The Fix-It Crew II

A DNA lesion is any change to a DNA molecule that disrupts its normal chemical structure. This excludes DNA polymerase so-called "errors." In the case of errors, while there is a non-complementary pair of nucleotides in the double helix, the chemical structure of the nucleotide molecules remains intact.

In this post, we'll discuss the the DNA repair pathways that correct the vast majority of DNA lesions. These include ribonucleotide excision repair (RER), base excision repair (BER), and nucleotide excision repair (NER). We'll conclude with translesion synthesis (TLS), which isn't a repair pathway per se, but rather a move that cells make that allows them to bypass a lesion and postpone its repair. Let's start with RER.

Ribonucleotide Excision Repair (RER)

We've covered errors: cases where a non-complementary deoxyribonucleotide is mistakenly added by a DNA polymerase. But polymerases make another kind of mistake much more frequently. They insert a ribonucleotide instead of a deoxyribonucleotide. Ribonucleotides are the monomers that make up RNA. So, a polymerase might insert an RNA version of an "A" nucleotide instead of a DNA version of an "A" across from a DNA "T" on the template strand. Ribonucleotides can't remain in DNA. They trigger strand breaks.

Ribonucleotide mis-incorporations occur frequently for two reasons: (1) the polymerase's ability to discriminate ribonucleotides from deoxyribonucleotides isn't perfect; the two molecules differ by only one small chemical attachment, and (2) ribonucleotides are present in the nucleoplasm at 10-100 times higher concentrations than deoxyribonucleotides. Ribonucleotides are the raw materials of mRNAs, which the cell manufactures constantly to build the proteins it needs. So ribonucleotides are everywhere! Since they're everywhere, they get mis-incorporated more often than non-complementary deoxyribonucleotides.

How frequently are ribonucleotides misincorporated? I said earlier that the deoxyribonucleotide mismatch error rate after polymerase proofreading (but before MMR) was one error every 10 million to 100 million nucleotides. In sharp contrast, the ribonucleotide insertion rate is once every 1,000 to 5,000 nucleotides! That's an enormous difference. The cell is constantly correcting mis-incorporated ribonucleotides!

Like MMR, RER is an example of a classic "cut and patch" repair system. In such a system, first a lesion is detected. Then a single strand incision is made near the lesion. Then the DNA near the removed and replaced using information from the "good" complementary strand. Finally, the two pieces of now-repaired single-stranded DNA are reconnected to restore the continuity of the double helix.

RER involves many of the same proteins that take part in the removal of the RNA portions of the primers used to make Okazaki fragments in lagging strand synthesis.Given that many of these proteins are replisome-linked, it's no surprise that RER, like MMR, occurs immediately after DNA synthesis just behind the replisome.

RER starts with the protein RNase H2. In RER, RNase H2 has two jobs. It uses its detector activity to identify ribonucleotides that have been incorporated into the growing chain. It also has an endonuclease activity which it uses to make an incision immediately 5' of the ribonucleotide it identified.

Recall that RNase H2 is also the enzyme that made endonucleolytic cuts to remove the first 10 or so ribonucleotides at the 5' end of Okazaki fragment primers. So in RER, RNase H2 perfoms the first step in "cut-and-patch" repair (identifying the lesion) as well as the second step (making a single strand incision near the lesion).

Once the incision is made, the PCNA sliding clamp recruits DNA polymerase delta to the nick. In a classic "cut-and-patch"repair process, the next step would be to chew up the strand with the mis-incorporated ribonucleotide. But RER is more like Okazaki fragment processing: DNA polymerase delta synthesizes off the DNA 3' end of the incision, using the complementary "good" strand as the template. As it synthesizes 5' to 3', it displaces the mis-incorporated ribonucleotide and several deoxyribonucleotides that follow, creating a flap.

The flap will be 1-10 nucleotides long. It will be cleaved off by FEN1 (Flap Endonuclease I), the same enzyme that performed that task in Okazaki fragment processing. Once FEN1 cleaves off the flap, only one more job remains: that of connecting, or ligating, the DNA. That falls to the enzyme that joined abutting Okazaki fragments: DNA ligase I. Once strand continuity is restored, the repair is complete.

As I mentioned, RER is tightly coupled to DNA replication. All the enzymes involved in RER tether to the back face of the PCNA sliding clamp including RNase H2, DNA polymerase delta, FEN1, and DNA ligase I. Because it is so closely linked with DNA replication, and because ribonucleotide mis-incorporations are so common, RER can be thought of as an obligatory quality control step in a DNA manufacturing process. RER succeeds in correcting 99% of mis-incorporated ribonucleotides during S phase.

Base Excision Repair (BER)

A ribonucleotide differs from a deoxyribonucleotide only by virtue of an extra chemical group attached to its ribose sugar. So, in effect, RER corrects the sugar part of a nucleotide. Base excision repair (BER), on the other hand, identifies and fixes a damaged or abnormal base component of a nucleotide--the A, T, G and Cs. If left uncorrected, these can cause mutations and they can also stall DNA polymerases. So base lesions must also be removed.

Many small chemical lesions arise on bases. I won't describe the chemistry here. There are too many and they have technical names like "deamination," "alkylation," "oxidation," and "hydrolysis." The discussion would lead us into a chemistry course, which is not our goal.

The important thing is all involve specific small chemical groups being added to, removed from, or changed on the base part of a nucleotide. The other is that base lesions occur extremely frequently: a single human cell experiences 20,000-50,000 of such events per day! So,, like RER, BER is an extremely active pathway in human cells.

BER--like RER--embodies a fairly classic "cut-and-patch" approach to DNA repair. I need to mention, though, that there are two versions of BER: short-patch BER (inserts a single correct base) and long-patch BER (inserts 2-10 bases). Short-patch BER is reminiscent of Polymerase Proofreading in that DNA polymerase exonuclease activity removes the bad nucleotide. Short-patch BER isn't associated with the replication. It takes place constantly and throughout the cell cycle.

Long-patch BER is more reminiscent of RER (and also, by association, with Okazaki fragment processing) in that the strand the DNA polymerase synthesizes displaces the strand with the lesion forming a flap that's then removed by FEN1. Long-patch BER uses many of the same enzymes as RER and Okazaki fragment processing. Thus it occurs at the replication fork.

Perhaps the most noteworthy aspect of BER is the number of detector proteins it uses. Because there are so many kinds of base lesions, the cell needs specialized base lesion detectors. With BER, there are 10-12 different detector molecules called DNA glycosylases, each of which detects one or a small number of specific base lesions.

As mentioned, these glycosylases are active throughout the cell cycle. They slide along one of the double helix's backbones individually flipping individual bases out of the helix and into a compartment in the enzyme, testing them for damage. This is a bit like pulling books from a shelf one by one to inspect the pages.

Nucleotide Excision Repair (NER)

Whereas BER fixes small chemical damage to bases, NER removes more severe lesions. It targets bulky, helix distorting lesions. Two examples would be: (1) UV-induced pyrimidine dimers, and (2) bulky chemical adducts caused by carcinogens in cigarette smoke.

Given its bulky targets, the NER detector protein recognizes DNA damage based on DNA shape distortion, not chemical identity. And while BER removes either one base (short-patch BER) or 2-10 bases (long-patch BER), NER removes 24-32 nucleotides--enough to excise the bulky lesions that NER targets.

Laike short-patch BER, NER takes place throughout the cell cycle. It's not specifically associated with DNA replication or the replisome. Its detector protein moves along genomic DNA constantly during all of the cell's growth phases, scanning for abnormal shapes. Once the detector protein binds to a damaged site, the rest of the repair machinery is recruited. (Note: I'm going to leave out the scientific names of the various repair proteins in this relatively short discussion about NER.)

NER stretches the "cut-and-stitch" model a bit more than the other repair pathways we've discussed. One difference is the requirement for helicases. Once the detector protein identifies the damaged site, a large enzyme complex containing two helicases arrives. One unwinds from the 3' direction and the other from the 5' direction. This opens up a 20-30 nucleotide repair bubble, exposing single-stranded DNA that's then coated and protected by RPA proteins. So helicases are needed for NER.

The second difference relates to the incision. To remove such a large stretch of DNA, not one but two incisions are made: one on each side of the lesion. One will be 20-22 nucleotides upstream of the lesion and the other will be 6-7 nucleotides downstream of the lesion. Once these incisions are made to the lesion strand, it dissipates away leaving just the template strand covered by RPAs.

The rest of NER takes the standard "cut-and-patch" approach. With the strand containing the lesion gone, DNA polymerase delta or epsilon arrives and, held in place by the PCNA sliding clamp, re-synthesize the now-missing strand (step 4). Then DNA ligase I arrives to stitch the two strands together, restoring a continuous double helix (step 5).

Translesion Synthesis (TLS)

Invariably, some serious lesions won't be repaired before a replisome happens upon them. What transpires when a DNA polymerase stalls at a lesion that hasn't yet been fixed? The cell--not surprisingly--has a solution for this scenario. It's called translesion synthesis (TLS).

In TLS, a stalled replication fork serves as a signal to recuit one of several special DNA polymerases called "TLS polymerases" that have the ability, unlike DNA polymerases delta and epsilon, to bypass large legions--to synthesize right past them--allowing the replisome to continue and putting the repair off for later. Let's get into how TLS works.

When DNA polymerase delta or epsilon comes upon a significant lesion--one that it can't bypass--it stalls. It stops synthesizing and waits there. But the CMG helicase in front of the polymerase continues unwinding the double helix. Because the polymerase has stalled, this creates a large stretch of single-stranded DNA, which becomes coated with a large number of RPA molecules.

This accumulation of RPAs is a signal to the cell that the polymerase has stalled. In response, the cell recruits the ubiquitin ligase RAD18 to the scene. RAD18 and its partner RAD6 arrive and attach a single small protein called ubiquitin to a specific amino acid on the PCNA sliding clamp: lysine 164. This single ubiquitin molecule attached to lysine 164 is in turn a signal to the cell to send a TLS polymerase to the scene.

TLS polymerases will supplant DNA polymerases delta and epsilon on the PCNA when that lysine is mono-ubiquitinated. The TLS polymerases have a PIP binding motif (facilitating its binding to the PCNA sliding clamp, generally) as well as a second binding motif specific for the ubiquitinated lysine on the PCNA. Because it's binding to two sites and not just to the PIP site, the TLS polymerases supplant the replicative helicases when the lysine is mono-ubiquitinated.

All good. But what is a TLS polymerase? First of all, there are several: DNA polymerase eta, kappa, iota, and zeta (and one more inaptly called Rev1). Whereas DNA polymerases delta and epsilon are highly accurate due to their unforgiving (tight) nucleotide binding sites, TLS polymerases have larger, more forgiving nucleotide binding sites that allow them to synthesize right past lesions.

Their more forgiving nucleotide binding sites do make the TLS polymerases more error prone than the replicative polymerases. That's OK. The priority in this situation is to bypass the lesion. Repairs--even of new mis-incorporations due to the lack of specificity of the TLS polymerases--can happen later, post-synthesis.

Also, like the DNA glycosylases that search for base lesions, there is some specialization among the TLS polymerases. For example, DNA polymerase eta is adept at bypassing UV-caused pyrimidine dimers. DNA polymerase kappa is good at bypassing bulky adducts. DNA polymerase zeta is extremely flexible (and thus very error prone!)

After the bypass, the PCNA is de-ubiquitinated and the normal polymerase (DNA polymerase delta or epsilon) resumes.