Although the presence of calcium in

Although the presence of calcium in RHA-P active site is confirmed by the homology modeling, some additional observations can be made following the metal depletion and reconstitution experiments performed on RHA-Phis. EDTA-treated RHA-Phis recovers its activity in the presence of divalent cations such as Mg2+ and Ca2+, but with an apparent higher efficiency when calcium is used, as suggested by the almost total recovery of the activity in the presence of a 3000-fold excess of this metal compared to the 56% obtained with a similar excess of Mg2+. The results obtained with Mg2+ indirectly suggest that the two alkaline earth metals might interchange in RHA active site, but that the two metal cations probably have different binding parameters. It is important to underline, however, that a detailed description of the binding kinetic of these two elements was beyond the scope of our work. By contrast, incubation of EDTA-treated RHA-Phis with manganese or zinc led to significantly lower values of recovery of the enzymatic activity. For Mg2+ and Zn2+, a relative decrease of the enzymatic activity was observed for the highest concentration used, i.e. 50 mM; more in detail, when using zinc a marked protein precipitation was observed in the sample, which may be responsible for the almost complete loss of enzymatic activity. As already underlined, the presence of calcium ions as essential cofactors is reported in some GH families, although with different functions [50]. Among the few α-RHAs crystal structures available in literature, two examples are provided bearing calcium ions involved in catalysis, besides GH106 BT0986: GH78 BsRhaB from Bacillus sp. GL1 and SaRha78A from Streptomyces avermitilis [29,30]. However, differently from what observed in BT0986, in these proteins calcium is bound in a separate domain located near the catalytic domain and contacts the rhamnose residue in the catalytic domain when the substrate is bound in the active site [30]. In particular, Fujimoto Z. and coworkers described for SaRha78A the identification of a novel carbohydrate-binding module (CBM67) highly specific for rhamnose residues [30]. Carbohydrate binding modules (CBM), are defined as non-catalytic domains found in many carbohydrate-active enzymes and contribute to the enzyme binding specificity towards different L-Kynurenine calculator and polysaccharides [50]. In RHA-Phis, and in general GH106, no separate CBM domains have been observed, and calcium is probably directly bound in the active site of the enzyme. In BT0986, as well as in RHA-Phis, calcium seems rather to play a key catalytic role, similar to the one reported from Zhu Y. and coworkers for the GH92 and GH47 α-mannosidase enzymes [51], where calcium is hypothesized to be essential for the inverting mechanism of α-mannose residues hydrolysis [51]. The role of the calcium ion in the active site of RHA-Phis undoubtedly needs further investigation; nonetheless, the data collected so far sustain its involvement as an essential cofactor at the active site of the enzyme. The model suggests that RHA-Phis and protein BT0986 have similar catalytic mechanisms, and all the residues involved in the calcium binding seem to be conserved in the two proteins. In addition, aminoacids contributing to rhamnose binding in the structure of BT0986 are also conserved in RHA-Phis (residues of interest are highlighted in green in Fig. S3 of the Supplementary Material). Noteworthy, the RHA-Phis model points out also some significant differences with the BT0986 template, which allows predicting a different substrate specificity. In particular, RHA-Phis lacks a loop region that in BT0986 is involved in the arabinose binding sub-site; this loop might be the key factor responsible for the accommodation of different substrates in the active site of these enzymes. To investigate RHA-Phis substrate specificity on natural substrates, kinetic experiments were performed using flavonoids such as naringin, rutin, hesperidin and quercitrin. Results confirm that RHA-Phis is able to hydrolyze both α-1,2 and α-1,6 glycosidic linkages [35]. However, the analysis of the k/KM values highlights some interesting differences in terms of substrate affinity and catalytic efficiency. In particular, rutin and quercitrin share the same flavonoidic aglyconic portion (named quercetin) and both have the saccharidic portion bound at position C3 on the C ring of the flavonoid. However, RHA-Phis shows a higher catalytic efficiency on rutin. On the other hand, rutin and hesperidin share the same rutinose α-1,6 disaccharidic residue, but bound at different positions of the flavonoidic ring (in position C3 on the C ring for rutin and at position C7 on the A ring for hesperidin). The higher catalytic efficiency of RHA-Phis towards rutin, might be related to the specific position of the flavonoidic ring to which the saccharidic portion is bound.