Damage recognition in nucleotide excision repair
Malta, E.
Citation
Malta, E. (2008, December 9). Damage recognition in nucleotide excision repair. Retrieved from https://hdl.handle.net/1887/13327
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/13327
Note: To cite this publication please use the final published version (if applicable).
Summary and general discussion
148
Summary and general discussion
Even though DNA has the very important task of storing all genetic information its integrity is constantly threatened by various agents. Upon reaction with the DNA these agents damage the DNA by formation of lesions that alter its molecular structure. Such damages can lead to cell death when blocking vital cellular processes like replication and transcription or can cause mutations and thereby alter the genetic information.
The accumulation of double strand breaks and base modifications blocking replication and/or transcription in cells results in an elaborate DNA damage response which triggers several cellular pathways to preserve the cell’s integrity. These pathways include cell cycle checkpoint control, which is initiated in eukaryotes to slow down the cell cycle and to give the cell time to repair the damage. In addition, a transcriptional response (both in prokaryotic and eukaryotic cells), which leads to the expression of numerous genes required to restore the cell’s genetic material, DNA damage repair pathways and chromatin remodeling pathways are induced. In case however that the cell experiences too much damage to efficiently restore its genetic information the cell might invoke a programmed cell death pathway, called apoptosis to preserve the integrity of the organism. Alternatively it might go into senescence, an irreversible state where the cell no longer divides and therefore cannot harm the organism.
DNA repair, one of the major mechanisms to counteract DNA damage, can be subdivided into several distinct DNA repair pathways as described in section 1.2. These pathways specialize in the removal of DNA damage and thereby aid to maintain its integrity.
Direct reversal is the easiest pathway in this regard since its proteins directly remove the aberrant group or catalytically revert the modified nucleotide to its original undamaged form.
It is also the only repair process that does not involve a disruption of the phosphodiester backbone. All other mechanisms that repair base damage do result in DNA backbone cleavage.
Examples of these are base excision repair (BER), UV damage endonuclease (UVDE) repair and nucleotide excision repair (NER). Base excision repair requires the presence of a relatively large number of proteins, glycosylases, each recognizing only one or a small subset of DNA damages. This system is used to repair the most commonly encountered DNA base modifications. NER on the other hand uses an invariant set of proteins to remove many different types of damage, and therefore repairs many damages that are not processed by other DNA repair mechanisms. UVDE, a single protein that is present in some specific bacteria and eukaryotes, is also capable of recognizing various structurally unrelated damage types, although its range of action is not as wide as that of the NER system.
Proteins recognizing only one or a small set of related damages do so by making use of a highly specific binding pocket. This pocket has the right dimensions to accommodate a modified nucleobase and excludes undamaged bases and differentially modified bases.
In order to gain access to these nucleotides and incorporate them in the size- and shape- complementary pocket they are actively or passively flipped out of the DNA helix (Section
1.3). This strategy however only grants recognition of a single modification and for that reason it is unlikely that the same strategy is followed by the NER proteins.
The NER process has been extensively studied in Escherichia coli and involves the action of only three NER-specific proteins to recognize the damage and incise the damaged strand: UvrA, UvrB and UvrC. The molecular mechanism of the UvrABC system has been extensively studied and is summarized in section 1.4. Both UvrA and UvrB are involved in recognition of the damage. In this regard, UvrA is believed to initially signal the presence of a lesion in a stretch of DNA after which UvrB is used to confirm its presence and to pinpoint the exact position of the damage. UvrC is the nuclease of the system and performs its dual incision after the presence of a lesion has been established. After both incisions have been performed the excised and damaged oligonucleotide is removed and the resulting gap is filled by DNA polymerase.
The initial step in the bacterial NER process is the formation of a complex between UvrA and UvrB, which in chapter 2 is shown to consist of two UvrA and two UvrB molecules.
This UvrA2B2-complex is formed in solution and binds to the DNA in search of a damage. In this situation the C-terminal domains of the two UvrB subunits interact. Since a homologous domain of UvrC is known to bind to this same C-terminal region of UvrB it was proposed that the UvrB-UvrB interaction prevents premature association of UvrC. Only when both UvrA subunits dissociate from the complex, which occurs at a damaged site, the C-terminal domains of UvrB become available to UvrC and incision can take place.
A crystal structure of UvrB revealed that in addition to helicase domains and an ATPase site the protein contains a highly conserved β-hairpin structure. Biochemical characterization of this domain has shown that it is essential for damage recognition by the UvrB protein. Mutation of residues located at the bottom of the hairpin showed that two conserved tyrosine residues (Tyr92 and Tyr93) prevent binding of UvrB to undamaged DNA by clashing with undamaged nucleotides.
In chapter 3, 2-aminopurine fluorescence measurements have shown that when both UvrA subunits have dissociated resulting in a complex of UvrB bound to a damaged DNA substrate, the nucleotide directly 3’ to the lesion and its base pairing partner become extrahelical. This suggests that also UvrB uses a base flipping mechanism to search for DNA abnormalities. In contrast to other DNA repair enzymes, however, the damage itself does not change position upon UvrB binding (chapter 4). A previously solved co-crystal structure of B. caldotenax UvrB bound to a small DNA loop revealed that one of the DNA strands threads behind the hairpin structure of the UvrB protein and is wedged between the hairpin and the body of the UvrB protein whereas the other strand passes in front of it. In chapter 4 we present evidence that the strand passing behind the hairpin corresponds to the damaged one suggesting that UvrB uses its helicase activity to translocate along the damaged DNA strand in the 3’→5’ direction. A mutation in a pocket of the UvrB protein (Y249A), previously proposed to be involved in damage recognition by excluding damaged nucleotides, showed
150
that insertion of a base into this pocket is important for dissociation of UvrB from an undamaged site. From our results we propose the following model for damage recognition by UvrB: incorporation of an undamaged nucleotide into the pocket of UvrB and stacking on Tyr249 places the adjacent DNA in such a position that it clashes with hairpin residues Tyr92 and Tyr93. This thereby explains the ability of mutants lacking these hydrophobic residues to bind to undamaged DNA. On damaged DNA the presence of the damage arrests translocation behind the hairpin blocking the DNA in such a position that incorporation of a nucleotide into the pocket of UvrB is prevented. In this situation the adjacent nucleotides have more conformational freedom and under the influence of Tyr92 and Tyr93 they will be able to occupy a position where clashing with these residues no longer occurs. This model predicts that lesions small enough to pass behind the hairpin and planar enough to pass through the pocket of the UvrB protein will escape its detection. In chapter 5 it is shown that DNA substrates containing such a planar nucleobase substitution indeed evade detection by the UvrB protein. In contrast to UvrB these small planar nucleobase lesions are efficiently recognized by UvrA, reflecting different damage recognition strategies for both proteins.
UvrA probably recognizes a more general distortion in the DNA helix, like the occurrence of a stretch of ssDNA at the site of damage. Additionally, binding of UvrA might facilitate ssDNA formation as was proposed to occur in chapter 5. UvrA and UvrB clearly probe for different properties of a lesion and only when damage recognition determinants of both UvrA and UvrB signal the presence of a lesion, repair will take place. This concerted action thereby prevents unwarranted and costly repair. Fluorescence studies using a pyrene modification have shown that upon binding of UvrA and/or UvrB to the damaged site, a conformational change occurs in the DNA.
Although the prokaryotic and eukaryotic NER proteins share little structural homologies (section 1.6) both systems recognize the same wide range of structurally different damages. Additionally, in both kingdoms these systems serve an analogous function i.e. to repair lesions that are not or not efficiently removed by other DNA repair systems. This suggests that in spite of their differences both systems apply a similar mechanism of damage recognition. Indeed, also in eukaryotes multiple damage recognizing proteins cooperate to prevent unwarranted repair. One of these proteins, the initial damage-recognizing protein XPC, has been shown to flip nucleotides out of the DNA, a mechanism also employed by the UvrB protein. However in contrast to UvrB XPC flips out the damaged nucleotide(s) and their respective base pairing partners. The damage is flipped from the outer DNA strand into solution and therefore becomes solvent-exposed. This suggests that XPC follows another strategy than the UvrB protein. Flipping by XPC probably provides damage recognition by being energetically more favorable at sites of DNA damage due to a lesion-induced reduction in base stacking interactions. However, in the XPC study, it became apparent that it was the presence of a three nucleotide mismatch and not the damage itself that positioned the protein on the DNA. It therefore remains to be determined which of the DNA strands normally contains
the damage, the inner or outer DNA strand. Fluorescence measurements using the newly synthesized DNA modifications described in chapter 5 might elucidate the orientation of the damaged DNA strand and help to further characterize XPC’s flipping mechanism. These modified DNA substrates might also help to unravel the damage recognition properties of the XPA protein. The purified protein is known to display a binding preference for damaged DNA over undamaged and has a high propensity to binding to DNA molecules exhibiting a strongly kinked structure. This might reflect a structure induced by XPA itself upon binding to damaged DNA to check for the presence of damage, since DNA bending is facilitated at a damaged site. A DNA bending mechanism to sense the presence of a damage could also be envisioned for the bacterial NER protein UvrA. Not only does it confer damage recognition, it also might facilitate strand opening in order to allow the UvrB protein to insert its hairpin.
However, since XPA is believed to associate with TFIIH it could also be that its preference for kinked DNA enables it to bind to a specific DNA architecture that is formed only when TFIIH encounters a lesion. In this scenario XPA does not serve as a damage recognition factor but is only required to orchestrate proper NER complex formation since it is known to interact with multiple components of the incision complex.
Another similarity between the prokaryotic and eukaryotic NER systems is the involvement of a helicase protein in both systems. In bacteria the UvrB protein possesses a limited DNA unwinding activity required for repair whereas in humans this function is executed by XPD. Both proteins open up the DNA helical structure at the site of damage resulting in an open structure recognized by the nucleases of the NER system: UvrC for bacteria and XPF•ERCC1/XPG for humans. The helicase activity of UvrB also serves a more direct role in damage recognition since its activity is probably used to stall at a damaged site (chapter 4). A similar mechanism has previously been proposed for the XPD protein. Recent determination of a crystal structure of XPD revealed the presence of a potential nucleotide binding pocket, a structural feature also observed in UvrB. In UvrB this pocket was proposed to be involved in discriminating damaged from undamaged DNA since incorporation of an (undamaged) nucleotide into this pocket results in DNA dissociation. The same mechanism can therefore be envisioned for XPD. Alternatively the proposed pocket of XPD might confer damage recognition by preventing incorporation of a damaged nucleotide. Conformational studies using 2-aminopurine and the fluorescent DNA modifications described in chapter 5 might further elucidate the proposed damage recognition mechanism of this protein.
