Protein Rossmann fold

As already discussed in Chapter 11, there are more than 10 000 protein structures known but only about 30 3D structure types. This might be traced to a limited number of possible stable polypeptide structures but most probably reflects the evolutionary history of the diversity of proteins. There are structural motifs which repeat themselves in a multitude of enzymes which are otherwise neither structurally nor functionally related, such as TIM barrel proteins, four-helix bundle proteins, Rossmann folds, or a/j3-folds of hydrolases (Figure 16.1). 

A more detailed breakdown of the fold abundance by individual genomes shows the same trends, as well as a number of unique features (Fig. 6, see color insert). The latter include, for example, the marked overrepresentation of Rossmann-fold domains in Mycobacterium, flavo-doxins in Synechocystis and methyltransferases in Helicobacter. Furthermore, the differences in fold distribution between the multicellular eukaryote Caenorhabditis elegans and the unicellular yeast become readily apparent. In the nematode, the protein kinases are the most common fold, with the P-loops relegated to the second position in contrast, the yeast distribution is more similar to that seen in prokaryotes (Fig. 6). 

The class III deacetylases, named sirtuins, are structurally and functionally different from other HDACs. In contrast to the zinc-dependent deacetylation of classic HDACs, sirtuins depend on NAD" to carry out catalytic reactions. A variety of sirtuin crystal structures have been published over the past few years. The structures of human Sirt2 and SirtS as well as several bacterial Sir2 proteins could be derived, whereas no 3D structure is available for Sirtl and the other subtypes [69]. All solved sirtuin structures contain a conserved 270-amino-acid catalytic domain with variable N- and C-termini. The structure of the catalytic domain consists of a large classic Rossmann fold and a small zinc binding domain. The interface between the large and the small subdomain is commonly subdivided into A, B and C pockets. This division is based on the interaction of adenine (A), ribose (B) and nicotinamide (C) which are parts of the NAD" cofactor. (Figure 3.5) Whereas the interaction of adenine and... 

Figure 2. Pharmacofamilies of the NADH cofactor (structure shown in panel A) binding to oxi-doreductases. Panel B shows an overlay of a subset of NAD(P)(H) geometries obtained from 288 crystal structures of oxidoreductases. The two largest pharmacofamilies are shown, corresponding to the two-domain Rossmann fold enzymes in pharmacofamilies 1 (anti) and 2 (syn). Panel C shows the corresponding pharmacophores with all protein heteroatoms indicated that are within hydrogen bonding distance of atoms in the cofactor. (Figure adapted with permission from original work of Sem ef o/. ).

The AdoMet binding motif is similar to the Rossmann fold, which is well known from the nucleotide binding proteins [22]. It has been shown that the known crystal structures of methyltransferases are strikingly similar in the AdoMet-binding regions [23], which indicates that all AdoMet-utilizing enzymes may share a common divergent evolution. 

Most dehydrogenases that use NAD or NADP bind the cofactor in a conserved protein domain called the Rossmann fold (named for Mchael Rossmann, who deduced the structure of lactate dehydrogenase and first described this structural motif). The Rossmann fold typically consists of a six-stranded parallel /3 sheet and four associated a helices (Fig. 13-16). 

Figure 2-27 Topologies of the folds of three families of nucleotide binding oc/p proteins. Cylinders represent a helices and arrows p strands. (A) The ATPase fold for the clathrin-uncoating ATPase (B) The G-protein fold that hinds GTP and is found in ras proteins (C) The Rossmann fold that hinds NAD in several dehydrogenases. From Branden.262...

Despite sharing only 25% sequence identity, structural analysis indicates that both proteins of E. coli NADP-IDH and T. thermophilus NAD-IMDH are homodimers which share a common protein fold that lacks the p p p motif characteristic of the nucleotide binding Rossmann fold [23], The strict and distinct specificities of these enzymes provide an attractive model system for engineering specificity, while the extensive knowledge of substrate and coenzyme binding and catalysis provide the sound foundation critical for rational design. 

Reitzer et al., 1999) and a MeCbl-binding fragment of E. coli methionine synthase (Drennan et al., 1994), the cofactor is sandwiched between two domains (Figure 8). The conserved domain possesses an a/ 3 structure reminiscent of the Rossmann fold of nucleotide-binding proteins (Rossmann et al., 1974) and consists of a twisted )-sheet of five parallel strands encased by five a-helices. It binds the lower, a-face of the corrin macrocycle and the substituents projecting idowni from this face, notably the dimethylbenz-imidazole ribofuranosyl nucleotide loop. 

The validity of applying the method of Chou and Fassman [79] to the membrane proteins is still controversial, but the estimated values of the a, b and c subunits of EFq are also shown in Table 5.2. The b subunit is also very rich in a-helices, and for the most part hydrophilic residues, except for the 22 residues at the N terminus [11,21]. The distribution of the secondary structure in these Fq subunits has been described in detail [67]. As shown in Fig. 5.6, a Rossmann fold [83], an alternating structure of a-helices and -sheets, is found in the subunit y8 around residues... 

A systematic assessment of the relationship between protein function and structure has been performed by Hegyi and Gerstein [259] via relating yeast enzymes classified by the Enzyme Commission (EC numbers) to structural SCOP domains. In this study it has been found that different structural folds have different propensities for various functions. Most versatile functions (hydrolases and O-glycosyl glucosidases) have been identified to be mounted onto seven different folds, whereas the most versatile folds (e.g. TIM-barrel and Rossmann folds) realize up to 16 different functions. 

The other two subsequences from residue 185 to 225 and from 240 to 270 are those recognizing NADPH. Glyl94-X-Glyl96 Is believed to be a part of the consensus sequence of nucleotlde-blndlng proteins and a part of a Rossmann fold (48). 

Figure 13.1. Crystal structure of Staphylococcus aureus isoleucinyl tRNA synthetase (IRS) in complex with tRNA (Protein Data Bank code 1QU2(6)). tRNA is shown in green, the CPI domain in orange, the Rossmann fold in blue, and rest of the protein is represented in silver.

---