11. MACROMOLECULES INTRODUCTION BY W. T. ASTBURY Dept. of Biomolecular Structure, University of Leeds Received 15th May, 1958 It is not easy to follow after the comprehensive General Introduction with which Bernal has already set the scientific stage of this meeting, but I believe there still remain certain rather more specialized and personal considerations to which I can usefully draw your attention. As I see it, this Discussion is particularly opportune in the stimulating part it is bound to play in bringing together, and at last in intelligible focus, a number of experimental and theoretical approaches that have long been groping for such contacts. Of these “ co-ordinations ”, so to call them, I should like to say something first in relation to recent progress to- wards recognizing the configurations of polypeptide chains in proteins.To the well established, if still not yet completely elucidated, a- and ,%con- figurations there is to be added now, what was always there by implication, I suppose, especially in solution, the randomly-coiled state. While calculations, for example by Schellman,l have been carried out with varying promise on the probability of interchange between cc-helix and randomly-coiled, notable suc- cesses have been achieved on the experimental side by the application of light- scattering and optical-rotation tests and by deuteration studies. The results obtained along these lines by Doty and his collaborators at Harvard, by Elliott and his collaborators at the Courtaulds Laboratory, and by the Linderstrrm-lang school at the Carlsberg Laboratory are most impressive, and there will be plenty of references to them in the papers and discussions to follow.The exciting stage has now been reached of extrapolating from the synthetic polypeptides, where reversible transition from a-helices (right-handed if the residues are left-handed) to random coils has been demonstrated conclusively, to the globular proteins, and already with these too the strong indications are-as suggested also by X-ray diffraction studies-that they can be compounded at least in part of cr-helices; and, moreover, the proportions of the latter can again be reversibly modified to some extent. The rest of the globular protein molecule though-the bends where the helical runs join up and the so-called tertiary structure as a whole-is in general still pretty much as vague as ever, and in any event is presumably more or less peculiar to each species, being defined by further linkages as yet mostly only guessed at.Kendrew’s X-ray model of the myoglobin molecule offers a fascinating glimpse of the sort of thing to be expected. I want to recall old evidence and present some revitalizing new evidence for another polypeptide configuration in proteins that I have the feeling may eventually turn out to be scarcely less significant than the a- and P-configurations-meaning by the latter the more familiar “parallel p” as produced from the a-form by stretching; for this new-old configuration is also a @-form, what we call “ cross /3 ”, in which the 4.65 A reflection associated with interchain CO .. . HN linkages is not in its usual place on the equator of the fibre diagram but is on the meridian. It is the form we have come to connect specially with the phenomenon of “ super- contraction ”, that property of fibres of the k-m-e-f group of being able to contract under appropriate treatment to a length even shorter than the a-configuration. It was in the early thirties that Mrs. Dickinson and I first obtained a cross-/3 80W. T. ASTBURY 81 diagram, from a frog sartorius muscle that had been immersed in water at 60” or more. At that time we had little idea what it meant, but soon afterwards, while working with Bailey in the X-ray investigation2 which revealed how globular proteins unfold on denaturation and can then be drawn out into fibres giving a diagram of the type of &keratin and /3-myosin, we were surprised to find the same sort of thing again on stretching (by about 100 %) thin strips of “poached” egg-white ; surprised, because stretched denatured edestin, for example, had given a normal P-diagram, with the polypeptide chains running in the direction of stretching.We could only interpret the egg-albumin effect in terms of bundles of ,&chains which were actually thicker than the chains were long, and which were therefore oriented by stretching so as to leave the chains lying transversely. All through the thirties, too, we were developing the idea, never really abandoned, since, that muscular contraction, in the last resort, is a specialized manifestation in myosin, of the power of supercontraction common to the whole of the k-m-e-f group, while at the same time we were led to the view that the phenomenon was the result of the polypeptide chains being able to fall into transverse folds more pronounced even than those that defined the cc-form-this was in the bad old days before the %-helix, be it remembered, when the favoured scheme for the w o n - figuration was a square fold which was in effect the shortest possible lateral /?-fold, and which there was little difficulty therefore in imagining to be able to lengthen t ransversely.3 We owe to Rudall4 the first systematic X-ray examination of supercontraction, particularly in epidermin, the fibrous protein he had extracted from the epidermis, and it was he who pointed out in 1941 that the transverse folds that had been proposed for the supercontracted state would most likely lead to a cross-p pattern (see fig.1). He also showed how the cross-13 form may be reconverted to the cc-form by the action of saturated urea, for instance. Since then, a number of people-Woods, Whewell, Mercer, Sikorski and others-have continued the attack from a variety of angles, and it still goes on, indeed with the added impetus lately that “ natural ” cross-P structures have been discovered. These latter have made all the difference in the world to a problem that was never lacking in interest, to say the Ieast, and I should like to explain now how beautifully they have come to link up with the important discoveries by Keller also to be reported upon at this meeting.The two known “ natural ” examples of cross-/? structures are provided by bacterial flagella5 and the egg-stalk of the green lace-wing fly Chrysopa.6 The former, to judge by their X-ray diagrams and motility properties, are a kind of monomolecular muscles : the diffraction pattern comprises not only an a-diagram with a higher periodicity (about 410 A) like that of skeletal muscle, but a prominent 4-65A reflection on the meridian besides (fig. 3a). This, of course, represents the combination we had been postulating all along as the ultimate molecular basis of muscular activity, even though we had never actually observed it in whole muscle undcr conditions of physiological contraction. No wonder then that we suggested with increased confidence that the rhythmic localized shortening and lengthening implied by the bending movements of flagella rest too on an interchange between the 01- and supercontracted configurations.Thcre is still a considerable element of speculation, it must be confessed, in this interpretation of the structure and movements of bacterial flagella, but the case of Chrysopa is beyond doubt : the egg-stalk is a natural silk in which /3 poly- peptide chains lie transverse to the fibre axis ; in folds too, for stretching the stalks pulls them out into the parallel-/? form of the more familiar silks. And this is not all : the basal pedestal, which consists of a film of protein spread by the laying insect on the leaf surface before it proceeds to draw off the 15-20 ,u thread which constitutes the egg-stalk, shows the ,8 polypeptide chains standing perpendicular to the film, with the a and c axes in the plane of the film.It is as though the struc- tural components of the fibre are drawn off in the form of long “ jumping crackers ”.82 INTRODUCTION The co-ordination with the “ classical ” observations on the cross-fl state and above all with Keller’s discoveries is clear, and to my mind of a significance for protein studies that could be far-reaching. Keller has found that polyethylene (and other chain-polymers such as nylon) can build comparatively large orthodox crystals just as if they were short-chain hydrocarbons or fatty acids. They do this by the trick, somehow energetically more favourable in their case-and one can see in a general way why for more flexible chains it should provide a more probable path to good packing-of fording into shorter lengths; they take up what I have just called the “ jumping-cracker ” form, a chain of regular transverse folds that in polyethylene, for instance, are of the order of lOOA long.Chrysopa silk behaves similarly, and the analogue of the polyethylene crystal is the basal pedestal. How long the transverse folds are in the polypeptide is being investig- ated, but fig. 2a shows some promising high-spacing reflections on the equator that could very well turn out to be counterparts of those found by Keller with polyethylene. (It is the high-angle pattern in fig. 2a that is the transverse equiv- alent of that given by other wild-type fibroins in the parallel-p form. The cor- respondence becomes direct and obvious, of course, when the Chrysopa silk passes over into the parallel-IS form on stretching-a change, it should be emphasized, that is in no sense a mere rotation of micelles but a genuine intramolecular trans- formation; * see fig.2b.) Looking back over the years during which we at Leeds have many times been preoccupied with the often difficult business of orienting macromolecules for X-ray diffraction purposes, it is noteworthy how the technique we developed, of first making thin films in which the long molecules lay down flat and then trying to stretch narrow ribbons cut from such films and in any case photographing them with the X-ray beam parallel to the surface, may have owed much of its success to the comparative inflexibility of most of the structures with which we happen to have operated.I am thinking of cellulose and its derivatives and other polysaccharides, gelatin, a-proteins (including bacterial flagella), and nucleic acids; and of how, as is realized now, it was only with polypeptides set free from the stiffening effect of specific intramolecular linkages that we came across the “ anomaly ” of long transverse folds. The question of what decides whether or not a long chain-molecule shall take up a transversely folded configuration, and the magnitude of the “ driving force ”, is intriguing enough on any count, but is pre-eminently important, to my idea, in relation to muscular contraction and its proposed interpretation in terms of the supercontraction of myosin. It is hardly to be doubted now that the latter is the expression of a marked tendency of the a-configuration, in hot water or by other means, to uncoil and fall instead into transverse folds; but what is the likelihood of such a transformation taking place under physiological conditions, and could it exert a suffciently useful pull? The lot of the physicist is still not entirely a happy one when he tries to persuade the physiologist that his (the physicist’s) unnatural experiments do really have something to do with what happens in vivo, and this has been a trouble certainly with our studies of the elastic properties of myosin 8 (actually an actomyosin, as Szent-Gyorgyi and Straub discovered later) isolated from its physiological environment and “ d e natured,” too, from the orthodox biochemical standpoint ; but quite recently we have been able to shed the stigma to some small degree by obtaining a cross-/3 diagram (fig.36) also from actomyosin gel contracted by the agency of adenosine triphosphate (ATP). This has been accomplished by Pautard, and it is peculiarly timely and gratifying in the present context to have succeeded at last with such a near-physiological demonstration of transverse folding in actomyosin under the * Though it was apparently not appreciated at the time, Palmer and Galvin,’ to judge by their published X-ray diagrams of fibres made from denatured crystalline egg-albumin, seem to have brought about a similar intramolecular transformation from an imperfect cross-/l state to well oriented parallel+ by means of a second stretching in live steam.(4 (6) FIG.1 .-X-ray fibre diagrams and skeleton representations of (a) the " parallel-/3 " form and (b) the " cross+? " form of epidermin (Rudall). [To face page 82(a) (b) FIG. 2.-X-ray fibre diagrams of the egg-stalk of the lace-wing fly Clzrysopa : (a) in the natural cross-/3 form ; and (6) in the parallel-/3 form produced from (a) by stretching. (Fibre axis in both diagrams parallel to short edge of page) (Rudall).(a) (b) FIG. 3.-(a) X-ray diagram, taken with the beam parallel to the surface, of a thin film prepared from the flagella of B. subtilis (Beighton); (b) ditto of a thin film prepared from actomyosin gel contracted by the action of adenosine triphosphate (Pautard).FIG. 4.-Electron micrograph showing the eventual disintegration of Polytoma flagella into chains of uniform particles (Millard).W.T. ASTBURY 83 action of what muscle itself uses to cause contraction. The effect as disclosed by X-rays is so far only feeble, but even so, I believe that it already has to be reckoned with as a possible first intimation of the biological pointer we are looking for. My second ‘‘ co-ordination ” links up automatically with my first, because I should like to say something next about the phenomena of biological contractility in relation to what may be called the corpuscle-fibre paradox; by which I mean the curious situation that whereas man-made fibres, on the strength of lessons learned primarily from studies of natural fibres, are essentially ‘‘ molecular yarns ” spun from more or less &awn-out chain-molecules, it has come to pass that at least the initial stages in the construction of natural fibres are found now frequently to involve a stringing-together of unit “ beads ” of coiled-up and folded chains.A classical early, but still outstanding, indication along this line of inference was given by feather keratin,g whose magnificent X-ray photograph had every appear- ance of a corpuscular origin yet was also a fibre diagram based on strikingly stretchable chains in a kind of crumpled p-configuration, but since then there have been, to mention only a few other examples: Waugh’s fibrous insulin; Bailey’s tropomysin, a member of the k-m-e-f group that can also grow into large single crystals ; F-actin, formed by the reversible polymerization of G-actin, and the source of the higher axial periodicities in the diffraction diagram of skeletal muscle ; 10 and latest and most dramatic of all, Andrew Szent-Gyorgyi’s “ proto- myosins ”, small units weighing only about 5000 into which he has succeeded in splitting myosin simply by the action of urea, under certain specific conditions.11 Pcrhaps the most impressive illustration of the paradox that we have observed at Leeds is the rapid breakdown-merely in the process of preparing a specimen for the electron microscope-of flagella detached from the alga Polytomn.They disintegrate fist into the familiar eleven sub-fibrils, next into filaments, then finally into chains of particles of diameter about 175 A (fig. 4).5 Szent-Gyorgyi (Albert) supposes 12 that since even the myosin “ molecule ” turns out after all to be no other than strings of particles held together by secondary forces, contraction must mean the collapse, somehow, of such strings into shorter, fatter collections, We cannot forget, though, that myosin is also effectively an elastic molecular yarn constructed from polypeptide chains normally in the a- configuration but which can be stretched into the parallel-P configuration and supercontracted into the cross$ configuration ; and of course to harmonize these two descriptions dynamically-they remind one of the corpuscle-wave dilemma that used to divide the theory of light-is the problem. I believe the difficulty can be smoothed out now, not too speculatively, with the aid of the newer findings on the globular proteins that I mentioned at the beginning of this Introduction.The concept of regular linear sequences of packets of suitably coiled-up and folded polypeptides to explain in a more static sense both the apparent biogenesis and the X-ray diagrams of natural protein fibres is not new-various people have thought of that ; but the tendency has been at the same time to picture the globular proteins as mostly all-or-none structures, so to speak, ready to unfold irreversibly at the slightest provocation; and from such a combined viewpoint it was not so easy to go on to explain long-range biological contractility-at any rate, not so easy as in terms of the folding and unfolding of the chain-molecules of straight- forward molecular yarns which the X-ray diagrams originally suggested, perfectly correctly in the case of many fibre structures.Now, however, we know from, for instance, the optical studies of Doty and his collaborators that the presence of cx-configurational components can be demonstrated not only in the fibrous proteins where they were first discovered by X-rays, but in the make-up of orthodox globular proteins too; and, what is very much more, the proportions of these cr-components can be reversibly altered to some extent by suitably altering the environment. The co-ordinating, and crucial, step in the argument that I want to make here is that these intra-globular configurational changes must, in principle,84 INTRODUCTION be accompanied by shape changes, and changes in the direction and mode of contact with neighbouring corpuscles, so that, in a linear polymer, spiralization and overall length changes, which may be very considerable, can follow as a matter of course.To clinch this inference by direct experiment we have, most apt among the observations to be discussed presently, Bresler’s arresting findings with human serum albumin-how the axial ratio of the molecule, considered as an ellipsoid, increases progressively from 4 to 16 when the water + dioxane solvent at pH 10 is made more and more hydrophobic by increasing the dioxane concentration; and this takes place, indeed, without any change in the optical rotation, and therefore even without affecting the a-helical components ; that is, presumably, by modifying only the tertiary structure. Clearly, chains of serum albumin, if there were such things, would be susceptible of violent contortions.The mitotic cycle of chromosomes, as viewed under the ordinary optical microscope, provides the classical example of the spiralization of chains of FIG. 5.-Infra-red absorption spectra of films of untreated and heat-denatured egg-white protein (Parker). (Inset : X-ray diagram of the heat-denatured egg-white.) - untreated (thin film); - - - - after boiling (thin film); - * - - difference curve. biological particles at the visual level, but in a closely similar connection I should like to make special reference to a just recently published paper by Ambrose,lJ in which, with the help of the interference microscope, he reveals as never before how minute fibrils are built by the linear aggregation of intracellular particles, and how readily these fibrils fall then into helical forms.My third and last “co-ordination” is a pendant to the paper by Elliott, Hanby and Malcolm in which they once and for all abandon the exciting general- ization suggested by the Courtaulds Laboratory several years ago to the effect that the a-helical configuration could be diagnosed always by a carbonyl stretching frequency round about 1665-1 660 cm-1 as opposed to 25-30 wave-numbers fewer for the corresponding band given by the p-configuration. This criterion would have been invaluable, but from the beginning doubts arose till eventually, even before the coup de g r k to be administered at this meeting, came the unanswer- able exception of polyglycine 11, for which they found 1648 cm-1 while Crick and Rich 14 interpreted its X-ray diagram in terms of a helix with three residues per turn and residue-length about 3.1 A, stabilized by intermolecular hydrogen bonds, as compared with the x-helix, which has 3.6 residues per turn and a residue-length of 1.5A and is stabilized by intramolecular hydrogen bonds.Also Parker here at Leeds, working in the same field, made the observation, among others hard toW . T . ASTBURY 85 reconcile with the Courtaulds thesis, that boiled fish myosin, for instance, gave an excelient /I diffraction pattern with no detectable trace of a, yet still seemed mostly or from the proposed infra-red test ; and similarly with egg-white. It is the latter experiment that is illustrated in fig.5 ; albeit, for sentimental reasons dating back to the cross-,f3 story, again with a film of “poached” egg-white! Altogether, it must be accepted now that, whether or not the /3-configuration is recognizable by a characteristic frequency of the carbonyl absorption band, the or-coniiguration is not, other states of coiling-perhaps almost any substantial deviation from the /I-configuration-resulting in a frequency in the neighbourhood of that given by the genuine ol.-helix. On second thoughts, it looks as if this, my concluding “coordination”, might be described more appropriately as a “ de-coordination ”. 1 Schellman, Compt. rend. Lab. Carlsberg, Sku. chim., 1955, 29, no. 15. 2 Astbury, Dickinson and Bailey, Biochem. J., 1935, 29, 2351. 3 Astbury and Bell, Nature, 1941, 147, 696. Astbury, Proc. Roy. SOC. B, 1947, 134, 4 Rudall, Synip. on Fibrous Proteins, J. SOC. Dyers and Colourists, 1946, p. 15 ; Ad- 5 Astbury afid Weibull, NGtnre, 1949, 163, 280. Astbury, Beighton and Weibull, 6 Parker and Rudall, Nature, 1957, 179, 905. 7 Palmer and Galvin, J . Amer. Chem. Soc., 1943, 65, 2187. 8 Astbury and Dickinson, Proc. Roy. Soc. B, 1940, 129, 307. 9 Astbury and Lomax, Nature, 1934, 133, 794. ‘0 Asttury, Nature, 1947, 160, 388. 11 Szent-Gjzorgyi, A. G. and Borbiro, Arch. Biochem. Biopltys., 1956, 60, 180. 12 Szent-Gyorgyi, A., Science, 1956, 124, 873 ; J . Cell. Comp. Physiol., 1957, 49, ’ 3 Ambrose, Proc. Roy. SOC. B, 1957, 148, 57. 14 Crick and Rich, Nature, 1955, 176, 780. Schellman and Harrington, Compt. rerld. Lab. Carlsberg, Sir. chim., 1956, 30, no. 3. 303 (Croonian Lecture, 1945). vatices in Protein Clieniistry, 1952, 7, 255. Symp. SOC. Expt. Biol., 1955, 9, 282. Astbury and Marwick, Nature, 1932, 130, 309. Bear and Rugo, Ann. N. Y. Acad. Sci., 1951, 53, 627. suppl. 1, 311.