Folding and Unfolding

The​ mad cow disease crisis began in 1984, with reports of cows ‘acting strangely’ on a farm in Sussex, and ended 32 years later, with the last reported death from a variant of Creutzfeldt-Jakob disease. It was clear early on that something in the brain matter of cows was causing the infection, and that its vector was the charnel houses in which the brains and spines of slaughtered cows were diced, rendered and minced to a powder, before being added to a variety of commercial foods, including animal feed. So cows ate cow brains and got Bovine Spongiform Encephalopathy. And it wasn’t long before house cats, elk, a lion in a West Country animal park and an endangered Arabian oryx at London Zoo all caught a form of BSE.

At first the government, supported by some scientists, insisted that humans could not contract BSE. But in the mid-1990s, instances of CJD, a spontaneously occurring neurodegenerative disorder that usually strikes near randomly at a rate of one case per million people, began to increase in Britain. Like BSE, CJD is a spongiform disease: in autopsies the brain resembles a dried-out sponge, pocked with tiny holes. Like BSE, it is incurable and invariably fatal. Many doctors and scientists realised that because of the clustering pattern, this variant of CJD (vCJD) was not occurring randomly and was probably a human form of BSE.

The outbreak appeared to have the makings of a pandemic. The mysterious infectious agent persisted for years after the death of the animal, and was resistant to burning, irradiation and most known disinfectants. One group of European researchers estimated that 500,000 people could be exposed to BSE from a single infected cow; Oxford’s Wellcome Trust Centre predicted between a hundred and 136,000 deaths. (In the end, the figure was 178.) These imprecise numbers reflected the fact that the science was far from settled. Parliament and the public were told that the disease was caused by ‘abnormal proteins’, which didn’t make things much clearer: after water, proteins are the most abundant molecules in the human body. This theory was known as the prion hypothesis (prion, pronounced ‘pree-on’, is short for ‘proteinaceous infectious particle’). It held that the proteins in a human or cow were capable of changing form, with the new molecule not only causing disease but also replicating itself and becoming infectious. The hypothesis turned out to be correct. But early in the crisis scientists could not give definitive answers about where BSE came from, or how it spread, or how to treat it. ‘Infectious proteins – once a biological heresy – have been blamed for mad cows and dead people,’ one journalist wrote in Science, but ‘nobody has proven that these prions really exist.’

In his new book the French neuroscientist and biochemist Michel Brahic recalls that frantic period. He was one of the scientists who fielded calls from journalists, and who told them – mistakenly – that it was extremely unlikely the disease could jump to humans. This is a surprisingly frank admission, but perhaps less so for a scientist working in Brahic’s field. Prion research is characterised by unexpected observations that seem to violate the scientific consensus. How and why do proteins, which are produced by our body to build structures (such as skin and muscle) or carry out tasks (like breaking down our food), abandon their set roles and develop a new one? Unlike other proteins, prions can gain the ability to replicate, causing damage as they spread; they can jump between bodies, and sometimes even between species. In this regard they are similar to a bacterium or virus, but both of those are complex assemblages made up of many different proteins and powered by an individual genome, which gives them the ability to replicate and the evolutionary impetus to survive and reproduce. Prions don’t have their own DNA, or obvious access to evolution. They don’t fit the category of foreign invader, or of an internal genetic renegade, like cancer.

The American biochemist Stanley Prusiner coined the term ‘prion’ in 1982, and won the Nobel Prize for his research in 1997. By the early 2000s, there was consensus on what prions were. This owed something to heightened interest in the wake of the BSE outbreaks, but more important were the revolutions in recombinant genetics and structural biology which gave researchers the ability to clone genes easily and to study proteins outside cells, as well as providing new methods of modelling molecular structures. The work on the existence of prions solved several scientific mysteries: not just BSE, but also scrapie and kuru, a laughing sickness that affected the Fore people of Papua New Guinea, who practised cannibalism. All were assigned a single cause: not just a protein, but the same protein – the PrP, or prion protein.

For many, Brahic writes, ‘the idea that proteins can behave as infectious agents was, at first, hard to accept.’ These scientists’ difficulty was in part a result of their adherence to the (admittedly loose, but still influential) central dogma of mid-20th-century biology, which held that information could only flow in one direction. Genes gave instructions for an organism to make proteins, which did the work of the cell. A mutated gene could create a harmful protein, but the ultimate cause of the problem lay elsewhere. But disease arising from something inherent and dynamic within a protein itself was another matter. It suggested a hidden landscape beneath what was thought to be well-mapped terrain. Outside biochemistry, the story of what Brahic calls a ‘revolutionary concept’ in science is either poorly understood or completely unknown, something his book seeks to correct.

After the discovery of prions, many biochemists fundamentally rethought their ideas about protein behaviour. Their focus shifted from establishing what particular disorder a prion might cause to understanding what the logic of the prion was, and where else it might be at work. This led to the discovery of a host of other proteins that acted like prions and induced several small revolutions in scientific thinking. (It was at this point that the term ‘prion’ became confusingly polysemous, referring both to the distinct protein PrP – present in humans and most mammals – that can cause BSE and CJD, and more generally to all proteins that act in an ‘infectious’ manner.) It is now believed that every known neurodegenerative disease – from Alzheimer’s to Parkinson’s to ALS to Huntington’s to general dementia – involves a prion-like protein. At the same time, biochemists researching other species hypothesise that disease-causing prions are either aberrant or account for a small minority of all prion-like proteins. Across countless species there are many helpful if poorly understood prions, which often record and pass down information about the environment to successive generations – making prions one of the methods by which living things supplement or circumvent genetic descent.

While it is​ easy enough to explain basic genetics to non-specialists, the question of how genes give instructions – to make blue eyes, say, or blood cells – is more complicated. Proteins are the first step of that process. The vast majority of genes, whether in a bacterium or a tuna or a human, are instructions to make a protein; working in aggregate, proteins build the structures of the living world. As the biochemist J.D. Bernal wrote in 1951, ‘proteins are the necessary basis for carrying out the processes that we call life.’ The physical structure of a protein molecule tells us most about its function; we can learn less about it from the instructions written in DNA than from the way the resulting molecule arranges itself in space. The process is relatively straightforward and usually illustrated in four orders of increasing structural complexity. DNA contains instructions for a cell to arrange amino acids in a linear chain which, once formed, is called a protein. There are twenty amino acids in the human body, and their main shared feature is that they can bind to each other in the manner of train carriages. The sequence of amino acids joined together in a straight line is the first level of organisation, known as the primary structure.

Each of the twenty amino acids has a slightly different chemical composition. Some are small, some are bulky; some repel water, others welcome it; some are ‘sticky’, binding tightly to other amino acids when they come in contact, others less so. It is these interactions that give the chain its increasingly complex structure. If the chain has repeating sequences of between three and ten amino acids that can bind with the sequence just ahead of or behind them, it will continually fold back on itself and form a sheet. If each amino acid in the sequence is trying to bind with the acid a few positions ahead of it, the resulting forces will cause the chain to curl into a helix. These simple chain patterns – sheets, helixes and loops – are called the secondary structure, and they can be used as the building blocks for more complex three-dimensional shapes, in a similar fashion to the knots and loops woven together in macramé.

This knitting and folding is the third level of organisation, the tertiary structure. The results can be simple and relatively inert – for example, the springy helixes of keratin proteins that are twisted together to form the fibres of mammalian hair (if, genetically, someone’s hair is also rich in the amino acid cysteine, which forms strong chemical bonds between the cables, it will tend to stick together and form curls). Or they can be staggeringly complex. The proteins that viruses use to enter human cells have multiple distinct areas, known as domains. A series of sheets of water-repelling amino acids can anchor the protein to the outside of the oily viral membrane. A bundle of helices might form a long needle-like cylinder reaching towards the human cell; the tip of the cylinder often contains a mimic of a human protein, which will bind to the human cell without alerting the immune system to the presence of a foreign invader. When it binds, the viral protein cylinder contracts like a piston, pulling the viral and human cells together and causing infection. (For some complex tasks like this, assemblages of multiple smaller proteins form a larger whole molecule. This is the final level of organisation, the quaternary structure.)

This progression from a linear string to increasingly complex three-dimensional structures is referred to as protein folding. The name calls to mind origami and, just like it, the order is crucial. If the wrong amino acids stick together, the protein may not fold properly in the subsequent steps; if two helices that were supposed to nestle into each other late in the folding process come in contact too early, the final shape may not form. This problem is solved by other proteins (in biochemistry there is always another protein) called chaperones. A chaperone helps other proteins to fold, either by acting as a scaffold that keeps intermediate structures stable, or by temporarily shielding sticky or reactive regions of the chain until the rest of the protein is in the right position. As the biochemist Susan Lindquist put it, chaperones prevent proteins from having ‘inappropriate liaisons’.

A complete, folded three-dimensional protein is remarkably stable. A disordered linear strand has arranged itself into an increasingly more ordered structure, and is about as likely spontaneously to revert to its previous state, or to radically rearrange itself, as a chair would be to transform into a table or to collapse into a pile of wood. That is, until a prion comes along. In a prion, a region of the amino acid chain fails by chance to fold into a stable position. Lacking the interactions with other parts of the chain that would stabilise it, this part thrashes about. It takes a new shape every few microseconds, trying to bind to itself, or to other proteins or molecules around it. These regions are called ‘intrinsically disordered domains’. Sometimes they result from errors in protein folding, which cause the cell to target them for destruction and recycling. In rare cases they are not random, and serve a purpose – keeping part of the protein structure inactive until the cell needs it, for example. A prion puts this disorder to novel uses. Like a reverse chaperone, it induces another protein not to fold but unfold, undoing the relentless push towards order and causing its target to have more inappropriate liaisons. The new protein now also acts like a prion, abandoning its previous role and turning more proteins into prions.

The prion protein​ has a normal function in the brain relating to memory (mice transgenically altered to have no PrP perform poorly in spatial tests that rely on memory, such as mazes). In this state, the PrP doesn’t cause problems and it’s not clear why, after years or decades, the infective form emerges. It could be random, but pathological prion protein has certain characteristics that seem protective or adaptive. Once the prion proteins in a cell start acting like prions, something about their disordered structure makes them highly resistant to protein-digesting molecules and other chaperones, which could destroy or refold them. As the prions make more and more of the prion protein in the cell infective, they form aggregates of long, crystalline fibres in the brain, winding around and between cells like an invasive plant. There is disagreement about whether this effect is part of their pathology. Are the fibres damaging the brain and its cells? Or is the problem that so much PrP is piling up in an unusable form, thereby starving the brain of the normal form?

By the early 2000s, when it had become accepted that a protein could transmit information about how to fold onto other proteins, researchers began looking for prion-like effects in other diseases. In 2008, two separate hospital research groups published findings on a failed Parkinson’s treatment that had unexpected results. Parkinson’s is caused by aggregates of a protein called α-synuclein, which disrupts dopamine signalling, causing tremors. Since Parkinson’s tends to emerge late in life and to progress with age, the groups had grafted stem cells from the brain of a foetus into the brains of Parkinson’s patients. This wouldn’t halt the degeneration of the patients’ own brains cells, but it was hoped that the healthy foetal cells might be able to run the brain’s dopamine system, heading off the dopamine deprivation symptoms. The cells were implanted in patients in the 1990s, but when the researchers examined the patients’ brains in autopsies nearly a decade later, they found that α-synuclein tangles from the aged brain cells had crept into the grafted foetal cells. In effect, this microcosm of a baby brain now had Parkinson’s. The disease wasn’t just degenerative – on the cellular level it was infective too.

Different research groups, including Brahic’s at Stanford in the early 2010s, began experimenting with human brain cells grown on petri dishes, and showed very clearly that the pathological α-synuclein strands could spread to a healthy neuron, implicating the protein itself as the infective agent. Prusiner, whose Nobel Prize derived partly from the brilliance of his work and partly from his willingness to make definitive statements, declared in 2015 that α-synuclein was ‘the first new bona fide prion to be discovered … in the last fifty years’, and suggested more would follow.

Prion-like behaviour was discovered in other brain diseases, including Alzheimer’s, where both Amyloid-β proteins (which form the brain ‘plaques’ associated with the disease) and Tau proteins (aggregates of which cause the dementia symptoms) have been shown to misfold and propagate themselves in the manner of PrP or α-synuclein. (Why the brain? Brahic wonders if we’ve simply looked hardest there.) And as more examples were found, researchers identified the protein sequences and structural shapes common in proteins that act as prions, enabling them to trawl databases of known proteins to find new candidates to study. The revelation that prions cause most neurodegenerative diseases made headlines, but for biochemical science a more significant result has been the discovery of a new behaviour for proteins – many argue it amounts to an entirely new class of proteins.

Yeast, for example, is now thought to have a prion-like protein that detects when nutrients become scarce, and induces all proteins of its type to refold themselves into new structures that work on different sources of food. This change carries an inheritance: when a yeast cell divides, the daughter cells share the original cell’s proteins – including the prion – and so any of the originally structured form of the protein that they produce will be refolded by the prions. Similar benign or helpful mechanisms have been found in other organisms. Humans and nearly all other animals have a protein called CPEB that can shift into a prion-like state in neurons involved in long-term memory; rather than harming the cells, the CPEB aggregates appear to protect the neuron from being reprogrammed, perhaps preserving the memory. Many non-pathological prions appear to be able to halt their spread or even reverse it, folding themselves back into their original structure and leaving things as they were before the prion went to work.

The basic logic of the prion – a protein that can radically change its structure and function, and force that change onto other proteins too – may turn out to be so common that it ranks among the vital biological processes. As is often the case with cascading scientific discoveries, this has set off a nomenclature crisis. It was initially thought that prions were only pathological, but in the last few years scientists have started referring to ‘good’ and ‘bad’ prions – a quick fix that is both imprecise and totalising. It would be useful to find a new term for them: in addition to the good/bad dichotomy, there’s the confusion over the word ‘prion’ being used to refer to both PrP and a whole class of proteins that can refold other proteins. Brahic likes ‘piaf’ (again borrowing letters from ‘protein’ and ‘infection’, with a nod to the French singer), which was originally mooted by Prusiner in the 1980s.

Treatment for neurodegenerative disease is still a long way off. But in time we might find ways to prevent prion proteins from transforming into their pathological form, or to target them for destruction the moment that happens: antibodies, which work by recognising the structures of certain proteins, are being trialled against the prion form of α-synuclein. For now the risk of another BSE-like outbreak remains, and most wealthy countries monitor CJD and livestock for early warning signs. (The mystery of how BSE jumped to humans, as well as the low death count, was solved in 2002: although millions of people were exposed to pathological PrP proteins from cattle, it could only jump to humans who had a rare variant of the human PrP protein, differentiated by a single amino acid. To anyone but the genetically unlucky, exposure to BSE was harmless.)

When Brahic began studying infectious neurodegenerative diseases as a medical student in the 1970s, he writes at the end of his book, the idea that proteins could transmit and replicate information between themselves was referred to as the ‘heretical hypothesis’. It’s still not clear how widespread this behaviour is, or how often it’s used to pass down inherited information outside of the genome. But in the past few years several papers about the origins of life on Earth have pushed the heretical hypothesis to a radical conclusion. The standard theory says that around four billion years ago the basic information-encoding chemical molecules of life formed, and somehow came to be self-replicating. DNA, or more likely RNA, entered a life-sustaining loop of reproducing itself, and then, through random iteration, began to make the proteins that do its bidding. Now it’s established that proteins can themselves transmit information, why couldn’t that order be reversed? At the beginning of all things a prion, copying itself over and over again, waiting for the rest of the world to catch up.