New & Noteworthy

Pulsating Yeast

November 19, 2015

---

Sometimes when you get a minor injury, doctors will recommend alternating heat and cold as a therapy. The heat opens things up and the cold shuts them back down again.

Now obviously it would be pretty useless to apply both at the same time. Adding a bit of lukewarm water to an injury is not going to be very helpful at all.

The same thing holds true for many genes. If activators and repressors all turned on at the same time, there wouldn’t be much of an effect on the expression of a gene regulated by both. It is no way to respond to something in the environment!

Instead, if you want a gene to go up and then go back down again, you’d have the activator turn on first, followed by the repressor. Another way to put this is you’d have a pulse where all of the activators activate their genes at once and then stop working followed by a pulse where all of the repressors work at once.

This is exactly what Lin and colleagues found in their recent study in Nature. There they looked at the effect of certain external stimuli on the timing of when the activator Msn2p activated genes and when the repressor Mig1p repressed genes in our favorite yeast S. cerevisiae. These transcription factors coregulate many of the same genes.

The authors found that in the presence of either lowered glucose concentrations or 100 mM NaCl, most of the Msn2p in the cell turned on first followed closely by the Mig1p repressors. In the absence of either stimulus, there was no coordination.

So there does seem to be a carefully choreographed dance between these two transcriptional regulators with these signals. But of course gene regulation is a bit more complex than a sprained ankle.

There may be situations where a cell wants both regulators to do their jobs at the same time. Sometimes lukewarm water may be just what the doctor ordered.

And this is what Lin and colleagues found with 2.5% ethanol. Under this condition, the pulses of the two regulators overlapped—both were on at the same time. Apparently different stimuli call for different responses which means different timing of transcription factor pulses.

The authors next wanted to get at why Mig1p repression lagged behind Msn2p activation. Since both transcription factors can only enter the nucleus and do their job after they lose a few key phosphate groups, the authors reasoned that perhaps Mig1p dephosphorylation lagged behind that of Msn2p.

They decided to look at the PP1 phosphatase, Glc7p, as previous work had shown that it can indirectly regulate both Msn2p and Mig1p. And indeed, when the authors lowered the expression of GLC7, Msn2p and Mig1p no longer pulsed one after the other at lower glucose concentrations. It looks like Glc7p is a key player in controlling the pulsing of these two regulators.

Even though much of this work was done with synthetic promoters with Mig1p and Msn2p binding sites, the results were not restricted to these artificial constructs. Lin and colleagues found that around 30 endogenous targets also responded to lowered glucose concentrations in a coordinated way just like their synthetic construct. Yeast regulates genes by controlling when activators and repressors pulse.

Finally, all of these studies were done using fluorescent proteins and filming single cells in real time. (Is biology cool or what?) This makes sense because subtle signs of synchronization can be lost when averaged over a large population.

This also allowed the authors to investigate what happens in unstimulated cells. In other words, what happens when both regulators enter the nucleus at the same time? Or if a repressor gets in first?

The first thing they found was that even in the absence of stimulation, there were still pulses. So at seemingly random times, suddenly all of the Msn2p would swoop into the nucleus at the same time and then all leave a short time later. Or the same thing would happen with Mig1p.

If by chance the two entered the nucleus at the same time, both the synthetic reporter and an endogenous gene, GSY1, were not activated. But if Msn2p happens to get in there first, both were activated.

And if the repressor Mig1p managed to get into the nucleus at least 4-5 minutes before Msn2p, activation by Msn2p was muted. The presence of Mig1p beforehand seemed to keep Msn2p from activating coregulated genes to as high a level.

Taken together these results confirm that just like a synchronized swim team, yeast regulates genes by controlling when activators and repressors can work. First there is a pulse where the all of the molecules of a certain activator are primed to do their job and then, after a short time, they all stop doing their job. This can then be followed later by a pulse of repressors shutting it all down.

And this isn’t just in yeast either. For example, these kinds of pulses are important in neuroscience as well.

This work suggests that in dissecting regulatory pathways, researchers may need to pay more attention to the timing of pulses. Then they can see that hot followed by cold makes much more sense than both together.

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

Yeast, the Spam Filter

November 11, 2015

---

Imagine what our email inboxes would look like if we didn’t have spam filters! To find the meaningful emails, we’d have to wade through hundreds of messages about winning lottery tickets, discount medications, and other things that don’t interest us.

When it comes to sorting out meaningful mutations from meaningless variation in human genes, it turns out that our friend S. cerevisiae makes a pretty good spam filter. And as more and more human genomic sequence data are becoming available every day, this is becoming more and more important.

For example, when you look at the sequence of a gene from, say, a cancer cell, you may see many differences from the wild-type gene. How can you tell which changes are significant and which are not?

SuperBud to the rescue! Because many human proteins can work in yeast, simple phenotypes like viability or growth rate can be assayed to test whether variations in human genes affect the function of their gene products. This may be one answer to the increasingly thorny problem of variants of uncertain significance—those dreaded VUS’s.

In a new paper in GENETICS, Hamza and colleagues systematically screened for human genes that can replace their yeast equivalents, and went on to test the function of tumor-specific variants in several selected genes that maintain chromosome stability in S. cerevisiae. This work extends the growing catalog of human genes that can replace yeast genes.

More importantly, it also provides compelling evidence that yeast can help us tell which mutations in a cancer cell are driver mutations, the ones that are involved in tumorigenesis, and which are the passenger mutations, those that are just the consequence of a seriously messed up cell. Talk about a useful filter!

The researchers started by testing systematically for human genes that could complement yeast mutations. Other groups have done similar large-scale screens, but this study had a couple of different twists.

Previous work from the Hieter lab had identified genes in yeast that, when mutated, made chromosomes unstable: the CIN (Chromosome INstability) phenotype. Reduction-of-function alleles of a significant fraction (29%) of essential genes confer a CIN phenotype. The human orthologs of these genes could be important in cancer, since tumor cells often show chromosome rearrangements or loss. 

So in one experiment, Hamza and colleagues focused specifically on the set of CIN genes, starting with a set of 322 pairs of yeast CIN genes and their human homologs. They tested functional complementation by transforming plasmids expressing the human cDNAs into diploid yeast strains that were heterozygous null mutant for the corresponding CIN genes. Since all of the CIN genes were essential, sporulating those diploids would generate inviable spores—unless the human gene could step in and provide the missing function.

In addition to this one-to-one test, the researchers cast a wider net by doing a pool-to-pool transformation. They mixed cultures of diploid heterozygous null mutants in 621 essential yeast genes, and transformed the pooled strains with a mixture of 1010 human cDNAs. This unbiased strategy could identify unrecognized orthologs, or demonstrate complementation between non-orthologous genes.

In combination, these two screens found 65 human cDNAs that complemented null mutations in 58 essential yeast genes. Twenty of these yeast-human gene pairs were previously undiscovered.

The investigators looked at this group of “replaceable” yeast genes as a whole to see whether they shared any characteristics. Most of their gene products localized to the cytoplasm or cytoplasmic organelles rather than to the nucleus. They also tended to have enzymatic activity rather than, for example, regulatory roles. And they had relatively few physical interactions.

So yeast could “receive messages” from human genes, allowing us to see their function in yeast. But could it filter out the meaningful messages—variations that actually affect function—from the spam? 

The authors chose three CIN genes that were functionally complemented by their human orthologs and screened 35 missense mutations that are found in those orthologs in colorectal cancer cells. Four of the human missense variants failed to support the life of the corresponding yeast null mutant, pointing to these mutations as potentially the most significant of the set.

Despite the fact that these mutations block the function of the human proteins, a mutation in one of the yeast orthologs that is analogous to one of these mutations, changing the same conserved residue, doesn’t destroy the yeast protein’s function. This underscores that whenever possible, testing mutations in the context of the entire human protein is preferable to creating disease-analogous mutations in the yeast ortholog.

Another 19 of the missense mutations allowed the yeast mutants to grow, but at a different rate from the wild-type human gene. (Eighteen conferred slower growth, but one actually made the yeast grow faster!)

For those 19 human variants that did support life for the yeast mutants, Hamza and colleagues tested the sensitivity of the complemented strains to MMS and HU, two agents that cause DNA damage. Most of the alleles altered resistance to these chemicals, making the yeast either more or less resistant than did the wild-type human gene. This is consistent with the idea that the cancer-associated mutations in these human CIN gene orthologs affect chromosome dynamics.

As researchers are inundated by a tsunami of genomic data, they may be able to turn to yeast to help discover the mutations that matter for human disease. They can help us separate those emails touting the virtues of Viagra from those not-to-be-missed kitten videos. And when we know which mutations are likely to be important for disease, we’re one step closer to finding ways to alleviate their effects. 

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

New SGD Help Video: Variant Viewer

November 5, 2015

---

Using SGD’s Variant Viewer, you can compare the nucleotide and protein sequences of your favorite genes in twelve widely-used S. cerevisiae genomes. This tool shows alignments, similarity scores, and sequence variants for open reading frames (ORFs) from the different strains relative to the S288C reference genome. Sequence data are derived from Song et al., 2015.

Take a look at our new video tutorial to get started with the Variant Viewer, and let us know if you have questions or suggestions.

SGD Help Video: Mutant Phenotypes

November 4, 2015

---

SGD’s Phenotype pages present detailed information about single mutant phenotypes for a particular gene, along with references for each observation. Phenotype pages are accessible from the ‘Phenotype’ tab of the Locus Summary and is also linked from the Mutant Phenotypes section of the Locus Summary, where the phenotype data are presented in summary form. Data are presented in tabular form on the Phenotype page.

This brief video will give you an overview of the contents and organization of SGD’s Phenotype pages.