Biotechnology at 25:
Perspectives on History, Science, and Society
Saturday, March 13, 1999
To view the video presentation of Dr. Cohen's
lecture click on a 'Play' button. The presentation will
start at that point in the transcript.
Stanley N. Cohen, M.D.
Co-developer of recombinant DNA technology and Professor of Genetics, Stanford
Stanley N. Cohen: Thank you very much. And it's a pleasure to
drive up from the farm to talk to you this morning about some of the
history that led to the creation of recombinant DNA. I've been asked
to spend the next fifteen minutes doing that.
term recombinant DNA was introduced in a paper that my colleagues and
I published in early 1974. It was intended to be synonymous with DNA
cloning but it's also been used in a narrower sense to refer simply to
composite DNA molecules that result from the physical joining of DNA
fragments in the test tube. In reality though, the usefulness of recombinant
DNA is for the analysis and manipulation of genes. And particularly,
its importance for biotechnology, the subject of today's celebration,
depends on both the ability to join DNA fragments and the ability to
propagate and clone DNA in a foreign host. The invention of DNA cloning
was made possible by two separate lines of basic research. One concerned
genetic and biochemical studies of bacterial plasmids, circles of DNA
that have the remarkable ability to reproduce themselves in bacteria
independently and separately from the chromosomes of the host bacteria
that harbor them. The other pillar of DNA cloning is the biochemical
manipulation of DNA and procedures to do that.
1968, when I joined the Stanford faculty, I began to investigate the
genetics and biochemistry of a mysterious group of plasmids that made
disease-causing bacteria resistant to the effects of drugs that ordinarily
inhibit their growth. Even then, antibiotic resistance was an important
clinical problem. This is a cover that recently appeared in Stanford,
MD. And genes carried by plasmids were responsible for this resistance.
While plasmids had been studied for fifteen years, little was known about
how they evolved and how they were propagated.
you think about it, you realize that the field of molecular biology developed
in the 1950's and '60's as an amalgam of the disciplines of genetics,
biochemistry and microbiology. And bacterial viruses, as Jim Watson has
already said, had been the principal focus of molecular biology, and
the reason for this was largely because the cloning of viruses occurs
during normal viral infection producing millions of genetically identical
viral copies. These cloned viruses can be analyzed genetically in biochemistry.
To study plasmids I realized it would be necessary to develop methods
for cloning plasmid DNA molecules in bacteria.
1971, Lesley Shuh, a first year Stanford medical student working in my
lab, found that by modifying a procedure developed initially at the University
of Hawaii for work with viruses, we could enable E. coli bacteria
to take up plasmid DNA and to produce offspring that contain self-replicating
plasmids. Genes on the plasmid made these transformed bacteria resistant
to antibiotics. They survived and grew when exposed to the drugs, whereas
bacteria lacking plasmids were killed. Since all descendants of a transformed
cell contain plasmids identical to the one that originally entered that
cell, it was now possible to biologically make replicas, clones of individual
molecules of plasmid DNA. The ability to clone plasmid DNA provided one
of the two ingredients for genetic engineering.
map and isolate plasmid genes I wished to take plasmids apart and put
them back together again one segment at a time. In nature, antibiotic
resistance genes, and other groups of genes as well, were linked in cells
to the replication machinery of plasmids by natural recombination processes.
Potentially, the replication regions of plasmids might be able to duplicate
DNA segments attached biochemically, as occurs with genes linked to plasmid
replication regions in cells. But how to accomplish this linkage?
important biochemical strategy for joining together DNA segments depends
on the ability of nucleotides in a single strand of DNA to pair with
complementary nucleotides on the other strand to form a DNA duplex. The
A's and G's pair with C's and T's, as Watson and Crick had found some
The late 1960's work in the laboratory of Gobind Khorana showed that
short segments of synthetic single strand pieces of DNA could be linked
in test tubes using pairing of overlapping nucleotides and building up
segments of DNA in this way. DNA ligase, an enzyme that repairs breaks
in DNA, can seal the molecules together. However, construction of overlapping
complementary DNA ends, as shown here by the stepwise biosynthetic addition
of single nucleotides, was both technically difficult and laborious.
Sgaramella and others in Khorana's lab had also reported that the DNA
ligase of a particular bacterial virus called T4 has the power to join
together blunt DNA ends that lack complementarity.
However, not withstanding this discovery, the scientific community still
focused on complementary nucleotides at DNA ends as a method of joining
DNA molecules. A strategy to circumvent the need to add these nucleotides
one at a time was described in 1969 in a Ph.D. thesis proposal by Stanford
graduate student, Peter Lobban, working in the laboratory of Professor
Dale Kaiser in the department of biochemistry.
stretch of A's could be added to one DNA fragment and a complementary
stretch of T's to the other using terminal transferase, an enzyme discovered
earlier at the Oak Ridge National Laboratory.
Quite independently, R. H. Jensen at the International Minerals and
Chemicals Company devised the same strategy. In mid-1971, a little noticed
paper by Jensen and his colleagues reported the first linking together
in test tubes of separate DNA molecules that contained stretches of A's
and T's added to their ends. However, Jensen's efforts to use DNA ligase
to clone the nicks in these recombined DNA molecules were not successful.
Ironically, it's now known that repair of such nicks occurs naturally
in cells if the molecules are introduced into living cells.
and his faculty advisor Dale Kaiser discovered the additional enzymatic
manipulations necessary for permanent fusion of DNA fragments linked
together by AT tails and used these manipulations for joining together
fragments of a bacterial virus named P-22. Their work was noted by Paul
Berg and his colleagues as enabling their use of the DNA AT joining approach
to biochemically link DNA molecules taken from two different sources,
the E. coli virus, lambda, and the mammalian virus, SV40.
Unbeknownst to Berg at the time, the insertion of SV40 in lambda DNA
interrupted a lambda gene essential for replication. The biohazard concerns
that led Paul and others to decide not to introduce the link DNA fragments
into bacteria and which led in significant part to the biohazard discussions
that Jim Watson had alluded to were moot.
by late 1972 several different methods were available to join separate
DNA molecules outside of living cells. This provided the second ingredient
for genetic engineering. However, the splicing of DNA for the first DNA
cloning experiments did not depend on synthetic DNA ends or even on blunt
ended joining by bacteriophage T-4 ligase. Nor did it require the enzymes
for biochemical advances used in the DNA manipulations I've just described.
Instead, it employed restriction endonuclease, enzymes that have the
ability to recognize specific nucleotide sequences in DNA and to cut
these sequences in one step in a way that produces projecting complementary
ends. These enzymes and their remarkable DNA cutting abilities had been
discovered respectively in the laboratories of Werner Arber in Switzerland
and Hamilton Smith at Johns Hopkins. And I think most of you know these
enzymes recognize sequences in DNA that are palindromes. They read the
same backwards and forwards, and here are some examples of word palindromes:
A man a plan a canal, Panama. Madam o madam.
Eco R1 endonuclease, one of these enzymes, recognized, as these enzymes
do in general, sequences of nucleotides which because of the polarity
on DNA read the same in the five prime to three prime direction on both
strands. In November 1972, Joe Hedgpeth, Howard Goodman, and Herbert
Boyer, at the University of California, reported the six nucleotide sequence
recognized by Eco R1, a restriction enzyme encoded by a gene on an antibiotic
resistance plasmid isolated from a patient at the University of California,
San Francisco. As seen here, Eco R1 cuts duplex DNA asymmetrically within
the sequence, generating projecting single strand ends. Simultaneously
published work at Stanford by Vittorio Sgaramella in the Department of
Genetics and by Janet Mertz and Ron Davis in the Department of Biochemistry
shows that DNA segments could be joined together using these complementary
That same month, at a scientific meeting in Honolulu, Hawaii, I reported
our newfound ability to clone individual molecules of plasmid DNA and
bacteria. I listened with excitement as Herb Boyer described his data
showing the nature of the DNA ends generated by Eco R1 cleavage. I had
been working to restructure plasmids. It should be possible, I thought,
to reshuffle plasmid DNA segments by cleaving the DNA with Eco R1 and
then joining different resistance gene segments to fragments having the
capacity for replication. This would be done using the projecting single
strand ends generated by the Eco R1 enzyme.
That evening at a delicatessen near Waikiki Beach, Herb and I discussed
the scientific collaboration that led to DNA cloning. This discussion
was depicted some years later in a cartoon in the Honolulu Advertiser.
One can see that this is Herb insisting that, yes, Jim Watson is right,
DNA does really have two strands, and I guess this bearded person is
swallowing a corned beef sandwich.
DNA fragments held together by complementary Eco R1 generated ends would, as
we discussed, be fused covalently and permanently by DNA ligase and then introduced
into bacteria. We reasoned that if one of the fragments carries machinery for
replication, this fragment might allow replication and propagation of the DNA
linked to it. While the concept was straightforward, no one knew at the time
whether such structurally modified plasmids could be propagated in living cells.
began the experiments shortly after the Hawaii meeting and by March 1973,
had established the feasibility of DNA cloning. Among my plasmids at
Stanford was a DNA circle, plasmid pSC101, which carries a resistance
gene for the antibiotic tetracycline. We found that this plasmid was
cut at only a single site by the Eco R1 endonuclease. Because cleavage
at this site did not interrupt either the replication origin of the plasmid
or the tetracycline resistance gene that it carried, pSC101 could be
used as a vector for the cloning of fragments of other plasmids that
had been similarly generated using Eco R1. As has since been done many
hundreds of thousands, if not millions of times, the plasmid DNA circle
was opened, as shown in the diagram here, by cutting it with a restriction
enzyme. When the linearized DNA vector was mixed with another DNA that
had been cleaved by the same enzyme, complementarity at the end held
the two fragments together. Ligation closed the nick still present in
the DNA circle, which we then introduced into calcium chloride-treated
bacteria. Spreading the bacterial population on culture media containing
tetracycline resulted in growth of only bacterial clones that contain
recombinant DNA molecules, allowing us to then clone these DNA molecules.
During this work, plasmids we isolated at Stanford were transported
to Boyer's lab in San Francisco where they were cut by Eco R1, analyzed
and then transported back to Stanford where the joined fragments were
introduced into bacterial cells by transformation and analyzed again.
Annie Chang, a research technician in my laboratory who carried out many
of these experiments, lived in San Francisco at that time and transported
our precious DNA samples between the two labs almost daily. Bob Helling,
who was on sabbatical leave in Herb's lab, participated in the DNA analyses.
The strategy that Boyer and I had devised worked better than anyone could
have expected. And the months in early 1973 were a period of almost unbelievable
excitement for all of us.
May 1973, a manuscript describing these results was prepared for submission
to the Proceedings of the National Academy of Sciences. This paper,
which was published in November 1973, forms the basis for three patents
just recently expired that are owned by Stanford and the University of
California and which underlie much of modern biotechnology. And this
is the basic DNA cloning patent. Although DNA cloning had been developed
to investigate basic biological phenomena, its practical significance
and potential applications in biotechnology were apparent. However, our
demonstration that fragments of DNA taken from E. coli plasmids
could be cloned in E. coli by inserting them into plasmid vectors
didn't necessarily mean that foreign DNA could also be propagated in
the same way. It had long been believed that natural barriers would prevent
gene transfer between all but the most closely related organisms, and
scientific colleagues offered cogent reasons why the transplantation
of DNA to unrelated organisms would not be possible.
While initial collaborative experiments with Boyer and Helling were
impressive, Chang and I found that DNA cloning allowed species barriers
to be breached. DNA from a plasmid from a totally different type of bacterium,
a Staphylococcus, was propagated in E. coli by linking
it to an E. coli plasmid vector. In our April 1974 report of these
results we wrote, "The replication and expression of genes in E. coli that
have been derived from a totally unrelated species now suggests that
interspecies genetic recombination may be generally obtainable." We suggested
it may be practical to introduce into E. coli genes specifying
metabolic or synthetic functions such as photosynthesis or antibiotic
production indigenous to other biological classes. These results, we
felt, support the view that antibiotic resistance replicons such as the
pSC101 plasmid may be of great potential usefulness for the introduction
of DNA derived from eukaryotic organisms into E. coli, thus enabling
the application of bacterio-genetic and biochemical methods for the study
of eukaryotic genes.
afterward, additional collaborative experiments between Boyer's lab and
mine, with the participation of John Morrow at Stanford and Howard Goodman
at UCSF, showed that eukaryotic DNA, segments of eukaryotic chromosomes,
could also be propagated in E. coli and its genes transcribed
there. Boyer and his collaborators succeeded then in expressing a mammalian
protein, somatostatin, in bacteria, and found that this protein showed
normal immunological reactivity. In still later experiments done collaboratively
with Robert Schimke at Stanford, Chang and I constructed the first bacteria
that synthesized a functional mammalian protein, the mouse enzyme, dihydrofolate
reductase, thus demonstrating that bacteria could produce biologically
active animal cell proteins, and helping further to set the scene for
the emergence of biotechnology. This emergence was quick to come with
the subsequent construction of plasmids encoding insulin by scientists
at Genentech, by Rutter and Goodman and their associates, and by Wally
as we've seen from Jim Watson's talk, and I hope is also evident from
mine, science isn't done in a vacuum. In common with most inventions,
the invention of DNA cloning was the fruit of many years of efforts in
multiple laboratories. As with most scientific discoveries, one can say
with some certainty that if Boyer and I had not collaborated to invent
DNA cloning in early 1973, the cloning of DNA would later have been achieved
by others, probably approaching this objective from a different perspective.
Such is and such should be the nature of science. Thank you.