Page 299 - Cascade_Biocatalysis_Integrating_Stereoselective_and_Environmentally_Friendly_Reactions

12.3 Structure–Function Relationships 275

these enzymes. Cyanide hydratases are typical for phytopathogenic fungi. This is
probably because these enzymes are able to make the plant environment suitable
for the fungus by detoxifying HCN, which is formed either by the breakdown of
cyanogenic glycosides or as a by-product in the biosynthesis of ethylene [21].

12.2.3
Nitrilases in Plants

Nitrilases in plants are comprised of two subtypes with different biological func-
tions. The first subtype involves nitrilases exhibiting mostly over 70% identity levels
to the biochemically characterized nitrilases 1–3 from A. thaliana, which act on
substrates such as 3-phenylpropionitrile, allylcyanide, or indole-3-acetonitrile. They
seem to be specific for Brassicaceae, which is in accordance with their preference
for the nitriles endogeneous in this family or structurally similar to them [22]. They
are able to produce plant auxins such as indole-3-acetic acid but the role of this
pathway in vivo is not fully understood [21].
Plant nitrilases of the other subtype are highly homologous to A. thaliana
nitrilase 4, with identity levels similar as in the group of nitrilase 1–3 and
homologs. The identities between the two subtypes are over 60% in most cases. The
preferential substrate of nit4 and its characterized homologs is β-cyano-l-alanine,
an intermediate in a common pathway of cyanide detoxification [18]. Accordingly,
these enzymes seem to be widely distributed within the plant kingdom.

12.3
Structure–Function Relationships
Investigation of the structure–function relationships in nitrilases was largely based
on sequence analyses, homology modeling, and mutational studies, as the crystal
structures of nitrile-hydrolyzing enzymes have not been available except for an
aliphatic nitrilase from Pyrococcus abyssi [13]. Other crystallized members of the
nitrilase superfamily (amidases, N-carbamoyl-d-amino acid amidohydrolases, etc.
[9]) shared only low levels of identity with experimentally confirmed nitrilases.

12.3.1
Sequence Clustering

The well-known classification of nitrilases into different substrate specificity sub-
types [2] was demonstrated to be in partial correlation with their aa sequence
similarities. A prediction of substrate specificities in putative nitrilases could be
made by grouping similar sequences. For instance, four probable arylacetonitri-
lases were selected because of their high similarity (over 50%) to the biochemically
characterized enzyme from Neurospora crassa, and their expected substrate speci-
ficities were confirmed [6]. Using this approach, it also seems possible to predict
cyanide hydratases [5, 6]. Aromatic nitrilases are not as easy to predict. Members of