New & Noteworthy

Sen1p Is the Traffic Cop for RNA Polymerase III

August 12, 2022

One way to imagine DNA is as a busy road with a lot of competing traffic. Say, a small village in southern Italy…where someone must mediate conflicts between competing vehicles to avoid disaster.

It turns out that the “someone” in yeast cells is Sen1p. Two recent papers from associated groups describe the intriguing detail of how the Sen1p helicase plays this role for RNA polymerase III transcription. The paper by Aiello et al. in Molecular Cell shows how Sen1p regulates transcription-driven conflicts between the several machineries all engaged with DNA. In the related paper by Xie et al. in Science Advances, the authors show how the Sen1p helicase mediates “fail-safe” methods of transcription termination for RNA Pol III, thereby promoting efficiency and avoiding conflict with other pieces of machinery.

The key conflict preventing RNA Pol III from transcribing noncoding genes is with RNA Pol II, which is busy transcribing coding genes. Aiello et al. show how Sen1p has two strategies for mediating these conflicts, both of which involve interactions between Sen1p and the replisome. One involves temporary release of RNA Pol II from DNA while the other resolves genotoxic R-loops in nascent RNA. Both are critical for preventing genome instability.

In the related paper by Xie et al., the authors focus on how termination of transcription of noncoding genes by RNA Pol III is achieved, and the role that Sen1p plays in termination. They show how Sen1p can interact with all three polymerases and also with the other two subunits (Nrd1p and Nab3p) of the NRD1 snoRNA termination (NNS) complex. More specifically, they show by mutation and co-immunoprecipitation that it is the N-terminal domain (NTD) of Sen1p that interacts with RNA Pol III and the replisome.

The authors use metagene analysis of RNA Pol II distribution at mRNA-coding genes to show how Sen1p can promote the release of RNA Pol II to resolve transcription-replication conflicts (TRCs). They further show how the association of Sen1p with the replisome is required for limiting TRCs at the ribosomal replication fork barrier, and how this action appears redundant with that of RNases H. The cooperation and redundancy in this role are key means to protect genome stability.

Not only is Sen1p required for termination of RNA Pol III transcription, but the authors show how this function is independent of the NNS complex. Unlike resolution of conflicts between RNA Pol II and RNA Pol III, the termination function of Sen1p does not require the replisome.

They asked the question of whether Sen1p acts via the primary termination site for RNA Pol III or, rather, a backup secondary termination that catches errors (i.e., when RNA Pol III reads through a weak termination site). Termination for RNA Pol III employs a tract of T nucleotides (T-tract) in the nontemplate strand and these T-tracts can be relatively weak or strong. When T-tracts prove insufficient to stop the polymerase, Sen1p plays a role by means of secondary structures in nascent RNAs, which act as auxiliary cis-acting elements. This backup method is termed the “fail-safe transcription termination pathway.” The RNA secondary structures are not absolutely required for RNAPIII termination, but can function as auxiliary elements that bypass weak or defective termination signals.

Once more, it is the power of the yeast model that has allowed investigation to such exquisite molecular detail. That cells preserve genomic stability and avoid pile-ups amid so much traffic along DNA remains truly remarkable–even when we know more of how it works.

Categories: Research Spotlight

Tags: replisome, RNA polymerase II, RNA polymerase III, Saccharomyces cerevisiae, transcription, transcription conflicts

Trouble with Triplets

April 06, 2018

---

This Research Spotlight also comes as a video animation. Be sure to check it out!

In the Star Wars movie Attack of the Clones, Padme and Anakin end up fighting Genosians on an automated assembly line. Padme manages to survive because the assembly line is set up in an ordered and predictable way. She knows when to jump, when to pause and so on to survive. If there was more chaos in the line, she might have been killed and then there’d be no Luke or Leia!

If the Genosian assembly line weren’t predictable, Padme might’ve ended up a pancake.

The same kind of thing can happen when cells read genes. An enzyme called RNA Polymerase II (Pol II) slides along the gene, spewing out a long string of messenger RNA behind it.

Just like the assembly line on Genosis, the gene is set up in an ordered and predictable way. Nucleosomes (barrel-shaped clusters of proteins) are spaced along the gene’s DNA to help keep it from getting tangled up. But when Pol II comes hustling along, the nucleosomes need to be carefully removed in front and then replaced after it passes.

Similar to the plight of poor Padme getting hurt if the assembly line were not predictable, genes can likewise get damaged if the nucleosome removal/replacement process goes haywire. In Star Wars, the result of a defective assembly line is damaged droids and TIE fighters. In cells, the result is a changed assembly line—the gene can end up longer! This change can be harmful for both the gene and the cell.

In a new study in GENETICS, Koch and coworkers use our favorite beast, the yeast Saccharomyces cerevisiae, to show that the ISW1 gene is a key player in making sure that the nucleosomes get back on DNA after being taken off. When yeast lack the ISW1 gene, nucleosomes don’t always return after being removed to make way for Pol II. And for some genes, this spells even more trouble.

The genes that have the most problems are those that have something called triplet repeats. Basically, this means the same 3 bases (e.g. CAG) are repeated many times in a row in the gene.

Using PCR, Koch and coworkers showed that a gene with triplet repeats ended up with more of them if the ISW1 protein (Isw1p) wasn’t there to shepherd the nucleosomes properly. And too many repeats in a gene can be harmful–not just to the yeast, but to us as well.

In fact, Triplet Repeat Expansion Disorders happen when the number of repeats increases from one generation to the next.  These diseases are some of the most devastating ones around.

They mostly cause slow but steady degeneration of nerve and brain cells, and the cruel symptoms (loss of body movement control, dementia) often only show up in mid-life. The most well-known is probably Huntington’s disease, which killed folk-singer Woody Guthrie.

So understanding how Isw1p helps keep the Pol II assembly line running smoothly in yeast might help us understand how triplet repeat expansion happens in humans, and may eventually give us ideas how to keep it from happening in the first place. This is especially likely because humans have proteins that are similar to Isw1p and probably do something similar.

To determine how Isw1p regulates nucleosome reassembly, Koch and coworkers used Southern blots to show that yeast that lacked Isw1p couldn’t replace their nucleosomes after Pol II had passed as efficiently as yeast that had the protein. This means that when cells are missing the ISW1 gene, a long stretch of the DNA is left bare after Pol II has passed by.

This stretch of nucleosome-free DNA can end up forming new structures called hairpins that can cause cells to send in DNA repair machinery to deal with it. Unfortunately this machinery isn’t always that great at fixing the DNA. Like the Three Stooges adding more pipe to try to fix a leak, the cell can end up adding more DNA to deal with the hairpin.

Like adding extra DNA, throwing extra pipes at a plumbing leak is a great way to make a bad problem worse.

But for the cell, the results are not as hilarious as they were for Moe, Larry, and Curly. That extra DNA can lead to deadly genetic diseases.

A yeast cell needs Isw1p to keep the cell from bringing in the Three Stooges to mess up its DNA and potentially cause devastating genetic diseases. If it turns out the same is true in people, then once again yeast will have shown us how human genetic diseases might happen. And perhaps provide a target for us to go after to prevent these diseases from happening. #APOYG!

by Barry Starr, Ph.D., Barbara Dunn, Ph.D., and Kevin MacPherson, M.S.

Categories: Research Spotlight

Tags: chromatin remodeling, DNA repair, ISW1, nucleosome, RNA polymerase II

Squeezing Out a Tiny Bubble of DNA

November 20, 2017

---

Have you ever tried to walk back to the lab with too many plates between your thumb and middle finger? And had them be so compressed that they shoot out of your hands, spilling all over the floor? Yeah, me too.

While this situation is bad for researchers, it turns out that the same sort of force may be helpful in cells by forcing the DNA open just a bit to get the ball rolling on transcribing a gene into messenger RNA (mRNA). The DNA is squeezed between two points so that a bubble of DNA pops open, leaving some single stranded DNA for RNA polymerase II (RNAP II) to get a hold of.

In a new study in Nature Structural & Molecular Biology, Tomko and coworkers show that Ssl2p, the double-stranded DNA translocase subunit of the basal transcription factor TFIIH, hydrolyzes dATP or ATP to force open around 5 or 6 base pairs of DNA. Once pried open, this small DNA bubble is expanded to 13 base pairs in the presence of NTP hydrolysis, presumably RNAP II beginning to transcribe the DNA. This stable open complex is now ready for promoter clearance and elongation.

These authors teased apart this mechanism using single-molecule magnetic-tweezers (henceforth referred to as tweezers). The idea is to stretch DNA out between two anchor points, add various components of the RNAP II machinery and various reagents, and to measure the changes in the stretched out DNA. Under the right conditions, these length changes directly reflect DNA unwinding.

Getting a gene read in a eukaryote like Saccharomyces cerevisiae is no easy task. A complicated mass of proteins called the preinitiation complex needs to form on the promoter first.

Tomko and coworkers loaded a subset of this complex, which included the TATA binding protein (TBP), TFIIB, TFIIF, TFIIH, and RNAP II, onto a piece of DNA with a single TATA box-containing followed by a strong initiation site. The authors arranged it so the 2.1 kilobases of DNA they used in their experiments was negatively supercoiled as is found in vivo.

In the first set of experiments, they added NTPs and dATP to their reactions. Most of the DNA did not show any changes as measured by the tweezers, but around 5% did.

There were two distinct populations of DNA in the subset that were active. Around 1/3 of the DNA showed  clear open and closed DNA transitions with the open DNA stretching over about 13 base pairs. The other 2/3 of the DNA showed longer-lived, smaller bubbles of around 5 or 6 base pairs of DNA. The authors interpreted the first as stable open complexes and/or elongation complexes and the second set as bubbles of DNA forced open by the preinitiation complex.

They repeated the experiment in the absence of NTPs and saw only the smaller bits of open DNA. And when they left out both NTPs and dATP, they saw no opened DNA at all.

So the small regions of opened DNA are dependent only on dATP hydrolysis making the Ssl2p subunit of TFIIH, a polypeptide known to hydrolyze dATP, the prime culprit for causing them. Which also makes sense, as Ssl2p is a DNA translocase that has been implicated in previous experiments in forcing the DNA to shorten between the preinitiation complex and downstream DNA. These experiments also showed that NTPs needed to be hydrolyzed to proceed to the next step, a stable open complex.

Tomko and coworkers next wanted to determine if these smaller bits of open DNA play a meaningful role in transcription. They tackled this question with a couple of different experiments that both rely on the idea that opening the 5-6 base pairs of DNA is a necessary first step on the way to transcription.

In the first experiment, the authors saw that these smaller regions of open DNA had longer lifetimes at lower NTP concentrations. Remember, NTP hydrolysis is required to proceed to the next step, the 13 bp, more stable open complex. By decreasing the concentration of NTPs, the authors slowed this step down, causing the preceding step of the smaller 5-6 base pairs of opened DNA to hang around longer.

In the next set of experiments, they engineered DNA bubbles into the DNA that varied between 3 and 12 base pairs and tested what size bubble would allow transcription to proceed in the absence of TFIIH, the basal or general transcription factor they implicated in forcing these DNA bubbles open. Previous work had shown that TFIIH is unnecessary if the DNA has a 12 base pair region that is already opened.

They found that while 3 base pairs of opened DNA was insufficient to allow transcription in the absence of TFIIH (and dATP), an open region of 4 base pairs was. This is consistent with the idea that the 5-6 base pairs of opened DNA they saw in their experiments are relevant and that they are caused by the action of TFIIH.

Reading a gene is nontrivial in eukaryotes like us, and our friend yeast. The DNA is squeezed by TFIIH within the preinitiation complex to open just enough of the DNA for RNAP II to get a toe hold and get transcription rolling. A much better result than when a similar force gets your stack of plates rolling on the lab floor.

Don Ho probably isn’t singing about tiny, Ssl2p-mediated DNA bubbles but they too are fine.

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

Categories: Research Spotlight

Tags: open complex, Preinitiation, RNA polymerase II, SSL2, TFIIH

Creating an Ethanol-Making ‘Super Bowl’ Championship Team

May 24, 2017

---

There are a few ways to turn a failing sports team around. One is to tailor individual training to make each player better. Now, the team is better overall because of the changes each player makes.

Another way to improve a team is to change a player in a key position who makes everyone better. A classic example of this is the American football team, the New England Patriots.  

On September 23, 2001, Drew Bledsoe, then the starting star quarterback of the New England Patriots, took a savage hit from New York Jets linebacker Mo Lewis. The Patriots replaced Bledsoe with his backup, Tom Brady, and some might argue, the team (whom Brady led to their first Super Bowl win that year) and the NFL, has not been the same since.

Quarterback Tom Brady, along with head coach Bill Belichick, makes whomever the New England Patriots bring in better. Wide receivers, tight ends, and running backs can be replaced in the lineup without the team missing a beat. He just makes the players around him better than they might be on another team.

In a new study, Qiu and Jiang take a “Patriots” approach to ethanol production in the yeast Saccharomyces cerevisiae. Rather than improving individual genes on their own, these authors instead decided to “bring in” a new version of RPB7, a gene that encodes a key subunit of RNA polymerase II, the molecular machine responsible for making messenger RNA (mRNA).

They hoped that changing this pivotal transcriptional player would cause lots of other genes to do “better” so that “team” yeast would make a lot more ethanol.  Their hopes were realized in their Tom Brady equivalent—a mutant they called M1. Yeast bearing this mutant RPB7 gene became the Super Bowl champs of ethanol production.

One of the keys to increasing ethanol production in yeast is to find strains that are more tolerant of high levels of ethanol. The more ethanol they can withstand, the more they can make.

These authors used error prone PCR mutagenesis of the RPB7 gene to find their game-changing mutant. They then took their library of ~10⁸ clones and cultured them in increasing amounts of ethanol, selecting for more ethanol-resistant strains.

After 3-5 rounds of subculture, they plated the cells onto media containing ethanol. Around 30 colonies were picked and sequenced with the best mutant being the one with two mutations—Y25N and A76T. They named this mutant M1.

This mutant grew a bit better than the parental strain background, S288C, in the absence of ethanol, but where M1 really shined was when ethanol was around. It grew around twice as fast in 8% ethanol and could grow at 10%, a concentration that completely inhibited the parental strain from growing.

Being able to withstand high levels of ethanol is important, but it isn’t all that yeast have to deal with. There are multiple other stressors around when you are swimming in 20 proof media.

For example, yeast can suffer from high levels of reactive oxygen species (ROS). M1 not only tolerated 3.5 mM hydrogen peroxide, a proxy for ROS, better than the parental strain, but it also had around 37% of ROS levels inside cells than that of the parental strain. M1 can deal with high levels of ethanol and ROS.

The authors then tested how this mutant dealt with other potential fermentation problems. For example, acetate, a fermentation byproduct, and high levels of NaCl both inhibit yeast growth. M1 tolerated 80 mM acetic acid and 1.5 M NaCl better than the parental strain did.

M1 appeared to be a champion mutant for making ethanol, and the fermentation studies bore this out.

Under a wide variety of conditions, M1 outperformed the parental strain in terms of growth rate, cell mass, and amount of ethanol made. For example, after 54 hours, yeast containing the M1 mutation of RPB7 managed to make 122.85 g/L of ethanol, 96.58% of the theoretical yield. This is a 40% increase over the control strain. Quite the ethanol producer!

Finally, Qiu and Jiang used microarray analysis of the parental and M1 strains at high levels of ethanol to discover the genes that M1 affected. They found 369 out of a total of 6256 genes behaved differently between the two strains. Of the 369, 144 were up-regulated and 225 were down-regulated.  

I don’t have time to go over all the genes they found but a great many of them make sense. As the authors write, “…a significant set of genes are associated with energy metabolism, including glycolysis, alcoholic fermentation, hexose transport, and NAD⁺ synthesis.”  M1 seems fine-tuned for making ethanol.

A mutant subunit in RNA polymerase II has made yeast better at making high levels of ethanol, most likely by affecting many key genes at once. It is a fascinating way to quickly affect a whole suite of genes involved in a process. In the ethanol-making Super Bowl, we have a new champion yeast strain, M1.

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

Categories: Research Spotlight

Tags: ethanol, fermentation, NFL, Patriots, RNA polymerase II, RPB7, Super Bowl, Tom Brady

A Scientist Sees Transcription

September 23, 2015

---

In the classic Dr. Seuss tale Horton Hears a Who, the elephant Horton thinks he hears voices coming from a speck of dust. He gets into all sorts of trouble over this until all the Whos in Whoville prove they are alive when they all shout at once. Now Horton’s jungle compatriots believe him and Horton can hang out with his new friends.

Horton’s companions never get to hear an individual Who. They are not blessed with Horton’s big elephant ears and so have to just hear all the Whos shouting at once.

Up until recently, we have been in the same situation as the kangaroo and everyone else in the jungle when it comes to transcription in a cell. We can use all sorts of tools to get at what goes on when RNA polymerase II (pol II) gets ready and then starts to transcribe a gene, but we can only get an aggregate picture of lots of cells where it is happening. We can’t hear the Mayor of Whoville amidst all of the other Who voices.

In a new study out in Nature, Fazal and coworkers use the equivalent of elephant ears, optical tweezers, to study the initiation of transcription by purified pol II machinery from Saccharomyces cerevisiae on single molecules. And what they find is that at least for one part of the process, our having looked at things in the aggregate may have fooled us about how the process worked. It was important that we be able to pick out individual voices from the cacophony of the crowd.

Not surprisingly, transcribing a gene is tricky work. It is often split into three steps: initiation, elongation, and termination. And each of these can be subdivided further.

Fazal and coworkers focused on transcription initiation. Previous work had suggested that the process goes something like this:

Top image via Wikimedia Commons

Basically, an alphabet soup of general transcription factors and pol II sit down on a promoter. This complex then pries open the DNA and looks for a signal in the DNA to start transcribing. The polymerase then transcribes short transcripts until it shifts into high gear when it escapes the promoter and enters elongation phase.

This theory comes from the study of transcription in bulk. In other words, it derives from looking at many cells all at once or many promoter fragments in a test tube.

Fazal and coworkers set out to look at how well this all holds up when looking at single genes, one at a time. To do this they used a powerful technique called optical tweezers.

Optical tweezers can “see” what is going on with moving enzymes by measuring the change in force that happens when they move. For this study, the preinitiation complex bound to a longish (2.7 kb) piece of DNA was attached to one bead via pol II, the moving enzyme. The other end of the DNA was attached to a second bead. Each bead is then immobilized using lasers (how cool is that!) and the DNA is stretched between the two beads. Watch this video if you want more details on the technique.

Depending on where you attach the DNA to the bead, you can either track polymerase movement or changes in DNA by precisely measuring changes in the forces keeping the beads in place. Using this technique the researchers found that the bulk studies had done pretty well for most every step. Except for the initial transcribing complex.

The earlier studies had suggested that an open complex of around 15 nucleotides was maintained until elongation began. This study showed that in addition to the 15 base pairs, an additional 32 to 140 base pairs (mean of about 70 base pairs) was also opened before productive elongation could begin. And that this whole region was transcribed.

This result paints a very different picture of transcription initiation. Rather than maintaining a constant amount of open DNA, it looks like the DNA opens more and more until the open DNA collapses back down to the 12-14 base pair transcription bubble seen during elongation.

It turns out that this is consistent with some previous work done in both yeast and fruit flies. Using KMnO₄, a probe for single stranded DNA, scientists had seen extended regions of open DNA around transcription start sites but had interpreted it as a collection of smaller, opened DNA. In other words, they thought they were seeing different polymerases at different positions along the DNA.

These new results suggest that they may have actually been seeing initial transcribing complexes poised to start processive elongation. Seeing just one complex at a time changed how we interpreted these results.

Fazal and coworkers were also able to see what happened to some of the 98% of preinitiation complexes that failed to get started. Around 20% of them did end up with an extensive region of open DNA of around 94 +/- 36 base pairs but these complexes were independent of transcription, as they didn’t require NTPs.

But since this opening did require dATP, they propose that it was due to the general transcription factor TFIIH, a helicase. It looks like in these failed complexes, TFIIH is opening the DNA without the polymerase being present.

A clearer picture of what might be going on at the promoter of genes starts to emerge from these studies. Once around 15 base pairs of DNA are pried open to form the appropriately named open complex, TFIIH unwinds an additional 70 or so base pairs. The polymerase comes along, transcribing this entire region. The whole 85 or so base pairs stays open during this process.

Eventually the polymerase breaks free and the opened DNA collapses back down to around 12-14 base pairs. Now the polymerase can merrily elongate to its heart’s content. Until of course something happens and it stops…but that is another story. 

Categories: Research Spotlight

Tags: optical tweezers, RNA polymerase II, Saccharomyces cerevisiae, transcription

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

Yeast, Smarter than a Train Wreck

January 30, 2014

Imagine you run a railroad that has a single track. You need for trains to run in both directions to get your cargo where it needs to go.

One way to regulate this might be to have the trains just go whenever and count on collisions as a way to regulate traffic. Talk about a poor business model! Odds are your company would quickly go bankrupt.

Another, more sane possibility is to somehow keep the trains from running into each other. Maybe you schedule them so their paths never cross. Or maybe you have small detours where a train can wait while the other passes. Anything is better than regulation by wreckage!

Turns out that at least in some cases, nature is a better business person than many people previously thought. Instead of trains on a track, nature needs to deal with nearby genes that point towards one another, so-called convergent genes. If both genes are expressed, then the RNA polymerases will barrel towards one another and could collide.

A new study in PLoS Genetics by Wang and coworkers shows just how big a deal this issue is for our favorite yeast Saccharomyces cerevisiae. An analysis of this yeast’s genome showed that not only did 20% of its genes fit the convergent definition but that in many cases, each gene in a pair influenced the expression of the other gene. Their expression was negatively correlated: when one of the pair was turned up, the other went down, and vice versa.

One way these genes might regulate one another is the collision model. When expression of one gene is turned up and a lot of RNA polymerases are barreling down the tracks, they would crash into and derail any polymerases coming from the opposite direction. A prediction of this model is that orientation and location matter.  In other words, the negative regulation would work only in cis, not in trans.  Surprisingly, the authors show that this is clearly not the case.

Focusing on four different gene pairs, Wang and coworkers showed that if the genes in a pair were physically separated from one another, their expression was still negatively correlated.  This was true if they just flipped one of the genes so the two genes were pointed in the same direction, and it was still true if they moved one gene to a different chromosome.  Clearly, collisions were not the only way these genes regulated one another.

Using missense and deletion mutation analysis, the authors showed that neither the proteins from these genes nor the coding sequence itself was required for this regulation.  Instead, the key player was the overlapping 3’ untranslated regions (UTRs) of the transcripts.  The authors hypothesize that the regulation is happening via an anti-sense mechanism using the complementary portions of the 3’ UTRs.

This anti-sense mechanism may be S. cerevisiae’s answer to RNAi, which it lost at some point in its evolutionary history.  Given the importance of RNA-mediated regulation of gene expression in other organisms, perhaps it shouldn’t be surprising that yeast has come up with another way to use RNA.  

Instead of RNAi, it relies on genomic structure and overlapping 3’ UTRs to regulate genes.  This may be a bit more cumbersome than RNAi, but at least yeast came up with a more clever system than polymerase collisions to regulate gene expression.  

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

Categories: Research Spotlight

Tags: RNA polymerase II, Saccharomyces cerevisiae, transcription, UTR

A GLAMorous New Role for Prefoldin

October 17, 2013

The prefoldin complex seemed like an ordinary housekeeper. It sat in the cytoplasm and folded protein after protein, just as Cinderella spent her days folding laundry for her stepsisters.

In the old story, the handsome prince searched the kingdom for a girl whose foot would fit the glass slipper. Using this crude screen, he finally found Cinderella and revealed her to be the true princess that she was.

In a new study, Millán-Zambrano and coworkers did essentially the same thing for the prefoldin complex.  They searched the genome of S. cerevisiae for new mutations that would affect transcription elongation. They found the prefoldin complex subunit PFD1 and went on to establish that in addition to its humdrum cytoplasmic role, prefoldin has a surprising and glamorous role in the nucleus facilitating transcriptional elongation.

The researchers decided to cast a wide net in their search for genes with previously undiscovered roles in transcriptional elongation. Their group had already worked out the GLAM assay (Gene Length-dependent Accumulation of mRNA), which can uncover elongation defects.

The assay uses two different reporter gene constructs that both encode Pho5p, an acid phosphatase. One generates an mRNA of average length, while the other generates an unusually long mRNA when fully transcribed. The acid phosphatase activity of Pho5p is simple to measure, and correlates well with abundance of its mRNA. If there is a problem with transcriptional elongation in a particular mutant strain, there will be much less phosphatase activity generated from the longer form than from the shorter one. So the ratio of the two gives a good indication of how well elongation is working in that mutant strain.

Millán-Zambrano and coworkers used this assay to screen the genome-wide collection of viable deletion mutants. They came up with mutations in lots of genes that were already known to affect transcriptional elongation, confirming that the assay was working. They also found some genes that hadn’t been shown to be involved in elongation before.  One of these was PFD1, a gene encoding a subunit of the prefoldin complex. As this deletion had one of the most significant effects on elongation, they decided to investigate it further.

Prefoldin is a non-essential complex made of six subunits that helps to fold proteins in the cytoplasm as they are translated. The authors tested mutants lacking the other subunits and found that most of them also had transcriptional elongation defects in the GLAM assay, although none quite as strong as the pfd1 mutant.

Since prefoldin is important in folding microtubules and actin filaments, the researchers wondered whether the GLAM assay result was the indirect effect of cytoskeletal defects. They were able to rule this out by showing that drugs that destabilize the cytoskeleton didn’t affect the GLAM ratio in wild-type cells, and that mutations in prefoldin subunits didn’t confer strong sensitivity to those drugs.

If prefoldin has a role in transcription, it would obviously need to get inside the nucleus. It had previously been seen in the cytoplasm, but when the authors took another look, they found it in the nucleus as well. Furthermore, Pfd1p was bound to the chromatin of actively transcribed genes! And besides its effect on transcription elongation, the pfd1 mutant has lower levels of RNA polymerase II occupancy and abnormal patterns of histone binding on transcribed genes.

There’s still a lot of work to be done to figure out exactly what prefoldin is doing during transcriptional elongation. Right now, the evidence points to its involvement in evicting histones from genes in order to expose them for transcription. But even before all the details of this story are worked out, this is a good reminder never to assume that an everyday housekeeper is only that.

With the right screen we can find new and exciting things about the most humdrum of characters. A glass slipper screen revealed the princess under that apron and chimney soot. And a GLAM assay revealed the sexy, exciting transcription elongation factor that is prefoldin.

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

Categories: Research Spotlight

Tags: prefoldin, RNA polymerase II, Saccharomyces cerevisiae, transcription elongation

Keeping the Noise Down

May 08, 2013

When you get down to a single cell, things can get really noisy. Instead of the nice, smoothed over data that you see in populations, you see some variation from cell to cell. This is even if all the cells are identical genetically.

Of course this makes perfect sense if you think about it. Part of the variation comes from slightly different environments. Conditions at the bottom of the flask are bound to be different from those at the top! This goes by the name of extrinsic noise.

Another source of variation has to do with levels of reactants within the cell and the chances that they encounter each other so they can react. These effects can be especially pronounced when there aren’t a lot of reactants around. This goes by the name intrinsic noise.

One process with a lot of noise is gene regulation. It is often affected by minor fluctuations in the environment and there are usually just one or two copies of the gene itself. This is the perfect recipe for noise.

The noisiness of gene expression can be split into two steps. One, called burst frequency, reflects how often RNA polymerase sits down and starts transcribing a gene. The second, burst size, has to do with how many proteins are produced each time a gene is turned on.

Of these two processes, the most sensitive to noise is usually burst frequency. A transcription factor (TF) has to find the promoter of the gene it is supposed to turn on and then bring the polymerase over to that gene. This is dependent on the amount of TF in a cell and the number of TF binding sites on the DNA. What this means is that most of the time, genes with low levels of expression tend to be very noisy.

There are some situations, though, where it is very important to have low expression and low noise: for example, where a cell needs at least a few copies of a protein, but can’t tolerate too many. For most promoters, low levels of expression mean high noise, which in turn means there will be some cells that lack this key protein entirely. But a new study out in PLOS Biology shows one way that a promoter can have the best of both worlds.

In this study, Carey and coworkers examined the noisiness of sixteen different naturally occurring promoters in the yeast S. cerevisiae, controlled by the TF Zap1p. This is a great system because the activity of Zap1p is determined by the concentration of zinc in the medium. This means the authors were able to look at the noisiness of these promoters under a broad range of gene activities.

Their research yields a treasure trove of information about the noisiness of these promoters at varying levels of expression. As we might predict, noise decreased at most (11/13) of the reporter genes as more active Zap1p was around. This makes sense, as cell to cell variability will decrease as genes are turned on more often. Higher burst frequency means less noise.

The opposite was true for most (2/3) of the reporters repressed by Zap1p. As more Zap1p was around, transcription of the reporter gene became less frequent, which meant that the noise effects became more prominent.

One of the more interesting findings in this study focused on an exception to this rule. The ZRT2 promoter showed a bimodal expression pattern, as it was activated at low levels of zinc and repressed at high levels. What makes it so interesting is that its noise level stays fairly constant.

As the zinc concentration increases and activity goes up, the noise goes down. This is what we would expect. But when zinc levels get high enough so that the gene is repressed, the noise levels do not increase. They stay similar to the levels seen with the activated gene.

The authors show that this promoter is repressed differently than the other two repressed promoters, ADH1 and ADH3. These promoters are repressed by decreasing the burst frequency: they fire less often when repressed. In contrast, the ZRT2 promoter fires at the same activated rate when repressed, but yields less protein with each firing: repression decreases burst size.

So this is how a cell can manage to get a gene turned on at low levels more or less uniformly through a cell’s population. If it can create a situation where the gene fires a lot but very little protein is made with each firing, then the cell will have relatively constant but low levels of that protein.

This study also provides a new tool for dissecting how a TF affects the expression of a gene. If a repressor decreases expression without an increase in noise, then it is probably affecting burst size. If on the other hand the noise goes up as expression goes down, then the repressor is affecting burst frequency.

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

Categories: Research Spotlight

Tags: cellular noise, RNA polymerase II, Saccharomyces cerevisiae, transcription

Old Genes, New Tricks

March 14, 2013

You can’t teach an old dog new tricks, or so the saying goes. But imagine you found that your old dog knew a complicated trick and had been doing it all her life, right under your nose, without your ever noticing it! You’d be surprised – about as surprised as the Hinnebusch group at NIH when they discovered that some long-studied S. cerevisiae genes had an unexpected trick of their own.

They were working on the VPS* (vacuolar protein sorting) genes. While known for a very long time to be important in protein trafficking within the cell, Gaur and coworkers found that two of these genes, VPS15 and VPS34, play an important role in RNA polymerase II (pol II) transcription elongation too. Now there is an unexpected new trick…like your dog learning to use a litter box!

There had been a few hints in recent years that the VPS genes, especially VPS15 and VPS34, might have something to do with transcription. Following up on these, the researchers tested whether vps15 and vps34 null mutants were sensitive to the drugs 6-azauracil and mycophenolic acid. Sensitivity to these drugs is a hallmark of known transcription elongation factors. Sure enough, they were as sensitive as a mutant in SPT4, encoding a known transcription elongation factor. Further experiments with reporter genes and pol II occupancy studies showed that pol II had trouble getting all the way to the end of its transcripts in the vps mutant strains.

There was a bit of genetic interaction evidence that had suggested that there might be a connection between VPS15, VPS34, and the NuA4 histone acetyltransferase complex. This is important, since NuA4 is known to modify chromatin to help transcription elongation. Looking more closely, the researchers found that Vps34p and Vps15p were needed for recruitment of NuA4 to an actively transcribing reporter gene.

Other lines of investigation all pointed to the conclusion that these VPS proteins have a role in transcription. They were required for positioning of several transcribing genes at the nuclear pore, could be cross-linked to the coding sequences of transcribing genes, and could be seen localizing at nucleus-vacuole junctions near nuclear pores.

One appealing hypothesis to explain this has to do with what both genes actually do. Vps34p synthesizes phosphatidylinositol 3-phosphate (PI(3)P) in membranes, while Vps15p is a protein kinase required for Vps34p function. The idea is that when Vps15p and Vps34p produce PI(3)P at the nuclear pore near transcribing genes, this recruits the NuA4 complex and other transcription cofactors that can bind phosphoinositides like PI(3)P. There are hints that this mechanism may also be at work in mammalian and plant cells.

There’s a lot more work to be done to nail down the exact role of these proteins in transcription. But this story is a good reminder to researchers that new and interesting discoveries may always be hiding in plain sight.

* These genes were also called VPL for Vacuolar Protein Localization and VPT for Vacuolar Protein Targeting

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

Categories: Research Spotlight

Tags: RNA polymerase II, Saccharomyces cerevisiae, transcription, VPS genes

Next