We found evidence of the interaction

We found evidence of the interaction between AR and hemiacetals by evaluating the effect of glucose on l-idose reduction. l-idose was recently proposed as the best AR substrate able to mimic glucose [38]. This aldose is structurally very closely related to d-glucose, from which it differs only in the stereo chemical configuration of the C5 carbon atom. The two aldoses, however, significantly differ in terms of the content of the free aldehyde, with l-idose, of the aldoses, having one of the highest contents of the free aldehyde form in aqueous solution (60 to 80 fold higher than glucose) [43]. Given that glucose free aldehyde represents approximately only 1.28×10% of total glucose, we verified the effect of the hemiacetal form of glucose on l-idose reduction by avoiding the interference exerted by the reduction of glucose itself. Using 0.4mM l-idose as substrate, a modest inhibition, progressively increasing up to approximately 10% of the initial enzyme activity, was observed when d-glucose was increased in the assay mixture from zero to 4mM (Fig. 6). Higher d-glucose concentrations could not be tested because of the contribution of glucose reduction to the overall AR activity. l-Glucose is a very poor substrate for the enzyme. It is thus an ideal tool to evaluate the hemiacetal inhibition, since no contribution is expected in the activity measurements for at least up to 30mM of the sugar. When l-glucose, rather than d-glucose, was used as a source of hemiacetal a progressive increase of the inhibitory effect on l-idose reduction was observed up to approximately 17%. The observed inhibitory effect may be underestimated, since it is impossible to measure the enzyme activity on long chain aldoses in the absence of the hemiacetal form. In fact, when the data related to the inhibition of l-glucose on l-idose reduction were analyzed by a nonlinear fitting program (see Section 2.4), the values of the rate extrapolated to a zero hemiacetal concentration (2.52×10mMmin) and to an infinitive hemiacetal concentration (1.86×10mMmin) revealed an inhibitory effect of approximately 26%. This result, together with the biphasic effect observed for several long chain aldoses (Fig. 3), suggests an apparent broad specificity of the enzyme for cyclic hemiacetals of different aldoses. The interaction, however, displays some specificity, since hemiketal structures, such as those generated by d-fructose and l-sorbose, did not affect the enzyme's activity (Fig. 6). The case of l-threose and d-erythrose is also worth noting; the reduction of these aldoses did not appear to be affected by their hemiacetal form. Either the furanosidic hemiacetal form, the only one compatible with 4 carbon 1347 aldoses, is unable to interact with the enzyme, or the hemiacetal/free aldehyde ratio is too low to generate a biphasic response. It is also possible that the molecular size and/or the hindrance of l-threose and d-erythrose enable the free aldehyde of these sugars to escape the inhibitory action of furanosidic hemiacetals. However, the fact that the reduction of these sugars, as also occurs for GAL, was not affected by l-glucose (data not shown), suggests that they are not affected by the perturbation of the AR active site induced by hemiacetals. On the basis of indications emerging from Eqs. (2), (5), (7), it was possible to have insight in the mechanism of action of the hemiacetal on the free aldehyde reduction. The analysis of kinetic parameters measured for l-idose at different fixed levels of hemiacetal (Fig. 7) indicates that some effect on the catalytic step of aldehyde reduction may occur. In fact, an increase in apparent , from 1.24±0.48s to 2.32±0.19s was observed (Fig. 7, inset). More evident is the increase in apparent up to three fold, from approximately 1μM (K) to 3.4μM (K′), with the increase in hemiacetal concentration. The absolute values of these kinetic parameters may be debatable, because of the resulting rather high error. However, because of the high confidence limits of the data fitting, the trend of the parameters with the increase in hemiacetal concentration is unequivocal. The emerging values for the kinetic parameters related to the hemiacetal-bound AR (i.e., k and K′) are essentially the same as those previously measured for l-idose reduction analyzed in the high range of substrate concentrations [38]. Indeed, using the above kinetic parameters measured for l-idose, a biphasic double reciprocal plot, generated by a computer-assisted simulation of Eq. (2), was able to fit the experimental results of Fig. 3A, by imposing a value of K of 2.5×10M (Fig. 5). In principle, a partial uncompetitive inhibition model can also generate an apparent cooperative behavior (see Appendix A Kinetic models for aldose reduction by aldose reductase, Section III Partial inhibition for an uncompetitive model of action of aldose hemiacetal on free aldehyde reduction). However, this inhibition model can be ruled out, since the biphasic curves in the double reciprocal plot would appear with an upward curvature, as would happen with an apparent positive cooperativity or substrate inhibition. This is strengthened by the progressive increase of apparent values of l-idose free aldehyde with the increase of hemiacetal concentration (Fig. 7, inset) which is incompatible with an uncompetitive inhibition model (Eq. (7)). As previously mentioned, our data are unable to discriminate between nonclassical competitive from mixed noncompetitive inhibition models. Thus the latter mechanism, which must be characterized in any event by a significant contribution of the competitive interaction (Fig. 5), cannot be ruled out.