Once biochemists could assay and purify an RNA polymerase from E. coli, it was important to know the biological role of the enzyme. For example, the bacterial cell might possess three different kinds of RNA polymerase: one for the synthesis of messenger RNA, one for the synthesis of tRNA, and one for the synthesis of ribosomal RNA. If that were the case, much effort could have been wasted in studying in vitro transcription from a gene if the wrong RNA polymerase had been used. Enzymologists’ failures to find more than one type of RNA polymerase in E. coli were no proof that more did not exist, for, detection of an enzyme in cells can be difficult. In other words, what can be done to determine the biological role of the enzyme that can be detected and purified?

Fortunately, a way of determining the role of the E. coli RNA polymerase finally appeared. It was in the form of a very useful antibiotic, rifamycin, which blocks bacterial cell growth by inhibiting transcription initiation by RNA polymerase. If many cells are spread on agar medium containing rifamycin, most do not grow. A few do, and these rifamycin resistant mutants grow into colonies. Such mutants exist in populations of sensitive cells at a frequency of about 10^-7. Examination of the resistant mutants shows them to be of two classes. Mutants of the first class are resistant because their cell membrane is less permeable to rifamycin than the membrane in wild-type cells. These are of no interest to us here. Mutants of the second class are resistant by virtue of an alteration in the RNA polymerase. This can be demonstrated by the fact that the RNA polymerase purified from such rifamycin-resistant cells has become resistant to rifamycin.

Since rifamycin-resistant cells now contain a rifamycin-resistant polymerase, it would seem that this polymerase must be the only type present in cells. Such need not be the case, however. Consider first the hypothetical possibility that cells contain two types of RNA polymerase, one that is naturally sensitive to rifamycin, and one that is naturally resistant. We might be purifying and studying the first enzyme when we should be studying the naturally resistant polymerase. This possibility of this situation can be excluded by showing that rifamycin addition to cells stops all RNA synthesis. Therefore cells cannot contain a polymerase that is naturally resistant to rifamycin. A second possibility is that cells contain two types of polymerase and both are sensitive to rifamycin. Because mutants resistant to rifamycin can be isolated, both types of polymerase would then have to be mutated to rifamycin resistance. Such an event is exceedingly unlikely, however. The probability of mutating both polymerases is the product of the probability of mutating either one. From other studies we know that the mutation frequency for such an alteration in an enzyme is on the other of 10^-7. Therefore, the probability of mutating two polymerases to rifamycin resistance would be about (10^-7)² (Fig. 1), which is far below the frequency of 10-7 that is actually observed.

Fig1. Mutation of a single RNA polymerase to rifamycin resistance occurs at a frequency of 10^-7, whereas if two independent mutational events are required to mutate two different polymerases to rifamycin resistance, the frequency is 10^-7 × 10^-7= 10^-14.

Thus far, then, we know these facts: The target of rifamycin is a single type of RNA polymerase in bacteria. This RNA polymerase synthesizes at least one essential class of RNA, and this polymerase is the one that biochemists purify. How do we know that this RNA polymerase synthesizes all the RNAs? Careful physiological experiments show that rifamycin addition stops synthesis of all classes of RNA, mRNA, tRNA, and rRNA. Therefore the same RNA polymerase molecule must be used for the synthesis of these three kinds of RNA, and this RNA polymerase must be the one that the biochemists purify.

Unfortunately there is an imperfection in the reasoning leading to the conclusion that E. coli cells contain only a single type of RNA polymerase molecule. That imperfection came to light with the discovery that the prokaryotic RNA polymerase is not a single polypeptide but in fact contains four different polypeptide chains. Therefore, the rifamycin experiment proves that the same polypeptide is used by whatever polymerases synthesize the different classes of RNA. Much more arduous biochemical reconstruction experiments have been required to exclude the possibility that bacteria contain more than one single basic core RNA polymerase.