PERSPECTIVE PERSPECTIVE DNA cloning: A personal view after 40 years Stanley N. Cohen1 Departments of Genetics and Medicine, Stanford University School of Medicine, Stanford, CA 94305

Edited by Joseph L. Goldstein, University of Texas Southwestern Medical Center, Dallas, TX, and approved August 13, 2013 (received for review August 2, 2013)

In November 1973, my colleagues A. C. Y. Chang, H. W. Boyer, R. B. Helling, and I reported in PNAS that individual genes can be cloned and isolated by enzymatically cleaving DNA molecules into fragments, linking the fragments to an autonomously replicating plasmid, and introducing the resulting recombinant DNA molecules into bacteria. A few months later, Chang and I reported that genes from unrelated bacterial species can be combined and propagated using the same approach and that interspecies recombinant DNA molecules can produce a biologically functional protein in a foreign host. Soon afterward, Boyer’s laboratory and mine published our collaborative discovery that even genes from animal cells can be cloned in bacteria. These three PNAS papers quickly led to the use of DNA cloning methods in multiple areas of the biological and chemical sciences. They also resulted in a highly public controversy about the potential hazards of laboratory manipulation of genetic material, a decision by Stanford University and the University of California to seek patents on the technology that Boyer and I had invented, and the application of DNA cloning methods for commercial purposes. In the 40 years that have passed since publication of our findings, use of DNA cloning has produced insights about the workings of genes and cells in health and disease and has altered the nature of the biotechnology and biopharmaceutical industries. Here, I provide a personal perspective of the events that led to, and followed, our report of DNA cloning.

restriction enzyme | pSC101 | EcoRI | genetic engineering | gene cloning

In a PNAS paper entitled “Construction of would prevent propagation of genes across a highly public controversy about potential Biologically Functional Bacterial Plasmids different biological domains. Stringent host hazards of “genetic tinkering,” a decision by In Vitro,” mycolleaguesA.C.Y.Chang, range limitations to virus propagation had Stanford University and the University of H. W. Boyer, R. B. Helling, and I reported been observed, and, in some instances, California to seek patents on the technology in November 1973 that individual genes impediments to survival of foreign DNA that Boyer and I had invented, and efforts can be cloned and isolated by enzymatically had been found even among subgroups of by entrepreneurs and industry to implement fragmenting DNA molecules, linking the the same species (9). Supporting the notion DNA cloning methods for commercial pur- pooled fragments to autonomously replicating that DNA was unlikely to survive in cells poses. In the 40 years that have now passed circular bacterial genetic elements known as of an unrelated species was the finding since publication of these PNAS papers, use of plasmids, and introducing the resulting re- that individual biological species maintain DNA cloning methods has produced impor- combinant DNA molecules into bacteria (1). characteristic ratios of A+TtoG+Cbase tant insights about the workings of genes and Boyer and I were young faculty at the Univer- pairs (10, 11). Our discovery that DNA can be cells in health and disease and has profoundly sity of California, San Francisco (UCSF) and transplanted to, and propagated in, a differ- altered the biotechnology and pharmaceutical Stanford, respectively. Annie Chang was a Re- ent species, and even in a different biological industries. I provide here a personal perspec- search Technician in my laboratory and Bob kingdom, by attaching it to a vector indig- tive of these events. Helling was a University of Michigan professor enous to the recipient led to the realization on sabbatical leave in Boyer’s laboratory. A that natural barriers to DNA survival are Plasmids and Antibiotic Resistance few months later, Chang and I reported that not so constraining after all, and that “ge- After the development of antimicrobial agents genes from totally unrelated bacterial species netic engineering”—atleastatthecellular in the 1940s, the notion was prevalent that can be combined and propagated using the level—is possible (8). It also provided a pro- these drugs would end infectious diseases same approach (2) and that interspecies tocol that enabled such engineering to be done caused by bacteria. Of course that did not recombinant DNA molecules can produce by virtually any laboratory having modest happen, and the reason was the occurrence a biologically functional protein in a foreign genetic and biochemical capabilities. of antibiotic resistance. Investigations carried host. Soon afterward, Boyer’s laboratory Our DNA cloning experiments resulted out primarily in laboratories in Japan and the and mine published collaborative experiments from the pursuit of fundamental biological United Kingdom in the early 1960s showed demonstrating that genes from eukaryotic questions rather than goals that most observers that antibiotic resistance in bacteria com- cells can be cloned in bacteria (3). might regard as practical or “translational.” I monly is associated with the acquisition of Bacterial viruses and plasmids had been was investigating mechanisms underlying genes—often multiple genes—capable of de- shown to pick up DNA from the chromo- the ability of plasmids to acquire genes con- stroying antibiotics or otherwise interfering somes of their hosts (4); hybrid viruses from ferring antibiotic resistance and to exist sep- animal cells also had been reported (5, 6). arately from bacterial chromosomes; Herb Author contributions: S.N.C. wrote the paper.

However, it had long been known that only Boyer was studying enzymes that restrict The author declares no conflict of interest.

closely related species can interbreed and pro- and destroy foreign DNA. The PNAS publica- This article is a PNAS Direct Submission. duce viable offspring, and hybrids displaying tions resulting from these pursuits generated fi — This article was invited in recognition of the 40th anniversary of heritable characteristics of very different spe- considerable scienti c excitement and work the November 1973 PNAS paper by S. N. Cohen, A. C. Y. Chang, cies exist only in mythology; thus, there was aimed at repeating and extending the findings H. W. Boyer, and R. B. Helling reporting a method for constructing uncertainty about whether so-called “nat- was undertaken almost immediately by and cloning biologically functional DNA molecules (1). ural barriers created during evolution” (7, 8) other researchers. The papers also prompted 1E-mail: [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1313397110 PNAS Early Edition | 1of9 Downloaded by guest on September 27, 2021 with their actions. The resistance properties Research in the burgeoning field of mo- that are distinguishable visually from thinner, commonly did not map genetically to the lecular biology during the 1960s focused on more kinky regions of single-stranded DNA. bacterial chromosomes, suggesting that the bacteriophages for an important reason: We expected that such experiments would genes encoding resistance were located on a bacterial cell infected by a virus generates provide information about the structural re- separate elements (some had called them thousands of identical copies—clones—of lationships between resistance genes that had episomes) analogous to the fertility factor a single infecting genome during the normal been picked up by plasmids during their (F-factor) discovered earlier (12). Like F- viral life cycle. Thus, phenotypic effects can meandering through bacterial populations. factors, resistance factors (R-factors) were be correlated with the results of biochemical The results of these experiments and also of capable of being transferred between bac- analyses. I realized that elucidation of how separate investigations from Davidson’slab- teria by cell-to-cell contact (13, 14). In 1952, resistance genes function and how R-plas- oratory (27, 28) showed remarkable sequence Joshua Lederberg had given the name “plas- mids evolve required a way to clone in- conservation among large segments of dif- mids” to such extrachromosomal genetic ele- dividual plasmid DNA molecules and to ferent R-plasmids and, importantly, provided ments (15). The antibiotic-inactivating genes isolate the resistance genes. Genetic map- direct physical evidence that plasmid se- carried by resistance plasmids provide a bi- ping of R-plasmid properties had led to the quences associated with interbacterial DNA ological advantage to host bacteria in pop- prediction that bacterial plasmids exist as transfer had become linked covalently to re- ulations exposed to antimicrobial drugs, DNA circles (19, 21, 22), and I proposed to sistance genes to form large circles of ’ and, in barely a decade after the intro- use circularity to isolate intact resistance R-plasmid DNA. Sharp s electron micros- duction of antibiotics to treat human in- plasmid DNA. If I could obtain R-plasmid copy also detected a phenomenon that we — — ’ fi fections, R-plasmid–mediated multidrug circles I reasoned IcouldapplyDNA didn t yet understand the signi cance of: resistance had become a major medical fragmentation approaches I had used to short inverted repeats of DNA sequences λ problem as well as a scientificenigma. study gene expression, together with ul- that bracket regions containing certain re- — The transfer of resistance properties be- tracentrifugation analysis and DNA hybrid- sistance genes presaging by a couple of years tween bacteria was found by centrifugation ization methods, to assess changes in circle the discovery of insertion sequence (IS) ele- analysis to be sometimes associated with the size associated with the gain or loss of re- ments and antibiotic resistance transposons, acquisition of heterogeneous DNA bands sistance phenotypes or transferability. Stan- which enter plasmids by illegitimate (i.e., (16, 17). However, the molecular nature of ley Falkow, whose published investigations nonhomologous) recombination and lead to this DNA was controversial. Particularly at Walter Reed Medical Center and Geor- rearrangements of plasmid structure (for getown University had been instrumental in uncertain was whether the resistance and reviews, see refs. 29 and 30). attracting me to plasmid biology, agreed to transfer components of R-plasmids were provide bacterial strains and plasmids for my Bacterial Transformation by Plasmid carried by the same DNA molecule or were initial experiments. However, Falkow also DNA located on separate molecules that can in- told me that he planned to stop working on The methods for heteroduplex analysis that I teract transiently during interbacterial trans- plasmids to pursue other scientific interests, learned from Sharp during our collaboration fer (18, 19). Importantly, nothing was known and the decision of an established expert to enabled me to use electron microscopy in my about the genetic recombination mecha- abandon a field that I had just decided to enter plasmid studies at Stanford when he and nisms that had enabled the accumulation of was disconcerting! However, there was also Davidson later moved on to other scientific multiple resistance genes on the same genetic good news: in October 1967, Donald Helinski interests. However, still missing from a element. Before moving to Stanford, I pre- and his colleagues at the University of Cal- growing collection of tools available to in- pared National Institutes of Health (NIH) ifornia, San Diego published the first molecular vestigate the molecular biology of plasmids and National Science Foundation (NSF) re- evidence for circular DNA forms of a trans- was a method for reintroducing plasmid search proposals aimed at addressing these missible plasmid (23), supporting the notion DNA molecules into bacteria. Genetic trans- questions. As a postdoctoral fellow in the that R-factors, too, would be DNA circles. formation using naked DNA had been shown laboratory of Jerard Hurwitz at the Albert I arrived at Stanford in March 1968, and, for pneumococcus, Bacillus species, and cer- Einstein College of Medicine, I had found the following year, Christine Miller, who was tain other bacteria, but not for Escherichia coli that genes on different segments of DNA of then a newly hired laboratory technician, and or other microbes that carry the R-plasmids I λ bacteriophage were controlled temporally Ireportedthepurification of intact circular was studying. Quite fortuitously, a critical at the level of transcription during the pro- DNA of the large antibiotic resistance plas- discoverybyMortonMandelandAkikoHiga duction of virus particles (20), and I thought mid R1 (24). Several months later, I was in- reported in the October 1970 issue of the that some of the approaches I used in those vited to Caltech by Jerome Vinograd, a Journal of Molecular Biology (31) pointed the experiments could be applied to the study of pioneer in the study of circular DNA of an- way toward such a method. R-plasmids. I had been trained as both imal cells, to give a seminar talk about our Mandel and Higa found that E. coli cells a physician and scientist and believed that an plasmid results. Discussions with Norman treated with calcium ions can take up DNA of understanding of resistance plasmids was Davidson and his graduate student Phillip the temperate bacteriophages λ and P22 and both medically and scientifically important. Sharp during that visit initiated a collabora- that such cells can release infective virus. The Bacteriophage λ was the most extensively and tion between our laboratories aimed at using procedure is robust, and Peter Lobban, a competitively investigated bacteriophage of electron microscope-based heteroduplex an- graduate student working with A. Dale Kaiser that era, but the role of R-plasmids in anti- alysis, which had been developed earlier in in the Stanford Department of Biochemistry, biotic resistance was being studied in only Davidson’s laboratory (25) and also at the had begun using it to introduce phage P22 a small number of microbiology labora- University of Wisconsin (26), to compare DNA into Salmonella typhimurium, aclose tories; I liked the prospect of working in regions of sequence similarity on different relative of E. coli.MandelandHigahad what was still a quiet backwater of scien- plasmids: annealing of homologous DNAs reported an attempt to genetically transform tific research. results in smooth thickened DNA segments E. coli to express an antibiotic resistance gene

2of9 | www.pnas.org/cgi/doi/10.1073/pnas.1313397110 Cohen Downloaded by guest on September 27, 2021 present on the bacterial chromosome and experiments carried out with Sharp and Herb Boyer was studying restriction en- PERSPECTIVE had met with failure. However, plasmids, like Davidson that R6 and R6-5 contained re- donucleases in his laboratory at UCSF. Col- phages, are extrachromosomal elements; if peats of some DNA sequences and hoped laborating with him was Assistant Professor R-plasmid DNA could be taken up by E. coli that ordinary genetic recombination be- in Residence Daisy Dussoix (then Roulland- at even a low frequency, and if the antibiotic tween these homologous segments would Dussoix), who as a student had participated resistance genes carried by circular R-plasmid lead to circularization of shear-generated with Arber in early investigations of the replicons were expressed in these cells, colo- fragments in calcium chloride-treated cells— restriction/modification phenomenon. Some nies of bacteria that acquire plasmids could and perhaps even the formation of novel plasmids had been found to encode restriction be selected using culture media containing plasmids lacking some of the original DNA enzymes (38, 39), and the Ph.D. thesis project appropriate antibiotics. Such selection might resistance determinants. If such recombi- of Robert Yoshimori, a graduate student enable cloning of single molecules of plas- nants occurred even infrequently, antibiotic working with Boyer and Roulland-Dussoix, midDNA.IaskedLeslieHsu,afirst-year selection might identify colonies of bacteria was to identify new restriction enzymes that medical student who had come to my lab- acquiring them. We did in fact identify might be encoded by E. coli plasmids. He oratory for research training, to undertake bacteria containing a small plasmid that ex- found two such enzymes (40), and Boyer such experiments. pressed only tetracycline resistance and that and his colleagues proceeded to purify them By late 1971, Leslie, with the help of Annie we thought had been derived from R-5 (34). and investigate their properties. One of the Chang, had shown that bacterial clones con- However, our later investigations (35) in- enzymes was EcoRI (E. coli Restriction en- taining autonomously replicating plasmid dicated that this was a natural plasmid that donuclease I). Like HindII, EcoRI cleaved molecules can be obtained using a modifica- had contaminated the DNA preparations; DNA site-specifically, and Boyer set out to tion of the Mandel and Higa procedure. Later we had underestimated the power of anti- determine the nucleotide sequence attacked work by others (32) indicated that the failure biotic selection to identify rare bacterial cells by the enzyme. Boyer also provided a sam- of chromosomal DNA to transform calcium transformed by resistance plasmids. Whereas ple of EcoRI to Norman Salzman at the chloride-treated E. coli in Mandel and Higa’s oureffortsinmid-1972tocloneresistance NIH, whose laboratory found that the on- earlier experiments had resulted from exo- genesfromplasmidsusedlargelythesame cogenic virus SV40 is cleaved once by this nucleolytic digestion of fragmented chromo- strategy that Boyer and I and our colleagues endonuclease and used the cleavage site to somal DNA by a bacterial enzyme. Luckily, used successfully less than a year later, they map the origin and direction of SV40 DNA the circularity of R-plasmid DNA molecules lacked a key ingredient of the later experi- replication (41). John Morrow, a Stanford had avoided this pitfall. The resulting E. coli ments: an enzyme that cuts each plasmid graduate student of Paul Berg, who also re- colonies each contained a circular DNA spe- DNA molecule identically and produces com- ceived EcoRI from Boyer for testing on SV40 cies having the resistance, fertility, and sedi- plementary sequences at the ends of DNA DNA, made the same observation (42). Ad- mentation properties of the parental genetic fragments it generates—i.e., the restriction ditional experiments with EcoRI samples that element. Publication of our paper reporting endonuclease EcoRI. Boyer had given to others yielded an un- these findings in August 1972 (33) interested expected dividend: evidence that EcoRI, un- plasmid researchers but, so far as I could de- Restriction Endonucleases like HindII, cleaves the DNA sequence it termine, was hardly noticed by others. There The ability of bacteria to restrict the growth recognizes asymmetrically, generating single- was scant awareness in the phage-oriented of phage that had been propagated on other strand extensions that contain nucleotides world of 1972 molecular biology of the strains had been known since the late 1930s, complementary to those present at the ends implications of being able to clone plasmid but work aimed at understanding the mech- of other EcoRI-generated fragments. DNA molecules, and our report did nothing anism underlying this phenomenon didn’t Hydrogen bonding between dA and dT to alter the backwater nature of the field of begin for another 20 years. Much of that work deoxynucleotides and between dGs and dCs plasmid biology. That was fine with me: I was was carried out by the Swiss microbiologist had been known for a decade to be able to a junior scientist whose laboratory included and geneticist Werner Arber and his student hold DNA strands together. Alfred Hershey just a few students and postdocs, plus two Daisy Dussoix, who showed that the DNA and his colleagues at the Cold Spring Harbor research assistants. My primary academic of restricted phage is enzymatically de- Laboratory had reported in 1963 that bacte- appointment at Stanford was then in the graded (9). In 1970, Hamilton Smith and his riophage λ DNA contains complementary Department of Medicine, and my clinical colleagues at Johns Hopkins University single-strand segments at its ends, enabling teaching responsibilities affected the time I reported that a restriction enzyme they linear DNA that had been packaged in a viral had available for laboratory research; the quiet named HindII—a protein isolated from the particle to become circular and insert into reception that our paper received allowed me bacterial pathogen, Haemophilus influen- the bacterial chromosome (43). Cohesive to proceed with less pressure to undertake the zae—recognizes particular nucleotide se- ends on λ DNA molecules were used as experiments I had long been planning. quences in DNA and cuts duplex DNA substrate by Martin Gellert (44) and others In May 1972, Annie and I began to break site-specifically at these sequences (36). The (45) to isolate an enzyme, DNA ligase, that apart molecules of the large multidrug re- following year, Karen Danna and Daniel covalently joins λ DNA segments held to- sistance plasmids R6 and R6-5 using the Nathans found that the HindII endonu- gether by complementary ends. Comple- mechanical shearing procedure I had used clease cleaves DNA of the mammalian tu- mentary ends were thus well recognized as 6 years earlier to separate and study the two mor virus SV40 into 11 fragments that can a device for joining together DNA mole- halves of the bacteriophage λ genome (20). be separated by acrylamide gel electro- cules (46). Attribution of credit for who first The fragmented plasmid DNA was in- phoresis, demonstrating the utility of re- made the observation that cleavage of du- troduced into calcium chloride-treated bac- striction endonucleases for DNA analysis plex DNA by EcoRI generates fragments teria, and transformants were screened for (37).Arber,Nathans,andSmithreceived that have complementary cohesive termini cells that acquired individual resistance de- the 1978 Nobel Prize in Physiology or has been controversial, but what is shown by terminants. I knew from the heteroduplex Medicine for these accomplishments. the published record is that three separate

Cohen PNAS Early Edition | 3of9 Downloaded by guest on September 27, 2021 papers simultaneously reporting this finding in DNA–DNA hybridization experiments. appeared in the November 1972 issue of Because of the asymmetry of cleavage of the PNAS: the papers were authored by Janet E. EcoRI recognition sequence, the ends of the Mertz and Ronald W. Davis of the Stanford multiple plasmid DNA fragments generated Department of Biochemistry (47), by Vit- by EcoRI would be complementary—and torio Sgaramella of the Stanford Department under the right conditions individual plas- of Genetics (48), and by Boyer and his co- mid DNA fragments in the mixture could workers Howard Goodman and Joe Hedg- join to each other in different combina- peth (49) at UCSF. tions. If cleavage by EcoRI left the repli- cation function of the plasmid intact, the The Hawaii Meeting and the Initial DNA region encoding this function might join Cloning Experiments randomly to different combinations of an- In early 1972, I agreed to join Tsutomu tibiotic resistance genes in the fragment Fig. 1. Cartoon by D. Adair in the Honolulu Advertiser Watanabe, a Keio University microbiologist mixture, forming DNA circles that could newspaper, September 26, 1988 accompanying an arti- who was a pioneer in studies of bacterial be sealed using DNA ligase, as had been cle reporting demolition of the Waikiki beach delicates- antibiotic resistance, and Donald Helinski in shown for the complementary extensions sen where the initial DNA cloning experiments were organizing a United States–Japan conference λ planned. The persons depicted clockwise are presumed at the ends of DNA (44, 45). And the to be H. Boyer (12 o’clock), S. Cohen, G. Brinton, on plasmids later that year. A few weeks plasmid DNA transformation procedure C. Brinton, and S. Falkow. Reprinted by permission. before the meeting, which was held in mid- would enable us to select for, and hopefully November at the University of Hawaii in to separately clone, specific resistance genes Honolulu, Don contacted me to suggest that using appropriate combinations of anti- demonstrating that such recombination is Herb Boyer, whose work on the plasmid- biotics on culture plates. independent of the sequence of the duplex encoded EcoRI enzyme he had just learned Herb initially didn’tappeartobeespe- segments being joined. Jensen et al. (51) had about, be added to the list of speakers. I cially interested in studying R-plasmid DNA used the strategy of bringing natural DNAs telephoned our invitation to Herb and thus structure and offered simply to provide the together by enzymatically adding stretches began a scientific interaction that less than enzyme that he generously had given to of complementary nucleotides to their ter- ’ 6 months later resulted in the cloning of others. However, he was excited about the mini, and Peter Lobban in Dale Kaiser s antibiotic resistance genes from plasmids. use of autonomously replicating R-plasmids laboratory and David Jackson in Paul Berg’s The actual collaboration began during a to clone EcoRI-generated DNA fragments. laboratory showed that disparate fragments long walk near Honolulu’s Waikiki Beach in By the time we encountered a small delica- from either the same genome or different search of a sandwich shop to have a late tessen having an enticing window sign that genomes that were held together by enzy- evening snack. Boyer and I were joined by read, “Shalom,” in place of the ubiquitous matically installed complementary single- Stanley Falkow, who recently had moved his “Aloha,” we had decided to proceed collab- strand segments can be ligated to create laboratory to the University of Washington, oratively and agreed on the basic design of covalently bonded junctions (52, 53). Berg Charles Brinton, a microbiologist from the the project that our laboratories would later commented: “it doesn’t take a genius to University of Pittsburgh, and Charles’swife, jointly carry out. We would target the R6-5 figure out that if you can create artificial Ginger. During that walk, Herb and I dis- plasmid, which Sharp, Davidson, and I had ends that are complementary to each other, cussed recent results from our laboratories. learned much about from heteroduplex the two DNA molecules will come together” I described our experiments showing that analysis, and which Chang and I had been (54). Moreover, results obtained by Mertz E. coli could be transformed genetically with shearing using a mechanical stirring device and Davis (47) and by Sgaramella (48) and naked plasmid DNA, and our plasmid DNA and metal blades, in our initial experiments. published during the month of the Hawaii shearing experiments, which had not yet A few minutes later, over warm corned beef meeting showed that the four nucleotide been published, and Herb described the sandwiches and cold beer (Fig. 1), Herb and I single-strand extensions generated by EcoRI similarly unpublished sequencing data that sketched out an experimental plan on nap- are sufficient in length to enable DNA he, Joe Hedgpeth, and Howard Goodman kins taken from the dispenser at our table. fragments to be spliced together in vitro. had obtained for the EcoRI cleavage site. Our strategy was straightforward (Fig. 2), However, notwithstanding our expectation As Herb and I talked, I realized that EcoRI but there was no assurance that it would that we would be able to biochemically join was the missing ingredient needed for mo- work. Yes, we knew that we could genetically the complementary ends of EcoRI-generated lecular analysis of antibiotic resistance plas- transform E. coli with plasmid DNA and use fragmentsofplasmidDNA,therewereim- mids. Large plasmids would be cut specifically antibiotic resistance genes to identify cells portant biological unknowns in the experi- and reproducibly by the enzyme, and this that acquire plasmids, and we expected ments that Boyer and I planned. Would method of cleavage would surely be better fromthenucleotidesequenceattheEcoRI cleavage of R6-5 with EcoRI disrupt regions than the haphazard mechanical shearing cleavage site that the restriction enzyme needed for plasmid DNA replication or ex- methods I had been using for fragmenta- would cut the DNA of our large plasmids pression of antibiotic resistance? And would tion of plasmid DNA circles. Because EcoRI reproducibly into multiple fragments. We recombinant DNA molecules created in the recognizes a six base pair sequence, cleavage knew from published earlier results that laboratory be reproduced and transcribed in sitesonduplexDNAwouldbeonaverage DNAs having complementary ends would bacterial cells? DNA junctions formed during about 4,100 base pairs apart, and each of the link together by base pairing: Khorana and legitimate genetic recombination in cells are DNA fragments produced would likely con- his colleagues had joined together double- generated by a process that has resulted from tain only a few genes. The number of frag- stranded fragments of synthetic DNAs in billions of years of evolution; would the ments would be few enough to separate vitro by chemically adding complementary random joining of DNAs by artificial means them by centrifugation, enabling their use nucleotides to them one at a time (50)— create anomalous chromatin conformations

4of9 | www.pnas.org/cgi/doi/10.1073/pnas.1313397110 Cohen Downloaded by guest on September 27, 2021 EcoRI-Treated and selection of kanamycin resistant cells, PERSPECTIVE Plasmid DNA the gene had been cloned. We were abso- lutely elated but knew that robust DNA C D Random cloning would require a plasmid vector that E includes both a selectable gene and replica- B R6-5 Association tion capabilities on a single EcoRI-generated fragment. We found that the tetracycline A F DNA Ligase resistance plasmid that Chang and I had G isolated during our mechanical shearing experiments—a replicon we named pSC101 (Fig. 3, Top, lane c and Fig. 3, Middle)—had Plasmid DNA Transformation exactly these properties. Using pSC101 as Isolation and + a vector, we were able to identify the specific Characterization R6-5 DNA fragment that carries the kana- mycin resistance gene (Fig. 3, Bottom). Bacterial CaCl -Treated By early May, we had shown that our A Colonies 2 G Bacteria cloning results were reproducible, and we met to decide on the figures for the manu- D script we would be preparing. I outlined the ’ Fig. 2. Schematic diagram of the strategy used for construction of biologically functional plasmids (1). R6-5 plasmid paper s format in a notebook that sometimes DNA fragments generated by cleavage using the EcoRI endonuclease were allowed to associate randomly in vitro and has been referred to as my “laboratory note- were then covalently joined by DNA ligase. DNA in the resulting mixture was introduced into calcium chloride-treated book” (57) but which was used for jotting E. coli, and bacterial colonies expressing individual antibiotic resistance phenotypes encoded by R6-5 were selected on down ideas and future plans rather than for media containing antibiotics. Plasmid constructs isolated from these E. coli clones contained DNA fragments carrying specific resistance genes. recording experimental data. The paper (1) was completed in early June and, after being modified to address small points raised by that prevent propagation of the molecules? DNA, did bacterial transformations and These multiple issues led Falkow, who to- selection, and characterized the products by gether with Charles Brinton had participated heteroduplex analysis and ultracentrifuga- in the discussion and who envisioned the tion in gradients. Data were analyzed at possibility of isolating an enteric bacterial both places, and results were discussed toxin gene he had been studying using the between laboratories almost daily. I’d procedure that Boyer and I had just sketched arrive in the laboratory early in the morn- out, to remark, “If it works, let me know” (55). ing to look at the culture plates when col- A senior Stanford colleague whom I spoke onies produced by cells plated late the with after my return to Palo Alto was con- previous evening were still tiny. I often siderably less sanguine, proffering the opinion wished that the bacteria would grow faster that nothing interpretable was likely to come so that we could obtain results sooner. from the “messy” experimental design. Annie lived in San Francisco and carried We began the experiments shortly after the materials between Stanford and UCSF. new year. They went more smoothly than we We’d hurry to isolate plasmid DNA so could have hoped, and by March 1973 we that she could carry some of it to Herb’s had demonstrated the feasibility of the DNA laboratory for gel analysis the next day. It cloning approach that Boyer and I had out- was an extraordinarily exciting time for lined a few months earlier on delicatessen all of us. Fig. 3. DNA analysis in the initial DNA cloning experi- (Top) napkins. During a visit to the Cold Spring By introducing a mixture of ligated EcoRI- ments. Agarose-gel electrophoresis of (lane a) the pSC102 plasmid containing three of the multiple EcoRI- Harbor Laboratory that winter to give a generated R6-5 DNA fragments into E. coli, generated fragments of R6-5 DNA (lane b). Lane c shows seminar talk, Herb learned about the still we recovered a plasmid that expressed kana- that EcoRI cleavage of the pSC101 vector produces unpublished agarose gel electrophoresis/ mycin resistance but not the other resistance a single DNA fragment of the expected size. (Middle) fi DNA staining method that Phillip Sharp, Bill genes of R6-5. This replicon included only Electron photomicrograph of pSC101, the rst plasmid used successfully as a vector for DNA cloning. (Bottom) Sugden, and Joseph Sambrook had developed three of the DNA fragments characteristic Agarose gel electrophoresis showing cloning of the to separate and visualize fragments of DNA of the parental plasmid (Fig. 3, Top,lanesaand kanamycin resistance gene of R6-5: (lane d) EcoRI- generated by restriction enzymes (56); this b). Further analysis indicated that one of cleaved DNA of the pSC101 plasmid vector, (lane c) advance offered a hugely important addi- these fragments encoded functions and sites EcoRI-generated fragments of a novel plasmid (pSC102) that had been constructed from R6-5 (see Top) and that tion to the centrifugation and heteroduplex necessary for autonomous DNA replication expresses the kanamycin resistance determinant of the methods we were using to analyze plasmids. but contained no detectable resistance gene; parental R6-5 replicon, (lane b) mixture of the DNAs In the collaboration, Herb’s laboratory puri- a second fragment lacking the capability for shown in lanes c and d, and (lane a) EcoRI-generated fied the restriction endonuclease we used autonomous replication, but carrying a kana- fragments of a novel plasmid (pSC105) expressing both the tetracycline resistance gene of the pSC101 vector and and characterized plasmid DNA in ethi- mycin resistance gene, had been attached in the kanamycin resistance gene, which had been cloned dium bromide-stained agarose gels. My the fragment mixture to the replication re- from pSC102 by attaching it to pSC101. Top and Bottom laboratory isolated and purified plasmid gion, and, during bacterial transformation are from ref. 1.

Cohen PNAS Early Edition | 5of9 Downloaded by guest on September 27, 2021 reviewers, was communicated to PNAS by cleaved pI258 DNA and pSC101 DNA using been accepted for publication. His pre- Academy member Norman Davidson. the EcoRI enzyme, characterized, and then sentation prompted Bill Sugden, one of the Herb and I had recognized that small an- combined, the DNA fragments, introduced inventors of agarose gel electrophoresis at- tibiotic resistance plasmids such as pSC101 the ligated mixture into calcium chloride- tending the Gordon Conference to comment, might enable the cloning of eukaryotic cell treated E. coli, and selected bacterial colo- “well, now we can put together any DNAs we genes in E. coli andincludedthisstatement nies that expressed both the ampicillin want to” (59). The following morning on the in our paper’s Discussion section. However, resistance of pI258 and the tetracycline re- last day of the meeting, a special session was which eukaryotic DNA should be used to test sistance encoded by pSC101 (Fig. 4, Upper). called by cochairs Maxine Singer and Dieter the notion? We couldn’tspecifically select Buoyant density ultracentrifugation (Fig. 4, Söll to discuss the implications of the data bacteria containing cloned eukaryotic DNA, Lower), agarose gel electrophoresis, and elec- that Boyer had shown. Seven months earlier, and only a few cellular genes of eukaryotes had tron microscope heteroduplex analysis the ability of EcoRI to generate cohesive DNA been purified and characterized well enough showed that the E. coli colonies contained ends that could be used for joining DNAs to identify them unambiguously in bacterial plasmids that included DNA sequences from biochemically had been reported; however, isolates. Besides, there were other experiments the two species (2). now presented with evidence that EcoRI- that each of us wanted to do, and our priorities generated, ordinarily nonreplicating DNA were different. Herb was eager to use a rapidly Gordon Conference Discussions About fragments can actually be propagated in expanding collection of restriction enzymes Biohazard Concerns and Cloning of bacteria by attaching them to plasmid DNA— to construct high copy-number vectors that Eukaryotic DNA in Bacteria and that hybrid DNAs created in this way offer increased flexibility of cloning sites. The Sometime in late May or early June 1973, are biologically functional—the attendees “ most pressing issue for me was to learn Boyer received an invitation to give an in- were concerned that hybrids with bio- whether DNA hybrids containing very dif- formal talk at the Gordon Research Confer- logical activity of unpredictable nature may ” ferent components derived from unre- ence on Nucleic Acids scheduled for mid- eventually be created (60). They voted to lated species could be propagated and cloned. June. He described his laboratory’swork communicate this concern to the presidents Our laboratories exchanged experimental at a session on “Bacterial Enzymes in the of the National Academy of Sciences (NAS) tools and set out separately to address our Analysis of DNA” and, near the end of his and its Institute of Medicine, suggesting different priorities. presentation, reported results from our col- that a study committee be established to con- sider the issue and to “recommend specific “ ” laborative experiments, which had not yet Testing of Interspecies Barriers actions or guidelines.” Design of specific experiments that I believed A discussion at the same Gordon Confer- could provide an initial test of the hypothe- ence between Boyer and John Morrow, who S. aureus sized barriers to interspecies gene transfer pl258 had completed his Ph.D. thesis project in DNA PcR(ApR) began after completion of the Cohen, Chang, Extraction EcoRI site Paul Berg’s laboratory at Stanford but had

Boyer, and Helling manuscript. Richard EcoRI Cleavage pSC101 TcR not yet moved to a postdoctoral position with Novick of the Public Health Research In- ApR + Donald Brown at the Carnegie Institution of stitute in New York and his colleagues had Washington laboratory in Baltimore, MD, EcoRI cleavage R Ap plus ligation ’ described an 18-kb plasmid named pI258 EcoRI EcoRI led to a second collaboration between Boyer s site site fi (58) that replicates autonomously in Staph- pSC112 Transformation of laboratory and mine. Brown had puri ed and yloccus aureus, but not in E. coli. pI258 had TcR E. coli K12 Selection for ApR characterized the ribosomal genes of the been shown to carry a β-lactamase gene en- African frog, Xenopus laevis,andMorrow coding resistance to penicillins, and such pSC112 pSC112 Untreated EcoRI Treated had found that this DNA was cleaved by the resistance might be used to select E. coli =1.679 EcoRI enzyme preparation that Boyer had transformants carrying hybrid plasmids =1.700 =1.691 providedforanalysisofSV40viralDNA. =1.679 expressing β-lactamase. Whether DNAs Morrow and Boyer discussed trying to clone known to be highly disparate in nucleotide =1.710 EcoRI-generated fragments of frog ribosomal composition (11) and taken from microbes DNA using the approach that Boyer’slab as different as the Gram-positive coccus and mine had employed to clone plasmid S. aureus and the Gram-negative rod-shaped DNA fragments. Brown agreed to allow the E. coli could be propagated as part of the DNA he had given to Morrow to be used for same replicon and whether the staphylo- Fig. 4. Cloning of S. aureus plasmid DNA in E. coli. the attempt. However, how to identify cloned coccal gene would be expressed in the new (Upper) Schematic diagram of strategy used for testing ribosomal RNA? When Herb returned to the viability of interspecies DNA hybrids (2). DNA of the host was questionable. However, if these pI258 plasmid, which carries a β lactamase gene en- UCSF,hephonedmetodiscussthisquestion events occurred, the density gradient analysis coding resistance to penicillins in S. aureus was cleaved andtoinvitemyparticipationintheproposed methods that Miller and I had used earlier by EcoRI endonuclease and mixed with similarly cleaved project. We agreed that multiple parameters (24) would aid in establishing the origin of DNA of the pSC101 vector encoding tetracycline re- would be needed to show unambiguously that DNAsegmentsthatdifferinA+T/G+Cra- sistance. After ligation, the mixture was introduced into DNA from another biological kingdom was E. coli cells, and colonies that expressed both resistance tio. E. coli cells resistant to both penicillin/ phenotypes were identified. (Lower) Centrifugation analy- being propagated in bacteria. Although there ampicillin and tetracycline were already sis in isopycnic density gradient of plasmid DNA (pSC112) were no phenotypic properties that would highly prevalent, so combining pI258 and isolated from an E. coli clone expressing both resistances enable bacterial colonies that acquired plas- pSC101 DNAs would not produce a novel and showing DNA species that band at buoyant densi- mids carrying ribosomal DNA inserts to be ties characteristic of E. coli (ρ = 1.710) and S. aureus resistance combination. (ρ = 1.68–1.69) DNAs and reflect the distinctly different selected, ribosomal genes, which were known The experiments themselves were not com- A+T/G+C nucleotide ratios of these unrelated bacterial to be extraordinarily rich in G+C base pairs plicated and the results were conclusive. We species. Lower is from ref. 2. (61), could be distinguished from E. coli

6of9 | www.pnas.org/cgi/doi/10.1073/pnas.1313397110 Cohen Downloaded by guest on September 27, 2021 DNA by buoyant density differences during had been attempting unsuccessfully to clone DNA molecules” and that “the plasmid is PERSPECTIVE isopycnic centrifugation, as well as by frag- Drosophila melanogaster DNA by using unstable and readily lost from its E. coli ment size during agarose gel analysis and by λdv, a nonlytic phage variant that Kenichi host” (67). It was later learned that insertion electron microscopy heteroduplex analysis. Matsubara and Dale Kaiser had shown can of foreign DNA into the λdvgal site that the I thought that it might be necessary to screen replicate autonomously in E. coli (62). Berg team had used for construction of hybrid a large number of individual bacterial colonies Attaching the Drosophila DNA to the DNA molecules (52, 54) disrupts a gene es- to findrecombinantplasmids,butitwas pSC101 plasmid enabled Hogness and his sential for λdvgal replication in bacteria (54, worth a try. coworkers to confirm in late 1974 the cloning 66, 69), possibly explaining the lack of success Morrow and Chang carried out the bulk of of eukaryotic cell genes in bacteria and to use of the DNA cloning attempts reported in the experiments, which were done largely in DNA cloning to map sequences in Drosophila Mertz’s dissertation (67). But Pollack’s ’ my laboratory at Stanford during the late chromosomes (63). Concurrently, other lab- concerns and Berg s decision had impor- summer and early fall of 1973—still several oratories focused on constructing new bacte- tantly raised awareness about possible bio- months before the November publication of riophage λ variants able to produce viable hazardous consequences of creating novel results from the initial collaboration with molecular hybrids (64, 65), and, in November DNA combinations (54). Boyer. As data began to accumulate to sup- 1974, Ron Davis and his coworkers at Stan- The committee that Berg formed to ad- port the conclusion that eukaryotic DNA ford reported that such hybrids can be pro- dress the biohazard concerns of the Gordon can actually be propagated in bacteria, our pagated as plaque-forming phage (66). Conference participants consisted mostly of examination of grids prepared for electron experts on oncogenic viruses, and it initially Biohazard Speculations Mount focused on issues related to the introduction microscope heteroduplex analysis removed In response to the vote of attendees at the any remaining uncertainty (Fig. 5): hetero- of mammalian cell virus sequences into bac- nucleic acids Gordon Conference and the teria. However, during discussions by the duplexes formed between the ribosomal RNA consequent letter from Singer and Söll, NAS group, its focus expanded to address the taken from frogs and recombinant plasmids President Phillip Handler chose Paul Berg possible construction of novel resistance- isolated from E. coli revealed homology at to form a committee to evaluate possible gene combinations, and Herb Boyer, David two locations that were spaced at the distance biohazardous consequences of constructing Hogness, Ron Davis, and I were invited to that Brown’s laboratory had shown to be the hybrid DNAs. Berg had thought deeply participate in the formulation of the com- spacing between sequence repeats present on about this issue and was a perfect choice: mittee’s final recommendations. These ribosomal RNA genes (3). And additional using complementary ends installed onto recommendations, which were published DNA/RNA hybridization experiments indi- DNA and further manipulations that Berg concurrently in July 1974 in PNAS, Science, cated that transcripts capable of interacting and his postdoctoral fellows David Jackson and Nature as a letter entitled, “Potential with ribosomal RNA genes of the frog were and Robert Symons credited to Dale Kai- Biohazards of Recombinant DNA Mole- produced in E. coli. ser’s graduate student Peter Lobban, Berg’s cules,” proposed a voluntary moratorium The news that eukaryotic DNA can be laboratory had biochemically attached DNA on the introduction of resistance genes into fi cloned and ampli ed in bacteria spread im- of the SV40 tumor virus to a version of bacterial species that do not already express fi mediately in the scienti c community, and the bacteriophage λdv replicon that in- that type of resistance and on the linkage of requests for the pSC01 plasmid began to arrive cludes the galactose (gal) operon of E. coli animal virus genes to plasmids (70–72). fi at my laboratory. The rst to receive pSC101 (52). Berg later received the 1980 Nobel An article by New York Times journalist was David Hogness, a distinguished Stanford Prize in Chemistry in recognition of “his Victor McElheny, who learned about our Department of Biochemistry professor who fundamental studies of the biochemistry DNA cloning experiments from Berg et al. of nucleic acids, with particular regard to committee member David Baltimore, ap- recombinant DNA.” Berg and his graduate peared in the Times a few weeks before re- student Janet Mertz planned to introduce lease of the committee’s moratorium proposal these SV40-λdvgal hybrid DNA molecules (73).Inthisarticle,whichwasheadlined, into mammalian cells to determine whether “Animal Gene Shifted to Bacteria; Aid Seen to the bacterial gene would function there (54, Medicine and Farm,” McElheny and the sci- 67), and he and Mertz have written that they entists he interviewed spoke optimistically also wished to propagate the SV40-λdvgal about the potential benefits of DNA cloning, hybrid DNA molecules in E. coli (67,68). which was by then increasingly referred to as However, at a Cold Spring Harbor Labora- “recombinant DNA technology.” However, tory summer course in 1971 where Mertz a press conference arranged by the NAS to described her proposed experiments, bi- announce the moratorium proposed by the ologist Robert Pollack raised biohazard Berg et al. letter resulted in an abrupt shift of concerns about the possibility of creating public focus to biohazard issues. The notion Fig. 5. Electric photomicrograph of heteroduplex oncogenic E. coli by such experiments, and that prompted the shift: “if the researchers X. laevis showing homology between DNA isolated from Berg was persuaded to forego attempts at themselves are concerned, then the dangers oocytes and plasmid DNA isolated from bacteria and fi ” containing fragments of ribosomal RNA genes that had cloning the biochemically spliced SV40- must be truly horri c. The unprecedented been cloned by attaching the eukaryotic cell DNA to the λdvgal DNA molecules in either eukaryotic effort of scientists to restrict their own re- pSC101 vector. (A) Single strand of X. laevis rDNA. (B) cells or bacteria (54, 68). Ironically with regard search in order to guard against hazards that Double-stranded regions of homology between X. laevis to Pollack’s scenario, Mertz’s1975PhDdis- were not known to exist was so novel that this rDNA and plasmid DNA isolated from E. coli.(C), Single- “ fi strand DNA regions corresponding in length to the sertation (67) stated that scienti c problems effort was widely interpreted as implying pSC101 vector, which shares no homology with X. laevis have been encountered during attempts to that danger was likely. A more extensive rDNA. (Scale bar: 1 μm.) Figure is from ref. 3. use λdvgal as a vector for replicating other personal perspective on the Berg et al. letter,

Cohen PNAS Early Edition | 7of9 Downloaded by guest on September 27, 2021 the Asilomar Conference that it led to, and applicable to multiple eukaryotic and pro- that have occurred during a four-decade the post-Asilomar period of interaction be- karyotic hosts. Immunologically reactive (79– time frame seem unprecedented, and the tween scientists and legislators is provided in 81) and then biologically functional (82) consequent growth of knowledge in bi- the oral history I recorded with University of eukaryotic proteins were produced in bac- ology and chemistry has been almost loga- California Berkeley historian Sally Smith teria. Specialized vectors were developed by rithmic. However, although this article has Hughes for the Bancroft Library in the mid- academic and industrial laboratories to ex- been retrospective, in reality, the accelerated 1990s (74). press human proteins such as insulin and scientific journey that has resulted from growth hormone in E. coli, to produce vac- the ability to clone DNA has only begun. Stanford/UCSF DNA Cloning Patents cines containing antigens expressed from McElheny’s upbeat article in May 1974 was cloned genes, and to identify genetic reg- ACKNOWLEDGMENTS. I thank Herb Boyer, Annie read by Niels Reimers, whose job as Director ulatory signals using reporter genes. To- Chang, and Bob Helling for their partnership in the fi fi of the Of ce of Technology Licensing at gether, these advances provided a foundation scienti c collaboration that resulted in successful DNA Stanford was to help fund the university’s cloning, and both Annie Chang and Christine Miller for for the creation of biotechnology companies. the other contributions they made to my laboratory and academic programs by promoting the li- Herb Boyer and entrepreneur Robert A. to science over decades. I also thank the many scientists censing of inventions made at the university. whose prior research made possible the experiments I’ve Swanson founded an enterprise that became described, especially plasmid biologists/microbiologists The day after the article appeared, I received preeminent among these companies: Gen- Stanley Falkow, Donald Helinski, and Richard Novick; a telephone call from Reimers indicating that entech. Efficient DNA sequencing methods biochemists Martin Gellert, Robert Lehman, Jerard Hur- he wanted to discuss patenting the technol- invented by Allan Maxam and Walter Gil- witz, and Phillip Sharp; geneticists/microbiologists Dale fi Kaiser, Peter Lobban, Ronald Davis, and Vittorio Sgara- ogy that Boyer and I had invented. My rst bert (83) and by Frederick Sanger and his mella; and biologist James D. Watson. The contributions fi reaction was quite negative. Could ndings colleagues (84) dramatically facilitated and support of now-deceased friends and colleagues, of basic research funded by the public be especially Joshua Lederberg, Norton Zinder, Jerome analysis of cloned DNA, and, together Vinograd, and Norman Davidson, have also been of patented, and should they be? I told him with the invention of the PCR by Kary major importance to me, both scientifically and person- that our work depended on years of funda- ally. I apologize to those whose relevant contributions Mullis (85), information that DNA se- have not been specifically mentioned. For more extensive mental research on plasmid biology by many quencing yielded about the structure and information about my perspective of the events I’ve related laboratories and on properties of DNA, function of cloned genes (86) led to the here, I refer readers to my Bancroft Library oral history DNA ligase, and restriction enzymes that fi (74), and for the perspectives of others, to the Bancroft birth of the eld of genomics. The number Library oral histories of Herbert W. Boyer (87), Paul Berg had been discovered by others. And would and breadth of the scientific discoveries (54), and Niels Reimers (75). a patent adversely affect advancement of the science? Reimers pointed out that prior knowledge is a pillar for every invention and 1 Cohen SN, Chang ACY, Boyer HW, Helling RB (1973) Construction 18 Anderson ES, Lewis MJ (1965) Characterization of a transfer that a well-honed legal process determines of biologically functional bacterial plasmids in vitro. Proc Natl Acad factor associated with drug resistance in Salmonella typhimurium. whether a particular advance is novel and Sci USA 70(11):3240–3244. Nature 208(5013):843–849. 2 Chang ACY, Cohen SN (1974) Genome construction between 19 Watanabe T, Fukasawa T (1961) Episome-mediated transfer of patentable, as well as the validity of the bacterial species in vitro: Replication and expression of drug resistance in Enterobacteriaceae. I. Transfer of resistance factors inventorship claimed in the application. He Staphylococcus plasmid genes in Escherichia coli. Proc Natl Acad Sci by conjugation. J Bacteriol 81:669–678. explained that only commercial entities USA 71(4):1030–1034. 20 Cohen SN, Maitra U, Hurwitz J (1967) Role of DNA in RNA 3 Morrow JF, et al. (1974) Replication and transcription of eukaryotic synthesis. XI. Selective transcription of gamma DNA segments in would pay royalties, that a patent would not DNA in Escherichia coli. Proc Natl Acad Sci USA 71(5):1743–1747. vitro by RNA polymerase of Escherichia coli. J Mol Biol 26(1): impede noncommercial use of DNA cloning 4 Zinder ND, Lederberg J (1952) Genetic exchange in Salmonella. J 19–38. methods, and that funds received by Stanford Bacteriol 64(5):679–699. 21 Campbell AM (1963) Episomes. Adv Genet 11:101–145. 5 Lewis AM, Jr., Prigge KO, Rowe WP (1966) Studies of adenovirus- 22 Jacob F, Brenner S, Cuzin F (1963) On the regulation of DNA and UCSF would aid research programs at SV40 hybrid viruses. IV. An adenovirus type 2 strain carrying the replication in bacteria. Cold Spring Harb Symp Quant Biol these institutions. I discussed Reimer’s pro- infectious SV40 genome. Proc Natl Acad Sci USA 55(3):526–531. 28:329–348. posal with Herb, and together we agreed to 6 Rowe WP, Pugh WE (1966) Studies of adenovirus-SV40 hybrid 23 Hickson FT, Roth TF, Helinski DR (1967) Circular DNA forms of viruses. V. Evidence for linkage between adenovirus and SV40 genetic a bacterial sex factor. Proc Natl Acad Sci USA 58(4):1731–1738. let our universities proceed with applications materials. Proc Natl Acad Sci USA 55(5):1126–1132. 24 Cohen SN, Miller CA (1969) Multiple molecular species of circular for patents that eventually had 461 licensees 7 Chargaff E (1977) Research with Recombinant DNA: An Academy R-factor DNA isolated from Escherichia coli. Nature 224(5226): – before their expiration in 1997. Reimers’ oral Forum March 7–9, 1977 (National Academy of Sciences, 1273 1277. Washington, DC), pp 56–57. 25 Davis RW, Davidson N (1968) Electron-microscopic visualization history is a source of further information 8 Sinsheimer RL (1977) Recombinant DNA. Annu Rev Biochem of deletion mutations. Proc Natl Acad Sci USA 60(1):243–250. about the events that led to these patents (75). 46:415–438. 26 Westmoreland BC, Szybalski W, Ris H (1969) Mapping of 9 Arber W (1965) Host-controlled modification of bacteriophage. deletions and substitutions in heteroduplex DNA molecules of Annu Rev Microbiol 19:365–378. bacteriophage lambda by electron microscopy. Science 163(3873): Converting Promise into Reality 10 Falkow S (1965) Nucleic acids, genetic exchange and bacterial 1343–1348. As experiments using DNA cloning pro- speciation. Am J Med 39(5):753–765. 27 Sharp PA, Cohen SN, Davidson N (1973) Electron microscope 11 Marmur J, Falkow S, Mandel M (1963) New approaches to heteroduplex studies of sequence relations among plasmids of cedures proceeded without adverse incident bacterial taxonomy. Annu Rev Microbiol 17:329–372. Escherichia coli. II. Structure of drug resistance (R) factors and F in laboratories around the world, biohazard 12 Lederberg J, Cavalli LL, Lederberg EM (1952) Sex compatibility in factors. J Mol Biol 75(2):235–255. fears dissipated (74). And multiple scientific Escherichia coli. Genetics 37(6):720–730. 28 Sharp PA, Hsu MT, Otsubo E, Davidson N (1972) Electron 13 Harada K, Kameda M, Suzuki M, Mitsuhashi S (1964) Drug microscope heteroduplex studies of sequence relations among advances in these laboratories began to turn resistance of enteric bacteria. 3. Acquisition of transferability of plasmids of Escherichia coli. I. Structure of F-prime factors. J Mol Biol the early promise of DNA cloning for pro- nontransmissible R(Tc) factor in cooperation with F factor and 71(2):471–497. ducing fundamental knowledge and practical formation of Fr(Tc). J Bacteriol 88:1257–1265. 29 Cohen SN (1976) Transposable genetic elements and plasmid 14 Watanabe T, Ogata C, Sato S (1964) Episome-mediated transfer evolution. Nature 263(5580):731–738. applications into reality. Better strategies for of drug resistance in Enterobacteriaceae. 8. Six-drug-resistance R 30 Starlinger P, Saedler H (1976) IS-elements in microorganisms. detecting cloned genes and their products factor. J Bacteriol 88:922–928. Curr Top Microbiol Immunol 75:111–152. 15 Lederberg J (1952) Cell genetics and hereditary symbiosis. Physiol 31 Mandel M, Higa A (1970) Calcium-dependent bacteriophage were devised (76, 77), and methods were Rev 32(4):403–430. DNA infection. J Mol Biol 53(1):159–162. soon developed for cloning enzymatically 16 Falkow S, Citarella RV, Wohlhieter JA (1966) The molecular 32 Cosloy SD, Oishi M (1973) Genetic transformation in Escherichia generated DNA sequences complementary nature of R-factors. J Mol Biol 17:102–116. coli K12. Proc Natl Acad Sci USA 70(1):84–87. 17 Rownd R, Nakaya R, Nakamura A (1966) Molecular nature of the 33 Cohen SN, Chang ACY, Hsu L (1972) Nonchromosomal antibiotic to mRNA (78). The strategies we had used drug-resistance factors of the Enterobacteriaceae. J Mol Biol 17(2): resistance in bacteria: Genetic transformation of Escherichia coli by for cloning DNA in E. coli were shown to be 376–393. R-factor DNA. Proc Natl Acad Sci USA 69(8):2110–2114.

8of9 | www.pnas.org/cgi/doi/10.1073/pnas.1313397110 Cohen Downloaded by guest on September 27, 2021 34 Cohen SN, Chang ACY (1973) Recircularization and autonomous Circular SV40 DNA molecules containing lambda phage genes and 71 Anonymous (1974) NAS ban on plasmid engineering. Nature PERSPECTIVE replication of a sheared R-factor DNA segment in Escherichia coli the galactose operon of Escherichia coli. Proc Natl Acad Sci USA 250:175. transformants. Proc Natl Acad Sci USA 70(5):1293–1297. 69(10):2904–2909. 72 Berg P, et al. (1974) Letter: Potential biohazards of recombinant 35 53 Cohen SN, Chang AC (1977) Revised interpretation of the origin Lobban PE, Kaiser AD (1973) Enzymatic end-to-end joining of DNA molecules. Science 185(4148):303. – – of the pSC101 plasmid. J Bacteriol 132(2):734 737. DNA molecules. J Mol Biol 78(3):453 471. 73 McElheny VK (May 20, 1974) Animal gene shifted to bacteria; 36 Smith HO, Wilcox KW (1970) A restriction enzyme from 54 Berg P (2000) AStanfordProfessor’sCareerinBiochemistry, aid seen to medicine and farm. NY Times, p 61. Hemophilus influenzae. I. Purification and general properties. J Mol Science Politics, and the Biotechnology Industry (typescript of an oral 74 Cohen SN (2009) Science, Biotechnology, and Recombinant Biol 51(2):379–391. history conducted in 1997 by Sally Smith Hughes) (Regional Oral History DNA: A Personal History (typescript of an oral history conducted in 37 Danna K, Nathans D (1971) Specific cleavage of simian virus 40 Office, The Bancroft Library, University of California, Berkeley, CA). fi DNA by restriction endonuclease of Hemophilus influenzae. Proc Natl 55 Falkow S (1977) Oral history with Charles Weiner, MIT Oral 1995 by Sally Smith Hughes) (Regional Oral History Of ce, The Acad Sci USA 68(12):2913–2917. History Program Recorded on May 29, 1976 and Feb 26, 1977 (MIT Bancroft Library, University of California, Berkeley, CA). 75 ’ fi 38 Takano T, Watanabe T, Fukasawa T (1966) Specific inactivation of Institute Archives, Cambridge, MA). Reimers N (1998) Stanford sOf ce of Technology Licensing and infectious lambda DNA by sonicates of restrictive bacteria with R 56 Sharp PA, Sugden B, Sambrook J (1973) Detection of two the Cohen/Boyer Cloning Patents (typescript of an oral history factors. Biochem Biophys Res Commun 25(2):192–198. restriction endonuclease activities in Haemophilus parainfluenzae conducted in 1997 by Sally Smith Hughes) (Regional Oral History Office, 39 Takano T, Watanabe T, Fukasawa T (1968) Mechanism of host- using analytical agarose—ethidium bromide electrophoresis. The Bancroft Library, University of California, Berkeley, CA), p 85. controlled restriction of bacteriophage lambda by R factors in Biochemistry 12(16):3055–3063. 76 Grunstein M, Hogness DS (1975) Colony hybridization: A – 57 fi Escherichia coli K12. Virology 34(2):290 302. Sharrer T (October 1, 2006) Recombinant DNA: The rst report. method for the isolation of cloned DNAs that contain a specific gene. 40 Yoshimori R, Roulland-Dussoix D, Boyer HW (1972) R factor- The Scientist, p100. Proc Natl Acad Sci USA 72(10):3961–3965. controlled restriction and modification of deoxyribonucleic acid: 58 Peyru G, Wexler LF, Novick RP (1969) Naturally occurring 77 Kedes LH, Chang ACY, Houseman D, Cohen SN (1975) Isolation Restriction mutants. J Bacteriol 112(3):1275–1279. penicillinase plasmids in Staphylococcus aureus. J Bacteriol 98(1): of histone genes from unfractionated sea urchin DNA by subculture 41 Fareed GC, Garon GF, Salzman NP (1972) Origin and direction of 215–221. – simian virus 40 deoxyribonucleic acid replication. J Virol 10(3): 59 Lear J (1978) Recombinant DNA: The Untold Story (Crown cloning in E. coli. Nature 255(5509):533 538. 78 484–491. Publishers, New York). Efstratiadis A, Kafatos FC, Maxam AM, Maniatis T (1976) 42 Morrow JF, Berg P (1972) Cleavage of Simian virus 40 DNA at 60 Singer M, Soll D (1973) Guidelines for DNA hybrid molecules. Enzymatic in vitro synthesis of globin genes. Cell 7(2):279–288. a unique site by a bacterial restriction enzyme. Proc Natl Acad Sci Science 181(4105):1114. 79 Goeddel DV, et al. (1979) Expression in Escherichia coli of USA 69(11):3365–3369. 61 Davison PF (1966) Isopycnic centrifugation for the isolation of chemically synthesized genes for human insulin. Proc Natl Acad Sci 43 Hershey AD, Burgi E, Ingraham L (1963) Cohesion of DNA DNA strands coding for ribosomal RNA. Science 152(3721):509–512. USA 76(1):106–110. molecules isolated from phage lambda. Proc Natl Acad Sci USA 62 Matsubara K, Kaiser AD (1968) Lambda dv: An autonomously 80 Itakura K, et al. (1977) Expression in Escherichia coli of 49(5):748–755. replicating DNA fragment. Cold Spring Harb Symp Quant Biol a chemically synthesized gene for the hormone somatostatin. Science 44 – Gellert M (1967) Formation of covalent circles of lambda DNA by 33:769 775. 198(4321):1056–1063. – 63 E. coli extracts. Proc Natl Acad Sci USA 57(1):148 155. Wensink PC, Finnegan DJ, Donelson JE, Hogness DS (1974) A 81 Villa-Komaroff L, et al. (1978) A bacterial clone synthesizing 45 Gefter ML, Becker A, Hurwitz J (1967) The enzymatic repair of system for mapping DNA sequences in the chromosomes of proinsulin. Proc Natl Acad Sci USA 75(8):3727–3731. DNA. I. Formation of circular lambda-DNA. Proc Natl Acad Sci USA Drosophila melanogaster. Cell 3(4):315–325. 82 Chang ACY, et al. (1978) Phenotypic expression in E. coli of 58(1):240–247. 64 Hershfield V, Boyer HW, Yanofsky C, Lovett MA, Helinski DR a DNA sequence coding for mouse dihydrofolate reductase. Nature 46 Kaiser AD, Wu R (1968) Structure and function of DNA cohesive (1974) Plasmid ColEl as a molecular vehicle for cloning and – ends. Cold Spring Harb Symp Quant Biol 33:729–734. amplification of DNA. Proc Natl Acad Sci USA 71(9):3455–3459. 275(5681):617 624. 83 47 Mertz JE, Davis RW (1972) Cleavage of DNA by R 1 restriction 65 Murray NE, Murray K (1974) Manipulation of restriction targets Maxam AM, Gilbert W (1977) A new method for sequencing endonuclease generates cohesive ends. Proc Natl Acad Sci USA in phage lambda to form receptor chromosomes for DNA fragments. DNA. Proc Natl Acad Sci USA 74(2):560–564. 69(11):3370–3374. Nature 251(5475):476–481. 84 Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with 48 Sgaramella V (1972) Enzymatic oligomerization of bacteriophage 66 Thomas M, Cameron JR, Davis RW (1974) Viable molecular chain-terminating inhibitors. Proc Natl Acad Sci USA 74(12): P22 DNA and of linear Simian virus 40 DNA. Proc Natl Acad Sci USA hybrids of bacteriophage lambda and eukaryotic DNA. Proc Natl 5463–5467. 69(11):3389–3393. Acad Sci USA 71(11):4579–4583. 85 Saiki RK, et al. (1988) Primer-directed enzymatic amplification of 49 67 Hedgpeth J, Goodman HM, Boyer HW (1972) DNA nucleotide Mertz JE (1975) Deletion mutants of simian virus 40. PhD DNA with a thermostable DNA polymerase. Science 239(4839): sequence restricted by the RI endonuclease. Proc Natl Acad Sci USA dissertation (Stanford University, Stanford, CA). 487–491. – 68 fl 69(11):3448 3452. Berg P, Mertz JE (2010) Personal re ections on the origins and 86 Nakanishi S, et al. (1979) Nucleotide sequence of cloned cDNA 50 Khorana HG (1966-1967) Polynucleotide synthesis and the emergence of recombinant DNA technology. Genetics 184(1):9–17. for bovine corticotropin-beta-lipotropin precursor. Nature 278(5703): genetic code. Harvey Lect 62:79–105. 69 Helling RB, Goodman HM, Boyer HW (1974) Analysis of 423–427. 51 Jensen RH, Wodzinski RJ, Rogoff MH (1971) Enzymatic addition endonuclease R-EcoRI fragments of DNA from lambdoid 87 Boyer HW (2001) Recombinant DNA Research at UCSF and of cohesive ends to T7 DNA. Biochem Biophys Res Commun 43(2): bacteriophages and other viruses by agarose-gel electrophoresis. J 384–392. Virol 14(5):1235–1244. Commercial Application at Genentech (typescript of an oral history 52 Jackson DA, Symons RH, Berg P (1972) Biochemical method for 70 Anonymous (1974) Potential biohazards of recombinant DNA conducted in 1994 by Sally Smith Hughes) (Regional Oral History inserting new genetic information into DNA of Simian Virus 40: molecules. Proc Natl Acad Sci USA 71(7):2593–2594. Office, The Bancroft Library, University of California, Berkeley, CA).

Cohen PNAS Early Edition | 9of9 Downloaded by guest on September 27, 2021