Matt Ridley

somebody had suggested that the scrapie agent might have no D N A or R N A genes at all. It might be the only piece of life on the planet that did not use nucleic acid and had no genes of its own. Since Francis Crick had recently coined what he called, half-seriously, the

'central dogma of genetics' — that D N A makes R N A makes protein

- the suggestion that there was a living thing with no D N A was about as welcome in biology as Luther's principles in Rome.

In 1982 a geneticist named Stanley Prusiner proposed a resolution of the apparent paradox between a DNA- less creature and a disease that moved through human D N A . Prusiner had discovered a chunk P O L I T I C S 2 7 7

of protein that resisted digestion by normal protease enzymes and that was present in animals with scrapie-like diseases but not in healthy versions of the same species. It was comparatively straight-forward for him to work out the sequence of amino acids in this protein chunk, calculate the equivalent D N A sequence and search for such sequences in amongst the genes of mice and, later, people.

Prusiner thus found the gene, called PRP (for protease-resistant protein) and nailed his heresy to the church door of science. His theory, gradually elaborated over the next few years, went like this.

PRP is a normal gene in mice and people; it produces a normal protein. It is not the gene of a virus. But its product, known as a prion, is a protein with an unusual quality: it can suddenly change its shape into a tough and sticky form that resists all attempts to destroy it and that gathers together in aggregate lumps, disrupting the structure of the cell. All this would be unprecedented enough, but Prusiner proposed something even more exotic. He suggested that this new form of prion has the capacity to reshape normal prions into versions of itself. It does not alter the sequence — proteins, like genes, are made of long, digital sequences - but it does change the way they fold up.1

Prusiner's theory fell on stony ground. It failed entirely to explain some of the most basic features of scrapie and related diseases, in particular, the fact that the diseases came in different strains. As he puts it ruefully today, 'Such a hypothesis enjoyed little enthusiasm.'

I vividly remember the scorn with which scrapie experts greeted the Prusiner theory when I asked them for their views for an article I was writing about this time. But gradually, as the facts came in, it seemed as if he might have guessed right. It eventually became clear that a mouse with no prion genes cannot catch any of these diseases, whereas a dose of misshapen prion is sufficient to give the diseases to another mouse: the disease is both caused by prions and transmitted by them. But although the Prusiner theory has since felled a large forest of ignorance - and Prusiner duly followed Gajdusek to Stockholm to collect a Nobel prize - large woods remain. Prions retain deep mysteries, the foremost of which is what on earth they 278 G E N O M E

exist for. The PRP gene is not only present in every mammal so far examined, but it varies very little in sequence, which implies that it is doing some important job. That job almost certainly concerns the brain, which is where the gene is switched on. It may involve copper, which the prion seems to be fond of. But - and here's the mystery — a mouse in which both copies of the gene have been deliberately knocked out since before birth is a perfectly normal mouse. It seems that whatever function the prion serves, the mouse can grow to do without it. We are still no nearer to knowing why we have this potentially lethal gene.2

Meanwhile we live just a mutation or two away from catching the disease from our own prion genes. In human beings the gene has 253 'words' of three letters each, though the first twenty-two and the last twenty-three are cut off the protein as soon as it is manufactured. In just four places, a change of word can lead to prion disease — but to four different manifestations of the disease.

Changing the 102nd word from proline to leucine causes Gerstmann—Straiissler-Scheinker disease, an inherited version of the disease that takes a long time to kill. Changing the 200th word from glutamine to lysine causes the type of C J D typical of the Libyan Jews. Changing the 178th word from aspartic acid to aspara¬

gine causes typical C J D , unless the 129th word is also changed from valine to methionine, in which case possibly the most horrible of all the prion diseases results. This is a rare affliction, known as fatal familial insomnia, where death occurs after months of total insomnia. In this case, it is the thalamus (which is, among other things, the brain's sleep centre), which is eaten away by the disease.

It seems that the different symptoms of different prion diseases result from the erosion of different parts of the brain.

In the decade since these facts first became clear, science has been at its most magnificent in probing further into the mysteries of this one gene. Experiments of almost mind-boggling ingenuity have poured out of Prusiner's and others' laboratories, revealing a story of extraordinary determinism and specificity. The 'bad' prion changes shape by refolding its central chunk (words 108-121). A P O L I T I C S 2 7 9

mutation in this region that makes the shape-change more likely is fatal so early in the life of a mouse that prion disease strikes within weeks of birth. The mutations that we see, in the various pedigrees of inherited prion disease, are peripheral ones that only slightly change the odds of the change in shape. In this way science tells us more and more about prions, but each new piece of knowledge only exposes a greater depth of mystery.

How exactly is this shape change effected? Is there, as Prusiner suspects, an unidentified second protein involved, called protein X, and if so, why can we not find it? We do not know.

How can it be that the same gene, expressed in all parts of the brain, behaves differently in different parts of the brain depending on which mutation it has? In goats, the symptoms of the disease vary from sleepiness to hyperactivity depending on which of two strains of the disease they get. We do not know why this should be.

Why is there a species barrier, which makes it hard to transmit prion diseases between species, but easy within species? Why is it very difficult to transmit by the oral route, but comparatively easy by means of direct injection into the brain? We do not know.

Why is the onset of symptoms dose-dependent? The more prions a mouse ingests, the sooner it will show symptoms. The more copies of a prion gene that a mouse has, the more quickly it can get prion disease when injected with rogue prions. Why? We do not know.

Why is it safer to be heterozygous than homozygous? In other words, if you have, at word 129, a valine on one copy of the gene and a methionine on the other copy, why are you more resistant to prion diseases (except fatal familial insomnia) than somebody who has either two valines or two methionines? We do not know.

Why is the disease so picky? Mice cannot easily get hamster scrapie, nor vice versa. But a mouse deliberately equipped with a hamster prion gene will catch hamster scrapie from an injection of hamster brains. A mouse equipped with two different versions of human prion genes can catch two kinds of human disease, one like fatal familial insomnia and one like C J D . A mouse equipped with 2 8 0 G E N O M E