It could only happen at Cambridge Consultants. An internal training course run by me – a passionate biologist – for a group of engineers inquisitive enough to want to gain some understanding of how it is possible to engineer biology. A challenge certainly… but also indicative of the unique blend of biology, engineering and advanced computation that is enabling our breakthrough work in bioinnovation.
Business success from world-changing science.
With the world facing considerable challenges right now, bionnovation has the potential to answer some of the biggest questions in vital areas like healthcare, agriculture and climate change. We’re working alongside our clients to develop game-changing products, services and concepts such as the CATALOG DNA data encoding machine and the BacillAi TB monitoring system.
To continue be successful, we need our biologists and engineers to be ‘looking down the same microscope’ – a favourite phrase of mine that sums up what’s going on in our newly refurbished cell culture labs. Here, biologists and engineers are manipulating the same materials and getting the same first-hand experiences as they develop a new wave of advanced therapies and medicinal products.
So how to inspire my engineering colleagues on day one of the course? My instinct was to take a leaf out of Albert Einstein’s On Cosmic Religions and Other Opinions and Aphorisms: “I believe in intuition and inspiration. Imagination is more important than knowledge. For knowledge is limited, whereas imagination embraces the entire world, stimulating progress, giving birth to evolution”.
I wanted to begin by giving them a flavour of biology that they could connect to. Offer them a link to their own skills and fascinations. I’m certainly fascinated by evolution. I have been since my teens when I first read The Blind Watchmaker by Richard Dawkins, which explored how the process of natural selection produced the vast complexity we see in nature.
Compelling reasons to engineer biology
I’m inspired by nature – fundamentally we know DNA can be programmed to create very complex machines, we have intuition about the nature of heredity, and we know that genetic information propagates and evolves. My challenge was to create a syllabus which balanced simple explanations with compelling reasons why we need to engineer biology – while inspiring the mechanical, electronic and optical engineers sitting in front of me.
Naturally then, I started the first session with a picture of a skateboard and a duvet jacket! Both incorporate materials produced by genetically modified microorganisms using modern molecular biology techniques. That’s perhaps a little frivolous, but for me – and I hope my audience – a more compelling reason to engineer biology is medicine. Indeed, the first commercial success , in 1979, of synthetic biology was to produce insulin in bacteria.
The Eagle in Cambridge is one of the UK’s most historically noteworthy pubs – it’s where Crick and Watson announced they had solved the structure of the DNA in 1953. I used this touch of trivia from a golden age of classic experimention to introduce a discussion about the structure of DNA and how the information it contains is structured. We looked at what genes are and how they code for proteins, the functional and catalytic materials of the cell.
I moved on to explain how the cost of reading and writing DNA has fallen dramatically in the years since the sequencing of the first human genome. Crucially, this now enables much of modern biotechnology. Discussing the advances in DNA sequencing and synthesis got the engineers back to the comfort zone of instrumentation. We talked at length about the CATALOG machine I mentioned above, marvelling at its capacity to encode the whole of Wikipedia into DNA.
Practical bioinformatics exercise
Sequencing technology has provided us with vast troves of data. There has been great effort to make this readily accessible. Stepping away from PowerPoint, we launched into a practical bioinformatics exercise. It involved exploring the function of drug and understanding its biochemical target, DNA and protein sequences and structure, using some of the many web portals that access this universe of sequence space. It’s easy to get lost but that’s somewhat the point. Biology is complicated and interconnected.
Actually, should that be interconnected or repetitive? It’s certainly true that biology itself reuses many functional and structural units. Proteins are made of a finite number of domains, architectures and topologies repeated and reinvented to create new functions. Body structure is controlled by master switches called hox genes. An image of the fruit fly Drosophila with legs in place of antennae illustrates to the group what happens when hox genes go wrong!
This brought us to a key question. If nature can reuse parts, then why can’t we as biologists? Can’t we just abstract them, take them out of context and reuse them? It’s all modular right? Well to a point yes; this is what molecular biologists have done for years, and it happens to be what engineers do when they standardise nuts and bolts, so everything fits together.
Current fashion labels this as synthetic biology. I made the point to the group that all recombinant DNA technology is synthetic biology but that nowadays an ever-broadening understanding of biology has given us a huge number of use cases from which to abstract parts and repurpose them. The insulin team at Genentech were no different. They synthesised DNA, cloned it into a vector and used a cell free system to produce the final drug product. That’s not so different to how we might approach things today. I would consider these people synthetic biologists – or bioinnovators to be more precise.
DNA is Lego is it not?
After explaining the fundamentals of Watson Crick Hypothesis – that familiar ladder of A-Ts and C-Gs – I furthered the argument that there’s a hidden truth there. If the two strands are separated, they can be copied, each acting as a template for the other in the ubiquitous PCR reaction. Sequences recognise one another – anneal – so different strands with complimentary ends can join to make new recombinant molecules. It’s easy to copy DNA, cut it and stick it back together. These methods can be standardised, with methods such BioBricks. My revelation – that DNA is Lego – went down well with the engineers.
Throughout the course, I used the age-old battle between bacteria and their viral invaders, bacteriophage, to illustrate a source of parts, tools and their abstraction. Bacterial immunity responses provide a great backdrop to show just how parts can be taken from nature and repurposed for biotechnological ends. Bacterial restriction enzymes, which cut invading DNA, have been a mainstay of genetic engineering for years.
CRISPR/Cas systems are also derived from a bacterial immune response. It doesn’t all go the bacteria’s way though. Phage can hijack the bacteria’s protein synthesis machinery because they have strong promoters and fast polymerases, or they hide by integrating their DNA within the host genome. This also allowed me to talk about gene circuits. I revealed how it’s possible to re-utilise recombinases (DNA integration enzymes) to create ‘state machines’ and recombinant DNA logic. We got deeper into building logic circuits using DNA logic gates – and how it’s possible to build molecular clocks and design biological circuits with the aid of the same programming language used to design electronics.
A key finding in the training course was that when we as biologists talk to engineers, we tend to get rather ‘hand wavy’! There always seems to be a ‘but’ – another layer of complexity. Biological systems are not binary. They respond to a myriad of inputs, each sub-system having its own response threshold. The affinity of DNA binding proteins, acting as on/off switches for genes is often dependent not just on the presence or absence of an effector molecule but also its concentration. Multiple metabolites affect multiple genes and vice versa. It’s the subtle interplay and feedback between many systems which leads to an emergence of complex behaviour.
Proteins are the catalysts of cellular behaviour; nature has endowed function to these macromolecules based on its needs, but we often repurpose them for biotechnological or medical purposes. For example, enzymes have been discovered which will break down PET plastic into its constituent monomers – they are functional enough to support microbial growth in the environment but not address the mountains of waste we produce as a society. How do we improve that protein to be useful?
Complex biological problems
Protein evolution, or rather the rational design of proteins, is a tough problem. The information flow from DNA to protein is one way. The DNA sequence encodes for a protein sequence, a linear chain of amino acids, which inherently contains the information for that protein to fold into a functional entity. However, how this happens is not fully understood. It’s very difficult to model a protein structure from sequence alone. Google’s Alpha fold gets pretty close and is an example of where AI can be used to help with complex biological problems.
To improve function, we need to change the properties of a protein, which means varying the structure by varying the sequence. For a chain of 250 amino acids, a modestly sized protein, there are 20250 possible sequence variations – more than the 1083 atoms in the universe. How do we address that complexity and intelligently navigate such a large sequence design space?
The Blind Watchmaker teaches us about a slow evolutionary tick, a gain of function, accumulated mutations which develop to fit a niche. As I’m blind to the tools of artificial intelligence, I asked the engineers in the room to help. The resulting discussion was stimulating but inconclusive. We can’t solve all the big questions all at once, after all. Thankfully, that means there’s plenty to get stuck into – and a lot more opinions to be shared – at my next ‘how to train an engineer in biology’ training course. I can’t wait.