Low activity of β-glucosidase in curing beans, resulting partially from heat de-activation (Marquez and Waliszewski 2008), may also stem from proteolytic destruction. Denaturation of the enzyme protein by phenolic compounds and perhaps by oxidants formed during the curing process might result in tagging of β-glucosidase for proteolytic degradation. These very conditions might be an impediment, however, for assessing the enzyme activity, because extraction and assay conditions may also lead to the enzyme denaturation and subsequent proteolytic degradation. For example, determination of enzyme-substrate affinity, measured by Km values in green pod extract, revealed that β-glucosidase affinity for natural or synthetic substrate was one order of magnitude lower than in other organisms (Dignum 2002b). These results suggest that low enzyme activity in a bean extract might reflect a dysfunctional state. Attempting to avert this problem, we found that protection of β-glucosidase from proteolytic degradation, using protease inhibitors in the extraction as well as the assay medium, resulted in increased enzyme-substrate affinity (lower Km) and increased reaction rate, compared to previously reported values (Dignum 2002b), and favorably comparable to those obtained by Hannum (1997). During curing, however, β-glucosidase and other glycosyl hydrolases, presumed to display resistance to proteolytic degradation (Nishi and Itoh 1992; Varki 1993), might persist at a level sufficient to carry out the hydrolytic release of vanillin or other glyco-conjugates in the vanilla pod. It is desirable to re-examine whether low β-glucosidase activity reflects the actual state of the enzyme protein in a curing bean or, alternatively, a dysfunctional state resulting from inappropriate extraction and assay protocols.
Substrate accessibility and subsequently enzyme-substrate interaction is yet another factor in the enzyme-catalyzed hydrolytic release of vanillin and perhaps other glyco-conjugates because glucovanillin, the vanillin parent compound and β-glucosidase that catalyzes the hydrolytic cleavage of glucovanillin, might reside in different regions of a vanilla pod. This view is the rationale and objective for killing, stated early on by Arana (1943) and confirmed by subsequent studies (Theodose 1973), namely, disorganization of vanilla pod tissue in order to establish contact between enzymes and their corresponding substrates, which are compartmentalized and separated in the green bean. This view is supported by studies showing that degradation of glucovanillin to vanillin, apparently by β-glucosidase as well as other flavor generating processes, is initiated by disruption of green bean tissue by mechanical means, tissue maceration by chopping or grinding, for instance (Towt 1952). Other studies indicate, similarly, that other killing methods lead to de-compartmentalization of enzymes and substrates and onset of flavor formation in curing vanilla bean (Odoux 2006). Assay for β-glucosidase activity, when protected against proteolytic degradation, revealed that the enzyme activity, expressed as mg product/hr/mg protein, was 75.2 in the green outer fruit tissue, 32.3 in the placental tissue, and 11.1 in the hair cells, respectively, suggesting enzyme localization mostly in the green outer region. Other studies (Odoux 2006) indicate that the enzyme is localized in the inner placental region. Knowledge of the enzyme localization in the vanilla pod is, therefore, contentious. While there is progress in the understanding of the site of synthesis and accumulation of glucovanillin, we do not have unequivocal knowledge on the localization of β-glucosidase in the vanilla pod. Molecular methods, for example, immuno-cytochemistry or β-glucosidase-green fluorescent protein fusion used for the enzyme visualization and localization in plants (Matsushima et