Full text loading...
Chapter 17 : DNA Damage Tolerance and Mutagenesis in Eukaryotic Cells
Category: Microbial Genetics and Molecular Biology
Ebook: Choose a downloadable PDF or ePub file. Chapter is a downloadable PDF file. File must be downloaded within 48 hours of purchase
This chapter discusses pathways of DNA damage tolerance and mutagenesis for UV radiation damage in eukaryotes, from being simple model organisms such as budding yeast (Saccharomyces cerevisiae) to more complex organisms, including mammalian cells. The scope of this chapter is confined largely to UV-C radiation photoproducts, since, as in Escherichia coli, they represent the type of DNA damage whose study has resulted in the most detailed insights into molecular mechanisms. Other types of lesions that are possibly of environmental relevance are mentioned briefly. The chapter first briefly summarizes important observations in search of molecular biological explanations. The latter have revealed multiple mechanisms and consequences of tolerating DNA lesions that alter normal coding capacity. The second half of the chapter revisits important concepts originally developed for E. coli, notably translesion DNA synthesis by specialized DNA polymerases.
Distribution of single-base-pair substitutions in a portion of the SUP4-o tRNA gene (bp 31 to 89 ) of an S. cerevisiae wild-type strain treated with 254-nm UV radiation with and without subsequent photoreactivation (PR). Only the transcribed strand is shown (the anticodon is at bp 36 to 38 ). Total base pair substitution frequencies are indicated in gold. The base substitutions shown represent the analyzed events. (Adapted from references 5 and 180 .)
Detection of untargeted mutations in a budding-yeast mating system. UV-irradiated haploid cells containing a CYC1 deletion (cyc1Δ-363) are fused with a nonirradiated partner carrying a point mutation (cyc1-91) in the region deleted in cyc1Δ-363. Mutants with reversion of the point mutation (cyc1-91 → CYC+) in the unirradiated genome are detected among the diploid progeny.
ade2 adeX mutational system for determining the timing of mutation fixation in S. cerevisiae. (A) ade2 mutants form red colonies as a result of the accumulation of an intermediate of the adenine biosynthetic pathway. Inactivation of a metabolic step upstream of ade2 by a second mutation (adeX) prevents the formation of the red pigment; hence, the colony is white. (B) If (stationary-phase haploid) ade2 cells are UV irradiated and an adeX mutation has undergone mutational fixation in both DNA strands prereplicatively (lower section), a pure-white colony will result. If the sequence alteration has affected only one DNA strand before replication, one daughter cell will show the ade2-conferred phenotype and (if this cell remains viable) a red-and-white sectored colony will result (upper section).
Dependence of UV radiation-induced cytotoxicity and mutagenicity on the length of time between irradiation and the onset of S phase. Here, normal fibroblasts (black lines) and XP-A cells (gold lines) are compared. Cells were released from confluence and irradiated at the times indicated (1, 18, or 24 h before S phase). An increased time window reduces mutagenicity in normal cells but not in the NER-deficient XP-A cells. For both cell lines, survival is not significantly affected, possibly because other repair pathways can substitute for NER to some degree. (Adapted from reference 197 .)
Model for replication of eukaryotic DNA containing photoproducts. Replication progresses bidirectionally from the three depicted origins (O in R1 to R3) in the presence of randomly located photoproducts (gold dots). When replicon R1 encounters a photoproduct during lagging-strand DNA synthesis, replication continues with the next Okazaki fragment and a gap is left. Progression of R2 is blocked on both sides of the replication bubble by photoproducts in the template strand for leading-strand synthesis. However, the adjacent replicon R1 is not blocked for leading-strand synthesis; hence, these replicons can fuse. On the other side, however, progression of the neighboring origin R3 is also arrested when the leading strand reaches a photoproduct. This situation results in a long-lived unreplicated region (LLUR), which may be replicated following the activation of an alternative replicational origin (R4). Finally, remaining gaps may be filled by translesion synthesis (gold lines). (Adapted from reference 398 .)
Alkaline sucrose density gradient profiles of DNA from unirradiated or X-irradiated (10 Gy) Chinese hamster ovary (CHO) cells. In the irradiated cells there is a decrease in the relative amount of small DNA molecules (fractions 5 to 10), reflecting regions in which replication was initiated soon after irradiation. However, the relative number of large molecules (fractions 15 to 20), which reflect the progression of replicons that were initiated prior to irradiation, is not significantly decreased. This suggests that growing DNA strands continue to be elongated but new chains cannot be initiated.
Detection of inhibition of replicon initiation in a defined gene (the DHFR cluster of methotrexate-resistant CHO cells) by a two-dimensional agarose gel technique. (A) After restriction digestion, DNA molecules are separated on the basis of their molecular mass in the first dimension (1st-D, x axis). The greater the extent of replication (increasing from 1n to 2n), the slower these fragments migrate in the gel. In the second dimension (2nd-D), separation is carried out under conditions in which the fragment shape contributes to the migration rate. Molecules of a defined region of the genome are then visualized by Southern hybridization. The dashed arrows indicate the putative positions of replication intermediates (note the position of bubble versus branched structures). Thus, the technique allows the distinction of replication intermediates originating within the immediate region of interest (bubble arc [gold line]) from forks that enter from flanking regions (grey line). (B) After exposure of cells to γ-radiation (9 Gy), a transient inhibition of replicon initiation is observed, as evidenced by the loss and reappearance of the bubble arc. The gold area in panel B corresponds to the gold arc in panel A, and the grey area in panel B corresponds to the (idealized) grey arc in panel A. (Adapted from reference 209 .)
Replication tracks visualized by fiber labeling. Cells are consecutively pulse-labeled with two distinguishable halogenated nucleotides (IdU, followed by CIdU). After fixing, denaturing, and immunolabeling of DNA fibers, ongoing forks, newly fired origins, and termination events can be visualized. When cells are treated with a DNA-damaging agent between IdU and CIdU labeling, the absence of new-origin firing and slowing or stalling of ongoing forks can be discriminated and quantitated. (Adapted from reference 279 .)
Detection of pyrimidine dimers in completely replicated UV-irradiated SV40 DNA. After replication of the irradiated BrdU- and [3H]deoxythymidine-prelabeled SV40 genomes, DNA of intermediate density (indicating fully replicated DNA molecules) was first isolated in neutral CsCl gradients and analyzed in alkaline sucrose gradients (shown here) following digestion with the pyrimidine dimer-specific T4 UV endonuclease (T4) (gold lines). The peak at fractions 12 to 13 represents fully replicated covalently closed circular DNA molecules (form I DNA). The increase in the peak of relaxed circles (form II DNA around fraction 19) after digestion of the irradiated sample with the T4 enzyme indicates the presence of UV photoproducts in the replicated molecules. (Adapted from reference 399 .)
Probabilities of mutations of different molecular structure in the supF gene recovered in E. coli following passage of the shuttle vector pZ189 through CV1 monkey kidney cells. The mutational spectra detected with unirradiated (spontaneous mutations) and irradiated (500-J/m2 UV radiation at 254 nm) plasmids are compared. The total area of the pie charts reflects the relative mutation frequencies (3.7 × 10—2 in the unirradiated plasmid versus 7.6 × 10—1 in the irradiated plasmid). (Data from reference 145 .)
Relative probabilities of UV-induced base pair substitutions detected in different target genes of various viral and shuttle vectors and in a chromosomal gene. (Data from references 29 , 84 , 85 , 145 , and 156 .)
Overall spectrum of UV-induced mutations and relative probabilities of the different types of base pair substitutions detected in the supF gene after passage of the shuttle vector pZ189 through SV40-transformed normal, XP-A, XP-D, and XP-V fibroblasts. Abbreviations are as in Fig. 17–10 (Ins/Del, insertions/deletions). In the third line, the respective base substitution frequencies have been converted to reflect the probabilities of misinsertion opposite T or C. (Data from references 40 , 385 , and 451 .)
Relative probabilities of UV-induced base pair substitutions and strand assignment of HPRT mutations recovered from diploid human fibroblasts and Chinese hamster cells. T and NT indicate a base change opposite a potential pyrimidine dimer site in the transcribed strand or nontranscribed strand, respectively. Normal and XP-A human fibroblasts synchronized in the G1 or S phase were treated with UV radiation fluences of 6.5 or 0.5 J/m2, respectively. Unsynchronized normal hamster lung fibroblast cells (V79) and V-H1 cells deficient in NER (the rodent equivalent of human XP-D cells) were UV irradiated with 2 J/m2. Note the predominance of G•C → A•T transitions in human but not in V79 cells, which becomes more pronounced in NER-deficient cells. Note also the strand bias in NER-competent cells and its reversal when NER is compromised (in XP-A cells, V-H1 cells, and normal human cells when irradiated in S). (Data from references 272 , 278 , 447 , and 450 .)
Mutant frequencies of a cll transgene in mouse fibroblasts expressing DNA photolyases. Shown are control cells (vector only), a cell line expressing CPD photolyase (from the rat kangaroo Potorous tridactylis; CPD-PL), and a cell line expressing (6–4)PP photolyase [from Arabidopsis thaliana, without signal sequences for mitochondrial and chloroblast targeting; (6–4)PL]. Mutant frequencies following treatment with or without 500-J/m2 UV-B (UV-B, no UV) and with or without photoreactivation at 360 nm (— PR, +PR) are indicated. Note that a significant reduction in UV-B radiation-induced mutant frequencies following photoreactivation is detected only in cells expressing CPD photolyase. (Adapted from reference 478 .)
Correlation of positions of mutational hot spots in the p53 gene derived from skin tumors with slow repair at individual dipyrimidine positions as measured by ligase-mediated PCR. (A) Triangles indicate the number of mutations recovered from individuals with or without XP as indicated. The regions marked reflect domains that are conserved in p53 proteins from different organisms. (B) Relative repair rates for individual dimer sites on the transcribed and nontranscribed strands are expressed as the amounts of CPD remaining in normal skin fibroblasts after 10 h following UV radiation at a dose of 12 J/m2. Note the overall faster removal of dimers from the transcribed strand but also the considerable site-specific variation. The numbers in panel B correspond to the map positions in panel A. Note also a certain correlation between sites of mutational hot spots and those characterized by relatively slow dimer removal (indicated by gold position numbers). (Panel A adapted from references 86 ; data for panel B from reference 428 .)
Comparison of the structure of budding yeast and human Rev3 proteins. Regions of homology and percent residue identity are depicted. Roman numbers refer to conserved sequence motifs found in B family DNA polymerases. (Adapted from reference 114 .)
Bypass of a cis-syn T<>T CPD by purified S. cerevisiae polymerase £ (Rev3-Rev7) in vitro. (A) Extension of a 15-nucleotide 5’-labeled primer strand by Pol α or Pol ζ is attempted. (B) When products are separated electrophoretically, stalling at the lesion position in the template is evident for both polymerases. The major 34-nucleotide product reflects stalling just before the lesion position. However, in contrast to Pol α, an increasing fraction of lesions is bypassed by Pol ζ with increasing amounts of added polymerase, and a small amount of full-length product (71 nucleotides) is detectable. (Adapted from reference 304 .)
Deoxycytidyltransferase activity of human REV1. In a primer extension assay, insertion activity of purified REV1 is tested with different template bases or an abasic site and added single deoxynucleoside triphosphate (dNTP) (A, C, T, G) or a mix of all four (N). Note that only C is inserted to extend the 17-nucleotide (nt) primer by 1 or 2 bases. (Adapted from reference 488 .)
Bypass of a cis-syn T<>T CPD by S. cerevisiae Rad30 in a primer extension assay. (A) Incorporation preference in the presence of a single deoxynucleoside triphosphate or all four (N) is shown. While some misincorporation occurs, the correct base is preferentially inserted. Opposite template TT or a T<>T CPD, AA is inserted with equal efficiency and preference. (B) In the presence of all four deoxynucleoside triphosphates, bypass is studied with template TT or a T<>T dimer in some distance from the primer-template junction (“running start”). Whereas replicative DNA Pol δ is stalled at the lesion, Pol η (Rad30) inserts deoxynucleotides with equal efficiency opposite the dimer as opposite an undamaged template. However, in contrast to Pol 8 , the template is hardly ever replicated for its entire length. (Adapted from reference 174 .)
Isolation of a protein from HeLa cell nuclear extracts complementing the CPD bypass deficiency of XP-V cell extracts. (A) Fractions from a MonoS column were tested for their activity to alleviate the defect of XP-V cell extracts in replicating a plasmid containing a chemically synthesized cis-syn TT dimer (I, column input). This assay follows a two-step procedure. DNA synthesis is started without added radioactively labeled dCTP. Label is added at a time when arrest at lesions is suspected. Following electrophoretic separation and autoradiography, the presence of labeled closed-circular plasmid molecules (form RF I) indicates completed DNA synthesis and thus lesion bypass. Similar to Fig. 17–9 , the continued presence of lesions is demonstrated by digestion with T4 endonuclease, resulting in nicked circular DNA (form RF II). (B) Compared to the silver-stained protein gel, bypass activity in fractions 26 to 29 correlates with the appearance of a 54-kDa protein band (indicated by arrow; M, molecular mass marker). Partial sequencing of the protein revealed a human homolog of budding yeast Rad30. (In the final analysis, however, this isolated protein has to be considered an active but truncated protein since the predicted molecular mass of full-length human Pol η was later shown to be significantly higher.) (Adapted from reference 262 .)
Pol η focus formation following UV radiation treatment. Time-lapse fluorescence microscopy of the same nucleus allows monitoring of distribution of green fluorescent proteinfused Pol η in a living SV40-transformed human fibroblast. Around 120 to 200 min following UV radiation, the previously diffusely distributed Pol η relocalizes into distinct foci that colocalize with PCNA and sites of BrdU incorporation. (Adapted from reference 181 with permission.)
Two-step bypass of a TT (6–4)PP in vitro. (A) In a primer extension assay (as shown in panel B), it is shown that purified S. cerevisiae or human Pol η is capable of single-base insertion only opposite the 3’ T. However, a combination of Pol η and Pol ζ can catalyze a complete bypass. (B) Pol η inserts mostly G opposite 3’ T of the (6–4)PP but does not extend further from the inserted base. Pol ζ, however, can elongate from an imperfectly matched primer-template junction by preferentially inserting the correct A opposite the 5’ T of the lesion. (Adapted from reference 171 .)
Watson-Crick base pairing determines the base insertion preference of Pol η. (A) Difluorotoluene is used as a nonpolar isosteric analog of thymine. Although difluorotoluene has similar geometry and electrostatic potential distribution to thymine, hydrogen bond formation with adenine is not possible. (B) If tested in an incorporation assay (see Fig. 17–18 ), the Klenow fragment of E. coli DNA Pol I can efficiently insert F opposite a template A and A (and F) opposite a template F. S. cerevisiae Pol η is not capable of using F as an incoming base or as a template base. Steric effects determine the insertion preference of the Klenow enzyme, whereas hydrogen bonding is essential for base pair formation by Pol η. (Adapted from reference 454 .)
Schematic view of one-step and two-step DNA lesion bypass. A bypass polymerase replaces a high-fidelity polymerase that cannot replicate past a lesion in the template. TLS is accomplished in a single step (left) or in two separate steps (right). The latter mechanism involves base insertion opposite a lesion and elongation that is carried out by a separate polymerase specialized in primer extension. Both steps can be error-prone. The low processivity of these translesion polymerases facilitates replacement by a replicative polymerase a short distance downstream of the lesion.
A misincorporation-slippage model can explain spontaneous base pair substitutions that are coupled with frameshifts in budding yeast. Following base misincorporation opposite a damaged G (a spontaneous lesion of unknown nature that is normally subject to NER), extension from the imperfectly matched primer-template is prone to slippage within the adjacent three-T run. Extension of the slipped and misaligned primer leads to insertion of an additional AT base pair and thus to a frameshift next to a G•C → T•A transversion. These events are dependent on Pol ζ in S. cerevisiae. (Adapted from reference 140 .)
Scheme of a serum immunoglobulin G molecule. An immunoglobulin G molecule is composed of two identical pairs of H (heavy) and L (light) chains. Two functional domains can be detected. The V (variable) region binds the antigen, and the C (constant) region (colored) regulates its disposition. Short stretches of highly variable regions within the V domains are the result of somatic hypermutation. These are in contact with the antigen and represent the complementarity-defining regions (CDR). (Adapted from reference 142 .)
Model for somatic hypermutation. Different pathways of introducing mutations are depicted. If uracil created by activation-induced cytidine deaminase (AID) stays unrepaired, replication will introduce a C•G → T•A transition. Uracil can be removed by uracil DNA N-glycosylase (UNG), and the abasic site can be subject to error-prone bypass. The abasic site can also be converted into a single-strand break and can initiate error-free or error-prone repair synthesis. The outcome of DNA synthesis events is influenced by mismatch repair. (Adapted from reference 23 .)
Enzymatic steps of protein degradation by ubiquitin conjugation. In an ATP-consuming process, ubiquitin is first attached to the ubiquitin-activating E1 enzyme (Uba1 in S. cerevisiae). The residue is then transferred to one of several ubiquitin-conjugating E2 enzymes (Rad6 and Cdc34 are examples in S. cerevisiae). The following transfer of ubiquitin to a target protein is mediated by one of many E3 enzymes (ubiquitin ligase, such as Ubr1) that mediate substrate recognition. Multiple ubiquitin residues are added, and the polyubiquitinated substrate is typically degraded by the proteasome, a large multisubunit protease complex. The released ubiquitin monomers can be recycled. (Adapted from reference 59 .)
Defective amino acid-end rule-dependent degradation pathway in rad6 deletion mutants of S. cerevisiae. The steady-state levels of different β-galactosidase (βgal) constructs depend on the N-terminal amino acid X to which ubiquitin is being fused. Compared with Met (left), Arg and Leu are destabilizing residues in the wild-type strain (RAD), but not if RAD6 is completely deleted (rad6A) or deleted of the 9 most N-terminal residues (rad6A1-9), which are highly conserved among Rad6 homologs. (Adapted from reference 459 .)
Crystal structure of the Rad6 ubiquitin-conjugating enzyme. Residues that are required for interactions of Rad6 with Rad18 and Ubr1 are located on the back side of the protein, away from the active-site cysteine (Cys88) that functions as a ubiqui-tin acceptor. (Adapted from reference 467 .)
Subpathways of replicative DNA damage tolerance defined by the proteins of the S. cerevisiae Rad6 group. Rad6 functions as an E2 enzyme in Ubr1-mediated, N-end-rule-dependent protein degradation, independently of DNA damage tolerance pathways. In DNA damage responses, complex formation of Rad6 with the single-stranded DNA-binding Rad18 protein is essential. Both proteins are required for all error-free and error-prone pathways, and cross talk with proteins mediating TLS must be assumed. Proteins of the error-free subpathway are being attracted to sites of stalled replication through the Rad18-Rad5-Ubc13 interaction. Ubc13-Mms2 represents another ubiquitin (Ub)-conjugating machinery that attaches Lys63-linked ubiquitin residues to unknown (?) substrates.
Intrachromosomal gene conversion as the consequence of replicative DNA damage tolerance by template switching. Two defective copies of hygromycin resistance genes (HYG) with mutations in different locations (black boxes) have been placed on the same chromosome. Assuming DNA damage (^) in HYG 2 , leading-strand replication may temporarily switch to the HYG 1 copy and thus transfer wild-type sequence information (dashed line). The next round of replication yields a chromosome copy with a wild-type (w.t.) hygromycin resistance gene. (Adapted from reference 241 .)
Inability to ubiquitinate PCNA results in a defect in DNA damage-induced mutagenesis in S. cerevisiae.(A) PCNA structure. Lys164 and Lys127 are sumoylated in S phase; Lys164 is monoubiquitinated by the action of Rad6-Rad18 in response to DNA damage. (B) The induction of canavanine-resistant (Canr) forward mutants is depicted as a function of UV dose. Mutants carrying PCNA versions that cannot be ubiquitinated (K164R) or that can be neither ubiquitinated nor sumoylated (K127l164R) are as defective in mutagenesis as are rev3 mutants. However, a SIZ1 deletion which abolishes SUMO ligase activity has no effect on induced mutagenesis. WT, wild type. (Panel A courtesy of J. Kuriyan, panel B adapted from reference 402 .)
Ubiquitination of PCNA mediates the choice of DNA damage tolerance pathway in S. cerevisiae. This modification is essential for TLS. Human Pol η interacts physically and functionally with Rad18 and with monoubiquitinated PCNA. Recruitment of Ubc13-Mms2 can result in further addition of ubiquitin (Ub) residues linked through Lys63. Polyubiquitinated PCNA appears to be important for error-free damage tolerance, possibly establishing template switching and a copy-choice mode of replication. (Adapted from reference 153 .)
Human Pol η interacts with monoubiquitinated PCNA following UV radiation. (A) Human SV40-transformed fibroblasts expressing polyhistidine-labeled ubiquitin (His-Ub) are UV treated following transfection. His-tagged ubiquitin is collected through affinity to Ni-agarose beads. (B) Following cross-linking, copurifying Pol η and PCNA species are detected by immunoblotting. In response to UV irradiation, PCNA with attached His-tagged ubiquitin (larger than PCNA and slightly larger than PCNA-ubiquitin) is associated with Pol η. (Adapted from reference 183 .)