Automated Evolution of a Microbe for Sustainable Bioplastics
What happens when you evolve a microbe to improve the production of sustainable bioplastics? When microbes are well adapted to their environment, they can grow faster and produce more. Today we're taking a closer look at a technology in the Ginkgo foundry for automating the process of evolution.
Transcript
Our story begins in 1941 in Paris - I bet you weren't expecting that, were you?
Jacques Monod, legendary microbiologist and handsome devil, is rushing to finish his doctoral dissertation so he can get back to working for the French resistance. He is studying the response of bacteria to different mixtures of sugar and discovers what we now call the diauxic shift.
When E. coli grow on mixed sugars, in this case glucose and xylose, they don't consume them both at the same time. Monod observed an initial burst of growth that was associated with the consumption of glucose, then a pause, then a second burst as the bacteria switch to eating xylose.
We now know that this pause comes from changes in metabolic regulation as E. coli expresses the enzymes to process the new carbon source. Eventually Monod's work would lead to the discovery of the first genetic regulation system, for which he got the Nobel prize in 1965.
The practical implication for people who want to grow E. coli at scale is that there is one set of enzymes for glucose, another for xylose, and E. coli prefers to eat them one at a time, rather than all at once.
Fast forward a few decades and the diauxic shift still matters for metabolic engineers. This time it was impacting a bioprocess to make sustainable plastics. A chemical company was looking to reduce fossil fuel inputs in their supply chain for plastics. Instead of petroleum, they were using sugarcane pulp as their source material. This is essentially a waste product, also called bagasse to my friends out there who know their feedstocks.
Sugarcane pulp is rich in both glucose and xylose. This makes it a good starting point for bioplastics, because these are chemically close to the important monomers needed for the process. This project started with a strain of E. coli that had been engineered to produce plastic monomers using xylose and glucose as inputs.
But because of the diauxic shift, E. coli don't usually consume these two sugars together. This behavior was slowing down our customer's bioprocess and reducing their yields.
Their microbe needed to be adapted to the new medium. The E. coli had evolved in an environment of scarce resources, where it pays to focus on one kind of sugar at a time. We needed to transition it to an environment of abundance, where it pays to consume both glucose and xylose rapidly.
For this application, we went with a technology called Automated ALE, where ALE stands for Adaptive Laboratory Evolution. In the Ginkgo foundry, our ALE devices are used to rapidly evolve microbial strains to adapt to new media conditions.
Here's how it works. We begin with a culture of microbes, in this case E. coli, and a growth medium. We connect inputs to supply oxygen and fresh medium, and outputs to remove waste. As the microbes replicate and multiply, they cause the growth medium to become cloudy. We continuously monitor the cell density and dilute the culture with fresh media to keep the cells growing.
If you work in this area, you might be familiar with the idea of a turbidostat - a device that can maintain microbes at constant density and constant growth. This system is a similar idea, but with optimizations that allow reliable long-term evolution and eliminate biofilm formation, which screws up a lot of conventional turbidostats.
The ALE device provides the optimal conditions for evolution. Any time a mutation occurs that improves growth rate, that mutant will gradually spread until eventually it represents the entire population. As a result, the overall growth rate will gradually increase over time.
Here's what that looked like for this project. We grew the microbes on a combination of glucose and xylose. Over the course of about 2 and a half months, the growth rate gradually increased about 6x. They started out dividing about once every 10 hours. We got that division time under 2 hours.
This improvement in growth rate was associated with an improved consumption rate for xylose. Here's a plot showing the consumption rate of xylose and glucose during the course of growing a single batch of E. coli. The original strain consumes the glucose first. It only switches to xylose after about 8 hours, once the glucose starts to run out, and even then the xylose uptake rate is pretty low.
After the evolution, the adapted strain uses both glucose and xylose in parallel and rapidly. We saw a 390% increase in the consumption rate for xylose. We saw much faster batch fermentation times, dropping from 23 hours to 7 hours. Not bad for a 6 month R&D project.
A good ALE project needs a good alignment between the requirements of your process and the requirements of evolution. Evolution acts to improve fitness, so it tends to work well on things that help bacteria replicate and survive: growth rate, nutrient use efficiency, stress tolerance. I like ALE for adapting cells to low pH, high salt, high temperature or unusual nutrient mixes.
This project happened to use E. coli, which is fun because it's exactly the microbe where Jacques Monod first described the diauxic shift. But our system can continuously culture basically any microbe that grows well in suspension: bacteria, yeast, fungi, microalgae, or even plant cells.
I think of ALE as being complementary to genetic engineering as a strategy to improve microbes. It is definitely less powerful in some ways. Evolution is not going to introduce completely new metabolic pathways, at least not in 3 months. We need synthetic biology for that.
But evolution is great at finding tweaks and mutations to improve an existing function. It searches the entire genome, including places where human synthetic biologists might not know to look. It can make major changes, including in processes that regulate central metabolism like the diauxic shift, so long as those changes produce fitter microbes.
We can use automated ALE in addition to an engineering campaign, as a kind of polishing step to make those engineered microbes healthy and robust. Or we can use ALE just by itself. Maybe your process uses a wild-type microbe that doesn't need to be engineered. Some microbes are difficult to engineer, because the necessary genetic tools haven't been developed yet. But anything that lives can grow and adapt to local conditions.
Basically evolution works. Our automated ALE system can make it work for you.