X-RAY crystallographic studies have indicated that four dehydrogenases1–4, including lactate1 and glyceraldehyde phosphate4 dehydrogenase, contain a central parallel β sheet consisting of six strands flanked by four α helices. This super-secondary structure5 is called the dinucleotide fold or nicotinamide adenine dinucleotide (NAD) domain since it forms an NAD-binding site at the C terminus of the β sheet. The dinucleotide fold can be considered as two roughly identical mononucleotide domains each containing about 60 residues arranged in an alternating βαβαβ sequence and related by an approximate twofold axis5. The aromatic specificity site of subtilisin6,7 and the flavin-binding site of flavodoxin8,9 seem5 to be formed by a secondary structure very similar to a mononucleotide domain of the dehydrogenases. Crystallographic studies have shown that the structures of phosphoglycerate kinase10,11, hexokinase12, adenylate kinase13, phosphoglycerate mutase14 and triosephosphate isomerase15 also contain β sheets containing at least five strands flanked by at least three α helices. The nucleoside phosphate or, in the case of the mutase and isomerase, the sugar phosphate-binding site for each of these enzymes is either known10–12, or strongly suspected13–15 to be located at the C terminus of the β sheet. Differences occur, however, in some of these supersecondary structures relative to the NAD domain of the dehydrogenase family as regards the direction of the β strands12, the location of the binding site relative to the plane of the β sheet12, and most significantly, the sequential order or connectivity of the individual strands in the β sheet12–14. Whether these related supersecondary structures represent convergent5,12 or divergent evolutionary processes4,16, or both17, is a topic of active debate.
Read full abstract