592 PETER MITCHELL 



the interpretation of observations on enzyme distribution in relation to 

 membrane transport in micro-organisms. When we conceive the cell wall 

 as a molecular sieve, preventing the loss of enzymes that may exist entirely, 

 or only partly, in the free state at the surface of the plasma membrane (i.e. 

 in the periplasm), we must give serious consideration to the possibility that 

 the enzymes and catalytic carriers of the protoplasm may be poised in an 

 equilibrium which may favour their segregation in the periplasm, in the 

 plasma membrane, or in the endoplasm according to the satisfaction of 

 mutual affinities. The distribution of a given enzyme, and its status as a 

 relatively "soluble" or "insoluble" protein would thus be seen as an 

 expression of locational affinities for bonding the protein to compli- 

 mentary substratum sites in the cell. In general, it would be expected that 

 such bonding would be due to secondary valencies, but it is also possible 

 that primary valencies might sometimes be involved. The studies of Keilin 

 and King [42] on the reversible bonding of the soluble succinic dehydro- 

 genase in the insoluble cytochrome system of heart muscle lends support 

 to this conception. Such a conception has obvious potentialities for helping 

 to explain induction and repression of enzyme synthesis by a mass action 

 type of effect [43, 44] not only thought of as being due to equilibration of 

 nascent enzyme with enzyme-substrate or enzyme-(substrate analogue) 

 complexes, but also due to the equilibration of nascent enzyme with 

 enzyme-substratum and (enzyme-substrate)-substratum complexes. These 

 attractive ideas, which are related to those of Catcheside [45], hinge, 

 however, on the experimental demonstration of the locational affinities 

 between enzymes and substratum sites in bacteria. For reasons that I shall 

 explain in a moment. Dr. Moyle and I decided to study the distribution 

 of the a-ketoglutarate dehydrogenase activity in Micrococcus lysodeikticiis 

 as an example of the possible participation of locational affinities in 

 determining the cytological distribution of an enzyme. 



When Micrococcus lysodeikticns is ruptured by shaking with glass beads 

 or by gentler enzymic and osmotic methods, and the plasma membranes 

 are separated from the "soluble" or "protoplasm" fraction on the centri- 

 fuge in the usual way [see 11, 30], about half the a-ketoglutarate dehydro- 

 genase activity is found in the plasma membrane fraction. The amount of 

 enzyme activity attached to the plasma membranes is not dependent upon 

 the distribution of dialyzable cofactors, nor is it appreciably affected by 

 the extent to which the membrane material is diluted during its isolation. 

 One can therefore infer that the enzyme must be strongly bound to the 

 membrane structure. The fact that, nevertheless, about half the enzyme 

 activity is present in the "soluble" fraction shows either that there are 

 two a-ketoglutarate dehydrogenase proteins with different solubilities or 

 affinities for membrane and protoplasm components, or that there is only 

 one type of a-ketoglutarate dehydrogenase, which is a soluble protein, and 



