by Andrew Brown
The characteristic way in which amino acids join together, the peptide bond, had been established as early as 1902. Although this provided the clue to how long peptide chains might form, a group of chemists persisted in believing that there was a limit of about 30 amino acid residues or links per chain, and that these threads then stuck to each other in some non-specified fashion to form proteins. The fatal blow to the aggregate theory of the colloid chemists was administered by one of their own ranks, Theodor ‘The’ Svedberg. He was the Professor of Physical Chemistry in Uppsala and in the early twenties invented a new machine, the ultracentrifuge, which remains an invaluable tool for separating proteins, nucleic acids and sub-cellular particles by sedimentation. The ultracentrifuge relies on the simple principle that if an emulsion containing dispersed colloidal particles is subject to rapid rotation, the heavier colloidal particles will be separated from the lighter molecules of the solvent. This is exactly the way a milk separator works, where skimmed milk migrates to the periphery and lighter cream accumulates at the centre. Whereas the centrifugal force necessary to separate cream can be achieved at manual speeds, Svedberg’s ultracentrifuge could rotate at over 40,000 revolutions per minute and generate a field of force 100,000 times stronger than gravity.
Amongst other colloids, Svedberg examined very dilute solutions of haemoglobin, the oxygen carrying protein present in red blood cells. He expected to find multiple red boundaries corresponding to the constituent peptides of haemoglobin, but instead observed a single line. The sedimentation pattern produced in the ultracentrifuge proved that haemoglobin consisted of individual molecules of a weight of 68,000. He studied other proteins such as egg albumin, again with precise molecular weights emerging, suggesting strongly that proteins ‘must be regarded as substantially uniform chemical individuals’ and not aggregates of smaller peptides.19
While Svedberg was busy spinning protein solutions at unprecedented speeds in Sweden, a determined lecturer, James B. Sumner, at Cornell University in upstate New York was finishing a series of intricate experiments that provided the first unequivocal proof that enzymes were proteins. Working with the Jack bean, he painstakingly separated out a metabolically active substance, the enzyme urease, in such a pure form that he was able to crystallize it. Sumner had no assistants and worked with a paralysed arm, the legacy of a boyhood shooting accident; the research took nine years before he published his results in a short paper in 1926. He carefully demonstrated the urease crystals had catalytic activity that was destroyed (as was their protein structure) when they were digested by other proteolytic enzymes. A biochemist at Cambridge, Frederick Gowland Hopkins, had already concluded that ‘all metabolic tissue reactions are catalysed by enzymes’, but Sumner brought the first proof that an enzyme corresponded to a protein molecule and was not some extrinsic factor absorbed by the protein.20
By the summer of 1928, it is likely that all of the above information had gelled in Sage’s mind. Enzymes, the molecules that governed every chemical reaction inside cells, were proteins. Proteins were true macromolecules, not aggregates of smaller molecules, and could in many instances be crystallized so that they could be studied by X-ray diffraction. Mark and his group had made a start at modelling macromolecules with cellulose, and even though cellulose was a monotonous chain of glucose residues, Sage was unperturbed by the bewildering variety of protein structures. He might not be able to achieve a full analysis, but even a partial understanding of proteins would revolutionize our understanding of the nature of life and transform physiology and medicine.
The only scientist in England in 1928 to have made an X-ray analysis of natural protein fibres was Bill Astbury, at the Davy–Faraday Laboratory of the Royal Institution. In September, he wrote to Bernal in gloomy tones: ‘I don’t know whether you have heard the sad news, but it seems possible that I have abandoned crystallography. I have accepted the new lectureship in Textile Physics at Leeds University and am leaving the Davy Faraday at the end of the present month. I am making one last despairing effort to keep in touch with crystallography by attempting to do some X-ray work on wool etc. and I want to get some apparatus together.’21 He asked Bernal for advice on setting up a new crystallography department, and for recommendations on the best X-ray tube for general photographic research and the best microscope for studying the structure of fibres. The textile industry department was in fact the largest department at Leeds University, and thanks to the support of the Worshipful Company of Clothworkers, Astbury’s new endeavours were going to be quite generously supported, certainly with far more resources in terms of research manpower and money than Bernal would ever be offered at Cambridge.22
It was fitting that Astbury should turn to Bernal for expert opinion; after his fact-finding visit to the Continent, Sage possessed an unrivalled grasp of the whole subject of X-ray crystallography. Sir William Bragg wrote to thank him for his ‘interesting and helpful’ report on the European laboratories, and asked permission to circulate it to his son, Lawrence, in Manchester and to Henry Tizard, the senior government scientist at the Department of Science and Industrial Research (DSIR), who had responsibility for supporting university research likely to bring benefits to the country as a whole. Bernal persuaded Bragg that the time was ripe to impose some shape on the burgeoning subject, and Bragg arranged a one-day conference for interested parties at the Royal Institution on 15 March 1929 to follow a meeting of the Faraday Society. The Faraday Society meeting was on ‘Crystal structures and chemical constitution’ and was expected to attract all the leading workers from Europe as well as from England.
In the months leading up to these events at the Royal Institution, there was a flurry of correspondence between London, Cambridge, Leeds and Manchester about producing a comprehensive multi-author textbook of crystallography to be edited by Sir William. Bernal was to write about the space groups and other topics in conjunction with Astbury and Kathleen Lonsdale (Yardley). In answer to an enquiry from Astbury about the division of labour, Bernal indicated that he wished to deal with ‘the vector theory of space groups, including the recent work of Ewald on lattices and Hermann’s derivation of the space groups. I would also unless you are keen on doing it yourself, like to deal more or less critically with the structural theories of Weissenberg and the topological ideas of Niggli.’23 Sensing that Bernal’s contribution might be too theoretical, Astbury emphasized that the section on space groups should be of practical use to crystallographers on a daily basis in their laboratories, and while elegant methods of deduction were fine for illustrative purposes, it would be ‘the things which we deduce that are going to count in the end’.24 Bernal accepted Astbury’s views, but was ‘so busy now over this metal business and the general organisation of the Conference’25 that he had no time to write anything for their chapter.
The metal business was a paper on ‘The problem of the metallic state’26 that he was to deliver at the Faraday Society meeting. In this paper, Bernal pointed out that the usual definition of a metal was in terms of a material transmitting electricity by flow of electrons. By considering other properties of metals and their alloys, he attempted to uncover new ways of classifying the metallic state that would be scientifically fruitful. Naturally he started with X-ray crystallography data that showed all metallic substances examined to date contained atoms packed very close together. From this he deduced that the atomic diameter, or atomic volume, of any given element in the metallic state is constant: ‘one of the primary facts to be taken into account in any theory of the metallic state.’ It had been well known for thousands of years that pure metals tend to be soft and are easily bent, whereas alloys, formed by melting two metals together and allowing them to cool into a solid solution, are much harder. To Bernal, this mechanical property was as fundamental as electrical conductivity but it had been largely ignored by scientists. The major clue to this age-old puzzle, again came from the young science of X-ray crystallography, and Bernal reminded his audience that in an impure metal crystal, the re
flection of X-rays by a crystal plane, instead of being extremely sharp is noticeably diffuse. This was an indication that the atoms no longer lie in absolutely parallel planes: the crystal lattice is distorted so that smooth gliding of one plane over another is no longer possible and malleability is lost. He went to discuss how such distortion might also affect the electrical conductivity of metals and its bearing on the pioneering work being carried out at the Cavendish Laboratory on superconductivity by Peter Kapitza.
Bernal’s paper was packed with detailed information in chart form and what he modestly described as ‘rather hazardous speculation’. In a conclusion that would have earned Rutherford’s approval, Bernal stated that ‘the confusion which exists in this field is quite as much due to lack of systematic experimentation as to the intrinsic difficulties of theory’. He called for an organized research effort because ‘without such a framework of experimental fact, the quantum theory on which we rely ultimately to explain the nature of the metallic state, will be working in the dark and will pile up useless formulae with immense labour’. Although he published nothing more on metals, Bernal encouraged many researchers in this field, which has proved so important in solid state physics and metallurgy, and he continued to contribute generously to their work without accepting any formal acknowledgement. Linus Pauling credited Bernal’s extensive 1929 review of the metallic state with stimulating his own research into the electronic structure of metals such as iron, copper, nickel and gallium.27
There was one technique that Bernal discovered that was not picked up by anyone else, but was rediscovered twenty years later and became one of the essential foundations of the semiconductor industry. Kapitza asked him to prepare a very pure crystal of the metal bismuth because he wanted to study its electrical resistance in strong magnetic fields at very low temperatures.28 Bernal found that by passing a hot wire through the bismuth he could draw all the impurities to one end of the crystal – a technique that would become known years later as zone refining. Kapitza gave no acknowledgement to Bernal for his help,29 but it seems clear that the Cambridge crystallographers used the technique of zone-refining to prepare pure organic and metallic single crystals for X-ray analysis.30
At the Royal Institution conference that followed on from the Faraday Society meeting, three committees were formed to look into various aspects of crystallography, and Bernal was a member and secretary of all three. The various national schools of crystallography had adopted different conventions to describe the atomic structures of crystals revealed by X-ray diffraction, and published tables of data contained significant discrepancies. A proposal from Bernal served as the basis for discussion: ‘At present the results of structural crystallography, being published in a great diversity of journals and in a variety of technical terms and symbols, are neither (a) generally accessible nor (b) easily comparable.’31
He suggested that future publication should be restricted to journals with wide circulation and employ unified nomenclature and standardized tables.
Six months after the conference, Astbury wrote to ask about Bernal’s progress with their textbook chapter (‘Has anybody written anything yet?’32) and to complain that he and Kathleen Lonsdale were still held up by the lack of an agreed nomenclature or even any accepted convention on how axes of symmetry should be labelled. Bernal was working hard to resolve these issues with Paul Ewald and visited him in Stuttgart in September 1929. They failed to untangle the mess resulting from having at least three published and disparate sets of tables for the space groups, but the following summer managed to convene a working conference, which brought all the leading figures together in Zürich. The host for the conference was Paul Niggli, the Professor of Mineralogy and the editor-in-chief of the world’s most prestigious crystallography journal, Zeitschrift für Kristallographie. Niggli himself had published an influential textbook in 1919, in which he proposed one method of representing the space groups. Ewald and Bernal prepared an agenda for the meeting, which the Tables Committee worked through, with growing agreement. Ewald, who was chairing the committee, noticed that Niggli took little part and sat tight-lipped. On the third day, he suddenly exclaimed: ‘Gentlemen, you are stealing my book.’33 It seemed that the progress the Committee had been making towards a unified set of International Tables would be blocked by the immovable Niggli, and there was general despair. The project was saved by Bill Astbury who talked to Niggli on his own and, in Bernal’s words, used his ‘transparent honesty and disinterestedness’34 to persuade Niggli to withdraw his objection.
Another of Astbury’s qualities – his irresistible enthusiasm – had already advanced the state of knowledge about fibrous proteins. He expressed almost a reverence for the properties of wool, writing to Bernal, for instance: ‘The wool is very exciting.’35 One of his new colleagues at Leeds, J.B. Speakman, had been studying the physical and chemical properties of wool for some years, and in particular had amassed data on the elasticity of wool fibre. Starting in the summer of 1929, Astbury took over a hundred X-ray diffraction photographs of wool in its unextended and stretched states. In his report for the Clothworkers Department of the University, Astbury wrote that these photographs had convinced him that keratin, the protein matter of wool, showed a reversible change in its molecular structure on stretching. He labelled the unstretched protein, α-keratin, and the stretched form, β-keratin, and recorded the different periodicities of X-ray diffraction along the fibres. He also soon realized ‘that natural silk finds its counterpart, not in normal wool, but in stretched wool. There is a close similarity between the X-ray photograph of silk and of β-wool, from which we may make the deduction that silk does not show the long-range elastic properties of wool because it is already in the extended state…’36 His former colleague at the RI, A.L. Patterson captured the essence of Astbury’s discoveries in the following verse:
Amino acids in chains
Are the cause, so the X-ray explains,
Of the stretching of wool
And its strength when you pull,
And show why it shrinks when it rains
Apart from stretched wool and silk, Astbury soon found surprising similarities in his X-ray photographs of other natural materials. He wrote to Bernal again: ‘…let me tell you at once of the really exciting discovery I have made. Last Saturday I took a photo of a common fishing float, and it is identical with that of wool and hair, however fine!!! The fishing float is, of course, the quill of a porcupine, but isn’t it staggering that such a large epidermal structure – anything up to a foot long – should give an indistinguishable X-ray photo? The implications seem to me very great.’37 Astbury, through his ingenious interpretation of rather vague X-ray diffraction photographs, was beginning to realize that the mechanically durable and chemically unreactive protein, keratin, formed the animal appendages of hair, nails, horn and hoof, and he would soon find that it was the common supportive element in tissues like tendon.
Astbury presented this work at a Royal Society meeting in November 1930, and its subsequent publication, in Bernal’s opinion, represented ‘the key paper of all Astbury’s work’.38 In the paper, Astbury advanced a structural model for α-keratin that was analogous to the Meyer and Mark model for cellulose, but instead of calling for glucose rings to be linked together in a chain, he proposed that the repeating ring in keratin results from two amino acid molecules condensing (fusing together), and that this ring was broken open when the fibre was stretched into the β-form. To arrive at this model, Astbury and his associates had made some chemical investigations, hydrolysing the keratin of wool into its constituent amino acids. They knew that the elastic properties of hair and wool ‘are intimately bound up with the state of combination of the sulphur atoms’39 and therefore had concentrated their efforts on cystine, the only amino acid which contains a sulphur atom, and the most abundant amino acid by weight in wool. Although they had not been able to obtain large crystals of cystine, they found that ‘a micro-crystalline specimen of pure l-cystine gave a rem
arkably fine powder photograph’.40 If single crystals of amino acids could be grown and studied, they would give far more complete and detailed maps of atomic structure than the rather smeared images that Astbury was obtaining from fibres.
Astbury recognized that Bernal was the master of the single crystal technique, and by the spring of 1930 the two had made an arrangement for the division of work along these lines. In the same letter where he disclosed his excitement about the porcupine quill, Astbury wrote: ‘I have already collected specimens of most of the important amino-acids found in hair proteins and, as I told you, have got some sort of structure out for cystine… But if I send these amino-acids on to you, will you harden your ridiculously soft heart and stop doing odd jobs for other people? That would be an essential condition, of course, if we are to keep this job in our hands and out of reach of those damned Germans (God bless ’em!). By the way, I suppose you know that these protein amino-acids are all alpha-amino-acids. You ask me if I am going to stick to the proteins of animal fibres, but, of course, in a job like this it is impossible to stick to anything. It is not crystallography, but a kind of higher detective work, searching for clues, however faint, day after day…. I feel considerably safer with you on my side, so long as you are not developing my photos!’41
The work on amino acids soon gathered momentum. In October 1930, Astbury sent Bernal ‘pure specimens of L-cystine, leucine and glutamic acid. The cystine is in the form of a crystalline powder, and if you can grow big crystals, I believe you will be the first to do it. It may not be impossible… These three acids, together with glycine and alanine (and I should say arginine) are very important biologically, and, of course, the three I am sending are the most important constituents of hair… I shall look forward to getting some really authentic information from you about these amino-acids. It is a job worthy of your steel, and should prove extremely valuable from the biological point of view.’42 Bernal obtained other single crystals from the Biochemical Laboratory in Cambridge and made great inroads in just three short months: he was able to suggest, for example, why certain amino acids were soluble and others insoluble. From a single crystal of cystine, he took the most beautiful series of X-ray photographs revealing the symmetry of the molecule. Following Astbury’s hypothesis that α-keratin contained ring forms of amino acids linked together in a chain, Bernal made the first pass at the structure of the diketopiperazine ring that is formed from the condensation of two glycine molecules and was already known to occur in intact proteins. He wrote to Astbury that: ‘Diketopiperazine is fairly simple. It seems to consist of fairly flat hexagonal rings attached to each other sideways by electro-static forces and lying nearly parallel in fairly close sheets. When I get the intensity measurements I’ll be able to measure the dimensions of the ring fairly accurately. All these structures give the possibility of polymer linkages such as you suggest with the keratins, particularly the diacid substances. Both cystine and diketopiperazine yield horny substances which I am trying to get for investigation.’43
