New & Noteworthy
August 26, 2015
If you’ve ever put together something from IKEA you know it can be a bear. So many parts need to connect up together perfectly to build that new bookcase—if you tried to do it all at once you’d go crazy.
Luckily the good folks at IKEA try to make it a bit easier (and more tolerable) by splitting the task into smaller, more manageable pieces. You can concentrate on one part without having to worry about the rest. Once that part is done you can work on the next part and so on. In the end you assemble all the pieces together into your new bookcase.
This is the approach Galanie and coworkers took in their recent Science paper where they engineered our favorite yeast S. cerevisiae to make a couple of different opioids. And it is a good thing they broke this problem up, because it was a way bigger undertaking than anything IKEA might have thrown at them. Engineering this yeast strain was a genetic tour de force.
The authors coordinated 21 different genes from mammals, plants, bacteria and yeast to get the opiate precursor thebaine made. And the semisynthetic opiate hydrocodone took an extra two genes for a grand total of 23! Trying to do all of these at once might have been very frustrating. Thank goodness they split this Herculean task into six (or seven for hydrocodone) smaller modules.
The first step was to get yeast to make (S)-reticuline, a key intermediate on the way to useful opiates. This took 4 modules made up of 17 different genes: six from rat, six from plants, four from yeast and one from bacteria.
And of course just putting these into yeast all at once would almost certainly have made a whole lot of nothing. Each gene needed to be selected from the right beast and then optimized to work in the yeast strain. Sometimes this meant picking the right variant from the right plant, and sometimes it meant mutating a gene to make it behave better. This all would have been overwhelming if the task weren’t split into four easier sections.
Even with all of their optimization, this iteration only made about 20 μg/liter of (S)-reticuline. They needed yeast to crank out more of this intermediate, so they designed a fifth module.
As its name implies, this “bottleneck” module was designed to overcome bottlenecks in the first four modules. After it was added to the strain, the yeast managed to make 82 μg/liter. This was something they could work with!
Except now they were stuck. They needed (R)-reticuline instead of the S form, but no one knew how poppies managed this feat. The gene that did this job hadn’t yet been discovered.
So Galanie and coworkers rolled up their sleeves and dug through plant transcriptome databases to find the gene they were looking for. They found a likely candidate, synthesized the gene in order to produce the enzyme, tested whether it could transform the S form of reticuline into the R form in vitro, and found that it could.
They could now make the right intermediate, which meant they could make their final module. As its name implies, this “thebaine” module would finally allow them to make the opiate precursor thebaine in yeast. This module consisted of their recently discovered gene and three other plant genes.
They had finally made thebaine from simple sugars in yeast! Except it didn’t work very well at all. There seemed to be a bottleneck right after the (R)-reticuline stage. Back to the drawing board!
Given where the bottleneck was, the researchers guessed correctly that the culprit was the SalSyn enzyme which converted (R)-reticuline to salutaradine. A Western blot showed three distinct forms of this enzyme in yeast and only one form, the lowest molecular weight one, when it was expressed in tobacco. Clearly something was happening to inactivate this protein in yeast.
A close look at the protein suggested yeast was glycosylating positions that it shouldn’t, and site directed mutagenesis of these sites confirmed this. The glycosylation was causing the protein to be sorted incorrectly so that it couldn’t do its job.
Unfortunately just mutagenizing away the glycosylation sites wasn’t good enough, because this severely affected the enzyme’s ability to do its job. So the researchers created a chimeric protein with parts of another P450 enzyme they knew did great work in yeast. After optimizing its codons for yeast, this chimera performed beautifully.
Now, finally, they had a yeast strain that could make thebaine. Not a lot of it, only around 6.4 μg/liter—but amazing nonetheless.
A final module was added that consisted of two plant enzymes that converted thebaine to the drug hydrocodone. This monster strain could crank out around 0.3 μg/liter of hydrocodone. Yes, that is as puny as it sounds; one dose of painkiller for an adult would contain 5 mg of hydrocodone.
To be competitive with poppies, they need a 100,000-fold improvement to around 5 mg/liter. In talking with Dr. Smolke, it sounds like this could happen within a couple of years. After scaling up for production, voila! An entirely new source of opiates for pain relief.
Of course the elephant in the room is a Breaking Bad-esque scene where a yeast biologist grabs ahold of an opiate-producing strain and supplies various cartels with illegal drugs. Our Walter White wannabe wouldn’t be able to use the current strain, as he would need thousands of liters of yeast to produce a single dose of Vicodin.
But this scenario will be a real concern in the next few years. Which is why the Smolke lab has crossed every t and dotted every i in setting up and creating this strain. They have made it as difficult as possible for the wrong people to get their hands on it.
This strain represents a stunning achievement in synthetic biology. Move over poppies, there’s a new opiate producer in town.
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
August 19, 2015
Most people assume that Jesus Christ was born around 1 A.D. or 1 B.C. or something like that. After all, that dating system is based on when he was born. 1 A.D. is by definition “the first year of the Lord.”
And as can sometimes happen when you look that far back in time, he didn’t quite get the date of Jesus’ birth right. In point of fact, Jesus was most likely born between 4 and 6 B.C.
While not nearly as momentous, yeast biologists may have done something similar with the genome of our friend Saccharomyces cerevisiae. For many years scientists have believed that this yeast underwent a whole genome duplication (WGD) event around 100 million years ago.
But if the conclusions reached by Marcet-Houben and Gabaldón in a new PLOS Biology article are correct, then what looks like the WGD event actually happened before there was a S. cerevisiae around to duplicate its genome.
Which would mean that there couldn’t have been a WGD. There is no way for a species to double its genome if it doesn’t yet exist! Instead the authors propose that what looks like a WGD was actually a hybridization of two related yeasts from longer ago than 100 million years.
Once you get past the vocabulary, the idea behind this study is actually pretty easy. Basically, they took pairs of genes that most likely came from a duplication event (ohnologs) and figured out when they diverged away from one another. This is the same idea behind figuring out when chimps and humans shared a common ancestor by comparing homologous genes.
When Marcet-Houben and Gabaldón compared every potential ohnolog in S. cerevisiae, they found that a WGD event could explain the origin of only 15% of these genes. These 15% could be traced back to a time after S. cerevisiae was already around.
The other 85% all looked to have been duplicated before S. cerevisiae yet existed. Which of course means these could not have come from a WGD. A genome that does not yet exist cannot be duplicated. This set of genes must have arisen in a different way.
An origin story that makes more sense than a WGD for these genes is one in which two related species hybridize to form one new species. This kind of thing definitely happens, especially with yeast (and if you like lagers, you can be glad it does!).
Here the idea is that the related species share a subset of their genes from when they had a common ancestor. When these species fuse, the new beast has both sets of genes. A cursory look might suggest that these genes were from a duplication event, especially if there are many large tracts of them. After all, many genes share a lot of homology between species.
It is only with a closer look that you might trace these genes back to a common ancestor that came before the species you are studying even existed. This is, in essence, what Marcet-Houben and Gabaldón found.
From the phylogeny of reference species that they created, the authors were able to get a general idea about which clades these prehybridization species may have come from. The largest peak of duplication from their analysis came from before Saccharomyces split from a clade containing Kluyveromyces, Lachancea, and Eremothecium (KLE). The other major peak came before Saccharomyces separated from a clade that contains Zygosaccharomyces rouxii and Torulaspora delbrueckii (ZT). So the simplest interpretation is that at some point long ago, an ancestor of modern S. cerevisiae was formed from the hybridization of a pre-KLE and a pre-ZT species.
Showing that something like this happened so long ago is fraught with peril. All sorts of things can happen to a genome in more than a hundred million years. And duplicated genes are even trickier because they can mutate at different rates when one copy is gaining a new function. (After all, that is one way that new genes are born.)
So Marcet-Houben and Gabaldón threw everything but the kitchen sink at the genome sequences. They tried at least three different methods for comparing the various sequences, using alternative reference phylogenies and a variety of techniques to show what might happen to their results if genes were mutating at different rates.
With each method they got similar results. Many or even most of the “duplication” events happened before S. cerevisiae was even a species. And on top of all of this they were able to show that they got a similar result when they used a known hybrid, S. pastorianus, and its two founding species, S. cerevisiae and S. eubayanus.
All of this taken together argues that S. cerevisiae did not undergo a WGD in the deep, dark past. Instead, it is the result of two closely related species getting together and creating a new species.
Now this does not necessarily mean there was no WGD in our favorite yeast’s past. There are at least two ways that a hybrid species might have formed, and as you can see in this image from Marcet-Houben and Gabaldón’s article, one of them involves a duplicated genome:
In the first scenario, shown on the top left, two diploids of different species fuse together to create a yeast with two sets of chromosomes. Eventually, through mutation, translocation, gene loss and whatever other genome sculpting mechanisms are handy, the yeast ends up with double the number of chromosomes of its predecessors.
In the second possibility, shown on the bottom left, two haploids fuse. This fused yeast then undergoes a whole genome duplication and then goes through similar processes as the first model to get to the current genome.
So, although there may have been a WGD, it looks unlikely that it happened to S. cerevisiae. Just as the placement of historic events in our calendar changes when more information is available, the generation of genome sequences for more and more yeast species and new methods for analyzing them are giving us deeper insight into the history of our friend S. cerevisiae.
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
August 13, 2015
For thousands of years, humans have used yeast as an essential part of recipes for bread, wine, and beer. But now we’re turning the tables on yeast. Instead of creating recipes with yeast, researchers are creating recipes for yeast.
Now of course we have been making minor tweaks here and there for years. But what we’re talking about now is changing out the whole recipe book: creating a whole new genome for S. cerevisiae.
The Synthetic Yeast 2.0 Project (Sc2.0) has the ambitious goal of re-designing and synthesizing the entire yeast genome, some 12 million base pairs. Along with the scientific challenges, the researchers face some tricky ethical issues as well. After all, they’re creating the blueprint for an entire living eukaryotic cell!
Fortunately, the Sc2.0 researchers have thought long and hard about these issues. They’ve issued a statement of ethics and governance in a new article in GENETICS that also reviews the current regulatory and ethical landscape for synthetic biology. The statement by Sliva and colleagues sets the course for Sc2.0 and serves as a model for oversight of other synthetic biology projects.
We wrote about the science in this space before, when the project published its first major milestone, the synthesis of chromosome III. It’s fascinating stuff: the scientists are not only re-synthesizing the genome, but are re-designing it to be leaner and more useful in the lab. They’re adding features like loxP sites that can be used to alter the structure of the genome for evolution experiments, and engineering the tRNA genes so that one codon can be repurposed to code for a novel amino acid.
But even though these are seemingly benign changes to a relatively harmless beast, there are ethical issues inherent in modifying a living organism in such a big way. While the authors focus on Sc2.0, the issues they discuss are relevant to other synthetic biology projects that combine genes from several organisms in novel pathways, such as the efforts to create an opiate biosynthetic pathway in yeast.
While we can only touch on the highlights of their statement here, one of the principles most strongly emphasized by Sliva and colleagues is that all Sc2.0 work will be done for peaceful purposes that benefit society. To promote transparency, they are making outreach a priority, engaging with and educating the public about the project. Sc2.0 will be a public resource, with no intellectual property rights or restrictions on data or materials.
The researchers are also committed to safety. They have engineered multiple auxotrophies into all working strains so that they need special media to survive, even though it seems unlikely that an Sc2.0 strain on the loose would be harmful or would have a competitive advantage over wild strains. And although it’s not required for working with organisms like yeast that are classified “Generally Regarded as Safe” by the FDA, all participants in the project receive biosafety training.
Currently, there is relatively little official policy in place for the field of synthetic biology. Two safety measures currently recommended for DNA synthesis companies by the U.S. Department of Health and Human Services are that the companies check that the sequences they synthesize don’t correspond to toxins or harmful organisms, and that they also verify the identities and institutions of their customers. While compliance with these guidelines is voluntary, the Sc2.0 project has decided to support only companies that follow these safety measures.
To ensure that the policies outlined in their Statement of Ethics and Governance are followed, the Sc2.0 project will maintain an Executive Committee comprised of people both internal and external to the project who have broad expertise in policy, ethics, and science. All of the participants in the project are accountable to this committee, which will actively monitor the work to ensure that the guidelines are followed.
It’s obvious that this is no half-baked scheme, but rather an impressively well-planned recipe for cooking up a yeast cell from scratch. But, we expect nothing less from our friend S. cerevisiae and the talented researchers in the yeast community than to be at the forefront of modern science!
by Maria Costanzo, Ph.D., Senior Biocuration Scientist, SGD
August 6, 2015
Some kids are born troublemakers, wreaking havoc and destruction everywhere they go. They can’t help themselves; it’s in their nature to be that way. But if they have concerned and protective adults in their lives, children can overcome this tendency and grow up to become productive members of society.
Within the cell, ribosomal proteins are problem children. Although they grow up to have essential and productive roles, as newborns they can cause big trouble.
Many of them have highly charged, unstructured regions that give them a tendency to aggregate with other proteins. And they have a complicated journey to adulthood, since ribosome assembly happens in multiple cellular compartments. With an estimated 160,000 ribosomal proteins synthesized every minute in rapidly growing S. cerevisiae cells, these troublemakers could cause major problems if left to their own devices.
To help control this unruly mob, certain proteins in the cell act as designated chaperones for ribosomal proteins. In a new paper in Nature Communications, Pausch and colleagues found that, surprisingly, these specialized nurses catch their client proteins even as they’re being born. They swaddle them from the first moment they start to emerge as nascent proteins, and keep them from causing any harm until they can be delivered safely to their final destination.
The researchers first looked at the proteins Rrb1 and Sqt1. Previous work had suggested they might act as specific chaperones for the ribosomal proteins Rpl3 and Rpl10, respectively. And Pausch and colleagues confirmed these results, showing that TAP-tagged Rrb1 pulled down only Rpl3 and Sqt1 only pulled down Rpl10. Each of these troublesome tots had its own personal chaperone!
But surprisingly, very little of the protein was needed for these chaperones to keep ahold of their respective charges. When the authors trimmed down the ribosomal proteins to shorter and shorter lengths, they saw that just the N-terminal 15-20 amino acids of each ribosomal protein were necessary and sufficient for interaction with its chaperone.
They decided to use X-ray crystallography to look in detail at the Sqt1-Rpl10 interaction. First they determined the crystal structure of Sqt1 on its own, and found that it forms an eight-bladed WD-repeat beta-propeller, looking much like a round electric fan. The amino acids positioned on the surface of the blades are negatively charged.
Next, the authors co-crystallized Sqt1 with a peptide corresponding to amino acids 2-15 of Rpl10. The structure showed that the positively charged peptide was cradled in the negatively charged surface.
To test whether these charged residues were important for the interaction, they mutated the charged residues of Sqt1 and of the peptide and combined them in various ways. Sure enough, changing the charged residues of either partner disrupted or diminished the interaction.
Pausch and colleagues went on to test whether those same charged residues are important in vivo. An sqt1 mutation changing glutamate residue 315 to lysine (E315K), that abolished the Sqt1-Rpl10 interaction in vitro, was lethal for yeast cells, confirming the importance of the interaction.
The researchers also detected many allele-specific genetic interactions between the charged residues of the two proteins, and even found that switching the charges in an interacting pair of amino acids (changing an Sqt1 residue to a positive charge and its Rpl10 binding partner to a negative charge) would improve growth compared to either single mutant.
The lethality of that sqt1-E315K mutation, and even the lethality of an sqt1 null mutation, were weakly suppressed by overproduction of Rpl10. So yeast cells can get by (just barely) with an un-chaperoned Rpl10, as long as there’s enough of it around. This result also confirmed that Rpl10 is the only client of Sqt1.
As yet another verification that Sqt1 acts as a chaperone, the authors looked to see what happens to the Rpl10 protein in sqt1 mutants. If cells carrying wild-type SQT1 are lysed and separated into a pellet and supernatant, most Rpl10 spins down in the pellet but a significant amount is soluble in the supernatant. However, if the cells carry any of several sqt1 mutant alleles that alter the charged residues and diminish the interaction with Rpl10, all of the Rpl10 is found glommed together in the pellet.
The two chaperone-ribosomal protein interactions that Pausch and colleagues investigated, Sqt1-Rpl10 and Rrb1-Rpl3, both involved the extreme N termini of the ribosomal proteins. Previous studies had also shown that two other chaperones for ribosomal proteins, Yar1 and Syo1, also interact with the N termini of their clients. So the authors wondered whether interactions between ribosomal proteins and their chaperones might even start during translation of the ribosomal proteins.
In a final experiment, the researchers treated yeast with cycloheximide to freeze translation and then pulled down each of the four chaperones via affinity tags. Each chaperone specifically pulled down the mRNA encoding its client protein, showing that it was binding to the nascent protein as it first started to emerge from the translating ribosome.
So this study has defined a new step in ribosomal biogenesis. Certain specific ribosomal proteins are such troublemakers that it’s too dangerous for the cell to just release them into the cytoplasm after they’re translated.
Instead, these bouncing baby proteins are caught by their individual nurses before they’re even fully born, and wrapped up to protect both the ribosomal proteins themselves and the rest of the cell. Since ribosomal biogenesis is highly conserved across species and since defects in it are associated with many different diseases, further study of these cellular midwives could have important implications for human health. Perhaps some gentle guidance could help put wayward ribosomes on the right track.
by Maria Costanzo, Ph.D., Senior Biocuration Scientist, SGD