Tobacco Documents Online

® rU

. 7: -1324 SCIENTIFIC OFFPRINTS AMERICAN . _ N. 0 o° SCIENTIFIC AMERICAN ~ JULY 1975 (Q C VOI. 233, NO. l PP. 24-33 Q [` WM W ~ .. ~ PUBLISHED BY W. H. FREEMAN AND COMPANY 660 MARKET'STREET, SAN FRANCISCO, CALIFORNIA 94104 Copyright0 1975 byScientifioAmeriasn, Inc. Atl rights reserved. Printed in the U.S.A. No part of this oNprintmay be reproduced by.any mechanicaf photographic or electronic process .or . In the form of.a phonographic recording, nor may it be stored in a retrieval system, transmitted or otherwise.copied for public or private.use without written permission of the.publisher. . . .. .. .. . -_... .. . .. . .._ .......~-..iwt~

e -- e) , - . - 'S7 C m ~U'l . C ~ y ~ d ~ ° ~ o d p; a f~ 0 m b o p; p ~ m ~ r~ m 0 0 0 0

, a ~ C c , THE MANIPULATION OF GENES Techniques for cleaving DNA and splicing it into a carrier molecule make it possible to transfer genetic•information from one organism ~- to an unrelated one. There the DNA replicates and expresses itself '_ .~., _ a, . M ythology is full of hybrid crea- tures such as the Sphinx, the Minotaur and the Chimera, but the real world is not; it is populated by organisms that' have been shaped not by the union of characteristics derived from very dissimilar organisms but by evolu- tion within species that retain their basic identity generation after generation. This is because there are natural barriers that normally prevent the exchange of genet- ic information between unrelated orga- nisms. The barriers are still poorly un- derstood, but they are of fundamental biological importance. --The basic unit of biological related- ness is the species, and in organisms that reproduce sexually species are defined by the ability of their members to breed with one another. Species are deter- mined' and defined by the genes they carry, so that in organisms that repro- duce asexually the concept of species de- pends on nature's ability to prevent the biologically significant exchange of ge- netic material-the nucleic acid DNA- between unrelated groups. The persistence of genetic uniqueness is perhaps most remarkable in simple organisms such as bacteria. Even when they occupy the same habitat most bac- terial species do not exchange genetic in- formation. Even rather similar species,of bacteria do not ordinarily exchange the genes on their chromosomes, the struc- tures that carry most of their genetic information. There are exceptions, how- ever. There are bits of DNA, called plas- mids, that exist apart from the chromo- somes in some bacteria. Sometimes a plasmid can pick up a short segment' of DNA ff om the chromosome of its own cell and transfer it to the cell' of a relat- ed bacterial species, and sometimes the plasmid and the segment of chromosomal DNA can become integrated into the chromosome of the recipient cell. This transfer of genes between species by extrachromosomal elements has surely played some role in bacterial evolution, but apparently it has not been wide- spread in nature. Otherwise the char- acteristics of the common bacterial spe- cies would not have remained so largely intact' over the huge number of bacterial generations that have existed during thee era of modem bacteriology. In 1973 Annie C. Y. Chang and I at the Stanford University School of Medi- cine and Herbert W. Boyer and' Robert B. Helling at the University of California School of Medicine at San Francisco reported the construction in a test tube of biologically functional DNA mole- cules that combined genetic information from two different sources. We made the molecules by splicing together segments of two different plasmids found in the colon bacillus Escherichia coli and then inserting the composite DNA into E. coli' cells, where it replicated itself and ex- pressed the genetic information of both parent plasmids. Soon afterward we in- troduced plasmid genes from an unre- lated bacterial species, Staphylococcus aureus, into E. coli, where they too ex- pressed. the biological properties they had displayed in their original host; then,, applying the same procedures with John F. Morrow of Stanford' and Howard M. Goodman in San Francisco, we were able to insert into E. coli some genes from an animal: the toad Xenopus laevis. We called our composite molecules DNA chimeras, because they were con- ceptually similar to the mythological Chimera (a creature with the head of a lion, the body of a goat and the tail of a serpent) and were the molecular counter- parts of hybrid plant chimeras produced by agricultural grafting. The procedure we described has since been used and extended by workers in several labora- tories. It has been called plasmid en- gineering, because it utilizes plasmids to introduce the foreign genes, and molecu- lar cloning„because it provides a way to propagate a clone, or line of genetically alike organisms, all containing identical composite DNA molecules. Because of the method's potential for creating a wide variety of novel genetic combina- tions in microorganisms it is also known as genetic engineering and genetic ma- nipulation. The procedure actually con- sists of several distinct biochemical and biological manipulations that were made possible by a series of independent dis- coveries made in rapid succession in the late 1960's and early 1970's. There are four essential elements: a method' of breaking and joining DNA molecules de- rived from different sources; a suitable gene carrier that can replicate both itself and a foreign DNA segment linked to it; a means of introducing the composite DNA molecule, or chimera, into a func- tional bacteriali cell, and a method of selecting from a large population of cells a clone of recipient cells that has ac- quired the molecular chimera. In 1967 DNA ligases-enzymes that can repair breaks in DNA and under cer- tain conditions can join together the loose ends of DNA strands-were dis - covered almost' simultaneously im five laboratories. A DNA strand is a chain of nucleotides, each consisting of a deoxy- ribose sugar ring, a phosphate group and one of four organic bases: adenine,, thy- mine, guanine and cytosine. The sugars and phosphates form the backbone of the strand, from which the bases pro- ject. The individual nucleotide building blocks are connected by phosphodiester bonds between the carbon atom at posi- tion No. 3 on one sugar and the carbon atom at position No. 5 om the adjacent sugar. Double-strand DNA, the form found in most organisms, consists of two 3 _

I NIcKED DNA =HIM REPAIRED DNA ~ i U 0 D DNA LIGASE is an enzyme that repairs "nicks," or breaks in one strand of a double.strand molecule of DNA (top). A strand of DNA is a chain of nucleotides (bottom), each consist- ing of a deoxyribose sugar and a phosphate group and one of four organic bases: adenine (A), thymine (T), guanine (G) and cytosine (C). The sugars and phosphates constitute the backbone of the strand, and paired bases, linked by hydrogen bonds (broken black lines), connect two strands. The ligase catalyzes synthesis of a bond at the site of the break (broken colored line)' between the phosphate of one nucleotide and the sugar of the next nucleotide. chains of nucleotides linked by hydrogen bonds between their projecting bases. The bases are complementary: adenine (A) is always opposite thymine (T)i and guanine (G) l is always opposite cytosine (C). The function of the ligase is to repair "nicks," or breaks in single DNA strands, by synthesizing a phosphodiester bond between adjoining nucleotides [see il- ltistration above]. In 11970 a group working in, the labo- ratory of H. Gobind Khorana, who was then at the University of Wisconsin, found that the ligase produced by the bacterial virus T4 could sometimes cata- lyze the end-to-end linkage of complete- ly separated double-strand DNA seg- ments: The reaction required! that the ends of two segments be able to find each other; such positioning of two DNA molecules was a matter of' chance,, and so the reaction was inefficient. It was clear that efficient joining of! DNA mole- cules required a mechanism for holding the two DNA ends together so that the ligase could act. An ingenious way of accomplishing this was developed and tested indepen- dently in two laboratories at Stanford: by Peter Lobban and A. Dale Kaiser and by David Jackson, Robert Symons and Paul Berg. Earlier work by others had shown that the ends of the DNA mole- cules of! certain bacterial viruses can be joined by base-pairing between comple- mentary sequences of nucleotides that are naturally present on single-strand segments projecting from the ends of those molecules: A's pair with T's,, G'ss pair with~ C's and the molecules are held together by hydrogen bonds that form between the pairs. The principle of link- ing DNA molecules by means of the single-strand projections had been ex- ploited in Khorana's laboratory for join- ing short synthetic sequences of nucleo- tides into longer segments of DNA. The Stanford groups knew too that' an enzyme;,terminal transferase, would cat- alyze the stepwise addition, specificalNy at what are called the 3' ends of single strands of DNA, of a series of identical nucleotides. If the enzyme worked also with double-strand DNA„then a block of identical nucleotides could be added to one population of DNA molecules and a block of! the complementary nucleotides coul& be added to another population from another source. Molecules of the two populations could then be annealed U 0 0 by hydrogen~bonding and sealed togeth- er by DNA ligase. The method was po- tentially capable of joining any two spe- cies of DNA. While Lobban and Kaiser testedthe terminal-transferase procedure with the DNA of the bacterial virus P22, Jackson, Symons an& Berg applied the procedure to link the DNA of the animal virus SV40 to bacterial-virus DNA. The SV40 and bacterial-virus DNA molecules Berg's group worked with are closed loops, and the loops had first to be cleaved to provide linear molecules with free ends for further processing and linkage [Isee illustration on opposite page]. (As it happened, the particular enzyme chosen to cleave the loops was the Eco RI endonuclease, which was later to be used in a different procedure for making the first biologically func- tional gene combinations. At the time, however, the enzyme's special property of producing complementary single- strand ends all by itself' had not yet been discovered.) The cleaved linear molecules were treated with an enzyme, produced by the bacterial virus lamb&, called an exonu- clease because it operates by cutting off nucleotides at the end of a DNA mole- cule. The lambda exonuclease chewed back the 5' ends of DNA molecules and thus left projecting single-strand ends that had 3' termini to which the blocks of complementary nucleotides could be added. The next step was to add, with the help of terminal transferase, a block of A's at the 3' end of one of the two DNA species to be linked and a block of T's at the 3' ends of the other species. The species were mixed together. Frag- ments having complementary blocks at their ends could find each other, line up and become annealed'by hydrogen bond- ing, thus forming combined molecules. To fill the gaps at the 5' ends of the orig- inal segments the investigators supplied nucleotides and two more enzymes: exo- nuclease III and DNA polymerase. Fi- nally the nicks in the molecules were sealed with DNA ligase. T he method of making cohesive ter- mini for joining DNA molecules in the first successful genetic-manipulation experiments was conceptually and op- erationally different from the teiminal- transferase procedure. It was also much simpler. It depended on the ability of one of a group of enzymes called restric- tion endonucleases to make complemen- tary-ended fragments during the cleav- age of DNA at a site within the mole- cule, instead of requiring the addition of new blocks of complementary nucleo- tides to DNA termini. 4

- Viruses grown on certain strains of E. coli were known to be restricted in their ability to grow subsequently on other strains. Investigations had shown that this restriction was due to bacterial en- ~ zymes that recognize specific sites on a"f'oreign" viral DNA and cleave that 3/ AAAA LAMBDA EXONUCLEASE 3' ® 5' DNA. (To protect its own DNA the bac- terial cell makes a modification enzyme that adds methyl groups to nucleotides constituting the recognition sites for the restriction endonuclease, making them resistant to clcavage.) Restriction endo- nucleases (and modification methylases) 5' AAA -3, 3' are widespread in microorganisms; genes for making them were found on viral chromosomes and extrachromosomal plasmid DNA as well as on many bac- terial chromosomes. During the early 1970's the nucleotide sequences at the cleavage sites recognized by severaU re- 5' ~ TERMINAL TRANSFERASE PLUS TTP 5 TTTT 3. .611111111111111111111 3/TTTTT 5 T T T T T I, EXONUCLEASE III AND DNA POLYMERASE TTTTT DNA LIGASE AAA TTTTT s TERMINAL•TRANSFERASE procedure for joining DNA mole- cules involves a number of steps, each dependent on a different en• zyme. If one of the molecules to be joined is a closed loop, it must first be cleaved. The linear molecules are treated with lambda exo- nuclease, an enzyme that cuts nucleotides off the 5' end of DNA strands (the end with a phosphate group on the No. 5 carbon). Then specific nucleotides are added to the 3' end (the end with an OR group on the No. 3 carbon) by the action of the enzyme termi- COMBINED DNA MOLECULE nal transferase. One DNA species is supplied with adenosine tri- phosphate (ATP), the other with thymidine triphosphate (TTP), so that A nucleotides are added to one species and complementary T nucleotides to the other. When the two species are mixed, the com- plementary bases pair up, annealing the molecules. Nucleotides and the enzymes DNA polymerase and exonuclease III are added to fill gaps and DNA ligase is added to seal the DNA backbones. The re- sult is a double molecule composed of two separate DNA segments. ..:..._:..: _ .. ... . -... . .. . ; , .,... .. . .. - .. ..- .... . ...,..,. . .a at-iT. .. _... . - _ .. _~~•......._ 3" 5' TERMINAL TRANSFERASE PLUS ATP

, striction endonucleases were identified. In every instance, it developed, the cleavage was at or near an axis of rota- tional symmetry: a palindrome where the nucleotide base sequences read the same on both strands in the 5'-to-3' direction [see illustration below]. In some instances the breaks in the DNA strands made by restriction en- zymes were opposite each other. One particular endonuclease, however, the Eco RI enzyme isolated by Robert N. Yoshimori in Boyer's laboratory in San Francisco, had a property that was of special interest. Unlike the other nu- cleases known at the time,, this enzyme introduced breaks in the two DNA strands that were separated by several nucleotides. Because of the symmetrical, palindromic arrangement of the nucleo- tides in the region of cleavage this sepa- ration of the cleavage points on the two strands yielded DNA termini with pro- jecting complementary nucleotide se- quences: "sticky" mortise-and-tenon ter- ~ _ .;.. , ._ ,... _ AXIS AXIS OF ROTATIONAL SYMMETRY I a ABLE WAS I E E I SAW ELBA I I 5' G W A A ~ T T C 3' 3' C T T ~ A A T G 5' i b AArrc C G J RESTRICTION ENDONUCLEASES cleave DNA at sites where complementary nucleo- tides are arranged in rotational symmetry: a palindrome, comparable to a word palin- drome (a). The endonuclease Eco RI has the additional property of cleaving comple- mentary strands of DNA at sites (colored arrows) four nucleotides apart. Such cleav- age (b) yields DNA fragments with comple- mentary, overlapping single.strand ends. As a result the end of any DNA fragment pro- duced by Eco RI cleavage can anneal with any other fragment produced by the enzyme. 6 ` mini. The Eco RI enzyme thus produced in one step DNA molecules that were functionally equivalent to the cohesive- end molecules produced by the compli- cated terminal-transferase procedure. The experiments that led to the dis- covery of the capabilities of Eco RI were reported independently and simulta- neously in November, 1972; by Janet Mertz and Ronald W. Davis of Stanford and by another Stanford investigator, Vittorio Sgaramella. Sgaramella found that molecules of the bacterial virus P22 could be cleaved with Eco RI and would then link up end to end to form DNA segments equal in length to two or more viral-DNA molecules. Mertz and Davis observed that closed-loop SV40-DNA- molecules cleaved by Eco RI would re- form themselves into circular molecules by hydrogen bonding and could; be sealed with DNA ligase; the reconsti- tuted molecules were infectious in ani- mal cells growing in tissue culture. Boyer and his colleagues analyzed the nucleo- tide sequences at the DNA termini pro- duced by Eco RI, an& their evidence confirmed the complementary nature of! the termini, which accounted for their cohesive activity. In late 1972, then, several methods were available by which one could join double-sfirand molecules: of DNA. That was a major step in the develop- ment of a system for manipulating genes. More was necessary, however. Most seg- ments of DNA do not have an inherent capacity for self-replication; in order to reproduce themselves in a biological sys- tem they need to be integrated into DNA molecules that can replicate in the particular system. Even a DNA segment that can replicate in its originalihost was not likely to have the specific genetic signals required for replication in, a dif- ferent environment. If foreign DNA was to be propagated in bacteria, as had long been proposed in speculative scenarios of genetic engineering, a suitablevehicle;e or carrier, was required. A composite DNA molecule consisting of the vehicle and the desired foreign DNA would have to be introduced into a population of functional host bacteria. Finally, it would be necessary to select, or identify,, those cells in the bacterial population that took up the DNA chimeras. In 1972 it still seemed possible that the genetic information on totally foreign DNA mol- ecules might produce an aberrant situa- tion that would prevent the propagation of hybrid molecules in a new host. Molecular biologists: had focused for many years on viruses and their rela- tions with bacteria, and so it was natu- ral that bacterial viruses were thought of as the most likely vehicles for genetic manipulation. For some time there had been speculation and discussion about using viruses, such as lambda, that oc- casionally acquire bits of the E. coli chromosome by natural recombination mechanisms for cloning DNA frQm for- eign sources. It was not a virus, hov.ever, but a plasmid that first served as a ve- hicle for introducing foreign genes into a bacterium and that provided a mecha- nism for the replication and selection of the foreign DNA. A ubiquitous group of' plasmids that confer on their host bacteria the ability to resist a number of antibiotics had been studied intensively for more than a dec- ade. Antibiotic-resistant E. coli isolated in many parts of the world, for example, were found to contain plasmids, desig- nated R factors (for "resistance"), carry- ing the genetic information for products that in one way or another could inter- fere with the action of specific antibiotics [see "Infectious Drug Resistance,° by Tsutomu Watanabe; SclErrrlr-IC ArsExl- cAiv, December, 1967]. Double-strand circular molecules of R-factor DNA had been separated from bacterial chromo- somal DNA by centrifugation in density gradients and had been characterized by biochemical and physical techniques [see "The Molecule of Infectious Drug Resistance," by Royston C. Clowes; ScIENTIFZC AMERICAN, April, 1973]. In 1970 Morton Mandel and A. Higa of the University of Hawaii School of Medicine had discovered that treatment of E. coli with calcium salts enabled the bacteria to take up viral DNA. At Stan- ford, Chang and I, with Leslie Hsu, found that if' we made the cell mem- branes of E. coli permeable by treating them with calbium chloride,, purifie& R- factor DNA could be introduced into them [see illustration on opposite page]. The R-factor DNA is taken up in this transformation process by only about one bacterial cell in a million, but those few cellfi can be selected! because they live and multiply in the presence of the anti- biotics to which the R factor conf'ers re- sistance, whereas other cells die. Each transformed cell gives rise to a clone that contains exact replicas of the parent plas- mid DNA molecules, and so we reasoned that plasmids might serve:as vehicles for propagating new genetic information in a line of E. coli cells. In an effort to explore the genetic and molecular properties of various re- gions of the R-factor DNA we had be- gun to take plasmids apart by shear- ing their DNA mechanically and then transforming E. coli with the resulting

c c C £ragments. Soon afterward we began to cleave the plasmids with the Eco RI en- zyme, which had been shown to produce multiple site-specific breaks in several viruses. It might therefore be counted on to cleave all molecules of a bacterial plasmid in the same way, so that any particular species of DNA would yield a specific set of cleavage fragments, and do so reproducibly. The fragments could then be separated and identified accord- ing to the different rates at which they would migrate through a gel under the influence of an~electric current. W hen the DNA termini produced by Eco RI endonuclease were foundto be cohesive, Chang and I, in collabora- tion with Boyer and Helling in San Francisco, proceeded to search for a plasmid that the enzyme would cleave without affecting the plasmid's ability to replicate or to confer antibiotic resist- ance. We hoped that if such a plasmid could! be found, we could insert a seg- ment of foreign DNA at the Eco RI cleavage site, and that it might be pos- sible to propagate the foreign DNA in E. coli. , In our collection at Stanford there was a small plasmid, pSC101, that had been isolated following the mechanical shear- ing of a large plasmid bearing genes f'or multiple antibiotic resistance. It was less than a twelfth as long as the parent plas- mid, but it did retain the genetic infor- mation for its replication in E. coli and for conferring resistance to one antibiotic, tetracycline. When we subjected pSC 101 DNA to cleavage by Eco RI and ana- lyzed the products by gel electrophoresis, we found that the enzyme had cut the plasmid molecule in only one place, pro- ducing a single linear fragment. We were able to join the ends of that frag- ment again by hydrogen bonding and re- seal them with DNA ligase, and when we introduced the reconstituted circular DNA molecules into E: coli by trans- formation, they were biologically func• tional plasmids: they replicated and con- ferred tetracycline resistance. The next step was to see if a fragment of foreign DNA could be inserted at the cleavage site without interfering with replication or expression of tetracycline resistance and thus destroying the plas- mid's ability to serve as a cloning ve- hicle. We mixed the DNA of another E. coli plasmid, which carried resistance to the antibiotic kanamycin, with the pSC101 DNA. We subjected the mixed DNA to cleavage by Eco RI and~ then to ligation, transformed E. coli with~ the re- sulting DNA and found that some of the transformed bacteria were indeed resist- TETRACYCLINE-RESISTANT CELL !1 o O n--- GESIUM GHLORIDE AND O 1 tI HIUIUM tSHUMIUt ` o Q O ~ TETRACYCLINE•RESISTANCE ~ P LASMID M DNA EXTRACTION CENTRIFUGATION' Q~.,~I d~0 O oa O, 00 PLASMID DNA TETRACYCLINE- SENSITIVE CELL I CALCIUM CHLORIDE TETRACYCLINE- RESISTANCE PLASMID -1-kaf CHROMOSOME TRANSFORMED, TETRACYCLINE-RESISTANT CELL PLASMID DNA can be introduced into a bacterial cell by the procedure called transforma- tion. Plasmids carrying,genes for resistance to the antibiotic tetracycline (top le/t) are sep- arated from bacterial chromosomal DNA. Because differential binding of ethidium bromide by the two DNA species makes the circular plasmid DNA denser than the chromosomal DNA, the plasmids form a distinct band on centrifugation in a cesium chloride gradient and can be separated (bottom le/t).The plasmid DNA is mixed with bacterial cells that are not resistant to tetracycline and that have been made permeable by treatment with a calcium salt. The DNA enters the cells, replicates there and makes the cells resistant to tetracycline. I DNA ~~o 0 o~

REPLICATOR . CLEAVAGE SITES 1. TTAA AATT DNA LIGASE FOREIGN DNA I CLEAVAGE BY ~~ ENDONUCLEASE rr.... **'o TM,,,, *^nr= AATT AATT TTAA TTAA PLASMID CHMERA TRANSFORMATION / REPLICATiION \ DAUGHTER CELLS FOREIGN DNA is spliced into the pSC101 plasmid and introduced with the plasmid into the bacterium Escherichia coli; The plasmid is cleaved by the endonuclease Eco RI at a single site that does not interfere with the plasmid's genes for replication or for resistance to tetracycline (top left): The nucleotide sequence recognized by Eco RI is present also in other DNA, so that a foreign DNA es- posed to the endonuclease is cleaved about once in every 4,000 to 8 16,000 nucleotide pairs on a random basis (top right) ~ Fragments of cleaved foreign DNA are annealed to the plasmid DNA by hydro- gen bonding of the complementary base pairs, and the new com- posite molecules are sealed by DNA ligase. The DNA chimeras, each consisting of the entire plasmid and a foreign DNA fragment, are introduced into E. coli by transformation, and the foreign DNA is replicated by virtue of the replication functions of the plasmid. c

c C ant to both tetracycline and kanamycin. The plasmids isolated from such~ trans- formants containe& the entire pSC101 DNA segment and also a second DNA fragment that carried the information for kanamycin resistance, although it lacked replication functions of its own. The results meant that the pSC 101 could serve as a cloning vehicle for introduc- ing at least a nonreplicating segment of a related DNA into E. coli. And the pro- cedure was extraordinarily simple. Could genes.from other species be in- troduced into E. coli plasmids, however? There might be genetic signals on for- eign DNA that would prevent its propa- gation or expression in E. coli. We de- cided to try to combine DNA from a plasmid of another bacterium, the p1258 plasmid' of Staphylococcus aureus, with our original E. coli plasmid. The staph- ylococcal. plasmid had already been studied in several laboratories; we had found that it was cleaved into four DNA fragments by Eco RI. Since p1258 was not native to E. coli or to related bac- teria, it could not on its own propagate in an E. coli host And it' was known to carry a gene for resistance to still another antibiotic, penicillin, that would serve as a marker for selecting any transformed! clones. (Penicillin resistance, like com- bined resistance to tetracyeline and kan- amycin, was already widespread among E. coli'strains in nature. That was impor- tant; if genes from a bacterial speciess that cannot normally exchange genetic information with the colon bacillus were to be introduced into its it was essential that they carry only antibiotic-resistance traits that were already prevalent in E. coli. Otherwise we would be extending, the species' antibiotic-resistance capabil- ities.) Chang and I repeated! the experiment that had been successfut with two kinds of E. coli plasmids, but this time we did it with a mixture of the E. coli's pSC- 101 and the staphylococcal p1258: we cleaved the mixed plasmidsmith Eco RD endonuclease, treated them with ligase and then transformed: E. coli. Next we isolated transformed bacteria that ex- pressed the penicillin resistance coded for by the S. aureus plasmid as well as the tetracycline resistance of the E. coli plasmidi These doubly resistant cells were found to contain a new DNA spe- cies that had the molecular characteris- tics of' the staphylococcal plasmid DNA as well as the characteristics of pSC101. The replication and expression in E. coli of genes derived from an organism ordinarily quite unable to exchange genes with E; coli represented a breach in the barriers that normally separate biological species. The bulk of the ge- netic information expressed in the trans- formed bacteria defined it as E. coli, but the transformed cells also carried repli- cating DNA molecules that had molecu- lar and biological characteristics derived from an unrelated species, S. aureus. The fact that the foreign genes were on a plasmid meant that they would be easy to isolate and purify in large quantities for further study. Moreover, there was a possibility that one might introduce genes into the easy-to-grow E. coli that specify a wide variety of metabolic or synthesizing functions (such as photo- synthesis or antibiotic production) and that'~ are indigenous to other biological classes. Potentially the pSC101 plasmid and the molecular-cloning procedure could serve to introduce DNA molecules from complex higher organisms in,to bac- terial hosts, making it possible to apply relatively simple bacterial genetic and biochemical techniques to the study of animal-cell genes. C ould animal-cell genes in fact be intro- duced into bacteria, and would theyy replicate there?' Boyer, Chang, Helling and I, together with Morrow and Good- man, immediately undertook to find out. We picked certain genes that had been well studied and characterized and were available, purified, in quantity: the genes that code for a precursor of the ribosomes (the structure on which proteins are syn- thesized) in~the toad Xenopus laevis. The genes had properties that would enable us to identify them if we succeed'ed in getting them to propagate in bacteria. The toad DNA was suitable f'or another reason: although we would be construct- ing a novel biological combination con- taining genes from both animal cells and bacteria, we and others expected! that no hazard would result from transplanting the highly purified ribosomal genes of a toad. Unlike the foreign DNA's of our ear- lier experiments, the toad genes did not express traits (such as antibiotic resist- ance) that coul& help us to select bac- teria carrying plasmid chimeras. The tetracycline resistance conferred by pSC101 would make it! possible to select transformed clones, however, and we could then proceed to examine the DNA isolated from such clones to see if any clones contained a foreign DNA having the molecular properties of toad ribo- somal DNA. The endonuclease-gener- ated fragments of toad ribosomal DNA have characteristic sizes and base com- positions; DNA from the transformed cells could be tested for those charac- teristics. The genes propagated in bac- teria could also;be tested'for nueleotide- sequenee homology with~ DNA isolated directly from the toad. When we did the experiment and ana- lyzed the resulting transforme6cells, we found that the animal-cell genes were in- deed reproducing themselves in gen- eration after generation of bacteria by means of the plasmid's replication func- tions. In addition, the nucleotide se- quences of the toad DNA were being transcribed into an RNA product in the bacterial cells. Within a very few months after the first DNA-cloning experiments the pro- cedure was being used in a number of laboratories to clone bacterial and ani- mal-cell DNA from a variety of sources. Soon two plasmids other than pSC101 were discovered that have a single Eco RI cleavage site at a location that does not interfere with essential genes. One of these plasmids is present in many copies in the bacterial cell, making it possible to "amplify," or multiply many times, any DNA fragments linked to it. Investigators at the University of Edin- burgh and at Stanford went on to de- velop mutants of the virus lambda (which ordinarily infects E. coli) that made the virus too an: effective cloning vehicle. Other restriction endonucleases were discovered that also make cohesive termini but that cleave DNA at different sites from the Eco RI enzymes, so that chromosomes can now be taken apart and put together in various ways. The investigative possibilities of DNA cloning are already being explored in- tensively. Some workers have isolated from complex chromosomes certain re- gions that are implicated in particular functions such as replicationa Others are making plasmids to order with specific properties that shoul& clarify aspects of extrachromosomal-DNA biology that have been hard to study. The organiza- tion of complex chromosomes, su& as those of the fruit fly Drosophila, is being studied by cloning the animal genes in bacteria. Within the past few months methods have been developed for selec- tively cloning specific genes of higher organisms through the use of radioac- tively labeled RNA probes: instead of purifying the genes to be studied before introducing them into bacteria, one can transform bacteria with a heterogeneous population of animal-cell DNA and then isolate those genes that produce a par- ticular species of RNA. It is also possible to isolate groups of genes that are ex- pressed concurrently at a particular stage in the animal's development. The potential seems to be even broad- er. Gene manipulation opens the pros- 9

pect of constructing bacterial cells, which can be grown easily and inexpen- sively, that will synthesize a variety of biologically produced substances such as antibiotics and hormones; or enzymes that can convert sunlight directly into food substances or usable energy. Per- haps it even provides an experimental basis for introducing new genetic infor- mation into plant or animal cells. It has been clear from the beginning of experimentation in molecular cloning that the construction off some kinds of novel gene combinations may have a po- tential for biological hazard, and the sci- entific community has moved quickly to make certain that research in genetic manipulation would not endanger the public. For a time after our initial experi- ments the pSC101 plasmid was the only vehicle known to be suitable for cloning foreign DNA in E. coli, and our col- leagues asked for supplies with which to pursue studies we knew were of major scientific and medical importance. In- vestigators normally facilitate the free exchange of bacteria and other experi- mental strains they have isolated or de- veloped, but Chang and I were con- cerned that manipulation of certain genes.could give rise to novel organisms whose infectious properties and ecologi- eal effects could not be predicted. In agreeing to provide the plasmid we therefore asked for assurance that our colleagues would neither introduce tu- mor viruses into bacteria nor create anti- biotic-resistance combinations that were not already present in nature; we also asked the recipients not to send! the plas- mid on to other laboratories, so that we could keep track of its distribution. When still other cloning vehicles were CLEAVED TOAD RIBOSOMAL DNA discovered, it became apparent that a more general mechanism for ensuring ex- perimental safety in gene-manipulation research was advisable. The groundwork for such control had been established earlier: the National Academy of Sci- ences had been urged to consider the "possibility that potentially biohazardous consequences might result from wide- spread or injudicious use" of these tech, niques and had asked Paul Berg to form an advisory committee that would con- sider the issue. Berg too had been con- cerned about the potential, hazards of certain kinds of experimentation for some years, and had himself decided to abandon plans to try to introduce genes from the tumor virus SV40 into bacteria • because of the possible danger if the ex- periment were successful. Berg brought together a number of i4- vestigators, including some who were then directly involved in molecular clon- ing, in the spring of 1974. In a report released in July and in a letter to leading professional jpurnals the members of the committee expressed their "concern about the possible unfortunate conse- quences of indiscriminate application" of the techniques and formally asked all, investigators to join themi in voluntarily deferring two types of experiments (which had, as a matter of fact, been avoided by informal consensus up until that time). Experiments of Type I in- volved the construction of novel orga- nisms containing combinations of toxin- producing capabilities or of antibiotic- resistance genes not found in nature. Type 2 experiments involved the intro- duction of DNA from tumor viruses or other animal viruses into bacteria; the committee noted'~ that "such recombinant molecules might be more easily dissemi- CD 42 CLEAVED PLASMID CHIMERAS nated to bacterial populations in humans and other species, and might thus in- crease the incidence of cancer or other diseases." The Academy committee was con- cerned largely because of our inability to assess the hazards of certain experi- ments accurately before the experiments were undertaken. Guidelines for safety had long been available in other areas of potentially hazardous research, such as studies invol+ving known disease-causing bacteria and viruses, radioactive isotopes or toxic chemicals. Because of the new- ness of the microbial gene-manipulation methods, no such guidelines had yet been developed for work in this area, however; there was the possibility that potentially hazardous experiments might proceed before appropriate guidelines could be considered and implemented. We recognized that most work with the new methods did not and would not in- volve experiments of a hazardous nature but we recommended the deferral of ' Type I and Type II experiments until the hazards were more carefully assessed, until it was determined whether or not the work could be undertaken safely and until adequate safety precautions were available. The committee also proposed that an international meeting be held early in 1975 to consider the matter more fully: Such a meeting was held in February at the Asilomar Conference Center near Pacific Grove, Calif. It brought together 86 American biologists and 53 investiga- tors from 16 other countries, who spent three and a half days reviewing progress in the field of molecular cloning and for- mulating guidelines that would allow most types of new hereditary character- istics to be introduced into bacteria and CD30 CD35 CD4 CD 18 -4.2 -3.9 4 •/ GEL ELECTROPHORESIS demonstrates the presence of toad DNA in chimeric plasmids. Fragments of DNA migrate through a gel at different rates under the influence of an electric current, de, pending on their size. Linear molecules of plasmid DNA (right) and the cleavage products of toad ribosomal DNA (Te/t) therefore 10 3.0 have characteristic sizes and migrate characteristic distances in a given time. The bands of DNA, visualized by a fluorescent dye, are photographed in ultraviolet. All five chimeric plasmids (center) contain a plasmid DNA molecule; in addition each chimera in- cludes one or more fragments characteristic of original toad DNA. CLEAVED pSClot PLASMID c

~_ viruses safely: Invited nonscientists from the fields of law and ethics participated in the discussions and decisions at Asilo- mar, along with representatives of agen- cies that provide Federal funds for sci- entific research; the meetings were open to the press and were fully reported. The issues were complex and there were wide differences of opinion on many of them, but there was consensus on three major points. First, the newly develope& clon- ing methods offer the prospect of deal- ing with a wide variety of important sci- entific and medical problems as well as other problems that trouble society, such as environmental pollution and food and energy shortages. Second, the accidental dissemination of certain novel biological combinations may present varying, de- grees of potential risk. The construction of such combinations should proceed only under a graded series of precau- tions, principally biological and physical barriers, adequate to prevent the escape of any hazardous organisms; the extent of the actual risk shoul& be explored'by experiments conducted under strict con- tainment conditions. Third, some experi- ments are potentially too hazardous to be carried out for the present, even with the most careful containment. Future re- search and experience may show that many of the potential hazards considered at the meeting are less serious and less probable than we now suspect. Never- theless, it was agreed that~ standards of protection should be high~ at the begin- ning and that they can be modified later if the assessment of risk changes. Physical containment barriers have long been used in the U.S. space-explo- ration program to minimize the possibil- ity of contamination of the earth by ex- traterrestrial microbes. Containment pro- cedures are also employed routinely to protect laboratory workers and the pub- lic from hazards associated with radio- active isotopes and toxic chemicals and in work with disease-causing bacteria and viruses. The Asilomar meeting for- mulated the additional concept of bio- logical barriers, which involve fastidious . cloning vehicles that are able to propa- c gate only in specialized hosts and equal- ly fastidious bacterial, strains that'~ are unable to live except under stringent laboratory conditions. In the past the scientific community has commonly policed its own actions in- formally, responding to ethical concerns with self-imposed restraint. Usually, but, not always;,society at large has also con- sidered the public well-being in deter- mining how knowledge: obtained by basic scientific research should be ap- plied. Extensive public scrutiny and TOAD DNA . HETERODUPLEX ANALYSIS identifies regions of a toad DNA (black) that have been in- corporated in a chimeric plasmid DNA molecule. DNA isolated from toad eggs and the DNA of the chimera are denatured„that is, each natural double-strand molecule is split into two single strands of DNA, by alkali treatment. The toad and the cbimeric DNA's are mixed together, and any complementary sequences are allowed to find each other. The toad DNA incorporated in the chimeras has nucleotide sequences that are complementary to sequences in the DNA taken directly from the animal source. Those homologous sequences anneal to form heteroduplex double-strand DNA..t6at can be identified in electron micrograpbs. open discussion by scientists andd non- lic discussions initiated by scientists scientists of the possible risks and bene- working in genetic manipulation will be. fits of a particular line of basic research One can hope that the forthright ap- has been rare, however, when (as in this proach and the rigorous standards that case) the hazards in question are only have been adopted! for research in the potential and; for some experiments, cloning of recombinant DNA molecules even hypothetical. As this article is be. will promote a sharper focus on other ing written it is still'too early to know issues relevant to public and environ- what the long-range outcome of the pub- mental safety. PRESENCE OF'TOAD DNA in two separate chimeric plasmid molecules is demonstrated by an electron micrograph made by John F. Morrow at the Stanford University School of Medicine. As is indicated in the drawing (bottom), there are DNA strands from two plas- mids and a strand of toad DNA. The micrograph shows thickened regions of DNA where nucleotide sequences are homologous and two single strands have been annealed. The toad DNA in the chimeras cddes for ribosomes, and the space between the two betero- duplex regions is compatible with the spacing of multiple ribosomal genes in toad DNA. II PLASMID DNA1ll1CKING

The Author STANLEY N. COHEN is associate professor of medicine at the Stanford University School of Medicine. A gradu- ate of Rutgers University and the Uni- versity of Pennsylvania School of Medicine, he joined the Stanford faculty in 1968 after spending several years teaching and doing research in molecular biology at the Albert Einstein College of Medicine. His research has also involved a stint at'the National Institute of Arthri- tis and Metabolic Diseases. A specialist in bacterial plasmids, Cohen was a mem- ber of the National Academy of Sciences committee that recently callod' for the voluntary deferral of' certain potentially hazardous experiments involving recom- binant DNA molecules. Bibliography CONSTRUCTION' OF BIOLOGICALLY FUNCTIONAL BACTERIAL PLASMIDS IN' VITRO. Stanley N. Cohen„Annie C. Y. Chang, Herbert W. Boyer and'Robert B. Helling in Proceedings of the Na- tional Academy o f Sciences o f the United States of America, Vol. 70, No. 11, pages 3240-3244; November, 1973. REPLICATION AND TRANSCRIPTION OF EUxARYoTIc DNA IN ESCHERICHIA COLI. John F. Morrow, Stanley N. Cohen, Annie C. Y. Chang, Herbert W. Boyer, Howard M. Goodman and Robert B. Helling in Proceedings of the National Academy o f Sciences o f the United States of America, Vol. 71, No. 5, pages 1743-1747; May, 1974. POTENTIAL BIOHAZARDS OF RECOMBI~ NANT DNA MOLECULES. P. Berg et al. in Proceedings of the National Acad- emy of Sciences of the United States of America, Vol. 71, No. 7, pages 2595-2599; July, 1974. FIRST ASM CONFERENCE ON E%TRA- CIiROMOSOMAL ELEMENTS IN BACTE- RIn in Microbiology-1974. American. Society for Microbiology, 1975. REPORT OF TIIE WORKING' PARTY ON THE EXPERIMENTAL MANIPULATION OF MICROORGANISMS. Her Majesty's Stationery Office, London;1975.