These type of proteins "PKSs" are one of the most promising and exciting areas of biotech research. They offer a path to make a sort of programmable biochemical factory, where bacteria can make a custom drug, plastic polymer, etc. in a predictable way. They can also do some tricky chemistry that is almost impossible any other way.
PKSs build a molecule piece by piece like a car assembly line, and follow a predictable logic where you can tell from the parts in the DNA exactly what chemical they will make. After 30+ years of working on it bioengineers are finally figuring out how to reliably mix and match the parts to make custom PKSs that work.
Can you elaborate on What’s a PKS? Have you read the science article and does it mention PKS? I don’t have full text access but from the article and the abstract it didn’t sound like it’s a new addition to an existing paradigm..
A PKS is a "polyketide synthase" which is a type of enzyme that acts sequentially, like an assembly line to build large complex molecules. It is best visualized- there are some good videos on youtube with animations and figures showing how they work.
Yes, I read the article- polyketide synthase (PKS) is mentioned right in the title. What makes this finding novel is just how massive this PKS, named PKZILLA is compared to previously discovered ones.
Can you elaborate on the reliability of the process? I imagine there are a lot of things that could go wrong when applying biological processes at any scale where they are able to produce meaningful quantities of stuff?
There are two main obstacles: actually making a new molecule with a PKS, and getting it to be active enough to make useful and financially viable quantities of a product. Both are challenging.
A substantial fraction of the currently used pharmaceuticals, especially big complex antibiotics are already made commercially at scale with PKSs- but mostly in cases where they are just using the natural organism with its natural PKS.
They are challenging to engineer, e.g. "get parts to work together" in an unnatural order but that has been getting better over time, with lots of recent breakthroughs, and a bunch of new molecules being made, but the production quantities do tend to be low.
There is no fundamental reason why PKSs have to be slow, or make small quantities of things, there are a few good examples of both natural products and engineered PKSs that make things in really high concentrations. So this is presumably something that can be improved over time, but is definitely the biggest weakness, which limits commercial viability right now to high value things like drugs.
The impressive thing about this is not the size -- making long peptide chains is easy -- it's the fact that this can be consistently folded into a useful structure, without misfolding or aggregating.
I have to wonder how much of the structure is functional, versus, for lack of a better term, structural "bulk" that is unaffected by misfolding.
Essentially 100% of it is functional! What you have here is a repetitive pattern of small simple protein domains, all connected together in a long "assembly line."
As this is a polyketide synthase (PKS), it is a long assembly line of "modules" that are like individual fatty acid synthase complexes, made from a bunch of small domains that each do a single chemical step as part of a massively complex many-step chemical synthesis.
So we're talking about a huge number of individual domains (140 of them) that each fold separately into a "blob" from this massive peptide, and then they have "linkers" which are not folded, but connect from one domain to the next, and are just long enough so they are each in the proper spot.
Fascinating! This is a completely new area to me. Can I ask what the advantage of this over "classical" biosynthesis pathways, where precursors float from one enzyme to another?
They have different strengths and weaknesses, and natural product pathways often naturally combine steps from both. I'm not sure how technical to make this, but from your username and comments I will assume I can make it pretty technical.
Some of the advantages of PKSs (both for bioengineering, and for the microbes using them to make molecules):
-Rapid evolution and/or engineering modifications: the predictable relationship between the DNA and resulting chemical structure means you can reliably change the DNA and know what will happen to the chemical structure. This reconfigurability is also useful for microbes- they can duplicate or delete regions of DNA to create new variants, such as new antibiotics when their competition develops resistance.
-You can change a part of the molecule without having to have anything unique in that region that an enzyme will recognize to act in the right spot- because a specific region of the assembly line relates to a specific part of the final product.
-Avoids a lot of the issues of enzyme promiscuity in engineered pathways, to make something really different from what exists naturally while reusing natural domains. Although PKS domains can be selective, can be promiscuous and still work (and many natural ones are), because the physical configuration of the PKS will only present them with the right substrate at the right time
-Overcomes lots of issues with reactant concentrations, and thermodynamics. Because the substrate is physically tethered, it is channeled right to the reaction site, and then snatched away afterwards- effectively simulating a very high reactant concentration, and a low product concentration, allowing reactions to occur in the right direction, that wouldn't otherwise be favorable/spontaneous
-Precise control over stereochemistry
There are also NRPSs, which are similar to PKSs, but make peptides instead of polyketides...
I'm imagining something a bit like a paint roller with a bunch of patterns around the rim, or an IBM Selectric ball, where the raw materials keep rolling around on the outside until the finished product hits a special protrusion that causes it to fall away.
That is a fairly accurate concept of how a single PKS "module" works- which extends a growing carbon chain by 2 carbons, plus whatever side groups or modifications it makes to those 2 carbons. Then the entire thing is passed to the next module to repeat, and so on down a long linear "assembly line" that you can think of like the conveyor belt that moves a car down a factory one station at a time.
How much computing power would it take to design and fold this protein the same way? And how much longer until we have software which can do that? It'd be really cool to see what sort of super proteins we could design, all the "bugs" that would need to be updated with updated versions, and what we build on a larger scale with them. This thing looks massive enough it could just eat a small microorganism.
This protein is only slightly larger than the previously largest protein ever, Titin. Which is ubiquitously found in all your muscles so making such large proteins isn’t rocket science for our cells.
This protein folds into 140 individual small domains- we can easily cut it up into small chunks, and computationally fold each individual folding domain.
Because of the modular nature of this type of protein (polyketide synthase or PKS) we actually can already design and build proteins like this.
PKSs build a molecule piece by piece like a car assembly line, and follow a predictable logic where you can tell from the parts in the DNA exactly what chemical they will make. After 30+ years of working on it bioengineers are finally figuring out how to reliably mix and match the parts to make custom PKSs that work.