|dc.description.abstract||In all known organisms, except some viruses, genetic information is contained in the form of DNA. Although genetically relatively stable, DNA is subject to continuous damage from the external- and cellular environment. Mechanisms that maintain DNA integrity, such as DNA repair, are therefore essential to the cell. It has long been known that defects in different repair systems may result in increased incidence of cancer. As one example, defective mismatch repair causes hereditary non-polyposis colon cancer (HNPCC). Base excision repair (BER) is a different type of excision repair. It is initiated by glycosylases that remove damaged or inappropriate bases from DNA. Until recently association of defects in DNA glycosylases with specific disease phenotypes in animals or humans has largely been missing. The human UNG-gene encodes two major uracil-removing DNA glycosylases in mammals, mitochondrial UNG1 and nuclear UNG2. Uracil is not usually considered a normal base in DNA. However, uracil can enter DNA either by spontaneous deamination of cytosine in DNA, or by incorporation of uracil (as dUMP) during replication. Deamination results in mutagenic U:G mispairs, while incorporation of dUMP results in U:A pairs that may have detrimental effects by other mechanisms, including altered binding of transcription factors. Surprisingly, it has recently become clear that uracil can also be formed in DNA by enzymatic deamination of cytosine in B-cells. This programmed generation of uracil is involved in class-switch recombination (CSR) and somatic hypermutation (SHM) required for affinity maturation of antibodies. Thus, inactivating mutations in the Ung (in mice) or UNG (in humans) in mice or humans give distinct phenotypes. Mice and patients carrying such mutations have a hyper IgM syndrome characterised by increased IgM, decreased IgG, impaired CSR and altered pattern of SHM. Ung-/- knockout mice also have been shown to have a 20-fold higher incidence of B-cell lymphoma. These findings have made UNG-proteins and other DNA glycosylases even more interesting to study.
We have for the first time been able to purify the full-length hUNG2 and have determined some of its enzymatic and kinetic parameters. hUNG2 is localised in replication foci and has a very high turnover number, making it an ideal enzyme keeping track of uracil close to the fast moving replication fork. Our study compares the enzymatic properties of hUNG2 with the probably second most efficient uracilremoving glycosylase in the cells, hSMUG1. hUNG2 has the ability to remove uracil both from single stranded DNA, U:A pairs and U:G mispairs indicating the hUNG2 could be responsible for the removal of deaminated cytosine as well as misincorporated uracil during replication. The activity of hUNG2 is strongly stimulated by Mg2+ present at physiological concentrations. However, the most pronounced effect is a 140-fold reduction in the KM-value when using ssDNA as substrate. We also speculate that UNG2 may be involved in removal of uracil in single stranded DNA close to the replication fork, thus initiating repair by recombination or fork regression. hSMUG1 has properties that makes it a likely candidate for a role as a broad specificity backup for hUNG2. hSMUG1 has relatively low catalytic efficiency and turns over substrate slowly. However, it is strongly inhibited by AP sites to which it binds, and is strongly stimulated by AP-endonuclease APE1 (also called HAP1) in assays with double stranded DNA.
The UNG gene is regulated by the two promoters PA and PB that drive the expression of nuclear UNG2 and mitochondrial UNG1, respectively. A role for hUNG at the replication fork fits very well with the way UNG2 mRNA is cell cycle regulated. After the release of serum starved HaCaT cells, mRNAs for UNG1 and UNG2 are increased 2.5 and 5-fold, respectively, in late G1/early S-phase. This is associated| with a 4-5-fold increase in enzyme activity. We have identified putative E2F-binding sites in both promoters. The strongest reduction in promoter activity was observed after mutations in E-box elements in both promoters, although no footprint corresponding to the position of the E-box was found by in vitro analysis. In contrast, footprint analysis of PA shows footprints that overlap with the putative E2F and CCAATT elements. Furthermore, we observed strong footprints that overlap with 3 putative AP2 binding sites in PA. Mutation of these AP2 elements results in a 2-3-fold increase in basal promoter activity, indicating binding of a negative regulator of PA to these sites. The precise nature of this regulator is not known. However, overexpression of either AP2 or a truncated AP2 lacking the activation domain, but retaining its DNA binding domain, results in a 2-4-fold stimulation of PA. Together with the mutation analysis, this indicates that AP2 and the truncated AP2 bind to elements that were occupied by a negative regulator. Binding of AP2 to these putative AP2 elements in PA demonstrated by in vitro footprint analysis using HeLa cell extracts. Furthermore, extracts fortified with purified AP2 enhanced the footprints.
During our work with the UNG2 promoter we sequenced a region upstream of the promoter area and identified a potential human AlkB homologue, that we later named hABH2. BLAST searches of the human genome databases identified yet another putative AlkB homologue, that we later named hABH3. We found that hABH2 and hABH3, like AlkB, are Fe2+ -and a-ketoglutarate-dependent oxygenases that directly revert 1-methyladenine and 3-methylcytosine to adenine and cytosine, respectively. These aberrant methylations are known to be cytotoxic to cells. The repair mechanism involves hydroxylation of the methyl group that is subsequently spontaneously released as formaldehyde. The identification of these enzymes in fact doubles the number of direct repair enzymes identified in mammals, the other ones being O6-methylguanine-DNA methyltransferase (MGMT) that uses an entirely different mechanism for removal of the methyl group, and DNA ligase that seals single strand nicks. We have also examined the transport of fluorescently tagged hABH2 and hABH3 in human HeLa cells. hABH2 is transported to the nucleus exclusively and shows some accumulation in nucleoli outside S-phase, while being associated with replication foci during S-phase. This indicates a possible role for hABH2 in repairing replication blocking 1-methyladenine and 3-methylcytosine near the replication fork. hABH3 is mainly transported to the nucleus, but is also present in the cytoplasm. hABH3 is largely excluded from the nucleoli, but occasionally we observed distinct spots in nucleoli and nucleoplasm. A surprising and exciting observation was that AlkB and hABH3 also have the ability to repair RNA both in vitro and in vivo in E. coli. Although the biological significance of this RNA repair remains to be determined, it opens a new field of research, and may suggest that mechanisms of macromolecular repair are more extensive than previously thought. Hopefully our work on RNA repair will inspire the initiation of new studies on how cells handle damaged RNA, and how damage to RNA will affect its functions in protein synthesis and regulation of cellular processes. It also should contribute a useful link between DNA repair and the highly important, but apparently not highly visible, research area protein repair. As discussed in the Introduction of this thesis, DNA, RNA and protein are all repaired, and it is perhaps time to consider these entities as a whole; that is: macromolecular repair.||nb_NO