This episode was produced and presented by Dr Marushka Soobben, a scientist from South Africa who’s been taking part in an internship here at the Naked Scientists.
And this week, a journey into the world of proteins. What are they? why do they matter? And why did predicting their shapes win a recent Nobel prize?
In this episode
What are proteins?
Alan Fersht, University of Cambridge
We usually think of protein as something on our plate. Chicken. Tofu. Perhaps a gym supplement. But what if I told you proteins are so much more? They’re the tools life uses to get anything done – breathing, thinking, healing, and growing. And inside you right now, there are tens of thousands of different proteins, all doing specialised jobs, all working together like a molecular orchestra.
Proteins are polymers, like plastics: massive molecules made from chains of smaller building blocks called amino acids. Think of them like beads on a string – except instead of just one type of bead, nature uses any number from a set of 20 different amino acid “beads”, which can be placed in any order. The specific order they're arranged in is determined by our genes, which act as recipes coding for the protein amino acid sequence. And because each amino acid has a unique shape and chemical properties, the amino acids that make up a protein, and the order they come in, dictates how the protein folds into its final 3D shape. And that shape is essential. It decides what the protein can do, where it fits in the body, and how it interacts with other molecules to perform its function.
One of the most important proteins we know about – insulin – was first decoded right here in Cambridge, and it changed how we understand life by figuring out which amino acids were present.
Later, from the amino acid sequence, we figured out the 3D structure of proteins like insulin and haemoglobin – the actual shape they fold into – which is crucial, because in the world of proteins, shape is everything. If a protein doesn’t fold correctly, it can’t do its job. And when that job is regulating blood sugar, as insulin does, or carrying oxygen around the body as haemoglobin does, then things get serious.
So, in this first part of the programme, we're diving into what proteins really are – not just nutrients, but the biological machines that run our bodies. Here’s Sir Alan Fersht, a protein chemist at the MRC Laboratory of Molecular Biology at Cambridge University…
Alan - Proteins are the workhorses of the cell. Somebody once described them as being where the rubber hits the road. And that's because virtually every chemical reaction in the body is catalysed by an enzyme from the dissolving of carbon dioxide to the synthesis of our genes. They're responsible for the structure of our body. Muscles are protein, hair is a protein. They're also responsible for the defense of the body. Antibodies, which are important in defending us against disease, are proteins. So what are they made of? Proteins are essentially chains of simple molecules called amino acids linked together to form sometimes short and sometimes very long chains.
Marushka - Can you tell us why a protein structure is so important to how it works?
Alan - What is important about a protein is basically its shape, its three-dimensional structure. What proteins have to do is to interact with other molecules in the cell. They do that like in, say, a key fitting a lock or two pieces of a jigsaw puzzle fitting together. So shape is very important for proteins to recognise other molecules. If something goes wrong, the shape changes, perhaps due to a mutation where amino acid residues in a protein are changed because of a mutation, or sometimes the structure gets a bit twisted, it can cause disease. So structural biology, which is the determination of the structure of proteins in three dimensions, has been a very important field.
Marushka - So as you mentioned, the key has to fit the lock with a protein. And insulin was one of the first proteins to be fully sequenced, meaning that it was the amino acid sequence was decoded. Why was this such an important breakthrough?
Alan - Insulin was the very first protein to have its sequence of amino acids determined, and that was done by Fred Sanger in 1954 in Cambridge. By sequencing, we mean the order in which the different amino acids join one to another. It was a tremendous breakthrough because until then, the structures of proteins in such a way were entirely unknown. And so it was one of the big breakthrough events in the last century in chemistry and biology. First of all, it showed that proteins had a defined structure, which you could determine, and just not a mystery. And it laid the foundations for modern molecular biology and biochemistry. The three-dimensional structures of proteins were first determined again in Cambridge at the MRC Laboratory of Molecular Biology. And in order to do that, people had to use various complex techniques, but they had to know the amino acids in the protein to solve that three-dimensional structure.
Marushka - So once the linear structure has been decoded, so the amino acid sequence, and we've got the 2D structure, we're then able to find the 3D structure for proteins such as haemoglobin. Why was finding the 3D structure so important in biology?
Alan - The three-dimensional structure of haemoglobin was determined by Max Perutz. It first of all showed us in detail what a protein looked like, and it also showed us how the different amino acid residues in the protein contributed to its function, how it worked. But it also set up the whole area of understanding many diseases, which are caused by mutations. As I mentioned earlier, mutations in proteins are the change of one amino acid residue to another because of effects of radiation and chemicals and things like that affecting our genes. And there are diseases such as sickle cell anaemia that occur just because of a single mutation. And Max Perutz's experiment was able to show this and explain it. That led to the whole area of looking at proteins to understand how mutations could cause disease and to help us think of ways of how to cure diseases because we knew the structure of the protein. So haemoglobin, apart from initiating the whole area of structural biology, it also set up the idea of having structure-based drug design, where you looked at a protein and you tried using computers and chemistry to make molecules to cure those diseases.
Marushka - So to end off, can you tell us about how much proteins influence our body? What types of diseases that they are involved in, showing why they're so important to study their 3D structure, not just to understand the disease, but also to solve the disease and to find cures for the disease?
Alan - There are many diseases that are caused by just tiny changes in a protein, simple mutations that change one amino acid residue to another. Tay-Sachs disease is an example of a single mutation causing a disease. Cystic fibrosis, another. And at a more general level, cancer is a disease of mutation. There are proteins that control the cell cycle, whether or not it's going to replicate, proliferate, or whether it stays as it is. And there are proteins in the cell that act like accelerator pedals that tell the cell to reproduce. They can be turned on by a mutation. So the cell wants to replicate indefinitely. There are other proteins that are brakes that try and stop it, and mutations can stop them doing that. And these mutations cause cancer. And by looking at those types of mutations, we can try and design drugs to get them working again properly.
09:58 - What happens when a protein goes wrong?
What happens when a protein goes wrong?
Michele Vendruscolo, University of Cambridge
Just one small mistake in the protein structure – a misfold or a mutation or even an incorrect amount of protein – can tip the balance from health to disease. Proteins are the body’s molecular machines, and when they go wrong, the consequences can be profound.
When Nobel Prize-winner Max Perutz solved the 3D structure of haemoglobin, he revolutionised protein research by revealing how even the tiniest change in structure can have life-or-death implications, as an affected protein may be rendered unable to perform its proper function in the body.
Michele Vendruscolo, at Cambridge University’s Centre for Misfolding Diseases, works on what happens when proteins such as haemoglobin go rogue, producing conditions like thalassaemia and sickle cell disease…
Michele – One should think of a protein as like a machine that carries out essential work in the body. For example, when we eat, the food has to be broken down into bits, and then these bits have to be used to build up our components. Proteins are the workers that carry out disassembly and reassembly so that we can actually use the food that we eat. And so, in order to function in this way, proteins – which are very complex molecules – have to acquire a very specific shape, which is called the native state. And this is all fine, because most of the time they do acquire this shape and then they can function, and everything goes well.
But at other times, especially with age or under stress, they are unable to achieve this state and so they so-called ‘misfold.’ So effectively, we are studying the process by which proteins fail to function normally, misfold and malfunction, and give rise to disease processes.
Marushka – So when proteins don't have the correct shape, this leads to a whole bunch of diseases. So shape is very important, as well as the levels of protein within your body.
Michele – The failure to fold is very common. There are, in fact, many diseases that are caused by protein misfolding and aggregation. And there are very common ones like Alzheimer's disease, Parkinson's disease, or even type 2 diabetes. Diseases that affect tens of millions of people worldwide are actually caused by misfolding and aggregation.
Marushka – So diseases like cancer, Alzheimer's – quite a lot of diseases – are dependent on how much protein there is in our body, whether it has folded correctly, and whether it's present where it needs to be. A quite well-known protein is haemoglobin. How can haemoglobin teach us about structure and function? What can you tell us about that?
Michele – Yes, well, historically, haemoglobin is a very important protein, not just because it carries out an important function in the body, but also because it has been much studied. And I guess you made two points about protein function. One is how much protein there is, and the other is what the correct conformation is for the function. Haemoglobin is a good example of both, because there are diseases like thalassaemia, which are related to how much protein there is, and other diseases like sickle cell that are related to a change in shape.
Marushka – How exactly does it work in sickle cell and thalassaemia? What exactly is going wrong with those two diseases in terms of haemoglobin?
Michele – Yes, in thalassaemia, there are mutations that decrease the production of the protein. So we simply have less, and there are all sorts of problems related to the fact that there is not enough. It's like in the example from before – if there are not enough workers, the production chain cannot actually function. In sickle cell, it's a different matter. There are still mutations, but these mutations affect the ability of haemoglobin to actually carry out its function. So there is enough, but it is not functioning properly.
The basic function of haemoglobin is to carry oxygen. It has to be able to change its shape in order to capture oxygen and release it – capture oxygen in the lungs, then travel to other parts of the body and release it in the right place. These mutations affect the ability to capture or release oxygen at the right time and in the right place.
Marushka – So that's what makes people anaemic – they don't have enough oxygen going to where it needs to go. And how can understanding the structure of protein help us fix these diseases?
Michele – Yes, the type of conformational change responsible for sickle cell is actually related to the fact that haemoglobin self-associates into, effectively, filaments that are dysfunctional. Instead of remaining monomeric and carrying out its function with oxygen, it self-assembles into these filaments that are dysfunctional. So the therapy is to keep the protein in the monomeric state – to reverse the self-association process. That is one of the main therapeutic strategies for sickle cell.
Marushka – And for other types of diseases – you work commonly with Alzheimer's and neurodegenerative diseases – how does understanding those specific structures, or just the structure in general of proteins, help us with medicine and drug discovery?
Michele – Yes, I mean, essentially a big part of the therapy is to maintain the protein in its functional state. And it's very interesting to consider the fact that because we have millions of different types of proteins, and even hundreds of thousands or millions of copies of each one, the processes of synthesis, folding, trafficking, and degradation are extremely complex. We have a very powerful quality control system, called protein homeostasis, that is incredibly effective and guarantees the correct functioning of proteins throughout our lives.
We experience disease when, in some way, this protein homeostasis control starts to fail. So the real common thread in therapies related to correcting protein misalignment and aggregation is to augment the protein homeostasis system in various ways to maintain proteins in their functional state.
How do scientists design proteins?
Wes Robertson, University of Cambridge
Proteins used to be mysterious molecules we didn’t understand, especially when it came to their structure, shape and function. But now we're learning to design them – to reprogram biology itself. From lab-made insulin, to antibody-based drugs, scientists are building custom proteins to treat disease, clean up pollution, and even tackle climate change. Here’s synthetic biology specialist, Wes Robertson, from the MRC Laboratory of Molecular Biology at Cambridge University, on the new generation of protein engineers that are turning nature’s machinery into tomorrow’s toolkit…
Wesley – We've learned enough from studying biology and nature to take that information as engineers and start making new proteins. We've studied tens of thousands of protein variants with different sequences and different structures. Now we can use that information to model how sequence leads to structure. In the lab, we're using that understanding of how structure affects function to help us design and generate novel proteins for use in biomedicine, materials science, and beyond.
Marushka – So we've come a long way since insulin was first discovered. How do we now build or redesign proteins in the lab?
Wesley – Well, in addition to understanding the amino acid sequence of proteins, we also understand how the sequence of DNA programs the amino acid sequence in the protein. So, as genetic engineers, taking a step back, we can program the DNA in cells to yield different sequences. We can now engineer new proteins with new sequences and new structures in a predictable manner. The advantage is that we can easily edit the DNA sequence in many different contexts to test many different types of protein structures at scale. In practice, most molecular biology labs in the world do this on a daily basis.
Marushka – So proteins are really involved in multiple facets of life. Are we now programming biology like software? What does that actually mean?
Wesley – Yes, I think that's a good analogy. We know the language of the genetic code – this is how cells use DNA to program the synthesis of proteins. With that understanding of the language, we can now use it in a predictable and programmable manner. A simple DNA input yields a reliable protein output. Now we can test many different proteins to perform specific functions, so we can understand what is the best sequence and structure to carry out a particular function.
We've also developed new tools for bioengineering, where we're now writing DNA at the gene – but also genome – level. We can safely program novel functions for synthetic proteins, but we can also safely program novel functions for entirely synthetic microorganisms.
Marushka – So basically, we start from DNA, we then get the amino acid sequence, and then we're getting new, novel proteins. Can you give us some everyday examples of lab-designed proteins – like medicines or enzymes – that people might know?
Wesley – For sure. As you alluded to earlier, insulin is the first great example, which really started the biotechnology industry in the late 1970s. Another great example is monoclonal antibodies. As a field, we've developed a robust pipeline to understand the sequences and then engineer these antibodies in the lab to deploy them in the clinic to target cancer, viral infections, or inflammation. But beyond just medicine, we also use proteins in materials science. For example, they can be used as adhesives, in washing powders, and they can have a variety of other applications.
Marushka – So could these custom proteins help us tackle bigger problems like global health or even climate change?
Wesley – Definitely. Scientists are tackling these difficult problems using our ability to engineer proteins – for example, to make proteins that can break down plastics so we can degrade them. We can also use proteins and engineer them to replace plastics themselves in a more green and sustainable way than how plastics are currently made. We can also engineer proteins to capture CO₂, and we can use proteins in plants to help them withstand infections and increase resistance to drought. We also use proteins in medicine, but we take inspiration from how nature uses proteins for many other applications beyond just medicine.
Why AI is the future of protein design
David Baker, Institute for Protein Design
Understanding protein shapes has helped us treat disease and even design new proteins from scratch. But what if we could take that one step further – and let computers do the heavy lifting? In this final part, we dive into how artificial intelligence is transforming biology itself. From painstakingly decoding the 3D models in the lab to near-instant predictions on the computer, AI systems like AlphaFold are changing the game. To find out what this means for the future of medicine, biology – and life itself – I put in a call to David Baker, who won a Nobel Prize for his work on protein design…
David - My lab at the University of Washington, many years ago, we were working on trying to understand the principles of protein folding and as experimentally, and as we work those out, we started developing methods for predicting protein structure from sequence. And as those improved, we realised we could go backwards. So instead of going from sequence to structure, we could go backwards from structure to sequence. And so, and really our first demonstration of that was in terms of designing a completely new protein. It was way back in 2003 with Brian Kuhlman's work on Top7. So it's been quite a while.
Marushka - So how does AI like AlphaFold actually figure out the shape of a protein?
David - All AI methods start with a very large training set of examples, in this case of protein sequences and the corresponding protein structures. And the methods basically learn patterns and relationships between sequence and structure. They learn many, many different types of patterns. And I think it's pretty fair to say now with deep learning methods, given a sufficiently large training set, you can basically solve any mapping problem. And so for example, recently there's been great progress in weather forecasting, I think we've all noticed, which is kind of the same principle. If you have enough data, you can train a model to learn relationships, even ones that are very subtle.
Marushka - And why is predicting a protein shape such a big deal in science? Why is that so important?
David - Well, the proteins are the miniature machines that carry out essentially all the important functions in our bodies and in all living things. And their shape is really important to their function. So predicting protein structures helps you understand what they do. And likewise, if you want to design new proteins for carbon sequestration or curing cancer or for preventing neurodegeneration, to solve those problems, you need design proteins, which have shapes, which are appropriate for solving those problems. And so that's why it's so important to go through protein structure for both prediction and design.
Marushka - So when we figure out the shapes and the protein design, this can basically help us with drug discovery. So can using computers like this really speed up how we find new drugs, new treatments?
David - Yes, well, it's important to two different problems. There's the structure prediction problem, going from amino acid sequence to structure and function. And then there's the design problem, which is starting with a function that doesn't exist, and designing a protein which has that function. And so in the design case, it has an immediate application of drug discovery, because you can literally design new drugs. And we have one medicine that's already been approved for use in humans, we have a number more in clinical trials. And these are medicines that are directly designed to be vaccines or therapeutics. So yeah, we've already proven that that's the case.
Marushka - And what does this mean for the future of biology and medicine?
David - Well, I think what it means for medicine is that we can now design cures for a lot of conditions which have been really difficult. So just to give you an example, so we're designing improved treatments for cancer that are more focused and targeted. We're designing treatments for autoimmunity, where you suppress the immune system at the place in the body where you want it to be suppressed and not throughout, which can lead to complications of infection. My colleague Neil King is designing vaccines, for example, a universal flu vaccine and coronavirus vaccines that could give very broad protection. We're designing proteins to combat newly emerging viruses for pandemic protection or protection against bioterror agents. And I mentioned neurodegeneration, we're trying to combat the proteins, which have been implicated in neurodegeneration. So I think all across medicine, we're going to see applications of protein design. And it doesn't stop there. In sustainability, we're working on proteins to break down plastic proteins to break down toxic compounds, like forever chemicals, new ways of sequestering greenhouse gases. And then in technology, they're really exciting new avenues we're pursuing for much more sensitive and general sensors, kind of like electronic noses. So there are really applications of protein design across a large fraction of the problems that humanity currently faces.
Marushka - So proteins are involved in quite a multitude, many aspects of life, basically. Are we getting to a future where we can design life from scratch?
David - Well, so I've described designing independent individual proteins to design life from scratch, there are many parts that need to be designed. So a living thing has to be able to replicate itself, it has to be able to get resources. And that requires designing many proteins or protein like structures, and then a coding system. And so towards that end, we're working on sort of new ways of synthesising proteins. In biology, it's the ribosome, but that's kind of limited in what the inputs it can take is. And so we're working on sort of more general systems. So you could make arbitrary molecules. But designing life itself will require a lot of coordinated design of a lot of different parts. And I think it's a very exciting area.
Comments
Add a comment