Heredity and Development: Second Edition (1972)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

HEREDITY AND DEVELOPMENT: SECOND EDITION 167 8 DNA—Structure and Function Few chemistry textbooks written before 1950 had much to say about nucleic acids. Little was known of their chemical nature and nothing of their func- tion. Apparently they were always associated with proteins, forming a class of compounds known as nucleoproteins. Their history began in 1868 when Friedrich Miescher obtained the first crude preparation by extracting used surgical bandages, which were permeated with pus cells. Later Miescher extracted nucleic acid itself from fish sperm. Sperm might be expected to be a rich source of nuclear substances because the nucleus occupies such a large part of the cell. In fact the ratio of nucleus to cell is higher for sperm than for any other cell. Later it was discovered that the thymus gland is also a rich source of nucleic acid. Many chemical studies were made on calf thymus glands, which were obtained from local slaughter houses. When nucleic acid of the thymus gland was hydrolyzed, it was found to consist of only a few components: adenine, guanine, cytosine, thymine, deoxyribose (a sugar), and phosphoric acid (Fig. 8–1). Yeast cells were also extensively studied. Their nucleic acid was found to be much like that of the thymus but it differed in having uracil instead of thymine and ribose instead of deoxyribose. Gradually the belief arose, now known to be incorrect, that animal cells have one type of nucleic acid (with thymine and deoxyribose) and plants have another (with uracil and ribose). The nucleic acids and nucleoproteins remained the orphans of the chemist for so long largely because they had no obvious importance either inside or outside the cell. Other proteins were clearly of enormous 

HEREDITY AND DEVELOPMENT: SECOND EDITION 168 importance: some were the enzymes that controlled the reactions of the living cells; others were the hemoglobins that carried oxygen; still others were hor- mones, with their dramatic effects on a variety of life processes. In the first third of the twentieth century nucleoproteins were not extensively studied because there was no urgent reason for doing so. The number of scientific problems that might be studied is always far greater than the number that can be studied—scientific manpower is always insufficient. The biochemists of this period concentrated largely on problems associ- ated with the release and utilization of energy within the cell. Here was a problem of clear and obvious importance. It was vigorously investi- 8–1 The hydrolysis products of DNA and RNA. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 169 gated and, at one level of analysis, essentially answered. By 1950 the numer- ous reactions, each with a specific enzyme, in the pathway from glucose to the end products, carbon dioxide and water, were thought to be known. Adenosine triphosphate (ATP) had been identified as a key substance in the storage and transfer of energy within the cell. These biochemical pathways were adorned with many Nobel Prizes. The experiments described in Chapter 7, however, suggested that DNA is vitally involved in inheritance. In 1944 it was established that DNA is the transforming substance in Diplococcus; evidence obtained in 1952 suggested strongly that the entire genetic information of the T2 phage is DNA. With leads of this sort, it is not surprising that many 

HEREDITY AND DEVELOPMENT: SECOND EDITION 170 biologists turned their attention to DNA. There working hypothesis was: DNA is the hereditary material. The possibilities for gaining new insights into genetic mechanisms were enormous with a hypothesis so specific. The hypothesis linked two fields, chemistry and genetics, creating the possibility of testing deductions of a chemical nature with genetic data. Conversely, genetic deductions could be tested with data on the chemical nature of DNA. It is worth a brief digression to emphasize the tremendous utility of hypotheses of this type. They have been characteristic of genetics since the early days, and one might even say that they were largely responsible for the rapid progress in the field. Recall that Sutton’s basic hypothesis was that genes are parts of chromosomes (Chapter 4). If this is so, then one should observe a parallel between the behavior of chromosomes in meiosis and fertil- ization and the behavior of the Mendelian factors in inheritance. Sutton found such a parallel and speeded genetics on its road to becoming a science. In later years geneticists and cytologists constantly checked the discoveries of one field against those of the other. Genetic data first suggested the hypothesis of crossing over. A basis for the event was then found in careful studies of the chromosomes during meiosis. Bridges advanced the hypothesis of non-disjunction on the basis of genetic data and tested his hypothesis by a study of the chromosomes of his experimental material. The hypothesis that pieces of chromosomes may become inverted was suggested by genetic data and confirmed by a study of the salivary gland chromosomes. This type of rigorous checking of the hypotheses of one field by the data of another has not been generally possible in biology. To return to the main argument: we can test our hypothesis about the DNA molecule with the well-established principles of genetics. Similarly we can anticipate that, as knowledge of the chemistry of DNA becomes available, new insights into genetic mechanisms will be obtained. In order to proceed, we shall accept as true the hypothesis that the gene is DNA. The following deduction follows logically: DNA must have a structure and a composition that will account for the known properties of genes. Let us recall some of these basic properties. Linkage data localized the gene as part of a chromosome. Experiments on crossing over showed that the genes are in a linear order and in a definable site on the chromosome. Genes were found to be exceedingly stable. Barring mutation, which is a rare phenomenon, the gene maintained its integrity gen- eration after generation. This stability continued even with frequent replica- tion. At each mitotic division every gene becomes two, and at anaphase one gene goes into each daughter 

HEREDITY AND DEVELOPMENT: SECOND EDITION 171 cell. A gene might replicate a hundred thousand times or more without mak- ing a mistake. Yet from time to time mistakes—or mutations—occur. Such mutations are essential for the welfare of the species, for they are the raw materials of evolutionary change. But genes do more than merely maintain themselves. They have specific effects that the geneticist observes as the phenotype of the cell or individual. These more important properties of genes can be summarized as follows: 1. Genes have the ability to make copies of themselves. 2. Genes carry hereditary information. 3. Genes are able to transfer this information to the rest of the cell. On the basis of our hypothesis, therefore, the DNA molecule must have a structure that can replicate, carry information, and translate this information into the phenotype of the cell. THE WATSON-CRICK MODEL Did DNA have the necessary properties to be the gene? No one knew in 1950, but an American biologist, James D.Watson (born 1928, now of Har- vard University), and his English associate, Francis Crick (born 1916, of Cambridge University), addressed themselves to the problem. They sought to devise a model of the DNA molecule that would satisfy the requirements imposed by the genetic data. In two papers, one published in April and the other in May of 1953, they showed how the few facts known about DNA could be used to construct a model of its structure. This model was, in terms of scientific methodology, a hypothesis. These were the facts: 1. DNA is composed of six kinds of molecules; adenine, guanine, thymine, cytosine, deoxyribose, and phosphoric acid. Each adenine, guanine, thymine, and cytosine combines with a molecule of deoxyribose and phosphoric acid. The four combinations are known as nucleotides. 2. Many of these nucleotides combine to form the huge DNA molecule. (Watson and Crick were attempting to determine the precise nature of their combination.) 3. The available data on the X-ray diffraction patterns suggested to Watson and Crick that the DNA molecule consists of two long fibers twisted around one another to form a double helix (like double spiral staircases, one for ascent and one for descent). 

HEREDITY AND DEVELOPMENT: SECOND EDITION 172 4. X-ray data indicated that the diameter of the double helix is about 20 Ångstrom units. 5. Each fiber of the double helix consists of phosphate and deoxyribose units alternating with one another: phosphate-deoxyribose-phosphate- deoxyribose, and so on. 6. The adenine, guanine, cytosine, and thymine units (collectively known as nitrogenous bases) are attached to the phosphate-deoxyribose chain. 7. In different cells of the same species, the relative amounts of adenine, guanine, thymine, and cytosine are the same. 8. In different species, the relative amounts of adenine, guanine, thymine, and cytosine vary greatly. 9. In all cells that had been studied, the amount of adenine was found to equal the amount of thymine and the amount of guanine was found to equal the amount of cytosine. This was discovered by Erwin Chargaff of Columbia University. Neither Watson nor Crick discovered even one of these facts about DNA, which had been slowly accumulating over the years and were available to all interested in DNA. It was Watson and Crick who first saw how the data could be unified into a model that would satisfy both the genetic and the chemical requirements for the molecular structure of DNA. Their triumph was of the mind, not of the laboratory. Half a century earlier, Sutton had made a similar contribution. Although he did study the chromosomes of grasshoppers, he merely confirmed what others had already established. He saw the relation of the data of Mendel and of the cytologists and combined them to arrive at the hypothesis: genes are parts of chromosomes. Sutton’s feat was an exercise of pure intellect, as was that of Watson and Crick. But what was their model? The critical aspect of the Watson-Crick model is the positioning of bases on the two entwined strands. Since the relative amounts of the different bases vary from species to species, there can be no single structure for all DNA. Yet a striking regularity exists: the amount of adenine always equals that of thymine and the amount of guanine always equals that of cytosine. This led Watson and Crick to predict that, wherever there is an adenine on one strand, there is a thymine opposite it on the other strand; and similarly, that guanine and cytosine are also opposite one another. Thus if we unwind the double helix, the arrangement of the bases will be as shown in Figure 8–3. The adenine and thymine, as well as the guanine and cytosine, were assumed to be held loosely to one another 

HEREDITY AND DEVELOPMENT: SECOND EDITION 173 8–2 Figures 1 and 2 (redrawn) from the original paper by Watson and Crick (Nature 171:965). by hydrogen bonds. These bonds would form if the bases were opposite one another in the positions suggested by the model. Further evidence for this specific pairing came from data on the relative sizes of the bases and of the diameter of the DNA molecule. Two of the bases, thymine and cytosine, are relatively small. Adenine and guanine are larger, as can be seen from the diagrams of the molecules in Figure 8–1. Thus the pairing that Watston and Crick assumed is always of one large and one small base. Such pairing fits well with the apparently uniform diameter of the DNA, which X-ray data revealed to be 20 Ångstroms. If the pairing were between a cytosine on one strand and a thymine on the other, the diameter would be less than 20 Ångstroms. Similarly if adenine and guanine were to pair, the diameter of the double helix would be more than 20 Ångstroms. The data, there- 

HEREDITY AND DEVELOPMENT: SECOND EDITION 174 8–3 A highly schematic reconstruction of the double helix formed by the DNA molecule. The lower part of the helix is enlarged to show the bases adenine (A), thymine (T), guanine (G), and cytosine (C) and how these bases are linked with deoxyribose (D), and phosphoric acid (P). Refer to Figure 8–1 for more details of the molecular structures. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 175 fore, were best explained on the assumption that adenine always pairs with thymine and guanine always pairs with cytosine. So far as size relations are concerned, adenine could pair with either of the smaller molecules—cytosine or thymine. Similarly, guanine could pair with either cytosine or thymine. One would then expect the amount of adenine+guanine to equal the amount of cytosine+thymine. Chargaff’s measurements showed this to be true but they showed more than that: he found that the amount of adenine is equal to the amount of thymine and the amount of guanine equals that of cytosine. Had each base also accounted for 25 per cent of the total, one could still main- tain that either large base could pair with either small base at random. But the amounts are not equal: the amount of adenine+thymine may differ consider- ably from the amount of guanine+cytosine. The data were best explained, therefore, by assuming a specific pairing of adenine with thymine and gua- nine with cytosine. The Watson-Crick model for DNA, then, consists of two long and closely associated strands wound around one another. The strands are complemen- tary to one another, in the sense that what is present on one strand automati- cally specifies what is on the other. Thus if the sequence of bases on one strand is adenine-adenine-cytosine-thymine-guanine-thymine, that of the other would have to be thymine-thymine-guanine-adenine-cytosine-adenine. Since the sugar-phosphate parts of the molecule are always the same and only the sequence of the bases can vary, Watson and Crick hypothesized ‘it therefore seems likely that the precise sequence of bases is the code which carries the genetical information.’ Accounting for Replication. The gene can make an exact copy of itself; if DNA is the gene it must have the same ability. This was possible with the Watson-Crick model and the argument was developed as follows: ‘Previous discussions of self-duplication have usually involved the concept of a tem- plate, or mould. Either the template was supposed to copy itself directly or it was to produce a “negative,” which in its turn was to act as a template and produce the original “positive” once again. In no case has it been explained in detail how it would do this in terms of atoms and molecules. Now our model for deoxyribonucleic acid is, in effect, a pair of templates, each of which is complementary to the other. We imagine that prior to duplication the hydro- gen bonds [between the bases opposite to one another in the two strands] are broken, and the two chains unwind and separate. Each chain then acts as a template for the formation on to itself of a new companion chain, so that even- tually we shall have two pairs of chains, where we only had one before. More- 

HEREDITY AND DEVELOPMENT: SECOND EDITION 176 8–4 The replication of DNA. During replication the two strands of DNA separate from one another. Each strand then serves as the template for the synthesis of a complementary strand. The available data indicate that replication occurs over short stretches at a time. In this diagram the upper portion of the DNA double helix has not started to replicate. In the central section the strands have separated 

HEREDITY AND DEVELOPMENT: SECOND EDITION 177 over, the sequence of the pairs of bases will have been duplicated exactly…. Despite [some] uncertainties we feel that our proposed structure for deoxyri- bonucleic acid may help to solve one of the fundamental biological problems —the molecular basis of the template needed for genetic replication. The hypothesis we are suggesting is that the template is the pattern of bases formed by one chain of the deoxyribonucleic acid and that the gene contains a complementary pair of such templates.’ Figure 8–4 shows in diagrammatic form how the replication is thought to occur. In this way the Watson-Crick model accounts for the important genetic fact that genes can make exact copies of themselves. The model was a hypothesis, which opposed no known chemical or genetic facts. During the next decade an ever-increasing amount of chemical and genetic data sug- gested that the hypothesis was indeed correct. In 1962 Watson and Crick shared the Nobel Prize with Maurice Wilkins, the physicist from Cambridge University who had supplied much of the X-ray data indicating that DNA is a double helix with a uniform diameter of 20 Ångstrom units. The other primary attribute of genes, their specific function in the cell, was not discussed by Watson and Crick in their first papers. This part of the prob- lem was studied subsequently and now a probable hypothesis is at hand. Genes and Protein Structure. The hypothesis that the chief function of genes is the control of protein synthesis became increasingly probable during the 1940s and early 1950s. A large body of work, much of it on the mold Neu- rospora, was best interpreted as indicating that genes produce or control the production of enzymes, which in turn control the numerous biochemical events that occur in all cells (Chapter 6). There was no clear understanding of how genes exert their control; do and nucleotides are combining to form the new strands. It is probable that the direction of synthe- sis, as indicated by the arrows, is anti-parallel. Replication produces two identical daughter strands. For example, note the site on the double helix just above the area of replication: there is a C-G unit. When these separate, the G will be united with a C in its new complementary strand and the C with a G in its new complementary strand. The two new double helices, therefore will be C- G and G-C. The lower section shows a section where replication has been completed. You will notice a few loose ends of DNA strands; these will be joined once replication has been completed in the central section. This diagram is based on DNA replication in bacteria; in higher organisms the process is similar but not identical. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 178 they produce enzymes and other proteins directly or indirectly? Considerable light was shed on this question by the intensive study of a disease of man— sickle cell anemia. Throughout much of central Africa, sickle cell anemia is common among the natives. The primary effect of the disease is on the hemoglobin of the red blood cells. When these cells are in capillaries where the oxygen concentra- tion is low, they may change from a round to an elongate or even to a sickle shape. These abnormally shaped cells may clog the capillaries and the small- est arteries. Many are destroyed, which causes the anemia. The number of red blood cells may be as few as two million per cubic millimeter, in contrast to the normal number of five million. Infant mortality is high and few individu- als with the disease live beyond 40 years. Genetic analysis has shown that the disease is caused by an autosomal gene, which is symbolized Hb1S (the normal allele is Hb1A). Homozygous individuals, Hb1S Hb1S, have the severe anemia already described. Heterozy- gous individuals, Hb1S Hb1A, are nearly normal, however their red blood cells do show abnormal shapes when subjected to very low oxygen concentrations. Hemoglobin is obviously an important protein and a great deal is known about its chemistry. Hemoglobin A is the common type in man. There are several other kinds, all differing only slightly from one another in the sequence of amino acids of which they are composed. Each molecule of hemoglobin A consists of about 600 amino acids, of 19 different kinds. These amino acids are arranged in four polypeptides—long chains of amino acids. Each molecule of hemoglobin is composed of two α polypeptides and two β polypeptides. The α and β chains differ from each other in length and in the sequence of their amino acids. The four polypeptides of the molecule are linked together and folded in a compact and specific manner to give the hemoglobin molecule a globular shape. Since the most obvious feature of sickle cell anemia is the abnormality of the red blood cells, it is reasonable to suppose that the hemoglobin of these cells might also be abnormal. Linus Pauling (born 1901) and his associates at the California Institute of Technology began an investigation to see if this was so. Their material was blood from three types of individuals: normal, Hb1A Hb1A; sufferers from sickle cell anemia, Hb1S Hb1S; and the heterozygotes, Hb1A Hb1S. The blood was fractionated and the hemoglobin obtained in a nearly pure form. In most features the three hemoglobins were identical. When they were compared in an electrophoresis apparatus, however, striking differences were observed. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 179 An electrophoresis apparatus consists basically of a long tube containing a liquid or semi-solid gel. An electric current is passed through the tube. Any charged substances placed in the gel will move, with the rate of movement depending on the size of the particle and its charge. Using this apparatus, it is frequently possible to separate different kinds of molecules in a mixture. Thus, if the three types of hemoglobin differed in their charges, they could be separated, and thereby shown to be different. This analytical device shows that normal hemoglobin, which we can call hemoglobin A, differs from sickle cell hemoglobin, which we can call hemoglobin S (Fig. 8–5). Furthermore, heterozygous individuals produce both kinds of hemoglobin. Pauling’s findings are striking evidence that genes can affect the structure of proteins. In the presence of the Hb1A gene, the protein hemoglobin A is synthesized in the red blood cells; in the presence of the Hb1S allele, hemoglobin S is synthesized. The next step is to compare the structures of hemoglobin A and hemoglobin S. This is an obvious step, perhaps, but one beset with tremen- dous difficulties. The problem was to determine the exact posi- 8–5 Electrophoretic patterns of hemoglobin from normal individuals, from heterozygotes, and from individuals homozygous for the sickle cell gene (modified from Pauling, Itano, Singer, and Wells, 1949). 

HEREDITY AND DEVELOPMENT: SECOND EDITION 180 tion of each amino acid in the total of 600. When Pauling carried out his experiments, the structure of not a single protein was known. It was not until 1954 that Frederick Sanger (born 1918), and his coworkers at Cambridge University, finally succeeded in determining the complete amino acid sequence for a protein. After ten years of intensive work they knew the posi- tion of each of the 51 amino acids in the insulin molecule. If it took ten years to determine the structure of insulin with its 51 amino acids, how long might it take to determine the structure for hemoglobin with its 600 amino acids? The task was begun by Vernon Ingram (born 1924), then also at Cambridge University and now at the Massachusetts Institute of Technology, and his associates. They were able to learn the answer by an analytical short cut—they did not have to determine the complete structure of hemoglobin. This, however, they have since accomplished. It was nearly impossible to study the huge hemoglobin molecule intact. A more practicable method of investigation was to break down the large molecule into smaller molecules and then to study the smaller molecules. If the structure of each of the smaller molecules could be determined, and if it could then be determined how these smaller molecules are combined, one would know the structure of hemoglobin. The older methods of the analytical chemist were not of much help. Typically, he hydrolyzes the proteins with acid and gradually breaks down the large molecules until only amino acids remain. The intermediate breakdown products are not uniform, however. A confusing mixture of large and small molecules is formed and this is not suit- able for the careful analysis that Ingram was attempting. He required a pre- cise method for breaking down the hemoglobin molecule into smaller molecules. The method he chose was to treat the protein with the enzyme trypsin. Trypsin is highly specific in action, hydrolyzing the protein only at the carboxyl side of arginine and lysine. Since these amino acids are always in the same places in the hemoglobin molecule, the hemoglobin will always be broken down in the same way. When the hemoglobin molecule is broken with trypsin, the result is 28 kinds of smaller molecules, averaging about ten amino acids each. In this manner, Ingram hydrolyzed both hemoglobin A from normal individuals and hemoglobin S from sufferers of sickle cell anemia. Next he separated the 28 smaller molecules by a process combining- electrophoresis and paper chromatography. He put a drop of the hydrolyzed hemoglobin mixture on a piece of paper and an electric cur- 

HEREDITY AND DEVELOPMENT: SECOND EDITION 181 rent was passed through the paper. This method is essentially the one used by Pauling to separate the entire molecules of hemoglobin A and hemoglobin S. In this case, Ingram was attempting to separate the 28 hydrolysis products on the basis of their electrical charges. He achieved considerable separation, but not enough for analytical purposes. Then he tried another method. The edge of the same piece of paper was put in a liquid that would dissolve the 28 hydrolysis products. As the liquid moved through the paper it carried the 28 kinds of molecules with it. The different kinds were carried at different rates. At the end of the experiment, the 28 types of molecules occupied different positions on the sheet of paper—each in a specific place (Fig. 8–6). This type of separation, using both electrophoresis and paper chromatography, gave consistent results. That is, in repeated experiments with hemoglobin A, the 28 spots always occupied the same positions relative to one another. When hemoglobin S was analyzed in a similar way, again there were 28 spots but when these spots were compared with those from hemoglobin A, there was an important difference. Twenty-seven of the spots occupied the same relative positions on two sheets of paper. The twenty-eighth pair, how- ever, occupied slightly different positions (Fig. 8–6). 8–6 The hydrolysis products of normal hemoglobin and sickle cell hemoglobin. The products have been separated by paper chromatography and electrophoresis (not all are shown here). In most cases the hydrolysis products of normal hemoglobin and sickle cell hemoglobin occupy equivalent positions. One spot in each, shown in black, occupies a slightly different position (modified from Ingram, 1958). 

HEREDITY AND DEVELOPMENT: SECOND EDITION 182 The implication is that hemoglobin A and hemoglobin S are identical in all but one of the units that result from hydrolysis with trypsin. It was, therefore, important to study the one instance of difference. If the structure of this vari- ant could be determined, one would know the difference between hemoglobin A and hemoglobin S without knowing the entire structure of either. The variant spots were found to consist of the amino acids at the end of the β chain. Ingram was able to determine the exact sequence which, using abbreviations for the amino acids, proved to be this: Hb A: -lys—glu—glu—pro—thr—leu—his—val Hb S: -lys—glu—val—pro—thr—leu—his—val The two sequences are identical except for the sixth position from the end: in hemoglobin A this is occupied by glutamic acid; in hemoglobin S by valine. This single difference in the β chains seems to be the ultimate cause of sickle cell anemia. When the hemoglobin molecule has valine instead of glu- tamic acid at this one site, it cannot function normally. The blood of het- erozygous individuals, having both the Hb1A and Hb1S alleles, has hemoglobin of both kinds. The conclusions that can be drawn from these observations are simple, specific, and staggering. In a normal individual, the β chain of hemoglobin A is made under the influence of one allele at the Hb1 locus. In individuals with sickle cell anemia, an abnormal hemoglobin, hemoglobin S, is made under the influence of a mutant allele. When formed by the mutant allele, the β chain of the hemoglobin differs from the normal in a single amino acid substi- tution: valine rather than glutamic acid is the sixth amino acid from the end of the long chain of 146 amino acids. Thus genes can be implicated in the very basic steps in protein synthesis: the insertion of single amino acids in a polypeptide chain. Since the cell produces its proteins by combining amino acids, genes can affect the most elementary steps in the process. The link between gene and protein was now secure. The line of investiga- tion that began uncertainly with Garrod and was renewed by Beadle and Tatum, was revealing the molecular biology of the gene. Having established that genes can control the synthesis of proteins, it then became essential to discover how they do it. This proved to be quite an under- taking. After all, the genes are in the nucleus and most of the cell’s proteins are in the cytoplasm. How, then, could one account for the origin of the cyto- plasmic proteins? One could hypothesize that they are produced in the nucleus and then pass into the cytoplasm. Alternatively, one could imagine that some influence, or information, passes from the nucleus to the cytoplasm and directs the synthesis of cyto- 

HEREDITY AND DEVELOPMENT: SECOND EDITION 183 plasmic proteins. The preliminary answers to these questions came from stud- ies of the distribution of nucleic acids in the cell and on purified fractions of cells that seemed to be able to carry out in vitro the same reactions that occur in vivo. Localizing Nucleic Acids. During the 1930s methods were perfected for detecting both DNA and RNA in situ. Some of the methods depended on the different staining reactions of DNA and RNA. The Feulgen method, already mentioned, was specific for DNA. Another cytological method involved the use of two dyes, methyl green and pyronin. Methyl green was found to stain DNA and pyronin to stain RNA. Spectrophotometric methods were also developed to detect the nucleic acids. The bases in these compounds (and in a few others in the cell, such as ATP) intensely absorb ultraviolet light having a wave length of 260 mµ. There are no other compounds in the cell having this specific absorption. Since the bases are situated largely in DNA or RNA, a peak absorption at 260 mµ indicates the occurrence of nucleic acids. This method will not distinguish between DNA and RNA; it measures total DNA plus RNA plus any other substances, such as ATP, that contain the bases. If one knows the total, however, and then measures the DNA by the Feulgen method, the difference will be largely the RNA. These early methods suggested that the nucleic acids are distributed differ- ently in the cell. The DNA appeared to be largely or entirely restricted to the nucleus. Most of the RNA was thought to be in the cytoplasm but nucleoli were rich in it and some observers thought they saw RNA closely associated with the chromosomes. Once it was known that DNA is of great importance in the cell, it was rea- sonable to suppose that the closely similar RNA was also of importance. The differences between the two nucleic acids are small: RNA has uracil instead of thymine as one of its four bases and its sugar is ribose instead of deoxyri- bose (Fig. 8–1). One of the first hypotheses of the role of RNA in the cell was based on the observation that cells that synthesize large amounts of protein are rich in RNA. In the liver and pancreas, where there is much protein synthesis, there is usually from two to eight times as much RNA as DNA. In the kidney, brain, spleen, and thymus, which synthesize much less protein, there are usu- ally equal amounts, or there may even be more DNA than RNA. Possibly this correlation has a causal significance. This was the view of Jean Brachet, of the Free University of Brussels. In the mid 1940s he observed repeatedly that cells synthesizing large amounts of protein stain heavily for RNA. The stain seemed to be taken up by tiny granules but these were too small to be studied with 

HEREDITY AND DEVELOPMENT: SECOND EDITION 184 the compound microscope. Later, when the electron microscope methods were perfected for studying cells, the existence of these granules was con- firmed. Since they are rich in RNA, they are called ribosomes. Some appear to be free in the cytoplasm; others are situated on the walls of a network of tubes known as the endoplasmic reticulum (Fig. 8–7). Brachet and others suggested that the ribosomes must have something to do with protein synthe- sis. Since the ribosomes are rich in RNA, possibly it is the RNA that is con- cerned with protein synthesis. If this hypothesis was to be tested, new methods had to be developed. These methods involved the isolation of different parts of the cell by such gentle procedures that they could still carry out some of their functions. Cell-Free Systems. In the century following the formulation of the cell theory, the cell came to be regarded not only as a unit of structure but 8–7 A portion of a cell from the pancreas of a bat. The large sausage-shaped structure is a mito- chondrion. The tubes below it are parts of the endoplasmic reticulum. The dark granules associ- ated with the endoplasmic reticulum, and also free in the cytoplasm, are ribosomes (photograph by Keith Porter). 

HEREDITY AND DEVELOPMENT: SECOND EDITION 185 also as a seemingly indivisible unit of function. Few of the complex events that occurred in cells could be duplicated apart from them. It was known that enzymes could act in vitro and that some small organic compounds could be synthesized from inorganic molecules. But a biochemist, who could easily synthesize fats, carbohydrates, and proteins in his body, was powerless to accomplish this feat in his laboratory. This inability to reproduce in vitro common events occurring in vivo did not suggest the need for some vitalistic principle. It was realized that the biochemical events that occur within cells frequently require dozens of different kinds of enzymes, specific sources of energy, and varied raw materials. The oxidation of glucose, for example, was found to require dozens of enzymes and many complex molecules such as the riboflavins, cytochromes, and so on. There was no reason to believe that even these intricate chains of reactions could not be eventually carried out apart from living cells. It was merely a matter of waiting for the slow accumulation of knowledge and for the perfecting of the necessary methods to reproduce the desired conditions. Beginning in the late 1940s there was increasing success in isolating cell fractions that were functional. The general method is to grind up, or homoge- nize, tissues or masses of cells and then to fractionate the homogenate, usu- ally by centrifuging. When this is done, the heavier particles, such as nuclei and unbroken cells, are thrown to the bottom of the centrifuge tube. Smaller and less dense particles form layers above the nuclear layer. For example, a distinct layer immediately above the nuclear layer contains nearly all of the mitochondria. Above this is a layer consisting almost solely of fragments of the endoplasmic reticulum with the attached ribosomes. The uppermost layer, or supernatant, is a liquid free of all but the smallest particles. The lay- ers are far from pure, but with this analytical procedure it is possible to obtain crude preparations of mitochondria or ribosomes (methods were developed for removing the ribosomes from the fragments of the endoplasmic reticulum). Some of these cell-free fractions are able to function in limited ways and for limited periods of time. Thus the fraction containing the mitochondria is capable of carrying out oxidative reactions and of forming adenosine triphos- phate (ATP), which is the immediate source of energy for all the cell’s reac- tions. Of greater interest to us, Paul C.Zamecnik and his associates discov- ered that protein synthesis could also occur in cell-free fractions of cells. In this case, the layers containing the ribosomes and the supernatant were necessary. At last it had become possible to study directly the synthesis of proteins and to learn how this might be controlled by genes. Probably more individu- als were involved in these investigations than took part in the 

HEREDITY AND DEVELOPMENT: SECOND EDITION 186 entire development of genetics before 1950. The critical element in this research was the generous financial support made available by the United States Government through its agencies such as the National Science Founda- tion, the Public Health Institutes, and the Atomic Energy Commission. Some of the key experiments involved large teams of investigators and the use of equipment costing hundreds of thousands of dollars. Without this help many of the experiments could not have been done. Interestingly, many outstand- ing scientists in other countries were also supported by the U.S. This was considered to be a proper use of public funds since discoveries in science would be for the benefit of all mankind. The theory of gene action that was eventually formulated is elegant and relatively simple, but a tortuous path led to the conceptual goal. It is useful, therefore, briefly to anticipate the conclusions so that the narrative can be followed more readily. The primordial substance is DNA. It serves as a template for the synthesis of three kinds of RNA: messenger RNA (mRNA), ribosomal RNA (rRNA), and transfer RNA (tRNA). Messenger RNA carries the message ‘how to make a protein’ from the DNA to the cytoplasm where it comes in contact with the ribosomes. Transfer RNA combines with amino acids and brings them to the mRNA bound to the ribosomes. It is here that the amino acids separate from the transfer RNA and join one another to form a polypeptide. The direction of this genetic control is from the gene to the amino acids but the argument can be better developed by beginning with the amino acids and working back to DNA. Amino Acid Activation. Proteins are not only huge molecules—recall that hemoglobin is composed of nearly 600 amino acids—but they are syn- thesized in an exact manner. Recall again that a difference in one amino acid in β chain of hemoglobin changes that molecule from the normal oxygen- carrying substance to the defective molecule associated with sickle cell ane- mia. The synthetic mechanisms in the cell, therefore, must be capable of pro- ducing extraordinarily precise products. The cell makes its proteins nearly exclusively from 20 kinds of amino acids, small amounts of which are present in the fluid portions of the cell. The initial step in the union of amino acids so they can ultimately form polypep- tides involves ATP and a group of enzymes known as the aminoacyl tRNA synthetases. Energy is required and, as is usual in cells, it comes from the ATP. For purposes of description let us assume that the cell is making a polypep- tide composed of only two kinds of amino acids, alanine and 

HEREDITY AND DEVELOPMENT: SECOND EDITION 187 glycine, and that these alternate with one another to form a chain of 100 amino acids. The amino acids are first activated by joining with ATP. This reaction occurs on the surface of one of the aminoacyl tRNA synthetases. The cell contains at least 20 kinds of these synthetases, one or more for each type of amino acid. Thus, in our example, 50 alanine molecules would combine with ATP on the surface of 50 molecules of alanine aminoacyl tRNA synthetase and the 50 molecules of glycine would do the same except that the enzyme would be glycine aminoacyl tRNA synthetase. Transfer RNA. The next step is the union of the activated amino acid, now on the surface of the aminoacyl tRNA synthetase, with a tRNA molecule. Transfer RNA molecules are made on the DNA template by regions that can be called the tRNA genes. They have a structure complementary to the DNA that makes them. They are comparatively short chains: some of the first for which the structure was determined consist of 77, 78, or 85 nucleotides. They are composed largely of the bases typical for RNA—arginine, guanine, cytosine, and uracil—plus small amounts of others such as inosine, psue- douridine, and ribothymidine. The cell has more kinds of tRNA than there are kinds of amino acids, yet each tRNA can react with only one kind of amino acid. Consequently in some cases there may be several kinds of tRNA for a specific amino acid. All of the tRNAs are alike in having the same sequence of bases at one end of the molecule: cytosine-cytosine-adenine. Let us return to our example. A reaction now occurs between each molecule of activiated alanine, joined to its synthetase, with a molecule of alanine tRNA. Similarly each molecule of activated glycine on the surface of its synthetase reacts with a molecule of glycine tRNA. The reaction is the same in both cases: the amino acid is transferred from the aminoacyl tRNA synthetase to the end of the tRNA molecule that has adenine (the C-C-A end). The end result will be 50 alanine+ alanine tRNA molecules and 50 glycine+glycine tRNA molecules. The synthetases, having been freed, can now combine with other amino acid molecules and repeat the process. The extraordinary specificity of these reactions is due to the composition of the molecules involved as well as to their precise form. The combination of the amino acid and the tRNA might appear to be non-specific, since the combination is always with an adenine at the C-C-A end of the tRNA molecule. However, the fact that a specific kind of amino acid can combine first with a specific kind of synthetase and then the combination join only with a specific kind of tRNA is due to the 

HEREDITY AND DEVELOPMENT: SECOND EDITION 188 three-dimensional structure of the molecules involved. The shapes of the molecules must ‘fit,’ otherwise there will be no reaction. The next step will involve the separation of the amino acids from tRNA and their union in precise ways to form proteins. Ribosomes, messenger RNA, and various sorts of enzymes are involved. Ribosomes. The ribosomes, which Brachet and others suggested were involved in protein synthesis, appear as tiny granules under the electron microscope. They may be extremely abundant: in rapidly dividing E. coli cells about one-third of the dry mass is composed of ribosomes. They are composed of two subunits, one large and one small. Ribosomes seem to be composed almost entirely of three types of ribosomal RNA (about 60 per cent) and protein (about 40 per cent). The rRNA is made directly on DNA by the ribosomal RNA genes. The ribosomal proteins are made indirectly by DNA in the process now being outlined. Two of the rRNAs, the longest and the shortest (they are called 23S and 5S), seem to be restricted to the larger ribosomal subunit and the middle size rRNA (called 16S) seems to be in the smaller subunit. Ribosomal RNA differs in composition from both tRNA and messenger RNA in having larger amounts of guanine and cytosine. Messenger RNA. Messenger RNA molecules are long chains of nucleotides formed on the surface of DNA. They are complementary in com- position to the DNA. That is, if the sequence of bases in a short section of DNA is: —thymine—adenine—cytosine—guanine then the mRNA made by this section of DNA will be: —adenine—uracil—guanine—cytosine. A molecule of mRNA may be formed by one gene or by several sequential genes. A special enzyme, RNA polymerase, is required for the joining of individual nucleotides to produce the mRNA molecules. After being formed, the mRNA molecules move from the surface of the DNA into the cytoplasm. When the mRNA reaches the cytoplasm, it becomes closely associated with ribosomes. The tRNA molecules with their attached amino acids then come in contact with the mRNA. Next the amino acids separate from their tRNA carriers and join one another to form a long polypeptide chain. Proteins are formed by the reactions just outlined but the account omits any mechanism that will lead to an exact specificity of protein 

HEREDITY AND DEVELOPMENT: SECOND EDITION 189 structure. Recall that we are attempting to synthesize a polypeptide chain composed of 100 amino acids, half alanine and half glycine. Let us make the reasonable assumption that our molecule cannot play its normal role in the cell unless the alanine and glycine are linked in some precise manner—we could say that the order for one end of the polypeptide might have to be: A— G—G—A—G—A—A—A—A—. How could this be done? The discovery of the manner in which the specificity of DNA becomes reflected in the specificity of proteins was one of the seminal events in mod- ern genetics. For this reason it is worthwhile to review the experiments that led to an understanding of what it is about a gene that permits it to have a spe- cific function. Some of the conclusions have been anticipated but they were stated as facts. Now we will learn the experimental basis of the facts. Most of the observations and hypotheses concerned bacteria and viruses but the gen- eral conclusions are now believed to apply to all organisms. The DNA in a bacterial cell is part of a chromosome that is free in the cyto- plasm; in more complex organisms the chromosomes are in the nucleus, which is separated from the cytoplasm by a nuclear membrane. One could imagine that proteins could be made in close proximity to DNA, in which case the supervision of protein synthesis by DNA could be fairly direct. It was discovered, however, that proteins are made largely in the cytoplasm, often at a relatively great distance from the DNA. Considerations such as these led Francis Crick, in 1958, and François Jacob and Jacques Monod (of the Pasteur Institute in Paris), in 1961, to the hypothesis that a substance must carry the message from DNA in the chromo- somes to the ribosomes in the cytoplasm. Theoretical considerations, as well as some data, suggested that the hypothetical substance might be a specific kind of RNA. Therefore the name messenger RNA seemed appropriate. The DNA of the gene, therefore, can be thought of as doing two things. First, it can make copies of itself in the manner suggested by Watson and Crick (Fig. 8–4). Thus, if one DNA strand has the base sequence: adenine-thymine-adenine-cytosine-guanine-thymine it can serve as the basis for making the complementary strand consisting of: thymine-adenine-thymine-guanine-cytosine-adenine. Second, the same strand of DNA can also serve as a basis for making 

HEREDITY AND DEVELOPMENT: SECOND EDITION 190 RNA. Remembering that RNA has uracil instead of thymine, the base sequence for a messenger RNA made from the DNA of our example will be: uracil-adenine-uracil-guanine-cytosine-adenine. The hypothesis of Jacob and Monod was this: genetic information is car- ried from genes to ribosomes by mRNA. They suggested two deductions: 1. Molecules having the assumed properties of mRNA must be present in cells that are making proteins. The molecules must be polynucleotides that are produced by the chromosomes and become attached to the ribo- somes. Furthermore, the base composition of the messenger RNA must be complementary to the base composition of the DNA that makes it. 2. The same ribosomal particle should be able to participate in the synthe- sis of different proteins at different times, depending on the type of mRNA. Both of these deductions have been tested and found to be true. Two experi- ments will be chosen from the many that establish the existence of mRNA— the first deduction. If one wishes to observe the formation of a substance in the nucleus and its movement into the cytoplasm, special techniques must be used. Presumably these events are rapid, occurring in seconds or minutes, and the amounts of materials involved are too small to be detected by the usual methods of the analytical chemist. In many instances technical problems of this sort can be surmounted by the use of radioactive isotopes. If one wishes to trace the movements of RNA it is necessary to mark the RNA in some specific way, so that it can be distinguished from all other substances in the cell. In 1960, M.Zalokar used a substance specific to RNA, namely, uridine (uracil plus ribose) which contains tritium (a radioactive isotope of hydrogen −H3). The mold Neurospora was given the radioactive isotope for short peri- ods of time, one to four minutes. Then he examined the cells at frequent inter- vals. During the first few minutes the radioactivity was restricted to the nucleus, indicating that the uridine was located in this part of the cell. After eight minutes the label began to appear in association with the ribosomes. These observations were interpreted as follows: the uridine enters the cell and, in the nucleus, is incorporated into RNA; later this RNA moves into the cytoplasm and joins with the ribosomes. This movement is exactly what one would expect of the hypothetical messenger RNA. Other data obtained with the 

HEREDITY AND DEVELOPMENT: SECOND EDITION 191 radioactive uridine showed that at least 99 per cent of the cell’s RNA is made in the nucleus and then migrates into the cytoplasm. Another experiment had been reported by E.Volkin and L.Astrachan in 1957. Recall that part of the messenger RNA hypothesis demands that the base composition of the messenger RNA must conform to the base composi- tion of the DNA by which it is formed. Recall also the results when the bac- terium Escherichia coli is infected with the T2 phage (page 161): almost immediately the bacterial cells stop making their own specific molecules and begin to make phage DNA and phage proteins. If the messenger RNA hypothesis is correct, the events would be as follows: mRNA would be made on the phage DNA instead of on the bacterial DNA; this new and different mRNA would move to the ribosomes where it would give the instructions for making phage proteins. If the bacterial DNA and the phage DNA differ in their base compositions, there should be a difference between the RNA pro- duced by an uninfected bacterial cell and that produced by a cell after it has been infected by T2 phage. This expectation was borne out. After the phage had entered the cell the synthesized RNA reflected the base composition of the phage DNA, not the bacterial DNA. The second deduction of Jacob and Monod, that the ribosome is non- specific, was also shown to be true. In 1961 S.Brenner (of Cambridge Univer- sity), F.Jacob (of the Pasteur Institute), and M.Meselson (then at the Califor- nia Institute of Technology) worked together at the California Institute of Technology along the general lines of those of Volkin and Astrachan. They also used E. coli and T2 phage. Bacterial cells were given various isotopes and the experiments were designed so that the ribosomes and RNA produced before and after the phage entered the cells could be distinguished. They were able to show that ‘(1) After phage infection no new ribosomes can be detected. (2) A new RNA with a relatively rapid turnover is synthesized after phage infection. This RNA, which has a base composition corresponding to that of the phage DNA, is added to pre-existing ribosomes…. (3) Most, and perhaps all, protein synthesized in the infected cell occurs in pre-existing ribosomes.’ Thus, ‘Ribosomes are non-specialized structures which synthe- size, at a given time, the protein dictated by the messenger they happen to contain.’ The Message. The sender of the message (DNA), the messenger (messen- ger RNA), the helphers (ribosomes), and the consequence of the message (a specific protein) have all been described—but what is the message? Part of the answer came from arm chair speculation and part from 

HEREDITY AND DEVELOPMENT: SECOND EDITION 192 some extraordinarily sophisticated experimentation. The speculation, which we shall consider first, has consisted, more or less, of playing the ‘numbers game.’ The basis of protein specificity lies in the sequence of amino acids that comprise the protein. The data available in 1960 suggested that, beyond a reasonable doubt, the sequence of amino acids is determined by the genes. If DNA were composed of 20 different bases, one would suspect that each base would correspond to an amino acid. Thus a thymine in a particular loca- tion in DNA might specify that leucine should occupy a specific spot in a protein molecule. Such a scheme cannot work, of course, because there are only four bases in DNA: thymine, adenine, guanine, and cytosine. Could two bases specify a particular amino acid? Thus thymine-guanine might be thought to be the code for leucine or for some other amino acid. This is also impossible: there can be only 42, or 16, permutations of pairs of the four bases—and there are 20 amino acids. With three bases, however, there are 43, or 64, possible permutations. Thus a code composed of triplets of bases would be the minimum number required. The total of 64 is more than three times the number required. There is the possibility that the code for some amino acids might consist of two bases and that for others of three bases. But the scientists’ love of symmetry and order led most of them to adhere to the hypothesis that groups of three bases in the DNA molecule must somehow contain the information for lining up specific amino acids in polypeptide chains. The hypothesis was expanded to suggest this model. Let us assume, for example, that the sequence adenine-guanine-cytosine is the code for serine. The mRNA formed on this part of the DNA molecule will be uracil-cytosine- guanine (UCG). (The mRNA molecule is large and we are now discussing one small portion of it—a single triplet.) This messenger RNA molecule, with its UCG triplet, becomes attached to a ribosome. Somehow the serine tRNA, with its attached serine, reaches that portion of the ribosome contain- ing the UCG triplet of the mRNA. The serine becomes detached from the transfer RNA and then attaches to the amino acid on the adjacent tRNA molecule on the messenger RNA. This process is continued and other amino acids are added one by one to the growing polypeptide chain. When the chain is complete it becomes detached. In this manner the code of DNA becomes reflected in the specific amino-acid sequence of the protein. Hypotheses suggested by this model could not be tested directly by the methods then available. As is often the case, indirect methods suggested the answer. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 193 It had been discovered by M.Grunberg-Manago and S.Ochoa (of New York University School of Medicine) that RNA could be made synthetically from mixtures of the four ribonucleotides and the enzyme polynucleotide phosphorylase. Any combination of nucleotides, or even one kind alone, can be used. Thus the enzyme, plus uracil nucleotides, will form a synthetic RNA composed solely of a long chain of these nucleotides. This particular syn- thetic RNA is called poly U. Methods had been perfected, as mentioned earlier, for obtaining the syn- thesis of proteins in cell fractions. The basic ingredients were the ribosomes and the supernatant. These methods were further refined by W.M.Nirenberg and J.H.Matthaei, of the National Institutes of Health. They were able to obtain a cell-free system, from fractionated E. coli cells, that would readily combine amino acids to form proteins. Available theory plus available techniques suggested a critical experiment. What would happen if the cell-free system, which could synthesize proteins, was given a synthetic RNA? Could this RNA serve as a messenger? Nirenberg and Matthaei added poly U to their system plus an abundance of each of the 20 amino acids. Protein was formed, but it consisted solely of long chains of the amino acid phenylalanine. The other 19 amino acids were not used. Thus it seems that a sequence of uracil alone carries all the information needed to ‘tell’ the ribosomes to join phenylalanines together. These experi- ments did not indicate how many uracils were needed but, if the code is a triplet, uracil-uracil-uracil (or UUU) is the code for phenylalanine. Subse- quent experiments proved that triplets of the RNA bases, which were named codons, constitute the code and, finally, it became possible to identify the amino acids specified by each codon (Table 8–1). Seemingly it requires only 20 codons to specify 20 amino acids but nature has not been content with this minimum number. Of the 64 possible permuta- tions of three bases, 61 appear to be functional in coding. Only UAA, UAG, and UGA are not (later we will learn that these triplets have other functions). This means that some of the amino acids must be specified by more than one codon. The actual number varies: for some amino acids there may be six codons and for others only one. Thus UCU, UCC, UCA, UCG, AGU, and AGC all code for serine. There are five amino acids coded by four codons each. Three codons serve for one, isoleucine. Nine amino acids are coded by two codons each. Finally, tryptophan and methionine have a single codon each. Thus all but two amino acids have more than one codon but no codon codes for more than one amino acid. This has been one of the most important discoveries in molecular biology. The code is said to be unambiguous since each codon specifies only one amino acid; it is said to 

HEREDITY AND DEVELOPMENT: SECOND EDITION 194 Table 8–1 The 64 triplet codons that can be formed with the four RNA bases together with the amino acids they specify. U=uracil, C=cytosine, A=adenine, and G=guanine. TRIPLET AMINO ACID TRIPLET AMINO ACID CODED CODED UUU phenylalanine CUU leucine UUC phenylalanine CUC leucine UUA leucine CUA leucine UUG leucine CUG leucine UCU serine CCU proline UCC serine CCC proline UCA serine CCA proline UCG serine CCG proline UAU tyrosine CAU histidine UAC tyrosine CAC histidine UAA (none) CAA glutamine UAG (none) CAG glutamine UGU cysteine CGU arginine UGC cysteine CGC arginine UGA (none) CGA arginine UGG tryptophan CGG arginine AUU isoleucine GUU valine AUC isoleucine GUC valine AUA isoleucine GUA valine AUG methionine GUG valine ACU threonine GCU alanine ACC threonine GCC alanine ACA threonine GCA alanine ACG threonine GCG alanine AAU asparagine GAU aspartic acid AAC asparagine GAC aspartic acid AAA lysine GAA glutamic acid AAG lysine GAG glutamic acid AGU serine GGU glycine AGC serine GGC glycine AGA arginine GGA glycine AGG arginine GGG glycine be degenerate because usually more than one codon codes for each amino acid. The data so far available indicate that the code is universal, that is, the specific relations between codons and amino acids are the same for all organisms. The specificity of the code depends not only on the base composition of the codon but on the sequence of these bases as well. One might not have expected a cell to distinguish between GGC and the reverse, CGG, 

HEREDITY AND DEVELOPMENT: SECOND EDITION 195 but it does. That portion of messenger RNA with GGC instructs the cell to incorporate glycine into the protein but, when the sequence is CGG, arginine is incorporated. Perhaps you have noticed that the argument being developed is incomplete in several important ways: it has not been explained how the sequence of codons in mRNA ensures a certain sequence of amino acids; neither has it been explained how it comes about that the proper mRNA is formed by a gene when, according to the Watson-Crick model, the gene should consist of a pair of complementary nucleotide strands. Each strand should produce a different mRNA and, hence, a different protein. The Anticodons of tRNA. If a specific sequence of amino acids in a polypeptide chain is the reflection of the sequence of codons in mRNA, the tRNAs with their attached amino acids must line up on the mRNA in one, and only one, sequence. The mechanism that assures this appears to be another case of complementarity. We have already learned of two types of complementarity. The replication of a strand of DNA involves the production of a complementary strand, not a duplicate. Similarly, the mRNA produced on the DNA template is also complementary. The relation of tRNA and mRNA appears to be a third example. Among the approximately 80 nucleotides of the tRNA molecules, there seems to be a triplet of bases, the anticodon, that is complementary to the codons of mRNA. The complementary sequence of bases of the codon and anticodon allow them to form weak bonds with one another (Table 8–2). Thus the anti- codons will be the same as the DNA triplets except that U replaces T. The details are far from established and this account should be regarded as a highly simplified model. For one thing the unusual bases found in tRNA (p. 187) are being ignored, yet they will probably turn out to be important. With this additional information, we can re-examine the interaction of ribosomes, mRNA, and the tRNAs that results in the synthesis of a polypeptide. The surface of the ribosomes is the place where the polypeptide chain is synthesized. The data seem to suggest that there are two active sites on the ribosome: the first where a tRNA molecule, with its amino acid, attaches to the mRNA; the second where the amino acid joins the polypeptide chain and is released from its tRNA. The events seem to be about as follows (Fig. 8–8). Assume that the prob- lem is to synthesize a polypeptide that has alanine and glycine in 

HEREDITY AND DEVELOPMENT: SECOND EDITION 196 8–8 Protein synthesis. A highly schematic representation of amino acids being attached to a grow- ing polypeptide chain. A single ribosome is shown at three different times. At I an alanine tRNA is attaching to a GCA triplet on the mRNA. A glycine tRNA is shown at the right in the cyto- plasm. The ribosome 

HEREDITY AND DEVELOPMENT: SECOND EDITION 197 Table 8–2 A simple model of the complementary relations of the DNA, mRNA, and tRNA triplets that code for alanine and glycine. IF CODING FOR… THE DNA THE mRNA AND THE ANTI- TRIPLETS WILL CODONS WILL CODONS OF BE… BE… tRNA WILL BE… Alanine Glycine this sequence near the middle of the molecule: A—G—G—A—G. We will assume that the cell has an ample pool of alanine that has been activated and attached to alanine tRNA and of glycine that has also been activated and attached to glycine tRNA. When one of the mRNA codons for alanine, such as GCA, is on the first site of a ribosome, the anticodon on an alanine tRNA recognizes the codon on mRNA. The mRNA and alanine tRNA are briefly held together by the attraction of codon for anticodon. The alanine is then attached to the polypeptide chain being formed. The alanine tRNA, now attached to the polypeptide chain by its alanine, moves to the second site on the ribosome. The first site of the ribsome is then in contact with the glycine codon, GGU. A molecule of glycine tRNA, with its attached glycine, will link to the mRNA—again the bond being between codon and anticodon. The glycine then becomes attached to the alanine at the growing end of the polypeptide. A shift now occurs: the glycine tRNA moves to the second posi- tion on the ribosome while the alanine tRNA is both ejected from moves to the right and the alanine tRNA is moved to the second site on the ribosome, as shown in II. The alanine molecule becomes attached to the polypeptide chain. The glycine tRNA now occupies the first site on the ribosome and attaches to the GGU triplet on the mRNA. The ribo- some moves again to the right shifting the glycine tRNA to the second site (III). Its glycine becomes attached to the alanine on the polypeptide chain. The alanine tRNA loses its connection with alanine and the ribosome and becomes free in the cell. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 198 this site and loses its connection with its alanine. This process continues until the message of mRNA is translated into a specific sequence of amino acids in a polypeptide. Is One Strand Read, or Two? In Table 8–2 the DNA triplets that code for alanine and glycine are given as CGT and CCA (there are other possible triplets; see Table 8–1). This is the sequence in one of the two DNA.strands. The complementary sequence in the other DNA strand would be GCA GGT. If mRNA is made on this strand, the codons will be CGU CCA. The first codon would carry the message for arginine; the second for proline. Since the gene is composed of two complementary strands of DNA, one might expect it to produce two quite different mRNA molecules. The result would be two entirely different proteins associated with each gene. It seems most unlikely that such a system would work, so the hypothesis was advanced that only one of the two DNA strands can serve as a template for mRNA synthesis. There are data to support this hypothesis. One of the viruses, known as ø X 174, is unusual in having its DNA in a single strand rather than the typical double helix—at least for part of its life cycle. The virus is single stranded when it invades its host bacterial cells, but once inside a complementary DNA strand is formed. The double-stranded virus then makes mRNA. One could surmise that the mRNA might be made: 1. Only from the original DNA strand that entered the bacterium. 2. Only from the complementary DNA strand. 3. From all or parts of both strands. Let us suppose that the single-stranded DNA that first enters the bacterial cell has a DNA sequence shown as the left DNA strand in Table 8–3. If so, then the newly synthesized complementary strand will have the sequence shown as the right strand. The two strands will form completely different mRNA molecules, as the table shows. Methods for extracting mRNA and comparing it with the virus DNA are available. When the experiment was done, the data showed that the mRNA was formed only by the DNA strand that was synthesized in the cell. This is interesting in itself but the important conclusion for us is that only one of the two strands serves as the template for mRNA synthesis. It is not fully known how this comes about but it does serve to prevent biochemical chaos in the cell. The controlling mechanism must be at the level of mRNA synthesis and involve the participating enzymes. RNA Polymerase. In 1960 an enzyme was discovered that could catalyze the synthesis of RNA in vitro. It was named RNA polymerase. The raw 

HEREDITY AND DEVELOPMENT: SECOND EDITION 199 Table 8–3 If both strands of DNA were involved in protein synthesis, this is what would happen: materials required for this synthesis are the four ribonucleoside triphos- phates. (A ribonucleoside is composed of one of the four RNA bases, gua- nine, cytosine, uracil, or adenine combined with ribose; this plus three phos- phate groups makes a ribonucleoside triphosphate; a nucleotide is a nucleo- side plus one phosphate group.) The four triphosphates are not enough: some DNA must also be present if the reaction is to occur, as well as traces of other substances. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 200 When RNA polymerase, the four ribonucleoside triphosphates, and DNA are together, the enzyme splits off two phosphates from the triphosphate nucleosides (making them nucleotides) and joins them to form a long RNA molecule. RNA polymerases (there are different kinds) seem to be the key enzymes in the synthesis of the three kinds of RNA: mRNA, rRNA, and tRNA. The template in all three cases is a single strand of DNA. However, RNA poly- merase seems to function better with double-stranded DNA even though it copies only one. The details of this reaction are not fully known but possibly the events are something like this. A RNA polymerase molecule moves along double- stranded DNA transcribing one of the strands into the language of RNA. That is, if the polymerase finds an adenine nucleotide at one site on the DNA, it will add uracil in the RNA that is being synthesized. If the sequence of nucleotides in DNA is TAGCGGA, the sequence in the RNA will be AUCGCCU. DNA Polymerase. There is a similar mechanism for the replication of DNA. When Watson and Crick proposed their model for DNA replication, it was assumed that enzyme(s) must be involved. About five years later such an enzyme was discovered by Arthur Kornberg and his associates at Stanford University. Here again the raw materials are triphosphates: in this instance the four deoxyribonucleoside triphosphates. In addition, double-stranded DNA is required as a primer. With all of these, new DNA can be synthesized in vitro. The two end phosphates are split from the nucleoside triphosphates, thereby producing nucleotides, which are joined to form the DNA molecules. The primer DNA serves as a template and the newly synthesized DNA is identical to it, as one would expect from the Watson-Crick scheme. Thus Kornberg was able to reproduce in vitro one of life’s most fundamen- tal in vivo events. A dramatic confirmation of this work came from his labora- tory in 1967: If the DNA of a virus is used as the primer, DNA polymerase joins the nucleotides in such a way as to produce strands of DNA identical to that of the virus. This can be tested by adding some of the synthetic viral DNA to bacteria: they are invaded and the consequences are identical to natu- ral infections. There are other enzymes, less well known, that are also involved in DNA replication. Some, for example, seem to be able to repair broken strands of DNA and even replace missing or incorrect nucleotides—abilities which may go far to explain why mutations are so rare, if we think of mutations as injuries to DNA that do not get repaired. Interim Summary: Structure and Function of Genes. We have reached the end of an important episode in the intellectual history of mankind. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 201 The accomplishments that began with Griffith’s discovery of transformation in Diplococcus and led to the cracking of the genetic code by Nirenberg and Matthaei are without parallel in biology and are difficult to match in any sci- ence. With a high degree of assurance, we can give a general definition of the molecular composition of the gene and explain how it performs its control- ling function in the cell. The gene is composed of a double helix of DNA, two long strands of nucleotides wound around one another. The uniqueness of any gene is a con- sequence of the sequence of these nucleotides. DNA has four main functions: 1. It serves as a template for making copies of itself. 2. It serves as a template for making transfer RNA. 3. It serves as a template for making ribosomal RNA. 4. It serves as a template for making messenger RNA. The fact that the gene can be duplicated exactly is the basis of constancy of inheritance. The rare errors in duplication are mutations. The three types of RNA are concerned with the synthesis of proteins, long chains of precise sequences of amino acids. The sequence of amino acids is a reflection of the sequence of nucleotide bases in DNA. Three bases are required to specify a single amino acid. If the sequence is guanine-thymine- adenine, that portion of messenger RNA formed by this triplet will be the codon cytosine-adenine-uracil. The messenger RNA molecule moves to the cytoplasm and becomes associated with ribosomes. Amino acids combine with specific tRNA molecules. The tRNA then attaches to the mRNA proba- bly by means of a specific region that is complementary to the codon of the mRNA. In our example, the tRNA would probably have the sequence gua- nine-uracil-adenine as the active site. Thus, opposite each codon of the mRNA molecule there would be a tRNA with its appropriate amino acid. The amino acids then join one another and sunder their connections with the tRNA molecules. Thus, a protein is synthesized. These events are summarized in Table 8–4 where it is shown how a short length of a polypeptide chain is formed. A more visual representation is shown in Figure 8–8. This has been a summary of some of the mechanisms by which DNA can supervise the synthesis of specific proteins. Nothing has been said about the control of this supervision, that is, whether a given section of DNA will be active. Regulation of Gene Function. A healthy, rapidly growing E. coli cell has about 10,000,000 protein molecules, representing a huge array of 

HEREDITY AND DEVELOPMENT: SECOND EDITION 202 Table 8–4 The relation of specificity in DNA, messenger RNA, and protein. IF THE SEQUENCE OF …THEN THE …AND THE PROTEIN BASES IN ONE DNA SEQUENCE OF BASES WILL HAVE THE STRAND IS… IN MESSENGER RNA SEQUENCE OF AMINO WILL BE… ACIDS G C T A HISTIDINE A U C G A U VALINE A U A U A U LEUCINE T A A U A U LEUCINE C G T A G C THREONINE G C G C G C PROLINE G C C G A U VALINE A U C G T A GLUTAMIC ACID T A T A T A LYSINE T A types, and each made according to the general scheme already described. Some kinds are abundant, others are rare. Such cells can divide into two daughter cells in half an hour, so in one hour the descendants of the original cell will have a total of 40 million protein molecules—an incredible synthetic feat. Think also of the complexity of cells in higher organisms. Your blood cells synthesize hemoglobin. Other cells do not synthesize hemoglobin but they do make many different kinds of proteins. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 203 What controls these vast and precise biochemical events? Clearly there must be mechanisms that can start a gene making RNA and others that can stop it. The genes responsible for hemoglobin synthesis are ‘turned on’ in red blood cells but, presumably, they never function in other cells. The answers to these questions are not completely known but there is a useful working hypothesis. The Operon Hypothesis. In Escherichia coli there is one gene that con- trols the synthesis of the enzyme tryptophan synthetase, and another that con- trols the synthesis of the enzyme β-galactosidase. Presumably the mechanism is the usual one: the tryptophan synthetase gene transmits through messenger RNA the code for joining amino acids in the specific way that makes trypto- phan synthetase. The events beginning with the β-galactosidase gene and ending with the formation of the specific enzyme β-galactosidase are appar- ently the same. Tryptophan synthetase has a specific role in the cell. Under its catalytic influence, indole and serine are combined to form tryptophan. The amino acid tryptophan is one of the 20 found in all living organisms. E. coli cells normally synthesize tryptophan synthetase but if they are grown in a medium containing tryptophan the enzyme is no longer made. Thus the gene functions when there is no tryptophan in the medium, but is inhibited when tryptophan is present. This is a clear example of a gene’s action being controlled by a non-nuclear substance. The situation with respect to the β-galactosidase gene is almost the reverse. E. coli can use either glucose or lactose as a carbon source for intra- cellular syntheses. When there is no lactose in the cell’s environment, no β- galactosidase is synthesized. If lactose is added to the medium in which the cells are growing, synthesis of the enzyme begins. β-galactosidase hydrolyzes lactose to glucose and galactose. We must assume that the gene does not function in the absence of lactose but functions when lactose is present. Once again, this is an example of non-nuclear control of a gene’s activity. Experiments of this sort were the basis of the operon hypothesis for the regulation of the functioning of genes. This hypothesis was proposed in 1961 by F.Jacob and J.Monod of the Pasteur Institute in Paris. It has stimulated a large amount of fruitful work and there are enough data to suggest that it is essentially correct. Possibly its relative precision can be compared to Mendelism in 1900—correct in broad outline, useful for organizing data, testable by experimentation, but destined to be greatly extended by future discoveries. An essential element of the operon hypothesis is that genes do not 

HEREDITY AND DEVELOPMENT: SECOND EDITION 204 control themselves. A gene does not suddenly decide, ‘It is time to make β- galactosidase’ and then proceed to produce the mRNA that brings this about. The stimulus for activity comes ultimately from the biochemical environ- ment in which the gene is situated. An operon is thought of as a portion of a DNA strand consisting of genes that serve as templates for the synthesis of one or more types of mRNA together with other areas that regulate this synthesis. Figure 8–9 shows the lactose operon (lac operon) of E. coli. It consists of three genes, z, y, and a; a promoter site, where RNA polymerase attaches to the DNA and begins the transcription of mRNA; and the operator, which either allows the RNA polymerase to move along the DNA strand, or blocks its passage. Each operon is assumed to have its own regulator gene, which produces a mRNA that controls the synthesis of a regulator protein. The regulator pro- tein exists in two forms, active and inactive, a difference apparently due to the shape of the molecule. The active form of the regulator protein combines with the operator and, as we shall see, blocks transcription of mRNA. When there is no lactose in the cell, no β-galactosidase is synthesized (Fig. 8–9). The regulator gene for the lac operon makes mRNA, and this leads to the synthesis of the regulator protein in the active shape. The regula- tor protein then binds at the operator site of the lac operon. If the RNA poly- merase attaches to the promoter site to begin the transcription of mRNA, its passage along the DNA molecule is blocked by the regulator protein attached to the operator. No mRNA can be transcribed from the z gene, the one for β- galactosidase mRNA. In the presence of lactose, however, mRNA is synthesized. The regulator protein is still formed but its shape is altered in the presence of some metabo- lite(s) associated with lactose (possibly the metabolite combines with the regulator protein; this would certainly alter its shape). It can no longer com- bine with the operator. The RNA polymerase attaches to the promoter site and begins to move down the DNA strand. The operator site is not blocked so the RNA polymerase passes along the z gene transcribing it into the mRNA for β-galactosidase. It continues on to make the mRNA from the y and a genes. The mRNA for the y gene serves as a template for the synthesis of a protein that makes the cell membrane of E. coli more permeable to lactose. The protein associated with the a gene is also related to lactose metabolism but its precise role is unknown. Thus, whether a gene makes β-galactosidase or not depends on an envi- ronmental condition: the presence or absence of lactose. The biochemical economy of this is obvious: β-galactosidase has the function of 

HEREDITY AND DEVELOPMENT: SECOND EDITION 8–9 A highly schematic representation of the lac operon and its functioning. The upper portion of the diagram shows what happens when there is no lactose in the cell: no mRNA can be made from genes z, y, and a and, hence, their specific proteins cannot be synthesized. When lactose is present, as shown in the lower portion of the diagram, mRNA can be made from genes z, y, and a and their specific proteins will be synthesized. The z gene is responsible for β-galactosidase. 205 

HEREDITY AND DEVELOPMENT: SECOND EDITION 206 catalyzing the hydrolysis of lactose to glucose+galactose. In the absence of lactose and, hence, of the need to hydrolyze it, no β-galactosidase is synthe- sized. The cell can do other things with those amino acids that might have been used for β-galactosidase. The contrasting effects of lactose and tryptophan were mentioned at the beginning of this section (The Operon Hypothesis). The presence of lactose stimulates the production of the enzyme that hydrolyzes this sugar. When there is tryptophan in the medium where E. coli cells are growing, synthesis of tryptophan is repressed. Here again the economy is obvious: when ade- quate concentrations of tryptophan are present in the cell, there is no need to synthesize more. The tryptophan operon is also controlled by a regulator gene, which pro- duces a regulator protein. In contrast to the situation with the regulator pro- tein for the lac operon, which is synthesized in an active form, the regulator protein for the tryptophan operon is synthesized in an inactive form. Hence, it cannot combine with the tryptophan operator and so prevent the formation of mRNA. Therefore, mRNAs will be produced that result in tryptophan synthe- sis. When there is an excess of tryptophan in the cell, however, the regulator protein assumes the active form. It then combines with the tryptophan opera- tor site and prevents the formation of mRNA. The result: no tryptophan synthesis. There is a third type of gene action. Some operons seem to form mRNA molecules at a fixed rate that appears to be uninfluenced by the exact molecu- lar composition of the cell. The molecules controlled by these operons are called constitutive proteins. The Need for Nonsense. Recall that 61 of the 64 possible codons specify an amino acid (Table 8–1). Three do not: UAA, UAG, and UGG. These were given the flattering name nonsense codons. It was not long until they began to make sense. Figure 8–8 suggests that a single long mRNA molecule is made from the z, y, and a genes of the lac operon. If this mRNA is translated into protein, the result will be one very long polypeptide chain, not three. It has been discovered that the nonsense codons have the function of ending a polypeptide chain and therefore of permitting a single mRNA to make more than one polypeptide. They are now called terminating codons. The last triplet in the z gene and the y gene (and possibly the a gene also) seems to be one of the nonsense codons. When the ribosomes are reading the mRNA, they add the amino acids called for by each codon until they reach one of the nonsense codons. At this point no amino acid can be added to the polypeptide chain, and the chain is terminated. Synthesis of a new molecule then begins again with the sequences beyond the nonsense codon. The 

HEREDITY AND DEVELOPMENT: SECOND EDITION 207 initiation codon for the new mRNA molecule generally seems to be AUG but some evidence suggests that GUG can also serve. Summary. The gene was a ‘black-box’ for most of its history, having started life as an abstract concept that helped geneticists think about the results of their experiments. Later it became a ‘thing’ that could be associated with chromosomes and share their odyssey during mitosis, meiosis, and fertil- ization. This level of understanding was sufficient to answer most of the non- chemical questions about variation and heredity and to serve as a powerful method for improving man’s food animals and food plants. There was even enough understanding to enable him to improve himself, had there been the desire to do so. In the 1940s the gene emerged from the black-box as DNA and in the 1950s its chemical nature was more precisely defined. In the 1960s a gener- ally satisfying model was formulated for the specific nature of the gene and for its mode of action in the cell. Some geneticists now feel that the broad conceptual framework of their science has been completed: there are no big questions remaining. To be sure there is a great amount of detail yet to be added, but the great problems remaining are the roles of genes in evolution and in differentiation. In the last part of this book we will explore some of the problems of the genetic control of differentiation but, before we do, there are some practical things to be said about the genetics of man. Suggested Readings Chapter 7 of the Readings includes Gunther Stent’s article ‘DNA’ and addi- tional references. Further information on the ideas developed in this chapter can be obtained from these books and articles. CARLSON, ELOF AXEL. 1966. The Gene: A Critical History. Philadelphia: W.B. Saun- ders. DU PRAW, ERNEST J. 1968. Cell and Molecular Biology. New York: Academic Press. DU PRAW, E.J. 1970. DNA and Chromosomes. New York: Holt, Rinehart and Winston. HARTMAN, PHILIP E., and SIGMUND R.SUSKIND. 1969. Gene Action. Second Edi- tion. Englewood Cliffs, N.J.: Prentice-Hall. NOMURA, MASAYASU. 1969. ‘Ribosomes.’ Scientific American. October 1969. pp. 28– 35. PTASHNE, MARK, and WALTER GILBERT. ‘Genetic repressers.’ Scientific American. June 1970. pp. 36–44. RAVIN, ARNOLD W. 1965. The Evolution of Genetics. New York: Academic Press. STAHL, FRANKLIN W. 1969. The Mechanics of Inheritance. Second Edition. Engle- wood Cliffs, N.J.: Prentice-Hall. 

HEREDITY AND DEVELOPMENT: SECOND EDITION 208 STENT, GUNTHER S. 1971. Molecular Genetics. An Introductory Narrative. San Fran- cisco: W.H. Freeman. WATSON, JAMES D. 1968. The Double Helix. New York: Atheneum. A personal account of the discovery of the structure of DNA. WATSON, JAMES D. 1970. Molecular Biology of the Gene. Second Edition. New York: W.A. Benjamin. These are some of the specific papers referred to in the text. ALLISON, A.C. 1956. ‘Sickle cells and evolution.’ Scientific American August 1956. pp. 87–94. BRENNER, S., F.JACOB and M.MESELSON. 1961. ‘An unstable intermediate carrying information from genes to ribosomes for protein synthesis.’ Nature 190: 576–81. CRICK, F.H.C. 1962. ‘The genetic code.’ Scientific American October 1962. pp. 66–75. HURWITZ, J., and J.J.FURTH. 1962. ‘Messenger RNA.’ Scientific American February 1962. pp. 41–9. INGRAM, V. 1958. ‘How do genes act?’ Scientific American January 1958. pp. 68–74. JACOB, F., and J.MONOD. 1961. ‘On the regulation of gene activity.’ Cold Spring Har- bor Symposia on Quantative Biology 26:193–209. NIRENBERG, M.W. 1963. ‘The genetic code: II.’ Scientific American March 1963. pp. 80–94. NIRENBERG, M.W., and J.H.MATTHAEI. 1961. The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polyribonucleotides.’ Proceed- ings of the National Academy of Science 47:1588–1602. PAULING, L., H.A.ITANO, S.J. SINGER, and I.C.WELLS. 1949. ‘Sickle cell anemia, a molecular disease.’ Science 110:543–8. WATSON, J.D., and F.H.C.CRICK. 1953. ‘Molecular structure of nucleic acid.’ Nature 171:737–8. WATSON, J.D., and F.H.C.CRICK. 1953. ‘Genetical implications of the structure of deoxyribonucleic acid.’ Nature 171:964–7. ZALOKAR, M. 1960. ‘Sites of protein and ribonucleic acid synthesis in the cell.’ Experi- mental Cell Research 19:559–76. ZAMECNIK, P.C. 1960. ‘Historical and current aspects of the problem of protein synthe- sis.’ Harvey Lectures 54:256–81.