ENZYME MOLECULE /966 the Three-Dimensional Structure of an Enzyme Molecule

SCIENTIFIC AMERICAN

ENZYME MOLECULE /966 The Three-dimensional Structure of an Enzyme Molecule

The arrangement of atoms in an enzyme molecule has been worked out for the first time. The enzyme is lysozyme, which breaks open cells of bacteria. The study has also shown how lysozyme performs its task

by David C. Phillips

ne day in 1922 Alexander Flem- determined and whose properties are corporated into a polypeptide chain a ing was suffering from a cold. understood in atomic detail. Among residue, and each residue has its own O This is not unusual in London, these properties is the way in which the characteristic side chain. The 129-resi- but Fleming was a most unusual man enzyme combines with the substance on due lysozyme molecule is cross-linked and he took advantage of the cold in a which it acts—a complex sugar in the in four places by disulfide bridges characteristic way. He allowed a few wall of the bacterial cell. formed by the combination of sulfur- drops of his nasal mucus to fall on a Like all enzymes, lysozyme is a pro- containing side chains in different parts culture of bacteria he was working with tein. Its chemical makeup has been of the molecule [see illustration on op- and then put the plate to one side to established by Pierre Jolles and his posite page}. see what would happen. Imagine his colleagues at the University of Paris The properties of the molecule cannot excitement when he discovered some and by Robert E. Canfield of the Co- be understood from its chemical con- time later that the bacteria near the lumbia University College of Physicians stitution alone; they depend most criti- mucus had dissolved away. For a while and Surgeons. They have found that cally on what parts of the molecule are he thought his ambition of finding a each molecule of lysozyme obtained brought close together in the folded universal antibiotic had been realized. from egg white consists of a single three-dimensional structure. Some form In a burst of activity he quickly estab- polypeptide chain of 129 amino acid of microscope is needed to examine the lished that the antibacterial action of subunits of 20 different kinds. A pep- structure of the molecule. Fortunate- the mucus was due to the presence in tide bond is formed when two amino ly one is effectively provided by the it of an enzyme; he called this substance acids are joined following the removal of techniques of X-ray crystal-structure lysozyme because of its capacity to lyse, a molecule of water. It is customary to analysis pioneered by Sir Lawrence or dissolve, the bacterial cells. Lyso- call the portion of the amino acid in- Bragg and his father Sir William Bragg. zyme was soon discovered in many tis- sues and secretions of the human body, in plants and most plentifully of all in ALA ALANINE GLY GLYCINE PRO PROLINE the white of egg. Unfortunately Flem- ARG ARGININE HIS HISTIDINE SER SERINE ing found that it is not effective against ASN ASPARAGINE ILEU ISOLEUCINE THR THREONINE the most harmful bacteria. He had to ASP ASPARTIC ACID LEU LEUCINE TRY TRYPTOPHAN wait seven years before a strangely CYS CYSTEINE LYS LYSINE TYR TYROSINE similar experiment revealed the exis- GLU GLUTAMIC ACID MET METHIONINE VAL VALINE tence of a genuinely effective antibi- GLN GLUTAMINE PHE PHENYLA1 ANINE otic: penicillin. TWO-DIMENSIONAL MODEL of the lysozyme molecule is shown on the opposite page. Nevertheless, Fleming's lysozyme has Lysozyme is a protein containing 129 amino acid subunits, commonly called residues (see proved a more valuable discovery than key to abbreviations above). These residues form a polypeptide chain that is cross-linked at he can have expected when its prop- four places by disulfide (-S-S-) bonds. The amino acid sequence of lysozyme was deter- erties were first established. With it, mined independently by Pierre Jolles and his co-workers at the University of Paris and by for example, bacterial anatomists have Robert E. Canfield of the Columbia University College of Physicians and Surgeons. The been able to study many details of bac- three-dimensional structure of the lysozyme molecule has now been established with the help of X-ray crystallography by the author and his colleagues at the Royal Institution in terial structure [see "Fleming's Lyso- London. A painting of the molecule's three-dimensional structure appears on pages 80 and zyme," by Robert F. Acker and S. E. 81. The function of lysozyme is to split a particular long-chain molecule, a complex sugar, Hartsell; SCIENTIFIC AMERICAN, June, found in the outer membrane of many living cells. Molecules that are acted on by enzymes I960]. It has now turned out that are known as substrates. The substrate of lysozyme fits into a cleft, or pocket, formed by the lysozyme is the first enzyme whose three-dimensional structure of the lysozyme molecule. In the two-dimensional model on three-dimensional structure has been the opposite page the amino acid residues that line the pocket are shown in dark green.

78 30 >. JL MAIN CHAIN NITROGEN CARBON OXYGEN

SULFUR SIDE CHAIN CARBON HYDROGEN BOND THREE-DIMENSIONAL MODEL of the lysozyme molecule, painted by Irving Geis, is based on an actual model assembled at the Royal Institution by the author and his colleagues. The painting enables one to trace and distinguish between the chemical bonds that hold together the main polypeptide chain and the bonds in the 129 side chains, one for each amino acid residue. The molecule is folded so as to form a cleft that holds the substrate molecule while it is being broken in two. The painting on the next page shows how the substrate fits into the cleft. The red balls represent oxygen atoms that are important in splitting the substrate.

\ L The difficulties of examining mole- according to whether the waves are in 1015 identical molecules in a regular cules in atomic detail arise, of course, or out of phase—in or out of step— array; in effect the molecules in such a from the fact that molecules are very with one another. (This effect is seen crystal diffract the X radiation as though small. Within a molecule each atom most easily in light waves scattered they were a single giant molecule. The is usually separated from its neighbor by a regularly repeating structure, such crystal acts as a three-dimensional dif- by about 1.5 angstrom units (1.5 X 10'8 as a diffraction grating made of lines fraction grating, so that the waves scat- centimeter). The lysozyme molecule, scribed at regular intervals on a glass tered by them are confined to a number which contains some 1,950 atoms, is plate.) In the second stage of image of discrete directions. In order to obtain about 40 angstroms in its largest di- formation, according to Abbe, the ob- a three-dimensional image of the struc- mension. The first problem is to find a jective lens of the microscope collects ture the intensity of the X rays scattered microscope in which the atoms can be the diffracted waves and recombines in these different directions must be resolved from one another, or seen sep- them to form an image of the object. measured, the phase problem must be arately. Most important, the nature of the im- solved somehow and the measurements The resolving power of a microscope age depends critically on how much of must be combined by a computer. depends fundamentally on the wave- the diffraction pattern is used in its The recent successes of this method length of the radiation it employs. In formation. in the study of protein structures have general no two objects can be seen sep- depended a great deal on the develop- arately if they are closer together than X-Ray Structure Analysis ment of electronic computers capable about half this wavelength. The short- of performing the calculations. They est wavelength transmitted by optical In essence X-ray structure analysis are due most of all, however, to the microscopes (those working in the ul- makes use of a microscope in which discovery in 1953, by M. F. Perutz of traviolet end of the spectrum) is about the two stages of image formation have the Medical Research Council Labora- 2,000 times longer than the distance been separated. Since the X rays can- tory of Molecular Biology in Cambridge, between atoms. In order to "see" atoms not be focused to form an image di- that the method of "isomorphous re- one must use radiation with a much rectly, the diffraction pattern is re- placement" can be used to solve the shorter wavelength: X rays, which have corded and the image is obtained from phase problem in the study of protein a wavelength closely comparable to it by calculation. Historically the meth- crystals. The method depends on the interatomic distances. The employment od was not developed on the basis of preparation and study of a series of pro- of X rays, however, creates other dif- this reasoning, but this way of regard- tein crystals into which additional heavy ficulties: no satisfactory way has yet ing it (which was first suggested by atoms, such as atoms of uranium, have been found to make lenses or mirrors Lawrence Bragg) brings out its essen- been introduced without otherwise af- that will focus them into an image. The tial features and also introduces the fecting the crystal structure. The first problem, then, is the apparently im- main difficulty of applying it. In re- successes of this method were in the possible one of designing an X-ray cording the intensities of the diffracted study of sperm-whale myoglobin by microscope without lenses or mirrors. waves, instead of focusing them to form John C. Kendrew of the Medical Re- Consideration of the diffraction the- an image, some crucial information is search Council Laboratory and in Pe- ory of microscope optics, as developed lost, namely the phase relations among rutz' own study of horse hemoglobin. by Ernst Abbe in the latter part of the the various diffracted waves. Without For their work the two men received the 19th century, shows that the problem this information the image cannot be Nobel prize for chemistry in 1962 [see can be solved. Abbe taught us that the formed, and some means of recovering "The Three-dimensional Structure of a formation of an image in the micro- it has to be found. This is the well- Protein Molecule," by John C. Kendrew, scope can be regarded as a two-stage known phase problem of X-ray crys- SCIENTIFIC AMERICAN, December, 1961, process. First, the object under exami- tallography. It is on the solution of the and "The Hemoglobin Molecule," by nation scatters the light or other radia- problem that the utility of the method M. F. Perutz, SCIENTIFIC AMERICAN, tion falling on it in all directions, form- depends. November, 1964]. ing a diffraction pattern. This pattern The term "X-ray crystallography" re- Because the X rays are scattered by arises because the light waves scattered minds us that in practice the method the electrons within the molecules, the from different parts of the object com- was developed (and is still applied) in image calculated from the diffraction bine so as to produce a wave of large the study of single crystals. Crystals pattern reveals the distribution of elec- or small amplitude in any direction suitable for study may contain some trons within the crystal. The electron density is usually calculated at a regular array of points, and the image is made visible by drawing contour lines MODEL OF SUBSTRATE shows how it fits into the cleft in the lysozyme molecule. All the through points of equal electron den- carbon atoms in the substrate are shown in purple. The portion of the substrate in intimate sity. If these contour maps are drawn contact with the underlying enzyme is a polysaccharide chain consisting of six ringlike struc- on clear plastic sheets, one can obtain tures, each a residue of an amino-sugar molecule. The substrate in the model is made up of a three-dimensional image by assem- six identical residues of the amino sugar called N-acetylglucosamine (NAG). In the actual bling the maps one above the other in a substrate every other residue is an amino sugar known as N-acetylmuramic acid (NAM). stack. The amount of detail that can be The illustration is based on X-ray studies of the way the enzyme is bound to a trisaccharide seen in such an image depends on the made of three NAG units, which fills the top of the cleft; the arrangement of NAG units in resolving power of the effective micro- the bottom of the cleft was worked out with the aid of three-dimensional models. The substrate is held to the enzyme by a complex network of hydrogen bonds. In this style of model- scope, that is, on its "aperture," or the making each straight section of chain represents a bond between atoms. The atoms them- extent of the diffraction pattern that has selves lie at the intersections and elbows of the structure. Except for the four red balls rep- been included in the formation of the resenting oxygen atoms that are active in splitting the polysaccharide substrate, no attempt image. If the waves diffracted through is made to represent the electron shells of atoms because they would merge into a solid mass. sufficiently high angles are included 83 (corresponding to a large aperture), the don. This is the laboratory in which when Roberto J. Poljak, a visiting work- atoms appear as individual peaks in Humphry Davy and Michael Faraday er from Argentina, demonstrated that the image map. At lower resolution made their fundamental discoveries dur- suitable crystals containing heavy atoms groups of unresolved atoms appear with ing the 19th century, and in which the could be prepared. Since then C. C. F. characteristic shapes by which they can X-ray method of structure analysis was Blake, A. C. T. North, V. R. Sarma, be recognized. developed between the two world wars Ruth Fenn, D. F. Koenig, Louise N. i The three-dimensional structure of by the brilliant group of workers led Johnson and G. A. Mair have played lysozyme crystallized from the white of by William Bragg, including J. D. Ber- important roles in the work. hen's egg has been determined in atom- nal, Kathleen Lonsdale, VV. T. Astbury, In 1962 a low-resolution image of ic detail with the X-ray method by our J. M. Robertson and many others. Our the structure was obtained that revealed group at the Royal Institution in Lon- work on lysozyme was begun in 1960 the general shape of the molecule and

LYSOZYME, MAIN CHAIN 102

LYSOZYME, SIDE CHAIN

SUBSTRATE, MAIN CHAIN

SUBSTRATE, SIDE CHAIN

HYDROGEN BOND

DISULFIDE BOND

MAP OF LYSOZYME AND SUBSTRATE depicts in color the nections for hydrogen bonds holding the substrate to the lysozyme. central chain of each molecule. Side chains have been omitted ex- The top three rings of the substrate \A, B,C) are held to the un- cept for those that produce the four disulfide bonds clipping the derlying enzyme by six principal hydrogen bonds, which are iden- lysozyme molecule together and those that supply the terminal con- tified by number to key with the description in the text. The lyso-

84 showed that the arrangement of the on the basis of nearly 10,000 diffrac- bridge). In the lysozyme helixes the hy- polypeptide chain is even more com- tion maxima, which resolved features drogen bond is formed somewhere be- plex than it is in myoglobin. This low- separated by two angstroms. Apart from tween two CO groups, giving rise to a resolution image was calculated from showing a few well-separated chloride structure intermediate between that of the amplitudes of about 400 diffraction ions, which are present because the an alpha helix and that of a more sym- maxima measured from native protein lysozyme is crystallized from a solution metrical helix with a three-fold symme- crystals and from crystals containing containing sodium chloride, the two- try axis that was discussed by Lawrence each of three different heavy atoms. angstrom image still does not show in- Bragg, Kendrew and Perutz in 1950. In 1965, after the development of more dividual atoms as separate maxima in There is a further short length of helix efficient methods of measurement and the electron-density map. The level of (residues 80 through 85) in which the computation, an image was calculated resolution is high enough, however, for hydrogen-bonding arrangement is quite many of the groups of atoms to be clear- close to that in the three-fold helix, and ly recognizable. also an isolated turn (residues 119 through 122) of three-fold helix. Fur- The Lysozyme Molecule thermore, the peptide at the far end of helix 5 through 15 is in the conforma- The main polypeptide chain appears tion of the three-fold helix, and the hy- as a continuous ribbon of electron den- drogen bond from its NH group is made sity running through the image with to the CO three residues back rather regularly spaced promontories on it that than four. are characteristic of the carbonyl groups Partly because of these irregularities (CO) that mark each peptide bond. in the structure of lysozyme, the pro- In some regions the chain is folded in portion of its polypeptide chain in the ways that are familiar from theoretical alpha-helix conformation is difficult to studies of polypeptide configurations calculate in a meaningful way for com- and from the structure analyses of myo- parison with the estimates obtained by globin and fibrous proteins such as the other methods, but it is clearly less keratin of hair. The amino acid residues than half the proportion observed in in lysozyme have now been designated myoglobin, in which helical regions by number; the residues numbered 5 make up about 75 percent of the chain. through 15, 24 through 34 and 88 The lysozyme molecule does include, through 96 form three lengths of "alpha however, an example of another regu- helix," the conformation that was pro- lar conformation predicted by Pauling posed by Linus Pauling and Robert B. and Corey. This is the "antiparallel Corey in 1951 and that was found by pleated sheet," which is believed to be Kendrew and his colleagues to be the the basic structure of the fibrous pro- most common arrangement of the chain tein silk and in which, as the name sug- in myoglobin. The helixes in lysozyme, gests, two lengths of polypeptide chain however, appear to be somewhat dis- run parallel to each other in opposite torted from the "classical" form, in directions. This structure again is stabi- which four atoms (carbon, oxygen, ni- lized by hydrogen bonds between the .118 trogen and hydrogen) of each peptide NH and CO groups of the main chain. group lie in a plane that is parallel to Residues 41 through 45 and 50 through the axis of the alpha helix. In the lyso- 54 in the lysozyme molecule form such zyme molecule the peptide groups in a structure, with the connecting resi- the helical sections tend to be rotated dues 46 through 49 folded into a hair- slightly in such a way that their CO pin bend between the two lengths of groups point outward from the helix comparatively extended chain. The re- axes and their imino groups (NH) in- mainder of the polypeptide chain is ward. folded in irregular ways that have no The amount of rotation varies, being simple short description. slight in the helix formed by residues 5 Even though the level of resolution through 15 and considerable in the one achieved in our present image was not formed by residues 24 through 34. The enough to resolve individual atoms, effect of the rotation is that each NH many of the side chains characteristic group does not point directly at the CO of the amino acid residues were readily group four residues back along the chain identifiable from their general shape. but points instead between the CO The four disulfide bridges, for ex- groups of the residues three and four ample, are marked by short rods of high back. When the NH group points di- electron density corresponding to the rectly at the CO group four residues two relatively dense sulfur atoms within them. The six tryptophan residues zyme molecule fulfills its function when it back, as it does in the classical alpha cleaves the substrate between the D and the helix, it forms with the CO group a hy- also were easily recognized by the ex- E ring. Note the distortion of the D ring, drogen bond (the weak chemical bond tended electron density produced by which pushes four of its atoms into a plane. in which a hydrogen atom acts as a the large double-ring structures in their

85 liquid. Such "polar" side chains are hydrophilic—attracted to water; they are found in aspartic acid and glutamic acid residues and in Iysine, arginine and histidine residues, which have basic side groups. On the other hand, most of the markedly nonpolar and hydrophobic side chains (for example those found in leucine and isoleucine residues) are shielded from the surrounding liquid by more polar parts of the molecule. In fact, as was predicted by Sir Eric Rideal (who was at one time direc- tor of the Royal Institution) and Irving Langmuir, lysozyme, like myoglobin, is quite well described as an oil drop with a polar coat. Here it is important to note that the environment of each molecule in the crystalline state is not signifi- cantly different from its natural environment in the living cell. The crystals themselves include a large proportion FIRST 56 RESIDUES in lysozyme molecule contain a higher proportion of symmetrically (some 35 percent by weight) of mostly organized regions than does all the rest of the molecule. Residues 5 through 15 and 24 watery liquid of crystallization. The through 34 (right) form two regions in which hydrogen bonds (gray) hold the residues in effect of the surrounding liquid on the a helical configuration close to that of the "classical" alpha helix. Residues 41 through 45 protein conformation thus is likely to be and 50 through 54 (left) fold back against each other to form a "pleated sheet," also held much the same in the crystals as it is in together by hydrogen bonds. In addition the hydrogen bond between residues 1 and 40 ties the first 40 residues into a compact structure that may have been folded in this way solution. before the molecule was fully synthesized (see illustration at the bottom of these two pages). It appears, then, that the observed conformation is preferred because in it the hydrophobic side chains are kept side chains. Many of the other residues the results of some further experiments out of contact with the surrounding also were easily identifiable, but it was we can begin to suggest answers to two liquid whereas the polar side chains nevertheless most important for the important questions: How does a mole- are generally exposed to it. In this way rapid and reliable interpretation of the cule such as this one attain its observed the system consisting of the protein image that the results of the chemical conformation? How does it function as and the solvent attains a minimum free analysis were already available. With an enzyme, or biological catalyst? energy, partly because of the large num- their help more than 95 percent of the Inspection of the lysozyme molecule ber of favorable interactions of like atoms in the molecule were readily immediately suggests two generaliza- groups within the protein molecule and identified and located within about .25 tions about its conformation that agree between it and the surrounding liquid, angstrom. well with those arrived at earlier in the and partly because of the relatively Further efforts at improving the ac- study of myoglobin. It is obvious that high disorder of the water molecules curacy with which the atoms have been certain residues with acidic and basic that are in contact only with other located is in progress, but an almost side chains that ionize, or dissociate, polar groups of atoms. complete description of the lysozyme on contact with water are all on the Guided by these generalizations, molecule now exists [see illustration on surface of the molecule more or less many workers are now interested in the pages 80 and 81]. By studying it and readily accessible to the surrounding possibility of predicting the conforma-

GROWING POLYPEPTIDE CHAIN

RIBOSOME

MESSENGER RNA

i i i i J L CODON NUMBER 1 10

FOLDING OF PROTEIN MOLECULE may take place as the grow- acid sequence of each protein is coded in "messenger" ribonucleic ing polypeptide chain is being synthesized by the intracellular par- acid (RNA). It is believed several ribosomes travel simultaneously ticles called ribosomes. The genetic message specifying the amino along this long-chain molecule, reading the message as they go. tion of a protein molecule from its it seems a reasonable assumption that, ing the obvious speculation is that chemical formula alone [see "Molecular as the synthesis proceeds, the amino there is no incentive to fold these hy- Model-building by Computer," by Cy- end of the chain becomes separated by drophilic residues in contact with the rus Levinthal; SCIENTIFIC AMERICAN, an increasing distance from the point of first part of the chain until the hy- June]. The task of exploring all possible attachment to the ribosome, and that the drophobic residues 55 (isoleucine) and conformations in the search for the one folding of the protein chain to its na- 56 (leucine) have to be shielded from of lowest free energy seems likely, how- tive conformation begins at this end contact with the surrounding liquid. ever, to remain beyond the power of any even before the synthesis is complete. It seems reasonable to suppose that imaginable computer. On a conservative According to our present ideas, parts at this stage residues 41 through 54 estimate it would be necessary to con- of the polypeptide chain, particularly fold back on themselves, forming the sider some 1011>0 different conformations those near the terminal amino end, may pleated-sheet structure and burying the for the lysozyme molecule in any gen- fold into stable conformations that can hydrophobic side chains in the initial eral search for the one with minimum still be recognized in the finished mole- hydrophobic pocket. free energy. Since this number is far cule and that act as "internal templates," Similar considerations appear to gov- greater than the number of particles in or centers, around which the rest of ern the folding of the rest of the mole- the observable universe, it is clear that the chain is folded [see illustration at cule. In brief, residues 57 through 86 simplifying assumptions will have to be bottom of these two pages]. It may are folded in contact with the pleated- made if calculations of this kind are to therefore be useful to look for the stable sheet structure so that at this stage succeed. conformations of parts of the polypep- of the process—if indeed it follows this tide chain and to avoid studying all the course—the folded chain forms a struc- The Folding of Lysozyme possible conformations of the whole ture with two wings lying at an angle molecule. to each other. Residues 86 through 96 For some time Peter Dunnill and I Inspection of the lysozyme molecule form a length of alpha helix, one side have been trying to develop a model of provides qualitative support for these of which is predominantly hydrophobic, protein-folding that promises to make ideas [see top illustration on opposite because of an appropriate alternation practicable calculations of the minimum page]. The first 40 residues from the of polar and nonpolar residues in that energy conformation and that is, at the terminal amino end form a compact part of the sequence. This helix lies in same time, qualitatively consistent with structure (residues 1 and 40 are linked the gap between the two wings formed the observed structure of myoglobin by a hydrogen bond) with a hydropho- by the earlier residues, with its hydro- and lysozyme. This model makes use bic interior and a relatively hydrophilic phobic side buried within the molecule. of our present knowledge of the way surface that seems likely to have been The gap between the two wings is not in which proteins are synthesized in folded in this way, or in a simply re- completely filled by the helix, however; the living cell. For example, it is well lated way, before the molecule was it is transformed into a deep cleft run- known, from experiments by Howard M. fully synthesized. It may also be im- ning up one side of the molecule. As we Dintzis and by Christian B. Anfinsen portant to observe that this part of shall see, this cleft forms the active site and Robert Canfield, that protein mole- the molecule includes more alpha helix of the enzyme. The remaining residues cules are synthesized from the terminal than the remainder does. are folded around the globular unit amino end of their polypeptide chain. These first 40 residues include a formed by the terminal amino end of The nature of the synthetic mecha- mixture of hydrophobic and hydro- the polypeptide chain. nism, which involves the intracellular philic side chains, but the next 14 resi- This model of protein-folding can be . particles called ribosomes working in dues in the sequence are all hydrophilic; tested in a number of ways, for example collaboration with two forms of ribonu- it is interesting, and possibly significant, by studying the conformation of the cleic acid ("messenger" RNA and "trans- that these are the residues in the anti- first 40 residues in isolation both di- fer" RNA), is increasingly well under- parallel pleated sheet, which lies out stood in principle, although the detailed of contact with the globular submole- environment of the growing protein cule formed by the earlier residues. In chain remains unknown. Nevertheless, the light of our model of protein fold-

Presumably the messenger RNA for lysozyme contains 129 "co- illustration shows how the lysozyme chain would lengthen as a ri- dons," one for each amino acid. Amino acids are delivered to the bosome travels along the messenger RNA molecule. Here, hypothet- site of synthesis by molecules of "transfer" RNA (dark color). The ically, the polypeptide is shown folding directly into its final shape. 87 rectly (after removal of the rest of the polysaccharide chains, being connected ride substrate, in the hope that we molecule) and by computation. Ulti- by bridges that include an oxygen atom would then be able to recognize the ac- mately, of course, the model will be re- (glycosidic linkages) between carbon tive groups of atoms in the enzyme and garded as satisfactory only if it helps us atoms 1 and 4 of consecutive sugar understand how they function. to predict how other protein molecules rings; this is the same linkage that joins Our studies began with the observa- are folded from a knowledge of their glucose residues in cellulose. The poly- tion by Martin Wenzel and his col- chemical structure alone. peptide chains that cross-connect these leagues at the Free University of Berlin polysaceharides are attached to the that the enzyme is prevented from func- The Activity of Lysozyme NAM residues through the lactyl side tioning by the presence of NAG itself. chain attached to carbon atom 3 in each This small molecule acts as a competi- In order to understand how lysozyme NAM ring. tive inhibitor of the enzyme's activity brings about the dissolution of bacte- Lysozyme has been shown to break and, since it is a part of the large sub- ria we must consider the structure of the linkages in which carbon 1 in NAM strate molecule normally acted on by the bacterial cell wall in some detail. is linked to carbon 4 in NAG but not the the enzyme, it seems likely to do this by Through the pioneer and independent other linkages. It has also been shown binding to the enzyme in the way that studies of Karl Meyer and E. B. Chain, to break down ehitin, another common part of the substrate does. It prevents followed up by M. R. J. Salton of the natural polysaccharide that is found in the enzyme from working by preventing University of Manchester and many oth- lobster shell and that contains only the substrate from binding to the en- ers, the structures of bacterial cell walls NAG. zyme. Other simple amino-sugar mole- and the effect of lysozyme on them are Ever since the work of Svante Ar- cules, including the trisaccharide made now quite well known. The important rhenius of Sweden in the late 19th cen- of three NAG units, behave in the same part of the cell wall, as far as lysozyme tury enzymes have been thought to way. We therefore decided to study the is concerned, is made up of glucose-like work by forming intermediate com- binding of these sugar molecules to the amino-sugar molecules linked together pounds with their substrates: the sub- lysozyme molecules in our crystals in into long polysaccharide chains, which stances whose chemical reactions they the hope of learning something about are themselves cross-connected by short catalyze. A proper theory of the en- the structure of the enzyme-substrate lengths of polypeptide chain. This part zyme-substrate complex, which under- complex itself. of each cell wall probably forms one lies all present thinking about enzyme My colleague Louise Johnson soon enormous molecule—a "bag-shaped mac- activity, was clearly propounded by found that crystals containing the sugar romolecule," as W. Weidel and H. Pel- Leonor Michaelis and Maude Men ton molecules bound to lysozyme can be zer have called it. in a remarkable paper published in prepared very simply by adding the The amino-sugar molecules concerned 1913. The idea, in its simplest form, sugar to the solution from which the in these polysaccharide structures are of is that an enzyme molecule provides a lysozyme crystals have been grown and two kinds; each contains an acetamido site on its surface to which its substrate in which they are kept suspended. The (-NH • CO • CH;i) side group, but one molecule can bind in a quite precise small molecules diffuse into the protein of them contains an additional major way. Reactive groups of atoms in the crystals along the channels filled with group, a lactyl side chain [see illustra- enzyme then promote the required water that run through the crystals. For- tion below]. One of these ammo sugars chemical reaction in the substrate. Our tunately the resulting change in the is known as N-acetylglucosamine (NAG) immediate objective, therefore, was to crystal structure can be studied quite and the other as N-acetylmuramic acid find the structure of a reactive complex simply. A useful image of the electron- (NAM). They occur alternately in the between lysozyme and its polysaccha- density changes can be calculated from

H O H H O \ I I I 4 C C 1 R, — C —OH —N—C—C—H R, — C — C-OH I •o H I I I I H H H —C —H I C H 6 POLYSACCHARIDE MOLECULE found in the walls of certain A, C and £ are NAG residues; li, D and F are NAM residues. The bacterial cells is the substrate broken by the lysozyme molecule. inset at left shows the numbering scheme for identifying the The polysaccharide consists of alternating residues of two kinds of principal atoms in each sugar ring. Six rings of the polysaccharide amino sugar: N-acetylglucosamine I NAG I and N-acetylmuramic fit into the cleft of the lysozyme molecule, which effects a cleav- acid (NAM I. In the length of polysaccharide chain shown here age between rings D and E (see illustration on pages 84 and 85).

88 measurements of the changes in ampli- 62; these deserve special mention be- tude of the diffracted waves, on the as- cause they are affected by a small sunrption that their phase relations have change in the conformation of the en- not changed from those determined for zyme molecule that occurs when the the pure protein crystals. The image trisaccharide is bound to it. The elec- shows the difference in electron density tron-density map showing the change in between crystals that contain the added electron density when tri-NAG is bound sugar molecules and those that do not. in the protein crystal reveals clearly In this way the binding to lysozyme that parts of the enzyme molecule have of eight different amino sugars was moved with respect to one another. 5 OXYGEN studied at low resolution (that is, These changes in conformation are through the measurement of changes in largely restricted to the part of the the amplitude of 400 diffracted waves). enzyme structure to the left of the cleft, The results showed that the sugars bind which appears to tilt more or less as a to lysozyme at a number of different whole in such a way as to close the cleft "CHAIR" CONFIGURATION (gray) is that places in the cleft of the enzyme. The slightly. As a result the side chain of normally assumed by the rings of amino sugar in the polysaccharide substrate. When investigation was hurried on to higher residue 62 moves about .75 angstrom bound against the lysozyme, however, the D resolution in an attempt to discover the toward the position of sugar residue B. ring is distorted {color) so that carbon exact nature of the binding. Happily Such changes in enzyme conformation atoms 1, 2 and 5 and oxygen atom 5 lie in these studies at two-angstrom resolution have been discussed for some time, no- a plane. The distortion evidently assists in (which required the measurement of tably by Daniel E. Koshland, Jr., of the breaking the substrate below the D ring. 10,000 diffracted waves) have now University of California at Berkeley, shown in detail how the trisaccharide whose "induced fit" theory of the en- made of three NAG units is bound to zyme-substrate interaction is supported with atoms in the enzyme molecule, un- the enzyme. in some degree by this observation in less this sugar residue is distorted a The trisaccharide fills the top half of lysozyme. little out of its most stable "chair" con- the cleft and is bound to the enzyme by formation into a conformation in which a number of interactions, which can be The Enzyme-Substrate Complex carbon atoms 1, 2 and 5 and oxygen followed with the help of the illustra- atom 5 all lie in a plane [see illustration tion on pages 84 and 85. In this illus- At this stage in the investigation ex- above]. Otherwise satisfactory interac- tration six important hydrogen bonds, citement grew high. Could we tell how tions immediately suggest themselves, to be described presently, are identified the enzyme works? I believe we can. and the model falls into place. by number. The most critical of these Unfortunately, however, we cannot see At this point it seemed reasonable interactions appear to involve the aceta- this dynamic process in our X-ray im- to assume that the model shows the mido group of sugar residue C [third ages. We have to work out what must structure of the functioning complex be- from top], whose carbon atom 1 is not happen from our static pictures. First tween the enzyme and a hexasaccharide. linked to another sugar residue. There of all it is clear that the complex formed The next problem was to decide which are hydrogen bonds from the CO group by tri-NAG and the enzyme is not the of the five glycosidic linkages would be of this side chain to the main-chain NH enzyme-substrate complex involved in broken under the influence of the en- group of amino acid residue 59 in the catalysis because it is stable. At low con- zyme. Fortunately evidence was at hand enzyme molecule [bond No. 1] and centrations tri-NAG is known to behave to suggest the answer. As we have seen, from its NH group to the main-chain as an inhibitor rather than as a substrate the cell-wall polysaccharide includes al- CO group of residue 107 (alanine) in that is broken down; clearly we have ternate sugar residues of two kinds, the enzyme molecule [bond No. 2]. Its been looking at the way in which it NAG and NAM, and the bond broken terminal CH3 group makes contact with binds as an inhibitor. It is noticeable, is between NAM and NAG. It was the side chain of residue 108 (trypto- however, that tri-NAG fills only half of therefore important to decide which of phan). Hydrogen bonds [No. 3 and, No. the cleft. The possibility emerges that the six sugar residues in our model could 4] are also formed between two oxygen more sugar residues, filling the remain- be NAM, which is the same as NAG atoms adjacent to carbon atoms 6 and 3 der of the cleft, are required for the except for the lactyl side chain append- of sugar residue C and the side chains formation of a reactive enzyme-substrate ed to carbon atom 3. The answer was of residues 62 and 63 (both tryptophan) complex. The assumption here is that clear-cut. Sugar residue C cannot be respectively. Another hydrogen bond the observed binding of tri-NAG as an NAM because there is no room for this [No. 5] is formed between the acetami- inhibitor involves interactions with the additional group of atoms. Therefore the do side chain of sugar residue A and enzyme molecule that also play a part bond broken must be between sugar residue 101 (aspartic acid) in the en- in the formation of the functioning en- residues B and C or D and E. We al- zyme molecule. From residue 101 there zyme-substrate complex. ready knew that the glycosidic linkage is a hydrogen bond [No. 6] to the oxy- Accordingly we have built a model between residues B and C is stable gen adjacent to carbon atom 6 of sugar that shows that another three sugar when tri-NAG is bound. The conclu- residue B. These polar interactions are residues can be added to the tri-NAG sion was inescapable: the linkage that supplemented by a large number of in such a way that there are satisfactory must be broken is the one between nonpolar interactions that are more dif- interactions of the atoms in the proposed sugar residues D and E. licult to summarize briefly. Among the substrate and the enzyme. There is only Now it was possible to search for the more important nonpolar interactions, one difficulty: carbon atom 6 and its origin of the catalytic activity in the however, are those between sugar resi- adjacent oxygen atom in sugar residue neighborhood of this linkage. Our task due B and the ring system of residue D make uncomfortably close contacts was made easier by the fact that John A.

89 Rupley of the University of Arizona hud One of the oxygen atoms of residue 52 acid residues, together with a contribu- shown that the chemical bond broken is about three angstroms from carbon tion from the distortion to sugar residue under the influence of lysozyme is the atom 1 of sugar residue D as well as D that has already been mentioned, is one between carbon atom 1 and oxygen from the ring oxygen atom 5 of that enough to explain the catalytic activity in the glycosidic link rather than the residue. Residue 35, on the other hand, of lysozyme. The events leading to the link between oxygen and carbon atom is about three angstroms from the oxy- rupture of a bacterial cell wall probably 4. The most reactive-looking group of gen in the glycosidic linkage. Further- take the following course [see illustra- atoms in the vicinity of this bond are more, these two amino acid residues tion on this page]. the side chains of residue 52 (aspartic have markedly different environments. First, a lysozyme molecule attaches acid) and residue 35 (glutamic acid). Residue 52 has a number of polar itself to the bacterial cell wall by in- neighbors and appears to be involved in teracting with six exposed ammo-sugar a network of hydrogen bonds linking it residues. In the process sugar residue D with residues 46 and 59 (both aspara- is somewhat distorted from its usual gine) and, through them, with residue conformation. 50 (serine). In this environment residue Second, residue 35 transfers its ter- 52 seems likely to give up a terminal minal hydrogen atom in the form of a hydrogen atom and thus be negatively hydrogen ion to the glycosidic oxygen, charged under most conditions, even thus bringing about cleavage of the when it is in a markedly acid solution, bond between that oxygen and carbon whereas residue 35, situated in a non- atom 1 of sugar residue D. This creates polar environment, is likely to retain its a positively charged carbonium ion (C +) terminal hydrogen atom. where the oxygen has been severed A little reflection suggests that the from carbon atom 1. concerted influence of these two amino Third, this carbonium ion is stabilized by its interaction with the negatively charged aspartic acid side chain of residue 52 until it can combine with a hy- droxyl ion (OH~) that happens to diffuse into position from the surrounding water, thereby completing the reaction. The lysozyme molecule then falls away, ^P CARBON leaving behind a punctured bacterial l|§| OXYGEN cell wall. |jfe HYDROGEN It is not clear from this description that the distortion of sugar residue D plays any part in the reaction, but in fact it probably does so for a very interesting reason. R. H. Lemieux and G. Huber of the National Research Council of Canada showed in 1955 that when a sugar molecule such as NAG in- 0H- _-'' WATER MOLECULE corporates a carbonium ion at the carbon-1 position, it tends to take up the same conformation that is forced on ring D by its interaction with the enzyme molecule. This seems to be an example, therefore, of activation of the substrate by distortion, which has long been a favorite idea of enzymologists. The binding of the substrate to the en- LYSOZYME, MAIN CHAIN zyme itself favors the formation of the carbonium ion in ring D that seems to play an important jsart in the reaction. It will be clear from this account that although lysozyme has not been seen LYSOZYME, in action, we have succeeded in build- MAIN CHAIN ing up a detailed picture of how it may work. There is already a great deal SPLITTING OF SUBSTRATE BY LYSOZYME is believed to involve the proximity and of chemical evidence in agreement with activity of two side chains, residue 35 (glutaniicacidi and residue 52 (aspartic acid ). It is pro- this picture, and as the result of all the + posed that a hydrogen ion (H ) becomes detached from the OH group of residue 35 and work now in progress we can be sure attaches itself to the oxygen atom that joins rings D and E, thus breaking the bond between that the activity of Fleming's lysozyme the two rings. This leaves carbon atom 1 of the D ring with a positive charge, in which form will soon be fully understood. Best of it is known as a carbonium ion. It is stabilized in this condition by the negatively charged side chain of residue 52. The surrounding water supplies an OH~ ion to combine with the all, it is clear that methods now exist carbonium ion and an H+ ion to replace the one lost by residue 35. The two parts of for uncovering the secrets of enzyme the substrate then fall away, leaving the enzyme free to cleave another polysaccharide chain. action.