signature=e48117269aff62d4a3fa5fe5baffd0d1,The nitrilase superfamily: classification, structure and ...

Branch 1: nitrilase

Members of the nitrilase branch (EC 3.5.5.1) are found in plants, animals (C. elegans), fungi (Saccharomyces cerevisiae's frequently inactivated NIT1 gene), and many types of bacteria. The best evidence that nitrilase functions in vivo to convert indoleacetonitrile to the plant growth factor indole-3-acetic acid (auxin) comes from Arabidopsis, in which it was shown that recessive mutations in a nitrilase gene resulted in reduced sensitivity to the auxin-like effects of indoleacetonitrile and that overexpression of a nitrilase caused increased sensitivity to indoleacetonitrile [32]. Bacterial nitrilases are often exploited for biochemical syntheses and for environmental remediation [33]. It is not clear whether bacterial nitrilases primarily function in ecological relationships with plants or whether they benefit isolated microbes.

Branch 2: aliphatic amidase

Aliphatic amidases (EC 3.5.1.4) [3,4] comprise a small branch of nearly identical proteins found in Pseudomonas, Bacillus, Brevibacteria and Helicobacteria. They hydrolyze substrates such as the carboxamide sidechains of glutamine and asparagine utilizing the conserved cysteine within the nitrilase superfamily.

Branch 3: amino-terminal amidase

The N-end rule is a means by which the rates of ubiquitin-dependent protein degradation is regulated. The S. cerevisiae Nta1 protein deaminates amino-terminal asparagine and glutamine residues to aspartate and glutamate, which lead to more rapid rates of protein turnover [34]. Nta1 has fungal homologs but mammalian amino-terminal amidases appear to be unrelated.

Branch 4: biotinidase

Biotinidases (EC 3.5.1.12) utilize specific amidase/esterase activity to release biotin from biotinamide, biotin-lysine and biotin-peptide conjugates and biotin methylester [35]. Biotinidase deficiency can result in an inability to recycle biotin that is manifested in neurological and cutaneous abnormalities in humans [36]. Biotinidases are secreted into serum and have a unique, conserved carboxy-terminal domain. Vanins [37] and GPI-80 [38] are members of the biotinidase branch that contain a similar carboxy-terminal domain containing, in addition, a GPI anchor and are involved in T-cell thymic homing and neutrophil adherence and migration. One member of this branch contains repeated nitrilase-related domains. Recently, porcine panthetheinase (EC 3.5.1.-), an amidase that converts pantetheine to panthothenate plus cysteamine in the dissimilative pathway of CoA, was sequenced and found to be nearly identical to vanins [39]. Although the biologically important substrate of vanins remains unproven, sequence and enzymatic similarity with biotinidases suggest that an amine molecule at least the size of an amino acid (that is, bigger than ammonia) may be the leaving group. Branch 4 enzymes are the only amidases in the nitrilase superfamily known to prefer secondary amine substrates of the form R-C=O(NHR') as opposed to simple acid amides. An extensive archive of vanins, including 118 expressed sequence tag (EST) sequences is available [40,41].

Branch 5: β-ureidopropionase

The β-ureidopropionases (EC 3.5.1.6) are enzymes involved in the catabolism of pyrimidine bases and the production of β-alanine [42]. Substrates of this enzyme are of the carbamylase type (see Figure 1c) and the amine product is usually a non-standard amino acid.

Branch 6: carbamylase

A variety of bacteria express hydrolases specific for the decarbamylation of D-amino acids. These enzymes have been exploited in the production of semisynthetic β-lactam antibiotics [43] and are now represented by the structure of the Agrobacterium enzyme [13].

Branches 7 and 8: glutamine-dependent NAD synthetase

As discussed earlier, the presence of a nitrilase-related domain appears to correlate with the ability of bacterial NAD synthetase (EC 6.3.5.1) to utilize glutamine as an ammonia source. Eukaryotic NAD synthetases always contain this novel, putative GAT domain and exhibit glutamine dependence. Substrate specificity of nitrilase-related proteins as glutamine amidases is not surprising given the specificity of the branch 2 and 3 enzymes. It remains to be seen how glutamine-dependent NAD synthetase may channel ammonia from the nitrilase-related active site to the NAD active site.

Branch 9: apolipoprotein N-acyltransferase

The modification and processing of Braun's lipoprotein, a major component of the outer membrane of E. coli, has been studied for decades [44]. Defects in this post-translational modification pathway are associated with copper sensitivity [45]. The apolipoprotein becomes proteolized, exposing an amino-terminal cysteine. After the cysteine is modified by diacylglycerol, branch 9 enzymes condense a fatty acid to the amino terminus of the modified cysteine residue.

Branch 10: Nit

Nit was originally identified as an approximately 300 amino acid amino-terminal extension on fly and worm homologs [46] of the human [47] and murine [48] Fhit tumor suppressor protein. Nit homologs are found in organisms with Fhit homologs [12] and, in the mouse, Nit1 and Fhit mRNA levels are highly correlated in seven of eight tissues examined [46]. Satisfaction of these criteria suggested that NitFhit is a Rosetta Stone protein, whose fusion might decode a previously unsuspected interaction between the proteins [18,19]. As Fhit is part of a cell-death pathway that is not clearly connected to known apoptotic players [49,50], identification of Nit as a Fhit-interacting protein was welcomed. The Fhit active site of NitFhit has been characterized and the structure of worm NitFhit has been elucidated [12], but the Nit substrate, cell biology and relationship to tumor suppression are not known.

The most striking feature of the Nit-Fhit interaction apparent from the crystal structure of the worm protein is that the complex assembles with a central Nit tetramer binding two Fhit dimers [12]. The carboxy-terminal β strands of Nit-conserved polypeptide sequences exit the compact Nit tetramer domain and physically interact with Fhit dimers. Fhit dimers are bound in a way that allows them to expose diadenosine polyphosphate-binding sites opposite from the Nit interaction surface [12]. Futhermore, the nucleotide kinetics of NitFhit active sites [12] were extremely similar to those of human Fhit dimers in the absence of Nit [51].

Concord between the phylogenetic profiles [52] of Fhit and Nit breaks down slightly with the discovery of Nit-related sequences in a small number of prokaryotes that have no Fhit homolog (see the Additional data file). The idea that nitrilase-related proteins spread from animals and plants to prokaryotes is, however, supported by the animal-associated ecology of these microbes.

Branches 11-13

Branches 11 and 12 contain distinct similarity groups with no characterized member. Branch 12 may contain Rosetta Stone [18,19] proteins in that a distinctive nitrilase-related domain is found fused to an amino-terminal domain of approximately 210 amino acids. The branch-12-associated domain is related to the RimI [53] superfamily of amino-terminal acetyltransferases, suggesting that branch 12 enzymes are involved in post-translational modifications. Branch 13 contains uncharacterized, nonfused nitrilase-related proteins that are difficult to place in a distinct similarity group.