This information was most recently updated November 9, 2006.

Base Excision Repair (BER)

The "pathway" most commonly employed to remove incorrect bases (like uracil) or damaged bases (like 3-methyladenine) is called "base excision repair". Actually, it's misleading to talk about this as a pathway, because there are numerous variations, each specific for a different type of incorrect base. Nevertheless, all of the variant pathways have features in common, and each of the pathways can be considered to consist of 3 steps, with steps 2 and 3 being common for all pathways:

  1. Removal of the incorrect base by an appropriate DNA N-glycosylase to create an AP site
  2. Nicking of the damaged DNA strand by AP endonuclease upstream of the AP site, thus creating a 3'-OH terminus adjacent to the AP site
  3. Extension of the 3'-OH terminus by a DNA polymerase, accompanied by excision of the AP site

Specificity of the various pathways is conferred by the DNA N-glycoslyases. These hydrolyze the N-glycosylic bond between the base and the deoxyribose, as illustrated here by the action of uracil DNA N-glycosylase:

Uracil DNA-N glycosylase

Notice that the AP site created by a DNA N-glycosylase is identical to that created by spontaneous DNA depurination or depyrimidination.

A large number of DNA N-glycosylases has been identified. The following table provides examples of glycosylases found in human cells. Note the abundance of glycosylases specialized for removing U. Although U:A is not mutagenic, the presence of U instead of T alters the affinities of DNA-binding proteins. Consequently removal of U paired to A is important for the health of the cell.

Similar glycosylases are also found in other organisms. Portions of eukaryotic glycosylases often resemble prokaryotic glycosylases, but eukaryotic glycosylases frequently have amino- or carboxy-terminal additions that are presumably involved in specifying intracellular location and in interacting with other proteins.

Some of the Human DNA Glycosylases

 Acronym Full Name Size (aa) AP Lyase Activity Substrates
 UNG Uracil DNA N-Glycosylase 313 No  ssU>U:G>U:A, 5-FU
 TDG Thymine DNA Glycosylase  410  No  U:G>ethenocytosine:G>T:G
 UDG2 Uracil DNA Glycoslyase 2 327 No U:A
SMUG1 Single-strand-selective Monofunctional Uracil-DNA Glycosylase 1  270 No  ssU>U:A, U:G
MBD4  Methyl-CpG-binding Domain 4   580  No U or T in U/TpG:5-meCpG 
 MPG Methyl Purine DNA Glycosylase 293 No 3-meA, 7-meA, 3-meG, 7-meG
MYH  MutY Homolog 535 Yes (+/-) A:G, A:8-oxoG
 OGG1 8-Oxo-Guanine Glycosylase 1   345  Yes 8-oxoG:C
NTH1  Endonuclease Three Homolog 1  312   Yes T-glycol, C-glycol, formamidopyrimidine 

Most glycosylases are monomeric enzymes with no cofactor or divalent cation requirement. Most are highly specific for a certain type of altered base, but some have more relaxed specificity.

The crystal structures of many of the DNA glycosylases have been determined. They are similar to each other, and they suggest a common mode of action (with variations, depending on the specific structure recognized by the glycosylase). It appears that the DNA glycoslylases gently pinch the DNA while scanning it, with the result that the DNA kinks (bends sharply) at positions of instability caused by mismatching. The glycosylases all possess attractive binding sites for the bases that they are designed to recognize. The DNA kinking, combined with additional pushing by the enzyme, encourages mismatched bases to flip out of the DNA double helix and enter the binding site. If the binding site offers a good fit, the base remains inside long enough for its bond to deoxyribose in the DNA backbone to be severed by the enzyme. Here is an example of a good fit (uracil in its binding pocket within uracil DNA N-glycosylase):

The binding site in the above picture is attractive for uracil, because, once inside, uracil can stack with Phe158, and it can hydrogen bond with Asn204. The binding site is specific for uracil as opposed to thymidine, because the bulky methyl group at the C5 position of thymidine would be excluded by the polypeptide backbone chain next to Phe158 and by Tyr147.

In addition to removing altered bases, some DNA glycosylases also possess an AP lyase activity, which allows them to cleave the DNA backbone on the 3' side of the AP site (see above table and see diagram below). The cleavage is between C and O, not between O and P, and for this reason is called a "lyase" activity rather than an "endonuclease" activity.

AP-lyase.gif

Regardless of whether the DNA glycosylase possesses lyase activity, the next step in BER is catalyzed by an AP endonuclease, which cleaves the DNA backbone on the 5' side of the AP site. The best characterized AP endonuclease is, interestingly, the E. coli enzyme called "exonuclease III". For many years, the 3' to 5' exonuclease and phosphatase activities of this enzyme were thought to be manifestations of its in vivo functions. Perhaps those activities do play roles in vivo, but it is now clear that the AP endonuclease activity of exonuclease III is also important and in fact accounts for about 90% of the total AP endonuclease in the E. coli cell. The remaining 10% is accounted for by a different enzyme, endonuclease IV. Similar enzymes are also found in eukaryotic organisms. In S. cerevisiae, the major AP endonuclease is related to endo IV, but in mammalian cells, AP endonucleases related to exonuclease III are more abundant.

The following diagram summarizes the initial events of BER, from introduction of base damage to action of the AP endonuclease. Uracil DNA N-glycosylase is used as an example of a glycosylase lacking AP lyase activity, and the NTH1 enzyme is used as an example of combined glycosylase and lyase activity. In the latter case, the damaged base is thymine glycol, which is represented by a T with two OH groups. Lyase activity opens the deoxyribose ring at the AP site, generating an aldehyde-terminated carbon chain. The aldehyde group is represented by an oxygen double-bonded to the carbon chain.

First steps of BER with glycosylase/lyase

In both cases above, after strand cleavage by AP endonuclease the upper strand on the left side of the nick retains a 3'-OH terminus that can easily be extended by a polymerase. In the second case, the combined action of lyase and AP endonuclease have excised the abasic deoxyribose, leaving a gap that can be easily filled in by a polymerase and sealed by a ligase. The diagram below illustrates this process using DNA polymerase beta, which is the major polymerase used for base excision repair in mammalian cells.

Note that in the first case above (no lyase), the strand on the right side of the nick retains a 2'-deoxyribose-5'-phosphate terminus that must be removed to permit ligation. This non-nucleotide terminus is removed by one of two alternative pathways. In mammalian cells the major pathway is mediated by DNA polymerase beta, which has two distinct enzymatic activities. Using its polymerase activity, polymerase beta incorporates the correct nucleotide at the AP site (indicated by red bond lines and by a bold C or T), and then, with a recently discovered deoxyribose phosphatase (dRPase) activity, it excises the deoxyribose phosphate moiety, as shown below.

BER-Polymerase Beta

Note that the above pathways utilizing DNA polymerase beta generate a repair "patch" (stretch of newly synthesized DNA) only a single nucleotide in length (indicated by red bond lines and bold C or T). An alternative longer patch pathway is also sometimes employed in mammalian cells and is usually employed in yeast cells.

When the short patch pathway is employed, the final ligation step is carried out by DNA ligase III in partnership with the XRCC1 protein.

The longer patch pathway depends on enzymes normally involved in DNA replication: DNA polymerase delta or epsilon, the "clamp" protein PCNA (a cofactor for both polymerases delta and epsilon), DNase IV/FEN-1/Rad27, and DNA ligase I. The mechanism (shown below) probably involves synthesis of a new stretch of DNA several nucleotides long (indicated in the diagram by red bond lines and bold C, A, G and T), which results in displacement of the strand terminated by the deoxyribosephosphate group. The displaced strand is then probably removed by a structure-specific nuclease (called DNase IV or FEN-1 in mammals and Rad27 in S. cerevisiae), to create a ligatable substrate. DNase IV/FEN-1/Rad27 appears to play a similar role in the maturation of Okazaki fragments behind replication forks.


Link to next page