In our previous CFTR-pancreatitis blog post, I discussed our plan of action for developing yeast models of pancreatitis-associated CFTR variants as a part of our Pancreatitis PerlQuest partnership with Mission: Cure. In this post, I will share with you the progress that we’ve made in characterizing and identifying which patient allele models are suitable to advance to a high-throughput drug screen.
Strain generation and assay development
As a reminder, we are modeling the 27 CFTR variants listed in Table 1. These specific CFTR variants were selected for two reasons. First, they are associated with pancreatitis (LaRusch et al., 2014) and second, they are conserved in the budding yeast ortholog YOR1 such that we can make the identical mutation.
Table 1. List of yeast YOR1 models of pancreatitis-associated CFTR variants and their clinical significance
First, we needed to generate 27 single-amino acid mutations in the yeast YOR1 gene. To do this quickly, we made these mutations on plasmids that we can later integrate into the yeast genome. Concurrently, we also generated yor1∆ in yeast to confirmed that it is sensitive to oligomycin as reported in the literature. The reported sensitivity of yor1∆ is in the range of 0.05 – 0.35 µg/mL oligomycin. We tested for sensitivity at 0.2, 0.3, and 0.5 µg/mL in liquid media and determined that 0.3 and 0.5 µg/mL showed good separation of growth between wildtype YOR1 and yor1∆ cells. Yeast cells prefer glucose as the carbon source in the media, but it was previously determined that the oligomycin sensitivity of yor1∆ is more apparent in media containing glycerol and ethanol. Thus, we grew all cells in glycerol and ethanol media with added oligomycin. Cells grow significantly slower in this media than in glucose-containing media because glycerol and ethanol are poor (non-fermentable) carbon sources. Therefore, these growth assays were done in 384-well plates over a seven-day period at 30oC and terminal cell density was determined with a plate reader.
Once the yor1∆ phenotype was confirmed, we independently knocked in each of the YOR1 variants into the yor1∆ background to determine how it affects sensitivity to oligomycin. All variants were generated on a plasmid containing a HIS3 (histidine biosynthesis gene) marker and integrated into the HIS3 locus of the yor1∆ cell. In addition, we included a wildtype YOR1 positive control plasmid and a vector only negative control that were also integrated into the yor1∆ cell.
Natural history study of yeast CFTR-pancreatitis models
We initially tested all strains at 0.3 µg/mL and 0.5 µg/mL oligomycin (Figure 1). As expected, yor1∆ and the previously characterized F670∆ (CFTR F508∆) are sensitive to oligomycin at 0.3 and 0.5 µg/mL. In addition to F670∆, we determined that G307R, R387G, I669∆, G704X, G713D, G713S, D741G, S939X, S964L, R1182X, G1274E, S1258P, G1370D, and N1306K are also sensitive to both 0.3 and 0.5 µg/mL oligomycin.
On the other hand, P67L, R426Q, V449L, I467K, I467T, D606Y, I668V, F670C, F1093L, and D1274N are resistant to oligomycin like wildtype cells. Interestingly, R722T and R1182L are partially defective. Both grew to an intermediate cell density between wildtype and yor1∆ at 0.3 µg/mL, and do not grow at 0.5 µg/mL.
Figure 1. Growth of YOR1 variants in 0.3 µg/mL oligomycin (blue) or 0.5 µg/mL oligomycin (orange). Cells were grown in 384-well plates in media containing glycerol and ethanol as carbon sources at 30oC for seven days. Growth is displayed as the change in cell density (OD600 reading) between day 7 and the initial day 0 reading. Each bar represents the mean of three independent samples
To determine if the oligomycin-sensitive variants might grow at a lower oligomycin concentration, we retested them at 0.2 µg/mL. We also retested the oligomycin-resistant variants at 0.65 µg/mL to determine whether they show sensitivity at a higher concentration. We found no difference in growth of all strains at 0.2 and 0.65 µg/mL oligomycin compared to 0.3 and 0.5 µg/mL. Therefore, we’re showing only the data for 0.3 and 0.5 µg/mL oligomycin doses.
Conservation between CFTR and YOR1
We expected a strong positive correlation between oligomycin sensitivity and variants that are associated with cystic fibrosis (CF), because CF-associated variants have mutations that severely affect protein function. This appears true for most cases: G307R, I669∆, F670∆, G704X, G713D, G713S, R722T, S964L, R1182X, S1258P, G1274E, and N1306K are CF-associated and oligomycin-sensitive (Table 2). Only P191L, I467K, and I467T are reported to be associated with CF, but are not sensitive to oligomycin. One explanation is that P191 and I467 are not as highly conserved over evolution as the other residues.
On the other hand, R387G, D741G, S939X, R1182L, and G1370D are not associated with CF based on LaRusch et al., but are sensitive to oligomycin (Table 2; CFTR2 does classify G1370D as CF causing). This indicates that these variants do in fact compromise protein function. They may not be severe enough to cause CF, but may be sufficiently defective to explain their association with pancreatitis. A confounding factor is that humans carry two copies of each gene as well as a genetically diverse background that may affect the severity of a given variant. We have carried out our experiments in yeast cells that are genetically identical except for the single amino acid variant in YOR1, and in haploid cells that contains only a single copy of YOR1. Therefore, we can determine the effect of any single variant on Yor1 protein function.
Table 2. Summary of the oligomycin sensitivity yeast models of pancreatitis-associated CFTR variants. Bold YOR1 variants are oligomycin sensitive, gray YOR1 variants are oligomycin resistant, and orange YOR1 variants shows a dose-dependent sensitivity to oligomycin. ++ indicates wild type level growth in oligomcyin, + indicates intermediate growth, and – indicates oligomycin sensitivity
Drug discovery screen: 17 of the 27 yeast CFTR-pancreatitis models show growth defects
F670∆ is well characterized and we will use it as a control in our high-throughput 20K compound discovery library screen. We will also screen an additional variant. Similar to F670∆, 14 of the other 26 variants are intolerant of oligomycin: G307R, I669∆, G704X, G713D, G713S, S964L, R1182X, S1258P, G1274E, N1306K, R387G, D741G, S939X, and G1370D. Any of these could be advanced to a screen, although we would expect a low hit rate because of the severity of the baseline growth defect. However, R722T and R1182L are interesting in that they have a less severe phenotype that might be more amenable to rescue and therefore, be able to better resolve potential small molecule suppressors. Although both variants are similarly sensitive to oligomycin, there are a couple of key differences between the two. First, R560T (R722T) is associated with severe CF, while R1162L (R1182L) isn’t reported to be associated with CF. Second, R560T (R722T) is in the nucleotide binding domain 1, the same domain as F670. R1162L (R1182L) on the other hand is in the intracellular loop 2. Since we are already screening F670∆, it might be better to screen a variant that is in a different region of the protein in order to widen our pool for small molecules with different suppressive mechanism of actions.
To summarize our CFTR-pancreatitis PerlQuest progress to date, we’ve determined that 17 of the 27 (63%) yeast CFTR-pancreatitis models have phenotypes that could be used in a high-throughput small molecule screen.
Stay tuned for more updates as we screen our CFTR-pancreatitis yeast models!
Meanwhile, head over to the Mission: Cure & National Pancreas Foundation Fundraiser “Party for a Cure” this Friday, October 5 in Los Angeles, CA. Contribute to two organizations who are leading research, advocacy, support and education for individuals and families affected by pancreatitis. While you’re there, introduce yourself and chat first hand about the pancreatitis PerlQuest with our founder and CEO, Ethan Perlstein.
Feature image modified from wallpapersafari.com