The Genetic Code, Part I

You may be familiar with the cryptic code puzzles that appear on the comic sections of newspapers. They involve some famous or amusing quote that has all of the letters in it substituted for other letters. For example, all the A’s may be substituted with W’s, all the V’s with P’s, etc. The substitutions are always consistent, and the puzzle is solved simply by figuring out the correspondence of letters. This code is very simple, because it just involves finding the equivalents of single symbols, which are both in the same alphabet. Unfortunately, the code needed to direct protein synthesis has to be more complex than that.

The first reason our code needs to be more complex is that DNA and proteins are very different kinds of molecules. Even if we could imagine a code that establishes a relationship between the bases in the nucleic acid and amino-acids, how could this correspondence work on a molecular level? Is there some molecular mechanism that could predict an interaction between a particular amino-acid, and one or more particular nucleic acid bases?

The fact that proteins and nucleic acids have different biochemical properties suggests that a direct molecular correspondence is unlikely to occur. What we need is to have nucleic acids and proteins communicated through a translator, who speaks the language of both. Francis Crick was the first to propose this solution.

Crick suggested that there must be a molecule with two functionally different ends. On the one end there must be a mechanism for attaching a specific amino-acid, and at the other end, there must be a mechanism for interacting with a specific sequence of nucleotide bases. Crick was correct indeed. There are such molecules. These are small, highly specialized, strings of RNA, called “transfer RNA” (tRNA).

Three-Letters Words

The second reason the genetic code needs to be more complex than newspaper puzzles is that there are 20 kinds of amino-acids, but only four kinds of bases in nucleic acids. We have A, G, T and C, and that’s it. If we had a simple substitution that was one for one, we could only have a code that was specific for four different amino-acids. We have 20 of them to account for, however.

What that suggests is that sequences of more than one nucleotide must be used to code for a single amino-acid. This would be like having the bases of a nucleic acid combined together to form code-words, where each base is a single letter. How many letters long must each word be?

Imagine that we had code-words that were made up of two letters each. That would not be sufficient to code for all the amino-acids. This is simple math. If we have four things, and we combine these four things in pairs, then we can make 4 squared combinations. That’s sixteen combinations. However, if we make combinations of three things, we would have 4 cubed combinations, that’s 64, more than enough. Actually, a three letter code suggests that the code has some sort of redundancy.

The logic in support for having a three letter code seems pretty obvious, but it simply suggests a testable hypothesis which then had to be demonstrated. Once again, it was Francis Crick along with colleagues who demonstrated experimentally that the genetic code must involve sequences of three bases. Crick used a technique in which they could cause a very particular kind of mutation in the DNA of a virus. This involved the elimination of just one base-pair from a DNA double helix. Alternatively, the mutation involved the addition of just one base-pair. If they applied this treatment in the appropriate fashion, they could be assured that either one base pair, or one was added.

Consider the simple sentence: “Old men are fun”, but without the spacing. Consider it as a string of characters. That sentence is composed of four words, each specified by three letters. If we delete the first letter, and try to read it, we’ve got: “ldmenarefun”. If we deleted two letters, we’d have: “dmenarefun”. Neither of these strings of characters makes any sense because we have shifted the place we start reading by one or two positions. The resulting remainder of the string becomes nonsensical.

Now, if we delete three letters, the words make sense again: “menarefun”. It’s not the same sentence as we’ve started with, but that’s not the point. The point is that the words in that sentence, comprised of three letters, only make sense if we take out three letters. We could do the opposite. If we add three letters, part of the string would make sense.

Crick used the same kind of logic, deleting or adding just one base-pair from a viral DNA, to show that there must a three letter code. If they induced one mutation in a gene, then all of the amino-acids coded by that gene would get changed. If they eliminated two base-pairs, they’d have the same results. If they deleted three base-pairs, however, they found that the remaining portion of that gen would make sense, in the sense that most amino-acid sequences would be intact.

It was pretty clear that the genetic code had to be made of three bases each. Many experiments since that have used a variety of techniques to verify the existence of this three letter code. The three-base sequences that serve as fundamental units for the code have come to be known as codons. Each codon corresponds to a unique amino-acid.

The next step was to establish what the correspondence is between particular codons and particular amino-acids. I will leave that for next time.

DNA Replication, Part II

Watson and Crick’s semi-conservative model contrasted with a couple of other possibilities for how DNA could possibly replicate. There is the conservative model, which suggests that both strands in the original DNA double helix stay together during replication. Then there is the dispersive model, which suggests that both strands are not only separated, but even broken up into smaller pieces during replication. Deciding which of these models was the correct one seemed to be pretty easy, because they make very different predictions. It was not obvious how to prove it in the laboratory, however.

The prediction of Watson and Crick’s semi-conservative hypothesis is that each of the two daughter double helixes, after one round of replication, should be made up of one old strand and one new strand. The conservative model predicts that after replication, one of the double helixes that result would be entirely old. The daughter helixes would be entirely new. Finally, the dispersive model predicts that both the daughter and parental helixes would be made up of just a mixture of old and new DNA.

This sounds simple, but the difficulty was figuring how to actually test that. We need some way to be able to determine what’s old and what’s new DNA after replication. It took several years before anybody figured out how to do this.

Meselson-Stahl Experiment

This brings us to a pair of researchers, Matthew Meselson and Franklin Stahl. In 1957, a few years after Watson and Crick’s work, they came up with a novel method for distinguishing new and old DNA during replication. Let me explain how they did that.

First, they grew bacteria in two different kinds of culture media. One of these culture media had normal nitrogen in it (N14). The other media had a heavier isotope of nitrogen in it (N15). This isotope of nitrogen is not radioactive, it’s just a little bit heavier. Not much heavier, just a little bit.

The point of culturing bacteria in these two different media is that the nitrogen in those media would be taken up and incorporated into any new biological molecules that were being synthesized. Specifically, the nitrogen would be taken up and incorporated into any new DNA that was being synthesized.

If you culture bacteria for some period of time, what would be many generations, then you can assume that all of the nitrogen that is incorporated in that DNA would either have N14 or N15 depending on the culture media in which you are growing it. In this way, Meselson and Stahl could essentially label old and new DNA by how heavy (dense, really) that DNA was.

As you can imagine, the density difference between DNA that had been made with N15, as compared to DNA that was made with N14, is really small. The really clever part was figuring out how to very accurately measure the densities of these kinds of DNA.

The Clever Part

To do this, Meselson and Stahl devised a new kind of procedure called “density gradient centrifugation”. This density gradient centrifugation allowed them literally to sort out DNA according to how dense it was.

The idea behind this is actually similar to the reason why swimmers don’t sink in the Great Salt Lake. If the density of a liquid and an object in it are more or less the same, then the object would neither sink nor float, it would just sort of stay where you put it. So, if you add a lot of salt to water, actually it becomes of the same density as our own tissues, and you don’t sink in it, you just sort of stay there.

It is more interesting, though, if you have a gradient of densities. In other words, if you have some range from high to low densities in some liquid mediums, then objects of slightly different densities would sort themselves out. The objects would end up at that gradient at exactly where their own density matches the density of that point in the density gradient.

That’s the idea that Meselson and Stahl had. How do you create a density gradient? After trying a number of different kinds of solutions, they found a compound called Caesium Chloride. This is a salt that when it is put into a solution has approximately the same density as DNA. What they then did was take a tube of caesium chloride solution and centrifuge it. If you centrifuge a tube, what happens is that the heavier stuff goes to the end of the tube, and the lighter stuff stays at the top.

What Meselson and Stahl had to do with this experiment was centrifuge the ceasium chloride solution enormously quickly. They actually spun it around so fast that they created a 100000 g-forces. This is really fast, so it is called ultra-centrifugation. They did it for a number of days. At the end, they would get as a density gradient of the caesium chloride along the tube. Remember, caesium chloride is about the same density as DNA.

That means that if you then take some DNA and put it in that tube and spin it around, the DNA would ordinarily just be dispersed in that tube because is more or less the same density as the ceasium chloride. As the density of ceasium chloride develops, however, the DNA would all coalesce in a single band. That band would be in a position along the length of the tube that corresponds exactly to the density of DNA at that point in the tube.

This method was so sensitive, that Meselson and Stahl determined that they could tell the difference between N15 and N14 DNA.

Watson and Crick Had it Right

So, that’s the technique. Armed with this technique, Meselson and Stahl then did the following experiment, which should be sort of obvious bases on what we talked about. They took a culture of N15 bacteria and transferred them to a culture flask that had N14. Now, those bacteria, when they started to replicate their DNA, would start incorporating the lighter nitrogen. Any new DNA produced by those bacteria would be lighter than the old DNA that they had.

They waited for about 20 minutes, which is long enough for just one round of DNA replication. Then they took the bacteria out, extracted the DNA from them and used their density centrifugation method to determine what the densities of the DNA in the sample was.

The conservative model predicts that at this point there should be two separate sets of DNA. There should be lighter DNA and heavier DNA. The new DNA is going to be lighter and the old DNA is going to be heavier. This is because, according to this model, the parent strand stays intact. After one replication, you should have some DNA that is heavy, and some DNA that is all light.

This is not the result they observed. What they saw was just one intermediate band. So, they could rule out the conservative model directly. They couldn’t rule out the dispersive model, thought. After just one round of replication, both the dispersive and semi-conservative models made the same prediction. Each daughter double helix should be composed of half old(heavy) DNA and half new(light) DNA. All of the DNA in the sample, after one replication, should be at some intermediate weight. This is what they saw. The dispersive hypothesis made exactly the same prediction.

If you wait for two replications, however, all of a sudden you get a distinct difference in the predictions made by the semi-conservative and dispersive models. After two replications (about 40 minutes), the dispersive model would still predict there would be only one band, all of new and old DNA is all mixed up. Therefore, all of the DNA would be about the same density. The thing that should change is the position of that band along the gradient.

What Meselson and Stahl saw was the creation of two bands after two rounds of replication, which confirmed the prediction of the semi-conservative model.

Well, it took some years to prove it, but Watson and Crick had it right.

DNA Replication, Part I

Lately I’ve been writing a lot about DNA, its history and structure. I think this is really important, I would say key, to understanding what is life about, and how it evolves. Here I’ll continue with this business. Here I want to look at the proposal Watson and Crick had for how the double helix might be replicated because of the complimentary base-pairing they discovered. This would be a series on DNA replication, which I think is one of the most fascinating and complex processes in the universe.

Watson and Crick suggested that the DNA molecule must unzip, and then, each half of the molecule could serve as a template for a newly formed half. This is a good hypothesis, but was it correct? As Watson and Crick proposed that, there were two alternative hypotheses on the scene.

The Alternative Hypotheses

The first alternative suggested that the DNA double helix must remain completely intact when it is replicated. That is, the two strands do not separate. The entire molecule is somehow used as a template for making more DNA. This alternative was called the conservative hypothesis of replication. The original DNA double helix molecule remained completely intact and conserved. The idea was that there must be some intermediate molecule that got information from the structure of the helix and used it to build a completely different helix.

A second alternative suggested that the original DNA molecule becomes completely broken down during replication, with the newly copied DNA assembled by some unknown mechanism. In other words, the DNA double helix would actually be irrelevant. This alternative was called the dispersive model. It was called dispersive obviously because the DNA in the original helix just becomes dispersed and incorporated in the new copies that were being created.

Based on what was known about molecular biology and DNA in the 1950’s, both of these hypotheses were reasonable. Neither offered a solution to the problem of duplicating the exact order of nucleotides, however. This order is the information that we are seeking. Based on what we know today, both of these alternatives seem unlikely. They’re value then was to serve as alternative hypotheses against which to test specific predictions made by the Watson and Crick model.

The mechanism that Watson and Crick proposed became known as the semi-conservative model of DNA replication. This was called semi-conservative because it predicts that during replication, the double helix unzips and the new daughter helixes would both have one strand of the old helix. We begin with one helix, it separates somehow, and the resulting daughter helixes that are formed maintain the original halves of that parent helix. Upon this old half, new halves are formed to create the new double helixes.

This hypothesis led to a specific prediction. The prediction was that if you could know which was old and new DNA after replication would occur, all the old DNA that was in the original parental double helix, would now be dispersed between the two daughter helixes equally. The daughter helixes would all be composed of one half of old DNA, and one half of new DNA.

Let’s contrast that to the prediction we might have if we look at the other two models. Let’s think about the conservative model first. That model suggests that the DNA helix just remains intact once you’ve got it. After replication, that model would suggest that the two daughter helixes would separately made up of, on the one hand all old DNA, and on the other hand, all new DNA. The old DNA in the original parent is still in the original parent, and the daughter DNA helix is completely new.

The dispersive model made yet another prediction. That prediction was that the old DNA that was found in the original parental helix would just randomly scattered across the two daughter helixes.

We have three specific and different predictions that could be used to distinguish between these three models of replication. The trick is figuring out how to know what’s old and new DNA. Actually, it was several years after Watson and Crick’s original proposal that anybody could figure out how to experimentally test it.

The way to test the hypothesis was pretty obvious conceptually. Let’s ask ourselves, after replication, what happens to the material in the original parent’s helix? We have very distinct predictions. The problem was figuring out how to know where old and new DNA was. This is often it is in science. An idea comes forward, people understand what they have to do, but the critical experiment isn’t actually done until somebody comes along and says not what’s the experiment, but how you can actually do it. This may take years, as it did in this case.

Eventually, researchers figured out an extremely clever way to know what the difference is between old and new DNA. I will talk about this interesting and brilliant experiment in my next article.

The DNA Structure, Part II

James Watson was a young American, who had just completed his PhD. He was interested in protein structure. He moved to Cambridge, England, and began working with Francis Crick, who was a physicist familiar with x-ray crystallography and how to interpret it. The story goes that Watson happened to visit London for a seminar, and saw the x-ray diffraction patterns that Rosalind Franklin had obtained from Maurice Wilkins’ purified DNA. Watson made some notes, rushed back to Cambridge and told Crick what he had seen.

Using Franklin’s data, Watson and Crick were able to deduce a number of key structural elements about how DNA must be shaped. These are things they figured out by looking at those dots on the x-ray crystallograph. First, they learned that the molecule had to form some kind of helix. It had to have a kind of spiral structure, similar to the alpha helix that is characteristic of many parts of proteins. Second, they figured out that the width of this helix was about two nanometers. The interesting thing about this is that, this width was twice the width of what you would have expected if there was only a single helix. That gave them the idea that there had to be more than one helix. A double helix, perhaps.

Another thing they learned was what the regular spacing of the repeated patterning along the length of the molecule is. They saw that there was a repeating pattern at about 0.3 nanometers. This corresponded to the size of one nucleotide. Then there was a larger repeating pattern that was ten times that size. From this they inferred that the number of nucleotides that would occur when the spiral went around just once and returned to the same point in the spiral had to be about ten nucleotides.

That isn’t a lot of information. If I gave you that information you couldn’t tell me the structure of DNA. What Watson and Crick did was to use that data and set out to figure out the structure of DNA the old-fashioned way. They made physical models of the molecule with metal rods. They made large-scale models of DNA several feet tall.

They built many models and asked each time: when we have this model, does it all fit together? They tried over and over again. Eventually they came up with a model that fit. The trickiest part of the modeling was to figure out how the nitrogenous bases fit into the picture. Remember, a polymer of DNA is a repeating pattern of sugars and phosphates, with different nitrogenous bases hanging off the side (the guanines, the adenines and so forth). Where did they fit? If you had two, or even more molecules of DNA that were spiraling together, where do the nitrogenous bases go?

Well, after a couple of failed attempts putting the nitrogenous bases on the outside, Watson and Crick realized they had to go on the inside. Why they might want to put them on the outside? It is the nitrogenous bases that vary along the length of the molecule. It is the variation of the different kinds of nucleotides that must somehow involve the code. If we’re going to get access to that code, we’ve got to make what’s different about that code available to the outside world. They couldn’t get the backbones to work together in any way that made sense with the nitrogenous bases on the outside.

If they turned those nitrogenous bases in, and had the nitrogenous bases connecting with each other, forming kind of stairs, with the backbones of these molecules forming the stringers that are holding the steps; the molecule began to fit together. This actually made sense, because these nitrogenous bases are chemical repelled by water. They want to be on the inside of the molecule because of that.

There was one interesting additional problem. This is actually the most interesting part of the story. That is, how did the bases fit together? If the put the nitrogenous bases on the inside, they could get a double helical structure that began to fit the data, but there was still a problem remaining, there are two kinds of bases, the pyrimidines and the purines, and they are of different sizes. Purines have two rings, and pyrimidines only have one.

If you just try to put these stair steps across the two sides of the double helix, you’ll have some steps that are wider, and some that are narrower. For example, if you got two purines together, you’ll have a relatively narrow step. The outside of this spiral would be going in and out, which is not structurally stable. That’s were Chargaff's rule came in. They realized the implications of Chargaff's rule, which says that the amount of the base adenine (A), always equals the amount of thymine (T). Similarly, the amount of guanine (G), always equals the amount of cytosine (C). This suggested that it may be that one always pairs up with the other when they are matching together on the inside of the helix.

It turns out that when they looked at how these kinds of nitrogenous bases would match up, they found that those peculiar combinations (A and T, C and G) would always maximize that potential weak bonding that occurs between the bases. With this, they actually solved two problems. They figured out how you can have a regular distance along the whole length of the staircase. Also, they figured out what could hold the staircase together. If you always match A with T and G with C, the bonding that holds two sides together is maximized.


In April, 1953, which is only a year after people became convinced that DNA was the information molecule, Watson and Crick published a one-page paper in the journal Nature, which described the double helix. They described the molecular structure and how they thought it would all fit together. The real significance of this work was not simply to describe the 3D structure of DNA, but to show how that 3D structure might actually say something about replication.

Watson and Crick’s paper ends with the following sentence: “It had not escaped our attention that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material.”

The fact that A and T, G and C were always paired together, meant that if you took the two sides of the molecule apart, you would always know what the other side has to be. That’s called complementary base-pairing. It was this fact that suggested that mechanism by which DNA was replicated. I’m very excited about this subject, but I will leave it to another article.

The DNA Structure, Part I

After the work of Hershey and Chase, biologists in the early 1950’s became convinced that DNA was what they needed to look at to understand the genetic code. They actually had no idea how DNA could possibly act as a mechanism for genetic inheritance. Let’s step back and remind ourselves of what this molecule has to accomplish. It needs to do two things. First, it needs to have some way of providing a code that can store information about proteins.

The linear structure of DNA, with variable bases along the chain was consistent with the idea that could provide such a code. Proteins also are linear chains. So, you can imagine that there could be some sort of mapping of the pattern of one molecule in the pattern of the other. It wasn’t clear what this mapping could be, but they thought they could figure that out. I will talk about that code in another article.

Second, the molecule had to be able to replicate. If we are going to transmit genetic information, from one generation to the next, it won’t work unless we make duplicate copies of the code, so we can handle one copy of the blueprint to the offspring. The real problem was that it wasn’t clear how DNA could be replicated. The linear structure of DNA offer some hope for a code, but it didn’t offer a clue about how replication might occur.

This is where the race, literally a race, for discovering the three-dimensional structure of DNA began. Scientists were convinced they needed to know the three dimensional structure of DNA to understand replication. It was this impetus that led to the discovery of the now famous DNA double-helix, by Watson and Crick. This was arguably the most important finding in biology in the 20th century.

Why biologists should be interested in the three-dimensional structure of DNA? Biochemists had begun to understand, in the 1950’s, that the function of proteins could be understood by figuring out something about their structure. So, it was hoped that some aspect of the function of DNA, specifically how it was replicated, could be understood by deducing its structure.

What kind of evidence could you use to deduce the three-dimensional structure? The first kind of evidence came from a procedure known as x-ray crystallography. The positions of atoms in a crystal can be inferred from the pattern that they create when you shoot a beam of x-rays through that substance. The x-rays would bend around those atoms, and then, when you look at the pattern on the other side, you see lines and dots in particular orientations and spacing. From them, if you’re familiar with the procedure and very clever, you’ll able to deduce something about the relative positions and orientations of the atoms that make up that structure.

If you shoot an x-ray through a relatively simple crystal, you’ll get a very regular pattern. If you shoot an x-ray through a more complex material, like an organic molecule, you’ll get a much more complex pattern. It is pretty difficult to pull information form that pattern, and infer something about the way the molecule must be structured.

You may be asking, what are we talking about here, crystals or organic molecules? Well, even the most complex organic molecules, in the hands of a good biochemist, can be crystallized. In fact, that’s where the story starts. Maurice Wilkins, a biochemist working at King’s College in London, was able to produce a remarkably pure crystal of DNA. Working with Wilkins was a woman named Rosalind Franklin. She was an expert x-ray crystallographer. She took Wilkins’ purified DNA crystal and was able to get what was to that point the most clear and accurate x-ray crystallograph of DNA that had yet been obtained.

As I said, these patterns are quite hard to interpret. Nowadays we can use computers and algorithms to deduce the structure of proteins from this kind of data, but back then was pretty much painstaking hand calculation and educated guesswork. As it turns out, Wilkins and Franklin puzzled over their x-ray crystallograph trying to deduce something about the structure of DNA. They didn’t quite figure it out. By the time, another person arrived on the scene, James Watson.

Before I introduce Watson, I want to introduce another kind of clue. There was a scientist working at Columbia University named Erwin Chargaff, who discovered a peculiar thing about DNA. He found that if you took the DNA from any organism, and decomposed it into its component nucleotides, you always found a quite interesting relationship. You always found that the amount of the base adenine (A), always equals the amount of thymine (T). Similarly, the amount of guanine (G), always equals the amount of cytosine (C). So, if you take apart the DNA of any organism, you always will find that the amount of A equals the amount T, and the amount of G equals the amount of C.

This was a very curious relationship that became known as Chargaff’s rule. Next time we’re going to talk about James Watson and his codiscovery of the double-helix with Francis Crick.