February 11, 2015
Finding a needle in a haystack would take a long time and would be very tedious (although it’s been done!) Finding a specific drug to fight malaria by testing the effect of each drug, one at a time, on a purified protein in vitro would be at least as tedious and maybe even more so.
Luckily, we don’t have to sift through a haystack. In a new study in ACS Chemical Biology, Frame and colleagues used our friend S. cerevisiae to find nine drugs out of a collection of more than 64,000 that are promising candidates for stopping the malaria parasite in its tracks. It is as if yeast allowed them to set fire to the haystack and see nine needles gleaming in the ashes.
Malaria is a huge problem for global health. Plasmodium falciparum, the organism that causes malaria, is fast developing resistance to the few effective drugs that we have left.
But P. falciparum has an Achilles heel—it can’t make its own purine nucleotides! Since these are the building blocks of DNA and, obviously, essential for life, if we can keep P. falciparum from being able to take them up, we can kill it.
P. falciparum imports purines via a major transporter protein, called PfENT1, located in the plasma membrane. So a drug that specifically inhibited this transporter could be a good way to attack the pathogen.
It’s possible to assay the activity of the transporter in vitro, adding different drugs one at a time and seeing which inhibits transport. But doing this for thousands of drugs might make you wish you were looking for a needle in a haystack. Frame and colleagues decided to harness the awesome power of yeast genetics to test a very large set of drugs more quickly.
The toxic nucleoside analog 5-fluorouridine (5-FUrd) is taken into yeast cells by the high-affinity uridine transporter Fui1. It kills normal yeast cells, but fui1 null mutant yeast can survive in the presence of 5-FUrd.
The researchers engineered a yeast codon-optimized version of pfENT1 and expressed it in the mutant, restoring 5-FUrd uptake. The nucleoside analog was again toxic to this strain, and the only way the yeast could survive was if the transporter activity of pfENT1 was inhibited.
This system allowed a simple and powerful screen for pfENT1 inhibitors. The yeast strain expressing pfENT1 would be able to grow in the presence of 5-FUrd only if pfENT1 transporter activity was blocked by the drug that was being tested.
Setting up the screen on a large scale, the scientists were able to test 64,560 compounds. They initially found 171 compounds that allowed the yeast to grow. They narrowed these down to 9 compounds that worked well and belonged to different structural classes of chemicals.
Because of the way the study was designed, it was likely that these compounds allowed yeast to grow because they prevented PfENT1 from pumping the toxic 5-FUrd into the cell. But what if the compounds were actually doing something different, and unexpected? To rule out this possibility, the researchers designed a secondary screen for the 9 top candidate drugs.
They used ade2 mutant yeast, which can’t make their own adenine and need to be fed it in order to survive. These mutants can make do with the related compound adenosine, but it can’t normally get inside the cell; yeast doesn’t have a transporter that will take it up. However, PfENT1 can transport adenosine, so ade2 mutants can grow on it if they are expressing PfENT1.
With this system, if the candidate drugs are working as expected, they should prevent yeast growth. And that is exactly what the researchers found. This confirmed that the drugs are working because they specifically inhibit PfENT1 and do not allow growth by some other, indirect mechanism.
To be completely sure of the mechanism, the scientists did a direct test. They found that the nine drugs prevented PfENT1-expressing cells from taking up radiolabeled adenosine.
This was all fine, but the ultimate goal of the study was to affect growth of the malaria parasite. So Frame and colleagues tested the drugs on P. falciparum.
In the presence of any of the nine drugs, the parasite couldn’t take up adenosine and also failed to grow. This even happened when the parasites were grown in medium containing much higher purine concentrations than found in human blood.
Even though PfENT1 was targeted by the drugs, all nine of the drugs also killed Pfent1 null mutants. This suggested that the drugs have a secondary target or targets in addition to PfENT1. This could be a real advantage, because it could help prevent the parasites from developing resistance to the drugs.
All nine of these diverse drugs are promising candidates for the treatment of malaria. And the same approach could be used to find chemicals that affect the function of other transporters from various organisms.
As usual, yeast is providing scientists with streamlined ways to find new treatments for serious human diseases. Instead of tediously rummaging about in a haystack, yeast lets us quickly and easily find the needles we need.
by Maria Costanzo, Ph.D., Senior Biocurator, SGD
Categories: Research Spotlight
April 09, 2012
Genomic scientists are quickly being overwhelmed by all of the data they are generating. As trillions of A’s, T’s, C’s and G’s come pouring out of sequencers all over the world, how is anyone going to make sense of it all?
One idea is to use yeast to quickly figure out what effect certain differences have on a gene’s function. Now this won’t be that useful for differences outside of genes or in genes that aren’t shared by yeast and humans. But that still leaves an awful lot of SNPs that we might be able to better understand using the awesome power of yeast genetics.
In the most recent issue of GENETICS, Mayfield and coworkers use yeast to study a large number of variants in the human cystathione-beta synthase (CBS) gene. They chose this gene because it is involved in the metabolic disease homocystinuria, different variants respond to treatment in unpredictable ways, and it can substitute for the yeast homolog, CYS4.
The hope was that they would be able to group CBS variants based on their phenotype in yeast and that this would let them predict which treatments would work for novel variants. They were definitely able to group variants based on phenotype. Time will only tell whether they can use this to better treat patients who come into the clinic with novel variants of the gene.
They looked at 84 known alleles of CBS that affected an amino acid with a single base pair change (81 were from homocystinuria patients). They grouped these alleles based on growth phenotypes in yeast under varying conditions. For example, they determined how well each grew in the absence of glutathione. Only those alleles that were still functional would support growth. They also varied the amount of glutathione, looked at the effect of heme and vitamin B6, studied metabolite profiles with mass spectroscopy and so on.
From this they were able to group many of the alleles in clinically meaningful ways. This means that when a novel allele comes up in a patient, they can screen it in this yeast assay to see if it falls within a known group. At least 38 never before seen missense mutations have been found in the CBS gene since 2010 and undoubtedly new ones will keep appearing as more DNA is sequenced.
The study also revealed alleles that were more difficult to interpret in this assay. For example, some alleles known to cause disease did not affect yeast growth. This might mean that their particular mutation needs something human and/or patient specific to manifest itself or that the enzyme function is fine but something else is wrong.
This study provided a powerful proof of principle. The next step will be to see how well it works in practice and if any patients can benefit.
Benjamin deals with his homocystinuria
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics