The Chemist Who Named the Bond

Karlsbad, the afternoon of 22 September 1902. The German Association of Naturalists and Physicians is holding its annual assembly, and the audience has already sat through a morning lecture by the Prague physiologist Franz Hofmeister arguing that proteins are long chains in which each amino acid is joined to the next by a linkage between an acid group and an amino group. Now Emil Fischer, the most celebrated organic chemist in Europe, stands up and makes the same argument from an entirely different direction — not from the behaviour of proteins, but from small molecules he has built himself. Two men, one morning apart, proposing the backbone of every protein on earth. Fischer also brings a word for the new compounds. He proposes to call them peptides.

That coincidence is where the vocabulary of a whole field enters the language. It is also a fair picture of how the idea arrived: not as a discovery but as a convergence, and one that then took another thirty years of argument to become settled fact. What Fischer added to it was not the hypothesis. It was the willingness to spend the rest of his working life making the molecules that would test it.

A prize he had already won

It is worth remembering how much of Fischer's reputation was made before any of this. Ten weeks after Karlsbad he was in Stockholm collecting the second Nobel Prize in Chemistry ever awarded, and it had nothing to do with peptides. It was for sugars and purines: the structures of glucose and its relatives, worked out by a combination of stubbornness and a reagent of his own invention, and the group of compounds that includes caffeine and uric acid. His Nobel lecture that December is a summary of that work, with the protein programme mentioned only as the direction he intends to go next 2.

He was fifty years old and starting again. The choice looks obvious in hindsight and was not: proteins in 1900 were widely regarded as beyond chemistry, colloidal substances too large and too ill-defined for the methods of the day to touch. Several serious people thought the question was not yet ripe. Fischer's judgement was that the way into an intractable problem is through the smallest possible piece of it, and that if proteins were chains, then two amino acids joined together were a protein in miniature and could be studied as one.

The first two links

The first dipeptide appeared in 1901, in a paper written with the young French chemist Ernest Fourneau, and the route to it says everything about the state of the art. They did not build glycylglycine by joining two molecules of glycine. They took the cyclic anhydride of glycine — a small ring formed of two glycines already joined at both ends — and boiled it with hydrochloric acid until the ring opened at one point and not the other, leaving a two-residue chain behind 1. The first peptide in the literature was obtained by breaking something rather than by making it.

There was a prehistory even to that. Theodor Curtius had prepared benzoyl-protected glycine chains twenty years earlier without recognising what class of compound he had, and one of the recurring lessons of this period is that making a molecule and knowing what you have made are separate achievements. What Fischer supplied was the frame that turned an oddity into a member of a family. A name does real work here. Once a class of compounds exists, its members can be counted, ordered, compared and predicted, and a result that would otherwise have been a curiosity in somebody's notebook becomes a data point in a programme.

Building chains deliberately required methods, and Fischer's group spent years inventing them. The workhorse was the halogen route: convert an amino acid to its acid chloride with the amino group replaced by a halogen, couple that to a second amino acid, then replace the halogen with an amino group using ammonia. It worked, and it produced peptides in a defined order, which was the essential point. But each round asked a great deal. The reactive intermediates were unstable. The products were often stubbornly soluble in water and difficult to crystallise, and crystallisation was the only purification worth the name. And the coupling conditions were harsh enough that residues could lose their optical form, converting a natural amino acid into the mirror-image version that no organism uses — an error invisible to the analysis of the day and fatal to any biological comparison.

The ceiling that lasted fifty years

By 1907 the group had reached its high-water mark: a chain of eighteen residues, fifteen of them glycine and three leucine, reported in the seventeenth paper of a series on polypeptide synthesis 3. It was, by a comfortable margin, the longest defined chain anyone had made, and it did the things a protein fragment ought to do — it gave the colour reactions associated with proteins and it was attacked by digestive enzymes, which was the closest thing available to evidence that the synthetic linkage and the natural one were the same linkage.

Then the programme stopped climbing. The reason is arithmetic and it is worth spelling out, because it governs the next half century of the field. Every addition to a chain in solution requires the product to be separated from the leftover reagents, the by-products and the failed material — and the longer the chain, the more the failures resemble the product, so the harder the separation and the greater the loss. Suppose each cycle recovers three-quarters of what went in, which for this chemistry would be doing well. Ten cycles leaves you about six per cent. Twenty leaves under half a per cent. Thirty leaves a trace you cannot characterise, from a starting quantity you cannot afford.

There was a second constraint, less obvious but just as binding. To join residues in a chosen order you must block the parts of each molecule that would otherwise react, and then unblock them afterwards without damaging what you have built. Fischer never had a protecting group that came off gently enough. The conditions required to remove his blocking groups tended to attack the chain as well. That particular lock was picked in 1932, when Max Bergmann — who had trained in Fischer's laboratory — and Leonidas Zervas introduced a group that could be stripped off under mild conditions, and it is no accident that serious progress on natural peptides dates from just after that.

What he thought a protein was

Fischer is often summarised as the man who proved proteins are polypeptides, and that is a little too tidy. He believed the chain structure and he built the evidence for it, but he was cautious in public about how far it went. His own chains were made mostly of glycine, the simplest and least protein-like of the amino acids, chosen because it behaved. He was aware that natural proteins were far larger than anything he could reach and that his molecules resembled them only in the linkage. He seems to have suspected that other kinds of bond might also appear in real proteins, and the historical record of the period shows a man defending a hypothesis rather than announcing a settled result 4.

He was right to be careful. The chain theory was seriously contested well into the 1930s, notably by chemists who preferred to see proteins as assemblies of small rings, and it took the arrival of X-ray crystallography and better molecular weight measurements to close the argument. Fischer did not live to see it; he died in 1919, worn down by the war, by the loss of two sons in it, and by years of handling reagents that a modern laboratory would treat with considerably more respect.

What survives is bigger than the molecules. The vocabulary is his: peptide, dipeptide, tripeptide, polypeptide, and the peptide bond itself, all still doing the same work in the same sense a century later. So is the method of attack — take the impossible object, find its smallest complete unit, make that unit deliberately, and let the rest follow — which is the strategy behind essentially every synthesis that came afterwards 5. And so is the ceiling, which turns out to be the more interesting inheritance, because the next fifty years of the field can be read as a sustained argument with it. Every advance from the 1930s to the 1960s is somebody finding a way to lose less material at each step.

The hormone syntheses of the 1950s were performed with tools Fischer would have recognised at once, and they were heroic precisely because the tools had not fundamentally changed. Nine residues was a Nobel Prize partly because eighteen residues of mostly glycine was still, half a century after Karlsbad, a respectable comparison. What finally broke the ceiling was not a better bond but a different way of holding on to the chain, and the story of the field really starts there — with a chemist who accepted the whole of Fischer's logic and refused only his washing-up.