DNA and genes

What does the thread look like up close? A DNA molecule is built like a ladder that has been twisted along its length. The two long sides are called strands, and they wind around each other like the rails of a spiral staircase. A spiral of this kind is a helix, and because DNA has two strands, its shape is called the double helix.

The shape was worked out in the early 1950s. In London in 1952, Rosalind Franklin and her student Raymond Gosling took an X-ray photograph of DNA whose cross-shaped pattern could only come from a helix. In Cambridge, James Watson and Francis Crick used her measurements to build a model that fitted, and they published it in 1953. The shape mattered because it showed at once how the molecule could be copied.

The rungs of the DNA ladder carry the message. Each rung is made of two small units called bases, one attached to each strand, meeting in the middle. There are only 4 kinds of base: adenine, thymine, guanine and cytosine, usually written A, T, G and C. Read along one strand and the bases form a long line of letters, such as ATTGCA. That order is the information. The DNA you inherited from one parent is about three billion letters long.

The bases do not pair at random. Adenine fits only with thymine, and guanine fits only with cytosine, because of their shapes and the weak bonds that form between them. So every rung is one of just two pairs, A with T or G with C. Know the letters on one strand and you know the letters on the other.

Every time a cell divides, each new cell needs its own full set of DNA, so the whole molecule must first be copied. The copying is called replication, and the pairing rule is what makes it possible.

First the two strands unzip: an enzyme works along the ladder, parting each rung down the middle. That leaves two single strands with their bases exposed. Each strand then serves as a template, a pattern to build against. Loose bases floating in the nucleus settle onto the exposed ones, A opposite T and G opposite C, and a second enzyme joins them into a new strand. The result is two double helixes where there was one. Each holds one old strand and one new, and both carry exactly the same sequence of letters as the original.

Copy three billion letters by hand and the page would be full of errors. A cell does far better: after all its checks, it makes about one mistake in every billion letters. How?

The first safeguard is the pairing rule itself. At every position, each base has one partner that fits, so a wrong base sits badly and is usually rejected before it is joined on. The second is proofreading. The enzyme that builds the new strand checks each base it has just added and, if the fit is wrong, cuts it out and tries again. A third system sweeps along once the copying is finished and repairs most of what was missed.

The few errors that slip past all three are permanent. Once copied, a changed letter is handed on to every cell that descends from that one.

DNA is not read as one long sentence. It is divided into stretches, each with a job, and a stretch that holds the instructions for one product is called a gene. What a typical gene describes is a protein, and humans have about twenty thousand genes of that kind.

Proteins do most of the work of a body. Haemoglobin, which carries oxygen in your blood, is a protein. So is keratin, which makes up your hair and nails, and so are the enzymes that digest your food. Every protein is a long chain of smaller units called amino acids, of which there are twenty kinds. The order of the amino acids decides how the chain folds, and its folded shape decides what the protein can do. So a gene’s task comes down to one thing: saying which amino acids go in which order.

How can four bases spell out twenty amino acids? One base per amino acid would not be enough, and nor would two. The cell reads the bases three at a time. Each group of three is a codon, and each codon stands for one amino acid. GCA, for instance, means the amino acid alanine. Four bases taken three at a time give sixty-four possible codons, more than enough, so most amino acids have several, and three codons mean stop.

The full table of which codon means which amino acid is the genetic code. It is nearly universal: bacteria, plants and people all read it the same way. That is why a human gene placed in a bacterium still works. Human insulin for people with diabetes has been made this way since 1982, by bacteria carrying the human insulin gene.

A gene sits in the nucleus, but proteins are built outside it, and the DNA itself never leaves. So the cell sends out a copy. The stretch of DNA holding the gene is unzipped, and a working copy of that one gene is made in RNA, a close chemical cousin of DNA. This copy is messenger RNA. It is a single strand, short and short-lived, and small enough to slip out through pores in the wall of the nucleus.

Outside, the message meets a ribosome, a tiny machine that a human cell holds by the million. The ribosome moves along the messenger RNA, reading it one codon at a time, and for each codon it adds the matching amino acid to a growing chain. When it reaches a stop codon it lets the finished chain go, and the chain folds into a working protein.

Two metres of thread in a space too small to see would tangle hopelessly if it were left loose. So the DNA is wound around spool-like proteins and coiled again and again into compact packages called chromosomes. Each chromosome is one enormously long DNA molecule carrying hundreds or thousands of genes.

A human body cell has 46 chromosomes, and they come in pairs: 23 from your mother and 23 from your father. In nearly every pair, the two chromosomes carry the same genes in the same order, though often in slightly different versions, which is why you can have your mother’s version of one gene and your father’s version of another. Egg and sperm cells are the exception. Each carries only 23, one from every pair, so that when the two join, the new cell is back to 46.

All of an organism’s DNA taken together, every chromosome, every gene and everything in between, is its genome. The human genome was first read almost in full in 2003, after thirteen years of work by laboratories around the world.

It held a surprise. Only about 2 per cent of it is made up of genes that code for proteins. The other 98 per cent is called non-coding DNA. Some of that has a clear job: it includes the switches that turn genes on and off, and the instructions for RNA molecules that are never made into protein. Much of the rest consists of repeated sequences whose function is still debated. It was once dismissed as junk. It is safer to call it the part of the manual we understand least.

A nerve cell and a skin cell look nothing alike and do entirely different work. Yet both carry the same genome, copied faithfully from the single cell you started as. How can identical instructions build such different cells?

The answer is that no cell uses all of its genes. Cells differ because they switch on different genes. A gene that is switched on is being copied into messenger RNA and made into protein; a gene that is switched off sits unread. This use of a gene is called gene expression. Certain cells in the pancreas express the gene for insulin, and cells in the roots of your hair express the genes for keratin, while each leaves the other’s genes silent. As an embryo grows, chemical signals between cells tell each one which genes to use, and most cells then keep that pattern for life.

Any change in the sequence of bases in DNA is a mutation. Some arise from the rare copying errors that escape repair. Others are caused from outside. Ultraviolet light from the Sun and the chemicals in tobacco smoke both damage DNA, and anything that raises the rate of change in this way is called a mutagen.

Most mutations make no difference, because they fall outside genes or leave the protein unchanged. But a single letter in the wrong place can matter a great deal. In sickle cell disease, a single base is changed in one of the genes for haemoglobin, which swaps one amino acid for another. The altered haemoglobin clumps and bends red blood cells into a stiff, curved shape that blocks small blood vessels. A mutation in a skin cell ends with that person. Only one in an egg or sperm cell can be inherited by a child.

Line up the DNA of any two people and about 999 letters in every thousand are the same. The remaining one in a thousand, a few million letters in all, is where people differ, and these inherited differences are called genetic variation. They go back to mutations in egg and sperm cells over many generations, and each child still adds some new ones of its own, about seventy on average. They are why no two people, apart from identical twins, have the same DNA.

That uniqueness is useful. In 1984 the British geneticist Alec Jeffreys found that certain repeated stretches of DNA vary so much in length from person to person that the pattern they make can pick out one individual. He called it a DNA fingerprint. A trace of blood or a single hair root can now identify one person, or show who is related to whom.