Researchers at the Whitehead Institute at MIT have used a forward chemical genetics approach to identify natural-product-like molecules that could be useful in finding new drug targets for Parkinson's disease. In the study, a team from Susan Lindquist's lab used yeast cells to synthesize an enormous library of cyclic peptides, and then found two that were capable of saving the cells from alpha-synuclein, a toxic protein implicated in the development of Parkinson's disease.
At the core of this method are cyclic peptides, small peptides joined head to tail that can target protein-protein interactions. They can bind to proteins where other drugs can't. Previous work out of the Lindquist lab showed that not only do Parkinson's neurons show a toxic buildup of the protein alpha-synuclein, but also that this can be modeled in genetically engineered yeast and can be switched on to produce the protein to kill the cells.
Taking advantage of this knowledge, a postdoc in Lindquist's lab, Joshua Kritzer, created a library of cyclic peptides that he then inserted into yeast cells. The cells were turned on to produce alpha-synuclein and then monitored to see which died, and which were somehow protected by the introduced cyclic peptide. Of the approximately 5 million yeast cells that were inserted with a cyclic peptide, Kritzer ended up with only two cyclic peptides able to rescue the cells from death. When the scientists expressed these peptides in a C. elegans Parkinson's model, they prevented death of dopaminergic neurons.
"Parkinson's disease is a beast to try to model, and that's a real problem if you want to find out what causes the disease," Kritzer says. "I think it's a really good technique that takes advantage of really important holes in drug discovery technology."
Other landmark work came out of the lab several years ago showing that alpha-synuclein is toxic to cells in that it affects vesicle trafficking and stress in the endoplasmic reticulum. However, Kritzer says, "[Parkinson's] is still such a black box that we thought that we might need nontraditional ways of getting at it. The availability of this yeast model sparked something in both Susan and my mind about using these single-celled organisms to make the molecules for us instead of synthesizing a bunch of molecules and testing them."
Point mutagenesis found that both peptides have a common, four-amino-acid motif and that this particular sequence is very similar to other molecules, including oxidation/reduction and metal-binding molecules. Watching fluorescently tagged cells using electron microscopy, they found that, unlike certain other genes that suppress alpha-synuclein toxicity and rescue vesicle trafficking, cyclic peptides don't. "The significance of that is there's some step downstream of the vesicle trafficking defect that can also be targeted," Kritzer says.
Their screening approach is both faster and cheaper than current methods, he adds. In fact, it was meant to eliminate the use of spatial arrays and high-throughput robotics equipment. "Nobody's going to do a full-fledged, robotic-mediated, high-throughput screen on a [protein-protein] interaction unless it's already been completely validated as a drug target," Kritzer says. While he does do separate genetic screens and works with the neighboring Broad Institute to do small molecule, high-throughput screening, the biggest advantage of this technology is that it can be done in a week and doesn't require the costly setup of a systems biology lab. "One of the benefits of this technology is that it does not require robotics because the screens are pooled — you just take your pooled DNA, you transform it into a pooled liter of yeast, and then you plate the entire pooled liter of yeast and only those that grow, grow, and everything else dies," Kritzer says. "So the yeast does the work for you, and you don't need robotics to separate them out."