This information was most recently updated October 31, 2006.
All organisms need to deal with the problems that arise when a moving replication fork encounters damage in the template strand. Obviously the best way to deal with this situation is to repair the damage by one of the excision mechanisms discussed above. In some cases, however, the damage may not be repairable, or the advancing replication fork may already have unwound the parental strands, thus preventing excision mechanisms from using the complementary strand as template for repair, or excision repair may not yet have had an opportunity to repair the damage.
There are two reasons why it is important for the cell to be able to move replication forks past unrepaired damage. First, long-term blockage of replication forks leads to cell death. Second, replication of damaged DNA provides a sister chromatid that can be used as template for subsequent repair by homologous recombination.
Notice that replication fork bypass mechanisms cannot, strictly speaking, be considered examples of DNA repair, because the damage is left in the DNA, at least temporarily. Nevertheless, experiments in yeast (see below) demonstrate that damage bypass is an important component of the overall cellular response to DNA damage. It contributes to cellular survival of radiation damage to roughly the same extent as the pathways for nucleotide excision repair and repair by homologous recombination.
Eukaryotic mechanisms for replication fork bypass of damaged sites in DNA are not as well understood as the pathways I have discussed previously, but considerable progress has been made in recent years. Furthermore, now that the excision repair pathways have been largely worked out, many scientists interested in cellular responses to DNA damage are turning their attention to replication bypass.
Eukaryotic damage bypass mechanisms have been studied most extensively in the budding yeast, Saccharomyces cerevisiae. Characterization of mutants sensitive to ultraviolet or ionizing radiation has identified three "epistasis" groups (groups of genes operating in the same pathway) that are important for cell survival. The major genes in these three groups are listed here:
Each of the three epistasis groups is named for one of the key genes within it. The first two groups contain genes in pathways we've already discussed. The RAD3 group corresponds to nucleotide excision repair, while the RAD52 group corresponds to repair (principally of ds breaks) by homologous recombination. It is now clear that the RAD6 group enables damage bypass.
Although the RAD6 epistasis group was first described in S. cerevisiae, homologs of the RAD6 proteins have been found in all eukaryotic organisms, indicating that the pathways are of general importance.
Cloning and sequence analysis of the genes in the RAD6 group, combined with in vivo and biochemical studies of the interactions between the gene products, have led to the surprising conclusion that ubiquitylation plays an essential role in regulating the RAD6 pathways:
The Rad18 and Rad6 proteins are essential for all aspects of damage bypass. Mutations in these genes have profound effects on cellular survival after radiation damage. Sequence comparison with other proteins and biochemical studies have revealed that Rad18 is an ATPase capable of binding ss DNA. In addition, Rad18 contains a "ring finger" domain that helps it to interact with Rad6. Rad6 is a ubiquitin conjugating enzyme—an enzyme capable of transferring ubiquitin from a ubiquitin-activating enzyme to a protein substrate. The fact that mutations in Rad6 that destroy its ubiquitylation activity also knock out the Rad6 replication bypass pathway make it clear that the ubiquitylation function of Rad6 is essential for this pathway. One, perhaps the only essential (in this pathway) target of Rad6 is PCNA, which is mono-ubiquitinated by Rad6 on its lysine 164.
The SRS2 gene encodes a helicase. This helicase promotes replication bypass by the Rad6 pathway in two ways. First, it blocks homologous recombination at the lesions where it binds. In the absence of the Srs2 helicase, the cell might attempt to directly repair these lesions by homologous recombination, rather than letting replication forks pass through them first. Can you think of a reason why replicating lesions before repairing them (by homologous recombination or other pathways) might be advantageous to the cell? Second, it actively stimulates both replication bypass subpathways (bypass synthesis and recombinational bypass).
Like the Rad18 protein, the Rad5 protein is a ring-finger-containing ATPase capable of binding ss DNA. In addition, its sequence contains helicase motifs. Also like Rad18, it uses its ring finger to help it bind to a ubiquitin-conjugating protein, Ubc13. Ubc13 forms a heterodimer with Mms2, which is structurally similar to ubiquitin-conjugating enzymes but does not by itself possess ubiquitin-conjugating activity. Ubc13 has somewhat different specificity than Rad6. Ubc13's preferred substrate is already-formed ubiquitin chains, and it adds new ubiquitin molecules onto lysines at position 63 of old ubiquitin molecules. Thus it can add additional ubiquitins via K63 linkages onto the single ubiquitin conjugated to PCNA by Rad6.
The Rad5, Ubc13 and Mms2 proteins are all important for damage bypass by recombination, but they have no effect on damage bypass synthesis. Thus it seems likely that the initial monoubiquitylation of PCNA by Rad6 assists PCNA in interacting with bypass polymerases for translesion synthesis (see below), while polyubiquitylation by Ubc13 modifies PCNA to optimize it for recombinational bypass (see immediately below). Future studies will tell us about the precise effects of ubiquitin modifications of PCNA.
The mechanism of eukaryotic recombinational bypass is not yet clear. Two possibilities are shown here. In both of these examples, synthesis of one red nascent strand (the leading strand in these examples) is blocked by DNA damage (closed circles). Synthesis of the other red nascent strand (the lagging strand in these examples) continues for a distance beyond the damaged region. This is important, because this complementary nascent strand will provide a template for recombinational bypass synthesis through the damaged region. In the left-hand example, the blocked leading strand polymerase is imagined to switch to use of the nascent sister strand as template. Sufficient leading strand DNA is synthesized using this sister strand template to bypass the damage. Then the leading strand is unwound from its sister strand template, and the polymerase switches back to using the top parental strand as template.
In the right-hand example, when leading strand synthesis is blocked by damage, it is imagined that branch-migration-promoting enzymes rewind the replication fork past the site of damage. This would cause separation of the nascent strands from their parental strand templates, permitting the nascent strands to anneal with each other. The 3' end of the leading nascent strand could then be extended by a polymerase (dashed red line in figure), using the longer lagging nascent strand as template. Reversal of the branch migration process would then restore the replication fork, but now the leading nascent strand would continue past the site of damage.
In both these examples, the DNA synthesis when using the sister strand as template is conservative: none of the nascent lagging strand or template strand nucleotides are incorporated into the nascent leading strand. Both models should also be error-free, consistent with yeast genetic data indicating that bypass by the Rad5-, Ubc13-, Mms2-dependent mechanism is non-mutagenic. Note also that each template-strand switch will require opening and closing of the PCNA clamp. Such opening and closing is likely to be stimulated by appropriate polyubiquitylation.
The major alternative bypass pathway, translesion synthesis, can be either non-mutagenic or mutagenic, depending on the type of damage and on the repertoire of translesion polymerases available to the cell. Some of these possibilities are summarized in the diagram below, which shows the major translesion synthesis pathways in budding yeast:
In this diagram, the blue lines represent the two strands of a DNA molecule. The arrowhead on each strand represents the 3' end of the molecule (the end onto which polymerases can add new nucleotides). In the top portion of the diagram, the normal DNA polymerases and accessory proteins (represented by a yellow oval) have stalled because the polymerases have encountered a thymine dimer in the template strand. Eukaryotic cells possess several alternative polymerases or polymerase complexes that are capable of extending the nascent DNA strand past the thymine dimer. Two of these are shown in the above diagram. Many more have been discovered in the past few years (see below). In yeast, polymerase zeta (green) consists of two subunits, Rev3 and Rev7. Polymerase zeta can replicate past a variety of lesions. For some types of lesions, it makes few mistakes. However, available evidence (see below) suggests that it frequently makes mistakes in replicating past thymine dimers. In vivo function of polymerase zeta also requires the Rev1 protein (orange), which is another type of DNA polymerase. In vitro, Rev1 can insert single C residues in situations where the template contains a G and also where the template lacks a base (an AP site). One current hypothesis is that, in vivo, Rev1 inserts the first nucleotide to initiate translesion synthesis, and synthesis is then continued for a few more nucleotides (sufficient to bypass the lesion) by pol zeta.
The alternative bypass polymerase is called polymerase eta (light blue). It appears to function as a single polypeptide chain, which is encoded by the RAD30 gene in yeast. A mammalian version of polymerase eta is encoded by the XPV gene (now called the POLH gene). XPV stands for Xeroderma pigmentosum "variant." Patients bearing XPV mutations display the same symptoms as patients with mutations in the XPA-XPG genes, but their ability to perform NER is normal. Now it is clear that the problem with XPV patients is that they are unable to accurately replicate past thymine dimers. The Rad30/Xpv/PolH protein (pol eta) has been demonstrated in vitro to be capable of error-free synthesis past a thymine dimer (that is, it inserts two A residues). In XPV patients lacking pol eta, alternative translesion synthesis pathways—such as that of pol zeta—make mistakes when replicating past thymine dimers. Consequently, although the mechanisms differ, XPV and XPA-XPG patients all suffer from the same final problem—a significant increase in mutations in response to UV light.
Neither pol eta nor pol zeta binds tightly to its DNA substrate in vitro; both enzymes dissociate from the DNA after inserting a single nucleotide. This makes it easy for the normal polymerases and accessory proteins (yellow in diagram) to displace pol eta or zeta after it has extended the nascent chain past the damaged region. At that point, normal DNA synthesis can resume. It seems likely that switching back to the normal replicative polymerases will prove to be signalled by de-ubiquitylation of PCNA.
It turns out that other eukaryotic cells possess several additional proteins that resemble Rad30 or Rev3. Some of these bypass polymerases are capable of replicating accurately past certain types of DNA damage. Some of these polymerases were independently discovered in different laboratories and assigned different names. Recently, the relevant laboratories agreed on consistent nomenclature for all currently known eukaryotic DNA polmerases. This information is summarized in the table below.
Notice that polymerases involved in bypass synthesis make up an appreciable fraction of all known eukaryotic DNA polymerases. Notice that Rev1 does not have a Greek name. That's because it's not considered a true DNA polymerase, since it can insert only one nucleotide at a time, usually in a template-independent fashion. Polymerases iota and kappa are related to pol eta but have somewhat different bypass specificities and are not usually error-free. Pol kappa is found in most other eukaryotic organisms, including the fission yeast, Schizosaccharomycs pombe, but is not found in S. cerevisiae. Pol iota, pol theta and pol nu appear to be found only in animals. Pol theta is especially competent at bypassing AP sites, where it inserts an A residue. Since most AP sites are consequences of purine loss, and A is a purine base, repair by pol theta is frequently error-free.
This concludes the 2006 lectures on DNA repair in RPN530. Please be sure to write to me if you have questions, and please return to the DNA repair home page for other study aids.