Full text loading...
Chapter 19 : Managing DNA Strand Breaks 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 eukaryotic repair and tolerance responses to DNA strand breaks, and addresses pathways that do not rely on extensive sequence homology. It also considers topics that defy classification since they appear to be relevant to a variety of different pathways. Having discussed the major pathways of homologous recombination (HR), and nonhomologous end joining (NHEJ), the chapter addresses a complex of proteins that functions in surprisingly diverse areas of repair and cellular regulation associated with double-strand breaks (DSB). This complex is referred to as the MRN (for “Mre11-Rad50-NBS1/Xrs2”) complex. Following the identification of human homologs, additional insights came from the discovery that one of the components (NBS1 = Xrs2 in yeast) is responsible for an inherited human DNA repair disorder, the Nijmegen breakage syndrome (NBS). This discussion also visits several important areas of investigation that have helped characterize the various roles of the proteins involved. The recent structural findings that have provided a basis for models of molecular function are presented. Apart from higher-order eukaryotic chromosome structures, the nucleosome core consisting of DNA wrapped around a histone octamere, the building block of chromatin, can limit accessibility and influence cellular strategies of DNA damage recognition. The histone modifications such as phosphorylation seem to play a more direct role in DSB repair. The chapter describes some findings that address factors that dictate the principal “decision” between HR and NHEJ. Lastly, it talks about the repair of single-strand breaks (SSB).
Hypothetical mechanism for joining of 5’ and 3’ single-stranded ends of broken DNA molecules. The ends are held together by an alignment protein (oval in the background), allowing continuation of fill-in DNA synthesis (triangles) primed at the recessed 3’ end in order to bridge the unligated overhangs. Establishing the first base pair at the 3’ overhang creates a substrate for DNA ligase (diamond). The fill-in synthesis is continued, and a second ligation completes the joining. (Adapted from reference 400 .)
Elements of V(D)J recombination. (A) The assembly of a variable region gene (see Fig. 17–26) by joining of segments from a large repertory of germ line V, D, and J segments is depicted. (B) The coding segments are flanked by recombination signal sequences (represented by arrowheads in A to C). These consist of conserved heptamer and nonamer motifs, separated by 12- or 23-bp spacers of variable sequence. (C) Joint formation results in a deletion or inversion of the intervening region. Joining of the noncoding regions originates from a heptamer-heptamer fusion, creating the signal joint. As outlined in more detail in Fig. 19–3 , a usually imprecise fusion of the ends of coding sequences results in formation of the coding joint. (Adapted from references 30 and 359 .)
Coding-joint and signal joint formation during V(D)J recombination. Following synapsis of coding and signal ends, RAG1-RAG2 recombinase introduces nicks. In a transesterification reaction, the created 3’-OH ends attack the complementary strand, seal the coding ends by creating hairpin structures, and release the blunt signal ends. The latter are normally ligated without loss of nucleotides to form the signal joint. The hairpins at the coding ends are opened by a structure-specific endonuclease. Since these are frequently not opened exactly at the apex, the following processing may result in small inverted duplications (named P sequences). Other processing steps such as trimming, fill-in DNA synthesis, and extension by TdT (creating N regions) contribute to the variability of coding joints. (Adapted from reference 146 .)
Detection of a DNA end-binding activity that is antigenically related or identical to the 70-kDa autoantigen Ku subunit. With labeled linearized DNA as a probe, protein-DNA complexes are detected as retarded bands in gel shift assays. An excess of unlabeled linearized DNA but not of uncut circular DNA can successfully compete for this DNA end-binding activity (compare lanes with cut and uncut competitor DNA). Addition of antiserum raised against the 70-kDa subunit of (human) autoantigen Ku results in a band of even lower mobility at the expense of the original retarded band. This “supershift” indicates binding of the complex by the antibody. In contrast to the wild type (AA8), the DNA end-binding activity is absent in an X-ray-sensitive mutant CHO cell line (XR-V15B, a mutant of the XRCC5 group). A double band is found if extracts of certain normal human cells (such as the IMR-90 line [right panel]) are probed. (Adapted from reference 334 ; courtesy of G. Chu.)
DNA end binding by Ku70/80. A crystal structure of the Ku70/80 dimer bound to dsDNA shows the protein embracing the DNA via an intertwined arrangement of subunits ( 424 ). Both protein subunits make nonspecific contacts with the backbone of the DNA. The Ku70 subunit (gold) lies close to the blunt end of the dsDNA. Panels A and B show different representations of the same structure.
Telomere capping in lower and higher eukaryotes. The DNA structures of telomeres and their binding proteins are depicted. Commonly, ssDNA-binding proteins (Cdc13 and Pot1) and capping proteins (Ten1-Stn1, TRF2, and its ortholog Taz1) are found. (Adapted from reference 91 .)
DNA-dependent protein kinase assay used for purification of the catalytic subunit of DNA-PK cs from HeLa cell extracts. (A) DNA-PK activity in crude extracts is detected by using unphosphorylated 35S-labeled transcription factor SP1 as a substrate. During 0 to 10 min of incubation, the presence of linear plasmid DNA results in the conversion of SP1 into a phosphorylated form (SP1•P) of lower electrophoretic mobility. The addition of supercoiled DNA is ineffective. (B) The assay was repeated in the presence of linearized DNA after extract fractionation on ssDNA cellulose columns. Simultaneous addition of two fractions (eluted with 0.2 and 0.4 M KCl) is required for activity. Purified Ku complex is not active alone but can replace the factor(s) present in the 0.4 M eluate. The 469-kDa catalytic subunit of DNA-PK was later purified from the 0.2 M fraction. (Adapted from reference 143 .)
Models for DNA-PK interaction with DNA ends. (A) Structure models of DNA-PKcs protein suggest openings that can accommodate ssDNA or dsDNA. Regulation of kinase activity by ssDNA binding may be important during the process of end trimming. (B) The DNA end-binding Ku heterodimer may move away from the end to facilitate binding of and interaction with the catalytic subunit, DNA-PKcs. (Adapted from references 229 [panel A] and 157 [panel B].)
Hairpin-opening activity of the Artemis-DNA-PKcs complex. A hairpin-forming ssDNA probe is 5’-end labeled at a 1-nucleotide overhang (*). Incubation with Artemis alone results in exonucleolytic degradation, releasing labeled DNA fragments of two nucleotides. In the presence of DNA-PKcs, hairpin-opening activity is observed. Addition of a DNA-PKcs inhibitor can reduce this stimulation. The site of cleavage varies. Most prominently, a 23-nucleotide product is released that represents a product in which 2 nucleotides remain annealed at the hairpin tip. The presence of Ku does not alter the spectrum of Artemis activities (M = size marker of 21 nucleotides). (Adapted from reference 255 .)
Chromatin immunoprecipitation to study the binding of DNA ligase IV and Lif1 to DSB in S. cerevisiae. A general scheme of the technique is given in Fig. 18–6. Following DSB introduction by HO endonuclease expression, cross-linked protein-DNA complexes are precipitated with anti-Lig4 or anti-Lif1 antibodies. PCR is used to detect coprecipitated DNA from the DSB vicinity, indicating binding of Lig4 or Lif1. A positive result is strictly dependent on DSB introduction (+HO). Lig4 binding does not occur in a LIF1 -deleted mutant (upper panel, —). On the other hand, Lif1 binding does not depend on Lig4 (lower panel). Lif1 binding, however, does depend on functional S. cerevisiae Ku70 (Yku70). (Adapted from reference 398 .)
Structure of XRCC4. (A) XRCC4 dimerizes through a long coiled coil, leaving the helix-turn-helix (HTH) motif of each subunit exposed for interactions with DNA. The crystal structure of an N-terminal fragment of XRCC4 suggests how the dimer might form a bridge between two dsDNA molecules ( 196 ). (B) At high concentrations, two dimers of XRCC4 associate through their coiled-coil domains to form a tetramer (a second XRCC4 dimer is shown as a grey surface representation). The formation of tetramers is at odds with binding of XRCC4 to DNA ligase IV and therefore may serve another function ( 284 ).
Scheme for coding-joint and signal joint formation during V(D)J recombination by NHEJ proteins in vertebrates. Following cleavage by the RAG1-RAG2 recombinase, the usually blunt-ended signal ends are joined through the action of Ku70-Ku80 and XRCC4-DNA ligase IV. The coding ends require additional activities provided by DNA-PKcs and also need to be processed by Artemis. TdT may extend free ends and contribute to the variability of coding-joint formation. (Adapted from reference 20 .)
The RAD50 gene is required for the induction and processing of meiotic DSB in S. cerevisiae. (A) Two DSB sites in the region of the LEU2 gene during meiosis I can be physically mapped by Southern analysis with the depicted probe. (B) Concomitant with a commitment to recombination, fragments corresponding to these break sites can be detected in DNA prepared at different times after induction of synchronous meiosis. The signals are diffuse and transient in the wild type (RAD/RAD), suggesting some kind of degradative processing. No signals are detected in a rad50 deletion mutant (rad50Δ/rad50Δ). However, certain point mutants of RAD50 (rad50S/rad50S) show distinct and persistent bands, suggesting the accumulation of unprocessed DSB. (Adapted from reference 51 .)
Linear map and domain structure of MRN complex components Rad50, Mre11, NBS1, and Xrs2. Selected structural features are referred to throughout the chapter. Rad50 contains N-and C-terminal Walker A- and B-type nucleotide-binding motifs that are separated by coiled-coil regions with a flexible hinge in between. The CysXXCys motif within this region is important for homomeric interaction (“zinc hook”). The Mre11 sequence reveals DNA-binding domains and a nuclease domain, containing four conserved phosphoesterase motifs (M1 to M4). The NBS1 map shows phosphopeptide-binding FHA and BRCT domains, and the region of mutations isolated from NBS patient cells is depicted. Important phosphorylated serine residues are also indicated. The C-terminal Mre11-binding domain is marked (MBD). (Adapted from reference 83 .)
Colocalization of Mre11 foci with areas of DSB. (A) Subnuclear X irradiation of human fibroblasts through a mask creates a stripe-like pattern of exposure when DSB are visualized through labeling with TdT incorporating bromodeoxyuridine. (B) Detection of Mre11 foci by immunofluorescence in the same nucleus. (Adapted from reference 299 with permission.)
Schematic view of repair protein colocalization in foci following ionizing radiation. Localization of MRN complex and Rad51 in foci at DSB sites appears to be mutually exclusive. BRCA1 colocalizes to both; in the case of Rad51, the interaction is most probably mediated by BRCA2.
Potential role of Mre11 in microhomology-dependent joining of incompatible DNA ends. It is suggested that the Mre11 dimer binds both ends simultaneously and 3’ → 5’ exonuclease activity of Mre11 results in DNA degradation until sequence homology in the undigested strands is uncovered. (Adapted from references 181 and 310 .)
Stimulation of intermolecular DNA end joining by the MRN complex. A linearized plasmid with cohesive ends is incubated with the purified S. cerevisiae proteins DNA ligase IV (Dnl4), Lif1, and the MRN complex. Increasing concentrations of the MRN complex lead to increasing amounts of intermolecular ligation products at the expense of intramolecular recircularization. (Adapted from reference 67 .)
Interaction of the budding-yeast MRN complex with linear duplex DNA. (Top) DNA fragments (400 bp) with 5’ cohesive ends visualized by atomic force microscopy; no protein added. (Bottom) Addition of purified MRN components (= Rad50, Mre11, and Xrs2) results in nucleoprotein complexes that align and may bridge several DNA fragments. Some MRN complexes are marked by arrows. (Adapted from reference 67 with permission.)
Antagonism between Ku complex and exonucleases in S. cerevisiae. Accelerated single-strand resection around a DSB is found in the absence of Ku. In many respects, the MRN complex and Exo1 function as competing exonucleases, with Exo1 stimulating homologous recombination.
Increased mitotic recombination between heteroalleles as a consequence of impaired single-strand resection in diploid S. cerevisiae MRN mutants. Open circles indicate inactivating point mutations at different locations within the same gene, which, for example, may be needed for amino acid synthesis. Diminishing the length of ssDNA that invades the duplex will limit the length of generated heteroduplex DNA and thus the probability of coconversion by mismatch repair (compare left and right). Conversion of one marker alone leads to a prototrophic cell. (Adapted from reference 321 .)
The Rad50 ATPase. (A) The N- and C-terminal domains of Rad50 form two halves of an ATPase domain that are separated by a 700-residue coiled coil (see Fig. 19–14 and Fig. 19–22 ) ( 179 ). Crystal structures of ATP-bound and ATP-free forms of the dimeric Rad50 ATPase reveal conformational changes that may be linked to ATP-regulated dimerization and DNA-binding activities. (B) Two prominent positively charged grooves are present on one surface of the Rad50 ATPase dimer, suggestive of a surface that could interact with two DNA molecules.
The zinc hook of Rad50. The coiled-coil domain of Rad50 is interrupted by a “knuckle joint,” termed the zinc hook, that facilitates the packing of two antiparallel helices in the coiled coil and mediates interactions between two Rad50 molecules ( 177 ) (see Fig. 19–22 ). Four cysteines from two Rad50 molecules coordinate a zinc ion, creating a compact but stable dimerization interface.
Zink hook-mediated interaction between MRN complexes on different broken DNA molecules may result in tethering and aligning of DNA ends. (Adapted from reference 90 .)
γ-H2AX foci detected by immunofluorescence in mammalian nuclei following γ-irradiation treatment. (Courtesy of E. Rogakou.)
Cell cycle stage-dependent IR sensitivity of chicken DT40 B-cells. G1 cells were isolated by elutriation. During the following synchronous cell cycle passage, survival after 2-Gy γ-irradiation was determined at the times indicated. KU70-deleted cells exhibit the highest relative IR sensitivity if irradiated in G1 or early S phase. This is in contrast to cells defective in homologous recombination (RAD54 deleted), which are more sensitive in late S/G2. WT, wild type. (Adapted from reference 392 .)
XRCC1 interacts with DNA and polymerase p. The structure of XRCC1 was determined by solution nuclear magnetic resonance spectroscopy methods ( 263 ), and residues that interact with a 13-mer DNA (left panel) or with DNA polymerase β (right panel) were identified by shifts in the nuclear magnetic resonances of XRCC1 residues caused by formation of either complex (interacting residues are shown in gold).