Whether the regulatory protein of an inducible operon for a catabolic pathway is a positive or a negative element in the control of expression is correlated with the frequent or infrequent presence, respectively, of the system's substrate in the natural environment.In several of the best-studied inducible operons of enteric bacteria, the regulatory protein acts as a negative (repressor) element in the control of gene expression. Examples are the galactose (1), glycerol (2), histidine utilization (3), and lactose (4) operons. In other well-studied inducible operons, like arabinose (5), maltose (6), and rhamnose (7), the regulatory protein is a positive (activator) element. These differences are represented schematically in Fig. 1. I will show that one can analyze the functional implications of these differences and predict that galactose, glycerol, histidine, and lactose are seldom in the organism's natural environment, whereas arabinose, maltose, and rhamnose are frequently present there. These predictions are confirmed by the absorption patterns for these substances in the mammalian intestine.The relative merits of genetic control by repressors and activators undoubtedly have been considered by many investigators. However, there have been no systematic attempts to compare these mechanisms on the basis of function and thus explore the reasons for their selection in nature. Clearly, the two types of mechanisms respond in opposite ways to mutations affecting (i) the synthesis of regulator, (ii) the degradation of regulator, and (iii) the mutual recognition of regulator and its binding site on the DNA. But there could be other functional differences, not necessarily intuitively obvious, responsible for the selection of these types of mechanisms. To examine this possibility one cannot compare directly two systems like the lactose and arabinose operons because of all their differences that are irrelevant to the comparison of the control mechanisms per se. Ideally, one would like a controlled comparison in which the two systems are identical in every respect except for the differences in the type of control mechanism utilized. Although this is difficult to do experimentally, it can be simulated by appropriate mathematical analysis. I have carried out such an analysis, which confirms the obvious behavioral differences mentioned above, but more importantly suggests that they are the only inherent differences between repressor and activator control (8). Therefore, it is these three kinds of mutations that must be examined further if we are to understand the natural selection of genetic control mechanisms.
Natural mutational tendenciesOf the possible mutations affecting the synthesis and degradation of regulatory proteins, those that reduce the rate of synthesis and those that increase the rate of degradation presumably are the most common. It is reasonable to expect that only certain specific nucleotide sequences can result in a promoter with increased efficiency, whereas many different sequences could be promo...