I was a little surprised to see that while other bloggers (here, here, here, and here) have been arguing about whether or not the mutation really increases the risk to the degree (20-80%) mentioned by Brin, no one has really looked into the structure and biochemistry of the LRRK2 protein to see if there's a biochemical explanation for Parkinson's risk. I guess that task is up to me.
Let's begin at the DNA...with sbSNP
I looked for the LRRK2 gene in dbSNP to see if I could find the mutation in question (G2019S). LRRK2 is huge! The median size of the transcribed region for a protein-encoding gene is 16,995 nucleotides (I got this value from Scherer. This is such a handy book!).
LRRK2 spans 144,273 nucleotides, has 51 (yes! 51 exons!) and makes an mRNA that's 9239 nucleotides long. The base change that causes glycine to be replaced by a serine at amino acid 2019, maps in exon 41. dbSNP shows me that quite a few SNPs and indels have been identified in this gene and many of the SNPs replace one amino acid with another. I was hoping to find that the SNP (rs34637584) had some frequency data, but no such luck.
One very important point, from OMIM, is that several of the mutations in the LRRK2 gene, including the G2019S mutation are autosomal dominant. This explains why someone with the mutation would be at a high risk of getting the disease.
An autosomal dominant mutation, by definition, is a mutation that will most likely show a phenotype in anyone who has it. If one parent has one copy of the mutation, then each child has a 50% chance of inheriting the mutation. If your children get the chromosome with the mutation, and they live long enough, they will most likely have disease.
So why would this mutation be dominant?
There are two reasons, first LuzÃ³n-Toro et. al. (1) found that this mutation creates a new phosphorylation site (the serine at 2019) and increases the kinase activity of this protein. Adding or taking phosphate groups away from proteins is a commonly used tool for controlling their activity. When phosphate groups get added, the activity is described as kinase activity.
Second, this protein is normally found as dimer (two parts). When a person has one copy of the mutant gene, and one copy of the normal gene, the proteins that get made will either have two normal chains, two mutant chains, or one mutant and one normal chain. Often the mutant chain can mess up the activity of the normal chain, leading to a case where two thirds of the protein show the altered activity.
Let's look at the structure!
I got temporarily excited by a reference I found in OMIM saying that the structure had been solved in 2008 (2). So, I looked for the structure and tried to find the mutation site. I knew that this structure didn't contain the whole protein since the chains in the structure are only 184 amino acids long and in LRRK2, each chain has 2527 amino acids, but I wasn't sure at first how this structure compared to the intact protein and if contained the site of the G2019S mutation.
So, I turned to blastp. I used the sequence of the LRRK2 protein as a query and used the PDB IDs for the structures as the Entrez queries.
The bad news was that the 3D structures did not contain the mutation site.
The good news was the image of the protein showing conserved domains, showed me the G2019 mutation maps in a conserved domain.
This meant that I could probably find a comparable structure in the conserved domain database and use Cn3D to import and align the region from LRRK2.
So I did.
I clicked the boxes in the image until I got to the conserved domain database, then I downloaded 5 of the aligned structures and imported and aligned the protein sequence from the LRRK2 protein.
Having the conserved domain and all the annotations helped me find out what happens in the mutation site.
The CDD annotations say that most kinases, are able to add a phosphate group to a tyrosine in this highlighted region (the activation loop). When the tyrosine gets phosphorylated, the kinase changes shape so that ATP and substrates are better able to get to the active site.
The G2019 mutation changes the highlighted glycine to a serine. This adds a new target for phosphorylation since serines can also be phosphorylated.
So that's what seems to be happening in the structure. In any case, I think if you have the mutation, you have a pretty good chance of getting Parkinson's, unless there's some other amino acid change that compensates for the more active kinase.
Whatever is going on, this change definitely seems to be harmful for brain cells. Lui et. al. (3) made transgenic fruit flies with the G2019S mutation. The flies had symptoms that are similar to those in humans with Parkinson's: they couldn't move as well and they died young. The researchers also found that the mutant proteins were selectively killing the cells that make dopamine. Some days it just doesn't pay to be a fly.
- B. Luzon-Toro, E. R. de la Torre, A. Delgado, J. Perez-Tur, S. Hilfiker (2007). Mechanistic insight into the dominant mode of the Parkinson's disease-associated G2019S LRRK2 mutation Human Molecular Genetics, 16 (17), 2031-2039 DOI: 10.1093/hmg/ddm151
- J. Deng, P. A. Lewis, E. Greggio, E. Sluch, A. Beilina, M. R. Cookson (2008). Structure of the ROC domain from the Parkinson's disease-associated leucine-rich repeat kinase 2 reveals a dimeric GTPase Proceedings of the National Academy of Sciences, 105 (5), 1499-1504 DOI: 10.1073/pnas.0709098105
- Z. Liu, X. Wang, Y. Yu, X. Li, T. Wang, H. Jiang, Q. Ren, Y. Jiao, A. Sawa, T. Moran, C. A. Ross, C. Montell, W. W. Smith (2008). A Drosophila model for LRRK2-linked parkinsonism Proceedings of the National Academy of Sciences, 105 (7), 2693-2698 DOI: 10.1073/pnas.0708452105