Part 1: What we did
What we build when we make a genetically engineered probiotic
While “genetic engineering” can take a few different forms and use a few different strategies, for simplicity’s sake I’ll just take you through what we do at ZBiotics, and in particular, what I did to make B. subtilis ZB183 – the specific probiotic used in our first product, ZBiotics.
You may be surprised to discover that genetic engineering is not just going in and monkeying about with the DNA until we get something we want. Rather, it’s careful and highly controlled, using naturally existing processes that life evolved over billions of years!
THE STARTING POINT
Food-grade bacteria and functional proteins
To make ZB183, I started with a safe, probiotic bacteria found on fresh fruits and vegetables – something you likely consume every day. It’s a bacteria called Bacillus subtilis (“B. subtilis”). In addition to already eating it in our food, humans have also famously (or infamously, if you’re like me and don’t love the taste) been consuming it in large quantities for centuries in a Japanese fermented soybean food called natto. B. subtilis’s interaction with the body has been carefully studied for years and its safety has been well proven. That’s one of the reasons we decided to use it as the foundation of ZBiotics.
Like pretty much all life on earth, the blueprint for B. subtilis is its DNA, containing individual clusters of code called genes. B. subtilis has thousands of genes, and each one is an instruction guide for how to make different proteins.
Don’t think of proteins like your grocery store chicken breast (though that is indeed made up of proteins). In biology, the word “protein” means way more. Proteins are the special tools of life. They enable B. subtilis to do all the things it needs to do to survive – like digesting nutrients, building components of their cells, sensing what’s going on in their environment, and defending themselves from attack.
But bacteria don’t just make proteins for the sake of making proteins. Bacteria, including B. subtilis, create proteins when they need them (and not when they don’t). They sense what’s going on around them and make different types and amounts of proteins based on what they need in that moment. For example, if there are many nutrients available to digest, the bacteria will create more digestion proteins than they will defense proteins. Overall, bacteria like B. subtilis are extremely “smart” and can adapt to their environment well.
Tweak existing bacterial blueprints to deliver new functionality.
With B. subtilis in hand, it was time to choose what I wanted to produce, then go out and build it. For ZB183, I knew I wanted to make B. subtilis perform a desired function. I wanted it to break down acetaldehyde, the primary reason why people feel so crummy the day after drinking. To do that, I needed to make B. subtilis produce lots of acetaldehyde-digesting protein (my “protein of interest”) and I needed it to be producing that protein constantly and consistently.
How to get it done? By changing B. subtilis’ genetic blueprints according to the following strategy:
- Find existing machinery in B. subtilis for making lots of a given protein
- Repurpose that machinery to make lots of my protein of interest, rather than the protein the bacteria usually makes
- Remove all the “stops” the bacteria use to regulate when this specific protein is made, so that it will make my protein of interest all the time
(1) Find existing machinery in B. subtilis for making lots of a given protein:
When I started thinking about building ZB183, the first thing I knew was that our bacteria needed to be able to make a lot of my desired protein of interest. Rather than engineer the bacteria to do this from scratch, it’s a whole lot easier to just start with the machinery for building proteins that B. subtilis already has: essentially, taking the bacteria’s blueprint and tweaking a portion of it, rather than creating something brand new.
To do that, I searched through the scientific literature for known proteins that B. subtilis already has the machinery for making a lot of. There were actually several options, but I ended up choosing one I was already familiar with from my PhD. It’s a structural protein called flagellin that B. subtilis makes tons of when the conditions are right.
So okay, step 1 is complete: I’ve found some existing machinery that I can work with!
(2) Repurpose that machinery to make lots of my protein of interest:
I next identified the specific gene encoding that flagellin protein (remember, genes are the instructions – or blueprints – for making proteins). This was the blueprint I needed to change. Once I had that critical gene, I could swap in a new gene to change the protein it was making. Allow me to explain:
First, I borrowed a gene from another naturally-occurring environmental bacteria called Cupriavidus necator. Cupriavidus necator has its own really cool properties, including the gene to make the different protein I was looking for when creating ZB183. That different protein is called acetaldehyde dehydrogenase, and it breaks down acetaldehyde.
Here’s a cool model of that protein:
Once I had the gene encoding acetaldehyde dehydrogenase, I went back to B. subtilis. There, I removed the unwanted gene for the flagellin protein and swapped in my desired gene for the anti-acetaldehyde protein, right in its place. This swapping of genes is done with extreme precision, and all of the changes are carefully verified. I’ll explain in more detail how that swap is done in a separate companion post.
Ok, step 2 complete. I’ve repurposed some robust machinery to make my protein of interest: the protein that breaks down acetaldehyde.
(3) Make my protein of interest nonstop
At this point, I had only completed 50% of the job at hand. I also needed to change when the bacteria manufactures its new anti-acetaldehyde protein.
Currently, the B. subtilis bacteria only use its protein-producing machinery to make a lot of protein at certain times. That doesn’t work for our purpose, because it would be really difficult to predict when the bacteria would manufacture this protein after you consume it. So to fix this unpredictability, I wanted to make sure it was always producing the protein, rather than making decisions about when to turn it on and off.
To accomplish this, there are bits of code in the protein’s genetic blueprint that function as on/off switches. It can be tricky to figure out precisely where these switches are and how they work. However, because I was already familiar with this particular system from my PhD work, I was able to identify and delete the “off” switches (in this case there were two different ones, and again, I’ll explain how I made these deletions in a companion post).
At the end of Step 3, I now have the bacteria I wanted: one that produces a TON of acetaldehyde-digesting protein, and produces that protein all the time!
So what about the rest of the bacteria?
In every other way, the bacteria is exactly the same as it was before, continuing to make the thousands of other proteins it uses to go about its normal business. It just has this one small adjustment – the ability to break down acetaldehyde consistently and robustly. It also happens to be a small adjustment that provides an enormous benefit to the people who consume it.
And, in short, that’s genetic engineering.
We just take the genetic code (the blueprints) and carefully and specifically change a small aspect of it to produce a desired change in the organism. That change is particular to the added function we want to create: in our case, breaking down acetaldehyde.
To learn about how we make those careful and specific changes – using the controlled application of natural mechanisms evolved over billions of years – check out Part 2 of this post.