Matthew Meselson Harvard The Semi Conservative Replicat… — Transcript

Hello, I'm going to talk to you about the semi-conservative replication of DNA, not so much about the technical details, but about how Frank Stall and I ended up.

The first was about the structure, which was based on model building and a little bit of information from X-ray diffraction.

The X-ray diffraction certainly did not tell you the structure.

It told you something about the repeat distance along the helix, it told you that it was a helix and very little more.

The model building was model building, so the structure certainly wasn't proven, this was a proposal, and many people didn't believe it or maybe didn't even pay any attention to it.

The second paper proposed how the molecule might replicate, the two chains would separate, each would guide on its surface the formation of a new chain, and so you end up with two double helices, each one has one of the old chains and one brand new chain, that's called semi-conservative replication.

At Caltech, Max Delbruck, who knew Jim Watson, was in correspondence with him, was pessimistic about this replication scheme, pessimistic about semi-conservative replication.

The chains were wound around each other, so to get them apart, unless you break them, you'd have to unwind the basic double helix to get the two arms separate. Max thought this would be impossible, hydrodynamically impossible, and so he proposed that maybe the two chains.

He and Gunther Stent published a paper which proposed three different methods for DNA replication, semi-conservative as Watson and Crick had predicted, conservative in which, at least conceptually, maybe there was a way that a double helix could guide the formation of another double helix just like it nearby, and so you never break the single chains, separate the pieces and then put everything.

Pauling, Max was over in biology, and he told me about this problem, and it occurred to me that maybe one could do an experiment to find out the mode of replication of DNA based on the use of heavy isotopes.

Nevertheless, the idea was to label DNA with something heavy, I don't think I thought about, oh yes, it was deuterium I thought about at the time.

And then to label it by growing bacteria or phages in heavy medium, all light down to the bottom if it was all heavy and in the middle if it was half heavy and half light, that's an oversimplification, but that's the way I thought about the experiment at that time, this was around 1954 sometime.

I then went to Woods Hole as a teaching assistant to for Jim Watson, he had lived at Caltech the previous term, uh in the physiology course, and one day as a teaching assistant in that course, Jim Watson and Sydney Brenner, who were teaching the course, were in an upstairs room in the building called the Lily building at the Marine Biological Lab, and Jim looked through the window and pointed down at a tree under which was sitting a man who was, uh you couldn't tell from a distance, but tonic for himself, it was called the gin and tonic tree.

But Jim was saying this guy by himself, well I thought that was pretty rotten, ganging up on this poor guy, so I went down and introduced myself to him and told him what lay in store for him, and he told me what he was doing in phage genetics, and he would be at Caltech the next year.

So we agreed to try to do this experiment together, which was X-ray crystallography.

Finally, I got the X-ray crystallography done and we could start the experiment.

Frank and I decided that we should develop the method for doing this, density gradient.

To turn it on, to see what was happening, and to our amazement, a density gradient was forming before our very, very eyes, this was because the cesium ion is quite heavy, quite dense, and it tends to settle to the bottom in a powerful centrifugal field.

And diffusion wants it to go back up to redistribute it, and at equilibrium you have a density gradient in which there's more cesium chloride near the bottom than near the top, and so the density gradient forms automatically, we didn't expect this, we saw it happen, we thought we'd have to make a pre-formed density gradient in the centrifuge cell, but the centrifuge itself makes the density gradient.

So to make a long story short, the bacteria at one generation had only one kind of DNA, it had a density halfway between heavy and light.

And at the second bacterial generation, there were two kinds of DNA, half heavy and fully light.

Now we had to write this paper up, actually we were rather sluggish in writing it up, and so Max Delbruck took us to the Marine Lab of Caltech at Corona Del Mar and locked us up in a tower room.

But there was a question, how to write this up, there were two ways that we discussed.

One way is to start with a hypothesis, the Watson and Crick hypothesis.

And say here's a test, we'll do this experiment and see if it works out the way they said it should.

The other way would be to write up exactly what your experiment said, no more, no less, without reference to any hypotheses at all, and then at the very end say whether it may agree with some hypothesis.

We chose the last way, partly because we thought, well, if you're trying to test a hypothesis, the only way to really be sure that you're going to be right is to know all possible hypotheses, and since nobody can know all possible hypotheses, just because your experiment might agree with one of them, doesn't prove that that's the right hypothesis, there could be another one that your experiment agrees with.

So it seemed to us more elegant to write our paper up in terms of sub-units and only at the end say, ah, these sub-units could be the single chains of the Watson and Crick.

No, which of course they were.

So that's what we did, and this diagram here shows the result in terms of sub-units, not DNA chains.

What we did was blessed by a lot of accidents.

When the DNA structure was proposed, a lot of people didn't believe it.

And there were not a lot of references to it in the literature for the first several years after 1953.

After all, it was based on model building, which doesn't prove anything.

It looked so beautiful that some people were convinced that it had to be right because it looked so right.

Other people, I think, were convinced that it couldn't be right because it looked too good to be true.

In any case, it was just a model.

The evidence from the X-ray diffraction was really not very supportive.

It showed that the repeat distance was a certain distance and that it was helical.

But you couldn't do it.

Here's this double helix.

Standing there, and it's saying, here I am.

I have two chains, go find out how I replicate.

Here I am, I have four kinds of bases.

Go figure out how that is converted into making proteins.

I sit in the nucleus, proteins are made out in the cytoplasm in eukaryotes.

Go figure out what it is that takes this information from me out to the cytoplasm.

My information being made up of these bases changes once in a while.

That's called mutation.

Go figure out how that happens.

Every once in a while, there's something called genetic recombination.

Go figure out how I'm.

What I'm trying to say here is that if I showed you a model of say a polysaccharide.

Would it tell you what experiment you should do next?

This molecule, DNA, is like the Wizard of Oz, standing there, except unlike the wizard, this is a true molecule.

Standing there and telling you essentially the whole agenda for the future of science for the next 20 years.

That's a completely different mood and attitude, I think, from the science that went before it.

And the biological science that came after it.

So that was the characteristic of this time.

It was the DNA molecule telling you what the problems were.

What you had to go out and solve.