New & Noteworthy

Anchors Aweigh for Peroxisomes

August 15, 2013

Suppose you had a fleet of rowboats that you wanted to split up evenly between two shores.  You probably wouldn’t just set them free and hope they drifted to the right places. A better way might be to run some ropes from the distant shore and pull half the rowboats across.  That way you’d be sure to get the distribution of boats that you wanted.

Dividing cells face a similar problem with their organelles.  When cells divide, they need to make sure mother and daughter receive the right number of organelles.  Otherwise one or both could die!

This is obviously too important to just leave to chance, which is why cells have devised ways to precisely control how many organelles end up in mother and daughter.  But not every organelle is divided in the same way.  In our boat analogy, some are pulled over with ropes, others with chains, some with winches and so on. 

We don’t know a lot about how some organelles are distributed between mother and daughter cells.  For example, the details have not been worked out for peroxisomes, the organelles that contain enzymes for beta-oxidation of fatty acids. Until now, that is…

In a recent study in The EMBO Journal, Knoblach and colleagues looked in great detail at how yeast cells distribute their peroxisomes. During budding, some peroxisomes stay tied up in the mother cell while others are transported into the bud and re-tied there.  

The authors found that the structure that ties up peroxisomes is like a rope with hooks at both ends. One hook attaches to the peroxisome, while the other hook attaches to the cortical endoplasmic reticulum (ER) near the cell wall. Surprisingly, the same protein, Pex3p, acts as the hook at both ends of the rope, connecting it to both ER and peroxisomes.

The authors already knew that some peroxisomes stayed anchored around the edges of the mother cell while others were “mobile” and moved to the daughter when yeast cells divided. They also knew that the protein Inp1p was important for anchoring peroxisomes. In the inp1 null mutant all the peroxisomes are mobile and end up in the bud, while overexpressing Inp1p causes all the peroxisomes to be anchored in the mother cell and stay there.

Knoblach and colleagues suspected that Inp1p might act as the rope that tethers peroxisomes. To test this, they fused Inp1p to a protein that sits in the mitochondrial outer membrane, Tom70p. Now peroxisomes in this strain were attached to mitochondria! This established that Inp1p is the tether.

Another major molecular player in this process is Pex3p. The pex3 null mutant phenotype looks a lot like the inp1 mutant phenotype: the mother cell loses all its peroxisomes and they end up in the bud. Pex3p is an integral membrane protein that is channeled through the ER on its way to the peroxisomal membrane – so it can be present in both places. The authors found that both the N terminus and C terminus of Inp1p bind to Pex3p. All this suggested that together, Inp1p and Pex3p might form a structure that links peroxisomes to the ER.

They were able to show that Inp1p and Pex3p interact directly both at the peroxisome and at the ER using a neat trick called bimolecular fluorescence complementation. This simply means that if the two halves of green fluorescent protein (GFP) are brought close to each other, they can fluoresce like the intact protein.  The basic idea is that they fused the first half of GFP with Inp1p and the second half with Pex3p and looked for green spots to turn up in the right place of the cell.  Of course this is easier said than done!

To pull this off, the authors had to first make two haploid strains of opposite mating types. The first had a pex3 mutation that caused all Pex3p to be stuck in the ER, anchoring all its peroxisomes there. It also carried a version of INP1 that was fused to half of GFP.

The second strain had a different pex3 mutation that set all peroxisomes adrift, and this mutant pex3 gene was fused to the other half of GFP. This strain also had an additional marker that made peroxisomes glow red.

When the cytoplasms of these two strains had a chance to mix after mating, the zygote had red peroxisomes with glowing green spots, showing that Inp1p-half GFP from the first strain was interacting with Pex3p-half GFP from the second strain.  Because the Inp1p-half GFP of the first strain was bound to ER-localized Pex3p, and the Pex3p-half GFP of the second strain was localized to peroxisomes, this result showed that Inp1p connects peroxisomes with the ER.

The authors studied the kinetics of this process in a lot more detail and even tracked the wanderings of individual peroxisomes. The model that comes from all this work is that at start of budding, some peroxisomes are bound to Inp1p and others aren’t. Those that aren’t bound to Inp1p move to the bud, and the others stay anchored to the mother cell’s ER. Meanwhile, during budding Pex3p passes through the ER and re-emerges at the ER membrane at the bud cortex. It can then bind Inp1p, which in turn binds to the Pex3p on the surface of the migrating peroxisomes to dock them in the bud.

Not only is this cool just from a basic biology perspective, but it may also help us deal with some human peroxisome biogenesis disorders.  For example, Zellweger syndrome and infantile Refsum disease are associated with specific mutations in the human ortholog of PEX3. Once again, little S. cerevisiae is helping us navigate the inner workings of eukaryotic cells.  

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

Categories: Research Spotlight, Yeast and Human Disease

Tags: peroxisomes, Saccharomyces cerevisiae, yeast model for human disease

Hate the CIN, Love the CINner

March 21, 2013

At first our favorite small eukaryote, S. cerevisiae, might not seem like a great model for cancer studies. After all, budding yeast can’t tell us anything about some of the pathways that go wrong in cancer, like growth factor signaling. And it clearly can’t help explain what happens in specific tissues of the human body. But in other ways, it actually turns out to be a great model.

For example, all the details of cell cycle control were originally worked out in yeast.  And now a whole new batch of genes has been found that influence a phenomenon, chromosome instability (CIN), that is important in both yeast and cancer cells.

As the name implies, chromosomes are unstable in cells suffering from CIN. Big chunks of DNA are lost, or break off and fuse to different chromosomes, turning the genome into an aneuploid mess. And this mess has consequences.

CIN can cause new mutations or make old ones have a stronger effect.  Eventually these mutations can affect genes that are important for keeping a cell’s growth in line.  Once these are compromised, a tumor cell is born. 

Since CIN is pretty common in yeast, we might be able to better understand it in cancer cells by studying it in yeast. The Hieter lab at the University of British Columbia has come up with a powerful screen to get yeast to confess why it CINs. 

A previous study from the group set the stage by finding a large group of mutants that have CIN phenotypes, implying that those genes are involved in keeping chromosome structure stable. In a new paper in G3: Genes, Genomes, Genetics, van Pel et al. uncovered the network of interactions among the genes in this set, using synthetic genetic array (SGA) technology. And they confirmed that the human homologs of some of these genes interact in the same way as in yeast, making them potential targets for cancer therapies.

The idea behind SGA studies is that if two proteins are involved in the same process, then a strain carrying mutations in both of their genes will be much worse off than a strain carrying either single mutation. In the worst case, the double mutant will be dead. This is known as a synthetic lethal interaction.

Yeast is a great model for doing these sorts of studies on a very large scale.  We can construct networks showing how lots of different genes interact, and most importantly, find the genes that are central to many interactions.   These “hubs” are likely to be the key players in those processes.

The researchers looked specifically for interactions between genes that are involved with CIN in yeast and are also similar to human cancer-related genes. They came up with various interaction hubs that will be interesting research subjects for a long time to come. In this study, they focused on one of these: genes involved with the DNA replication fork.

One of these in particular, CTF4, is a hub for both physical and genetic interactions. Unfortunately, Ctf4p doesn’t look like a good target for chemotherapy. It’s thought to act as a scaffold, and lacks any known activity that could potentially be inhibited by a drug. However, the interaction network around CTF4 that van Pel et al. uncovered suggests another way to target this hub. If a gene that interacts with CTF4 itself has a synthetic lethal interaction with another gene, and we could re-create the synthetic lethal phenotype in a cancer cell, we might be able to knock out the whole process. And that is just what they found in human cells.

First the authors identified a couple of human genes that were predicted from the yeast screen to be close to human CTF4 in the interaction network and to have a synthetic lethal interaction with each other. They then lowered the expression of one using small interfering RNA (siRNA), and reduced the activity of the other with a known inhibitor. Neither treatment alone had much effect, but combining them significantly reduced cell viability.

Since cancer cells frequently carry mutations in CIN genes, it should be possible to create a synthetic lethal interaction, guided by the yeast interaction network, where one partner is mutated in cancer cells (equivalent to using siRNA in this study) and the other partner is inhibited with a drug. Since it relies on a cancer-specific mutation, this approach has the potential to selectively target cancer cells while not disturbing normal cells, the ultimate goal for chemotherapy.

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

Categories: Research Spotlight, Yeast and Human Disease

Tags: cancer, DNA replication, Saccharomyces cerevisiae, yeast model for human disease

Ghosts of Centromeres Past

January 28, 2013

Every cell needs to correctly divvy up its chromosomes when it divides.  Otherwise one cell would end up with too many chromosomes, the other with too few and they’d both probably die.

Cells have developed elaborate machinery to make sure each daughter gets the right chromosomes.  One key part of the machinery is the centromere.  This is the part of the chromosome that attaches to the mitotic spindle so the chromosome gets dragged to the right place. 

Given how precise this dance is, it is surprising how sloppy the underlying centromeric DNA tends to be in most eukaryotes.  It is very long with lots of repeated sequences which make it very tricky to figure out which DNA sequences really matter.  An exception to this is the centromeres found in some budding yeasts like Saccharomyces cerevisiae.  These centromeres are around 125 base pairs long with easily identifiable important DNA sequences.

The current thought is that budding yeast used to have the usual diffuse, regional centromeres but that over time, they evolved these newer, more compact centromeres.  Work in a new study published in PLOS Genetics by Lefrançois and coworkers lends support to this idea.

These authors found that when they overexpressed a key centromeric protein, Cse4p (or CenH3 in humans), new centromere complexes formed on DNA sequences near the true centromeres. The authors termed these sequences CLR’s or Centromere-Like Regions.  And they showed that these complexes are functional.

When Lefrançois and coworkers kept the true centromere from functioning on chromosome 3 in cells overexpressing Cse4p, 82% of the cells were able to properly segregate chromosome 3.  This compares to the 62% of cells that pull this off with normal levels of Cse4p.  The advantage disappeared when the CLR on chromosome 3 was deleted.

A close look at the CLRs showed that they had a lot in common with both types of centromeres.  They had an AT-rich 90 base pair sequence that looked an awful lot like the kind of sequence that Cse4p prefers to bind and a lot like the repeats found within more traditional centromeres.  They also tended to be in areas of open chromatin and close to true centromeres. The obvious conclusion is that these are remnants of the regional centromeres budding yeast used to have. 

The hope is that the yeast CLRs might make studying regional centromeres easier.  They are so long and complicated that it is very difficult to pick out which sequences matter and which don’t, but the yeast CLRs could be a simpler model system.  Even better, the CLRs might shed some light on the process of neocentromerization – the formation of new centromeres outside of centromeric regions, which happens a lot in cancer cells. Once again, simple little S. cerevisiae may hold the key to understanding what’s going on in much larger organisms.

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

Categories: Research Spotlight, Yeast and Human Disease

Tags: centromeres, evolution, Saccharomyces cerevisiae, yeast model for human disease

Identifying the Unstable

November 08, 2012

One of the many stumbling blocks in finding better treatments for genetic diseases is figuring out the cause of the disease.  These days, this doesn’t necessarily mean simply identifying the gene with the mutation.  No, nowadays it can mean figuring out what each specific mutation does to the gene it damages.

See, many genetic diseases are not caused by single mutations.  Instead, lots of different mutations can all damage the same gene in different ways.  And each class of mutation may require different treatments.

Cystic fibrosis (CF) is a great example of this.  While most cases of this ultimately fatal disease are caused by mutations in the CFTR gene, not every mutation does the same thing to the CFTR protein.  Because of this, scientists have found different drugs to treat people with different classes of CFTR mutations.

So one drug, Ivacaftor, targets CFTR proteins that can’t open up as well as they should, while another investigative drug, PTC124, targets prematurely stopped CFTR proteins.  Each only treats a specific subset of CF patients who have the correct CFTR mutation.

All of this screams out for a quick and easy assay to figure out how a mutation actually disables a certain protein.  And this is where a new study by Pittman and coworkers just published in the journal GENETICS can help.

The authors have come up with a sensitive in vivo assay in S. cerevisiae that allows scientists to quickly identify mutations that lead to unstable proteins.  This kind of instability isn’t rare in human disease either.  Some of the more famous examples include a kidney disease called primary hyperoxaluria type 1 (PH1), Lou Gehrig’s disease (ALS), Parkinson’s disease, spinal muscular atrophy (SMA), and even some forms of cancer.

The assay basically inserts wild type and mutant versions of the gene of interest into the middle of the mouse dihydrofolate reductase (DHFR) gene, individually adds these chimeric genes to yeast lacking DHFR, and then measures growth rates.  The idea is that if the mutation leads to instability, the DHFR chimeric protein will be unstable too and the yeast will show growth defects under certain conditions.  This is just what they found.

Initially they focused on a gene involved in PH1, the AGT gene encoding alanine: glyoxylate aminotransferase.  They were able to show that disease causing mutations known to affect protein stability affected growth in this assay.  Not only that, but there was a strong correlation between growth and level of protein stability.  In other words, the more unstable the protein, the more severe the growth defect.

They then expanded their assay beyond known AGT mutations.  First they were able to identify a subset of disease-causing AGT mutations as affecting the stability of the AGT protein.  But the assay ran into trouble when they switched to the more stable SOD1 protein.  This protein, which is involved in most cases of ALS, is so stable that mutations that destabilized it were invisible in the assay.  The authors solved this problem by introducing a mutation into DHFR that destabilized it.  Now they could identify mutants that destabilized SOD1.

As a final step, they used their assay to screen a library of stabilizing compounds to identify those that specifically stabilized their mutant proteins.  Unfortunately, in this first attempt they only found compounds that stabilize DHFR, but the assay has the potential to find drugs that stabilize disease-related proteins as well.

Whether or not that potential is realized, this technique should still be a very useful way to determine whether a mutation affects protein stability.  Then, when drugs that stabilize the protein have been found, using this or other screens, doctors will know which patients can be helped by these compounds.  And this will be a boon for scientists and patients alike.

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

Categories: Research Spotlight, Yeast and Human Disease

Tags: protein stability, Saccharomyces cerevisiae, yeast model for human disease

Using Yeast to Discover Treatments for ALS

November 01, 2012

What do Lou Gehrig, Stephen Hawking, David Niven and Mao Zedong have in common?  They all suffered (or in Hawking’s case, continue to suffer) terribly from a disease called amyotrophic lateral sclerosis or ALS.  And now the humble yeast S. cerevisiae may help scientists find new treatments so that others do not need to suffer similarly.

Patients with ALS gradually lose use of their motor neurons and generally die within 3-5 years of diagnosis.  While there are some rare forms that run in families, most are sporadic.  There is no history of the disease in the family and then suddenly, it just appears.

The causes of ALS have remained a mystery for many years but recent work has suggested that RNA binding proteins and RNA processing pathways are somehow involved.  In particular, an RNA-binding protein called TDP-43 appears to be a key player.  Mutations in its gene are associated with ALS, and aggregates of the protein are found in damaged neurons of ALS patients. Unfortunately, since this protein is needed for cell survival it is not an easy target for therapies.  This is where yeast can help.

Scientists have managed to mimic the effects of TDP-43 in yeast.  When this protein is overexpressed, the yeast cells die just like the motor cell neurons do. In a recent Nature Genetics paper, Armakola and coworkers use this model system for finding better therapeutic targets.  And it looks like they may have succeeded.

These authors used two different screens to systematically look for proteins that when deleted or expressed at lower levels rescued yeast overexpressing TDP-43.  They found plenty.  One screen yielded eight suppressors while the other yielded 2,056 potential suppressors.  They decided to focus on one of the stronger suppressors, DBR1.

The first thing they wanted to do was to make sure this wasn’t a yeast specific effect.  If lowering the amount of DBR1 has no effect in mammalian models, it is obviously not worth pursuing!

To answer this question, they created a mammalian neuroblastoma cell line with an inducible system for making a mutant version of TDP-43, TDP-43 Gln331Lys, found commonly in ALS patients.  As expected, these cells quickly died in the presence of inducer.  They could be rescued, though, when DBR1 activity was inhibited with siRNA.  The authors confirmed that decreasing the activity of DBR1 in primary neurons decreased TDP-43 toxicity as well.

So decreasing the amount of DBR1 appears to rescue cells that die from the effects of mutant TDP-43.  This suggests that targeting DBR1 may be useful as a therapy for ALS.  But this study doesn’t stop there.  It also tells us a bit about how lowering DBR1 levels might be rescuing the cells.

DBR1 is an RNA processing enzyme involved in cleaning up the mess left behind by splicing.  It cleaves the 2’-5’ phosphodiester bond of the spliced-out intron (called a lariat).  Previous studies in yeast have shown that when Dbr1p levels are reduced or its catalytic activity is disrupted by a mutation, there is a build up of these lariats. This study showed directly that the accumulated lariats interact with TDP-43 in the cytoplasm to suppress its toxicity. So in ALS, the accumulated lariats may serve as a decoy for the mutant TDP-43 protein, preventing it from binding to and interfering with more essential RNAs.

This last result may also suggest another potential therapy.  If scientists can find other ways to increase the amount of decoy RNA, then they may not need to depend on reducing levels of DBR1.  There may be many possible approaches to soaking up rogue TDP-43.

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

Categories: Research Spotlight, Yeast and Human Disease

Tags: ALS, DBR1, Lou Gehrig's Disease, RNA binding, Saccharomyces cerevisiae, yeast model for human disease

Next