Genes for bacterial and mitochondrial ATP synthase.

The structures of the ATP synthases in bacteria and mitochondria are very similar (for reviews see Senior, 1979; Fillingame, 1981 ; Racker, 1981 ; Walker et al., 1982). They are membrane-bound enzymes composed of two distinct domains: a domain called F, which is buried in the membrane and attached to it a domain known as F, which lies outside the membrane (see Fig. 1). F, is composed of five different polypeptides, a, b, 7, 6 and E, and bacterial Fo is made up from three proteins, a, b and c. Mitochondria1 F, may be more complex, but contains counterparts to the bacterial subunits. The catalytic sites of the enzyme are in F, and Fo has a proton channel through which the membrane proton-potential gradient generated by electron transport is coupled to phosphorylation of ADP. In mammals, all of the subunits of the enzyme are encoded in nuclear genes except those for subunit a (also called ATPase-6) and for a smaller protein that appears to be important for assembly of the enzyme complex (Macreadie et al., 1983). These two proteins are encoded in mitochondrial DNA (Anderson et al., 1981,1982). A similar arrangement is found in yeast (but not Neurospora), except that the gene for the c subunit (also called the proteolipid or DCCD-binding protein) is also in mitochondrial DNA (Macino & Tzagoloff, 1980). The genes for the Escherichiu coli enzyme are grouped in an operon, the unc operon (Walker et al., 1983). As illustrated in Fig. 1 this arrangement is related to the structure of the enzyme in so far as genes for Fo components a, c and b are clustered at the promoter proximal end and are followed by genes for F, subunits in the order 6 :a :y : b : E. It is noteworthy that the order : E is also found in maize and spinach chloroplast DNA (Krebbers er al., 1982; Kurawski et al., 1982). Clustering of genes with related functions is a feature of the genetic maps of coliphages such as T4 and lambda (Wood & King, 1979). For example, in lambda, morphogenetic genes for head proteins form a cluster adjacent to a cluster for tail genes (Katsura, 1983); in T4, genes for base plate, head and tail also form clusters. It has been suggested that these gene clusters might reflect the evolutionary origin of present-day assembly genes by tandem duplication and divergence of ancestral assembly genes (Casjens & Hendrix, 1974). In addition, or alternatively, clustering could reflect the selective advantage of minimizing recombination between genes for proteins that interact structurally, so as to decrease the probability of non-viable hybrids in interstrain matings (King & Laemmli, 1973; Casjens & Hendrix, 1974; Kikuchi & King, 1974). In lambda the similarity between the order