New & Noteworthy

Retooling Yeast Factories

December 13, 2017

Imagine you own a rifle factory and you want to retool it to make solar panels instead. Or vice versa.


It would be silly to retool a factory one step at a time and yet that is what we do with microbial factories. (Wikimedia Commons)

It wouldn’t make a lot of sense to make small, incremental changes. Make one change, test to see how it works, make a second change, test to see how that works, and so on.

Not only is this inefficient and time consuming, but you may not end up with the best factory for making your new product. The best approach is to rip out the old machinery and replace it all at once. Or at the very least make many changes at a time.

The same sorts of problems exist for cellular factories—microbes we engineer to churn out useful products. Ideally we would want to get to an optimized, retooled cellular factory as soon as possible by making as many changes as we can at once. And yet, that is not how we currently do it.

Most of the time when we repurpose a microbial organism like Saccharomyces cerevisiae to make something new, we do it the slow way. We either do a screen to find genes that affect the process or change a known gene and analyze the results. We then use this mutant strain as a starting point to repeat the process. Over many cycles, we hopefully get to a yeast strain that can make something new or more efficiently.

Not only is this time consuming and inefficient, but we might miss some synergies that can happen when two or three genes are tweaked at once. This is no way to retool a factory!

In a new study out in Nature Communications, Lian and coworkers came up with a clever way to make many changes all at once in yeast. And they aren’t limited to just deleting genes either. They can transcriptionally activate, transcriptionally interfere, or delete specific genes. Which is where they got their clever acronym for their process—CRISPR-AID.

As you can tell from the name, their system uses the seemingly ubiquitous gene-editing tool CRISPR. And they use it in some very cool ways to get all three processes happening in a cell at the same time with different genes.

Conventional CRISPR works through a nuclease being guided by an aptly named guide RNA (gRNA) to a precise place in cell’s DNA through base pairing between the DNA and the gRNA.  Once there, the nuclease makes a double-stranded break and at least in yeast, the cell invariably repairs it using homologous directed repair (HDR). (Other eukaryotes tend to “fix” their DNA with the more error prone non-homologous end joining [NHEJ] pathway. Yes, yeast is superior in yet another way!)

Scientists have been able to tweak this system to activate or repress gene transcription by knocking out the nuclease activity of the CRISPR protein and adding back either an activation or repression domain. Now they have an RNA-guided, sequence specific transcription activator or repressor.

This is a great tool chest of gene editing tools but there is one problem—how to get the right protein to the right spot when all three use a gRNA for specificity. Like Mork from Ork turning to Mindy to help guide him on Earth, Lian and coworkers turned to a PAM to help guide these proteins to the right place in the yeast genome.

The protospacer adjacent motif (PAM) is a small part of the gRNA that binds the DNA and helps to pry it open so the rest of the gRNA can bind. Every gRNA has to have a PAM or the CRISPR protein won’t bind.

It turns out that different bacteria have different CRISPR systems that use different PAMs. The idea then is to use a CRISPR protein from different bacteria for each of the three different functions of the CRISPR-AID system.

The most common form of CRISPR protein is Cas9 from Streptococcus pyogenes (Sp), which has NGG for its PAM sequence. Lian and coworkers used a nuclease deficient form of this protein attached to a repressor domain to create their transcription interference tool dSpCas9-RD1152. (The d means the nuclease in inactivated, the Sp is the bacteria it came from, and the RD1152 is the repression domain.)


Like actress Pam Dawber’s character Mindy helping Mork find his proper place on Earth, so too does the PAM sequence guide the CRISPR protein to its proper place in the genome. (Joe Haupt)

For their deletion tool they turned to the Staphylococcus aureus (Sa) Cas9, SaCas9, which has NGRRT or NGRRN for its PAM sequence.

They turned to a different CRISPR protein, Cfp1 from Lachnospiraceae bacterium ND2006 (Lb), for their activation tool. This protein recognizes TTN as its PAM. They attached an activation domain to a nuclease-deficient form to generate dLbCfp1-VP.

All three proteins were stably integrated into a yeast cell. Now that their toolbox was full and their yeast cell ready, Lian and coworkers set out to show that these tools were effective.

They first used it on a known system, making beta-carotene in yeast. Previous work had shown that increasing the expression of the HMG1 gene, decreasing the activity of ERG9, and deleting ROX1 led to increased beta-carotene production.

They used CRISPR-AID to accomplish all three in one fell swoop. Not only did expression from HMG1 go up, expression from ERG9 go down, and ROX1 get deleted, but the resulting yeast also made more beta-carotene. They got a 2.8-fold increase in beta-carotene compared to only a 1.7 fold increase by hitting just one of the genes.

They next set out to increase expression of recombinant Trichoderma reesei endoglucanase II (EGII). Previous work had identified 14 genes that increased expression when activated, 17 that increased expression when inhibited and 5 that increased expression when deleted. Each of these was done as a single gene experiment.

Lian and coworkers created a library of gRNAs that could target each of the 1620 possible combinations to see if any combination would lead to a synergistic increase in activity. They found a combination that did just that—increased expression of PDI1, decreased expression of MNN9, and deletion of PMR1 led to the highest yield of EGII. The increased expression of EGII was beyond what any of the three did on their own.

So the system appears to be working well. Multiple targets can be upregulated, downregulated or deleted all at once.

While we are getting closer to being able to retool microbial factories as efficiently as brick and mortar ones, we aren’t there yet. That will require a deeper understanding of how a particular microbial organism works. And there is no better understood eukaryote than our old friend S. cerevisiae.

Like a good PAM sequence, Mindy helps Mork find his place in the world.

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

Categories: Research Spotlight

Tags: gene editing , microbial factory , CRISPR