Recombinational Repair (RR) of DNA

DNA double-strand breaks (DSB) are potentially harmful damages that can arise either spontaneously or from exposure to DNA-damaging agents such as ionizing radiation. Estimates have predicted that ~1 in every 6,000,000 cells suffers from a spontaneous DSB in an hour. However, this number is based solely on the frequency of visible, gross chromosomal rearrangements, and does not take into account small deletions or insertions or repair, and is thus likely an underestimation of the actual DSB events. Studies using ionizing radiation or restriction enzymes to generate DSBs in vivo found that these damages are cytotoxic and mutagenic. In fact, it would appear that the cytotoxicity of ionizing radiation is directly related to the production of DNA DSBs, as experiments have shown that a single DSB event corresponds to a single lethal hit. Two major forms of RR exist to cope with DSBs: (1) homologous recombination, which occurs in S and G2 phase cells and relies on extensive nucleotide sequence complementarity between the intact chromatid and the damaged chromatid as the basis for strand exchange and repair, and (2) non-homologous (i.e. "illegitimate"), which requires little or no sequence homology and is mediated by direct end-joining.

Homologous recombinational repair (HRR)

In the yeast Saccharomyces cerevisiae the major pathway for repairing DSBs involves a set of genes, known as the Rad52 epistasis group, that consists of RAD50-52, RAD54, RAD55, RAD57, RAD59, XRS2, and MRE11/RAD58. The Rad51 protein, which is a functional homolog of RecA in E. coli, acts a key player to mediate homologous pairing and transfer of strands between two DNA duplexes. The possibility that mammalian cells also possess this pathway was suggested as early as 1985 by the finding that cultured cells perform recombination in gene targeting experiments and more recently by the fact that homologs of the yeast HRR genes could be readily identified in human and rodent cells. In addition, hamster cell mutants that are complemented by the human XRCC2 and XRCC3 genes, which are structurally related to RAD51, have properties consistent with defective recombination. In particular, these mutants show chromosomal instability in terms of visible breaks and exchanges at metaphase, IR sensitivity consisting of extreme sensitivity in S phase, and extreme sensitivity to DNA inter-strand crosslinking agents such as mitomycin C or cisplatin. The XRCC2 and XRCC3 genes are members of a recently identified Rad51-family that includes RAD51B, RAD51C, and RAD51D, all of which show low levels of sequence similarity to both human and yeast RAD51. The biochemical functions of the proteins encoded by these RAD51-like genes are presently unknown, but they may act as accessory factors for mammalian Rad51 protein, akin the role of Rad55 and Rad57, which enhance yeast Rad51’s transferase activity. The quantitative contribution of HRR in cellular response to IR is currently under investigation.

In mice a knockout mutation in MmRAD51 resulted in early embryonic lethality, indicating a function essential for cell viability in mammals. Similar observations have been made in chicken cells having conditional RAD51 expression. In contrast, a knockout mutation in the MmRAD54 gene was compatible with normal development and had a phenotype of IR sensitivity and modest sensitivity to mitomycin C. The repair of crosslinks by HRR is of particular interest because they may have an absolute requirement for this pathway. Unless crosslinks are converted to DSBs as an intermediate (as they appear to be in yeast), they would not likely be removed by another mode of repair. Moreover, knockout of the MsRAD52 resulted in discernible phenotype with respect to DNA damage sensitivity. Thus, it appears that quite different phenotypes are associated with the different genes in the HRR pathway. So far, no mouse knockout mutants have been reported for the Rad51-family members.

Other mammalian genes that likely participate in HRR are RAD50, MRE11, and RPA1, RPA2, and RPA3, as well as those encoding other nucleases, helicases, and polymerases.

Nonhomologous recombinational repair

In humans, the predominant mechanism of recombinational repair is non-homologous (illegitimate) DNA end-joining, where the entire reaction involves the following steps: recognition, recruitment and physical alignment, processing of the termini, gap filling and ligation.

Biochemical studies and analysis of radiation-sensitive cell lines have identified Ku, a heterodimer consisting of a 70 kDa and 86 kDa subunit, as the major DNA end-binding activity in eukaryotes. Ku has also been shown to promote circularization or joining of the ends of linearized duplex DNA. In binding to the ends of DNA, Ku serves as a barrier against degradation, as a scaffolding protein, and as a recruiting factor for other repair players. One such player, DNA-PKcs (catalytic subunit of DNA-dependent protein kinase), initiates a repair response by phosphorylating downstream target proteins, many of which are unknown. The DNA-PK complex is also essential for processing of DSBs that normally arise during immunoglobulin and T-cell receptor gene rearrangement.

The Xrcc4 protein is an essential component of the Ku-depending rejoining pathway and serves as a cofactor for DNA ligase IV (LIG4), which appears to be complexed with Xrcc4. LIG4 is required for V(D)J recombination and the repair of DSBs induced by ionizing radiation, and the other three known DNA ligases do not seem to substitute in this reaction.

Mre11/Rad50/NBS complex, which includes the protein encoded by the NBS gene that is defective in Nijmegen Breakage Syndrome, is thought to participate in the recognition and processing of DSBs, but its mechanism of action is not yet understood. This complex likely also has a role in HRR, based on the properties of the homologous proteins in yeast. Human Mre11/Rad50 complex possesses 3' to 5' nuclease activity, which may represent an early step in DSB repair.

Most in vitro studies to date have used DNA substrates that contain a restriction enzyme-generated DSB, i.e. a break harboring normal 3'-hydroxyl (OH) termini. However, in vivo 3'-termini are likely composed of 3'-phosphate and 3'-phosphoglycolate residues, common products of free radical attack. How these damages are repaired is poorly understood. One piece of this puzzle may be Ape1, the major human apurinic endonuclease, which possesses 3'-repair activity. Yet, Ape1 has a limited substrate range for 3'-damages in vitro, suggesting that alternative 3'-repair factors exist. In fact, fractionation of human cell extracts has found multiple 3'-repair activities, but their identities remain unknown. An area of future interest will be how these lethal and potentially mutagenic 3'-oxidative damages are dealt with.