Unlocking Precision Fermentation Via Protein Engineering
My submission to the Homeworld Collective writing challenge
Precision Fermentation
Yeast is a delicious miracle. Our partnership with these single-celled eukaryotes has produced many of the things that make life worth living, such as bread, beer, and wine. Our centuries of experience using yeast has culminated in the ability to produce these products at an industrial scale.
But perhaps most remarkable is that yeast is finding a new role as a workhorse of synthetic biology. Despite its small size, yeast is a complex organism capable of supporting all sorts of biochemistry. We’ve been using yeast to make alcohol for centuries, why not ask it to make medicine instead?
Along with the perpetual frenemy of the gut e. coli, yeast is the co-star of the field of precision fermentation, which uses genetic modifications to give these microbes the ability to produce a dizzying array of chemicals and biologics. All they ask for in return is a little sugar and a safe place to grow.
It’s hard to overstate how broad these capabilities are. Yeast biochemistry can support a wide range of reactions, and in theory, virtually any organic molecule, protein, or biopolymer could be made by augmenting yeast with new genes. With sufficient scale, precision fermentation could produce medicines, replace our food system, and supplant much of the chemical industry using the same fermentation tank.
Everything made here (source).
As always, the limiting factor is price. Modifying yeast to produce a product is difficult, requiring a team of scientists to determine which genes to add, and a team of engineers to scale the process. Naturally, most precision fermentation companies focus on selling high-value products like medicine. If we want to fully realize the potential of precision fermentation, we need to dramatically reduce costs.
So what are the major cost drivers? Despite all of the fancy staff and biotech that goes into precision fermentation, the real issue is dirty water. Specifically, separating your product from the complex broth that microbes grow in, ideally in a manner that doesn’t disturb your yeasty employees.
Let’s look at how researchers are working on the problem of purifying one important product: proteins.
Overview of the purification process
Separating your desired protein out of a batch of cells is challenging. For one, the media that the cells grow in is filled with a mess of different sugars, proteins, and nutrients. Making matters worse, the cells themselves are teeming with different proteins, some of which end up outside of the cell.
Often, the cells in your bioreactor produce the target internally, so the first step is to break down the cells in a process known as lysis. From a cost perspective, this is already a problem; the cells we worked so hard to modify and grow are already being destroyed. This is why many companies use an extracellular process where proteins get secreted outside the cell, so that the solution containing the protein can be separated from the yeast without destroying them.
Regardless of whether you can save your microbes, we come to the problem of separating the target protein from a complex solution. This usually begins with a bulk filtration step, designed to remove particles that are much larger than our target.
Next we turn to centrifugation, spinning a batch of your solution at incredible speeds to separate products out by density. After spinning, the proteins in the sample will form a band towards the bottom of the tube while the leftover solution can be taken out. The problem with centrifugation is that only a small amount of solution can be handled at any one time, creating a bottleneck in production.
After centrifugation, we use a series of sophisticated purification steps that remove virtually all contaminants from the solution (this often involves column chromatography, which we will discuss in the next section). By this point, you have a very pure solution containing your protein. After a final freeze drying step to remove the water, your protein can be shipped to customers.
You have successfully produced a pure, engineered protein, but at a high cost. Each step requires detailed research to identify the right production conditions, adding time and expense. You can see why precision fermentation is mostly used to produce medicine, few other products have the profit margins to support it!
Of all of these phases, the purification step adds the most cost and complexity, so let’s look at how we can improve it or eliminate it entirely. Ideally, our method achieves high-purity without sacrificing too much of our product in the process. Methods that avoid killing our microbes are an added bonus.
Column Chromatography
One of the most common purification methods is to use a set of techniques known as column chromatography. These techniques flow the solution through a tube filled with a substance that binds each protein in a slightly different way. Proteins with a high affinity for the filter move slowly through the column, while proteins with a low affinity pass right through. As the solution travels the length of the tube, similar proteins gather into bands like runners of different skill levels in a marathon. All we have to do is wait at the finish line to grab only the band containing our desired protein.
Thin-layer chromatography demonstrating the separation of different colored bands (source).
Sometimes we add a protein tag to our desired product and use a column that specializes in grabbing that tag. For instance, immobilized metal ion affinity chromatography appends a repeating histidine unit to the protein which can be retrieved by nickel atoms bound to the column material. The tagged product will stick to the column, while all of the contaminants pass through. Afterwards we can add a special solution to release the protein, leaving us with a very pure product.
Column chromatography is a remarkable technique, able to purify just about anything. For this reason, it has become a standard technique in protein purification, but it also comes at a high cost. The filter material is expensive, the throughput is low, and the tag can interfere with protein function, requiring an additional removal step. Usually, multiple rounds of chromatography are needed, and research has to be done to find the right way to separate your target from a given mixture of contaminants.
Self-cleaving tags
So chromatography can purify a wide range of proteins, if we’re willing to pay the price. Can we do better?
One clever idea is to use special proteins called inteins that can snip other proteins. For this to work, we need our target, an intein, and a tag connected to each other. Once the target is captured on the column, a chemical signal tells the intein to cut itself and the tag away from the target protein. This means that we can use milder cleavage conditions and can avoid the tag removal step.
It’s definitely an improvement, but self-cleaving tags don’t solve our problem on their own. We still have to use expensive chromatography techniques, and designing a new intein for each protein adds more complexity.
Self-aggregating proteins
Self-aggregating proteins hold more promise. One idea is to use inclusion bodies, insoluble aggregates of proteins that cells produce normally. If you can coax the cells to produce the target protein, you can break the cells down, separate out the inclusion bodies, and then get the target out by dissolving the aggregates with a solvent like urea. However, with inclusion bodies, we still have to kill our cells and the refolding process can be difficult; 60% of proteins don’t refold back to their native state on their own and need additional chemical steps to fold properly, if at all.
Building on this, we can turn to controlled aggregation to isolate the product. Similar to inclusion bodies, we design our target protein to clump up into insoluble aggregates, but this time, we make sure the aggregation only happens outside the cell. Now we can avoid breaking down the cells to get the product.
One way to do this is to add elastin-like polypeptides (ELP’s) to the protein. These peptides have a 5 amino acid repeating unit which tends to form aggregates in a specific temperature range. Now we can harvest some of the solution, heat it to the aggregation temperature, and collect the resulting aggregates.
But I think we can do better. ELP’s are finicky because the aggregation temperature depends on the surrounding environment and the protein they’re attached to. Finding the right conditions for aggregation adds time to the development cycle and must be repeated for each target.
Several research groups have used short tags that self-aggregate under various conditions. For our purposes, the trick is to find a tag that only aggregates outside the cell so that we can avoid destroying the yeast. For example, the L6KD peptide (6 leucines followed by a lysine and an aspartic acid) acts like a surfactant, forming small bubbles similar to how soap surrounds grease. As long as the cell pushes these proteins out, the aggregates will form in he media, saving the yeast. Unfortunately, many of these aggregates are pretty small, requiring a centrifugation step to get them out of the solution. This isn’t a deal-breaker, but centrifugation makes it difficult to scale to large volumes.
With this in mind, I want to add a new idea to the mix, inspired by how our blood can form large clots. These clots are formed when a protein called fibrinogen is cleaved by another protein called thrombin. The gel this forms can be pretty large: fibrin glue, a mixture of thrombin and fibrinogen, is used to patch wounds during surgery. If we attach fibrinogen to our desired product and add thrombin to the growth media, a clot should slowly accumulate outside the cells. This clot may grow large enough to be fished out of the fermenter without the need for filtration. From there, retrieving the product is a simple matter of cleaving it away from fibrin using a protease of your choosing. TEV protease seems like a good fit because of its self-limiting behavior and low solubility, making it easy to remove by adding something that renders it insoluble.
Can we eliminate the need for a protease entirely? As we saw before, inteins let us create self-cleaving proteins. Building on this, several papers have demonstrated cleavable, self-aggregating tags. These combined functionalities mean that our target protein can agglomerate outside the cell and release the desired product precisely when we need it.
A lot of work needs to be done to make these techniques practical, but cleavable self-aggregating tags hint at a solution with no filters, no centrifuge, and no column. Using their latent abilities, proteins can purify themselves.
Conclusion
If we can use the power of biochemistry to lower purification costs, precision fermentation can be used to make food, textiles, chemicals, and pharmaceuticals all while using fewer resources than modern industry. I’m not sure which trick will ultimately solve the purification problem, but I have a hunch that the solution will use proteins themselves to do the work for us, before the product leaves the reactor.
This echoes a pattern Laura Deming noticed in biology research:
Something special happens when a field becomes self-reinforcing. Previously, biology looked to physics and other disciplines for tools to break open new frontiers. But, empirically, since the 1950s, that has all changed. We don’t make mutant mice with x-rays and microscopes - we figure out the gene we want to go after, and we use high-precision biological tools to change it.
In the industrial revolution, coal itself powered the trains and steam engines that shipped coal to the rest of the world. Moore's law has continued for decades by leveraging computers to design lithography techniques and computer chips. Revolutions happen when industries use their products to make their own tools; it’s time for synthetic biology to do the same.
Appendix: Selected References
Affinity Tags:
Aptamers:
Aptamer-Based Affinity Chromatography for Protein Extraction and Purification | SpringerLink
Inteins:
Inclusion Bodies:
Inclusion Bodies: Status Quo and Perspectives | SpringerLink
Elastin-like Polypeptides:
Cleavable self-aggregating tags:
New trends in aggregating tags for therapeutic protein purification | SpringerLink
Cleavable Self-Aggregating Tags (cSAT) for Therapeutic Peptide Expression and Purification
"Revolutions happen when industries use their products to make their own tools; it’s time for synthetic biology to do the same."
Seems like a self-reinforcing cycle of innovation. We see this with Tesla's Semi transporting EVs and superchargers supporting its vehicles. We see this with SpaceX rockets launching Starlink satellites, which are helping fund new, cheaper rockets.