Sequence Comparisons Based on Structural Alignments

A structure such as the six-stranded coenzyme binding domain in the dehydrogenases would be disrupted by insertions or deletions of amino acids (see Fig. 7 for elaboration). Hence, sequence comparisons of parallel pleated sheet regions are particularly reliable. Structural methods of alignment of sheet areas have been discussed in Section II. The corresponding amino acid comparisons are made in Table IV. For tbe purpose of this chapter, the present LDH numbering scheme (4) will be used as the generalized reference system. 

In Table IV the AMP and NMN mononucleotide binding units have also been aligned below each other. The secondary structural features are shown at the top. One row indicates the functions of certain residues. 

A parallel pleated sheet (a) with hydrophobic residues pointing up and hydrophilic residues pointing down. If a deletion should occur in one strand it will cause a disruption in the physical properties of the protein (b), making the event of such a deletion within a sheet most improbable. 

the residues heavily circled in Fig. 6, representing those sheet residues in the hydrophobic pockets facing the helices, have been labeled H. Similarly, those residues in helices which are in this pocket, and therefore face the sheet, have been labeled S. Still further, residues which are fimc-tional in the binding of coenzyme have been labeled F. 

The conservation of B is well demonstrated by the constant number of residues between the aligned strands )8A and (see Table II). It has been necessary to assume a deletion in LADH between residues 216 and 216. A deletion in the equivalent position 24 of lobster GAPDH with respect to pig and yeast GAPDH has been observed. Thus with respect to this particular residue, lobster GAPDH is similar to LADH, whereas pig and yeast GAPDH are similar to LDH. A deletion in GAPDH of residue LDH 130 in E has had to be assumed. Neither GAPDH nor LADH has the functional loop present in LDH between residues 

---

Table 1. Comparison of amino acid residues which are involved in the binding of the cofactors ThDP and Mg2+ in PDCS.c. and PDCZ.m., respectively. The corresponding amino acid residues from PDCZ.m. have been deduced from sequence alignments (Figs. 1 and 2) based on the X-ray structure of PDCS.u. [35] and PDCS.c. [51]. Different residues are underlined. Those which have been exchanged by site-directed mutagenesis are marked with (see Tables 4 and 5)...

Panchenko, A. R. and Bryant, S. H. (2002) A comparison of position-specific score matrices based on sequence and structure alignments. Protein Sci. 11, 361-370. 

Siebert, S. and Backofen, R. (2005) MARNA multiple alignment and consensus structure prediction of RNAs based on sequence structure comparisons. Bioinformatics 21, 3352-3359. 

The global superposition of enzymes of family 13 on the common (/ /a)8 barrel is remarkably efficient to start /Joe loops comparison in the 3D space. This 3D superposition can focus directly to the new structural information for similar enzymes in term of functionality when compared to predictions based on sequence alignment. Interestingly, it has been seen in some comparisons, between less similar enzymes, that our 3D superposition is more powerful and more accurate to match segments of residues whose location is tightly related to functionality. In the shown examples of three enzymes of family 13, this 3D superposition is necessary to detect similarities and discrepancies between subfamilies of strict hydrolase and hydrolase-transferase. 

In order to compare the three-dimensional structures of such families of homologous proteins, automatic methods have been developed which consider both global and local structural features. Based on sequence alignments deriving from structure comparisons of homologous proteins, the occurrence of sequence identities... 

---