Matt Ridley

entirely different genes. Cerebellar ataxia is one.

There is even a bizarre report that a long C A G repeat deliberately inserted into a random gene in a mouse caused a late-onset, neurological disease rather like Huntington's disease. C A G repeats may therefore cause neurological disease whatever the gene in which they appear. Moreover, there are other diseases of nerve degeneration caused by other stuttering repeats of 'words' and in every case the repeated 'word' begins with C and ends in G. Six different C A G

diseases are known. C C G or C G G repeated more than 200 times near the beginning of a gene on the X chromosome causes 'fragile X', a variable but unusually common form of mental retardation (fewer than sixty repeats is normal; up to a thousand is possible).

C T G repeated from fifty to one thousand times in a gene on chromosome 19 causes myotonic dystrophy. More than a dozen human diseases are caused by expanded three-letter word repeats -

the so-called polyglutamine diseases. In all cases the elongated protein has a tendency to accumulate in indigestible lumps that cause their cells to die. The different symptoms are caused by the fact that different genes are switched on in different parts of the body.

What is so special about the 'word' C*G, apart from the fact that 6 0 G E N O M E

it means glutamine? A clue comes from a phenomenon known as anticipation. It has been known for some time that those with a severe form of Huntington's disease or fragile X are likely to have children in whom the disease is worse or begins earlier than it did in themselves. Anticipation means that the longer the repetition, the longer it is likely to grow when copied for the next generation. We know that these repeats form little loopings of D N A called hairpins.

The D N A likes to stick to itself, forming a structure like a hairpin, with the Cs and Gs of the C*G 'words' sticking together across the pin. When the hairpins unfold, the copying mechanism can slip and more copies of the word insert themselves.8

A simple analogy might be helpful. If I repeat a word six times in this sentence — cag, cag, cag, cag, cag, cag — you will count it fairly easily. But if I repeat it thirty-six times - cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag, cag — I am willing to bet you lose count. So it is with the D N A . The more repeats there are, the more likely the copying mechanism is to insert an extra one. Its finger slips and loses its place in the text. An alternative (or possibly additional) explanation is that the checking system, called mismatch repair, is good at catching small changes, but not big ones in C*G repeats.9

This may explain why the disease develops late in life. Laura Mangiarini at Guy's Hospital in London created transgenic mice, equipped with copies of part of the Huntington's gene that contained more than one hundred repeats. As the mice grew older, so the length of the gene increased in all their tissues save one. Up to ten extra C A G 'words' were added to it. The one exception was the cerebellum, the hindbrain responsible for controlling movement.

The cells of the cerebellum do not need to change during life once the mice have learnt to walk, so they never divide. It is when cells and genes divide that copying mistakes are made. In human beings, the number of repeats in the cerebellum falls during life, though it increases in other tissues. In the cells from which sperm are made, the C A G repeats grow, which explains why there is a relationship F A T E 6 1

between the onset of Huntington's disease and the age of the father: older fathers have sons who get the disease more severely and at a younger age. (Incidentally, it is now known that the mutation rate, throughout the genome, is about five times as high in men as it is in women, because of the repeated replication needed to supply fresh sperm cells throughout life.)10

Some families seem to be more prone to the spontaneous appearance of the Huntington's mutation than others. The reason seems to be not only that they have a repeat number just below the threshold (say between twenty-nine and thirty-five), but that it jumps above the threshold about twice as easily as it does in other people with similar repeat numbers. The reason for that is again a simple matter of letters. Compare two people: one has thirty-five C A G s followed by a bunch of C C A s and C C G s . If the reader slips and adds an extra C A G , the repeat number grows by one. The other person has thirty-five C A G s , followed by a CAA then two more C A G s . If the reader slips and misreads the C A A as a C A G , the effect is to add not one but three to the repeat number, because of the two C A G s already waiting.11

Though I seem to be getting carried away, and deluging you with details about C A G s in the huntingtin gene, consider: almost none of this was known five years ago. The gene had not been found, the C A G repeat had not been identified, the huntingtin protein was unknown, the link with other neurodegenerative diseases was not even guessed at, the mutation rates and causes were mysterious, the paternal age effect was unexplained. From 1872 to 1993 virtually nothing was known about Huntington's disease except that it was genetic. This mushroom of knowledge has grown up almost over-night since then, a mushroom vast enough to require days in a library merely to catch up. The number of scientists who have published papers on the Huntington's gene since 1993 is close to 100. All about one gene. One of 60,000-80,000 genes in the human genome. If you still need convincing of the immensity of the Pan-dora's box that James Watson and Francis Crick opened that day in 1953, the Huntington's story will surely persuade you. Compared 6 2 G E N O M E

with the knowledge to be gleaned from the genome, the whole of the rest of biology is but a thimbleful.

And yet not a single case of Huntington's disease has been cured.

The knowledge that I celebrate has not even suggested a remedy for the affliction. If anything, in the heartless simplicity of the C A G

repeats, it has made the picture look even bleaker for those seeking a cure. There are 100 billion cells in the brain. How can we go in and shorten the C A G repeats in the huntingtin genes of each and every one?

Nancy Wexler relates a story about a woman in the Lake Maracaibo study. She came to Wexler's hut to be tested for neurological signs of the disease. She seemed fine and well but Wexler knew that small hints of Huntington's can be detected by certain tests long before the patient herself sees signs. Sure enough this woman showed such signs. But unlike most people, when the doctors had finished their examination, she asked them what their conclusion was. Did she have the disease? The doctor replied with a question: What do you think? She thought she was all right. The doctors avoided saying what they thought, mentioning the need to get to know people better before they gave diagnoses. As soon as the woman left the room, her friend came rushing in, almost hysterical.

What did you tell her? The doctors recounted what they had said.

'Thank God', replied the friend and explained: the woman had said to the friend that she would ask for the diagnosis and if it turned out that she had Huntington's disease, she would immediately go and commit suicide.