Developing One of the First Methods for Sequencing DNA

00:00:13.22 I'm Ron Vale, and today I have the pleasure to be speaking
00:00:16.26 with Dr. Wally Gilbert. Dr. Gilbert started off as a
00:00:21.08 theoretical physicist, and then through an introduction
00:00:24.19 from Jim Watson, moved into the area of molecular
00:00:28.11 biology, where he's made many important discoveries
00:00:31.18 over the years. He isolated a protein that can control
00:00:36.14 a genetic element, the lac repressor. He's also
00:00:39.28 made many important contributions in our understanding
00:00:42.14 of mRNA. But today we're going to be talking about
00:00:46.06 his seminal work on DNA sequencing. Work that
00:00:50.04 rapidly went to his being awarded a Nobel Prize
00:00:53.26 in 1980. So thank you for joining us today, and
00:00:58.19 maybe we can just start off by going through the history
00:01:03.07 of what led you into tackle the problem of DNA sequencing.
00:01:08.08 Okay, in the early 70s, we knew of course that the
00:01:15.13 sequence of the DNA carried the ultimate code that's passed
00:01:18.14 from parent to offspring. But in fact, because it's essentially impossible
00:01:23.21 to actually work out the DNA sequence. I had isolated
00:01:28.10 the lac repressor, the protein that binds to a region of DNA.
00:01:31.28 In the early 70s, I actually then tried to isolate that region
00:01:36.04 of DNA to which the protein bound, and we did that by
00:01:40.02 attaching the protein to DNA, dissolving all of the rest of
00:01:42.27 the DNA with DNase and then isolating the little fragment
00:01:45.05 left, protected by the protein. So I had that physical fragment in my hand.
00:01:49.22 And I wanted to know its sequence. We tried developing
00:01:54.11 DNA sequencing methods. Nothing seemed to work.
00:01:56.14 And we then worked out its sequence, my technician,
00:02:00.17 Allan Maxam and I, for about two years. By copying that little fragment
00:02:05.01 of DNA to RNA, and then working out the sequence of
00:02:07.21 the RNA by breaking it with enzymes and using gigantic
00:02:12.00 electrophoretic apparatuses to separate the RNA spots on
00:02:16.22 paper electrophoresis. It took us two years to work out
00:02:20.21 essentially 24 bases of DNA. That's about one base
00:02:24.07 a month. We knew that a typical gene in bacteria, we thought,
00:02:29.12 the genes in our body would be in the order of 1000
00:02:31.22 to 15,000 bases long. We now know of course, the genes in our body are
00:02:36.29 30,000 bases long, far bigger than anyone could've imagined at the time.
00:02:41.19 But at one base a month, you would never get to a thousand
00:02:45.12 bases of DNA. You would never be able to sequence that in any way.
00:02:49.02 And that was actually the nature of the problem.
00:02:52.01 In the early 70s. Another group worked out the sequence
00:02:58.23 across the entire beta-galactose genes. Again by copying
00:03:04.00 DNA and RNA. And I thought, well, that's all I interested in,
00:03:08.13 the way it sequences. Go on and work on something else.
00:03:12.14 A Russian, Andrei Mirzabekov, came to visit me. And he wanted
00:03:17.00 to convince me to do an experiment. He wanted to convince me
00:03:19.11 to look at whether the repressor bound to DNA in the large
00:03:23.05 groove or the small groove by modifying the DNA with chemical
00:03:26.10 agents. And the agent he liked was dimethyl sulfate,
00:03:30.00 which puts a methyl group on the guanine in the large groove.
00:03:32.27 I thought it would be an awfully interesting experiment, but I
00:03:36.12 didn't see a way of doing it, so I didn't do anything.
00:03:39.08 Six months later, the part of his trip to America, came through
00:03:42.26 again. We had lunch, he tried again to convince me I should
00:03:45.11 really do an experiment of that kind. I had the repressor,
00:03:48.10 I had the DNA, I had all access to all of this. The question
00:03:55.07 -- the problem was, how could I recognize where these
00:03:59.19 modified bases were in the DNA, which the repressor would bind.
00:04:03.15 In the conversations with Andrei, we thought of a way of
00:04:08.28 actually trying to identify the positions of these modified bases.
00:04:12.29 We had available ways of breaking the DNA with restriction enzymes,
00:04:19.21 such that we could isolate a piece of DNA about a hundred
00:04:23.15 nucleotides long. Which could be labeled at one end or the other
00:04:27.09 with radioactive phosphate, depending on the way in which we
00:04:30.14 isolated the DNA and broken it sequentially with restriction enzymes.
00:04:33.28 So we could have available a DNA molecule we knew would bind
00:04:38.13 the repressor with a label at one end, and the modification
00:04:42.20 with dimethyl sulfate we knew would weaken the base,
00:04:46.03 which the modification occurred in such a way that we could
00:04:48.24 break the DNA at that point. So we reasoned since we knew
00:04:53.25 the whole sequence of this tiny little molecule, we should be able
00:04:57.21 to break the DNA at the point at which the methylations
00:05:02.02 occurred, and separate those molecules by size on a gel.
00:05:05.23 And since we knew the sequence, we should be able to
00:05:07.21 decipher which molecule came from what. So I did that
00:05:11.14 experiment. I put the repressor on the DNA, modified it,
00:05:14.15 second tube had no repressor on the DNA, modified it
00:05:18.14 with dimethyl sulfate, broke the DNA, ran gels, and the experiment
00:05:24.04 was a great success second time I did it. First time it didn't work
00:05:27.21 at all. Second time I did it, it worked perfectly. The guanine
00:05:33.00 modified the large groove, the repressor blocks the modification of
00:05:37.16 guanine. The adenines were modified in the small groove,
00:05:40.00 you can see yes, the repressor blocks certain adenines.
00:05:42.17 But the striking thing about that pattern, which I had expected
00:05:47.07 to have to deduce from the sequence what the pattern should
00:05:51.03 be, the striking thing about the pattern on the gel was
00:05:55.03 that it's separating the DNA fragments from the smallest
00:05:58.00 to the largest, you've got a very characteristic pattern.
00:06:00.24 And you could see, oh here's DNA broken at position 30.
00:06:05.28 Here's the next band, just a little higher, broken at position 31.
00:06:09.22 Well, you look at that pattern and you say, this must
00:06:12.28 be the sequencing method. It's sufficiently clear that
00:06:16.04 in one go I could read essentially a hundred bases of DNA.
00:06:19.06 So was there a eureka moment?
00:06:21.07 Well, there's a moment when you look at that and you say,
00:06:22.25 clearly this is a sequencing method. The only problem is
00:06:25.11 can you actually sequence just by looking at the A's and
00:06:30.03 G's? We actually sequenced something which we never
00:06:32.04 published just by looking at the A's and G's coming down
00:06:34.14 the two strands. And we started looking for chemical reagents that
00:06:38.03 would break the DNA at cytosines and thymines. Now, in fact,
00:06:42.15 we knew of such reagents immediately, because I knew hydrazine
00:06:45.29 broke DNA at cytosines and thymines. That's part of the
00:06:51.00 old chemistry that we tried five years earlier, trying to work out
00:06:55.08 a sequencing method. So we could immediately get a display
00:06:58.09 of the C's and the T's. Allan Maxam then spent about 9 months
00:07:02.26 trying to find a way of changing that display so you could
00:07:06.11 distinguish the C's and the T's. Essentially by throwing
00:07:09.04 every chemical in the laboratory into the reaction to see
00:07:12.04 if it made a difference. And he came up with a modification
00:07:15.13 by adding salt to the reaction. And we ended up with effectively
00:07:21.06 four reactions, each one of which was specific enough
00:07:24.26 to reveal a breakage of DNA at a certain base. And
00:07:29.22 the trick behind the sequencing was this idea of measuring
00:07:32.28 a distance from the wavelength of the labeled phosphate
00:07:37.13 to the position of each base. The first gels would display
00:07:42.04 molecules on the order from 10 to 50-60 long, you could read that
00:07:47.14 sequence. We very quickly went to gels that could display molecules
00:07:51.02 hundreds long. And you could read in an afternoon
00:07:55.11 more than one base a month, and in an afternoon
00:07:58.13 you could read several hundred bases. The next
00:08:02.16 I had had a previous student, Nancy Maizels, who had worked out
00:08:07.25 63 bases of DNA by copying an RNA for her thesis.
00:08:11.12 The next thesis was about 1500 bases of DNA, was
00:08:15.17 the sequence of the repressor gene itself worked out from the
00:08:20.29 DNA. In that case, correcting the protein sequence, which
00:08:24.13 there were mistakes. And that was the next student's thesis.
00:08:27.24 The student after that did 5000 bases of DNA that did an entire
00:08:32.04 plasmid. We stopped giving PhDs for sequencing. The method was actually
00:08:38.19 developed around 1976, '75 or '76, and we gave the method
00:08:43.27 away at Gordon Conferences. The Gordon Conference that
00:08:47.14 summer. We gave a little handout and people began to sequence
00:08:50.29 DNA immediately. The paper, the official paper published on it
00:08:55.11 was probably not published until 1976, or 1978. By that time,
00:08:59.17 people came to the laboratory to learn the sequencing
00:09:02.07 methods and took it away.
00:09:03.07 How long did it take from the time when you were working
00:09:08.06 on the problem from the repressor point of view to developing
00:09:11.20 this general method?
00:09:13.14 The total time was 9 months, to find the last bit of the
00:09:16.24 chemistry. And the sequencing came -- the fact that you could
00:09:21.12 do the sequencing came immediately out of that first experiment.
00:09:24.24 That immediately said, oh it's clear that the patterns are distinct
00:09:28.24 enough. Fred Sanger was working on the problem
00:09:32.21 simultaneously, trying to develop sequencing methods based
00:09:36.06 again on ultimately the same idea, can I measure a distance
00:09:41.13 along the DNA to a position at which a base occurs?
00:09:45.08 The gels that he used came out of my laboratory, gels that he
00:09:50.17 also used came out of the technology that we had developed
00:09:54.05 in terms of sizing DNA molecules. And most of that technology
00:09:59.01 had been developed earlier just to size DNA molecules
00:10:01.26 for the whole problem of finally isolating the DNA fragment
00:10:05.02 that was bound to the repressor.
00:10:06.04 And what was the reception from the community when
00:10:11.09 you first showed these gels at meetings?
00:10:15.01 People began to sequence immediately. I mean the sequencing
00:10:18.11 took off at that point, both Fred's method became available
00:10:21.23 and our method at essentially the same time in 1976.
00:10:24.17 And people around the world began to sequence actively
00:10:29.04 DNA. The rate of DNA sequencing started to go up following
00:10:34.12 Moore's Law, doubling every 18 months. So from that time
00:10:38.17 until probably 2010, it may be going even faster now.
00:10:43.25 So just looking at how the field has progressed since then,
00:10:50.03 when you first developed this method, did you have any
00:10:57.11 idea or anticipation of how the field of DNA sequencing
00:11:03.04 might develop?
00:11:05.00 It's a hard question. In the middle '70s, when we developed the method,
00:11:11.26 and by the late '70s, it was clear you could decipher
00:11:15.11 any gene. And that's what people were doing all over, everywhere.
00:11:21.20 And the Nobel Prize is actually for the recombinant technology
00:11:26.08 It was given half to Paul Berg for standing in for all the
00:11:30.17 people who showed you could move the DNA fragment
00:11:32.20 around from one cell to another, put it in a plasmid,
00:11:36.09 grow a human DNA fragment in a bacterial cell.
00:11:40.25 And to Fred and myself for the ability to decipher the DNA.
00:11:44.25 It wouldn't do much good if you could move DNA around
00:11:47.09 if you couldn't decipher it. Across that period people were
00:11:51.10 then sequencing genes like mad. Around 1980, it was essentially
00:11:59.05 the workhorse, the sequencing methods of those times. Fred's
00:12:02.19 and mine were being used equally. And continued to be
00:12:06.22 equally until the middle '80s, when the machines came in and
00:12:12.11 which the dideoxy method was mostly adapted to.
00:12:16.07 And once the machines came in, Fred's method took over.
00:12:20.20 Almost entirely, the chemical method is used only for very special
00:12:24.20 cases. It could do certain things that the dideoxy method
00:12:27.26 can't do. It can reveal the methylation pattern and things like that.
00:12:32.00 But beyond that, it became sort of for specialty use.
00:12:36.29 I did not anticipate at the time we discovered the DNA
00:12:41.01 sequencing, the genome. In fact in 1985, I was invited
00:12:46.08 to a meeting in California by Bob Sinsheimer, and
00:12:53.27 with a sort of idea, could one sequence the human genome?
00:12:59.14 And I went to that meeting thinking, that's the silliest
00:13:01.21 thing I ever heard of. That's so much larger than things we're
00:13:05.04 sequencing up till now. And at the meeting, I thought about
00:13:09.13 where the technology has gotten to. I thought finally,
00:13:14.13 it's probably possible to sequence the genome. I became,
00:13:17.04 at that time in 1985, an advocate of the idea that yes, you
00:13:20.24 should sequence the genome. Originally, I thought it would be
00:13:24.04 a 30 year problem, as I suggested to Bob. And he said
00:13:28.15 to sequence the genome, 10 years for the first 1%,
00:13:33.03 10 years for the next 10%, and 10 years for the last. And do it
00:13:37.11 on an industrial scale, but a 30 year problem.
00:13:40.26 It actually took about 15 years, it was a purely industrial
00:13:45.28 scale problem. And I realized in '85, the technology that
00:13:51.04 then existed was improving at a rate such that you could do
00:13:55.09 the entire human genome in a reasonable amount of time.
00:13:58.04 Of course by now, the technology has changed radically from
00:14:02.02 the first ones we had. We're now at least at third or fourth
00:14:07.09 generation of sequencing, the machines and technologies
00:14:10.24 out there. And whereas, originally it was one base a month,
00:14:16.17 it's now billions and billions of bases in an afternoon.
00:14:19.13 I find it sort of unimaginable to have been part of this entire
00:14:25.27 effort, but in fact, to have seen it go from nothing. At the time,
00:14:30.09 we were doing this first DNA sequence, Jim would be
00:14:34.14 only too willing to assure me that people would never be able,
00:14:37.19 off in a decade, be able to sequence genes.
00:14:41.10 It gives an amazing example of how fast science is moving
00:14:48.04 these days.
00:14:48.22 That's right.
00:14:49.21 Especially in the biological sciences.
00:14:52.26

This Talk

Speaker: Walter Gilbert

Audience:

General Public
Educators of H. School / Intro Undergrad
Student
Educators of Adv. Undergrad / Grad
Researcher
Educators

Recorded: July 2013

Talk Overview

Walter “Wally” Gilbert talks to Ron Vale about developing one of the first methods for sequencing DNA. In the 1970s, Gilbert sought to decipher the lac repressor DNA binding sequence. It took him and his colleague two years to sequence 24 bases. Later, Gilbert realized that he could use chemical agents to both modify and break DNA at different nucleotides, creating fragments that he could separate by electrophoresis. This resulted in a sequencing pattern which could be used to determine long sequences of DNA. Because of this technique, along with the Sanger method, DNA sequencing took off. Gilbert was awarded the 1980 Nobel Prize in Chemistry for this discovery.

Speaker Bio

Walter Gilbert

Walter Gilbert

Walter Gilbert is a Professor Emeritus at Harvard University. Gilbert has made many significant contributions to biology, including being the first to propose the existence of introns and exons and the RNA world hypothesis, and making major breakthroughs in elucidating the mechanism of mRNA transcription and translation. Gilbert developed one of the first methods for… Continue Reading