April 18, 2004

Junk, Rotations, and Chicken

Posted by Susan

Like Matt's, my recent lack of posting has been due to business rather than laziness. I am rotating once again (i.e. working for ten weeks in a lab to see if I want to do my thesis work there), this time while taking classes, which is rather taxing.

I never know how much it's kosher to say about ongoing research, but to put things broadly, I'm looking at BRCA1 in DT40 (chicken pre-B) cells. To the general public, BRCA1 is best known (if at all known) for being one of the two main genes associated with familial breast and ovarian cancer. The BRCA1 gene product is an extremely large protein involved in many processes--to name a few, regulation of growth and proliferation, RNA transcription, and (my focus) DNA repair. (Here's the OMIM entry if you're at all interested in BRCA1).

Due to my work, I've been thinking a lot lately about noncoding DNA*, so I was pleased to see the subject come up at The Panda's Thumb (due to the aforementioned business, I am writing about this quite a long time after seeing the article). I suppose I can't be touting The Panda's Thumb as a great new blog at this point, but it's certainly a great blog and worth checking out if you have any interest in evolutionary theory or science education. It's also fun if you like to tear apart anti-evolutionists--I always admire people who can attempt to argue with creationists or "intelligent design" devotees or the like. Anyway, the discussion at The Panda's Thumb focused more on introns and related topics, whereas my thoughts on noncoding DNA have concerned DNA repair.

Much of the early work on DNA repair was carried out in yeast and bacteria, whose genomes contain very little noncoding DNA (about 1% for yeast as compared to about 43% for the human genome). Both of these organisms primarily repair double-strand breaks (DSBs) by recombinational repair (aka homologous recombination or HR). In this process, the resected ends of DNA invade the intact homologous region on the sister chromatid to copy it, resulting in repair without loss of genetic information (though gene conversion may occur). One important feature of HR is that the resolution of the double Holliday junction formed from the strand invasions (helpful picture from Nat Rev Mol Cell Bio (2002) 3:430) can result in crossover or noncrossover products, depending on the resolution of the junctions. An interesting process (single strand annealing) shares some similarities with HR while giving only noncrossover products, but it's not germane to this discussion.

Nonhomologous end-joining is a pathway present only in eukaryotes. It is the predominant pathway of DSB repair in humans. Unlike HR, nonhomologous recombination involves loss of genetic material. Basically, once a DSB forms, the ends are recognized, resected a bit, and then joined together. Microhomology (of two or three base pairs) is required in some forms of NHEJ. While this process is critical for generating genetic variety in T-cell receptors and antibodies, it can damage genes and may be involved in generating some of the chromosomal translocations found in cancers like chronic myelogenous leukemia.

So why do humans (and other higher eukaryotes) use such an error-prone method to fix their DNA? Most attempts to answer this question have focused on the difficulty of performing the homology search, a key component of HR, on the human genome. Two problems are evident: first of all, the human genome (6 billion base pairs) is considerably larger than either the yeast (12 million bp)** or E. coli (5 million bp) genome. However, the chicken genome (1.1 billion base pairs), which is more comparable in size to the human genome, undergoes a significantly larger portion of its repair through HR***, so genome size is unlikely to be the sole factor dictating the preference for HR in repair of human DNA. This leads us to the other homology search problem--repetitive DNA elements, most commonly found in noncoding DNA. As I mentioned earlier, roughly 43% of the human genome is thought to consist of noncoding DNA. Genomes of lower eukaryotes and avians , which use HR preferentially over NHEJ, have much less. The theory is that, in an organism whose genome contains lots of repetitive elements, the homology search can lead to false matches of repetitive elements on different chromosomes (or on the same chromosome in the wrong place). If HR were performed on these false matches and crossover products arose, this would give rise to translocations, breakage cycles, circular chromosomes, chromosome loss, and other sorts of genomic instability that could be fatal for the cell (or, should the cell survive the changes and become cancerous, fatal for the organism).

Obviously, this sort of thing is rather hard to test, let alone prove, but I enjoy the speculation.

*Or, as Francis Crick would have it, "junk DNA". It's interesting to look at Crick's track record--he has managed to be both utterly right (structure of DNA, codons as triplets of bases) and utterly wrong (central dogma), and I suspect the whole noncoding DNA business is not going to go his way either. Regardless of how important noncoding DNA turns out to be, I agree with the commenters at the Panda's Thumb that the characterization of noncoding DNA as "junk" has probably retarded progress in the field.

**Fun fact of the day: the Saccharomyces genome database FAQ, in its miscellaneous questions section, addresses the question "I think I may have a yeast infection. What should I do?" thusly:

Unfortunately, we cannot directly help you because SGD is a scientific database that provides information about the molecular biology and genetics of the yeast Saccharomyces cerevisiae to researchers. We are not medical doctors and cannot give medical advice. You should speak to a qualified physician about any medical concerns. To find out more information about pathogenic yeast infections such as Candidiasis, you can go to a medical library at a local university, search the PubMed database for relevant literature, or browse the Candidiasis information at MEDLINE plus.

This is, to the best of my knowledge, the second most amusing example of references to fungal diseases in FAQs. (After this.)

***Among the delightful consequences of this preference is the high specificity of gene targeting (an HR-dependent process) in chicken cell lines like DT40 cells.

Posted by Susan at April 18, 2004 05:11 PM

You're not going to point people to the most amusing example of references to fungal diseases?

Posted by: Matt at April 18, 2004 06:43 PM
Post a comment

Remember personal info?