The evolutionary appearance of non‐cyanogenic hydroxynitrile glucosides in the Lotus genus is accompanied by the substrate specialization of paralogous β–glucosidases resulting from a crucial amino acid substitution

Abstract

The evolutionary appearance of non-cyanogenic hydroxynitrile glucosides in the Lotus genus is accompanied by the substrate specialization of paralogous b-glucosidases resulting from a crucial amino acid substitution SUMMARYLotus japonicus, like several other legumes, biosynthesizes the cyanogenic a-hydroxynitrile glucosides lotaustralin and linamarin. Upon tissue disruption these compounds are hydrolysed by a specific b-glucosidase, resulting in the release of hydrogen cyanide. Lotus japonicus also produces the non-cyanogenic c-and b-hydroxynitrile glucosides rhodiocyanoside A and D using a biosynthetic pathway that branches off from lotaustralin biosynthesis. We previously established that BGD2 is the only b-glucosidase responsible for cyanogenesis in leaves. Here we show that the paralogous BGD4 has the dominant physiological role in rhodiocyanoside degradation. Structural modelling, site-directed mutagenesis and activity assays establish that a glycine residue (G211) in the aglycone binding site of BGD2 is essential for its ability to hydrolyse the endogenous cyanogenic glucosides. The corresponding valine (V211) in BGD4 narrows the active site pocket, resulting in the exclusion of non-flat substrates such as lotaustralin and linamarin, but not of the more planar rhodiocyanosides. Rhodiocyanosides and the BGD4 gene only occur in L. japonicus and a few closely related species associated with the Lotus corniculatus clade within the Lotus genus. This suggests the evolutionary scenario that substrate specialization for rhodiocyanosides evolved from a promiscuous activity of a progenitor cyanogenic b-glucosidase, resembling BGD2, and required no more than a single amino acid substitution.

Lotus japonicus flowers are defended by a cyanogenic β-glucosidase with highly restricted expression to essential reproductive organs

“…A leaf-based genetic screen for cyanogenesis deficient (cyd) mutants identified several alleles of the cyd2 mutant, defective in the BGD2 gene, demonstrating its exclusive role in cyanogenesis in vegetative tissues (Takos et al 2010). Biochemical assays of recombinant expressed proteins confirmed that only BGD2 could hydrolyse the cyanogenic a-HNG lotaustralin, whereas both BGD2 and BGD4 could hydrolyse the non-cyanogenic c-HNG rhodiocyanoside A. Site-directed mutagenesis studies demonstrated that a single amino acid polymorphism in the aglycone binding region of the active site, specifically G211 in BGD2 and V211 in BGD4, was sufficient to explain the difference in substrate specificity between both enzymes for the aHNGs lotaustralin and linamarin (Lai et al 2014).…”

“…These fragments spanned the region encoding the substrate binding cavity and contained a glycine amino acid at the corresponding position to G211 of BGD2. This residue was previously shown to be critical for the acceptance of lotaustralin and linamarin as substrates by BGD2, which suggested the ability of this candidate gene to hydrolyse a-HNGs (Lai et al 2014).…”

“…5a). Prunasin, the a-HNG from almond, harbors a planar benzene ring molecular structure, which can be accommodated into the substrate binding pocket of the L. japonicus BGD isoforms (Lai et al 2014) and so its hydrolysis by each of the BGDs indicates functional enzyme expression. Cyanogenesis was detected for both BGD3 and BGD2 expressing leaf discs when incubated with the endogenous a-HNGs lotaustralin and linamarin, but not for BGD4.…”

“…Homology modelling and model quality validation of BGD2 and BGD4 using the crystal structure of the cyanogenic BGD from white clover (TrCBG) (Barrett et al 1995) and the rice b-glucosidase OsBGlu12 (Sansenya et al 2011) as a template has been performed as previously described using Modeller available at the Max Planck Institute for Developmental Biology (http://toolkit.tuebin gen.mpg.de/hhpred) (Lai et al 2014). The BGD3 sequence was modelled using the same TrCBG template (70 % amino acid sequence identity to BGD3) and the model was validated similarly.…”

“…japonicus biosynthesizes the a-HNGs lotaustralin and linamarin, as well as non-cyanogenic c-and bHNGs branching from the same pathway, named rhodiocyanoside A and rhodiocyanoside D, respectively (Forslund et al 2004;Takos et al 2011). Two paralogous HNGs metabolising BGDs, BGD2 and BGD4, are expressed in the leaves of L. japonicus, but differ in their ability to hydrolyse a-HNGs (Takos et al 2010;Lai et al 2014). A leaf-based genetic screen for cyanogenesis deficient (cyd) mutants identified several alleles of the cyd2 mutant, defective in the BGD2 gene, demonstrating its exclusive role in cyanogenesis in vegetative tissues (Takos et al 2010).…”