I am happy to speak about molecular structure in relation to medicine. This is a small part of the great subject chemistry in relation to medicine, a subject that I shall not attempt to cover as a whole. I shall talk only about molecular diseases and about a new theory of general anesthesia.
In one sense almost all diseases are molecular, in that the human body is made up of molecules and the vectors of disease, viruses, rikettsiae, bacteria, are made up of molecules. There are some diseases that clearly are not molecular diseases. If as a result of an accident a man loses a large part of his brain tissue and has mental disease. This injury is macroscopic in nature; the molecules that remain to him are no different from those that he had before. Or even if his brain is injured by anoxia so that damage is done, although we might say that the injury is chemical in nature, I would not say that the disease that results is a molecular disease.
Nearly twenty years ago when I was serving as a member of the Medical Research Committee investigating support of medical research by the U. S. Government, one of the members of the Committee, Dr. William B. Castles, talked about the disease sickle-cell anemia. This is a hereditary disease in which the red cells in the blood are twisted out of shape. As a consequence of the deformation of the red cells in the blood, the patient has a serious anemia. His red cells are destroyed so rapidly by the spleen that he is not able to manufacture now ones fast enough to keep him in good health. Also these deformed red cells clog up the capillaries in crises of the disease in such a way that the blood ceases to flow through some organ and the organ is damaged by anoxia. The statement that was made that was most interesting to me and that immediately brought a response from me, was that the red cells are twisted out of shape in the venous circulation, but resume their normal form in the arterial circulation.
It seemed clear to me that there was a high probability that it was the hemoglobin molecule that was involved in this difference in behavior of the red cells of these patients in the venous blood and in the arterial blood. The Hemoglobin molecule contains about 10,000 atoms. Four of them are iron atoms. In the lungs oxygen molecules can attach themselves to each of the four iron atoms, the blood charged with oxygen in this way circulates out to the tissues, and the oxygen molecules are given up. The difference between venous blood and arterial blood is that venous blood contains molecules of hemoglobin without oxygen attached to the iron, while the arterial blood contains molecules of oxyhemoglobin.
The suggestion that it is the hemoglobin that is involved brought the further idea that the patients with this disease manufacture an abnormal sort of hemoglobin that is self complementary so that the molecules clamp onto one another to form long rods which line up side by side as a crystal or perhaps tactoid liquid crystal of hemoglobin which, as it grows longer and longer, twists the red cell out of shape and leads to the manifestations of the disease.
When the iron atoms are oxygenated, the self-complementariness may be destroyed in such a way that the crystals go back into solution and the cells resume their normal shapes.
When I returned home to Pasadena a few months later, a young man, Harvey Itano, who had received his M. D. degree came to work with me, and I asked him to examine the hemoglobin of patients with this disease and to compare the hemoglobin of these patients with the hemoglobin of normal individuals. For three years he made comparisons of the hemoglobins of these two sources, but always with the same result, in every experiment that he carried out.
With a substance so complicated that its molecule contains 10 000 atoms, the identity of certain properties was no assurance that the molecules are completely identical in structure. Finally he and two other young men, Drs. J.S. Singer and I.C. Wells carried out an electrophoresis experiments in which the two samples of hemoglobin behaved differently. This one result was enough to show that they are different. With such complicated substances any number of experiments in which they behave the same would not be proof of their identity, but a single possible experiment in which they behave differently is proof of non-identity.
Many other abnormal human hemoglobins have been discovered since 1949, and many abnormal or different hemoglobins are manufactured by other animal species, under the control, of course, of genes.
The pitch, the distance of the helical axis per residue is 1.49 A. The dimensions of the α-helix have been very well verified by experiments for the synthetic polypeptides with the α-helix structure, for α-keratin, fibrous proteins, hair, horn, fingernails, porcupine quill and so on, and also through the work of Kendrew for the polypeptide chain of the globular protein myoglobin.
Myoglobin contains 8 α-helix segments one of which is shown here. The iron atom is up to the upper right-hand corner surrounded by the atoms of the porphyrin group that constitute with the iron atom the heme group where the oxygen is attached.
Here we have a patient with the disease sickle-cell anemia. This disease seemed to be a disease of the red cells and it turned out to be a disease of the hemoglobin molecule.
In the oxygenated blood of these patients the red cells have a normal appearance as seen through the microscope, in the deoxygenated blood, in the veins, they are sickled. When this investigation was first carried out showing that the patients produced abnormal hemoglobins, the parents of the patients were studied. It was found that the father contained in his red cells a mixture: 50 % of the hemoglobin was normal and 50 % abnormal. Similarly for the mother.
Here we have a paper chromatographic study. The diagram on the far right is the best. It corresponds to 4 hours of electrophoresis. At the bottom is the hemoglobin of a normal individual. It migrates rapidly and has a negative electric charge. Directly above it is the hemoglobin of the father or mother of a sickle-cell patient. Here there are two hemoglobins, normal hemoglobin and sickle-cell hemoglobin. The latter migrates at a rate showing that it is different in its electric charge by two electronic units. It has two electric charges, less than the normal hemoglobin. Dr. X brought into the laboratory a sample of blood from another person and found that this person had sickle-cell hemoglobin in his red cells and another hemoglobin still more abnormal than the sickle-cell hemoglobin, with 4 units of charge difference from normal hemoglobin. One parent of this interesting individual was a sickle-cell heterozygote, the other was a heterozygote in this new abnormal hemoglobin, manufacturing both, normal hemoglobin and this new abnormal hemoglobin. In the theory of Mendelian heredity one would expect that parents of this nature, i.e. two heterozygote’s, have children of several kinds: 1/4 of the children will inherit the abnormal gene of the father and also the abnormal gene of the mother and will then have a double abnormality, each present in single dose. This person has a disease of a new kind, a disease involving the inheritance of two differently abnormal genes, which separately do not produce any serious disease. But they cooperate with one another to produce a new type of anemia, the disease is called sickle-cell hemoglobin-C disease. (The new abnormal hemoglobin was named hemoglobin C). Many other hemoglobins have since been discovered, D, E, G, H and so on.
A few abnormal hemoglobins are indicated here. Along the diagonal of the matrix we have some of the homozygote’s: the sickle-cell patients with two sickle-cell genes, hemoglobin-C patients with two hemoglobin-C genes and so on. At the top are the carriers of the genes in single dose. They in general do not have serious diseases, and we have on the diagonal some of the complex diseases involving the inheritance of different abnormal genes and the manufacture of two abnormal hemoglobins.
Dr. X in our laboratory and his collaborators using the method of Sanger were able to show that there are two kinds of polypeptide chains in the normal hemoglobin molecule. One chain contains 141 amino acid residues, it is called the α-chain, it begins with a residue of valine and continues lysine, serine, proline, alanine, asparagine and so on. The other chain, the β-chain, begins valine, histidine, lysine,. threonine, proline, glutamic acid and continues on.
V. Ingram developed a technique of two-dimensional paper electrophoresis and chromatography and by splitting of hemoglobin into several simple peptides was able to show that the abnormality of the sickle-cell hemoglobin is in the β-chain. He and Dr. X tied it there, onto the 6th position where valine replaces glutamate. The glutamate residue carries a negative electric charge; its carboxyl group is ionized. Valine, on the other hand, has a hydrocarbon side-chain with no electric charge. Consequently, one negative electric charge is lost in this substitution.
There are two α-chains and two β-chains in the sickle-cell hemoglobin, just as in normal hemoglobin, and the β-chains are changed by the substitution in the 6th position in the chain, giving them a difference in electric charge of two units and a difference in molecular structure such as to produce the complementariness and insolubility characteristic of the sickle-cell hemoglobin. We do not yet, despite the work of Perutz in Cambridge, know the structure of the hemoglobin molecule well enough to be able to explain in terms of structure the formation of the tektoids by deoxygenated hemoglobin S, but we can expect, that this will occur soon.
Here there is a symbol given for hemoglobin F, the letter F stands for fetus. Hemoglobin F is the hemoglobin that is manufactured by the fetus. It contains two normal α-chains resembling the adult and two γ-chains which are rather cliff ere from the β-chain. At about the time of birth of the infant, the infant begins to manufacture β-chains whereas in prenatal life he was manufacturing γ-chains.
This is the technique that Ingram used of hydrolyzing hemoglobin with trypsin to produce about 26 peptides, each with 10-14 amino acid residues, and separating these on paper by electrophoresis in one direction, and by chromatography in the other. Only one of the 26 peptides is different in sickle-cell hemoglobin from normal hemoglobin. When this peptide was investigated it was found to be the first peptide in the 13-chain and to have glutamate replaced by valine.
Here we have results indicated for some other abnormalities in human hemoglobin molecules involving the β-chain. Many abnormal human hemoglobins are known in which the α-chain is abnormal. They are not shown here. There are 146 amino acid residues in the human hemoglobin β-chain. In the case of hemoglobin S in the 6th position glutamate is replaced by valine. With hemoglobin C glutamate is replaced by lysine. Lysine has an amino group attached to the δ-carbon atom of the side-chain, and it becomes an ammonium ion group at physiological pH so that the lysine side-chain carries a positive electric charge. With two β-chains in the molecule this means a difference of four units in electric charge between hemoglobin C and normal hemoglobin.
Hemoglobin G has a substitution in the 7th position, hemoglobin E a substitution in the 26th position, hemoglobin A2 a substitution in the 22nd, and it is accident that all these substitutions involve replacement of a glutamic acid residue. Other kinds of substitution are known.
In each case, as for example in hemoglobin S, there is only one amino acid residue changed, all of the other 140 are exactly the same as in the normal adult human hemoglobin. No variant of human hemoglobin has been discovered in which either the α- the β-chain differs from normal by more than one amino acid residue.
Here is shown the geographical distribution of the gene for sickling: Madagascar, United States, Sicily, Southern Italy, Greece, and Portugal. One can ask why the gene has spread so widely among the human population. There must be some advantage in carrying the gene in order for a mutation to begin to spread. The answer was suggested by a British physician, Dr. X who noticed that there were more sicklers in malarial regions in Africa than in non-malarial regions. Then a young physician, Dr. X carried out a crucial experiment in Kenya: he got 30 healthy adult male Africans who were shown by skin tests not to have developed any immunity to malaria. When they were inoculated with malaria 14 out of 15, who were normal people so far as their hemoglobin goes, became ill with malaria. But of those who had one sickle-cell gene, i.e. of the heterozygote’s in the sickle-cell gene, only two came down with malaria. There was a great degree of protection against malaria by a single gene. Their red cells contain a 50:50 mixture of sickle-cell hemoglobin and normal hemoglobin, and this provides them with protection against malaria. There is a molecular mechanism, of course: ordinarily the red cells of these individuals, the heterozygote’s, do not sickle in the venous circulation. But if the blood is completely deoxygenated, then the red cells are twisted out of shape. The crystal forms, even though the hemoglobin is diluted with an equal amount of normal hemoglobin. The malarial parasite lives inside the red cell and he has a high metabolic rate, he uses up the oxygen inside the red cell so that the partial pressure of oxygen becomes so low that the hemoglobin crystallizes, twists the red cell out of shape and squashes the parasite to death. So we have a molecular explanation not only of the lethal manifestations of the abnormal hemoglobin in the homozygote’s but also of the protection against malaria that is provided to the heterozygote’s.
This shows the incidence of hemoglobin C: high incidence in Northern Ghana, low in Southern Ghana and then still smaller along the coast. It seems likely that the hemoglobin-C mutation occurred only a short time ago, perhaps a thousand years ago, whereas the mutation producing the sickle-cell hemoglobin may have occurred 5000 or 10000 year ago and then has spread over Africa.
Now I want to discuss another type of disease, also related to hemoglobin, and I begin by showing again the structure of myoglobin. I point out the iron atom and a group that is rather close by the iron atom. This group is a histidine residue. It is in the 58th position of the α-chain or the 63rd position of the β-chain. This group is, I believe, responsible for retaining the iron atom in the ferrous state. Hemoglobin can also be called ferro-hemoglobin. The iron is bipositive, bivalent. Sometimes hemoglobin is oxidized to the tripositive state; the iron becomes ferric rather than ferrous. This ferri-hemoglobin, also called methemoglobin, does not have the power of combining reversibly with oxygen so that the power of transferring oxygen from the lungs to the tissues is lost. Ordinarily ferrous compounds are easily oxidized to the ferric state. This residue of histidine has an imidazolium ring in its side-chain. This ring at physiological pH adds a proton and assumes a positive charge. I believe that this positive charge in the neighborhood of the iron atom stabilizes the ferrous state by repelling the additional positive charge that would be added to the iron atom on conversion from Fe2+ to Fe3+. There is good evidence for this now through the investigation of the hemoglobins of certain people who have a disease in which two of the iron atoms in their hemoglobin molecules are easily converted to the tripositive state. Dr. X has been responsible for much of the investigations of the hemoglobin of these patients with the disease methemoglobinemia or ferri-hemoglobinemia.
Here we have the group of some 60 atoms that is called the heme group with the iron atom at its center bounded to four pyrrole rings, in the way that Hans Fischer showed sixty years ago, without having knowledge, of course, of the dimensions that we have now.
Here we have the heme group with the side-chain of histidine, the imidazolium cation carrying a positive charge that stabilizes the ferrous state of the iron, as one finds it in normal persons.
Here is a type of variant hemoglobin, hemoglobin Zurich, in which in the 62nd position of the β-chain in place of histidine there is a residue of tyrosine. Tyrosine with the p-hydroxybenzen ring does not pick up a proton, and does not carry a positive charge. The iron atom easily oxidizes to the ferric state, when the patient has the disease ferri-hemoglobinemia.
This shows another abnormal hemoglobin in which the same histidine residue is replaced by arginine. Arginine has a guanidine group in its side-chain that picks up a proton to produce the guanidinium cation, the positive charge stabilizes the iron atom, it does not oxidize to the tripositive state, and these people, even though they carry an abnormal hemoglobin, do not have the disease ferri-hemoglobinemia.
Here we have a few of the abnormal hemoglobins related to this state. Hemoglobin Boston has tyrosine in the 58th position of the α-chain Emmery has tyrosine in the 62nd position of the β-chain. They both lead to ferri-hemoglobinemia. Zurich has arginine in the 62nd position of the B-chain but without producing ferri-hemoglobinemia.
An interesting, abnormal hemoglobin is Milwaukee that has glutamic acid in a position four removed from histidine. The α-helix has nearly four residues per turn and this residue of glutamate is near the iron atom, too. Normally, there is valine in this position which does not carry a charge. The negative charge of glutamate apparently attracts an extra positive charge to the iron atom which becomes the ferric iron atom and produces ferri-hemoglobinemia. Here we have a disease then, for which there is a simple detailed chemical explanation of the manifestations of the disease. It will be hard to go beyond this, deeper than we have gone now, with this group of diseases in understanding disease on a molecular basis.
I should like to mention some evolutionary considerations, which involve studies carried out in our laboratory by X. We have here the peptide patterns using the method of Ingram, for human hemoglobin, fish hemoglobin, shark, hogfish. It is clear that there are great differences in these hemoglobins. In fact, the differences look to be greater than they actually are, because there are great similarities, too, even between human and fish hemoglobins.
We see now a comparison of human, cow, and pig hemoglobin. It is evident that the patterns are somewhat similar. The hemoglobin structures are rather similar.
This shows human, chimpanzee, gorilla, orangutan, rhesus monkey hemoglobins. If we compare human and gorilla, or human and chimpanzee, it is nearly impossible to find a difference, and a more detailed investigation shows that the α-chain of the gorilla differs from the α-chain of the human by two amino acid residues out of 141. The other 139 are the same, and in the same positions. The β-chains of human and gorilla differ by one amino acid residue only.
In the case of chimpanzee, the β-chain differs from that of human by one amino acid residue, and it is identical with the type of human β-chain called Norfolk hemoglobin. Probably an independent mutation, however, in the case of the people in Norfolk who have this identity with the chimpanzee.
With rhesus monkey instead of one or two differences there are six or eight differences. With horse there are 18 differences.
Dr. X and I thought that we could make some statements about the process of evolution. We took horse and human hemoglobins and assumed that the line leading to humans and horses separated 130,000,000 years ago. And then we assumed that there is a constant rate of evolutionarily effective mutation. With this assumption, the gorilla chain corresponds to 14,000,000 years ago, the β-chain to 7,000,000, the average being about 11,000,000 years ago. I don’t have comparisons of human and rhesus monkey here, but it would come out about 40,000,000 years ago, and this at once answers the question that students of evolution have asked: at what stage did the monkeys of different kinds and the anthropoid apes and man separate from one another. The monkeys separated off from the common ancestors of gorillas and human beings about 40,000,000 years ago, the humans and gorillas separated roughly 10,000,000 years ago. 260,000,000 years ago human fetus separated from human adult. Of course, there weren't humans then. A human fetus is more like a horse fetus in its hemoglobin, than like an adult human being. In a sense, so far as the α- and β-chains go, the fetuses of mammals are more closely related to one another than they are to their corresponding adults.
The α-chain and the β-chain have about 78 differences and 60 identities, corresponding to about 600,000,000 years ago. The identities are so numerous that we can be sure that they originally represented a single gene, a single chain.
And we have attempted to determine the nature of the single polypeptide chain of the hemoglobin manufactured by the ancestor of all vertebrates, some 600,000,000 years ago. Here we have just four different chains indicated, and as we move along we see that in this position all four have lysine, in this position three of them have leucine and the fourth has phenylalanine. It looks as though the feto-gene has undergone a weight mutation: after the separation of the genes for β- and γ-chains, gene duplication followed by independent mutation. The mutation occurred in the γ-gene.
This shows our present knowledge about the amino acid sequence in the polypeptide chain for hemoglobin manufactured by the common ancestor of all vertebrates some 600,000,000 years ago. There are some uncertainties, great uncertainties about half of the positions are not filled. But I am sure, that in a few years more all positions will be filled. It might then be possible to synthesize the polypeptide chain, add the heme to it, determine the oxygen combining power of this, primeval form of hemoglobin and in that way reach a conclusion about the partial pressure of oxygen in the atmosphere of the earth 600,000,000 years ago.
Now I shall talk briefly about another application of molecular structure to a problem that we can call a medical problem, the problem of the nature of general anesthesia. There has not been any satisfactory theory of general anesthesia until the theory I shall describe was developed about 4 years ago. I published this in June of 1961 and a few months later a young chemist, Prof. Stanley Miller of the University of California at San Diego, published essentially the same theory, quite different words and different calculations, but I think it is the same theory.
I have here a drawing of the structure of ordinary ice. Each water molecule forms four hydrogen bonds with its four neighbors. It looks as though there are holes in this crystal, but these holes are really not very large, no molecules except He and hydrogen, I think, could fit into it.
This shows another view of ordinary ice looking down the hexagonal axis. The atoms are really much larger so that the holes are small.
Here we have an aggregate of 20 water molecules forming 30 hydrogen bonds with one another. The bond angles within the pentagonal dodecahedron are 108° so that no bond angle strain from the tetrahedral angle of 109.5° is involved. Investigations of hydrate crystals showed that these pentagonal dodecahedra of 20 water molecules occur in many of them.
These are the basic structures of chlorine hydrate, methane hydrate, and xenon hydrate. These structures were determined by Dr. Marsh in our laboratory. I was especially interested in xenon hydrate, in the fact that xenon forms crystals with this framework. Xenon atoms occupy positions of the centers of the polyhedra. There are six extra water molecules in addition to the 40 of the two polyhedra.
This allows that there are not only the dodecahedra, but also tetrakisdecahedra in the centers of which somewhat lighter molecules such as methylchloride or chlorine can fit. These crystals with a rather open hydrogen-bounded framework of water molecules are stabilized by the van der Waals interactions of the xenon or cyclopropane molecules that occupy the cavities with the water molecules.
This is a larger cavity formed by 28 water molecules. It has 4 hexagonal faces and 12 pentagonal faces and it is large enough to permit a molecule of chloroform to fit inside it. In the chloroform hydrate crystal, CHCl3 • 17 H20 there is one of these polyhedra for every two dodecahedra. If xenon is present, the decomposition point of the crystal is raised by 14 ºC because of the van der Waals interaction of the xenon molecules with the neighboring molecules. There is a cooperation then in this crystal 2Xe • CHCl3 • 17H2O, a cooperative effect involving the xenon in stabilizing the crystal.
This is a still larger opening in a hydrogen-bonded framework in which there is a tetraisoamylammonium ion and a fluoride ion also present. I was reading a manuscript describing a crystal of this sort (not yet published, sent to me by Prof. X back in 1959) when I thought to myself "I understand the mechanism of general anesthesia.” Here we have an ion, something like the side-chain of a lysine residue in the proteins. Perhaps in the brain these electrically charged side-chains and some ions interact with the water molecules to form small crystals of a hydrate. In the presence of an anesthetic agent which can fill other cavities, xenon for example which serves as a good general anesthetic (it is not used because it is so expensive), xenon atoms could occupy some of the smaller cavities and stabilize the crystal, which however could also involve some ions from the solution and some of the side-chain groups of proteins - this would trap the electrically charged side-chains and ions which normally oscillate back and forth contributing to the electrical oscillations in the brain that constitute consciousness and memory and. by decreasing the amplitude of electrical oscillations would lead, to unconsciousness. Then, as the anesthetic agent is allowed to leave the body through the lungs the crystals would become unstable, would decompose again and the consciousness would be regained. If this theory were right, then it would be possible to anesthetize people by cold, just by cooling the brain. And of course, I discovered that it is possible to produce anesthesia just by cold. At 30°C mild anesthesia is produced, at 26°C deep anesthesia.
Also, if this theory is right, there should be a close relation between the van der Waals forces between the anesthetic molecules and the anesthetic activity. Here I have plotted the logarithm of the equilibrium partial pressure of the hydrate crystals of various anesthetic agents against the molecular polarizability (expressed here in cm3/mole). High pressure, nearly 100,000 mm Hg, is required to produce argon hydrate crystals, methane, krypton, and so on, xenon, methyl chloride, ethyl bromide, chloroform, carbon tetrachloride. Over here is the curve, similarly plotted against the molecular polarizability, of the logarithm of the anesthetizing partial pressure for mice at 37ºC. The curve has the same smooth character.
Here I have plotted the logarithm of the anesthetizing partial pressure against the logarithm of the equilibrium partial pressure against the logarithm of the hydrate crystals. There is a reasonably good linear relationship which goes over a great range of pressures. This is then some indication that it is the molecular polarizability that is responsible for the anesthetic activity, not necessarily the formation of the hydrate microcrystals; there might be some other way in which this molecular property could be active, but I think that the hydrate microcrystals idea is a good one.
Here are some experiments, no yet published, carried out with gold fish in which here we have the reciprocal temperature, temperatures running from 0ºC - 35ºC. For several anesthetic agents the temperature coefficient of the anesthetizing partial pressure (the logarithm is plotted here against the reciprocal temperature) corresponds over a wide range to nearly constant entropy of the reaction with nearly the same value for the different anesthetic agents. Then there is a rapid drop, the gold fish are anesthetized at 1.6ºC even in the absence of the anesthetic agent. This sort of catastrophe, of course, indicates a cooperative phenomenon such as crystallization, a phenomenon in which a large number of molecules take part, a change in phase. I think that this is a good indication that something that we might call microcrystal formation is taking place.
This is another representation of the crystal of xenon hydrate in which there is the hydrogen-bonded framework of water molecules with the xenon atom occupying each of the dodecahedral and tetrakistetrahedral cavities.
I am pleased that it is possible to propose an explanation of the extraordinary property of xenon, a highly unreactive substance which surely does not enter into ordinary chemical reactions in the human body, of producing anesthesia when it is inhaled. I think that it is possible to understand the molecular structure of the human body, to understand physiological phenomena, even psychic phenomena, I believe that it will be possible to get a penetrating and deep understanding of the nature of mental disease in the course of time, such as to permit great progress to be made in the control and treatment of mental disease, which is one of the great scourges, of course, in the world today because of a tremendous amount of human suffering. We are just entering now in the period of development of molecular biology and medicine, the ideas are necessarily rather crude ones that have been proposed so far, but I think that we can have great hope for the future.