New & Noteworthy

Be Good, For Adaptation’s Sake

December 17, 2012

You may never have herded cows. But in one way or another, you’ve certainly experienced the tragedy of the commons.

When herders get away with cheating, everyone loses. The same thing is often true for yeast.
Image from Wikimedia Commons

This happens when a village shares a pasture that can only feed a certain number of cows. For the system to work, everyone has to cooperate and keep the total number of cows under that limit. But inevitably, one cheater comes along and adds extra cows to his herd. At no immediate cost to himself, he gets all the benefits of the extra cows. But then tragedy kicks in as the pasture is overgrazed until no one can have any cows.

This doesn’t just happen out in the village. We can see it in the overfishing of the oceans, the production of carbon dioxide contributing to global warming, the milking of investors on Wall Street, and many other aspects of modern life. But perhaps surprisingly, we can even see it in cultures of the humble yeast S. cerevisiae.

In a recent issue of the Proceedings of the National Academy of Sciences, Waite and Shou set up a yeast system to look at the factors influencing the tragedy of the commons. In human society, sometimes cheaters cause a collapse, but other times, cooperators get together and exile the cheaters from the village. You might think that since yeast aren’t quite as smart as humans, in yeast “society” the cheaters would always win. However, Waite and Shou found that sometimes the cheaters were marginalized or even driven out, and the cooperators thrived!

To do these experiments, the researchers set up a very clever pasture in miniature. They engineered three strains, each marked with a different-colored fluorescent marker so they could be distinguished from each other.

The two “cooperator” strains needed each other to survive on minimal medium: one required lysine and produced excess adenine, while the other required adenine and produced excess lysine. The “cheater” strain required lysine but it didn’t provide any nutrients. So the cheater needed one of the cooperators to survive, but didn’t contribute anything to the common good.

As we might expect, being cooperative has a cost. The generous production of extra nutrients made the cooperator grow slower than the otherwise identical cheater strain. So you would predict that if you mixed the cooperators and the cheater in equal numbers and grew them together, the cheater would take over and collapse the culture every time.

However, when the researchers mixed all three strains in a 1:1:1 ratio and grew lots of replicate cultures, they found that cheaters didn’t always prosper. Sometimes the nice guys finished first.

Of course some of the cultures did collapse under the influence of the rapacious cheaters. After a while these cultures stopped growing and turned out to be made up of mostly dead or dying cheater cells. The cheaters had taken over the culture, selfishly using up the lysine until eventually there was not enough to continue growing.

But unexpectedly, other replicate cultures were growing much faster, at rates similar to cultures without any cheaters. In these cultures, the two cooperator strains had either dominated the culture or even driven the cheaters extinct!

To explain this, the researchers proposed that the intense selection pressure led to an adaptive race between cooperators and cheaters. In surviving cultures, the rare cooperator with a small advantage had outcompeted the cheaters.

To confirm this, they took a close look at the winning cooperators. They found that the fitness advantage could be inherited, so they used whole-genome resequencing to find out why the cooperators were outcompeting the cheaters. They kept finding mutations in the same five genes.

These genes all made sense, as mutating them would help in an environment with limited amounts of lysine. For example, most of the mutations were found in ECM21 and DOA4. Both of these gene products are important in pathways that break down proteins like permeases. Knocking them out would keep the permeases around longer, making for better lysine uptake. But this newfound advantage did not come without a price.

Dogs playing cards

While cheating is usually a good short term strategy, it doesn’t always work out so well in the long term.
Image from Wikimedia Commons

The researchers tested directly whether the adaptive mutations improved growth in limiting amounts of lysine. Without exception, they did. But almost all these strains grew more slowly in abundant lysine than did their ancestor strains. That explains why these mutant strains only became a significant proportion of the population late in the life of the culture, when lysine levels were very low.

The same mutations can arise in both cooperators and cheaters, of course. But when cheaters become better at growing in low lysine levels, they just become that much better at making themselves extinct. When cooperators get better at growing in low lysine levels, they are better able to keep growing and keep the cheaters at bay.

So the take-home lesson is that cooperation does pay, after all. Especially in a constantly changing environment, cooperators can often win the adaptive race and squelch the cheaters. Maybe we should take a hint from little S. cerevisiae that being kind to each other is not only a nice thing to do, it’s in all of our best interests!

SGD Fall 2012 Newsletter

December 14, 2012

SGD sends out its quarterly newsletter to colleagues designated as contacts in SGD. This Fall 2012 newsletter is also available online. If you would like to receive this letter in the future please use the Colleague Submission/Update form to let us know.

New Strain Sequences Available

December 12, 2012

SGD has incorporated two new S. cerevisiae strain genomes: CEN.PK113-7D and ZTW1. CEN.PK113-7D is a laboratory strain derived from parental strains ENY.WA-1A and MC996A, and is popular for use in systems biology studies. ZTW1 was isolated from corn mash used for industrial bioethanol production in China in 2007. We have also incorporated an updated genomic sequence for W303-derivative K6001, a key model organism for research into aging. The previous version of W303 was an early view courtesy of Markus Ralser. SGD has updated the ORF DNA and protein sequence alignments to include these two additional strains, and the updated W303. We have also included the new genomic, coding, and protein sequences in both the BLAST Search and Pattern Matching (PatMatch) tools. The alignments are accessible via the Analyze Sequence section on the Locus Summary pages. Also available are retrieval and download options.

As Good as the Original

December 7, 2012

Unlike the rest of us, cells are happy with either a copy or the original.

If you got to choose between an original and a copy of that original, you’d undoubtedly choose the original. Because of mistakes during the copying process, the original is bound to be superior.

Cairns was the first scientist to try to apply the same logic to cells like mother cells of the budding yeast S. cerevisiae, or stem cells. The idea is that since these cells make lots of copies of themselves, they might have some mechanism for keeping the original DNA and sending the copies to the new cells. This would protect the original DNA from building up mutations.

While this seems reasonable, the many studies done to date have failed to provide any compelling evidence to support the idea. When scientists look at a cell’s DNA in bulk, they see it randomly dividing between mother and daughter. But this is not the end of the story.

Another idea people have had is that one of the strands of the double helix is kept in its original form in the mother while the other is free to be passed on to the daughter. There is mounting evidence in yeast and mammalian cells that there is some strand-specific selection going on. But in a new study out in GENETICS, Keyes and coworkers show that this selection is not dependent on a strand being an original as opposed to a copy.

If mother cells preferentially keep one of the strands, then the red ones should only end up in the MD and DM cell types in this experiment. The real results were that all four cell populations had the red DNA. (Image from Keyes et al. used with permission from GENETICS)

They use a couple of cool tricks to be able to follow specific strands of DNA in specific cell types in these experiments. First, they engineered a strain that would incorporate BrdU (bromodeoxyuridine) into newly synthesized DNA, so that they could distinguish between the original and copied strands. Second, they used a technique to separate mother cells from daughter cells that involves arresting the mother cells with alpha factor, labeling them with biotin, and then allowing them to divide: the mothers can be pulled away from the daughters by their biotin labels.

As you can see from the image on the right, Keyes and coworkers let a population of biotin-labeled mother cells divide once in the presence of BrdU. Next they separated mothers (M) from daughters (D) and biotinylated the daughters.

Next they let the two populations divide separately one more time in unlabeled thymidine (TdR) instead of BrdU, and separated mothers from daughters again. As shown in the figure, this allowed them to isolate four different cell populations:

1) MM: the original mother cells
2) MD: daughters of the original mothers
3) DM: mother cells derived from the first daughter
4) DD: daughters of the first daughter

The image shows the prediction if the Watson strand (W) is preferentially kept by the mother. As you follow along, remember that the red strand represents the labeled DNA.

If mother cells inherit the Watson strand specifically (W) and the daughter cell always gets the Crick strand (C), then we would predict that only MD and DM cells should have BrdU DNA. The other two cell types should only have unlabeled DNA.

Let’s follow the mother side to see why. The mother cell would inherit the unlabeled Watson strand and a labeled Crick strand because she would keep the original from the first division. The labeled and so copied Crick strand would then be passed preferentially to the daughter.

The same sort of logic applies to the daughter cell. The daughter inherited the labeled Crick strand. Since this daughter now becomes the mother in the next division, the labeled strand now becomes the Watson strand. She would keep this labeled strand and pass the unlabeled Crick strand to her daughter.

These are not the results they got. Instead, all the cell types had pretty close to the same amounts of BrdU. The moms did not preferentially hang on to either the original Watson or Crick strand.

They then followed this up with a separate, similar experiment that looked at each individual chromosome on a microarray. The results were that there was no bias for either strand for any of the chromosomes.

It seems that mom doesn’t hang onto the original DNA even at a single chromosome. In the cases of strand-specific selection that are now being studied, there are most likely other ways to pick a strand that have nothing to do with whether it is an original or a copy. Another great idea foiled by data…

Next Page »