New & Noteworthy

Getting the Big Picture from 100 Genomes

May 20, 2015

Like the Peruvian Hairless dog, in some ways the S288C genome looks quite different from other members of its species. Image via Wikimedia Commons

Imagine if aliens visited the earth to learn about dogs, but they stumbled upon a colony of the very rare Peruvian Hairless. Taking a sample for DNA analysis, they would retreat to their home planet, do their studies, and conclude that all dogs had smooth, mottled skin and a stiff mohawk—as well as whatever crazy mutations the Peruvian Hairless happens to carry. 

Until recently, S. cerevisiae researchers have been a bit like those aliens. The genomic sequence of the reference strain S288C was completed in 1996, and for a long time it was the only sequence available. Scientists knew a lot about the S288C genome, but they didn’t have any perspective on the species as a whole.

In the past few years, genomic sequences have become available from a handful of other strains. But now, as described in a new paper in Genome Research, Strope and colleagues have determined the genomic sequences of 93 additional S. cerevisiae strains to make the number an even hundred.

This collection of strains and sequences has already provided new insights into yeast phenotypic and genotypic variation, and represents an incredible resource for future studies. And the comparison with this collection of other strains suggests that in some ways, S288C may be just as unusual as the Peruvian Hairless.

This collection of strains and their sequences gave the researchers a much broader perspective across the whole S. cerevisiae species. It’s as if the aliens discovered Golden Retrievers, Great Danes, Chihuahuas, and more. We only have space here to touch upon a few of the highlights.

First off, they confirmed what many yeast researchers have suspected for a while—S288C is a bit odd.  We already knew that a S288C carries polymorphisms in several genes that affect its phenotype. For example, the MIP1 gene in S288C encodes a mitochondrial DNA polymerase that is less efficient than in other strains, making its mitochondrial genome less stable.

Back when fewer strain sequences were available, it wasn’t clear whether the S288C polymorphisms in other genes like MKT1, SSD1, MIP1, AMN1, FLO8, HAP1, BUL2, and SAL1 were the exception or the rule. Now that Strope and colleagues had 100 genomes in hand, they could see that these differences are indeed peculiar to S288C and its close relative W303.  They might have arisen because of the long genetic isolation of the strains, or because of special selective pressures they faced during growth in the lab.

They also found a lot of variation in how often S. cerevisiae strains have acquired whole chromosomal regions from other Saccharomyces species. This process, known as introgression, happens when related species mate to form hybrids. Stretches of DNA that are transferred in this way are recognizable because gene order is preserved, but all the genes they contain are highly diverged.

The researchers found 141 of these regions containing 401 genes. Many showed similarity to S. paradoxus, which is known to hybridize with S. cerevisiae, but others apparently came from unknown, as yet un-sequenced Saccharomyces species. In a couple of cases that the authors looked at closely, the introgressed genes had slightly different functions from their native S. cerevisiae counterparts.

Another notable finding by Strope and colleagues concerned some genes that exist in multiple copies. The ENA genes, encoding an ATP-dependent sodium pump, are present in 3 copies in S288C (ENA1ENA2, and ENA5), while the CUP1-1 and CUP1-2 genes, encoding metallothionein that binds to copper and mediates copper resistance, are present in 10-15 copies.

To get perspective on a whole species, you need to look at lots of different examples. Image by Sue Clark via Flickr

The sequence coverage in these regions relative to their flanking regions allowed the researchers to see exactly how many repeats are present in each strain. All had between 1-14 copies of ENA genes and 1-18 copies of CUP genes. Interestingly, the strains of clinical origin had significantly higher copy numbers of CUP genes than the non-clinical strains, suggesting that copper resistance is an important trait for virulence.

So, instead of being confined to the S288C genome, S. cerevisiae researchers can now get a much fuller idea of the range of genetic and phenotypic variation within the species. The strains (available at the Fungal Genetic Stock Center), along with their genome sequences (available in GenBank), are an amazing resource for classical and quantitative genetics and comparative genomics.

Unlike those aliens, we won’t end up thinking of yeast as a mostly bald dog with a mohawk. No, we will have a fuller picture of S. cerevisiae strains in all their glory.

A few technical details

In selecting the strains to sequence, Strope and colleagues chose from a wide variety of yeast cultures isolated from the environment and from hospital patients with opportunistic S. cerevisiae infections. But they faced a problem: many of the cultures had irregular numbers of chromosomes or genome rearrangements, which would complicate both interpretation of the sequence data and any future genetic analysis.

To avoid this problem, the researchers selected only strains that were able to sporulate and produce four viable spores—showing that their genomes weren’t messed up. They also wanted strains with no auxotrophies (nutritional requirements), since these can negatively affect growth and complicate the comparison of phenotypes. In some cases, they corrected specific mutations in the strains to increase their fitness.

They ended up with 93 homozygous diploid strains to sequence. Producing paired-end reads of 101 bp, they generated genome assemblies that had 22- to 650-fold coverage per strain.

Because the sequence reads were relatively short, they didn’t provide enough information to assemble the sequence across repetitive regions. So Strope and colleagues used a genetic method to determine gene order. They crossed haploid derivatives of the strains to the reference strain S288C; if their genomes were not colinear with that of S288C, then some of the resulting spores would be inviable.

This analysis showed that 79 of the strains had chromosomes colinear to those of S288C, and allowed assembly of their genomes across multicopy sequences. The remaining strains had chromosomal translocations relative to S288C. Twelve of these carried the same reciprocal translocation between chromosomes 8 and 16.

by Maria Costanzo, Ph.D., Senior Biocuration Scientist, SGD

If it Swims Like a Duck…

May 6, 2015

It may swim like a duck, but this beast is obviously not a duck. Just like the glycine patch of Pxr1 looks like an interaction region when it isn’t. Image by Yotujonoo via Creative Commons

Back in the 1980’s some U.S. politicians were proposing to raise money by something they called “revenue enhancements”. Richard Darman, the budget director at the time, correctly pointed out that a revenue enhancement really is just a tax increase by another name. 

To make his point, he used the expression, “If it looks like a duck and quacks like a duck, it’s a duck.” In other words, just because politicians call it something else, if a revenue enhancement does everything a tax increase does then it really is just a tax increase.

This same reasoning is often used in biology. If two regions of a protein look the same (are homologous) and the proteins do similar things, then the two similar regions do the same thing. Except, of course, when they don’t.

This probably isn’t what Conan O’Brien had in mind when he changed the famous expression a bit to say, “If it looks like a duck and quacks like a duck, it’s a little person dressed as a duck,” but as is often the case with Team Coco, he was right in both biology and life. Not everything that looks and quacks like a duck is a duck, and not every homologous region in proteins that do similar things does the same thing.

Conan’s point is borne out in a new study out in GENETICS where Banerjee and coworkers show that even though the yeast Prp43 RNA helicase shares glycine patches with three of the proteins with which it interacts, this doesn’t mean the glycine patches are used the same way in each case. They may all look and act like ducks, but they are not all ducks! 

Glycine patches are short, glycine-rich protein motifs that are thought to help proteins recognize other proteins or RNAs. Two of the proteins that the researchers looked at, Spp382 and Sqs1, have glycine patches that are only subtly different from that of Prp43. In both of these, the glycine patch is important for interacting with Prp43, but that isn’t its only role. The patches really are ducks in this case, just different kinds of ducks—maybe a mallard and a mandarin duck.

In the case of the third protein, Pxr1, the glycine patch seems to have a completely different (albeit important) role. In this case, it really is a little person in a duck costume!

Prp43 is involved in two different kinds of RNA processing in the yeast cell—pre-mRNA splicing and rRNA maturation. It is one of the few proteins shared between the two complexes involved in each process.

Previous work had shown that different factors in each complex are important for bringing Prp43 to each party. For rRNA maturation, Sqs1 and Pxr1 are the critical players, while for pre-mRNA splicing, Spp382 is key. Since all four proteins share little else beyond a shared weakly conserved, 45-50 amino acid glycine-rich patch, one idea was that all of these proteins use the patch to interact with one another. As is true of much in life, the real answer is a bit more complicated than that.

The first set of experiments was to determine how well Prp43 interacts with each of the other glycine patches, using yeast two-hybrid assays. With full length proteins, the authors found that Spp382 interacted most strongly with Prp43, Pxr1 was the weakest, and Sqs1 was intermediate. They got a similar order of interaction when using just the glycine patches of each of these three proteins, with one small difference: the Pxr1 glycine patch did no better than the empty vector control.

This last result suggested that the glycine patch of Pxr1 was insufficient on its own to interact with Prp43. This was confirmed when they found no difference in the interaction of full length Pxr1 and Pxr1 deleted for the glycine patch.

The Pxr1 glycine patch apparently plays no role in interacting with Prp43—it really isn’t a duck at all. But that doesn’t mean it is dispensable! They showed later that it is critical for snoRNA processing, an important step needed for rRNA maturation.

Mandarins and mallards look like ducks and quack like ducks…and they are ducks. Like these ducks, the glycine patches of Spp82 and Sqs1 look and act like interaction regions, and in fact they are interaction regions. Image via Wikimedia Commons

Of course, sometimes if it looks and quacks like a duck, it is indeed a duck. This was the case for Sqs1 and Spp382.

As shown by two-hybrid and glycine patch swap assays, each of these glycine patches do seem to be important for interacting with Prp43. But each patch was more than just a way for two proteins to hook up.

To show this, Banerjee and coworkers looked for chimeras of Spp382, Pxr1, and Sqs1 that could rescue the lethal phenotype of a Spp382 deletion. First off, they showed that deleting the glycine patch from Spp382 was equivalent to deleting the whole protein—it was a lethal event. And as expected, replacing the Spp382 glycine patch with the one from Pxr1 was still lethal. But the Sqs1 glycine patch was able to rescue the deletion strain although it grew more slowly. So the Spp382 and Sqs1 glycine patches could to some extent substitute for one another.

One way to interpret the difference in growth rates is that it has to do with the fact that the glycine patch of Spp382 bound more strongly to Prp43 than did the one from Sqs1. The glycine patch from Sqs1 can’t fully rescue the Spp382 deletion strain because it is a weaker binder. But a set of mutagenesis experiments suggests that this is not the case.

The authors basically took the Spp382/Pxr1 chimera in which the Pxr1 glycine patch replaced the one from Spp382 and made a series of point mutations that slowly converted the glycine patch back to the one from Spp382. What they found was that the strength of interaction in the two-hybrid assay does not correlate with the level of rescue in the complementation assay. One interpretation is that the Spp382 glycine patch is doing more than recruiting Prp43.

Taken together, these results are a bit of a biological cautionary tale. Just because a protein region looks like another one, do not assume they are doing the same thing. Sometimes what looks and acts like a duck is just a man dressed as a duck.

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

A Factory Without Doors

April 29, 2015

Every factory needs raw materials. Without steel, this is just a pretty factory. And without incoming xylose, a yeast cell set up to make ethanol from biomass is just a pretty cell. Image by Steve Jurvetson via Flickr

Imagine you have built a state-of-the-art factory to make a revolutionary product. The place is filled with gleaming assembly lines and you have hired the best talent in the world to run the place.

Unfortunately there was a glitch in the factory design—the builders forgot to put doors in! Now you can’t get the raw materials in to make that killer product that will change everything.

This may sound contrived or even silly, but it is sort of what is happening in attempts to use yeast to make biofuels from agricultural waste. Scientists have tweaked yeast cells to be able to turn xylose, a major sugar found in agricultural waste, into ethanol. But yeast has no transporter system for this sugar. A bit can get in through the windows, so to speak, but we need to put in a door so enough can get inside to make yeast a viable source for xylose-derived ethanol.

An important step was taken in this direction in a new study by Reznicek and coworkers. They used directed evolution to transform the Gal2 transporter of Saccharomyces cerevisiae into a better xylose transporter. And they succeeded.

After three successive rounds of mutagenesis, they transformed Gal2 from a transporter that prefers glucose into one that prefers xylose. When put in the right background, this mutant protein opens the door for getting yeast to turn agricultural waste into ethanol. Perhaps yeast can help us stave off cataclysmic climate change for just a bit longer. 

The first step was to find the right strain for assaying xylose utilization. They needed a strain lacking 8 hexose transporters, Hxt1-7 and Gal2, because these transporters can take up xylose (albeit at a very low efficiency). Deleting these genes “shuts the windows” and completely prevents the strain from utilizing xylose as a substrate (as well as impairing its ability to use glucose).

This strain was also engineered to be able to utilize xylose. It contained a xylose isomerase gene from an anaerobic fungus and also either overexpressed or lacked several S. cerevisiae genes involved in carbohydrate metabolism. With this strain in hand, the researchers were now ready to add a door to their closed off factory.  

The authors targeted amino acids 292 to 477 in Gal2. This region is thought to be critical for recognizing sugars, based on homology with other hexose transporters. They used mutagenic PCR conditions that generated an average of 4 point mutations in this region, and screened for mutants that grew better than others on plates containing 0.1% xylose.

In their initial screen they selected and replated the 80 colonies that grew best. They then chose the best 9 to analyze further. Of these 9, one mutant which they dubbed variant 1.1 grew better on xylose than a strain carrying wild-type GAL2. Variant 1.1 had a single amino acid change, L311R.

They repeated their assay using variant 1.1 as their starting source. Out of the 14,400 mutants assayed, they found four that did better than variant 1.1. These variants, dubbed 2.1-2.4, all shared the same M435T mutation.  Variant 2.1 had three additional mutations—L301R, K310R, and N314D.

These four new mutants showed better growth on 0.45% xylose, and after 62 hours, all the strains had pretty much used up the xylose in their media. Of the four, variant 2.1 appeared to be the best xylose utilizer: after 62 hours the authors could detect no xylose in the media at all. This variant also grew faster than the others in 0.1% xylose.

Reznicek and coworkers had definitely made Gal2 a better xylose transporter, but they weren’t done yet. They wanted to try to make a door that only let in the raw supplies (xylose) they wanted and not other sugars (glucose).

Up until now, the screens had been done with xylose as the sole carbon source. When they grew variant 2.1 in the presence of both 2% glucose and 2% xylose, they found that it preferentially used the glucose first. Their evolved transporter still preferred glucose over xylose!

Now in some ways this wasn’t surprising, as the mutations had not really affected the part of the protein thought to be involved in recognizing sugars. They next set out to evolve Gal2 so that it would transport xylose preferentially over glucose.

This time they used a slightly different background strain for their screen. This strain, which was deleted for hxk1, hxk2, glk1, and gal1, was unable to use glucose although it could transport it.

They repeated their mutagenesis and looked for mutants that grew best in 10% glucose and 2% xylose. We would predict that any growing mutants would have to transport xylose better than glucose. And this is just what they found.

When they analyzed the mutants, they found that the key mutation in making Gal2 prefer xylose over glucose in the variant 2.1 background was T386A. Based on homology with Hxt7, this mutation happens smack dab in the middle of the sugar recognition part of the protein. Most likely this mutation compromised the ability of Gal2 to recognize glucose, as opposed to improving recognition for xylose.

These experiments represent an important but by no means final step in engineering yeast to make fuel from biomass. We are on our way to a smaller carbon footprint and perhaps a world made somewhat safer from climate change.

First, beer, wine, and bread; next, keeping coral alive and saving countless species from extinction. Nice work, yeast.

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

Sharing the Health

April 22, 2015

When yeast are forced to eat a meager diet, they not only live longer themselves but they also make a mysterious chemical that helps nearby yeast live longer. If they stay away from all-they-can-eat buffets, that is… Image by Andreas Praefcke via Wikimedia Commons

A study published a few years ago made a big splash in the health news by showing that obesity is socially contagious. If one person gains weight, their friends tend to gain weight too—even if they don’t live in the same town! This works the opposite way too: thinner people are more likely to be socially connected with thinner people.

You might think this is because people tend to make friends with others of a similar size, but this doesn’t seem to be the case. The researchers concluded that there is actually a cause-and-effect relationship: we all influence the weight of our friends.

Well, S. cerevisiae cells are not so different. They may not have social lives, but since they can’t move on their own, they do tend to live together in colonies. And within these colonies, they influence each other: not in terms of weight, but in terms of the effect that calorie intake has on the length of their lives.

Turns out that like nematodes, fruit flies and even mice, living on a meager diet makes yeast live longer. And in a new study published in PLOS Biology, Mei and Brenner found that yeast cells actually share the life-extending benefits of calorie restriction with their neighbors, probably via a still-unidentified small molecule.  

Yeast are normally grown in the lab on medium containing 2% glucose. To a yeast cell, this is like an all-you-can eat buffet that goes on for its entire lifetime. Media with a glucose content of 0.5% or less represent a meager diet. But that deprivation comes with a benefit, in the form of an extended lifespan.

Mei and Brenner already had some hints from previous studies that yeast cells might excrete a substance that promoted lifespan extension. To study this systematically, they devised an experiment to test whether mother cells change the media surrounding them as they divide.

The researchers placed individual mother cells in specific spots on Petri plates containing an all-you-can-eat buffet (2% glucose), a restrictive diet (0.5% glucose), or a near-starvation diet (0.2% glucose). They watched as the cells budded, and removed each new daughter cell as it separated from the mother, counting the buds. The lifespan of a mother yeast cell, termed the replicative lifespan, is measured as the number of times she can bud during her lifetime.

After the mother cells had budded 15 times, half of them were physically moved to fresh parts of the same plate, while the other half were left in place. For the mothers on the 2% glucose plates where calories were abundant, the move didn’t change anything. The mothers that were moved had exactly the same replicative lifespan as those that stayed put.

On the plates where calories were restricted, it was a different story. The cells that stayed in place had extended lifespans, as expected under these low-calorie conditions. But the cells that were moved to new locations lost most or all of the life extension—even though calories were still restricted in their new locations. This suggested that the mother cells had secreted a “longevity factor” into the medium surrounding them, which then extended their lifespan when they got older.

There were a couple of metabolites that were prime candidates for the longevity factor: nicotinic acid (NA) and nicotinamide riboside (NR). NA and NR are precursors to nicotinamide adenine dinucleotide (NAD+), a compound that acts as an essential cofactor for many important enzymes. They had already been implicated in lifespan extension because mutating genes involved in their metabolism can affect how long various creatures live.

When the scientists tried supplementing calorie-restricted cells that had been moved to fresh medium with either NA or NR, they found that supplying these metabolites could restore the longevity benefit.  This finding strengthens the idea that NAD+ metabolism is involved.

But was the longevity factor actually NA or NR? To test this, Mei and Brenner grew yeast in liquid media with the different glucose concentrations and then tested for NA and NR in the medium using liquid chromatography-mass spectrometry analysis.  They found that under all the conditions, the amount of NA secreted by the cells didn’t change and secreted NR was undetectable, suggesting that neither was the factor induced by calorie restriction.

To ask directly whether there is a diffusible longevity factor, the researchers grew cells in liquid medium containing 2% or 0.2% glucose until all the glucose was used up, then separated out the cells and freeze-dried the remaining liquid. They suspended the dried “conditioned” medium in water and spread it on plates to repeat the cell-moving assay.

Just like before, cells grown in 2% glucose had the same lifespan after being moved to a fresh spot, and the addition of resuspended conditioned medium to the plate didn’t change that. However, the starved cells grown on 0.2% glucose not only kept their lifespan extension when moved to conditioned media, but actually lived 10% longer compared to starved cells on un-conditioned media that were not moved.

When the researchers dialyzed the conditioned medium so that molecules smaller than 3.5 kDa were lost, the longevity factor was lost too. So it looks to be a small molecule, and of course they are actively pursuing its identity. Intriguingly, this would explain why other scientists have been unable to detect calorie restriction-induced lifespan extension in yeast using microfluidic technology, where immobilized yeast cells are grown with a constant exchange of growth medium. Under these conditions, a small molecule that promotes longevity would be washed away.

So, even though they don’t have Facebook friends, yeast cells influence the health of their peers. Rather than spreading the influence through social interactions as we humans do, they broadcast a chemical that is the key to long life. 

It’s tempting to think that the identity of this chemical will tell us something about human aging. But if this mysterious molecule worked in humans the same way as it does in yeast, people would still have to eat just enough food to stay alive to get the benefits. Still, perhaps the molecule can point us towards finding a treatment that will let us live longer while enjoying lots of good food. We could have our cake and eat it too!

by Maria Costanzo, Ph.D., Senior Biocuration Scientist, SGD

« Previous Page
Next Page »