Catalytic residue prediction

In Ref. [52] it was demonstrated that experimentally derived structural information such as the existence of S-S bonds, protein side-chain ligands to iron-sulfur cages, cross-links between side chains, and conserved hydrophobic and catalytic residues, can be used by GAs to improve the quality of protein structure prediction. The improvement was significant, usually nudging the prediction closer to the target by more than 2 A. However, even with this improvement, the overall prediction quality was still insufficient, usually off by more than 5 or 6 A from the target structure. This was probably due to the small number and the diverse nature of the experimental constraints. 

A related approach integrating sequence information (conservation), geometric information (cleft detection), and data on local stability calculated by Poisson-Boltz-mann methods was reported by Ota et al. [54], The method was used for predicting catalytic residues (polar atoms only) in enzymes. A number of putative active sites for a series of hypothetical proteins were found and are discussed in the study. 

Pseudokinases are a protein family that constitute approximately 10% of the human kinome (for reviews on this topic, see Ref. 51-53). These proteins are characterized by the presence of a kinase-homology domain predicted to lack enzymatic activity due to the absence of at least one of the three conserved critical catalytic motifs (1) the Val-Ala-Ile-Lys (VAIK) motif in subdomain II, in which the side-chain of Lys interacts with the a and p phosphates of ATP (2) the His-Arg-Asp (HRD) motif in subdomain Ylb, in which the aspartic acid is the catalytic residue and (3) the Asp-Phe-Gly (DFG) motif in sub-domain VII, in which the carboxylic moiety of aspartic acid binds the Mg11 ion that coordinates the p and y phosphates of ATP. Owing to their lack of intrinsic phosphoryl-transfer catalytic activity, pseudokinase domain-containing... 

The protein is a 12-stranded anti-parallel p-barrel with amphipathic P-strands traversing the membrane (Fig. 4). The active-site catalytic residues are similar to a classical serine hydrolase triad except that in addition to the serine (Ser-144) and histidine (His-142), there is an asparagine (Asn-156) in place of the expected aspartic acid. Calcium at the active site is predicted to be involved in the reaction mechanism facilitating hydrolysis of the ester. 

Three-dimensional models are usefiil to study and predict mutual matches between proteins or protein docking to varied molecules. This procedure is crucial for the identification of putative catalytic residues and the search for new inhibitor/substrate complexes. Moreover, the improvement of protein stability and the detection of amino acid targets by experiments of site-directed mutagenesis can be pursued. 

A seventh vitamin K-dependent protein, protein Z, has been isolated from plasma (123). Protein Z is closely related to the factor IX-like proteases, factors VII, IX, X and protein C, consisting of the Gla domain, the two EGF units and a single Hya residue. However, similar to the heme binding protein haptoglobin, protein Z has no associated proteolytic activity as the two of the three essential catalytic residues have been modified (123). In contrast to protein sequencing, cDNA analysis allows the prediction of the amino acid sequence of the primary translation product of the mRNA species. Post-translational modifications are not detected, but precursor structures are elucidated. The vitamin K-dependent clotting factors are synthesized as precursors with leader sequences of 38 to 46 amino... 

Specificity for a particular charged substrate can be engineered into an enzyme by replacement of residues within the enzyme-active site to achieve electrostatic complementarity between the enzyme and substrate (75). Protein engineering, when coupled with detailed stmctural information, is a powerful technique that can be used to alter the catalytic activity of an enzyme in a predictable fashion. 

Figure 4. Alignment of PelZ and PelC amino acid sequences. The vertical lines indicate identical amino acids and the two points indicate homologous amino acids. The bold letters correspond to the residues probably involved in Ca + binding or catalytic function(s). The two aspartate residues probably involved in Ca binding are indicated with an asterisk. The invariant residues, probably involved in PGA cleavage, are indicated with an open circle. The folding in p-sheets is characterised by the underlined amino acids. Double underlining of PelZ residues is deduced from Chou Fasman and Robson Gamier folding predictions.

The specificity determinants surrounding the tyrosine phospho-acceptor sites have been determined by various procedures. In PTK assays using various substrates, it was determined that glutamic residues of the N-terminal or C-terminal side of the acceptor are often preferred. The substrate specificity of PTK catalytic domains has been analyzed by peptide library screening for prediction of the optimal peptide substrates. Finally, bioinformatics has been applied to identify phospho-acceptor sites in proteins of PTKs by application of a neural network algorithm. 

Fig. 11.2. Schematic representation of the primary structure of secreted AChE B of N. brasiliensis in comparison with that of Torpedo californica, for which the three-dimensional structure has been resolved. The residues in the catalytic triad (Ser-His-Glu) are depicted with an asterisk, and the position of cysteine residues and the predicted intramolecular disulphide bonding pattern common to cholinesterases is indicated. An insertion of 17 amino acids relative to the Torpedo sequence, which would predict a novel loop at the molecular surface, is marked with a black box. The 14 aromatic residues lining the active-site gorge of the Torpedo enzyme are illustrated. Identical residues in the nematode enzyme are indicated in plain text, conservative substitutions are boxed, and non-conservative substitutions are circled. The amino acid sequence of AChE C is 90% identical to AChE B, and differs only in the features illustrated in that Thr-70 is substituted by Ser.

To rationalize the stereospecificity of PLE toward a large variety of monocarboxylic and dicarboxylic esters, Tamm and co-workers have proposed the general formula displayed in Fig. 7.5 [5 5] [67]. Here, no representation of the active site is implied, but the model does rationalize numerous data and allows some qualitative predictions. A qualitative topographical model of the active site of PLE has been proposed by Jones and co-workers [68] [69], As shown in Fig. 7.6, substrate binding is defined by a carboxylate group that interacts with the catalytic serine residue, and by one or two hydrophobic groups that bind to sites 1 and/or 2. 

The enzyme catalyzing the formation of retinal 2 by means of central cleavage of P-carotene 1 has been known to exist in many tissues for quite some time. Only recently, however, the active protein was identified in chicken intestinal mucosa (3) following an improvement of a novel isolation and purification protocol and was cloned in Escherichia coli and BHK cells (4,5). Iron was identified as the only metal ion associated with the (overexpressed) protein in a 1 1 stoichiometry and since a chromophore is absent in the protein heme coordination and/or iron complexation by tyrosine can be excluded. The structure of the catalytic center remains to be elucidated by X-ray crystallography but from the information available it was predicted that the active site contains a mononuclear iron complex presumably consisting of histidine residues. This suggestion has been confirmed by... 

---