Molecular Biology 12: 'Base excision repair (BER) and nucleotide excision repair (NER)'

These are my notes from lecture 12 in Harvard’s BCMP 200: Molecular Biology course, delivered by Johannes Walter on October 3, 2014.

Whereas RER and MMR from last time repair replication errors, base excision repair (BER) and nucleotide excision repair (NER) repair chemical damage. Here is a comparison of BER and NER.

Base excision repair (BER)

Hydrolysis is the most common form of DNA damage. Often it causes spontaneous de-purination or de-pyrimidation - i.e. removal of the entire aromatic base, resulting in an “abasic site,” often abbreviated an “AP site”. Each day in one cell 18,000 depurination and 600 depyrimidation events occur. Another form of hydrolysis is the deamination of cytosine, converting it to uracil. This creates a G:U base pair, which on replication will become one G:C copy and one A:U copy. When the A:U copy further replicates, it will create a fixed A:T mutation. The net result is a C→T transition. Uracil-DNA glycosylase (UNG; UNG gene in humans) is responsible for repairing the uracil mutation before it leads to a transition.

Oxidation is another common form of DNA damage. Most often, it converts guanine to 8-OxoG. If 8-OxoG then rotates, it is able to base pair with A instead of C. The replicaton of the 8-oxo-G:A pair will lead to one 8-OxoG:A daughter duplex, and one T:A daughter duplex. The net result is a G→T transversion.

In addition to uracil, 8-oxo-G and abasic sites as explained above, other bases resulting from chemical damage include hypoxanthine, thymine glycol, 3me-A, and εA. BER is responsible for repairing all of these.

“Short patch” BER involves a DNA glycosylase, AP endonuclease (APEX1 gene in humans; serves to remove AP sites), AP lyase, pol β and DNA ligase XRCC1. An alternate pathway is called “long patch” BER.

The enzyme 8-OxoG deglycosylate (OGG1, OGG1 gene in humans) rotates the 8-OxoG so it is pointing outward from the DNA strand, and stacks it with the ring of residue F319 in the enzyme. A hydrogen bond with residue G142 is only possible for 8-OxoG and not regular G, conferring specificity. And a K residue attacks the 8-OxoG to cleave it [Bruner 2000]. The mechanism by which OGG1 locates 8-OxoG in the first place is an area of continuing investigation. Another enzyme, MYH (MYH gene in humans) removes the A paired with 8-OxoG, or the 2-OH-A paired with G.

Apex1 knockout is embryonic lethal in mice, presumably because Ape1 is involved in all BER processes, whereas the Ogg1 and Ung knockouts have comparatively mild phenotypes.

Nucleotide excision repair (NER)

NER removes diverse, bulky lesions.

UV light can damage DNA by causing neighboring pyrimidines to covalently bond into either a “cyclobutane pyrimidine dimer” (~90% of the time) or a “6-4 photoproduct” (~10% of the time). Early studies showed that UV irradiation induced a small amount of DNA replication in bacteria, so it was inferred that an enzyme removes chunks of DNA that contain damage, and mutants were identified that lacked this activity [Setlow & Carrier 1964, Pettijohn & Hanawalt 1964, Howard-Flanders & Theriot 1966]. In humans, loss-of-function of any of eight different genes involved in UV damage repair causes a recessive disease in humans called xeroderma pigmentosum (XP) [form A is OMIM# 278700] characterized by extreme sensitivity to UV radiation, causing skin damage and early onset skin cancer. Cultured cells from XP patients can be killed (about 99% of cells dead) by ~1 J/m² of UV light, whereas comparable cell death is not seen in control cells until about 10 J/m² of UV. Using this cellular assay, researchers performed complementation experiments by fusing different patients’ cells to form heterokaryons and then testing the heterokaryon’s phenotype. These experiments revealed eight complementation groups, with the diseases dubbed XPA through XPG, plus XPV.

The next round of experiments tested whether these eight genes operated in the same or different pathways. The reasoning was that if, say, XPA and XPF they function in a single pathway, and loss of either gene completely disables the pathway, then the double knockout XPA^-/-, XPF^-/- should have a phenotype no more severe than either of the single knockouts (this situation is called epistasis in this context). Whereas if they operate in different pathways, then the double knockout should be more severe than either mutation alone (this situation is called additive in this context). (As long as we are defining terms, note that a third possibility, which had already been ruled out in the case of XP, was redundancy, in which two pathways can substitute for one another, so that either single knockout is well-tolerated but the double knockout has a phenotype). In the case of XP, it was revelaed that seven types were epistatic to each other (same pathway), while one was in a different pathway.

The mechanism for the UV damage repair pathway was finally worked out in [Sugasawa 2001]. XPC binds DNA even in the absence of damage. At sites with damage, the duplex is distorted, exposing a small strand of ssDNA, and XPC preferentially binds these ssDNA segments, thus indirectly recognizing sites of distortion, without any specificity for the specific chemical modification causing the distortion. This means that XPC will also bind ssDNA exposed at moments of transcription or replication. XPB protein is a helicase which unwinds the DNA at a site where XPC has bound, allowing XPD protein - also a helicase - to bind and start unwinding. (Note that XPB and XPD act as part of a larger complex called TFIIH). XPD stalls if it hits a bulky adduct indicating DNA damage. Therefore, XPB and XPD function to confirm the presence of DNA damage, whereas XPC by itself can only hypothesize that there might be damage at a particular site. XPG will then bind and recruit XPF; only when both are present, they will start destroying the damaged strand.

There is also transcription-coupled NER, in which RNA pol II stalls at a damaged base, recruiting proteins CSA and CSB (ERCC6 and ERCC8). Mutations in these two genes to a different recessive disease called Cockayne syndrome [OMIM # 216400 and 133540].

If replication begins before a thymine dimer has been repaired, the replication fork will stall at the dimer, and its presence there will occlude NER from occurring. If uncorrected, this problem would lead to incomplete replication by the time of mitosis. To solve this problem, DNA pol η (eta; POLH gene in humans) has evolved, which can supplant DNA pol δ and can replicate right over a thymine dimer, inserting two As in the correct location across from the dimer so that replication can complete [Johnson 1999]. DNA pol η is able to achieve this by lacking geometric selection in its nucleotide binding site, so that although it plays a crucial role of copying over thymine dimers, it has a far higher error rate than DNA pol δ overall, so it cannot be suffered to engage in general DNA replication. Instead, its recruitment to DNA is tightly regulated [reviewed in Waters 2009]. Loss-of-function of POLH causes variant XP (XPV) and is additive (not epistatic) with the other causal XP mutations.