A REPRINT FROM JAMA The Journal of the American Medical Association Selected Articles on MOLECULAR MEDICINE Commemorating the 40th Anniversary of the Watson and Crick Papers on the DNA Double Helix American Medical Association 4#™ Physicians dedicated to the health of America Ay What the Double Helix (1953) Has Meant for Basic Biomedical Science A Personal Commentary Joshua Lederberg, PhD THE ARTICLE published by Watson and Crick in 1953' was the landmark pointer to our contemporary model of DNA as a macromolecular structure. This lay on a well-worn path of biophys- ica] analysis, reducing microscopic anat- omy to the molecular level. It also helped inspire an enormous body of biochemi- cal research that has defined DNA as the informational molecule, a disconti- nuity that has been labeled the Biolog- ical Revolution of the 20th Century. As a piece of structural analysis, the idea of the double helix includes the concepts () that DNA is a duplex structure, com- prising two paired complementary strands, associated by secondary, non- covalent bonds; (2) that the strand pairs are coiled, forming a double helix; and (3) that these are antiparallel—the ori- entation of one strand being in the op- posite polarity from the other. The most novel features of DNA are associated with its duplicity, rather than its helicity. Linear polymers rarely form stiff straight rods; folding into coils is the norm. The genetic functions of DNA are inextricably associated with its du- plex structure, and hardly at all with its helical shape; this is reflected in the pre- occupation of DNA research with its role as an informational molecule. How- ever, we shall see a recent concentra- tion of interest in supercoiling. Inevita- From The Rockefeller University, New York, NY. Dr Lederberg is a consultant for Affymax Technolo- gies, which uses DNA combinatorics. Reprint requests to The Rockefeller University, 1230 York Ave, New York, NY 10021-6399 (Dr Lederberg). JAMA, April 21, 1993—Vol 269, No. 15 bly, the biochemical interactions of DNA with other molecules, be they regulato- ry proteins or chemotherapeutic inhib- itors, will often be intimately wound up with the precise three-dimensional con- formation of the helix. This is also proxy for higher orders of coiling, interactions with histones and other DNA-binding proteins, and the organization of DNA into chromosomes. DNA ean be built in either an anti- parallel or a parallel format, although the former adds a note of symmetry that may account for the prevalence of the antiparallel in nature. For parallel DNA adifferent enzyme would be need- ed to recognize and replicate the left- compared with the right-hand strand. Recognizing this asymmetry, Watson and Crick! speculated that DNA was antiparallel prior to concrete observa- tional evidence for this conformation. Rarely has a structural determina- tion been coupled so promptly with func- tional implications. Watson and Crick! immediately inferred that DNA duplex- es were formed automatically when each strand was replicated, and that this in- volved the assembly of nucleotides, one by one, complementary to the existing structure.” They overreached the mark by suggesting that this might be possi- ble even without the intervention of spe- cific anabolic enzymes, the discovery of which we owe to the prodigious labors of Arthur Kornberg and his school in the 1960s. But in imputing autocatalytic powers to the DNA double helix, Wat- son and Crick! might lay claim to having anticipated the enzymatic functions of RNA (if not DNA), an iconoclasm that earned the Nobel Prize in 1989 for Sid- ney Altman and Thomas Cech. Despite the intellectual revolution ini- tiated by Watson and Crick,' we might still ask the question, At what point was the welfare of any patient altered by specific knowledge of the double helix? This is a question I agonized over dur- ing the 1970s, and its first answer was perhaps the work of Y. W. Kan on the prenatal diagnosis of hemoglobin disor- ders, using DNA hybridization (1978). How rapidly we have moved in the in- terval is recounted by Caskey® in the companion article. Why did that take 25 years? One may simply point to the enor- mous edifice of contributory knowledge that now bridges the most reductionist aspects of DNA structure to patholog- ical manifestations. HISTORICAL BACKGROUND OF WATSON AND CRICK The biological role of DNA was still enmeshed in controversy in 1953. Nu- cleic acids had been extracted from pus cells by Miescher in 1869, and from the beginning were associated with cell nu- clei. These substances are now known to be macromolecules composed of a lin- ear array of nucleotides joined by phos- phodiester bonds. Cytologists writing in the early 1900s remarked on the as- sociation of nucleic acids with chromo- somes and speculated that this baso- philic material in chromatin might be the substance of genetic continuity. This brilliant anticipation was, however, sub- merged by a misleading observation, The Double Helix—Lederberg 15 namely, the apparent loss of basophilia in the chromosomes of oocytes, leading B. B. Wilson (1925) to remark “That the continued presence of ‘chromatin’ [ie, basi-chromatin] is essential to the ge- netic continuity of the chromosome has, however, become an antiquated notion.” We now know that these chromosomes become remarkably unraveled in keep- ing with their massive involvement in transcription, associated proteins then overshadowing the continuity of the DNA. This skepticism was reinforced by the apparent monotony of DNA structure embedded in Phoebus Levene’s first analyses of DNA. They contained only four constituent nucleotides—each com- prising a phosphate group, a sugar, and one of the four bases: cytosine (C), thy- mine (T), adenine (A), or guanine (G). Within the limited analytical precision available in the 1920s, these appeared to be present in exact stoichiometric equiv- alence. Hence the provisional hypothe- sis of DNA asa tetranucleotide, although it was well recognized that its molecular weight and other key parameters had yet to be ascertained. Nor was there any biological system or array of soure- es to tell that one DNA preparation was in any way different from any other. Such a simple molecule seemed a poor candidate for the miraculous capabili- ties of the gene. On the other hand, pro- teins contained an abundant variety of constituent amino acids (eventually 20). More important, dozens, even hundreds of proteins were isolated with vastly different biological, physical, and chem- ical properties, including wide dispari- ties in composition. The 1920s saw the most exciting developments in protein chemistry, even the crystallization of urease and of pepsin and the demon- stration that enzymes were pure pro- teins (Sumner, 1926; Northrop, 1930). The cap seemed to be a similar charac- terization of the tobacco mosaic virus, claimed to be pure protein by Wendell Stanley in 1985. This was, however, soon to be corrected by Bawden and Pirie in 1937, who found phosphorus and carbo- hydrate in infectious concentrates of to- bacco mosaic virus and inferred the pres- ence of RNA. Stanley, nevertheless, re- ceived the Nobel Prize in chemistry in 1946, together with Sumner and Northrop. By that time, Stanley ac- knowledged “that the nucleic acid could not be removed without causing loss of virus activity and there was general agreement that the virus was a nucleo- protein.” Thus, this prize was a noble reinforcement of the primacy of proteins as the seat of biological specificity. The breakthrough challenge to that dogma was thrust forth in 1944 by Os- 16 JAMA, April 21, 1993—Vol 269, No. 15 wald T. Avery, Colin MacLeod, and Maclyn McCarty. They had studied the diverse serological types of the pneu- mococcus and followed up Griffith’s re- port (1928) that these could be altered or transformed by extracts of other strains. The gist of the 1944 study was that the transforming substance was DNA! This was contrary to expecta- tions that the carbohydrate antigen or some associated protein would be the transforming substance. Avery, a mem- ber of the same Rockefeller Institute as Wendell Stanley, was intimately famil- iar and impressed with the difficulties of characterizing biopolymers. Though ful- ly cognizant of the biological implica- tions of the discovery, he was even more hesitant to dwell on them—but did in- elude a remark that “The inducing sub- stance has been likened to a gene... .” Their claims, of course, aroused in- tense critical controversy, largely around the obvious question whether their DNA preparations were still contaminated with traces of biologically active pro- tein. Avogadro’s number, 6X10“ per mole, would allow a residuum of 10% protein molecules per microgram of a preparation that was 99.99% protein free, at the limit of analytical detect- ability. The sensitivity of the active ma- terials to deoxyribonuclease might be ascribed to a protective rather than in- formational function of the DNA. Like- wise, the insensitivity to proteases might be an attribute of a nucleoprotein complex. My own role in the debate was a will- ingness, even desire, to believe—but a sense of responsibility that the issue was too important to be regarded as closed until there was no escape. It was not clear what feasible experiments (short of ab initio synthesis of DNA) could ul- timately seal all these infinitesimal loop- holes. One might go along with “DNA” as a working hypothesis, and some did. Most biologists blurred their judgments by talking about nucleoproteins—not necessarily informed by the distinction they were implying. Some might have meant something like “protein” or “nu- cleic acid” or a combination thereof, but please do not ask the role of the con- stituents. A rare few gambled on the DNA-—as in some sense did Watson and Crick,’ although they would have en- joyed working out its structure regard- less of its biological implications. In the event, the final elucidation of DNA struc- ture was a horse race. By Watson’s own account, only a few weeks would have separated their priority from the loom- ing insights of Maurice Wilkins and Rosalind Franklin (who had provided the critical experimental data) or of Li- nus Pauling. The biological significance of the pneu- mococcus transformation was also prob- lematical. It looked like a transfer of genetic information; but until 1951, the only markers tested were the serotype antigens. Could one extrapolate from those to genes in general, particularly given that the very idea of a bacterial genetics was in its infancy? After the 1944 bombshell, more chem- ical attention was given to the tetranu- cleotide model, and signs of greater chemical complexity emerged. Of par- ticular import were the deviations of the four bases from the simplistic 1:1:1:1 ratio, found by Erwin Chargaff. Fur- thermore, DNA from different sources exhibited different base composition. So perhaps DNA could be more complex, more diversified than previously thought—could be rehabilitated as a can- didate for the gene. During the 1940s, the Feulgen cytochemical test for DNA and analyses indicating constancy of DNA per genome in somatic cells and a halving in germ cells also added to DNA’s respectability. But these findings did not necessarily prove more than a struc- tural or scaffolding role for the DNA. The pneumococcus transformation re- mained the only biological assay for a genetie role for DNA—in contrast to the innumerable enzyme and immuno- logical assays available for candidate proteins. This impasse was alleviated by the broadening of phage research, sternly governed by Max Delbruck’s genius, to embrace a wider range of chemical stud- ies of phage infection. A critical one was the 1952 double-labeling experiment of Hershey and Chase. Most of the S-35 label (capsid protein) was excluded from infected cells; most of the P-32 (DNA) entered and was transmitted to the phage progeny. This experiment is of- ten cited as the crowning blow on behalf of the “DNA-only” model. But Hershey himself did not go so far—well aware that “most” is not “all,” he was still re- ferring to “nucleoprotein” in 19583—and this at the same Cold Spring Harbor Symposium that sponsored a critical dis- cussion of the paper by Watson and Crick.’ The article by Watson and Crick' did not, of course, bear directly on the loop- holes in Avery’s claims. It did add a further note of plausibility to a DNA- only concept of the gene. In the absence of any serious contradiction, this grad- ually hardened from working hypothe- sis to central dogma. The most serious challenge today is the prion hypothesis: that some “infectious” agents may be devoid of nucleie acid. This is still con- tentious at an experimental level: the hypothesis least in conflict with nucleic The Double Helix—Lederberg doctrine is that the infectious prion is a sort of epitaxial primer of aggregation of a host-determined protein. This still leaves obscure how and whether differ- ent prions could maintain and propa- gate their identity in a genetically de- fined host. Long after many other lines of evi- dence converged to support an infor- mational role of DNA—eg, the colinear- ity of DNA sequences with protein prod- ucts (Yanofsky), genetically active DNA was eventually synthesized in the chem- ical laboratory (Khorana) and repli- cated enzymologically (Kornberg), fully vindicating Avery et al and those who gave their faith to these propositions. THE FLOWERING OF MOLECULAR GENETICS Since the rediscovery in 1900 of Men- del’s 1865 work, genetics has had an extraordinary development, even with- out the benefit of tangible physical and chemical models of the genetic material. The biological phenomena of mutation and of sexual crossing (genetic recom- bination) opened the door to experiments in which existing organisms were the reagents. Genomes could be mixed by crossing, and new combinations of fac- tors segregated into the offspring. Like- wise, fruit flies could be subjected to radiation, and variant or mutant forms discovered. Genetic information is or- ganized into linear chromosomes, and the processes of meiosis in gametogen- esis: precise synapsis of homologues and crossing-over or segmental exchange of chromosome parts allowed powerful dis- section of fine structure on a scale that rivals that of microchemical analysis. These methods continue to play an in- dispensable role in the denomination and mapping of mutant genes. By 194], through the work of Beadle and Tatum, the groundwork of biochemical genetics had been laid—the role of genes in the prescription of protein products, and the use of mutations in the dissection of met- abolic pathways. Indeed, many of these ideas had been anticipated by Archibald Garrod’s studies of human biochemical defects at the very dawn of genetics. Since 1953, we have had a new lan- guage for the description of genes: they are now segments of DNA that can be defined and manipulated as chemical en- tities. The linguistic transition has been conceptually smooth, though marked by occasional generational quarrels. Under- standably, very few individuals can com- bine erudite knowledge of the life his- tories of a wide range of organisms in their natural habitats with focused and specialized knowledge of biochemical ma- nipulations in the Jaboratory. Nor have many radical revisions of genetic doc- JAMA, April 21, 1993—Vol 269, No. 15 trine issued from the molecular perspec- tive. We have had to acknowledge that genes, as bits of DNA, are subject to a wider range of chemical and biological interactions than was _ previously thought—especially with other DNA. The icon of stability of genomes has been shaken by the discovery of transposable elements, first noted in maize by McClintock in 1951; these remained in- explicable until they could be studied as DNA molecules. And concentrating on DNA now allows us to inject genes with viruses, needles, even “shotguns,” into a range of cellular targets including the germ line, providing a technical revo- lution in the construction of new geno- types in all kinds of organisms—bacte- ria, plants, and mammals. Meanwhile, other advances, notably the extension of recombination analysis to somatic cells in culture by cell fusion, have extended the technical power of genetic analysis in ways compatible with, but not dependent on, the double helix. It is paradoxical that the human chro- mosome number, 2n = 46, was not cor- rectly understood until 1956 (Tjio and Levan), and that for about 20 years thereafter this was at least as important in the development of human genetics as was the structure of DNA. The adumbration of DNA-based re- search, molecular genetics, since 1953 would embrace a substantial fraction of world science. Many encyclopedic mono- graphs struggle to record the details and promptly become obsolete. We can hardly do more herein than summarize the major headings, following an impre- cise dichotomy distinguishing topologi- eal DNA—an informational duplex— from mechanical DN A—a three-dimen- sional geometric object. DNA AS AN INFORMATIONAL DUPLEX Denaturation and Hybridization The most elementary aspect of the du- plex is the separability of its strands, us- ing temperature or chemical denaturants. A-T base pairs melt (separate from one another) at a lower temperature than G-C pairs, so melting curves can distinguish DNA ofdifferent base composition. Single strands once separated can also be rean- nealed, allowed to rejoin, the kinetics al- lowing the discovery that much DNA (Gin eukaryotes) has a repetitive or a redun- dant sequence. Radioactively labeled probes can be used to ferret out target ho- mologous DNA with high precision. Homology and Evolution; Polymorphism Within the Species These and related methods can be used as quantitative indices of the genetic relatedness of diverse species, supplant- ing the subjectively evaluated morpho- logical criteria used in systematics here- tofore. Within the species, genetic poly- morphism can now be described at the DNA level—one astonishing finding is that humans are typically heterozygous with a prevalence of two or three per 1000, ie, almost once in every gene. As most of these base substitutions have no perceptible phenotypic effect, random drift (rather than selectible or adaptive change) may predominate in evolution- ary change (Kimura). Mutagenesis and DNA Repair The vulnerability of genes to muta- tional change in response to x-rays was known empirically since 1927 (Muller), and to chemicals since 1944 (Auerbach). Early hopes that chemical mutagenesis would be a direct path to the chemistry of the gene were not substantiated. Most chemical mutagens react with amino ac- ids as well as DNA bases. The excep- tions are nuclein base analogues, which may be misincorporated into DNA; but these were discovered much later. Above all, we now understand that the initial lesions in DNA would usually be lethal, and that eventual mutations are the re- sult of intricate repair metabolism that occasionally misfires. Transcription; Genetic Code The “central dogma” of information flow has emerged, that DNA- (transcription) RNA-} (translation) protein. The base sequence of DNA is transcribed faithful- ly into a messenger RNA copy. This in turn governs the assembly of a polypep- tide sequence, each three-base frame of RNA encoding one particular amino acid. The polypeptide then folds (perhaps with the guidance of a chaperone) into a pre- ordained protein three-dimensional shape, which can then function as an enzyme, antibody, hormone, structural unit, and so forth. This folding process is not yet fully computable. There may even be cir- cumstances where a given polypeptide might have alternative foldings—but this is not accepted dogma. The details of messenger RNA syn- thesis have become much more intri- cate. Primary transcripts are usually processed, only some of the RNA tracts being spliced together to form the final message. The other “intervening se- quences,” or introns, may be the major part of the RNA—their functions re- main obscure. As with repeated sequenc- es, they may reflect “selfish DNA,” whose presence in the genome has little to do with their adaptive value to the overall organism. In other examples, RNA may be edited in other ways be- fore translation is completed. The Double Helix—Lederberg 17 Enzymology: Nucleases, Ligase, Replication; Reverse Transcriptase For a legion of brilliant and tireless investigators, the DNA structural mod- el has been the platform for isolating a host of enzymes involved in every as- pect of DNA metabolism. Besides giv- ing us that metabolic map, explaining how DNA is replicated, sliced, stitched, spliced, and repaired, these enzymes are the vital technical tools for further study of DNA and for the engineering of new constructs. Some viruses, notoriously the retrovi- ruses (including human immunodeficien- cy virus), exhibit areverse transcriptase, whereby RNA->DNA. This knowledge is indispensable to the virologist. It has also given some of the most valuable tools for studying RNA, eg, messenger, by al- lowing the production of DNA copies for input into other technology. Tools for Engineering: DNA Splicing; PCR These sempstering tools have found- ed the multibillion-dollar biotechnology industry. DNA tailored in vitro, with inserts from human or a variety of other sources, can be patched into convenient host garments (from bacteria to cows) for the easier exhibition of a variety of products—growth factors, enzymes, im- munizing antigens, replacement thera- peutics (like clotting factors)—in unlim- ited variety. Related technology is used to target specific host genes, to eluci- date their functions in physiology and development. The PCR (polymerase chain reaction) has been the instrument of the “democ- ratization of molecular biology.” With it a single DNA molecule in a messy mix- ture can be fished out and amplified ad libitum, most importantly at low cost and with simple instruments. High school students do experiments today that would have been doctoral disser- tations 15 years ago. The applications range widely, from forensics and diag- nosis of genetic disease to the hunt for new viruses and the revival of fossil DNA. At its heart, a synthetic DNA probe is a rational, linear, digital signa- ture to locate any counterpart in the analysand. Its core of combinatorial spec- ificity can be contrasted with that of antibodies, which is founded on three- dimensional shapes of the immunoglo- bulin and its targets. Drug Discovery DNA combinatorics has reached anew peak in a paradigm for drug discovery that mimics natural evolution.’ Random- ized DNA sequences are expressed on host cells (or phages), and these are then 18 = JAMA, April 21, 1993—Vol 269, No. 15 selectively screened for specificities of binding to specific reagents—usually receptors for which agonists or an- tagonists are sought. The cell express- ing the desired epitope can then be grown out for larger scale production and testing. In one application, the mam- malian antibody-forming mechanism can be emulated, and mutant immuno- globulin polypeptides selected for the desired specificity. RNA can fold into stereospecific objects; hence, random- ized RNA molecules can be directly selected and replicated with reverse transcriptase. Human Genome Project With the availability of all of these tools, the image has firmed of establish- ing the complete DNA sequence of the human genome. As a scientific objec- tive, this is uncontroversial. The con- troversy pertains to the primacy given to the staging of the effort. Should it be a once and for all technological produc- tion, mindless of the ancillary interest in some genes or DNA tracts compared with others? Does it need to be a cen- tralized project, administered top-down with the trappings (and political appeal) of other Big Science? Or can it be left to the cumulative efforts of hundreds or thousands of laboratories, each digging more deeply at some features of the ter- rain, and intent on going much further than establishing a sequence of bases? In fact, we are seeing the emergence of constructive compromise among these visions; and at the same time the tech- nologies of mapping and sequencing are advancing to where the costs of a uni- fied project need no longer prejudice more individualized efforts. In any case, sequence information is but the beginning of more intensive in- quiry into the polymorphisms, regula- tory factors, and gene functions associ- ated with any DNA segment. DNA AS A HELIX Higher Orders of Organization The visible chromosome is a packag- ing of DNA, histones, and accessory pro- teins three or four orders of coiling be- yond the double helix. Cytological ob- servation leaves no doubt that the mor- phological expression of the chromosome reflects functional allocation of different genes; but we are at the mere beginning of understanding. Gene Regulation and Morphogenesis The basic outlines of the central dog- ma now consensually agreed, the core challenge of molecular biology has been the path from the gene to the organism. Given that, to some approximation, each somatic cell has the identical genotype, (1) how is gene expression differentially modulated, and (2) how is this trans- mitted in cell lineages? A multitude of DNA-binding proteins have been found that do modulate gene expression: transcriptional regulators. As a three-dimensional interaction, protein binding is fully sensitive to three- dimensional shape and the major and mi- nor grooves of the double helix, as well as the base sequences contained therein. In addition, if not in consequence of bound proteins, some tracts of DNA are meth- ylated shortly after DNA replication, in ways correlated with gene activation. How these properties are locally transmitted remains a matter of spec- ulation, but may well be bound up with local methylation. DNA Supercoiling; Topoisomerases; Other Conformations The standard double helix exhibits a pitch of about 10 base pairs per com- plete turn. If nothing else, the processes of replication and transcription would entail the unraveling and rewinding of the helices: this is the task of enzymes generically called topoisomerases. These can transiently cut single strands to per- mit the relief of torsional stress, then rejoin them. In its natural habitat, DNA is often found in states of positive or negative supercoiling, often correlated with maintained gene expression. In ad- dition, many cytotoxic and cancer che- motherapeutic agents seem to be topoi- somerase inhibitors, and most owe some of their specificity to the momentary DNA-supercoil status of a given cell. It is particularly intriguing that environ- mental signals can modulate that sta- tus, often by regulating the production of the various topoisomerases. At least in vitro, DNA can undergo a spontaneous transition to a totally dif- ferent, kinked and left-handed confor- mation called Z-DNA. Tracts rich in G-C pairs are especially prone to this shift. The importance of Z-DNA in vivo is hotly contested. DNA conformations plainly confer dif- ferent chemical reactivity on the bases, a principle exploited by the footprinting methods used to study conformation. This must have some implications for localized chemical mutagenesis—a mat- ter not yet systematically studied. TRIUMPH OF MECHANISM The dominion of the DNA paradigm has been the triumph of mechanistic in- terpretation in 20th-century biology. It is sometimes remarked that human per- sonality is nothing but the individual’s 3 billion base pairs—an assertion that fascinates some, terrifies others, and has The Double Helix—Lederberg much to do with the debate about the Human Genome Project. If we could be- lieve that existing genotypes had achieved more than a tiny fraction of the human potential—in culture, in intel- lect, in compassion, in a sane ordering of affairs—we could elevate the genome to that pedestal of nemesis. On the other hand, we do know that many, probably most, individuals labor under some po- tentially remediable burden of heredi- tary origin. As much to understand the better nurturing of human develop- ment, a euphenics, as to intervene in genetic constitution, eugenics, it does behoove us to learn all we can about genetic polymorphisms and their impact on human health and capability. It is particularly important to distinguish in- terventions in germ cells from those in the somatic cells, and to communicate that it is only the latter that are intend- ed to be the targets of the new gene therapies. JAMA, April 21, 1993—Vol 269, No. 15 SELECTED READINGS It would be a precious exercise to provide specific documentation for every detail of this commentary; it would be both arduous and redundant—many single points would deserve a library. The up-to-date detail can be found in standard texts of molecular biology (a few are listed) and in the volumes of Annual Review of Biochemistry. The leading historical monographs on DNA are also listed. Molecular Biology Alberts B, Bray D, Lewis J, Raff M, Roberts K, Wat- son JD. Molecular Biology of the Cell. New York, NY: Garland Publishing; 1989, Berg P, Singer M. Dealing With Genes: The Language of Heredity. Mill Valley, Calif: University Science Books; 1992. Darnell JE, Lodish HF, Baltimore D. Molecular Cell Biology. New York, NY: Scientific American Books; 1990. Kornberg A, Baker TA. DNA Replication. 2nd ed. New York, NY: WH Freeman & Co; 1992. Stryer L. Biochemistry. New York, NY: WH Freeman & Co; 1988. Watson JD, Hopkins NH, Roberts JW, Steitz JA, Weiner AM. Molecular Biology of the Gene. Menlo Park, Calif: Benjamin/Cummings; 1987. Wells RD, Harvey SC, eds. Unusual DNA Structures. New York, NY: Springer-Verlag NY Inc; 1988. History of DNA Crick FHC. What Mad Pursuit: A Personal View of Scientific Discovery. New York, NY: Basic Books Ine Publishers; 1988. Dubos RJ. The Professor, the Institute, and DNA. New York, NY: Rockefeller University Press; 1976. Fruton JS. Molecules and Life. New York, NY: Wiley- Interscience; 1972. Judson HF. The Eighth Day of Creation. New York, NY: Simon & Schuster; 1979. Lederberg J. Genetic recombination in bacteria: a dis- covery account. Annu Rev Genet. 1987;21:23-46. McCarty M. The Transforming Principle. New York, NY: WW Norton & Co; 1985. Olby RC. The Path to the Double Helix. Seattle: Uni- versity of Washington Press; 1974. Portugal FH, Cohen JS. A Century of DNA. Cam- bridge, Mass: MIT Press; 1977. Sayre A. Rosalind Franklin and DNA. New York, NY: WW Norton & Co; 1975. Watson JD. The Double Heliz. New York, NY: Athe- neum Publishers; 1968. Watson JD, Tooze J. The DNA Story. San Francisco, Calif: WH Freeman & Co; 1981. References 1, Watson JD, Crick FHC. Molecular structure of nu- cleie acids: a structure for deoxyribose nucleic acid. Nature. 1953;171:737-738. 2. Watson JD, Crick FHC. Genetical implications of the structure of deoxyribonucleic acid. Nature. 1953;171:964- 967. 3. Caskey CT. Molecular medicine: a spin-off from the helix. JAMA. 1993;269:1986-1992. 4, Pei D, Ulrich HD, Schultz PG. A combinatorial ap- proach toward DNA recognition. Science. 1991;253: 1408-1411. The Double Helix—Lederberg 19