DNA Repair 101

You don't have to be a biologist to know that anything that can damage DNA is potentially harmful, because it can cause mutations (which are, in fact, mostly harmful; very few mutations are beneficial). Fortunately, cells contain dozens of different kinds of repair enzymes, and most DNA damage is repaired quickly. When damage isn't repaired quickly (or properly), you have a mutation.

It's not much of a stretch to say that DNA repair enzymes play a front-and-center role in evolution (or at least the portion of evolution that's driven by mutations). Which is why molecular geneticists tend to pay a lot of attention to DNA repair processes. Anything that can affect the composition of DNA can change the course of evolution.

DNA is remarkably stable, chemically. Nonetheless, it is vulnerable to oxidative attack (by hydroxyl radicals, superoxides, nitric oxide, and other Reactive Oxygenated Species generated in the course of cell metabolism—never mind exogenous poisons).

Of the four bases in DNA—guanine (G), cytosine (C), adenine (A), thymine (T)—guanine is the most susceptible to oxidative attack. When it's exposed to an oxidant, it can form 7,8-dihydro-8-oxoguanine, OG for short. What can happen then is, the OG residue in DNA pivots around its ribosyl bond until the amino group is facing the other way (see diagram), and when that happens, OG can pair up with adenine instead of guanine's usual partner, cytosine.

When guanine is oxidized to form 7,8-dihydro-8-oxoguanine,
it mispairs with adenine instead of its usual partner, cytosine.

Rest assured, there are proofreading enzymes that can and will detect such funny business in short order. But if OG isn't detected and replaced with a normal guanine before replication occurs, OG may get paired up with an adenine during replication (and then it'll eventually be swapped out with thymine, adenine's usual partner). That's bad, because what it means is that a G:C pair ended up getting changed to a T:A pair. (The place of the G got taken first by OG and then T. The place of G's opposite-strand partner, C, eventually got taken by A.) In so many words: that's a mutation.

It turns out there's a special enzyme designed to prevent the G↔T funny business we've just been talking about. It's called oxoguanine glycosylase, or Ogg1 for short. You'll sometimes see it called 8-oxoguanine-DNA-glycosylase, and from a capabilities standpoint it's often (wrongly) compared to the Fpg enzyme (formamidopyrimidine-DNA glycosylase), which is not the same as Ogg1 at all. 

Just about all higher life forms have an Ogg1 enzyme (which clips OG out of DNA and ensures it gets replaced with a brand-new guanine before any funny business can happen). Surprisingly few bacteria have this enzyme, instead preferring to let the more general-purpose Fpg (MutM) take its place. If you run a Blast search of a reference Ogg1 gene (the Drosophila version works well) against all bacterial genomes, you'll get only a few hundred matches (out of around 10,000 sequenced bacterial genomes), the vast majority belonging to members of the class Clostridia (a truly fearsome group of anaerobic spore-formers containing the botulism germ, the tetanus bacterium, the notorious C. difficile—also known as C. diff—and some other creatures you probably don't want to meet). If you run the same Blast search against Archaea (this is the other major "germ-like" microbial domain, along with true bacteria), you'll get hits against almost every member species of the Archaea. Personally, I think it's likely the Ogg1 enzyme originated with a common ancestor of today's Archaea and Eukaryota, and arrived in Clostridia by lateral gene transfer (not terribly recently, though).

One thing is certain: E. coli does not have Ogg1, nor does Staphylococcus, nor Streptococcus, nor any germ you've ever heard of (other than the aforementioned Clostridia members, plus Archaea). And yet, every yeast and fungus has it, every plant, every fruit fly, every fish, every human—every higher life form. Ironically, only five members of Archaea turned up positive for the Fpg enzyme when I did a check, whereas almost all Eubacteria ("true bacteria") have it, including Clostridia. Bottom line, Clostridia have the best of both worlds: Fpg, plus Ogg1. Belt and suspenders, both.

This is just a tiny intro to the subject of DNA repair, which is a vast subject indeed. For more, see this article, or just start rummaging around in Google Scholar.