Evolution of Proteins in Nature by Domain Swapping

The following sections explore nature s use of domain swapping to evolve new function. These include the formation of multifunctional proteins, tandem duplication, domain recruitment, and cicular permutation (Fig. 1). The evolution of several enzymes in the purine (Fig. 2) and pyrimidine (Fig. 3) de novo biosynthetic pathways, as well as other enzymes, are discussed as illustrative examples. 

Multidomain proteins tend to occur more frequently in eukaryotes than in prokaryotes. Often the eukaryotic counterpart to a set of individual prokaryotic enzymes that catalyze successive reactions is a single, multidomain protein. The theoretical advantages proposed for such an arrangement include (1) a geometry for the direct transfer of substrates from one active site to another, in a process known as substrate channeling, in order to increase the overall flux of the pathway, (2) the protection of intermediates that may be unstable in aqueous environments or may be acted on inappropriately by other enzymes, (3) the facilitation of interactions between domains for purposes of allosteric regulatory functions, and (4) the establishment of a fixed stoichiometric ratio of the 

CPSase of arginine biosynthesis anthranilate synthetase component II p-aminobenzoaie synthetase GMP and CTP synthetases  

CPSase catalyzes the formation of carbamyl phosphate from glutamine, bicarbonate, and two equivalents of ATP. The biosynthesis involves four partial reactions. GLNase catalyzes the formation of ammonia from glutamine. The remaining three partial reactions are catalyzed by SYNase. Bicarbonate is activated by ATP to form carboxyphosphate, which reacts with ammonia to form carbamate. The ATP-dependent phosphorylation of carbamate results in the production of carbamyl phosphate. 

This begs the question as to why the SYNase domain has evolved by tandem duplication, since both SYN.A and SYN.B seem to be functionally equivalent. However, evidence that the two ATP-dependent reactions occur at different sites in SYNase is very strong (Guy et al., 1996), and it has been subsequently shown that the functional form of SYN.A and SYN.B domains in the absence of the other is a homodimer (Guy et al., 1998). It appears that it is somehow advantageous to have separate, but 

---