type(s) of nondimer lesions, e.g., (6-4) photoproducts, which is responsible for the PHR-resistant UDS immediately following UV irradiation. After the rapid removal of these photoproducts, the bulk of UDS is due to dimer repair. From the rapid effect of dimer removal by PHR on UDS it can be deduced that the excision of dimers up to the repair synthesis step takes considerably less than 30 min. Also in XP fibroblasts of various complementation groups the effect of PHR was investigated. The immunochemical dimer assay showed rapid PHR-dependent removal comparable to that in normal cells. However, the decrease of (residual) UDS due to PHR was absent (in XP-D) or much delayed (in XP-A and -E) compared to normal cells. This supports the idea that in these XP cells preferential repair of nondimer lesions does occur, but at a much lower rate.

Characterization of an 8-hydroxy-5-deazaflavin:NADPH oxidoreductase from Streptomyces griseus.

An 8-hydroxy-5-deazaflavin-dependent oxidoreductase was isolated from the actinomycete Streptomyces griseus and purified 590-fold with 72% overall yield. The enzyme catalyzes electron transfer between 8-hydroxy-5-deazaflavins and NADPH. It seems to be more specific than methanogenic oxidoreductase as it has an absolute requirement for both the 5-deazaflavin structure and the presence of an 8-hydroxy group in the substrate. A molecular weight of 42,000 was found with gel permeation chromatography, while SDS gel electrophoresis indicated the presence of two identical subunits. Maximal enzymatic activity was found at 0.32 M NaCl and pH 5.9 for reduction of 8-hydroxy-5-deazaflavin and pH 7.9 for the reverse reaction. From the kinetic constants it was estimated that the main function of this oxidoreductase is probably to provide cells with reduced 8-hydroxy-5-deazaflavin to be used in specific reduction reactions. These results indicate the occurrence of 8-hydroxy-5-deazaflavin-dependent electron transfer in microorganisms not belonging to the archaebacteria.

Studies on a DNA photoreactivating enzyme from Streptomyces griseus. II. Purification of the enzyme.

A procedure is described for the isolation of a DNA photoreactivating enzyme from Streptomyces griseus. Application of chromatography on spherosil, ultraviolet-irradiated DNA-cellulose, DEAE-cellulose and single stranded-DNA-agarose resulted in a 22 000-fold purification with 46 percent recovery on the initial activity. According to polyacrylamide gel electrophoresis the final preparation was virtually homogeneous. The absorption spectrum of the enzyme exhibited a marked absorption band in the 400-460 nm region in addition to protein absorption.

Crystallization and preliminary X-ray diffraction studies of photolyase (photoreactivating enzyme) from the cyanobacterium Anacystis nidulans.

Photolyase (photoreactivating enzyme) from the cyanobacterium Anacystis nidulans was crystallized by the hanging drop vapor diffusion procedure using ammonium sulfate as a precipitant. The pale-yellow crystals were grown to a size of 0.4 mm in length and 0.1 mm in diameter. They belong to the tetragonal space group P4(1)2(1)2 or P4(3)2(1)2 with unit cell dimensions of a = b = 90.7 A and c = 135 A. Assuming that the asymmetric unit contains one molecule, the Vm value is calculated as 2.6 A3/dalton. The crystals are stable towards X-ray exposure and diffract beyond 2.5 A resolution.

DNA photolyase from the fungus Neurospora crassa. Purification, characterization and comparison with other photolyases.

A phr-gene from the filamentous fungus Neurospora crassa was overexpressed in Escherichia coli cells, yielding a biologically active photolyase. After purification till apparent homogeneity, the 66 kDa protein was found to contain equimolar amounts of 5,10-methenyltetrahydrofolic acid (MTHF) and FAD, classifying it as an MTHF-type photolyase. Compared to other MTHF photolyases the absorption maximum of Neurospora photolyase is shifted from ca 380 nm to 391 nm (epsilon = 34,800), while an additional shoulder is present at 465 nm. In dark-adapted enzyme the FAD chromophore is predominantly present in the oxidized form, in contrast with E. coli and Saccharomyces cerevisiae photolyase, which contain mainly semiquinone or fully reduced FAD, respectively. Preillumination or dithionite treatment converted oxidized FAD in Neurospora photolyase into the fully reduced form, with a concomitant shift of the absorption maximum from 391 to 396 nm and disappearance of the 465 nm shoulder. The action spectrum of photoreactivation coincides with the absorption spectrum of preilluminated (reduced) photolyase, extending the spectral region of MTHF-type photolyases from 380 till 396 nm. A quantum yield of 0.57 was obtained for the overall repair reaction. Comparison of spectral properties of FAD in Neurospora photolyase and the model compound lumiflavin points to an apolar microenvironment of photolyase-bound FAD. Neurospora photolyase has distinct advantages over E. coli photolyase as it is more stable and contains a full complement of chromophores.

Expression of an Anacystis nidulans photolyase gene in Escherichia coli; functional complementation and modified action spectrum of photoreactivation.

The Anacystis nidulans photolyase gene inserted in an expression vector plasmid was introduced into Escherichia coli cells and the production of Anacystis photolyase protein was confirmed by reaction with antibodies raised against photolyase purified from A. nidulans cells. The Anacystis photolyase functioned in photoreactivation repair defective E. coli cells. The E. coli transformants exhibited an action spectrum with a maximum around 380 nm similar to that of E. coli photolyase in contrast with the action spectrum of A. nidulans cells which has a maximum at 437 nm. These results indicate that the Anacystis photolyase produced in E. coli cells has enzymatic activity in spite of the apparent lack of its intrinsic 8-hydroxy-5-deazaflavin cofactor.

A putative blue-light receptor from Drosophila melanogaster.

A gene encoding a 62.5 kDa homolog of Drosophila melanogaster photolyase was isolated. Purified recombinant protein contained a flavin adenine dinucleotide chromophore. The recombinant protein did not show photolyase activity for either cyclobutane pyrimidine dimers or 6-4 photoproducts in vitro as well as in vivo in Escherichia coli host cells, suggesting that the protein is not a DNA repair enzyme but a blue-light photoreceptor. Reverse transcription polymerase chain reaction analysis showed that the gene is more expressed in head than in body and that it is more expressed in antennae than in legs, wings and mouth appendages. In a phylogenetic tree of the photolyase family, the Drosophila photolyase homolog is located in a cluster containing 6-4 photolyases and mammalian photolyase homologs, which is only distantly related to the clade of higher plant blue-light photoreceptors. The mammalian photolyase homologs are more closely related to Drosophila 6-4 photolyase than to the Drosophila photolyase homolog, suggesting different roles of the photolyase homologs.

Existence and expression of photoreactivation repair genes in various yeast species.

Photoreactivation repair (Phr) activities in cell extracts of 13 different yeast species were measured by the Haemophilus influenzae transformation assay. Five species including Schizosaccharomyces pombe showed no or low enzymatic activity. In contrast to the other species, chromosomal DNAs of these 5 species did not show detectable hybridization using a DNA fragment of the photolyase PHR1 gene of Saccharomyces cervisiae as a probe even at a low stringency condition. When the PHR1 gene was attached to the 5'-flanking sequence of the iso-1-cytochrome c (CYC-1) gene of S. cerevisiae and introduced into S. pombe cells, the transformants acquired a high Phr activity, indicating that the PHR1 gene alone can provide a Phr-negative species with this repair activity and the light-absorbing cofactor(s) must be present in S. pombe. Our results also demonstrated that the 5'-flanking sequence of the S. cerevisiae CYC-1 gene works in S. pombe as a regulatory element.

Mitochondrial DNA repair by photolyase.

Photolyase genes of Saccharomyces cerevisiae and Escherichia coli were expressed in S. cerevisiae and photoreactivation in nuclei and mitochondria of the host cells was analyzed by determination of survival and petit rates. Yeast photolyase was able to repair mitochondrial DNA effectively, whereas E. coli photolyase could reduce only a small fraction of the petit rate produced by UV irradiation. Analysis using fusion between yeast photolyase and E. coli lacZ genes as well as a chimeric gene between yeast and E. coli photolyase genes suggests the importance of the protruding amino terminal region of the yeast photolyase for its transport into mitochondria. A significant similarity between the protruding amino termini of yeast photolyase and yeast uracil-DNA-glycosylase suggests a common functional importance of the terminal sequences for both DNA repair enzymes.

Use of in vivo and in vitro assays for the characterization of mammalian excision repair and isolation of repair proteins.

Elucidation of the molecular mechanism of mammalian nucleotide excision repair requires the availability of purified proteins, DNA substrates with defined lesions and suitable repair assays. Repair assays introduced in recent years vary from testing individual steps and successions of steps in vitro to systems that closely reflect the entire process in vivo. In the first part of this review, an in vivo microinjection system is discussed. The second part of the article reviews an in vitro system for study of repair synthesis promoted by cell extracts. Both systems can be utilized as assays during the purification of protein factors that complement repair-defective xeroderma pigmentosum cells. The effect of purified repair proteins from other organisms on mammalian repair is also considered.

Unscheduled DNA synthesis in xeroderma pigmentosum cells after microinjection of yeast photoreactivating enzyme.

Photoreactivating enzyme (PRE) from yeast causes a light-dependent reduction of UV-induced unscheduled DNA synthesis (UDS) when injected into the cytoplasm of repair-proficient human fibroblasts (Zwetsloot et al., 1985). This result indicates that the exogenous PRE monomerizes UV-induced dimers in these cells competing with the endogenous excision repair. In this paper we present the results of the injection of yeast PRE on (residual) UDS in fibroblasts from different excision-deficient XP-strains representing complementation groups A, C, D, E, F, H and I (all displaying more than 10% of the UDS of wild-type cells) and in fibroblasts from two excision-proficient XP-variant strains. In fibroblasts belonging to complementation groups C, F and I and in fibroblasts from the XP-variant strains UDS was significantly reduced, indicating that pyrimidine dimers in these cells are accessible to and can be monomerized by the injected yeast PRE. The UDS reduction in the XP-variant strains is comparable with the effect in wild-type cells. In cells from complementation groups C, F and I the reduction is less than in wild-type and XP-variant cells. Fibroblasts belonging to groups A, D, E and H did not show any reduction in UDS level after PRE injection and illumination with photoreactivating light. These results give evidence that the genetic repair defect in some XP-strains is probably due to an altered accessibility of the UV-damaged sites.

Microinjected photoreactivating enzymes from Anacystis and Saccharomyces monomerize dimers in chromatin of human cells.

Photoreactivating enzymes (PRE) from the yeast Saccharomyces cerevisiae and the cyanobacterium Anacystis nidulans have been injected into the cytoplasm of repair-proficient human fibroblasts in culture. After administration of photoreactivation light, PRE-injected cells displayed a significantly lower level of UV-induced unscheduled DNA synthesis (UDS) than non-injected cells. This indicates that monomerization of the UV-induced pyrimidine dimers in the mammalian chromatin had occurred as a result of photoreactivation by the injected PRE at the expense of repair by the endogenous excision pathway. Purified PRE from yeast is able to reduce UDS to 20-25% of the UDS found in non-injected cells, whereas the in vitro more active PRE from A. nidulans gives a reduction to only 70%. This suggests that the eukaryotic enzyme is more efficient in the removal of pyrimidine dimers from mammalian chromatin than its equivalent purified from the prokaryote A. nidulans.

Partial characterization of the DNA repair protein complex, containing the ERCC1, ERCC4, ERCC11 and XPF correcting activities.

The nucleotide excision repair (NER) protein ERCC1 is part of a functional complex, which harbors in addition the repair correcting activities of ERCC4, ERCC11 and human XPF. ERCC1 is not associated with a defect in any of the known human NER disorders: xeroderma pigmentosum, Cockayne's syndrome or trichothiodystrophy. Here we report the partial purification and characterization of the ERCC1 complex. Immunoprecipitation studies tentatively identified a subunit in the complex with an apparent MW of approximately 120 kDa. The complex has affinity for DNA, but no clear preference for ss, ds or UV-damaged DNA substrates. The size of the entire complex determined by non-denaturing gradient gels (approximately 280 kDa) is considerably larger than previously found using size separation on glycerol gradients (approximately 120 kDa). Stable associations of the ERCC1 complex with other known repair factors (XPA, XPC, XPG and TFIIH complex) could not be detected.

Xeroderma pigmentosum group A correcting protein from calf thymus.

A proteinous factor was purified from calf thymus and HeLa cells, which specifically corrects the excision repair defect of xeroderma pigmentosum complementation group A (XP-A) cells. Recovery of UV-induced unscheduled DNA synthesis after microinjection of XP-A cells was used as a quantitative assay for the correcting activity of protein preparations. XP-A correcting protein appears to be very stable as it withstands heating to 100 degrees C and treatment with SDS or 6 M urea. A molecular weight of 40-45 kD was found both under native (gel filtration) and denaturing (SDS-PAGE) conditions. Calf XP-A protein binds to single-stranded DNA more strongly than to double-stranded DNA, but shows no clear preference for UV-irradiated DNA. Polyclonal antibodies raised against human recombinant XP-A protein, which strongly inhibit UV-induced unscheduled DNA synthesis of normal human cells, completely abolished XP-A correcting activity when mixed with calf thymus preparations. This indicates a close relationship between human gene product and the calf protein. In the final preparation two main protein bands were present. Only one band at approx. 41 kD showed both DNA binding activity in Southwestern blots and immune reaction with human XP-A antibody, suggesting that this is the active calf XP-A correcting factor.

Photoreactivation of the thymine dimer containing DNA octamer d(GCGTTGCG).d(CGCAACGC) by the photoreactivating enzyme from Anacystis nidulans.

Irradiation of the double-stranded octamer d(GCGTTGCG).d(CGCAACGC) with UV light causes dimerization of the two central thymine residues. Proton NMR data reveal that this photodimer has the same chemical structure as the photodimer, which is formed upon UV irradiation of the single strand d(GCGTTGCG), a cis-syn-cyclobutane-type thymine dimer. Irradiation of the purified thymine dimer double-stranded octamer d(GCGTTGCG).d(CGCAACGC) with visible light in the presence of photoreactivating enzyme isolated from Anacystis nidulans leads to an increase in absorbance at 260 nm, which is characteristic for the repair of thymine dimers. The NMR spectrum recorded after the photoreactivating treatment indeed shows that a complete conversion to the starting octamer has occurred.