New & Noteworthy

Runaway Polymerases Can Wreak Havoc in Cells

October 16, 2014

A train without working brakes can cause a lot of destruction if it careens off the tracks. And it turns out that a runaway RNA polymerase II (pol II) can cause a lot of damage too.  But it doesn’t cause destruction, so much as disease.

Unlike a train, which has its brakes built right in, pol II has to count on outside factors to stop it in its tracks. And one of these brakes in both humans and yeast is a helicase: Sen1 in yeast and Senataxin, the product of the SETX gene, in humans. 

Mutations in SETX are associated with two devastating neurological diseases: amyotrophic lateral sclerosis type 4 (ALS4) and ataxia oculomotor apraxia type 2 (AOA2), both of which strike children and adolescents.  One idea is that these mutations may short circuit the brakes on pol II, causing it to keep on transcribing after it shouldn’t. And this is just what Chen and colleagues found in a new paper in GENETICS.

The researchers used the simple yet informative yeast model system to look at some of these mutations, and found that they disrupted the helicase function of Sen1 and caused abnormal read-through of some transcriptional terminators.  Looks like bad brakes may indeed have a role in causing these devastating diseases.

Some human proteins can function perfectly well in yeast. Unfortunately, Senataxin isn’t one of those; it could not rescue a sen1 null mutant yeast, so Chen and coworkers couldn’t study Senataxin function directly in yeast. But because Senataxin and Sen1 share significant homology,  they could instead study the yeast protein and make inferences about Senataxin from it.

First, they sliced and diced the SEN1 gene to see which regions were essential to its function. They found that the most important part, needed to keep yeast cells alive, was the helicase domain. But this wasn’t the only key region.

Some flanking residues on either side were also important, but either the N-terminal flanking region or the C-terminal flanking region was sufficient. Looking into those flanking regions more closely, the researchers found that each contained a nuclear localization sequence (NLS) that directed Sen1 into the nucleus. This makes perfect sense of course…the brakes need to go where the train is!  If we don’t put the brakes on the train, it won’t matter how well they work, the train still won’t stop.

These flanking sequences appeared to do more than direct the protein to the nuclear pol II, though.  When the authors tried to use an NLS derived from the SV40 virus instead, they found that it couldn’t completely replace the function of these flanking regions even though it did efficiently direct Sen1 to the nucleus.

Next the researchers set out to study the disease mutations found in patients affected with the neurological disease AOA2.  They re-created the equivalents of 13 AOA2-associated SETX mutations, all within the helicase domain, at the homologous codons of yeast SEN1.

Six of the 13 mutations completely destroyed the function of Sen1; yeast cells could not survive when carrying only the mutant gene. When these mutant proteins were expressed from a plasmid in otherwise wild-type cells, five of them had a dominant negative effect, interfering with transcription termination at a reporter gene. This lends support to the idea that Sen1 is important for transcription termination and that the disease mutations affected this function.

The remaining 7 of the 13 mutant genes could support life as the only copy of SEN1 in yeast. However, 5 of the mutant strains displayed heat-sensitive growth, and 4 of these showed increased transcriptional readthrough.

Taken together, these results show that the helicase domains of Senataxin and Sen1 are extremely important for their function. They also show that Sen1 can be used as a model to discover the effects of individual disease mutations in SETX, as long as those mutations are within regions that are homologous between the two proteins.

It still isn’t clear exactly how helicase activity can put the brakes on that RNA polymerase train, nor why runaway RNA polymerase can have such specific effects on the human nervous system. These questions need more investigation, and the yeast model system is now in place to help with that.

So, although it might not be obvious to the lay person (or politician) that brainless yeast cells could tell us anything about neurological diseases, in fact they can. Yeast may not have brains, but they definitely have RNA polymerase. And once we learn how the brakes work for pol II in yeast cells, we may have a clue how to repair them in humans.

by Maria Costanzo, Ph.D., Senior Biocurator, SGD

Categories: Research Spotlight, Yeast and Human Disease

Tags: ALS, helicase, RNA polymerase II, Saccharomyces cerevisiae, transcription

The Baby Bear of Proteins

July 31, 2013

In the story of Goldilocks and the Three Bears, Goldilocks always likes Baby Bear’s stuff best. Baby Bear has the most comfortable bed, the best porridge, and so on.

The reason Goldilocks likes Baby Bear’s things are that they are just right. They are neither too hard nor too soft, too hot nor too cold, too big nor too little.

Turns out that when it comes to certain proteins, the yeast S. cerevisiae is sort of like Goldilocks…it likes to have them at just the right levels. Too much or too little protein can throw things out of whack.

This idea is supported in a new study in GENETICS, where Sasanuma and coworkers find that a key helicase in yeast, Srs2p, needs to be present in just the right amounts for meiotic recombination to go off without a hitch. In particular, they show that this protein affects meiotic recombination by interfering with the assembly of filaments containing another protein, Rad51p.

Meiotic recombination starts off with Spo11p making a double stranded (DS) break in the DNA. This DS DNA is then trimmed back so that there is a 3′ overhang of single stranded DNA which is then coated with replication protein A (RPA), Rad51p, and Dmc1p. The coated single stranded DNA then invades a stretch of homologous DNA and recombination can begin. 

One of the first things Sasanuma and coworkers did was to show that toying with Srs2p levels has a negative effect on meiosis in general. Too little Srs2p brings spore viability down to 36.8% of wild type, and overexpressing it brings spore viability down to 22.4%.  Clearly Srs2p is a bit like Goldilocks…the amount has to be just right.

The authors next set out to determine how overexpressing Srs2p affects meiosis so profoundly. They showed that too much Srs2p delays the start of meiosis, causes chromosomes to end up in the wrong places, and stunts the repair of DS DNA breaks. Basically, extra Srs2p inhibits meiotic recombination.

They next looked at areas on the DNA where both Rad51p AND Dcm1p were bound, and found that too much Srs2p keeps Rad51p but not Dcm1p off the DNA. When either of these proteins binds to DNA, it forms foci that are visible as dots when the proteins are detected with fluorescent antibodies. While a wild type strain had roughly equal numbers of Dcm1p and Rad51p foci, there were four fold fewer Rad51p foci when Srs2p was overexpressed. Clearly Srs2p was keeping Rad51p-DNA complexes from forming. 

Srs2p can act as a translocase, and it can also bind Rad51p. Sasanuma and coworkers asked which of these functions is essential to its ability to disassemble Rad51p filaments on DNA. Using srs2 mutants that were blocked in just one of these functions, they showed that the translocase mutant was completely unable to remove Rad51p from DNA during meiosis. The Rad51p-binding mutant could still cause Rad51p to dissociate from chromosomes, although at a reduced rate compared to wild type. So the translocase activity is essential, while Rad51p binding is not.

Although it was known that in vitro Srs2p can cause Rad51p-DNA filaments to disassemble, this study is the first to establish that it actually happens in vivo during meiosis. The requirement for the translocase activity suggests that Srs2p may actually move along the filaments as it disassembles them. And this work also shows that just like Goldilocks with her bowl of porridge, the cell needs an amount of Srs2p that is not too big, not too little, but just right.

by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics

Categories: Research Spotlight

Tags: helicase, meiotic recombination, Saccharomyces cerevisiae

Next