Introduction

Crystallography ranks with astronomy as one of the oldest sciences. Crystals, in the form of precious stones and common minerals, have attractive properties on account of their symmetry and their refractive and reflective properties, which result in the undefinable quality called beauty. Natural philosophers have long pondered the unusual properties seen in the discontinuous surface morphologies of crystals. Hooke (1665) and Huygens (1690) came close to grasping the way repeating objects create discrete crystal faces with reproducible interfacial angles. The symmetry of mineral crystals was explored systematically in the 18th and 19th centuries by measuring the angles between crystal faces, leading to the classification into symmetry systems from triclinic to cubic and the construction of symmetry tables (Schoenflies, 1891; Hilton, 1903; Astbury et al., 1935) – the predecessors of today's International Tables. International Tables for Crystallography

1912 to the 1950s

It was not until the interpretation of the first X-ray diffraction experiments by Max von LaueLaue, M. von and Peter EwaldEwald, P. P. in 1912 that it was possible to ascertain the size of the repeating unit in simple crystals. Lawrence BraggBragg, W. L., encouraged by his father, William BraggBragg, W. H., recast the Laue equations into the physically intuitive form now known as `Bragg's lawBragg's law' (Bragg & Bragg, 1913). This set the stage for a large number of structure determinations of inorganic salts and metals. The discovery of simple structures (Bragg, 1913), such as that of NaCl, led to a good deal of acrimony, for crystals of such salts were shown to consist of a uniform distribution of positive and negative ions, rather than discrete molecules. These early structure determinations were based on trial and error (sometimes guided by the predictions of Pope and Barlow that were based on packing considerations) until a set of atomic positions could be found that satisfied the observed intensity distribution of the X-ray reflections. This gave rise to rather pessimistic estimates that structures with more than about four independent atomic parameters would not be solvable.

The gradual advance in X-ray crystallography required a systematic understanding and tabulation of space groups. Previously, only various aspects of three-dimensional symmetry operations appropriate for periodic lattices had been listed. Consequently, in 1935, the growing crystallographic community put together the first set of Internationale Tabellen Internationale Tabellen zur Bestimmung von Kristallstrukturen (Hermann, 1935), containing diagrams and information on about 230 space groups. After World War II, these tables were enlarged and combined with Kathleen LonsdaleLonsdale, K.'s structure-factor formulae (Lonsdale, 1936) in the form of International Tables International Tables for Crystallography Volume I (Henry & Lonsdale, 1952). Most recently, they have again been revised and extended in Volume A (Hahn, 2005).

Simple organic compounds started to be examined in the 1920s. Perhaps foremost among these is the structure of hexamethylbenzene by Kathleen LonsdaleLonsdale, K. (Lonsdale, 1928). She showed that, as had been expected, benzene had a planar hexagonal structure. Another notable achievement of crystallography was made by J. D. BernalBernal, J. D. in the early 1930s. He was able to differentiate between a number of possible structures for steroids by studying their packing arrangements in different unit cells (Bernal, 1933). Bernal (`Sage') had an enormous impact on English crystallographers in the 1930s. His character was immortalized by the novelist C. P. Snow in his book The Search (Snow, 1934). By the mid-1930s, J. Monteath RobertsonMonteath Robertson, J. and I. Woodward had determined the structure of nickel phthalocyanine (Robertson, 1935) using the heavy-atom method. This was a major crystallographic success and perhaps the first time that a crystallographer had succeeded in solving a structure when little chemical information was available.

Another event which had a major impact was the determination of the absolute hand of the asymmetric carbon atom of sodium tartrate by BijvoetBijvoet, J. M. (Bijvoet, 1949; Bijvoet et al., 1951). By indexing the X-ray reflections with a right-handed system, he showed that the breakdown of Friedel's law in the presence of an anomalous scattererAnomalous scattering (dispersion)sodium tartrateEarly structure determinationssodium tartrate was consistent with the asymmetric carbon atom having a hand in agreement with Fischer's convention. With that knowledge, together with the prior results of organic reaction analyses, the absolute hand of other asymmetric carbon atoms could be established. In particular, the absolute structure of naturally occurring amino acids and riboses was now determined.

Until the mid-1950s, most structure determinations were made using only projection data. This not only reduced the tremendous effort required for manual indexing and for making visual estimates of intensity measurements, but also reduced the calculation effort to almost manageable proportions in the absence of computing machines. However, the structure determination of penicillinPenicillinEarly structure determinationspenicillin (Crowfoot, 1948; Crowfoot et al., 1949), carried out during World War II by Dorothy HodgkinHodgkin, D. C. and Charles BunnBunn, C. W., employed some three-dimensional data. A further major achievement was the solution of the three-dimensional structure of vitamin B₁₂ by Dorothy Hodgkin and her colleagues (Hodgkin et al., 1957) in the 1950s. They first used a cobalt atom as a heavy atom on a vitamin B₁₂ Vitamin B₁₂Early structure determinationsvitamin B₁₂ fragment and were able to recognize the `corrin' ring structure. The remainder of the B₁₂ structure was determined by an extraordinary collaboration between Dorothy Hodgkin in Oxford and Kenneth TruebloodTrueblood, K. N. at UCLA in Los Angeles. While Dorothy's group did the data collection and interpretation, Ken's group performed the computing on the very early electronic Standard Western Automatic Computer (SWAC). Additional help was made available by the parallel work of J. G. White at Princeton University in New Jersey. This was at a time before the internet, before e-mail, before usable transatlantic telephones and before jet travel. Transatlantic, propeller-driven air connections had started to operate only a few years earlier.

Many technical advances were made in the 1930s that contributed to the rapidly increasing achievements of crystallography. W. H. BraggBragg, W. H. had earlier suggested (Bragg, 1915) the use of Fourier methodsFourier methods to analyse the periodic electron-density distribution in crystals, and this was utilized by his son, W. L. BraggBragg, W. L. (Bragg, 1929a,b). The relationship between a Fourier synthesis and a Fourier analysis demonstrated that the central problem in structural crystallography was in the phase. Computational devices to help plot this distribution were invented by Arnold BeeversBeevers, C. A. and Henry LipsonLipson, H. in the form of their `Beevers–Lipson stripsBeevers–Lipson strips' (Beevers & Lipson, 1934) and by J. Monteath Robertson with his `Robertson sorting boardRobertson sorting board' (Robertson, 1936). These devices were later supplemented by the XRAC electronic analogue machine of Ray PepinskyPepinsky, R. (Pepinsky, 1947) and mechanical analogue machines (McLachlan & Champaygne, 1946; Lipson & Cochran, 1953) until electronic digital computers came into use during the mid-1950s.

A. Lindo PattersonPatterson, A. L., inspired by his visit to England in the 1930s where he met Lawrence BraggBragg, W. L., Kathleen LonsdaleLonsdale, K. and J. Monteath RobertsonMonteath Robertson, J., showed how to use Fourier syntheses for structure determinations (Patterson, 1934, 1935). When the `Patterson' synthesisPatterson synthesis was combined with the heavy-atom method, and (later) with electronic computers, it transformed analytical organic chemistry. No longer was it necessary for teams of chemists to labour for decades on the structure determination of natural products. Instead, a single crystallographer could solve such a structure in a period of months.

Improvements in data-collection devices have also had a major impact. Until the mid-1950s, the most common method of measuring intensities was by visual comparison of reflection `spots' on films with a standard scale. However, the use of counters (used, for instance, by Bragg in 1912) was gradually automated and became the preferred technique in the 1960s. In addition, semi-automatic methods of measuring the optical densities along reciprocal lines on precession photographs were used extensively for early protein-structure determinations in the 1950s and 1960s.

The first investigations of biological macromoleculesEarly structure determinations

Leeds, in the county of Yorkshire, was one of the centres of England's textile industry and home to a small research institute established to investigate the properties of natural fibresFibre diffractionearly studies. W. T. AstburyAstbury, W. T. became a member of this institute after learning about X-ray diffraction from single crystals in Bragg's laboratory. He investigated the diffraction of X-rays by wool, silk, keratin and other natural fibrous proteins. He showed that the resultant patterns could be roughly classified into two classes, α and β, and that on stretching some, for example, wool, the pattern is converted from α to β (Astbury, 1933).

Purification techniques for globular proteinsEarly structure determinationsglobular proteins were also being developed in the 1920s and 1930s, permitting J. B. Sumner at Cornell University to crystallize the first enzyme, namely urease, in 1926. Not much later, in Cambridge, J. D. Bernal and his student, Dorothy Crowfoot (Hodgkin), investigated crystals of pepsinPepsinEarly structure determinationspepsin. The resultant 1934 paper in Nature (Bernal & Crowfoot, 1934) is quite remarkable because of its speed of publication and because of the authors' extraordinary insight. The crystals of pepsin were found to deteriorate quickly in air when taken out of their crystallization solution and, therefore, had to be contained in a sealed capillary tube for all X-ray experiments. This form of protein-crystal mounting remained in vogue until the 1990s when crystal-freezing techniques were introduced. But, most importantly, it was recognized that the pepsin diffraction pattern implied that the protein molecules have a unique structure and that these crystals would be a vehicle for the determination of that structure to atomic resolution. This understanding of protein structure occurred at a time when proteins were widely thought to form heterogeneous micelles, a concept which persisted another 20 years until Sanger was able to determine the unique amino-acid sequences of the two chains in an insulinInsulinEarly structure determinationsinsulin molecule (Sanger & Tuppy, 1951; Sanger & Thompson, 1953a,b).

Soon after Bernal and Hodgkin photographed an X-ray diffraction pattern of pepsin, Max PerutzPerutz, M. F. started his historic investigation of haemoglobinHaemoglobinEarly structure determinationshaemoglobin. Such investigations were, however, thought to be without hope of any success by most of the contemporary crystallographers, who avoided crystals that did not have a short (less than 4.5 Å) axis for projecting resolved atoms. Nevertheless, Perutz computed Patterson functions that suggested haemoglobin contained parallel α-keratin-like bundles of rods (Boyes-Watson et al., 1947; Perutz, 1949). Perutz was correct about the α-keratin-like rods, but not about these being parallel.

In Pasadena, PaulingPauling, L. (Pauling & Corey, 1951; Pauling et al., 1951) was building helical polypeptide models to explain Astbury's α patterns and perhaps to understand the helical structures in globular proteins, such as haemoglobin. Pauling, using his knowledge of the structure of amino acids and peptide bonds, was forced to the conclusion that there need not be an integral number of amino-acid residues per helical turn. He therefore suggested that the `α-helixα-Helices', with 3.6 residues per turn, would roughly explain Astbury's α pattern and that his proposed `β-sheetβ-Sheets ' structure should be related to Astbury's β pattern. Perutz saw that an α-helical structure should give rise to a strong 1.5 Å-spacing reflection as a consequence of the rise per residue in an α-helix (Perutz, 1951a,b). Demonstration of this reflection in horse hair, then in fibres of polybenzyl-L-glutamate, in muscle (with Hugh Huxley) and finally in haemoglobin crystals showed that Pauling's proposed α-helix really existed in haemoglobin and presumably also in other globular proteins. Confirmation of helix-like structures came with the observation of cylindrical rods in the 6 Å-resolution structure of myoglobin in 1957 (Kendrew et al., 1958) and eventually at atomic resolution with the 2 Å myoglobinMyoglobinEarly structure determinationsmyoglobin structure in 1959 (Kendrew et al., 1960). The first atomic resolution confirmation of Pauling's β structure did not come until 1966 with the structure determination of hen egg-white lysozyme (Blake, Mair et al., 1967).

Although the stimulus for the Cochran et al. (1952) analysis of diffraction from helical structures came from Perutz's studies of helices in polybenzyl-L-glutamate and their presence in haemoglobin, the impact on the structure determination of nucleic acids was even more significant. The events leading to the discovery of the double-helical structure of DNADNA have been well chronicled (Watson, 1968; Olby, 1974; Judson, 1979). The resultant science, often known exclusively as molecular biology, has created a whole new industry. Furthermore, the molecular-modelling techniques used by Pauling in predicting the structure of α-helices and β-sheets and by CrickCrick, F. H. C. and WatsonWatson, J. D. in determining the structure of DNA had a major effect on more traditional crystallography and the structure determinations of fibrous proteins, nucleic acids and polysaccharides.

Another major early result of profound biological significance was the demonstration by BernalBernal, J. D. and FankuchenFankuchen, I. in the 1930s (Bernal & Fankuchen, 1941) that tobacco mosaic virus (TMV)Tobacco mosaic virus (TMV)Early structure determinationstobacco mosaic virus had a rod-like structure. This was the first occasion where it was possible to obtain a definite idea of the architecture of a virus. Many of the biological properties of TMV had been explored by Wendell StanleyStanley, W. working at the Rockefeller Institute in New York. He had also been able to obtain a large amount of purified virus. Although it was not possible to crystallize this virus, it was possible to obtain a diffraction pattern of the virus in a viscous solution which had been agitated to cause alignment of the virus particles. This led Jim Watson (Watson, 1954) to a simple helical structure of protein subunits. Eventually, after continuing studies by Aaron KlugKlug, A., Rosalind FranklinFranklin, R., Ken HolmesHolmes, K. C. and others, the structure was determined at atomic resolution (Holmes et al., 1975), in which the helical strand of RNA was protected by the helical array of protein subunits.

Globular proteins in the 1950sEarly structure determinationsglobular proteins

In 1936, Max Perutz had joined Sir Lawrence BraggBragg, W. L. in Cambridge. Inspired in part by Keilin (Perutz, 1997), Perutz started to study crystalline haemoglobinHaemoglobinEarly structure determinationshaemoglobin. This work was interrupted by World War II, but once the war was over Perutz tenaciously developed a series of highly ingenious techniques. All of these procedures have their counterparts in modern `protein crystallography', although few today recognize their real origin.

The first of these methods was the use of `shrinkage' stages (Perutz, 1946; Bragg & Perutz, 1952). It had been noted by Bernal and Crowfoot (Hodgkin) in their study of pepsin that crystals of proteins deteriorate on exposure to air. Perutz examined crystals of horse haemoglobin after they were air-dried for short periods of time and then sealed in capillaries. He found that there were at least seven consecutive discrete shrinkage stages of the unit cell. He realized that each shrinkage stage permitted the sampling of the molecular transform at successive positions, thus permitting him to map the variation of the continuous transform. As he examined only the centric (h0l) reflections of the monoclinic crystals, he could observe when the sign changed from 0 to π in the centric projection (Fig. 1.2.4.1). Thus, he was able to determine the phases (signs) of the central part of the (h0l) reciprocal lattice. This technique is essentially identical to the use of diffraction data from different unit cells for averaging electron density in the `modern' molecular replacement method. In the haemoglobin case, Patterson projections had shown that the molecules maintained their orientation relative to the a axis as the crystals shrank, but in the more general molecular replacement case, it is necessary to determine the relative orientations of the molecules in each cell.

The second of Perutz's techniques depended on observing changes in the intensities of low-order reflections when the concentration of the dissolved salts (e.g. Cs₂SO₄) in the solution between the crystallized molecules was altered (Boyes-Watson et al., 1947; Perutz, 1954). The differences in structure amplitude, taken together with the previously determined signs, could then map out the parts of the crystal unit cell occupied by the haemoglobin molecule. In many respects, this procedure has its equivalent in `solvent flattening' used extensively in `modern' protein crystallography.

The third of Perutz's innovations was the isomorphous replacementIsomorphous replacement method (Green et al., 1954). The origin of the isomorphous replacement method goes back to the beginnings of X-ray crystallography when Bragg compared the diffracted intensities from crystals of NaCl and KCl. J. Monteath Robertson explored the procedure a little further in his studies of phthalocyanines. Perutz used a well known fact that dyes could be diffused into protein crystals, and, hence, heavy-atom compounds might also diffuse into and bind to specific residues in the protein. Nevertheless, the sceptics questioned whether even the heaviest atoms could make a measurable difference to the X-ray diffraction pattern of a protein. Perutz therefore developed an instrument which quantitatively recorded the blackening caused by the reflected X-ray beam on a film. He also showed that the effect of specifically bound atoms could be observed visually on a film record of a diffraction pattern. In 1953, this resulted in a complete sign determination of the (h0l) horse haemoglobin structure amplitudes (Green et al., 1954). However, not surprisingly, the projection of the molecule was not very interesting, making it necessary to extend the procedure to noncentric, three-dimensional data. It took another five years to determine the first globular protein structure to near atomic resolution.

In 1950, David HarkerHarker, D. was awarded one million US dollars to study the structure of proteins. He worked first at the Brooklyn Polytechnic Institute in New York and later at the Roswell Park Cancer Institute in Buffalo, New York. He proposed to solve the structure of proteins on the assumption that they consisted of `globs' which he could treat as single atoms; therefore, he could solve the structure by using his inequalities (Harker & Kasper, 1947), i.e., by direct methods. He was aware of the need to use three-dimensional data, which meant a full phase determination, rather than the sign determination of two-dimensional projection data on which Perutz had concentrated. Harker therefore decided to develop automatic diffractometers, as opposed to the film methods being used at Cambridge. In 1956, he published a procedure for plotting the isomorphous data of each reflection in a simple graphical manner that allowed an easy determination of its phase (Harker, 1956). Unfortunately, the error associated with the data tended to create a lot of uncertainty.

In the first systematic phase determination of a protein, namely that of myoglobin, phase estimates were made for about 400 reflections. In order to remove subjectivity, independent estimates were made by KendrewKendrew, J. C. and Bragg by visual inspection of the Harker diagram for each reflection. These were later compared before computing an electron-density map. This process was put onto a more objective basis by calculating phase probabilities, as described by BlowBlow, D. M. & CrickCrick, F. H. C. (1959) and DickersonDickerson, R. E. et al. (1961).

One problem with the isomorphous replacement method was the determination of accurate parameters that described the heavy-atom replacements. Centric projections were a means of directly determining the coordinates, but no satisfactory method was available to determine the relative positions of atoms in different derivatives when there were no centric projections. In particular, it was necessary to establish the relative y coordinates for the heavy-atom sites in the various derivatives of monoclinic myoglobin and in monoclinic horse haemoglobin. Perutz (1956) and Bragg (1958) had each proposed solutions to this problem, but these were not entirely satisfactory. Consequently, it was necessary to average the results of different methods to determine the 6 Å phases for myoglobin. However, this problem was solved satisfactorily in the structure determination of haemoglobin by using an Patterson-like synthesis in which the vectors between atoms in the two heavy-atom compounds, H1 and H2, produce negative peaks (Rossmann, 1960). This technique also gave rise to the first least-squares refinement procedure to determine the parameters that define each heavy atom.

Perutz used punched cards to compute the first three-dimensional Patterson map of haemoglobin. This was a tremendous computational undertaking. However, the first digital electronic computers started to appear in the early to mid-1950s. The EDSAC1 and EDSAC2 machines were built in the Mathematical Laboratory of Cambridge University. EDSACEDSAC1 and 21 was used by John Kendrew for the 6 Å-resolution map of myoglobin (Bluhm et al., 1958). EDSAC2 came on-line in 1958 and was the computer on which all the calculations were made for the 5.5 Å map of haemoglobin (Cullis et al., 1962) and the 2.0 Å map of myoglobin. It was the tool on which many of the now well established crystallographic techniques were initially developed. By about 1960, the home-built, one-of-a-kind machines were starting to be replaced by commercial machines. Large mainframe IBM computers (704, 709 etc.), together with FORTRAN as a symbolic language, became available.

The first protein structures (1957 to the 1970s)Early structure determinations

By the time three-dimensional structures of proteins were being solved, Linderström-LangLinderström-Lang, K. U. (Linderström-Lang & Schellman, 1959) had introduced the concepts of `primary', `secondary' and `tertiary' structures, providing a basis for the interpretation of electron-density maps. The first three-dimensional protein structure to be solved was that of myoglobinMyoglobinEarly structure determinationsmyoglobin at 6 Å resolution (Fig. 1.2.5.1) in 1957 (Kendrew et al., 1958). It clearly showed sausage-like features which were assumed to be α-helices. The iron-containing haem group was identified as a somewhat larger electron-density feature. The structure determination of haemoglobinHaemoglobinEarly structure determinationshaemoglobin at 5.5 Å resolution in 1959 (Cullis et al., 1962) showed that each of its two independent chains, α and β, had a fold similar to that of myoglobin and, thus, suggested a divergent evolutionary process for oxygen transport molecules. These first protein structures were mostly helical, features that could be recognized readily at low resolution. Had the first structures been of mostly β structure, as is the case for pepsin or chymotrypsin, history might have been different.

The absolute hand of the haemoglobinHaemoglobinEarly structure determinationshaemoglobin structure was determined using anomalous dispersion (Cullis et al., 1962) in a manner similar to that used by Bijvoet. This was confirmed almost immediately when a 2 Å-resolution map of myoglobin was calculated in 1959 (Kendrew et al., 1960). By plotting the electron density of the α-helices on cylindrical sections (Fig. 1.2.5.2), it was possible to see not only that the Pauling prediction of the α-helix structure was accurately obeyed, but also that the atoms were consistent with laevo amino acids and that all eight helices were right-handed on account of the steric hindrance that would occur between the atom and carbonyl oxygen in left-handed helices.

The first enzyme structure to be solved was that of lysozymeLysozymeEarly structure determinationslysozyme in 1965 (Blake et al., 1965), following a gap of six years after the excitement caused by the discovery of the globin structures. Diffusion of substrates into crystals of lysozyme showed how substrates bound to the enzyme (Blake, Johnson et al., 1967), which in turn suggested a catalytic mechanism and identified the essential catalytic residues.

From 1965 onwards, the rate of protein-structure determinations gradually increased to about one a year: carboxypeptidaseCarboxypeptidaseEarly structure determinationscarboxypeptidase (Reeke et al., 1967), chymotrypsinChymotrypsinEarly structure determinationschymotrypsin (Matthews et al., 1967), ribonucleaseRibonucleaseEarly structure determinationsribonuclease (Kartha et al., 1967; Wyckoff et al., 1967), papainPapainEarly structure determinationspapain (Drenth et al., 1968), insulinInsulinEarly structure determinationsinsulin (Adams et al., 1969), lactate dehydrogenaseLactate dehydrogenaseEarly structure determinationslactate dehydrogenase (Adams et al., 1970) and cytochrome c Cytochrome cEarly structure determinationscytochrome c (Dickerson et al., 1971) were early examples. Every new structure was a major event. These structures laid the foundation for structural biology. From a crystallographic point of view, Drenth's structure determination of papain was particularly significant in that he was able to show an amino-acid sequencing error where 13 residues had to be inserted between Phe28 and Arg31, and he showed that a 38-residue peptide that had been assigned to position 138 to 176 needed to be transposed to a position between the extra 13 residues and Arg31.

The structures of the globins had suggested that proteins with similar functions were likely to have evolved from a common precursor and, hence, that there might be a limited number of protein folding motifs. Comparison of the active centres of chymotrypsin and subtilisinSubtilisin showed that convergent evolutionary pathways could exist (Drenth et al., 1972; Kraut et al., 1972).

The variety of structures that were being studied increased rapidly. The first tRNA structures were determined in the 1960s (Kim et al., 1973; Robertus et al., 1974), the first spherical virus structureVirusesearly studiesEarly structure determinationsviruses was published in 1978 (Harrison et al., 1978) and the photoreaction centre membrane protein structureMembrane proteinsphotosynthetic reaction centresEarly structure determinationsmembrane proteins appeared in 1985 (Deisenhofer et al., 1985). The rate of new structure determinations has continued to increase exponentially. In 1996, about one new structure was published every day. Partly in anticipation and partly to assure the availability of results, the Brookhaven Protein Data Bank (PDB) was brought to life at the 1971 Cold Spring Harbor Meeting (H. Berman & J. L. Sussman, private communication). Initially, it was difficult to persuade authors to submit their coordinates, but gradually this situation changed to one where most journals require coordinate submission to the PDB, resulting in today's access to structural results via the World Wide Web.

The growth of structural information permitted generalizations, such as that β-sheets have a left-handed twist when going from one strand to the next (Chothia, 1973) and that `cross-over' β-α-β turns were almost invariably right-handed (Richardson, 1977). These observations and the growth of the PDB have opened up a new field of science. Among the many important results that have emerged from this wealth of data is a careful measurement of the main-chain dihedral angles, confirming the predictions of Ramachandran (Ramachandran & Sasisekharan, 1968), and of side-chain rotamers (Ponder & Richards, 1987). Furthermore, it is now possible to determine whether the folds of domains in a new structure relate to any previous results quite conveniently (Murzin et al., 1995; Holm & Sander, 1997).

Technological developments (1958 to the 1980s)

In the early 1960s, there were very few who had experience in solving a protein structure. Thus, almost a decade passed after the success with the globins before there was a noticeable surge of new structure reports. In the meantime, there were persistent attempts to find alternative methods to determine protein structure.

BlowBlow, D. M. & RossmannRossmann, M. G. (1961) demonstrated the power of the single isomorphous replacement method. While previously it had been thought that it was necessary to have at least two heavy-atom compounds, if not many more, they showed that a good representation of the structure of haemoglobin could have been made by using only one good derivative. There were also early attempts at exploiting anomalous dispersion for phase determination. Rossmann (1961) showed that anomalous differences could be used to calculate a `Bijvoet PattersonBijvoet Patterson mapPatterson mapsBijvoet' from which the sites of the anomalous scatterers (and, hence, heavy-atom sites) could be determined. Blow & Rossmann (1961), North (1965) and Matthews (1966) used anomalous-dispersion data to help in phase determination. Hendrickson stimulated further interest by using Cu Kα radiation and employing the anomalous effect of sulfur atoms in cysteines to solve the entire structure of the crambin molecule (Hendrickson & Teeter, 1981). With today's availability of synchrotrons, and hence the ability to tune to absorption edges, these early attempts to utilize anomalous data have been vastly extended to the powerful multiple-wavelength anomalous dispersion (MAD) method (Hendrickson, 1991). More recently, the generality of the MAD technique has been greatly expanded by using proteins in which methionine residues have been replaced by selenomethionine, thus introducing selenium atoms as anomalous scatterers.

Another advance was the introduction of the `molecular replacementMolecular replacement' technique (Rossmann, 1972). The inspiration for this method arose out of the observation that many larger proteins (e.g. haemoglobin) are oligomers of identical subunits and that many proteins can crystallize in numerous different forms. Rossmann & Blow (1962) recognized that an obvious application of the technique would be to viruses with their icosahedral symmetry. They pointed out that the symmetry of the biological oligomer can often be, and sometimes must be, `noncrystallographic' or `local', as opposed to being true for the whole infinite crystal lattice. Although the conservation of folds had become apparent in the study of the globins and a little later in the study of dehydrogenases (Rossmann et al., 1974), in the 1960s the early development of the molecular replacement technique was aimed primarily at ab initio phase determination (Rossmann & Blow, 1963; Main & Rossmann, 1966; Crowther, 1969). It was only in the 1970s, when more structures became available, that it was possible to use the technique to solve homologous structures with suitable search models. Initially, there was a good deal of resistance to the use of the molecular replacement technique. Results from the rotation function were often treated with scepticism, the translation problem was thought to have no definitive answer, and there were excellent reasons to consider that phasing was impossible except for centric reflections (Rossmann, 1972). It took 25 years before the full power of all aspects of the molecular replacement technique was fully utilized and accepted (Rossmann et al., 1985).

The first real success of the rotation function was in finding the rotational relationship between the two independent insulin monomers in the P3 unit cell (Dodson et al., 1966). CrowtherCrowther, R. A. produced the fast rotation function, which reduced the computational times to manageable proportions (Crowther, 1972). Crowther (1969) and Main & Rossmann (1966) were able to formulate the problem of phasing in the presence of noncrystallographic symmetry in terms of a simple set of simultaneous complex equations. However, real advances came with applying the conditions of noncrystallographic symmetry in real space, which was the key to the solution of glyceraldehyde-3-phosphate dehydrogenaseGlyceraldehyde-3-phosphate dehydrogenaseEarly structure determinationsglyceraldehyde-3-phosphate dehydrogenase (Buehner et al., 1974), tobacco mosaic virus disk protein (Bloomer et al., 1978) and other structures, aided by Gerard Bricogne's program for electron-density averaging (Bricogne, 1976), which became a standard of excellence.

No account of the early history of protein crystallography is complete without a mention of ways of representing electron density. The 2 Å map of myoglobin was interpreted by building a model (on a scale of 5 cm to 1 Å with parts designed by Corey and Pauling at the California Institute of Technology) into a forest of vertical rods decorated with coloured clips at each grid point, representing the height of the electron density (Fig. 1.2.6.1). Later structures, such as those of lysozyme and carboxypeptidase, were built with `Kendrew' modelsKendrew models (2 cm to 1 Å) based on electron-density maps displayed as stacks of large Plexiglas sheets. A major advance came with Fred Richards' invention of the optical comparator (a `Richards boxRichards box' or `Fred's folly') in which the model was optically superimposed onto the electron density by reflection of the model in a half-silvered mirror (Richards, 1968). This allowed for convenient fitting of model parts and accurate placement of atoms within the electron density. The Richards box was the forerunner of today's computer graphics, originally referred to as an `electronic Richards box'. The development of computer graphics for model building was initially met with reservation, but fortunately those involved in these developments persevered. Various programs became available for model building in a computer, but the undoubted champion of this technology was FRODO, written by Alwyn Jones (Jones, 1978).

Meetings

The birth of protein crystallography in the 1950s coincided with the start of the jet age, making attendance at international meetings far easier. Indeed, the number and variety of meetings have proliferated as much as the number and variety of structures determined. A critical first for protein crystallography was a meeting held at the Hirschegg ski resort in Austria in 1966. This was organized by Max Perutz (Cambridge) and Walter HoppeHoppe, W. (München). About 40 scientists from around the world attended, as well as a strong representation of students (including Robert HuberHuber, R.) from the Münich laboratory. The first Hirschegg meeting occurred just after the structure determination of lysozyme, which helped lift the cloud of pessimism after the long wait for a new structure since the structures of the globins had been solved in the 1950s. The meeting was very much a family affair where most attendees stayed an extra few days for additional skiing. The times were more relaxed in comparison with those of today's jet-setting scientists flying directly from synchrotron to international meeting, making ever more numerous important discoveries. A second Hirschegg meeting occurred two years later, but this time the number of participants had doubled. By 1970, the meeting had to be transferred to the village of Alpbach, which had more accommodation; however, most of the participants still knew each other.

Another set of international meetings (or schools, as the Italians preferred to call them) was initiated by the Italian crystallographers in 1976 at Erice, a medieval hilltop town in Sicily. These meetings have since been repeated every six years. The local organizer was Lodovico Riva di Sanseverino, whose vivacious sensitivity instilled a feeling of international fellowship into the rapidly growing number of structural biologists. The first meeting lasted two whole weeks, a length of time that would no longer be acceptable in today's hectic, competitive atmosphere.

It took time for the staid organizers of the IUCr triennial congress to recognize the significance of macromolecular structure. Thus, for many years, the IUCr meetings were poorly attended by structural biologists. However, recent meetings have changed, with biological topics representing about half of all activities. Nevertheless, the size of these meetings and their lack of focus have led to numerous large and small specialized meetings, from small Gordon Conferences and East and West Coast crystallography meetings in the USA to huge international congresses in virology, biochemistry and other sciences.

The publication of this volume by the International Union of Crystallography, the first volume of International Tables devoted to macromolecular crystallography, strongly attests to the increasing importance of this vital area of science.