A beta-glycosidase (Mr 50,000) from Spodoptera frxgiperda larval midguts was purified, cloned and sequenced. It is active on aryl and alkyl beta-glucosides and cellodextrins that are all hydrolyzed at the same active site, as inferred from experiments of competition between substrates. Sequence alignment indicates that the amino acids directly involved in the catalysis are El87 (proton donor -pKa 7.5) and E399 (nucleophile -pKa 4.9), whereas chemical modification using tetranitromethane and phenylglyoxal and tertiary structure data suggest that R97 and Y331 affect their ionization. kcat and Km values obtained for different p-nitrophe-no1 beta-glycosides and sequence alignment indicate that the glycone moiety is stabilized in the transition state by a hydrophobic region around the hydroxymethyl group and by hydrogen bonds between the other equatorial hydroxyls and polar amino acids (439, E451, H142, N186 and E399) from the -1 subsite. Ki values for a range of alkyl beta-glucosides and oligocellodextrins reveal that the aglycone binding site is a hydrophobic cleft made up of 3 subsites with a polar amino acid only at its first (+I) subsite.This work concerns the comparison of the kinetic constants (Vmax and Km) obtained by different methods applied to a twosubstrate two-product enzymatic reaction of two aspartate aminotransferase (AAT) isoenzymes from Lupinus albus L. AAT catalyses the reversible transfer of an amino group from aspartate to 2-oxoglutarate to form oxaloacetate and glutamate, following a non-sequential kinetic mechanism named ping-pong bi-bi. With this mechanism, the high number of required analytical determinations makes the repetition of measurements often impractical. The following methods were used to calculate the kinetic constants: direct linear plot, least-squares fit to a hyperbola (followed by a secondary plot), fitting of an adequate overall rate equation (Alberty equation) and the Lineweaver-Burk plot (for comparison purposes). Our results show that there are some differences in the calculated values using these different methods. The results obtained by fitting the Alberty equation to our data are very similar to the ones obtained by the Lineweaver-Burk method (that should not be used to determine kinetic constants). This is probably due to the difficulty in judging the appropriate weighting scheme to use in the fitting of the rate equation. The direct linear plot (with primary and secondary plots) is thus considered the best method, as it does not need a weighting scheme or replicate measurements. If the primary plot is obtained by the least-squares fit to a hyperbola, the kinetic constants deviate a little from the values calculated from the direct linear plot.