Read "Biotechnology: Science, Engineering, and Ethical Challenges for the Twenty-First Century" at NAP.edu (2024)

Page 1

Part 1—
Basic Research:
The Foundation Of Biotechnology

Biotechnology has so suddenly been thrust to the center of public consciousness that the casual observer could be forgiven for thinking that this technology had been developed overnight. This is, of course, not the case.

In a sense, the set of knowledge and techniques now collectively known as biotechnology had its beginnings more than 100 years ago with the discovery of the cellular molecule deoxyribonucleic acid (DNA). But almost a century passed before scientists discovered that DNA is the genetic material that transmits the characteristics of an organism from one generation to the next. In 1953, Francis Crick and James Watson elucidated the structure and function of DNA.

Beginning in about 1970, the pace of innovation increased exponentially; all of the scientific breakthroughs and applications associated with modern biotechnology have occurred in the past 25 to 30 years. However, this revolution would not have been possible without the body of basic scientific knowledge that had accumulated before that time.

Like most technological advances, biotechnology has its roots in scientific curiosity: the desire to better understand the biological basis of human life. The importance of basic scientific research—research aimed solely at increasing knowledge—is a theme that runs through these first chapters. In the words of Anna Marie Skalka, of the Fox Chase Cancer Center's Institute for Cancer Research and author of Chapter 2, ''research designed simply to answer general problems posed by nature often yields the most far-reaching rewards."

Page 2

This point is reinforced by Alexander Wlodawer of the National Cancer Institute, author of Chapter 3, who notes that basic research conducted years ago on "obscure chicken viruses" is the foundation for almost everything modern science knows about the human immunodeficiency virus (HIV) that causes acquired immune deficiency syndrome (AIDS).

In Chapter 1, Kathleen Matthews of Rice University describes the nature of the genetic material DNA and indicates how basic research studies on bacteria led to the discovery of restriction enzymes that recognize specific DNA sequences and act as molecular scissors. These proteins make it possible to generate reproducible sets of DNA fragments and led quickly to the ability to separate and analyze DNA, to combine DNA from different organisms, to amplify or clone DNA, and to map genes.

Recombinant DNA technology makes possible the use of living cells as factories for the production of proteins (the basis of the biotechnology industry) as well as the development of transgenic plants and animals (which contain genes transplanted from other species). The ability to fragment and analyze DNA has led to the Human Genome Project, an international effort to map the entire human genome.

Skalka and Wlodawer describe ways in which DNA analysis and recombination technologies are being used to advance understanding and treatment of two of the most complex, mystifying diseases of the modern era: AIDS and cancer. Although these two chapters are probably the most difficult in the book for nonscientists, they offer a glimpse at the frontiers of current research, where answers are being sought to basic questions about cell and molecular biology.

In the final chapter of Part 1, C. Thomas Caskey of Baylor College of Medicine discusses the implications of biotechnology for the diagnosis, treatment, and prevention of disease, raising some of the ethical questions that will be addressed in greater depth later in the book.

Page 3

1—
Overview of Terminology and
Advances in Biotechnology

KATHLEEN S. MATTHEWS

Over the past 50 years, the face of biological science has been transformed by discoveries that began at midcentury with the identification of genetic material. The pace of discovery has increased significantly in the past two decades with our ability to manipulate and examine genetic material. The barriers between subdisciplines within the biological sciences have been eroded, and we can now speak of common tools and insights that transcend disciplinary boundaries.

Genetic material is the information that is passed from generation to generation within organisms. These stored and then transmitted data determine all of the activities carried out within a living organism. Genetic information is required for development, differentiation of cells, maintenance of cell function, and reproduction, not just of each cell within an organism, but for the organism itself. Because changes in this genetic material will be passed to subsequent generations, an alternation that affects function will also be passed from one generation to the next.

DNA Identified As Genetic Material

Deoxyribonucleic acid (DNA) was known to exist in cells long before it was identified as the genetic material. It was named on the basis of its properties: it contains the sugar ribose missing an oxygen at a specific position (deoxyribose), it is located in the nucleus (nucleic), and it has the properties of an acid. DNA was not identified as genetic material until almost 100 years after its discovery in the cell.

Page 4

Avery et al. (1944) were the first to demonstrate that this material actually contained information that was passed from generation to generation. They accomplished this feat by using two different strains of Pneumococcus bacteria. The first strain caused pneumonia and death when injected into mice. In contrast, the second, a very closely related strain, did not cause disease when injected into mice. After fractionating the pathogenic bacterial strain into the different components that are found in cells (protein, DNA, lipids, etc.) and mixing these components individually with the nonpathogenic strain, Avery and colleagues monitored for an ability to transform the bacteria from nonpathogenic to pathogenic. They found that only DNA could accomplish this transformation. In addition, the bacteria that were transformed were permanently transformed: not only had the genetic material been altered, but this alteration was passed to subsequent generations. This experiment provided the first direct demonstration that DNA could serve as genetic material.

Composition Of DNA

DNA is composed of four different components called "bases." These are assembled into structures like beads on a string. The bases have the chemical composition and chemical structure shown in Figure 1-1; these components are linked together in various sequences in different DNAs through the sugar moiety and phosphate to form a deoxyribose-phosphate backbone that forms a scaffold for the bases. In DNA, two single strands come together and form a double-helical structure in which the individual strands run in opposite directions; that is, their relative orientation is antiparallel. The connections to form the sugar-phosphate backbone within each strand are covalent (strong interactions that are difficult

FIGURE 1-1 Components of DNA.

Page 5

to break), whereas base pairing interactions are weaker (noncovalent bonds) and occur between bases on different strands. The latter interactions stabilize the double-stranded structure of DNA. The individual DNA strands can be separated by changing conditions (e.g., stressing the molecule by increasing temperature), a property that can be useful experimentally.

Another property of DNA is that the four bases, thymine (T), adenine (A), cytosine (C), and guanine (G), are able to form only certain base pairs. T and A always pair, and C and G always pair. Given the antiparallel arrangement of two strands in DNA, if the sequence of one strand is known, the sequence of the other strand can be deduced from these base-pairing rules. The diameter of the DNA molecule is constant regardless of the molecule's length. The length of the bacterial genome is on the order of millions of base pairs; in humans, the length of the DNA chain is about 30 billion base pairs.

DNA Sequence Encodes Proteins

Genetic information is stored in the sequence of the base pairs within the DNA (Figure 1-2). This information is copied, or transcribed, into an intermediary message that is used as a template for the assembly of amino acids into the class of molecules called proteins. Proteins are composed of 20 different amino acids that are also assembled somewhat like beads on a string. Three bases in DNA—and consequently in the messenger ribonucleic acid (RNA) intermediary—are used to specify an amino acid. The adjacent triplet of bases specifies a different amino acid and so on throughout the entire sequence of the protein. This translation process alone, however, is not sufficient to generate a functional species.

FIGURE 1-2 DNA makes RNA makes protein.

Page 6

Each of the amino acids contains a side chain that confers specific chemical properties such that specific interactions can form between amino acids. Amino acids in a particular sequence are able to fold into a specific three-dimensional structure that gives rise to function. In this way, the information that is contained in DNA is converted to functional form. Because the sequence of amino acids varies according to DNA sequence, proteins come in a variety of sequences, shapes, and sizes. Most importantly, proteins are able to perform a variety of functions based on their size and shape.

Protein molecules carry out the myriad of functions characteristic of living organisms. For example, enzymes are proteins that catalyze metabolic reactions in cells. Other proteins serve structural roles; others control differentiation, development, and growth; still others are used in defense. Proteins are the functional expression of genetic material.

Mutations

A gene is the segment of DNA that encodes a protein (or polypeptide). A gene is basically a blueprint for the construction of a protein sequence. Any change in DNA sequence is called a mutation. A mutation can be as small as a single base pair change or it can be an insertion or a deletion or inversion of segments of the sequence. A change in the DNA changes the mRNA and consequently the protein that is assembled from this sequence. When a different amino acid or set of amino acids is incorporated into the protein product, folding into a functional structure may be altered. A mutation can result in an incompletely folded and therefore nonfunctional protein or a protein that is folded in a way that diminishes function.

Mutations can have positive effects, although what are usually observed are negative effects on proteins that result in some deleterious effect on the organism. Perhaps one of the most illustrative examples is the mutation that results in sickle cell anemia. A single base pair change alters the gene that codes for one of the two chains in hemoglobin, the oxygen carrier in red blood cells. This change results in the alteration of the properties of the intact hemoglobin molecule such that the molecules now aggregate within the cell. This aggregation alters the shape of the cell from its normal jelly donut shape, which moves easily through the capillaries, into a sickle shape, hence the name of the disease. When this change happens, the capillaries are occluded, and the cells in blood cannot circulate effectively. As a consequence, oxygen concentration in the blood drops, an event that further promotes hemoglobin aggregation, and the person enters a sickle cell crisis. All of these effects result from a single change in the DNA sequence.

Page 7

Genes must be identified within the DNA for the transcription process to occur. There are controlling sequences and signaling sequences that indicate the start of a gene and which strand should be copied in which direction. Furthermore, there are regulatory sequences that determine when, how much, and under what conditions a particular protein is produced. It is the constellation of proteins produced over time and within the different types of cells in an organism that determines developmental processes, controls cell differentiation, maintains cell type, etc.

Fragmentation Of DNA

The fact that DNA is a very long and skinny molecule complicated its chemical and physical analysis for many years after its structure was known. By analogy, managing a detailed study of a train that stretches from here to the sun (comparable length:diameter relationship for human DNA) would be simpler if it were possible to fragment the train reproducibly into smaller and more manageable pieces. Until the mid-1970s, there was very little hope of doing that with DNA because the methods available were not very successful and, of utmost importance, were not very reproducible. Out of basic research studies on bacteria came the discovery of a family of enzymes, called restriction nucleases, that can recognize a specific sequence of bases in DNA act as molecular scissors. The importance of this discovery was that for any given DNA the products were consistently the same. It was therefore possible to generate a reproducible set of fragments every time a particular DNA was treated with a particular restriction enzyme. There are now several hundred different restriction enzymes in our arsenal that recognize different sequences reproducibly.

Separation Of DNA And Analysis

It is equally important to be able to separate the pieces of DNA generated by fragmentation. The technique of electrophoresis provides the opportunity for such separation (Figure 1-3). Mixtures of DNA fragments

FIGURE 1-3 Separation of DNA fragments.

Page 8

FIGURE 1-4 Restriction fragment length polymorphism (RFLP) analysis (A),
probing (B),patterns (C), and applications (D).

can be loaded onto thin gels and subjected to an electric field. Because the charge-to-mass ratio of DNA is independent of fragment length, the largest fragments do not move very far, whereas the smallest fragments move very rapidly through the gel. Standards can be used to identify the size of the DNA fragments, and parts of the gel can be removed to isolate a specific DNA fragment.

Page 9

DNA from a human cell can be treated with a restriction enzyme or a set of restriction enzymes to produce thousands of fragments. These can be separated by size via electrophoresis to generate a smear of DNA fragments (Figure 1-4A). If the gel in which this smear is found is blotted onto a support membrane, the DNA will stick to the membrane as the same smear of many individual bands. The membrane can then be treated under conditions that separate the two strands of DNA from one another (Figure 1-4B).

Page 10

A DNA strand does not discriminate as to whether it is interacting with its natural partner or other DNA with a complementary sequence. Thus, it is possible to take another DNA and mark it with radioactivity or a visible marker. If this marked DNA (probe) is mixed with the single-stranded DNA on the membrane, it will interact with any complementary sequence if the conditions are favorable for base pair formation, and a double-stranded hybrid structure will be generated (Figure 1-4B). After excess probe DNA is removed, a detection system identifies the probe signal and thus the bands that contain the complementary sequences (Figure 1-4B).

Because the DNA of all individuals is not identical, one individual can have restriction enzyme sites that do not appear in another. When restriction enzymes are used to generate fragments that are separated by size, blotted, and probed, different patterns can be detected (Figure 1-4C,D). This method is called restriction fragment length polymorphism (RFLP) analysis (Lander and Botstein, 1986). Significant information can be derived from differences (polymorphisms) in the lengths of restriction fragments. If, for example, the appearance of a particular band correlates to the presence of a specific genetic disease, this type of analysis can be used to identify individuals who have a genetic disease before symptoms appear. Given enough markers and sufficient information about the statistical distribution of those markers in the population, it is possible to use this methodology to identify uniquely an individual (Jeffreys et al., 1986, 1990). In the future it will probably be our DNA rather than our fingerprint that is used for identification.

This method can also be applied in forensic settings where, for example, samples are obtained from a victim or a crime scene and the patterns produced are compared with those from suspects (Figure 1-4D). If there is no match, an individual can be exonerated; if there is a match, a suspect can be implicated in a specific crime. Rules for how this type of analysis is to be done are now becoming generally accepted, and the conditions under which it can be accepted in a court of law are being considered in many states. DNA information can also be used to establish familial relationships (Figure 1-4D). This approach can determine whether a child is the progeny of a particular father and mother.

Beyond the human applications, this type of analysis can be used to determine whether a bacterium carries specific types of drug resistance. This analysis is much faster than the traditional technique of allowing a bacterium to grow in the presence of different drugs as a means of determining drug resistance. There are numerous examples of the application of RFLP analysis.

A method that allows RFLP analysis to be used more widely resulted from being able to amplify specific sequences of DNA. The method is

Page 11

FIGURE 1-5 DNA amplification: polymerase chain reaction (PCR).

called polymerase chain reaction (PCR), and a Nobel Prize was awarded to Kary Mullis for its development (Mullis, 1990). The method is based on the ability to separate the two DNA strands. Because DNA strands are held together by weak forces to form a double-stranded structure, it is possible to separate the two strands. Short sequences of DNA can be made chemically and matched to the ends of each of the strands (Figure 1-5). These short sequences can also form base pairs. By using an enzyme that can recognize this structure and extend, by base pair interactions, each of the short strands, the products are two pieces of DNA that are identical to the original double-stranded DNA. This cycle can be repeated many times to increase the amount of a particular DNA sequence by about a millionfold. Out of a mixture of sequences, a sequence of interest can be uniquely amplified. This amplification allows RFLP analysis on DNA from extremely small samples.

Recombinant Dna

DNA that has been taken apart at specific sequences can be put back together. A sealing enzyme, called ligase, recognizes a gap in DNA and connects the ends, but the action of this enzyme is not sequence specific. DNA from two sources can be recombined into one intact segment (called recombinant DNA; Figure 1-6). Thus, the coding sequence for a particular

Page 12

FIGURE 1-6 Connecting different DNAs: Recombinant DNA.

protein can be combined with DNA that belongs to a different organism. This recombinant DNA can be put into the cells of yet another organism for which the product encoded is a foreign protein. Cells of different types can be used as factories for the production of specific proteins. This type of technology forms the basis for the biotechnology industry, producing many types of specific proteins, e.g., hormones, insulin, hemoglobin, tissue plasminogen activator, interleukin, and cytokines.

Obviously, there are important questions about who receives these proteins and under what conditions. There are many issues connected with recombinant DNA technology, and there is even more complicated technology. Not only can a foreign gene be put into the cells of an organism: the gene can actually be incorporated into the DNA derived from germ cells or embryonic cells of another organism. From this combination, an embryo can be produced that contains this gene that came originally from another species (called a transgene). Transgenic embryos can be put into an adult female (e.g., a mouse), which will then give birth to mice permanently carrying the transgene. Various lines of transgenic mice are being used for numerous research efforts, from studying human diseases to examining the effects of the expression of a particular protein in a specific organ. There are also transgenic cows, sheep, pigs, and plants.

What is forbidden, at the moment, is the production of transgenic

Page 13

humans. The National Institutes of Health (NIH) currently forbids any efforts to alter the germ line in humans. Instead, what is being conducted in humans at this time is somatic gene therapy. The altered cells used are not germ-line reproductive cells. The general purpose of these efforts is to use a properly functioning gene to replace a gene that is defective and not operating properly. A retroviral vector is usually used to introduce the transgene into the DNA of cells isolated from an individual. The transgenic DNA then carries the foreign gene that encodes a functional protein that can carry out the specific functions missing in the individual. The cells that are put back into the individual generally home to their organ type (e.g., muscle or blood cells), and the needed protein can then be produced in the recipient. The first such trials were on children with a deficiency in adenine deaminase (ADA), which is required for immune function. The disease produced by ADA absence is called severe combined immune deficiency disease, and children with this disease cannot be exposed to the environment because their immune systems are not functioning. NIH has carried out trials in which the ADA gene was introduced into the cells isolated from an individual, and these cells were then reintroduced into the body. The needed protein was produced in amounts sufficient to allow the treated children to be out in the environment and to assume more normal lives. Now there are at least tens and no doubt soon to be hundreds of different trials for different kinds of gene therapy in attempts to correct a specific genetic deficiency.

Searching For Specific Genes

Gene identification, that is, identifying the coding regions for proteins that have particular functions, is one major purpose for mapping the human genome. A genetic map charts a sequence of characteristics that we know are derived from a particular segment of DNA (Figure 1-7). The Human Genome Project is concerned with putting all human genetic and physical information together in an integrated map. Sequences are also being determined for DNA from other organisms: mice; fruit flies (Drosophila); worms (Caenorhabditis), where we know a lot about differentiation and development; plants (Arabidopsis is a good example); bacteria; yeast; and various others. One surprise is that there is a lot of similarity between mice and humans. What distinguishes humans from mice is much less extensive than might be expected, given the morphology and the characteristics that we associate with these organisms. The commonality as well as the differences in sequence will give us information about living organisms, their development, and the mechanisms that generate the diversity inherent in the living world.

Page 14

FIGURE 1-7 Mapping the genome.

References

Avery, O. T., C. M. MacLeod, and M. McCarty. 1944. Studies on the chemical nature of the substance inducing transformation of pneumococcal types. Inductions of transformation by a deoxyribonucleic acid fraction isolated from pneumococcus type III. J. Exp. Med. 79:137-158.

Jeffreys, A. J., V. Wilson, S. L. Thein, D. J. Wetherall, and B. A. Ponder. 1986. DNA ''fingerprints" and segregation analysis of multiple markers in human pedigrees. Am. J. Hum. Genet. 39:11-24.

Jeffreys, A. J., M. Turner, and P. Debenham. 1990. The efficiency of multilocus DNA fingerprint probes for individualization and establishment of family relationships, determined from extensive casework. Am. J. Hum. Genet. 48:824-840.

Lander, E. S., and D. Botstein. 1986. Strategies for studying heterogeneous genetic traits in humans by using a linkage map of restriction fragment length polymorphism. Proc. Natl. Acad. Sci. 83:7353-7357.

Mullis, K. B. 1990. The unusual origin of the polymerase chain reaction. Sci. Amer. April: 56.

Page 15

2—
Advances in Cell and
Molecular Biology

ANNA MARIE SKALKA

The era of molecular biology traces its origin to the study of microorganisms, tiny bacteria and the still smaller viruses, which had aroused the interest of a small group of born-again physicists. This group of scientists was interested in such microorganisms because of their relative simplicity, which seemed to offer the potential for understanding the physical basis of heredity and an answer to the question, What is life?

The work of these molecular biology pioneers was enormously successful. It led to the revelation that the blueprint for clearly inherited genetic traits and for the biochemical functions common to all living cells is encoded in DNA. The efforts of these early molecular biologists also helped to delineate the major metabolic pathways that all cells use to maintain and reproduce themselves. Finally, the knowledge gained from these studies of simple microorganisms also produced the major tools of today's biotechnology industry: bacteria that serve as factories for the production of virtually any desired protein and viruses or plasmid DNAs that serve as vehicles for amplification of any desired gene.

The reason for the success of their efforts is illustrated in Table 2-1, which contains a comparison of the approximate genetic content of these microorganisms and other more complex organisms, as well as information concerning their reproduction potential. The enormous range in the amount of genetic material is apparent from estimates of genome sizes. The early molecular biologists believed that they had a hope of understanding a genome of 45 thousand or even 4 or 10 million base pairs, but there seemed to be no chance of coming to grips with genomes of higher

Page 16

organisms, which were orders of magnitude larger in size. The other advantage of microorganisms is their rapid reproduction rate; for example, yields of approximately 100 viruses from one infected bacterium in 30 minutes or a doubling of bacterial or yeast cells every 20 minutes to 2 hours. Thus, it would be possible to analyze the effects of genetic changes in such organisms in a matter of hours or days. The gestation periods of higher organisms, however, are measured in hours, weeks, and months. It is clear from this comparison why the fast-reproducing fruit fly was a favored model of early genetics for studying multicellular organisms.

Despite the extraordinary success of these early studies, there are certain fundamental questions that these experiments could not address. The living, single-cell microorganisms that were used as models do not face the problems of community living that exist for the cells of multicellular, complex organisms such as humans, mice, or fruitflies. The profoundly different cells that make up the tissues and organs of our bodies all come from a single fertilized egg. They reach their final form and number through a complex series of tightly regulated proliferation cycles and morphological changes. We now know that these changes are a response to signaling between individual cells and between cells and their environment. The details of these processes of cellular differentiation and embryonic development are still largely obscure. These are the mysteries that we are now just beginning to unravel in the modern era of cell and molecular biology.

The Genetic Revolution

It is difficult to discuss the expansion in our knowledge of biological processes without acknowledging the technological advances that have made such progress possible. Scientific discovery and technology are inextricably connected. Methods are often created to address specific scientific

Page 17

questions, and the use of novel techniques invariably uncovers new biological phenomena. The unique advantages afforded by modern scientific technological breakthroughs are nowhere more apparent than in the enormous increase in our capacity to perform sophisticated genetic and molecular analyses in higher organisms. Heralded as the genetic revolution, application of new techniques for gene mapping, isolation, and sequencing is fueling efforts to identify and understand the genetic components of human disease. Announcements in the literature, or even in the press, of some new, important discovery occur almost daily.

Our increased ability to manipulate genomes—to introduce extra genes or to modify existing genes in chromosomes—has made it possible to investigate gene function not only at the level of an isolated cell, but also in the context of the whole organism. Thus, we can now look forward to understanding the genetic and biochemical basis of normal embryonic development and cellular differentiation with the same success as the early molecular biologists elucidated the genetic determinants of pathways of metabolism in single-celled microorganisms. We can also hope to understand the molecular basis of defects in these processes caused by genetic lesions, environmental insults, or infectious agents. Medical sciences have already begun to use this knowledge for diagnosing, treating, and preventing disease; for improving the environment; and in agriculture. Examples are numerous. Because my scientific expertise is in the areas of cancer and virology, I will use a few examples from these disciplines to illustrate the current state of our knowledge and the range of challenges and opportunities that seem to be on the horizon.

Cancer (A Genetic Disease)

One of the most important contributions of the modern era of molecular biology is an understanding of the genetic basis of cancer. The fact that cancer incidence rises dramatically with age suggested to epidemiologists some time ago that cancer may be caused by an accumulation of genetic insults (estimated at three to seven; Vogelstein and Kinzler, 1993) over an individual's lifetime. We now know that cancer does arise from succeeding populations of cells in which more and more of these mutations have accumulated. We also know that the critical mutations are those that interfere at various steps in the pathways that regulate intracellular communication, proliferation, and differentiation. The result is uncontrolled cell growth, increasing disorganization, and, finally, malignancy.

A well-documented example of genes altered in colon cancer, uncovered through analysis of tissues isolated from different stages in the gradual evolution of tumors of the colon, is summarized in Figure 2-1. Genes that are the sites of the most frequent mutations observed in colon

Page 18

FIGURE 2-1 Genes altered in colon cancers. Source: Vogelstein and Kinzler (1993).

cancer and the stages at which they seem to be most critical are shown at the top of the figure. They include two genes known to be involved in two types of familial colon cancer, HNPCC (hereditary nonpolyposis colon cancer) and APC (adenomatus polyposis coli), and others including the ras, DCC (deleted in colon cancer), and p53 genes. Also included in Figure 2-1 are some less frequently observed mutations, including mutations in oncogenes and tumor suppressor genes. Oncogene mutations cause dominant, gain-of-function changes. By dominant we mean that only one of the pair of genes that we inherit from our father and mother need be affected to trigger the process of uncontrolled cell growth; the mutant gene function overrides that of the normal gene. Because tumor suppressor genes are negative regulators of cell proliferation, both genes must be damaged to cause a release of the critical brake they impose on tumor growth. One analogy for describing the effects of these two types of mutations is an automobile with two sets of controls, the kind used to teach new drivers. There are two ways in which such an automobile might speed out of control. One would be if one of the gas pedals got stuck. If this happened, it would not matter that the second pedal (or normal gene) functioned appropriately; the car would still be out of control. The gas pedal problem would be analogous to an oncogene mutation. However, if the car were out of control because one of the brake pedals did not function, it could still be stopped with the second brake pedal. Thus, as with

Page 19

the tumor suppressor genes, for all braking function to be lost, both brake pedals must fail. Both types of mutations, but more frequently mutations in one copy of a tumor suppressor gene, can occur in our germ cells and be passed on to our descendants. These lead to inborn, heritable predispositions to cancer. One brake is already gone and a second hit on the other will be disastrous. Although such inherited cancers are relatively rare, study of their genetic components has provided great insight into the etiology of the more common sporadic, or acquired, cancers because the very same genetic players are ultimately involved.

At first, the list in Figure 2-1 seems daunting and helps to explain why the war on cancer has been waging so long. We do know that at least in some cases, when a single, normal tumor suppressor gene is provided to a cancer cell with numerous mutations, proliferation and invasive properties can be inhibited dramatically. Therefore, the multiple players may, in fact, provide multiple targets for intervention. In any case, information of the sort provided by studies such as these has already proved useful for diagnosis, for determining prognosis, and in managing the care of cancer patients.

Some of these mutations, for example, those involving gene p53, are common to many malignancies. They are currently the focus of intense study. The HNPCC defect is estimated to occur in one of every 200 individuals in the population (Lynch et al., 1993). Affected individuals can develop tumors of the colon, uterus, ovary, and other organs. The responsible gene (hMSH2), whose identity and function were discovered recently, is required to repair copying mistakes in DNA replication (Fishel et al., 1993; Kinzler et al., 1993; Leach et al., 1993; Parsons et al., 1993). In its normal role this gene can be thought of as a text editor. The loss of its function shows up as an accumulation of mistakes (mutations at many sites in the genome), some of which are oncogenic. Efforts are already underway to develop screening tests for this mutation to use for members of families in which colon cancer is prevalent. Presumably individuals found to carry a mutant form of this gene would then undergo routine periodic testing so that malignancies could be detected at the earliest stage, when treatment would be most effective.

What about treatment? How can modern cell and molecular biology make a contribution in this area? On the basis of types of action deduced for the genes involved in cancer, two types of treatments may be envisioned. One, a seemingly rational approach for compensating for the loss of tumor suppressor gene function, would be to replace the gene. The other, for oncogenes where mutations trigger aberrant function, would be to counter malfunction. Developing such strategies will require a comprehensive understanding of the mechanism of action and the roles of the normal counterparts of these genes in the molecular biology of the cell.

Page 20

There are some promising starts, for example, the progress in our understanding of the function of the protein that is encoded by ras; ras is one of the most commonly detected oncogenes in human tumors and has been studied by many workers for more than a decade. Figure 2-2 shows a simplified illustration of the type of signaling pathways maintained by the cells of higher organisms that allow them to judge the status of their neighbors, to ensure proper attachments to tissue contact points, and to respond to the presence of growth factors, hormones, and other external signals. The pathway consists of chains of communicating proteins that receive signals from upstream (starting at the surface of the cell) and pass them to downstream targets known as effectors. The process is complex and includes chemical changes that are catalyzed after direct contact of the protein members. Furthermore, most (or all) of the proteins in the pathway can interact with multiple effectors. The Ras protein was long known to be a key player in these pathways. Despite this knowledge it is only recently that all of the steps could be put together in a fully connected framework (Egan and Weinberg, 1993).

The Ras protein (coded by ras gene) is a universal signal receptor, receiving messages initiated at a number of different cell surface receptors after they are stimulated by specific ligands, such as growth factors.

FIGURE 2-2 The ras signaling pathway.

Page 21

The protein comes in two forms; one is silent (no signal is passed) and the other is activated for signal transduction. The protein produced by the mutant ras oncogene is activated inappropriately, without having received a signal from an upstream receptor. The Ras protein is anchored in the inner membrane of the cell, ready to make contact with various receptor complexes and to relay the message of this interaction through a cascade of phosphorylating enzymes (kinases) to the nucleus.

Many of the members of the pathway are proteins encoded by previously identified oncogenes (e.g., raf, myc, jun), illustrating that defects anywhere along the pathway can have the same ultimate effect. The final targets of this pathway are nuclear transcription factors that respond by turning on the expression of specific genes. In the case of growth factors, genes required for proliferation are turned on. They then start up the cell cycle clock, which leads to DNA replication (the S phase), mitosis (the M phase), and division into two new cells. The proteins encoded by the tumor suppressor genes such as RB and p53 also act in the nucleus, but their roles are to put brakes on the cell cycle by preventing DNA synthesis, allowing repair of DNA damage before cell division, or, when repair cannot be effected, programming the cell to self-destruct.

To receive messages from cell surface receptors, ras must be anchored at the inner surface of the cell membrane (Figure 2-3). This anchoring depends on the use of a small 15-carbon molecule called farnesyl, which also has another role in the cell as a precursor in the pathway of cholesterol biosynthesis. Mutational analyses and studies with inhibitors of cholesterol biosynthesis have shown that Ras protein molecules lacking farnesyl do not attach to the cell membrane and cannot participate in signal transduction. The enzyme that performs this attachment, called farnesyltransferase, recognizes and binds the precursor, farnesyl-diphosphate, as well as a short amino acid sequence at the very end of the Ras protein (Figure 2-4). The enzyme then attaches the farnesyl moiety to the Ras protein. This reaction and farnesyltransferase are the target of current approaches to the design of inhibitors of Ras.

The approaches are based on the synthesis of molecules that look like either of the two substrates and will bind to the active site of the enzyme, thus interfering with attachment (Gibbs, 1991). Recent results from several laboratories look very encouraging (Kohl et al., 1993; James et al., 1993). These new inhibitors appear to interfere dramatically with the growth of tissue culture cells that express the activated form of Ras and proliferate in multilayered clumps that mimic cancerous growth. Most encouraging is the relative lack of toxicity of these inhibitors for normal cells or cells that are cancerous because of the expression of some other oncogene. Although much more needs to be done before such drugs can be used to treat people, it now seems increasing possible that inhibitors of

Page 22

FIGURE 2-3 Ras anchoring at the inner surface of the cell membrane. Source: Touchette (1990).

FIGURE 2-4 Action of farnesyltransferase. Source: Tamanoi (1993).

Page 23

this kind, designed through knowledge of molecular aspects of critical intracellular reactions, will be among the first to emerge as new clinical candidates for anticancer therapy.

Gene therapy is another experimental approach to cancer. As can be seen from Table 2-2, together with the indicated heritable genetic disorders, cancers are a major target for gene therapy protocols in the United States. These treatments attempt to manipulate oncogene or tumor suppressor gene expression in cancer cells to ameliorate or reverse the process. Another strategy for gene therapy is to transfer genes that will stimulate immunity against cancer cells. Many of these approaches to cancer treatment also look promising, although initial attempts have shown that there is still much to be learned about gene delivery systems and the immune system before gene therapy can be considered a straightforward exercise (Miller, 1992).

Page 24

Acquired Immune Deficiency Syndrome

Nowhere has our lack of sufficient fundamental knowledge about the body's immune defenses been more apparent than in our confrontation with acquired immune deficiency syndrome (AIDS; caused by the human immunodeficiency virus, HIV), which is another major target for gene therapy. HIV is particularly pernicious because it attacks the cells (called T cells) needed to mount an effective defense against infection. It enters these cells by attaching to a cell surface receptor (called CD4) that is a key component in the cell-to-cell communication that must occur during an immune response. This inappropriate interaction and others that occur after viral entry trigger responses in these cells that adversely affect the immune system both directly and indirectly. Some indirect effects appear to be responsible for damage to the central nervous system, giving rise to AIDS-related dementia. Efforts to develop vaccines have also been stymied because it still not clear what features of an immune response actually confer even transient protection against AIDS. Presumably, the immune system responds to components in the outer coat of the virus. However, HIV can rapidly mutate its genome during replication, allowing it to alter its outer coat. These two properties, direct attack on the immune system and variation in the surface it presents, eventually allow the virus to escape immune recognition entirely. Such problems have further confounded efforts to develop vaccines against AIDS.

The HIV life cycle is quite complex (Figure 2-5). Furthermore, gaps in our understanding of the host cell machinery that the virus takes over and of the nature of the control that the virus exerts on its own gene expression have made it difficult to intervene in the intermediate stages in the viral life cycle. Here and in other aspects of viral pathogenesis, the need for more fundamental knowledge is clear. Although not yet entirely satisfactory, the only approach that has yielded drugs that have been approved in the United States for the treatment of AIDS is targeted against the viral-specific enzymes encoded in the virus genome and essential at very early and very late replication stages (Katz and Skalka, 1994).

The targeted enzymes include the reverse transcriptase that the virus brings with it into the cell that it infects (and uses to convert its genetic information into a chemical form that is compatible with that of the host) and the protease that is required for processing and activating the virus' structural proteins and enzymes. The currently used reverse-transcriptase inhibitors (e.g., AZT, ddI, DDC, and d4T) are all analogs of the building blocks of DNA that interfere with viral DNA synthesis. Design of inhibitors for this enzyme and for the protease (discussed in more detail in this symposium by Dr. Wlodawer) have been aided significantly by the availability of detailed structural information concerning these proteins and

Page 25

FIGURE 2-5 The life cycle of HIV.

their mechanisms of action. Although the challenge of turning structural insights into effective treatments remains formidable, these enzymes offer some of the best models for modern, rational drug design. Studies of the third enzyme, the integrase that the virus uses to splice its DNA into that of the host cell, are not as far along. This enzyme functions at a key step in the viral life cycle, one that allows the infection to persist for the life of the infected cell and all of its daughter cells. Because of its importance, our laboratory and others are striving to obtain more detailed

Page 26

mechanistic and structural information. We have devised simple test tube assays for integrase that can be used to screen for and test inhibitors.

Unfortunately, HIV has already shown that it can mutate rapidly enough to eventually escape treatment with one anti-reverse transcriptase drug. The hope of current efforts is that attack at multiple targets will produce additive or synergistic effects that will eventually halt virus spread, prolong lives, and prevent establishment of infection in people newly exposed to the virus.

Summary

Modern advances in cellular and molecular biology have expanded opportunities and delineated new challenges in biomedical research. However, despite the best of efforts, this glimpse into the future can only be considered limited. Two additional variables will surely be part of the picture but, unfortunately, are impossible to assess.

For predicting scientific advances the missing variable is serendipity. We know that research in one area often produces conceptual breakthroughs and important insights in another, sometimes seemingly unrelated area. Thus, no matter how comprehensive our discussion is today, it is likely—even probable—that we are in for some big surprises from unsuspected directions in the future. If past experience holds true, some of these breakthroughs will have a big practical payoff. Thus, despite the current, public demands for more and more targeted research, we must continue to interpret our biomedical research mandates as broadly as possible. As the history of research into cell and molecular biology illustrates, research designed simply to answer general problems posed by nature often yields the most far-reaching rewards.

For predicting technological advances, the missing variable is improvement in methodologies. Perhaps the most striking example is in the computer industry. The computer currently in use probably was obsolete the day it appeared on the market. Technology, fueled by new knowledge, always seems to get better. With such improvement comes an ever-increasing capacity to explore the myriad of fundamental questions that remain in both cell and molecular biology. With these final thoughts in mind we can look forward to the next new era of cell and molecular biology.

Acknowledgments

I am grateful to Dr. Allen Oliff of Merck Research Laboratories for discussions and ideas concerning anticancer treatments based on oncogene protein inhibitors; to Dr. Ken Tartof of Fox Chase Cancer Center for providing the table summarizing gene therapy as well as helpful ideas

Page 27

concerning the usefulness of this approach; to my colleague Dr. Richard Katz for constructive suggestions concerning the manuscript; and to Marie Estes for her excellent secretarial assistance. Work in my laboratory is supported by National Institutes of Health grants CA-47486, CA-06927, and RR-05539; a grant from the Pew Charitable Trust; a grant for infectious disease research from Bristol-Myers Squibb Foundation; and an appropriation from the Commonwealth of Pennsylvania. The contents of this manuscript are solely the responsibility of the author and do not necessarily represent the official views of the National Cancer Institute or any other sponsoring organization.

References

Egan, S. E., and R. A. Weinberg. 1993. The pathway to signal achievement. Nature 365:781-783.

Fishel, R., M. K. Lescoe, M. R. S. Rao, N. G. Copeland, N. A. Jenkins, J. Garber, M. Kane, and R. Kolodner. 1993. The human mutator gene hom*olog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75:1027-1038.

Gibbs, J. B. 1991. R∼ C-terminal processing enzymes—new drug targets? Cell 65:1-4.

James, G. L., J. L. Goldstein, M. S. Brown, T. E. Rawson, T. C. Somers, R. S. McDowell, C. W. Crowley, B. K. Lucas, A. D. Levinson, and J. C. Marsters, Jr. 1993. Benzodiazepine peptidomimetics: Potent inhibitors of ras farnesylation in animal cells. Science 260:1937-1942.

Katz, R. A., and A. M. Skalka. 1994, The retroviral enzymes. Annu. Rev. Biochem. 63:133-173.

Kinzler, B. Vogelstein, and P. Modrich. 1993. Hypermutability and mismatch repair deficiency in RER+ tumor cells. Cell 75:1227-1236.

Kohl, N. E., S. D. Mosser, S. J. deSolms, E. A. Giuliani, D. L. Pompliano, S. L. Graham, R. L. Smith, E. M. Scolnick, A. Oliff, and J. B. Gibbs. 1993. Selective inhibition of rr-dependent transformation by a farnesyltransferase inhibitor. Science 260:1934-1937.

Leach, F. S., N. C. Nicolaides, N. Papadopoulos, B. Liu, J. Jen, R. Parsons, P. Peltomaki, P. Sistonen, L. A. Aaltonen, M. Nystrom-Lahti, K. Y. Guan, J. Zhang, P. S. Meltzer, J.-W. Yu, F.-T. Kao, D. J. Chen, K. M. Cerosaletti, R. E. K. Fournier, S. Todd, T. Lewis, R. J. Leach, S. L. Naylor, J. Weissenbach, J.-P. Mecklin, H. Jarvinen, G. M. Petersen, S. R. Hamilton, J. Green, J. Jass, P. Watson, H. T. Lynch, J. M. Trent, A. de la Chapelle, K. W. Kinzler, and B. Vogelstein. 1993. Mutations of a mutS hom*olog in hereditary nonpolyposis colorectal cancer. Cell 75:1215-1225.

Lewin, B., ed. 1974. Structure of the chromosome in gene expression. Pp. 1-47 in Vol 2. Eucaryotic Chromosomes. New York: John Wiley & Sons.

Lynch, P. M., R. J. Cavalieri, and C. R. Boland. 1993. Genetics, natural history, tumor spectrum, and pathology of hereditary nonpolyposis colorectal cancer: an updated review. Gastroenterology 104:1535-1549.

Marx, J. 1993. New colon cancer gene discovered. Science 260:751-752.

Miller, A. D. 1992. Human gene therapy comes of age. Nature 357:455-460.

Parsons, R., G. M. Li, M. J. Longely, W.-H. Fang, N. Papadopoulos, J. Jen, A. de la Chapelle, K. W. Kinzler, B. Vogelstein, and P. Modrich. 1993. Hypermutability and mismatch repair defiicency in RER+ tumor cells. Cell 75:1227-1236.

Tamanoi, F. 1993. Inhibitors of ras farnesyltransferases. Trends Biochem. Sci. 18:349-353.

Touchette, N. 1990. Cholesterol and cancer studies get a rise out of yeast. J. NIH Res. 2:61.

Vogelstein, B., and K. W. Kinzler. 1993. The multistep nature of cancer. Trends Genet. 9:138-141.

Page 28

3—
Structural Biology as
It Applies to Biotechnology

ALEXANDER WLODAWER

Structural biology is a large field that has a lot to contribute to biotechnology. I am going to discuss two examples of how information gained by structural biology techniques can be used in structure-based drug design to illustrate the potential of the field. One example has already led to drugs in clinical trials; the other has not led to any drugs as yet and shows how difficult it is to design drugs on the basis of structural data.

Many scientists still say that there is not a truly designed drug on the market and that every drug results from serendipity and medicinal chemistry, with design taking a secondary role. It is necessary, however, to understand what is meant by ''design." For example, design could be interpreted as designing a compound to inhibit a particular protein so that disease X will be cured. Imagine that a scientist uses his computer to create such a compound. It inhibits this protein and the disease is gone. If this is how drug design is interpreted, then of course it is true that nothing has ever been designed that way. What is really meant by drug design can be illustrated by using proteins encoded by human immunodeficiency virus (HIV) as an example.

Design Of Anti-HIV Drugs

Retroviruses have a unique way of encoding their proteins as so-called polyproteins, in which a number of different proteins are synthesized as contiguous polypeptide chains (Figure 3-1). This property is very

Page 29

FIGURE 3-1 Schematic diagram of the polyproteins interacting with the
cell membrane during budding of human immunodeficiency virus (HIV).

Page 30

important in the retroviral life cycle, because polyproteins attach themselves to the cell membrane before the budding of viral particles begins; after budding, one of the enzymes encoded by the virus (protease) starts the cascade of cleaving these polyproteins into individual molecules. Exactly how the process starts is still not completely known, but there is no question that protease is responsible for the creation of infectious viral particles. If the enzyme is disabled either by mutation or inhibition, the viral particles that are formed may look normal but are not infectious.

The first structure that was determined was of that of the enzyme without an inhibitor or substrate in its active site (Figure 3-2). The molecule has a very large cavity, and this cavity was suspected of being the site for the catalytic reaction. This enzyme belongs to the family of aspartic proteases. It has two aspartic acids that are in contact, and even before the structure was determined it was speculated that the enzyme would resemble other aspartic proteases, some of which are quite important in humans. One such protease is human renin, which has been a target for drug design for almost two decades. So, a lot was known about how to make inhibitors for this class of proteins even before we knew that HIV would carry one of them.

This is a good time to reiterate the importance of basic research versus targeted research. We would not have been anywhere in the field of retrovirology, as applied to infection by HIV, if many years ago people had not been willing to study obscure chicken viruses. A lot of what we learned about retroviruses came from studies for which utility appeared at that time to be zero. This work was done long before anybody ever heard of acquired immune deficiency syndrome (AIDS), and it is an excellent example of why we should not try to concentrate on the disease of the day and only do targeted research.

Design of Enzyme Inhibitors

How are inhibitors for these types of enzymes designed? First, it is important to understand something about how inhibitors work. Knowing the actual three-dimensional structure may not be completely necessary, but something must be known about the enzymatic properties and function. HIV protease can cleave many different sequences, including all linkages between the proteins in the polyprotein. These sequences can give rise to ideas about designing inhibitors by replacing the peptide bond between the two residues being cleaved by some other chemical group with a structure similar to a peptide bond but resistant to cleavage by the enzyme (Figure 3-3). Again, the choice of all of these inserts is the result of many years of work on the design of inhibitors of human renin

Page 31

FIGURE 3-2 HIV protease with an inhibitor, PD134922.

Page 32

FIGURE 3-3 Peptide bond and its nonscissile replacements.

Page 33

and these data were all in place when scientists started working on HIV protease in the late 1980s.

Many inhibitors of HIV protease have been made, and a selection of them was studied in our laboratory. Some of these compounds are similar to the original substrates; for example, MVT-101 seen in Figure 3-4 uses only standard amino acids. It has the reduced peptide bond in the center and blocking groups at the ends, and that is it. This is a very simple-minded way of producing an inhibitor; other inhibitors contain groups that are not standard amino acids. Peptides usually do not make good drugs because they are susceptible to other proteases present in the organisms, and they do not transport well through membranes. Figure 3-4 is a list of inhibitors studied by crystallography in my laboratory. It is a random list of compounds, not a progression leading to drugs, but all of these structures teach us something about the properties of the target system and the properties of the interactions between the enzyme and the inhibitor.

At least 20 different laboratories have determined at least one structure of HIV protease (Table 3-1), but the number of structures that were actually published and released is rather limited. A very important point emerging from this table is the lack of correlation between the number of structures from a particular laboratory and the number of compounds in clinical trials.

The structure of HIV protease shown previously was that of the apoenzyme. Shown in Figure 3-5 is the structure of an inhibitor complex. What happened in this case was something unexpected. Most enzymes do not change their structure appreciably upon binding of the inhibitors, but some do. In this particular case, there was a large change in the structure: about one-quarter of the molecule showed movements of -carbon of more than 1 angstrom (0.1 nanometers).

Comparison of Inhibitors

The next step of analysis involves comparing the inhibitors (Figure 3-6, see color plate). Once a series of structures has been created, the structures can be superimposed on one frame and educated guesses about the importance of parts of the molecule can be made. It becomes apparent which parts present in a particular inhibitor are not essential. This is when you can start looking at a number of your different examples and start drawing some overall conclusions from them.

When do you have a number of structures sufficient for beginning analysis? When can you rely on computer analysis and not worry, for example, about crystals or collecting data? Figure 3-7 (see color plate) is an example of a crystal structure of an inhibitor that we determined in

Page 34

FIGURE 3-4 HIV protease inhibitors studied by crystallography in the author's laboratory.

Page 35

FIGURE 3-5 Native HIV protease.

Page 36

collaboration with our colleagues from Upjohn who developed the model. Even though many of the features of the model are accurate, examination of the actual hydrogen bonds and the very detailed interactions between the inhibitors and the enzyme shows that the model as a whole is not accurate. Once the structure was determined, the computational chemists were able to recalculate the model with the additional information.

Our predictions of the structure of compounds that will bind to HIV protease are becoming quite good. However, when all experimentally determined structures will be released into public domain, the wealth of experimental structures available will be completely unprecedented. This level of activity has never happened before and is unlikely to happen in the future. We have a chance to learn a lot about how an enzyme works, how the inhibitors work, and about the process of inhibition.

Learning from Design

Have we found anything really unexpected, without which the design of compounds would not be possible or not easy? The answer to this

Page 37

FIGURE 3-8 HIV1 PR—Inh (U85548e): water 301.

question is yes. What was really unexpected was one water molecule, usually called water-301, that was found in the active site of HIV protease and that is always present in the complex between these specific inhibitors and the protease (Figure 3-8). Analysis of the structure near this particular water molecule reveals that it makes very good hydrogen bonds between two nitrogen atoms belonging to the enzyme and two oxygen atoms belonging to the inhibitor. These interactions were observed in almost all the structures of known inhibitors.

Page 38

As soon as we saw that there was something very unusual about and very specific to this complex, we and others suggested that it could be useful to try to make compounds that would include an equivalent of this water in the inhibitor itself. That is, of course, easy to postulate but difficult to do. Lam et al. (1994) recently published the structure of such a compound. This inhibitor has a cyclic urea group, where an oxygen atom replaces this particular water. This is one of the first examples of a non-peptidic HIV protease inhibitor that was designed, constructed, and shown to be a very potent antiviral agent. Unfortunately, this compound turned out not to be a good drug. Even though the inhibitor was very highly bioavailable, it was sequestered by the liver, resulting in an unsuccessful clinical trial. Follow-up compounds are being worked on, and there is hope that this approach will ultimately succeed. The problem of whether inhibitors will eventually turn into drugs is very complex, and bioavailability and pharmaco*kinetics are not easily seen in the structures. The next step involves extensive medicinal chemistry, which is still being done to a large extent by trial and error. That is the story of the HIV protease, which is conceptually a very easy example of drug design, because designing enzyme inhibitors is comparatively simple.

Design Of Cytokine Agonists And Antagonists

Now let us go to something that is not so simple: making agonists and antagonists of cytokines. Understanding cell signaling is one of the most important areas in biology now, because signaling pathways in organisms are elaborate, complicated, and not yet completely understood. The steps in some of these pathways, however, are known.

Knowledge of cytokine structures is relatively recent (Table 3-2). Except for preliminary structures of porcine growth hormone and interleukin 2 (IL-2), only in the past few years have we started seeing this particular class of cytokines in atomic detail. All that we know about the interactions between helical cytokines and their receptors, or actually extracellular parts of the receptors, comes from one publication by de Vos et al. (1992). This publication describes the interactions of human growth hormone with the receptor (Figure 3-9, see color plate). It was an extremely important development, but the choice was unfortunate from the point of view of trying to generalize it for the other systems. It turns out that a single molecule of human growth hormone causes dimerization of two identical molecules of the receptor. Each of the receptors interacts with a completely different part of the cytokine and the receptors use exactly the same part to interact with two different targets. Unfortunately, this is the only known example of these types of interactions, and it has already led us astray trying to understand other cytokines.

Page 39

As the structures of cytokines started appearing, we and others noted that they share many similarities. Superposition of IL-2, IL-4, and GM-CSF shows that the core of the protein is very similar although some parts are quite different (Figure 3-10, see color plate). This gives rise to a question of whether the similarity between different cytokines is important and can be used in trying to design cytokine agonists and antagonists. Bazan (1990) showed that regardless of the low or sometimes undetectable similarity among the cytokines themselves, the receptors are very closely related and their extracellular domains have almost the same structure, which is that seen for human growth hormone receptor.

Once we had access to a number of the structures in addition to the amino acid sequences of these cytokines, it was possible to structurally align their sequences. There is a fair amount of similarity, at least in the conserved core, where the superposition is meaningful. In the absence of the structural data, this type of superposition is very difficult; the super-positions of these particular cytokines found in the literature were uniformly wrong.

The parts of the primary and tertiary structure that are most similar include two of the helixes, helix A and helix D (Figure 3-11). Other studies also showed that these particular regions of the cytokines seem to be directly involved in receptor binding, which is why we hope to be able, by modifying these particular areas, to make other molecules that can modify the activities of cytokines.

The interplay between different cytokines and their receptors can get very complicated. The laboratory of Warren Leonard created a picture showing something that became obvious only recently, which has thrown a monkey wrench into this whole field (Figure 3-12, see color plate) (Russell et al., 1993). It turns out that a number of different cytokines,

Page 40

FIGURE 3-11 Sequence comparison of IL-2, IL-4, and GM-CSF.
Helix A is given first and helix D is given last.

Page 41

certainly IL-2, IL-4, and IL-7 and possibly some others, share the gamma C (common gamma receptor) subunit (formerly called IL-2 gamma receptor). Some of these cytokines are involved in pathways that regulate the activity of other cytokines. For example, the IL-2 and IL-4 pathways are intertwined, and they also use some of the identical components. Influencing the levels of the common gamma subunit in hope of modifying the activity of IL-4 may also modify the activities of IL-2 and IL-7 in such a way that the result may be the opposite of what was intended.

It is obvious that this is a very difficult field. How do we go about even postulating how to make drugs that could modify the activity of these types of molecules? First, by just looking at mutants of the naturally occurring proteins and identifying which parts of the proteins are important. A particular mutation of IL-4, for example, converts tyrosine 124 into aspartic acid, producing a molecule that is a full antagonist of IL-4 (Kruse et al., 1992).

To explain the action of antagonists, we have to understand how these signaling molecules work. The accepted mechanism is that the signaling event must involve dimerization of the extracellular domains of the receptors, leading to dimerization of the respective intracellular domains and then to a very complicated cascade of events. The dimerization of the extracellular domain is crucial and can sometimes occur without cytokines (e.g., antibodies can sometimes dimerize the receptors). Molecules that bind to the receptors are antagonists. Binding to the first molecule of the receptors but preventing the binding of the second molecule would prevent dimerization. In the case of tyrosine-124 mutant of IL-4, the specific receptor activity is not affected at all by this mutation but the biological activity is affected. We expect that this particular residue must interact with the common gamma subunit. A similar situation was reported for IL-2: a mutation of glutamine 126 into aspartic acid (Imler and Zurawski, 1992). Clearly this part of the structure is crucial to proper signaling by both cytokines.

Experimental structures of these complexes are not available, and we had to resort to modeling (Figure 3-13, see color plate). Unfortunately, it turns out that tyrosine 124 seems to be pointing halfway between the two receptor molecules, so trying to decide on which side to put the gamma receptor and on which side to put the IL-4 receptor is not obvious. However, some indirect interactions may be taking place. The residue arginine 121 corresponds to the glutamine in IL-2, which seems to be have properties similar to those of the mutant IL-4. So, mutation may be influencing the binding indirectly rather than directly.

Regardless of the interaction, modification of this area of the molecule may produce potential leads for drugs. However, drugs that are proteins are not very good because they cannot be given orally and are suitable

Page 42

only for conditions that are severe enough to warrant drugs given by injection. Of course, the next step would be to try to design small molecules that would correspond to the epitopes that are involved in receptor binding, thus antagonizing cytokines. We are still quite a long way from being able to go from these structures of proteins to the structure of small molecules that would replace those proteins.

In the long run, the approach described for cytokines may turn out to be important for the pharmacology of the future. We do not know today how to accomplish the goals, but sometimes posing questions is very important, even if there are no easy answers.

References

Bazan, J. F. 1990. Structural design and molecular evolution of a cytokine receptor superfamily. Proc. Natl. Acad. Sci. USA 87:6934-6938.

de Vos, A. M., M. Ultsch, and A. A. Kossiakoff. 1992. Human growth hormone and extracellular domain of its receptor: crystal structure of the complex. Science 255:306-312.

Imler, J.-L., and G. Zurawski. 1992. Receptor binding and internalization of mouse interleukin-2 derivatives that are partial agonists. J. Biol. Chem. 267:13185-13190.

Kruse, N., H.-P. Tony, and W. Swbold. 1992. Conversion of human interleukin-4 into a high affinity antagonist by a single amino acid repacement. EMBO J. 11:3237-3244.

Lam, P. Y. S., P. K. Jadhav, C. J. Eyermann, C. N. Hodge, Y. Ru, L. T. Bacheler, J. L., Meek, M. J. Otto, M. M. Rayner, Y. N. Wong, C.-H. Chang, P. C. Weber, D. A. Jackson, T. R. Sharpe, and S. Erickson-Viitanen. 1994. Rational design of potent, bioavailable, non-petide cyclic ureas as HIV protease inhibitors. Science 263:380-384.

Russell, S. M., A. D. Keegan, N. Harada, Y. Nakamura, M. Noguchi, P. Leland, M. C. Friedmann, A. Miyajima, R. K. Pure, W. E. Paul, and W. J. Leonard. 1993. Interleukin-2 receptor chain: a functional component of the interleukin-4 receptor. Science 262:1880-1883.

Page 43

4—
The Future of Biotechnology

C. THOMAS CASKEY

The technology that has evolved since the discovery of the structure of the DNA molecule will drive many of the opportunities that exist in medicine, society, and industry. We must be careful in evaluating the technology and the information that comes from it so that we make wise societal decisions.

Highlights Of Recent Discoveries

From the time line of the discoveries in this field in the past 30 or more years (Figure 4-1, Table 4-1), we can see that advances were made rapidly. Discovery and use of bacterial restriction enzymes (Kelly and Smith, 1970; Smith and Wilcox, 1970) and the development of plasmid technology (Cohen et al., 1973) meant that complex molecules of DNA could be dissected into small elements and analyzed individually.

Ed Southern (1975) gave us the first technique for diagnosis via a molecular method. This technique was DNA-based and could detect specific regions of DNA even when diluted 108-fold. We had not previously seen such sensitive technology in medicine. Two other marvelous techniques in the 1970s were the independent discoveries of methods for sequencing DNA (Maxam and Gilbert, 1977; Sanger et al., 1977). These methods enabled us to look at base sequences in DNA. Then in the 1980s, the polymerase chain reaction (PCR) technology came out of the biotech industry, from the Cetus Corporation (Mullis et al., 1986; Saiki et al., 1985). PCR enabled us to dissect the human genome efficiently by working

Page 44

FIGURE 4-1 Key advances in biotechnology. See Table 4-1 for details.
RFLP, restriction fragment length polymorphism; PFGE, purified field
gel electrophoresis; PCR, polymerase chain reaction; ES, embryonic
stem; YAC, yeast artificial chromosomes; FISH, fluorescence in situ
hybridization; GDB, Genome Data Base; EST, expressed sequence tag.

with large cloned molecules. We could now clone elements of DNA up to 1 million base pairs long, which was much longer than the few thousand base pairs contained within a plasmid (Burke et al., 1987). Another highlight was the remarkable integration of technology using PCR and common sequences within the human genome to enable us to lift sequences easily and rapidly out of a very complex molecule, such as the yeast artificial chromosome (Olson et al., 1989).

The introduction of genome and sequence databases occurred in the

Page 45

Page 46

1980s. An investigator anywhere could sit down at a computer and have access to data being generated worldwide. Then came the development of genetic and physical maps of the human genome. In 1993 came the announcement by the Généthon group in Paris of the development of a genetic linkage map containing more than 2,000 markers (Cohen et al., 1993). This map essentially fulfilled one of the short-term mapping goals of the Human Genome Project 1 year before the target date (U.S. Department of Health and Human Services and Department of Energy, 1990).

The Human Genome Project

Scientists working on the Human Genome Project have developed a set of markers along each chromosome that enable them to quickly and easily map regions of the genome and then actually isolate the material. The linear order of the material is then determined so that the chromosome can be reconstructed.

A new development in the Human Genome Project is the ability to seek disease genes. There are an estimated 80,000 genes in the human genome, of which about 2,000 have been isolated (McKusick et al., 1994). The ultimate goal of the Human Genome Project, determining the base sequence of the human genome, is currently beyond our available technology. However, identifying and characterizing all human genes (approximately 3 percent of the total sequence) is fully within our capacity.

We are adding entries to McKusick et al.'s (1994) Mendelian Inheritance of Man at an increasing pace, but this is still slow compared with what will be seen in the next few years. About 5 percent of the estimated 80,000 genes have been mapped, and at least 770 cloned and mapped genes have been associated with clinical abnormalities. Disease genes are found by looking for abnormal genes in affected patients. This process, however, will change with the increased discovery of genes. In the future, newly isolated genes will be studied and then patients will be found who fit their characteristics.

Trinucleotide Repeat Diseases

The discovery of a gene can lead to insights into a previously unknown system. The fragile X syndrome is so called because under certain culture conditions affected X chromosomes break at the q27.3 region, near the end of the long arm. Fragile X syndrome is the most common form of inherited mental retardation worldwide. From observing chromosomal breakage in culture and then tracking genome markers through affected families, we knew that the gene or genes responsible for this disease resided in Xq27.3. Classical positional cloning technology, used to identify

Page 47

FIGURE 4-2 Expansion of unstable repeat regions resulting in disease.
Some disease genes contain a region of repeated triplet nucleotides
that is polymorphic (that is, the number of repeats is variable).
Reproduced with permission from Rossiter and Caskey (In Press).

the gene, revealed that a new form of mutation was found at the site of a variable trinucleotide repeat (Verkerk et al., 1991) (Figure 4-2). All of us have a few cytosine-cytosine-guanine trinucleotide repeats at this location, but children with the mental retardation syndrome have markedly expanded trinucleotide repeats (Fu et al., 1991). The expanded trinucleotide repeats are very unstable when transmitted from generation to generation. Never before had we seen such unstable DNA that rendered such a high susceptibility to disease.

A recent study analyzed individual molecules of DNA from the sperm of an asymptomatic male for a disease called myotonic dystrophy (Monckton et al., 1995). This disease is also caused by unstable trinucleotide repeats. The size of the repeat region in the man's blood cells was 75

Page 48

FIGURE 4-3 Myotonic dystrophy pedigree, showing trinucleotide repeat expansion.
The number of (GCT)n repeats in each allele is given for each individual in this five-
generation pedigree. The upper number is the normal allele and the lower number
is the expanded allele. This family illustrates the phenomenon of trinucleotide
expansion but also has one of the rare examples of a reduction in size of the repeat
region (from III-4 to IV-5). Reproduced with permission from Monckton and Caskey (1995).

units, which caused him no health difficulty. However, the size of the repeat region in his sperm ranged from 30 to 450 units. Larger repeat sizes tend to result in more severe disease (Redman et al., 1993).

The fragile X syndrome and myotonic dystrophy findings were new discoveries in human genetics. They constitute molecular understanding of what clinicians for years had called ''anticipation," the increase in severity and earlier onset of disease in later generations (Fu et al., 1991). Figure 4-3 shows an example of expansion of a trinucleotide repeat within a myotonic dystrophy family. As the repeat region expands from generation to generation, the disease appears in a more severe form. This family was unaware of their myotonic dystrophy until the birth of individual V-1. His cousin had died as a newborn after 10 days on a respirator, but her myotonic dystrophy was not recognized.

How many other trinucleotide repeat diseases are there? At the moment we know of spinal and bulbar muscular atrophy, two forms of fragile X mental retardation, myotonic dystrophy, Huntington disease,

Page 49

spinocerebellar ataxia type 1, dentatorubral-pallidoluysian atrophy, and Machado-Joseph disease (Willems, 1994). All these disease are neurological disorders, but whether this is significant is not known.

Diagnostic Procedures

Physicians and patients can benefit from the accuracy and precision in diagnosis provided by genetic diagnostic procedures. In the future this information may enable us to intervene by medication or lifestyle alterations to preempt the onset of a disease in a person at risk. Accurate diagnosis will identity couples at risk for bearing children affected by a genetic disease.

duch*enne muscular dystrophy is a common and severe childhood disease. The dystrophin gene, responsible for the disease, is very large and is particularly prone to deletion mutations (Clemens and Caskey, 1992), and most patients do not have the same deletion. Scanning the huge dystrophin gene may seem to be an impossibility but in fact is very simple. By using PCR technology, many regions of the gene can be analyzed simultaneously and more than 98 percent of deletions can be detected with a simple procedure called multiplex PCR (Chamberlain et al., 1990) (Figure 4-4). Multiplex PCR analysis has proved to be readily transferable to clinical laboratories worldwide, a particular benefit for laboratories that could not afford the more complex diagnostic procedures used previously (Group study, 1992).

FIGURE 4-4 Scanning for deletions in the duch*enne muscular dystrophy gene.
Nine fragments from the dystrophin gene (lanes a,c,f,i) are generated by
simultaneous polymerase chain reactions and separated by size. The absence
of one or more bands (lanes b,d,e,g,h) indicates that there is a deletion of that
portion of the dystrophin gene. This method is used for diagnosing duch*enne
muscular dystrophy at Baylor College of Medicine and many other institutions.
Reproduced with permission from [study group] (1992).

Page 50

A second example of improved diagnosis followed the discovery of the cystic fibrosis gene, responsible for the most common recessively inherited disease in Caucasians (Beaudet, 1992). The diagnostic problem with cystic fibrosis is that many (more than 300) disease-causing mutations have been identified throughout the gene. Just as duch*enne muscular dystrophy required a scanning method of deletion detection, cystic fibrosis requires a method of scanning for small (mostly single base pair) alterations. We are doing this by using a robot to analyze 22 different positions in the gene. In a single assay, one technician can run up to 90 samples. Thus, molecular diagnosis for cystic fibrosis has rapidly moved from research to general use at a reasonable cost. Families can find out if they have a risk for this disease before bearing a child with cystic fibrosis. Or, if they have a child with cystic fibrosis, they can find out who else in the family is at risk and what their future reproductive options are.

The Human Genome Project is a very ambitious project that has the capacity to identify significant disease risks in all of us. Every individual probably carries 5 to 10 significant genetic mutations with the potential to cause severe disease at various stages of life. In several laboratories a "chip" technology is being developed that will scan large portions of the genome for mutations (Fodor et al., 1993). Attached to a chip would be DNA sequences that correspond to the sites of mutations for common diseases. A person's DNA could then be hybridized to that chip and the resulting signal would indicate which mutations were found. It is possible to place 100 or more DNA molecules on a single 1-centimeter chip and to use a detection device to read those signals. It is going to be possible with this or a similar technology to look at targeted regions of the genome and diagnose a disease or determine susceptibility for bearing an affected child.

Possibilities For Disease Prevention

How these diagnostic advances can be used to prevent disease will affect how we view health care and what we will allow to take place in our society. For example, let us consider alpha-1-antitrypsin deficiency, type II hyperlipidemia, Huntington disease, and certain types of cancer. These are all diseases in which it is possible to run a DNA-based analysis and determine if a person has the DNA alteration that is going to lead to the disease. Each of these diseases has a different time of onset in life, but the diagnostic test can be done at any age.

Alpha-1-antitrypsin deficiency is a common cause of adult-onset emphysema (Blank and Brantly, 1994). If you knew from birth that you had this deficiency and if you abstained from smoking, you would prolong your life by 10 years (Larsson, 1978). This is an example of a lifestyle

Page 51

choice improving the outcome of a condition that is determined genetically.

Coronary artery disease is a major killer in the United States and one that has received a lot of attention, but most of the focus has been on treatment after the disease develops. However, detection of a genetic predisposition to hypercholesterolemia permits intervention with diet and medication before the damage is bad enough to cause symptoms (Bild et al., 1993).

Not all presymptomatic diagnosis has the option of treatment. Huntington disease, a devastating neurological disorder, can be diagnosed very precisely by using DNA methods (Huntington's Disease Collaborative Research Group, 1993). However, at present there is no treatment or cure. Because of the dominant inheritance of this gene, any child with an affected parent has a 50 percent chance of possessing the same defective gene. For Huntington disease, presymptomatic testing might help in eliminating uncertainty and planning the future, but there is no option for improving health. This will also be true for other diseases until we have a greater understanding of mechanisms of disease and have developed better therapeutic measures.

Numerous genes throughout the genome are responsible for predisposing an individual to developing cancer (Figure 4-5). Recently discovered genes include those that contribute to breast and colon cancer. Offspring of parents with cancer live in uncertainty, not knowing whether or not that predisposition has been inherited. In the future there will be less uncertainty. However, this information must be handled carefully. Are you prepared to learn whether you are at high risk for developing cancer? What effect would that have on your life? What would your doctors do? These are issues to which we will be increasingly exposed as presymptomatic diagnosis becomes more accurate and commonly used.

Privacy Issues

What is the potential down side to this prior information? There may be loss of privacy resulting from increased knowledge about your future medical condition. You and your physician will now share some information that was previously unknown. How many others would you want to share that information with? Your spouse? Your children? Your siblings? These would all seem rather logical to me. Perhaps you have an employment health care plan that requires you to report any health difficulties. Is it possible that information might begin to slip from that ultimate privacy between you and physician? I would say yes.

Although many people would argue strongly that we have to protect the privacy of an individual's genetic information, I would argue a slightly

Page 52

FIGURE 4-5 Genomic location of some genes involved in human cancer. ABL1, Abelson murine leukemia viral (v-abl) oncogene hom*olog 1; APC, adenomatosis polyposis coli; BCR, breakpoint cluster region (chronic myeloid leukemia); BRCA1, breast cancer 1, early onset; CMM, cutaneous malignant melanoma/dysplastic nevus; COCA1, colon cancer 1; HRAS, Harvey rat sarcoma viral (v-Ha-ras) oncogene hom*olog; KRAS2, Kirsten rat sarcoma 2 viral (v-Ki-ras2) oncogene hom*olog; MEN1, multiple endocrine neoplasia 1; MLM, melanoma; MYC, Avian myelocytomatosis viral (v-myc) oncogene hom*olog; NF1, neurofibromin 1 (neurofibromatosis, von Recklinghausen disease, Watson disease); NF2, neurofibromastosis 2 (bilateral acoustic neuroma); NRAS, neuroblastoma RAS viral (v-ras) oncogene hom*olog; RARA, retinoic acid receptor, alpha; RB1, retinoblastoma 1 (including osteosarcoma); RET, ret proto-oncogene (multiple endocrine neoplasia MEN2A and medullary thyroid carcinoma 1, and Hirschsprung disease); TP53, tumor protein p53 (Li-Fraumeni syndrome); VHL, von Hippel-Lindau syndrome; and WT1, Wilms tumor 1. Redrawn from Caskey and Smith (1994).

Page 53

different point. I am going to accept that this information might become widespread and argue that we should make sure that there is no ill effect from this acquired information (e.g., exclusion from insurance, employment, education, or other opportunities). An individual's rights should be protected. I think this would be a more logical way to proceed in what I think is going to be an increasingly open society regarding information. Many disagree with this and insist that such information be kept private, but I have worked in medicine and dealt with families for too long to think that privacy can be achieved. Therefore, the protection of the information against misuse is very important.

Everyone carries some genetic error; some have a few illnesses and others are spared sickness. The only way that we can support the total societal burden of illness is to have a uniform underwriting of the entire population. The healthy people have to pay for the ill health of the rare individuals in the population. It is extremely important that we be able to gain access to presymptomatic information without damage to the patient, so that we can implement new procedures for disease prevention in the future.

Drug Production

Genetic research has also affected the area of drug production. In the past, insulin could only be obtained from abattoir material, growth hormone was extracted from cadaver brains, and clotting factor VIII was prepared from pooled blood. Cellular growth factors were not available until the era of recombinant DNA. All of these now can be produced very effectively by industry by using recombinant DNA techniques.

Page 54

Despite early resistance to the use of genetic engineering techniques, these methods have proved lifesaving in many situations. Virtually all hemophilia A patients who were treated before 1985 with factor VIII made from blood now carry the human immunodeficiency virus or have died from acquired immune disease syndrome. Another common complication is hepatitis resulting from contaminating viruses. However, the ability to make factor VIII from an isolated gene has eliminated the use of blood products and the problem of viral contamination (Kaufman, 1989). Another example of a disease caused by contamination of therapeutic materials with infectious agents is Creutzfeldt-Jakob disease, a neuro-degenerative disease leading to dementia and death. Some people taking growth hormone made from the brains of cadavers have contracted this disease. Again, the growth factor is now manufactured from an isolated gene (Goeddel et al., 1979).

Many recombinant drugs are coming onto the market. It is important that these drugs undergo careful clinical evaluation to determine their utility and efficacy. They are powerful drugs with tremendous utility in certain settings, but it is very important to prove their clinical application before they move into the marketplace. We cannot automatically accept drugs just because they are recombinant drugs; they must earn their place in the medical armament.

Gene Therapy

Gene therapy is a technology that will have a major effect on the practice of medicine. In most cases the chosen delivery system is a viral vector that has been disabled. The essential functions are removed from the virus and replaced with DNA sequences that correspond to the human disease of interest. The engineered virus can then be safely produced in large quantities and used for delivering those genes (McCabe, 1993).

One disease that is a candidate for gene therapy is the severe combined immune deficiency resulting from deficiency in the enzyme adenosine deaminase. Bone marrow transplantation can be used to treat this disease but is limited by the availability of matched donors and the problems associated with immune-rejection reactions. At least two young girls have been successfully provided with functional adenosine deaminase genes, enabling them to lead relatively normal lives instead of being kept in complete isolation (Blaese et al. 1994).

In our laboratory we are developing gene therapy for a common liver disease called ornithine transcarbamylase deficiency. Mice with an equivalent of this disease have no hair and are therefore easily distinguished. These mice can be completely corrected by a single injection given in the newborn period of a virus carrying the ornithine transcarbamylase

Page 55

gene (M. A. Morsy, T. Ngo, A. W. Warman, W. E. O'Brien, F. L. Graham, and C. T. Caskey, unpublished data). This is an example of gene therapy administered once that results in a cure in mice. This form of treatment is permanent and blood borne and is probably going to be the future direction for gene therapy, although the vector may not necessarily be a virus.

Genetic Personal Identification

Personal identification based on genetic information is currently used in cases involving paternity, missing individuals, and forensic evidence. It is also used extensively by population geneticists. It is estimated that any two people will have differences in DNA structure in about 1 nucleotide per 300 to 1,000 (Cooper et al., 1985). Huge diversity exists in the population, and it is this variation in DNA structure that is used for personal identification.

Let me use forensic analysis in a rape case to illustrate the method (Gill et al., 1985). A DNA sample from the victim is obtained from blood and also from the secretions taken when the victim seeks medical attention. Male and female fractions from the vagin*l sample can be separated, resulting in distinctive DNA signals. Eyewitness accounting of violent crimes is notoriously inaccurate, but this DNA "fingerprinting" precisely demonstrates a match or mismatch between a suspect and biological material left at the scene of a crime. Several regions of the genome are scanned during the analysis by using different genetic probes. With each probe that displays a match between two samples, the statistical probability that they came from the same person is increased. If a mismatch is shown, the samples cannot have come from the same individual. This is a very powerful technique, particularly when used to prove innocence.

The use of DNA fingerprinting was fought extensively in the courts at the federal and local levels, but technology has now withstood the test of time through the court systems. It was reviewed carefully by the National Academy of Sciences, which issued guidelines on how this technology should be used in the courts (Committee on DNA Technology in Forensic Science, 1992). This technology has also been used in archeological digs to solve crimes of many years ago. The Romanov family assassinations that took place in Russia have now been clarified by this technology (Debenham, 1994; Gill et al., 1994).

DNA variation can also be used to investigate the migration of human populations in the past (Torroni et al., 1994). Such studies have taken advantage of the fact that there is sequence variation in mitochondrial DNA and that mitochondrial DNA is inherited entirely from the mother.

Page 56

DNA analysis will enable us to discover (and hopefully understand) our genes and identify those that cause disease. The capture of some genes will enable us to develop pharmaceutical agents that can benefit individuals with disease. The use of this technology will enable us to better understand our past and our diversity.

References

Adams, M. D., J. M. Kelley, J. D. Gocayne, M. Dubnick, M. H. Polymeropolous, H. Xiao, C. R. Merril, A. Wu, B. Olde, R. F. Moreno, A. R. Kerlavage, W. R. McCombie, and J. C. Venter. 1991. Complementary DNA sequencing: expressed sequence tags and Human Genome Project. Science 252:1651-1656.

Beaudet, A. L. 1992. Genetic testing for cystic fibrosis. Pediatr. Clin. North Am. 39:213-228.

Bild, D. E., R. R. Williams, H. B. Brewer, J. A. Herd, T. A. Pearson, and E. Stein. 1993. Identification and management of heterozygous familial hypercholesterolemia: summary and recommendations from an NHLBI workshop. Am. J. Cardiol. 72:1D-5D.

Blaese, R. M., K. Culver, W. F. Anderson, J. Ramsey, C. Carter, A. D. Miller, C. Mullen, D. Kohn, and R. Morgan. 1994. Gene therapy for adenosine deaminase (ADA) deficiency: a 4 year follow-up of the original trial. P.3 in Abstracts of papers presented at the 1994 meeting on gene therapy, September 21-September 25, 1994, T. Friedmann, Y. W. Kan, and R. Mulligan, eds. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Blank, C. A., and M. Brantly. 1994. Clinical features and molecular characteristics of alpha 1-antitrypsin deficiency. Ann. Allergy 72:105-120.

Breathnach, R., J. L. Mandel, and P. Chambon. 1977. Ovalbumin gene is split in chicken DNA. Nature 270:314-319.

Burke, D. T., G. F. Carle, and M. V. Olson. 1987. Cloning of large segments of exogenous DNA into yeast by means of artificial chromosome vectors. Science 236:806-812.

Caskey, C. T., and J. M. Smith. 1994. Sizing up the enemy. A call for papers on cancer [editorial]. JAMA 271:1448-1449.

Chamberlain, J. S., R. A. Gibbs, J. E. Rainer, and C. T. Caskey. 1990. Multiplex PCR for the diagnosis of duch*enne muscular dystrophy. Pp. 272-281 in PCR Protocols: A Guide to Methods and Applications, M. Innis, D. Gelfand, J. Sninsky, and T. White, eds. Orlando, FL: Academic Press.

Chumakov, I., P. Rigault, S. Guillou, P. Ougen, A. Billaut, G. Guasconi, P. Gervy, I. LeGall, P. Soularue, L. Grinas, L. Bougueleret, C. Bellanné-Chantelot, B. Lacroix, E. Barillot, P. Gesnouin, S. Pook, G. Vaysseix, G. Frelat, A. Schmitz, J. L. Sambucy, A. Bosch, X. Estivill, J. Weissenbach, A. Vignal, H. Riethman, D. Cox, D. Patterson, K. Gardiner, M. Hattori, Y. Sakai, H. Ichikawa, M. Ohki, D. le Paslier, R. Heilig, S. Antonarakis, and D. Cohen. 1992. Continuum of overlapping clones spanning the entire human chromosome 21q. Nature 359:380-387.

Clemens, P. R., and C. T. Caskey. 1992. duch*enne muscular dystrophy. Curr Neurol. 12:1-22.

Cohen, D., I. Chumakov, and J. Weissenbach. 1993. A first-generation physical map of the human genome. Nature 366:698-701.

Cohen, S. N., A. C. Y. Chang, H. W. Boyer, and R. B. Helling. 1973. Construction of biologically functional bacterial plasmids in vitro. Proc. Natl. Acad. Sci. USA 70:3240-3244.

Committee on DNA Technology in Forensic Science, Board on Biology, Commission on Life Science, and National Research Council. 1992. DNA Technology in Forensic Science. Washington, D.C.: National Academy Press.

Page 57

Cooper, D. N., B. A. Smith, H. J. Cook, S. Niemann, and J. Schmidtke. 1985. An estimate of unique DNA sequence heterozygosity in the human genome. Hum. Genet. 69:201-205.

Debenham, P. G. 1994. Genetics leaves no bones unturned. Nat. Genet. 6:113-114.

Edwards, A., H. Voss, P. Rice, A. Civitello, J. Stegemann, C. Schwager, J. Zimmermann, H. Erfle, C. T. Caskey, and W. Ansorge. 1990. Automated DNA sequencing of the human HPRT locus. Genomics 6:593-608.

Fodor, S. P. A., R. P. Rava, X. C. Huang, A. C. Pease, C. P. Holmes, and C. L. Adams. 1993. Multiplexed biochemical assays with biological chips. Nature 364:555-556.

Foote, S., D. Vollrath, A. Hilton, and D. C. Page. 1992. The human Y chromosome: overlapping DNA clones spanning the euchromatic region. Science 258:60-66.

Fu, Y. H., D. P. A. Kuhl, A. Pizzuti, M. Pieretti, J. S. Sutcliffe, S. Richards, A. J. M. H. Verkerkk, J. J. A. Holden, R. G. Fenwick, Jr., S. T. Warren, B. A. Oostra, D. L. Nelson, and C. T. Caskey. 1991. Variation of the CGG repeat at the fragile X site results in genetic instability: resolution of the Sherman paradox. Cell 67:1047-1048.

Gill P., A. J. Jeffreys, and D. J. Werrett. 1985. Forensic application of DNA 'fingerprints.' Nature 318:577-579.

Gill, P., P. L. Ivanov, C. Kimpton, R. Piercy, N. Benson, G. Tully, I. Evett, E. Hagelberg, and K. Sullivan. 1994. Identification of the remains of the Romanov family by DNA analysis. Nat. Genet. 6:130-135.

Goeddel, D. V., H. L. Heyneker, T. Hozumi, R. Arentzen, K. Itakura, D. G. Yansura, M. J. Ross, G. Miozzari, R. Crea, and P. H. Seeburg. 1979. Direct expression in Escherichia coli of a DNA sequence coding for human growth hormone. Nature 281:544-548.

Gordon, J. W., G. A. Scangos, D. J. Plotkin, J. A. Barbosa, and F. H. Ruddle. 1980. Genetic transformation of mouse embryos by microinjection of purified DNA. Proc. Natl. Acad. Sci. USA 77:7380-7384.

Group study. 1992. Diagnosis of duch*enne and Becker muscular dystrophies by polymerase chain reaction. A multicenter study. JAMA 267:2609-2615.

Huntington's Disease Collaborative Research Group. 1993. A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington's disease chromosomes. Cell 72:971-983.

Kan, Y. W., and A. M. Dozy. 1978. Polymorphism of DNA sequence adjacent to human b-globin structural gene: relationship to sickle mutation. Proc. Natl. Acad. Sci. USA 75:5631-5635.

Kaufman, R. J. 1989. Genetic engineering of factor VIII. Nature 342:207-208.

Kelly, T. J., Jr., and H. O. Smith. 1970. A restriction enzyme from Hemophilus influenzae. II. Base sequence of the recognition site. J. Mol. Biol. 51:393-409.

Larsson, C. 1978. Natural history and life expectancy in severe alpha1-antitrypsin deficiency, Pi Z. Acta Med. Scand. 204:345-351.

Lawrence, J. B., C. A. Villnave, and R. H. Singer. 1988. Sensitive, high resolution chromatin and chromosome mapping in situ: presence and orientation of two closely integrated copies of EBV in a lymphoma line. Cell 52:51-61.

Littlefield, J. W. 1964. Selection of hybrids from matings of fibroblasts in vitro and their presumed recombinants. Science 145:709-710.

Mandel, M., and A. Higa. 1970. Calcium-dependent bacteriophage DNA infection. J. Mol. Biol. 53:159-162.

Maxam, A. M., and W. Gilbert. 1977. A new method for sequencing DNA. Proc. Natl. Acad. Sci. USA 74:560-564.

McBride, O. W., and H. L. Ozer. 1973. Transfer of genetic information by purified metaphase chromosomes. Proc. Natl. Acad. Sci. USA 70:1258-1262.

McCabe, E. R. B. 1993. Clinical application of gene therapy: emerging opportunities and current limitations. Biochem. Med. Metab. Biol. 50:241-253.

Page 58

McKusick, V. A., C. A. Francomano, S. E. Antonarakis, and P. L. Pearson. 1994. Mendelian Inheritance in Man. A Catalog of Human Genes and Genetic Disorders, 11th edition. Baltimore, MD: Johns Hopkins University Press.

Meselson, M., and F. W. Stahl. 1958. The replication of DNA in Escherichia coli. Proc. Natl. Acad. Sci. USA 44:671-682.

Monckton, D. G., and C. T. Caskey. 1995. Unstable triplet repeat diseases. Circulation 91:513-520.

Monckton, D. G., L. J. C. Wong, T. Ashizawa, and C. T. Caskey. 1995. Somatic mosaicism, germline expansions, germline reversions and intergenerational reductions in myotonic dystrophy males: small pool PCR analyses. Hum. Mol. Genet. 4:1-8.

Mullis, K., F. Faloona, S. Scharf, R. Saiki., G. Horn, and H. Erlich. 1986. Specific enzymatic amplification of DNA in vitro: the polymerase chain reaction. Cold Spring Harb. Symp. Quant. Biol. 51:263-273.

Nelson, D. L., S. A. Ledbetter, L. Corbo, M. F. Victoria, R. Ramirez-Solis, T. D. Webster, D. H. Ledbetter, and C. T. Caskey. 1989. Alu polymerase chain reaction: a method for rapid isolation of human-specific sequences from complex DNA sources. Proc. Natl. Acad. Sci. USA 86:6686-6690.

NIH/CEPH Collaborative Mapping Group. 1992. A comprehensive genetic linkage map of the human genome. Science 258:67-86.

Nirenberg, M. W., and J. H. Matthaei. 1961. The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polyribonucleotides. Proc. Natl. Acad. Sci. USA 47:1588-1602.

Okazaki, R., T. Okazaki, K. Sakabe, K. Sugimoto, and A. Sugino. 1968. Mechanism of DNA chain growth. I. Possible discontinuity and unusual secondary structure of newly synthesized chains. Proc. Natl. Acad. Sci. USA 59:598-605.

Oliver, S. G., Q. J. M. van der Aart, M. L. Agostoni-Carbone, M. Aigle, L. Alberghina, D. Alexandraki, G. Antoine, R. Anwar, J. P. G. Ballesta, P. Benit, G. Berben, E. Bergantino, N. Biteau, P. A. Bolle, M. Bolotin-f*ckuhara, A. Brown, A. J. P. Brown, J. M. Buhler, C. Carcano, G. Carignani, H. Cederberg, R. Chanet, R. Contreras, M. Crouzet, B. Daignan-Fornier, E. Defoor, M. Delgado, J. Demolder, C. Doira, E. Dubois, B. Dujon, A. Dusterhoft, D. Erdmann, M. Esteban, F. Fabre, C. Fairhead, G. Faye, H. Feldmann, W. Fiers, M. C. Francingues-Gaillard, L. Franco, L. Frontali, H. f*ckuhara, L. J. Fuller, P. Galland, M. E. Gent, D. Gigot, V. Gilliquet, N. Glansdorff, A. Goffeau, M. Grenson, P. Grisanti, L. A. Grivell, M. de Haan, M. Haasemann, D. Hatat, J. Hoenicka, J. Hegemann, C. Herbert, F. Hilger, S. Hohmann, C. P. Hollenberg, K. Huse, F. Iborra, K. J. Indge, K. Isono, C. Jacq, M. Jacquet, C. M. James, J. C. Jauniaux, Y. Jia, A. Jimenez, A. Kelly, U. Kleinhans, P. Kreisi, G. Lanfranchi, C. Lewis, C. G. van der Linden, G. Lucchini, K. Lutzenkirchen, M. J. Maat, L. Mallet, G. Mannhaupt, E. Martegani, A. Mathieu, C. T. C. Maurer, D. McConnell, R. A. McKee, F. Messenguy, H. W. Mewes, F. Molemans, M. A. Montague, M. Muzi Falconi, L. Navas, C. S. Newlon, D. Noone, C. Pallier, L. Panzeri, B. M. Pearson, J. Perea, P. Philippsen, A. Pierard, R. J. Planta, P. Plevani, B. Poetsch, F. Pohl, B. Purnelle, M. Ramezani Rad, S. W. Rasmussen, A. Raynal, M. Remacha, P. Richterich, A. B. Roberts, F. Rodriguez, E. Sanz, I. Schaaff-Gerstenschlager, B. Scherens, B. Schweitzer, Y. Shu, J. Skala, P. P. Slonimski, F. Sor, C. Soustelle, R. Spiegelberg, L. I. Stateva, H. Y. Steensma, S. Steiner, A. Thierry, G. Thireos, M. Tzermia, L. A. Urrestauazu, B. Valle, I. R. Warmington, D. von Wettstein, B. L. Wixksteed, C. Wilson, H. Wurst, G. Xu, A. Yoshikawa, F. K. Zimmermann, and J. G. Sgouros. 1992. The complete DNA sequence of yeast chromosome III. Nature 357:38-46.

Olson, M., L. Hood, C. Cantor, and D. Botstein. 1989. A common language for physical mapping of the human genome. Science 245:1434-1435.

Page 59

Redman, J. B., R. G. Fenwick, Jr., Y. H. Fu, A. Pizzuti, and C. T. Caskey. 1993. Relationship between parental trinucleotide GCT repeat length and severity of myotonic dystrophy in offspring. JAMA 269:1960-1965.

Robertson, E., A. Bradley, M. Kuehn, and M. Evans. 1986. Germ-line transmission of genes introduced into cultured pluripotential cells by retroviral vector. Nature 323:445-448.

Rossiter, B. J. F., and C. T. Caskey. In Press. Molecular biology of human genetic predisposition to disease In Encyclopedia of Molecular Biology and Biotechnology, R. A. Meyers, ed. Weinheim: VCH Publishers, in press.

Saiki, R. K., S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Erlich, and N. Arnheim. 1985. Enzymatic amplification of b-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science. 230:1350-1354.

Sanger, F., and A. R. Coulson. 1975. A rapid method of determining sequences in DNA by primed synthesis with DNA polymerase. J. Mol. Biol. 94:441-448.

Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467.

Schwartz, D. C., and C. R. Cantor. 1984. Separation of yeast chromosome-sized DNAs by pulsed field gradient gel electrophoresis. Cell 37:67-75.

Smith, H. O., and K. W. Wilcox. 1970. A restriction enzyme from Hemophilus influenzae. I. Purification and general properties. J. Mol. Biol. 51:379-391.

Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517.

Torroni, A., J. V. Neel, R. Barrantes, T. G. Schurr, and D. C. Wallace. 1994. Mitochondrial DNA ''clock" for the Amerinds and its implications for timing their entry into North America. Proc. Natl. Acad. Sci. USA 91:1158-1162.

U.S. Department of Health and Human Services, and U.S. Department of Energy. 1990. Understanding Our Genetic Inheritance. The U.S. Human Genome Project: The First Five Years, FY 1991-1995. DOE/ER-0452P. Washington, D.C.: U.S. Government Printing Office.

Verkerk, A. J. M. H., M. Pieretti, J. S. Sutcliffe, Y. H. Fu, D. P. A. Kuhl, A. Pizutti, O. Reiner, S. Richards, M. F. Victoria, F. Zhang, B. E. Eussen, G. J. B. van Ommen, L. A. J. Blonden, G. J. Riggins, J. L. Chastain, C. Kunst, H. Galjaard, C. T. Caskey, D. L. Nelson, B. A. Oostra, and S. T. Warren. 1991. Identification of a gene (FMR-1) containing a CGG repeat coincident with a breakpoint cluster region exhibiting length variation in fragile X syndrome. Cell 65:905-914.

Vollrath, D., S. Foote, A. Hilton, L. G. Brown, P. Beer-Romero, J. S. Bogan, and D. C. Page. 1992. The human Y chromosome: a 43-interval map based on naturally occurring deletions. Science 258:52-59.

Watson, J. D., and F. H. C. Crick. 1953. Molecular structure of nucleic acids. Nature 171:737-738.

Willems, P. J. 1994. Dynamic mutations hit double figures. Nat. Genet. 8:213-215.