[MUSIC] Today we are going to discuss nonsense, nonsense codons. Before we do that, I'd like to remind you a very, very simple basic fact. A gene is made of DNA. Here is one strand. Here is your other strand. These DNAs copy into an RNA, messenger RNA. So that's the DNA. That's the messenger RNA. And that is, in turn, copied into a protein. So with this very simple picture, you don't really need a beginning and an end. Because the molecule of DNA, it starts at the beginning of the gene and ends at the end of the gene. That's the very simplistic view. We know for instance that small organism, like viruses, like phage T4, have many genes on one molecule of DNA. So we know that the DNA is much longer than what is shown here. Still, you are making one message per gene, in the first approximation, and one protein per message. That means that there has to be, on the DNA, information to tell the machine whether you start. How come I start here to make them RNA? And then, where do I stop making the RNA? This is called a promoter, and this is called, at the end of the message, this is called a terminator. Promoter and terminator apply to transcription of the RNA from the DNA. On the RNA, the RNA is in fact, almost, always larger than the piece that codes for the protein. So on the RNA, you'll also have to have signals. We've seen last time with the Crick paper on the genetic code, that the code is read from a start point, start, by three consecutive residues. Called, that make a triplet, or a codon, that called for a single amino acid. So a triplet, which I will summarize like this. This is three nucleotide codes for one amino acid. That's very, very, very simple. So we need a start. And today we're going to concentrate not on the start, but on the stop, on the end of the coding information. There is a start which, defines where you put the first amino acid, and there is a stop, which defines where you arrest translation of the message. Start and stop. Now, the existence of a stop codon, which is very often called nonsense codon, is in fact the discovery of these codons. Results from a piece of luck. The piece of luck is the fact that E coli K12, originally isolated in Stanford Hospital from a health recovering patient, from the stool of a recovering patient. This E coli K12, has three fundamental properties that were extremely helpful for the scientists, who made molecular biology.. One of these property, is the presence of a factor, called fertility factor, which is called F, and we will discuss that in two weeks from now. The other is the presence of a lysogenic virus called lambda, which we'll discuss toward the end of the class, the course. And the last piece of luck is the fact that this particular strain had a mutation that is not harmful when the bacterium is growing. That allows the bacterium to sit and stay quietly in what is called a stationary phase. But it survives less well in the stationary phase. The gene in which this mutation is present is called rpoS and it's a component of the transcription machine. So the K12 strain has a mutation, this rpoS. Now we know that this mutation is a stop codon. So I'm going to put the start codon here, which is AUG. And the stop codons, I'm going to put in them in red again. The three stop codons in their classical code, UAG, UAA and UGA. Normally, in the rpoS gene, there is a residue that is codes for amino acid, that is a GAG. That's a normal codon in the rpoS gene of practically every antiviral bacteria, except K12. In K12, this, because of a mutation, becomes a stop code on UAG. This is not lethal, otherwise we wouldn't have K12. But this provides a selective force for avoiding the effect of this mutation. And now, if you look at the various E coli K12 strains that have been derived since the 40s, by many investigators, very often using x-rays and pretty high doses of x-ray to induce mutation. It turns out that in these strains, three out of nine well-characterized strains have acquired something which we call a suppressor. It's a second mutation that avoid the detrimental effect of the first mutation. And so, some have suppressors and some have just gone back to the GAG, or to another codon that's acceptable, like a CAG codon. Now these suppressors were present without people knowing them. In many of the strains that were being used by different scientists. Of course, the people didn't not know about these suppressors. But they were around. And because they were around, they were picked. So we can go now, now that we've put the cast in perspective, we can go to the last paper by Benzer, and his student, Champe, which appeared in 1962. And which deals with this notion of nonsense. So why did Benzer work on this? Well he stumbled on this. You remember, probably, that Benzer had isolated many sites with mutations in the R2 locust, hundreds of sites. But what he wanted was to run the map of the gene into the ground. That means having a mutation every single nucleotide, if possible, using various mutagens. That was his goal at the time. When he came back from Paris to Purdue, he found that one of his colleagues, Tessman, Irwin Tessman, had been isolating R2 mutants. Not because he was interested in R2. Or in phage T4, but he was interested mutagens, in characterizing mutagens. So, basically, what Tessman had done was to use a various number, large number of mutagens to isolate R2 mutants. And so, Benzer says, why should I waste my time identifying new sites, maybe Tessman has new sites. So I can ask Tessman his mutants, and see whether they map at the same sites as my mutants, or whether they identify new sites. That's perfectly legitimate experiment to do. Now, Tessman gave him his R2 mutants, and some of them, one or two, I don't remember the number, but some of them, some of those did not behave like R2. They grew a lambda lysogen strain. That is not the phenotype expected for an R2 mutant. So, they grow on the K lambda from Benzer, but they are mutant on the K lambda of Tessman. Which, by all criteria available at the time, were identical strains or very similar strains. And he categorized several of the mutants and he discovered a category of mutant, which he called the ambivalent mutant. It was not the property of a single mutation or a single site. It was a property of many mutations. They were ambivalent because they grow on one host, and they don't grow on the other host. So here you see a little drawing of the R2 locus. And at that time, it was believed that the two genes were separate. It was not, there were two separate messenger RNA. And there were two separate proteins. There still are two separate proteins, but now we know that the message is, in fact, a message that comes from way back on the pheA genome, and extends a lot after the R2 locus. Message is much longer, it has no importance for the discussion of the paths today. So he knew that. He knew that all the R2A mutants have no effect on R2B. All the R2B mutants have no effect on R2A. And he had isolated and characterized the 1589 deletion, this magic deletion, that makes a chimeric protein. A bit of A, and a large piece of B. And this chimeric protein has the B, R2B activity. And Crick and his colleagues, if you remember, had actually shown that if you could cross mutations upstream of 1589, some frame shift would abolish the activity and, if you compensate for the frame shift, you recover that activity. So Crick had used the 1589 deletion to study frame shift, the effect of frame shift. What Benzer was going to do was to study the effect of point mutant, not insertion or deletion, substitution mutants on the behavior of 1589. So that's how it started. In the text of the paper, he starts by saying that the R There's a code. If the code is a triplet, the 64 possibilities, because they're four ATGC times 416 times 464. So the 64 possible codons. Now, you can imagine two extreme situation. The minimal situation, 20 codons code for amino acids, and the 44 others are nonsense. Or the 64 are sense and there's no nonsense. Those are the two extremes. Plus anything in between you want. So are they nonsense? And the prediction is based on this drawing, which shows you the 1589 deletion that makes a fusion protein, a chimeric protein with B activity, and what he did is he crossed them. Point mutation upstream of the 59, 1589 deletion. If the point mutation is a nonsense, creates a nonsense, this will block protein expression and you will not have R2B activity. If the mutation changes one amino acid into another amino acid, which is unacceptable for the R2A function. But, irrelevant when you make the chimeric protein. You should still have the activity. That's a prediction with a nonsense, no activity, with the sense, yes R2B activity. Very simple.