After the discovery of DNA polymerase by Arthur Kornberg, the properties of the enzyme became quite well known. One of the most critical is that DNA polymerase (in fact all known DNA polymerases) synthesize DNA in one direction only: 5' to 3'. This fact led to a dilemma regarding how the semiconservative model would work for a DNA molecule.
Reiji Okazaki was a brilliant experimenter who took on this problem. As an aside, Okazaki was born near Hiroshima, Japan, in 1930. He was a teenager there at the time of the explosion of the first of two nuclear bombs that the US dropped at the end of World War II. Reiji's scientific career was cut short by his untimely death from cancer in 1975 at the age of 44, perhaps related to his exposure to the fallout of that blast.
Okazaki reasoned that there were three possibilities for replicating a double-stranded DNA molecule, shown in Figure 20.6 of your text:
The continuous model is impossible, based upon the nature of DNA polymerases (replication in only one direction). Therefore he needed to demonstrate that one of the other two models was actually taking place.
Any discontinuous synthesis requires that there be, at least transiently, small pieces of DNA in a replicating structure. Okazaki decided to look for these small pieces. He employed the ultracentrifuge to do this. This time, the separation is based upon size, so that he could see these smaller pieces and also follow what happens to them during replication.
Let's look at the Svedberg equation again:
This time, notice that the relative motion of the molecules (the S value) is directly related to the size (M). In fact molecules that have the same shape (same frictional coefficient, f) will separate solely as a function of their size. This is called velocity gradient separation. In order to assure that all his molecules had the same shape, Okazaki denatured the DNA with alkali.
His experimental strategy was to use a pulse-labeling technique. In this case, a culture of bacterial cells infected with a bacterial virus is given radioactively labeled DNA precursor. This is a different kind of gradient than before. In this case, using sucrose, the DNA molecules never find their equilibrium position (sucrose solutions are much less dense than CsCl solutions) and so the molecules are always in motion. You have to stop the ultracentrifuge at an experimentally determined time to do the experiment. In this case, only DNA synthesis that has taken place during the time of the pulse will produce radiolabeled molecules that can be located in the gradient.
Some of Okazaki's data is shown in Figure 20.7 in your text:
On the left you see that at very short times of labeling (short pulses) very short pieces of DNA are found (2 sec, 7 sec, 15 sec). However, with longer and longer times, the pieces of DNA get increasing longer (120 sec). He then tried the same experiment with a mutant virus that was defective in a gene called DNA ligase. We will see that this is the enzyme that joins pieces of DNA together into larger structures. In this case (on the right) the labeled pieces of DNA remained short, even after long times of radiolabeling.
These data suggested to Okazaki that DNA replication occurred by the synthesis of small pieces that were later linked together by DNA ligase into larger pieces. In order to prove this, he did what is called a pulse-chase experiment. He radiolabeled his uninfected bacterial culture for a short time, and then followed this by adding a large excess of unlabeled precursor. This resulted in a great decrease in the amount of radiolabel incorporated and allowed him to follow the fate of the short pieces. One of his figures from his 1968 Cold Spring Harbor paper is shown here:
In this case a 30 second pulse (A) is followed by a 5 minute chase (B) and then by a 60 minute chase (C). The data show that the DNA that was labeled during the 30 second pulse eventually winds up in very large DNA, equivalent to the size of the genome of the bacterial cell in this case (shown with a 14C label as a marker).
Okazaki concluded that DNA replication proceeds by a discontinuous mechanism. His data actually suggested that both strands are copied discontinuously. It wasn't until he used a mutant deficient in a particular repair process (uracil excision) that he understood that fragments of one strand produced by this repair had nothing to do with the actual replication process. The small fragments that can be observed during the short radiolabeling periods are called Okazaki fragments in his honor.