Advertisement

Studies on the Mechanism of Myrosinase

INVESTIGATION OF THE EFFECT OF GLYCOSYL ACCEPTORS ON ENZYME ACTIVITY (∗)
  • M. Grazia Botti
    Affiliations
    School of Chemistry, The Purdie Building, The University of St. Andrews, St. Andrews, Fife KY16 9ST, United Kingdom
    Search for articles by this author
  • Malcolm G. Taylor
    Affiliations
    School of Chemistry, The Purdie Building, The University of St. Andrews, St. Andrews, Fife KY16 9ST, United Kingdom
    Search for articles by this author
  • Nigel P. Botting
    Correspondence
    To whom correspondence should be addressed. Tel.: 44-01334-463856; Fax: 44-01334-463808
    Affiliations
    School of Chemistry, The Purdie Building, The University of St. Andrews, St. Andrews, Fife KY16 9ST, United Kingdom
    Search for articles by this author
  • Author Footnotes
    ∗ The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Open AccessPublished:September 01, 1995DOI:https://doi.org/10.1074/jbc.270.35.20530
      Myrosinase (EC 3.2.3.1) is the β-thioglucosidase enzyme responsible for the hydrolysis of glucosinolates, a group of naturally occurring plant metabolites. The enzyme catalyzes the hydrolysis of these S-glucosides to give D-glucose and an aglycone fragment, which then rearranges to give sulfate and an isothiocyanate. As part of ongoing mechanistic studies on myrosinase, the ability of the enzyme to catalyze transglycosylation reactions has been examined. Enzyme activity and stability were both decreased in the presence of various organic solvents, including simple alcohols, but not sufficiently to prevent reaction taking place. However, in contrast to most other β-glycosidases, myrosinase did not catalyze transglycosylation reactions either with the alcohols or other suitable glycosyl acceptors. Although a wide range of potential acceptors were investigated, none proved to be effective. Even when appropriately charged side chains were included in the acceptor molecule to mimic the sulfonic acid in the glucosinolate structure, transglycosylation did not take place. The putative enzyme-glycosyl intermediate therefore appears to be unavailable for reaction, possibly because D-glucose is the first product released from the enzyme. The transition state analogue, glucono-δ-lactone, a potent competitive inhibitor of β-glucosidase, was found to be a poor noncompetitive inhibitor of myrosinase. Myrosinase is specifically activated by ascorbic acid, and it is proposed that the inhibitor is binding at this alternative site.

      INTRODUCTION

      Myrosinase (EC 3.2.3.1) is the trivial name for the β-thioglucosidase enzyme responsible for the hydrolysis of glucosinolates (I) (Fig. 1), a group of sulfur-containing glycosides that occur in all members of the Cruciferae, including the brassica vegetables(
      • Fenwick G.R.
      • Heaney R.K.
      • Mullin W.J.
      ). Over 100 different examples have been isolated and characterized, containing a variety of substituents in the side chain R including allyl (sinigrin), benzyl, and indoyl(
      • Fenwick G.R.
      • Heaney R.K.
      • Mullin W.J.
      ). Enzymic hydrolysis usually occurs when cells are damaged as a result of plant injury or food processing giving as products β-D-glucose (II) and the aglycone fragment (III). The aglycone is unstable and reacts further giving the isothiocyanate (IV) by means of a Lossen-type rearrangement (Fig. 1). Some evidence implies that the rearrangement occurs spontaneously and is not enzyme catalyzed, but it is not conclusive(
      • Ettlinger M.G.
      • Dateo Jr., G.P.
      • Harrison B.W.
      • Mabry T.J.
      • Thompson C.P.
      ).
      Figure thumbnail gr1
      Figure 1:Metabolism of glucosinolates as catalyzed by myrosinase.
      Considerable work has been carried out in this area because of the agricultural and dietary importance of the Cruciferae. Products of glucosinolate hydrolysis are responsible for the distinctive flavor and aroma characteristics of crucifers. However, they can also have deleterious effects due to their pungency and in some cases goitrogenic activity and hepatoxicity(
      • Tookey H.L.
      • Etten C.H.
      • Daxenbichler M.E.
      ,
      • Larsen P.O.
      ). This is a particular problem with oilseed rape and its food and feed derivatives. EC recommendations concerning the maximum permitted levels of glucosinolates in rapeseed have resulted in major programs to breed safer varieties. However, the chemistry of the system is still not fully understood and the mechanism of myrosinase has not been elucidated.
      We were interested in studying the chemical mechanism of myrosinase and, in particular, comparing it with the much more widely studied β-glycosidases. In recent years, there has been a massive increase in interest in glycosidases(
      • Withers S.G.
      ). Inhibition of the enzymes that process carbohydrates has potential therapeutic applications for the treatment of cancer, AIDS, and other viral infections. Crystallographic studies, identification of active sites by mechanism-based labeling, and determination of transition state structure by kinetic analysis have allowed the construction of a detailed picture of glycosidase action. Myrosinase has been much less studied, and it is not known how the enzymatic hydrolysis of S-glycosides fits into the established picture of enzymatic O-glycoside hydrolysis.
      Myrosinase has been isolated from a number of plant sources including Lepidium sativum L. (light grown cress)(
      • Durham P.L.
      • Poulton J.E.
      ), Sinapis alba (white mustard)(
      • Bjorkman R.
      • Lonnerdal B.
      ), and Brassica napus (rapeseed) seeds(
      • Bjorkman R.
      • Lonnerdal B.
      ). The mustard and rapeseed (
      • Bjorkman R.
      • Lonnerdal B.
      ) enzymes have molecular weights of 120,000-150,000. The activity of the L. sativum L. enzyme (
      • Durham P.L.
      • Poulton J.E.
      ) was tested with 29 different glycosides, and only 4 were found to be good substrates. These were, in order of decreasing activity, sinigrin (the natural substrate), benzylglucosinolate, p-nitrophenyl-β-D-glucoside (PNPG),
      The abbreviation used is: PNPG
      p-nitrophenyl-β-D-glucoside.
      and o-nitrophenyl-β-D-glucoside. β-D-Glucose could not be replaced by any other sugar. α-Glucosides were inactive, as were unreactive β-glucosides, such as methyl or phenyl. Interestingly, there appears to be no discrimination between different glucosinolates, and all are hydrolyzed at similar rates regardless of the nature of the side chain. Early work by Ettlinger et al.(
      • Ettlinger M.G.
      • Dateo Jr., G.P.
      • Harrison B.W.
      • Mabry T.J.
      • Thompson C.P.
      ) demonstrated that the sulfate group was required for optimum activity as the desulfoglucosinolate did not act as a substrate with the Sinapis enzyme. The various isozymes have broad pH optima, from pH 5 to pH 7(
      • Durham P.L.
      • Poulton J.E.
      ,
      • Bjorkman R.
      • Lonnerdal B.
      ). No metal ion requirement has been found for the enzyme. For most plant enzymes, inhibition has been observed by -SH-directed reagents, implying that a cysteine residue is essential for catalysis. Myrosinase is specifically activated by ascorbic acid. The rate of hydrolysis of sinigrin, catalyzed by the mustard enzyme, is increased more than 25-fold by 1 mM ascorbic acid(
      • Ohtsuru M.
      • Hata T.
      ), although high concentrations begin to competitively inhibit reaction. Analogues of ascorbic acid are not effective. Hydrolysis of PNPG, by the same enzyme, however, is not increased by ascorbic acid(
      • Tsuruo I.
      • Hata T.
      ). Spectroscopic studies implied that there was a conformational change in the enzyme on binding ascorbic acid, and it was proposed that this brought about the rate enhancement.
      We report the results of our initial studies on the chemical mechanism of myrosinase. These comprise examination of the activity of the enzyme in the presence of alternative glycosyl acceptors and inhibition studies.

      RESULTS

      It is well documented that many glycosidases are able to catalyze transglycosylation reactions, namely the transfer of residues from glycoside substrates to acceptor molecules other than water(
      • Dale M.P.
      • Kopfler W.P.
      • Chait I.
      • Byers L.D.
      ,
      • Sinnot M.L.
      • Vitarelle O.M.
      ). Synthetic organic chemists have frequently exploited this useful property and employed β-glycosidases in the synthesis of novel glycosides(
      • Toone E.J.
      • Simon E.S.
      • Bednarsta M.D.
      • Whitesides G.M.
      ). For example, when o-nitrophenyl-β-D-galactoside (V) is hydrolyzed by β-galactosidase in the presence of the epoxy alcohol (VI), the galactosyl moiety is transferred to the acceptor hydroxyl group to give the new β-galactoside (VII) (
      • Bjorkling F.
      • Godtfredsen S.E.
      ) (Fig. 2).
      Figure thumbnail gr2
      Figure 2:The use of β-galactosidase-catalyzed transglycosylation in the synthesis of novel galactosides.
      The ability of myrosinase to catalyze similar transglycosylation reactions had not been examined. Therefore, preliminary preparative scale experiments were carried out using myrosinase under similar conditions with PNPG as the substrate and allyl alcohol as the glycosyl acceptor. Allyl alcohol was chosen because it resembles the side chain of the glucosinolate, sinigrin, a natural substrate for myrosinase, and should have some chance of binding at the enzyme active site. However, no allyl-β-D-glucoside, the product of transglycosylation, was observed in any reaction. The same substrate was also hydrolyzed in a 9:1 ethanol:water mixture to produce the ethyl-β-D-glucoside. Again, no product was observed. These early failures to observe any transglycosylation prompted us to study the activity of myrosinase in alcohol:water mixtures in more detail and show whether the reactions failed due to mechanistic reasons or simply because of reduced enzyme activity.
      Two myrosinase activities were monitored. The hydrolysis of sinigrin was followed by UV spectroscopy, monitoring the disappearance of the sinigrin absorbance at 227 nm. In separate experiments, PNPG hydrolysis was also followed by UV spectroscopy monitoring the appearance of the p-nitrophenol absorbance at 430 nm. The kinetics for the hydrolysis of sinigrin were determined under the standard assay conditions (pH 7.0, 33 mM phosphate buffer, 37 ± 0.1°C). The Km for sinigrin was found to be 0.42 ± 0.05 mM, in good agreement with previous workers(
      • Ohtsuru M.
      • Tsuruo I.
      • Hata T.
      ), and Vmax was also comparable (
      • Palmieri S.
      • Leoni O.
      • Iori R.
      ) at (48 ± 0.2) × 10-3 mol dm-3 min-1. The measurements were then repeated in the presence of increasing amounts of methanol, from 0.2 to 12.5 M (40% by volume) (Table 1). The hydrolysis of PNPG was then examined under similar conditions but using 50 mM phosphate buffer. In aqueous solution, a high Km of 89 ± 30 mM was obtained, as well as a Vmax of (0.59 ± 0.2) × 10-3 mol dm-3 min-1. Thus, the rate of PNPG was much lower than that of sinigrin as expected(
      • Ohtsuru M.
      • Tsuruo I.
      • Hata T.
      ). The experiments were then repeated in methanol:water mixtures as for sinigrin (Table 2).
      For both activities, it can be seen that the rate of hydrolysis decreases as the amount of methanol increases. Values of Km for sinigrin remained constant and the observed rate decrease was all due to changes in Vmax. Thus, binding of the substrate to the enzyme active site was unaffected by the change in the reaction medium, while the rate of the reaction decreased. Even with 12.5 M (40% v/v) methanol, there was an appreciable rate of reaction with only a 50% decrease from the rate in aqueous solution. Data for PNPG showed similar trends. Therefore, the enzyme is not terribly sensitive to the presence of the organic solvent, and so the chances of observing transglycosylation were reasonable. The measurements were also repeated in a range of other water-miscible organic solvents (ethanol, dioxan, acetonitrile) (1, 2), again showing reduced, although significant, activity.
      The kinetic results were all initial rate measurements and so gave no information concerning the long term stability of myrosinase under the reaction conditions. Therefore, the stability of myrosinase was examined in the methanol:water mixtures by incubating the enzyme in buffer at 37°C, removing aliquots at various time intervals, and measuring the myrosinase activity using the standard sinigrin assay (Fig. 3). If the enzyme proved to be very unstable, then the chances of transglycosylation reactions may be significantly reduced. However, myrosinase was in fact very stable. The half-life in the presence of 6.25 M (20%) methanol at 37°C was 40 h but was reduced to 8 h with 12.5 M (40%) methanol. Myrosinase was generally found to be very stable. Concentrated stock solutions used for kinetic experiments, stored at 4°C, did not lose appreciable amounts of activity even after 2-3 weeks.
      Figure thumbnail gr3
      Figure 3:Stability of myrosinase in methanol:water mixtures. Myrosinase (4 units) was incubated at 37 ± 0.1°C at pH 7.0 in 33 mM potassium phosphate buffer (1 ml), containing the appropriate amount of methanol. Aliquots (30 μl) were withdrawn at various time intervals, and the enzyme activity was measured using the standard assay. The incubation conditions were as follows: ▪, no methanol; ♦, 6.25 M (20%) methanol; •, 12.5 M (40%) methanol.
      Once the activity and stability of myrosinase in the presence of glycosyl acceptors had been examined, it was then necessary to determine the degree of transglycosylation that was taking place. This required measurement of the relative amounts of D-glucose and p-nitrophenol produced on hydrolysis of the PNPG under various conditions. A coupled enzyme assay was used to measure D-glucose concentrations, while the p-nitrophenol was determined spectrophotometrically. In the presence of a glycosyl acceptor, the amount of transglycosylation product can be calculated from the difference in these two values, i.e. less glucose will be produced than p-nitrophenol. Using methanol as the glycosyl acceptor, the transglycosylation product would be methyl-β-D-glucoside. A control experiment incubating this with myrosinase under the standard conditions resulted in no release of glucose. This demonstrated that methyl-β-D-glucoside was not a substrate for myrosinase and was stable under the reaction conditions. Indeed, the myrosinase preparation used in these studies has not been found to hydrolyze any of the alkyl O-glycosides or S-glycosides that have thus far been tested.
      M. G. Taylor and N. P. Botting, unpublished results.
      This also implies that there are no contaminating β-glucosidase enzymes, which would cause complications in the interpretation of our results.
      For the transglycosylation experiments, incubations containing 20 mM PNPG were employed. To test the procedure, a reaction was run in aqueous solution, and, as expected, equal amounts of D-glucose and p-nitrophenol were produced. When the experiment was then repeated in 0.2, 6.25, and 12.5 M methanol, it was observed that equal amounts of D-glucose and p-nitrophenol were also obtained, implying that no observable transglycosylation was taking place. This was in agreement with the qualitative observations of the synthetic scale experiments but was still a very surprising result. For most of the β-glycosidases, significant amounts of transglycosylation have been observed under similar conditions(
      • Dale M.P.
      • Kopfler W.P.
      • Chait I.
      • Byers L.D.
      ,
      • Sinnot M.L.
      • Vitarelle O.M.
      ).
      A range of other glycosyl acceptors were then investigated, all at 0.2 M concentrations. The amounts of D-glucose and p-nitrophenol produced after a 24-h incubation under standard conditions were determined in the presence of each potential glycosyl acceptor (Table 3). The larger alcohols, ethanol, propan-1-ol, and butan-1-ol, proved to be ineffective in transglycosylation reactions. As the glucosinolates are actually S-glycosides, a thiol acceptor was also tried. However, mercaptoethanol did not appear to trap out any of the glucose. The other part of the glucosinolate structure that probably provides important binding interactions is the sulfate group. This charged moiety is presumably balanced by some positively charged residue at the enzyme active site. It therefore seemed feasible that glycosyl acceptors containing a negatively charged group might have a better chance of binding at the active site and intercepting the glycosyl moiety. Three examples were chosen, which were commercially available, namely, 3-hydroxy-1-propanesulfonate (VIII), glycerol 2-phosphate (IX), and 2-mercaptoethanesulfonate (X) (Fig. 4). When PNPG was incubated with myrosinase in the presence of these compounds at 0.2 M concentration, equal amounts of glucose and p-nitrophenol were produced, again implying that no transglycosylation had taken place. Interestingly, it can also be observed that none of these compounds seems to be a particularly effective inhibitor of myrosinase, as the amounts of D-glucose and p-nitrophenol produced do not seem to be very sensitive to the presence of the glycosyl acceptor. The yields are all approximately 30% after 24 h. In the case of 12.5 M methanol, this drops to 21%, as may be expected from the effect of high methanol concentrations on the enzyme activity. The three charged glycosyl acceptors also have a significant effect, giving 22-25% yields, reflecting some degree of inhibition of the hydrolysis reaction. The detailed inhibitory properties of these compounds are currently under investigation. Preliminary results using sinigrin as the glycosyl donor, rather than PNPG, have indicated that transglycosylation does not take place with this substrate either. Crude measurements of glucose yields on completion of sinigrin hydrolysis have been unaffected by the presence of glycosyl acceptors.
      The above results call into question the existence of a long lived glycosyl-enzyme intermediate that can be trapped by nucleophilic species. In the case of the β-glycosidases, this intermediate has been thought to be either a stabilized oxocarbonium ion or a covalently bound glycosyl group, resulting from attack at the anomeric carbon by an active site nucleophile. In the former case, the oxocarbonium ion structure (XI) can been employed as a model for the transition state for the reaction. Enzyme inhibitors have then been designed that mimic this structure. One simple, readily available example is glucono-δ-lactone (XII) (Fig. 4). This molecule possesses a planar sp2 hybridized carbon atom at the 1-position of the sugar, which mimics the planar structure of the transition state at that position. As a result, glucono-δ-lactone has been found to be a very tight binding competitive inhibitor for β-glucosidase, with a KI of 0.2 ± 0.2 mM(
      • Dale M.P.
      • Ensley H.E.
      • Kern K.
      • Sastry K.A.R.
      • Byers L.D.
      ). We were interested in examining the interaction of myrosinase with this inhibitor to see if a similar type of inhibition was observed.
      Thus, the rates of hydrolysis of both sinigrin and PNPG by myrosinase were measured in the presence of a range of concentrations of glucono-δ-lactone. The results are shown in Lineweaver-Burk form in Fig. 5and Fig. 6. These show that for both substrates, glucono-δ-lactone acts as a noncompetitive inhibitor, that is, the Km is unaffected, while Vmax decreases with increasing inhibitor concentration. The variation of Vmax with inhibitor concentration is not simple, but it can be estimated that KI is approximately 5 mM for each substrate. Therefore, glucono-δ-lactone appears to bind much less tightly to myrosinase than to β-glucosidase.
      Figure thumbnail gr5
      Figure 5:Inhibition of myrosinase-catalyzed sinigrin hydrolysis by glucono-δ-lactone. Incubations were carried out under the standard conditions containing increasing concentrations of glucono-δ-lactone: ▪, 0 mM; •, 1 mM; ▴, 5 mM; ♦, 10 mM.
      Figure thumbnail gr6
      Figure 6:Inhibition of myrosinase-catalyzed p-nitrophenyl-β-D-glucoside hydrolysis by glucono-δ-lactone. Incubations were carried out under the standard conditions containing increasing concentrations of glucono-δ-lactone: ▪, 0 mM; •, 5 mM; ▴, 10 mM; ♦, 20 mM.
      The effect of the specific activator, L-ascorbic acid(
      • Ohtsuru M.
      • Hata T.
      ), on the degree of inhibition of myrosinase by glucono-δ-lactone was examined (Table 4). At a concentration of 1 mM, L-ascorbic acid exerts its maximum effect, increasing Vmax for the hydrolysis of sinigrin by a factor of 1.5 but leaving Km unchanged. When the measurements are repeated in the presence of glucono-δ-lactone, the degree of activation decreases as the concentration of the inhibitor increases. Thus, at 1 mM, glucono-δ-lactone Vmax increases by a factor of 1.4, but at 10 mM, Vmax actually decreases by a factor of 0.8. This raises the interesting possibility that the glucono-δ-lactone is actually binding at the activator site rather than at the substrate site, which would fit with the noncompetitive inhibition that is observed.
      Methyl-β-D-glucoside (XIII) was found to be a simple competitive inhibitor of myrosinase, albeit with a rather poor KI of 120 mM. The product, β-D-glucose, was also a very poor inhibitor of the enzyme. At a 1 mM concentration, it gave a 20% decrease in Vmax, but as the concentration was increased, the rate also increased, approaching its original value.

      DISCUSSION

      The reaction catalyzed by myrosinase superficially resembles the hydrolysis of glycosides catalyzed by many β-glycosidase enzymes. It is obviously different in two respects. First, the substrate is a sulfur rather than an oxygen glycoside, and second, there is a subsequent rearrangement of the aglycone fragment, which may or may not be catalyzed by the enzyme. The initial results of our study of the chemical mechanism of myrosinase, reported here, have shown that the resemblance may indeed be only superficial, as there seem to be some important mechanistic differences. One interesting feature is that myrosinase is a very stable enzyme. This probably results from the role of the glucosinolate/myrosinase system as a defensive mechanism in the plant. In the cell, myrosinase is normally separated from the glucosinolates, and they only come together after cell damage, such as attack by pests or pathogens(
      • Larsen P.O.
      ). Glucosinolate hydrolysis then releases isothiocyanates, which are noxious to pests and deter them from further attack. Immediate response is vital, and therefore there must always be active enzyme present in the cell. As a result, myrosinase must be stable and long lived. This is in contrast to many other enzymes, which are synthesized by the cell machinery when they are required, in response to the current needs of the cell.
      Transglycosylation is commonly observed with β-glycosidases; however, we have been unable to obtain evidence for this reaction taking place with myrosinase. For example, Gopalan and co-workers (
      • Gopalan V.
      • Vander Jagt D.J.
      • Libell D.P.
      • Glew R.H.
      ) observed significant transglycosylation using mammalian cytosolic β-glucosidase in the presence of alcohol acceptors. The rate of hydrolysis of PNPG was increased 8-fold by the addition of 0.2 Mn-butyl alcohol, and the ratio of n-butyl-β-D-glucoside to D-glucose was 24:1. The proposed reaction pathway for the transglycosylation is given in Fig. 7. In the case of the β-glucosidase, it was proposed that for PNPG, k3 was rate-limiting. So on addition of butan-1-ol, which is more nucleophilic than water, an overall increase in rate was observed as k4 > k3.
      Figure thumbnail gr7
      Figure 7:Reaction pathway for hydrolysis of β-D-glucosides in the presence of competing glycosyl acceptors.
      However, the rate of p-nitrophenyl-β-D-glucoside hydrolysis catalyzed by myrosinase was not increased by the addition of alternative, and more reactive, glycosyl acceptors, and in fact decreased. Also, there was no decrease in the amount of D-glucose produced as a result of transglycosylation. A wide range of potential glycosyl acceptors were employed, none of which were successful. Even the very nucleophilic thiols did not trap out the glycosyl moiety. It was reasoned that the acceptors simply were unable to gain access to the active site. The natural substrate contains a negatively charged group in the aglycone, and so acceptors containing similar charged side chains were employed. Again, these proved to be unsuccessful at transglycosylation.
      The ease with which transglycosylation takes place with myrosinase is obviously much less than with β-glucosidase and β-galactosidase, despite the apparent similarities between the reactions that they catalyze. In the case of the β-glycosidases, reaction is thought to proceed either via an oxocarbonium ion (XI) or a covalent enzyme-glycosyl intermediate (XIV) (Fig. 8). This is formed following expulsion of the aglycone from the substrate. A similar pathway is feasible for myrosinase except that there is then a subsequent rearrangement of the aglycone.
      Figure thumbnail gr8
      Figure 8:Proposed chemical mechanisms for β-glucosidases.
      Inhibition of myrosinase by glucono-δ-lactone, a tight binding competitive inhibitor of β-glucosidase that mimics the oxocarbonium ion intermediate, was examined. However, this compound was not a competitive inhibitor for myrosinase and instead showed noncompetitive inhibition with a poor KI. This gives a second example of the unusual behavior of myrosinase. One possible reason for the observed noncompetitive inhibition is that the glucono-δ-lactone is not binding at the active site but at some other site on the enzyme. In the case of myrosinase, there exists a candidate for this alternative binding site. The enzyme is specifically activated by ascorbic acid(
      • Ohtsuru M.
      • Hata T.
      ), which apparently binds to the enzyme causing a change in its conformation. Therefore, it is possible that the glucono-δ-lactone binds at the ascorbic acid site, but in this case causes a decrease in the rate of reaction. Some preliminary experiments (Table 4) indicated that the degree of rate enhancement due to ascorbic acid is reduced in the presence of glucono-δ-lactone, decreasing as the inhibitor concentration increases. It certainly appears that the glucono-δ-lactone is not acting as a transition state analogue from its mode and degree of inhibitory activity. However, the simple glucoside, methyl-β-D-glucoside, was a competitive inhibitor, albeit with a poor KI.
      When discussing the implications of these findings on the chemical mechanism, it must always be remembered that myrosinase could also be involved in catalysis of the rearrangement of the aglycone. The early evidence for this reaction being spontaneous and not involving the enzyme is certainly not conclusive(
      • Ettlinger M.G.
      • Dateo Jr., G.P.
      • Harrison B.W.
      • Mabry T.J.
      • Thompson C.P.
      ). A particularly interesting observation is that the desulfoglucosinolates are only inhibitors of myrosinase and are not hydrolyzed by myrosinase(
      • Hanley A.B.
      • Kwiatkowska C.A.
      • Fenwick G.R.
      ). This may imply that the driving force for the reaction is actually the rearrangement step. If the enzyme was indeed required to effect this rearrangement, then it is very possible that the aglycone is not the first product released during the catalytic cycle. If the other product, D-glucose, is released first, then the ease of transglycosylation would be expected to be greatly reduced, as there would be little room for the glycosyl acceptor to bind at the active site while the aglycone was still present. Determination of the product debinding order, using product inhibition studies, would allow this possibility to be examined. Unfortunately, D-glucose gave very poor inhibition, which could not be assigned as either competitive or uncompetitive. Work is now underway to prepare the aglycone fragment, and stable analogues thereof, to assess their inhibitory properties and address this possibility in more detail.

      ACKNOWLEDGEMENTS

      We thank the BBSRC for an earmarked studentship (to M. G. T.) and the Royal Society of Edinburgh for a SOED/RSE Personal Research Fellowship (to N. P. B.).

      REFERENCES

        • Fenwick G.R.
        • Heaney R.K.
        • Mullin W.J.
        CRC Crit. Rev. Food Sci. Nutr. 1983; 18: 123-201
        • Ettlinger M.G.
        • Dateo Jr., G.P.
        • Harrison B.W.
        • Mabry T.J.
        • Thompson C.P.
        Proc. Natl. Acad. Sci. U. S. A. 1961; 47: 1875-1880
        • Tookey H.L.
        • Etten C.H.
        • Daxenbichler M.E.
        Liener I.E. Toxic Constituents of Plant Foodstuffs. Academic Press, New York1980: 103-140
        • Larsen P.O.
        Conn E.E. The Biochemistry of Plants. Vol. 7. Academic Press, New York1981: 501-525
        • Withers S.G.
        Curr. Opin. Struct. Biol. 1994; 4: 885-892
        • Durham P.L.
        • Poulton J.E.
        Z. Naturforsch. 1990; 45: 173-178
        • Bjorkman R.
        • Lonnerdal B.
        Biochim. Biophys. Acta. 1973; 327: 121-131
        • Ohtsuru M.
        • Hata T.
        Biochim. Biophys. Acta. 1979; 567: 384-391
        • Tsuruo I.
        • Hata T.
        Agric. Biol. Chem. 1968; 32: 1425-1431
        • Palmieri S.
        • Leoni O.
        • Iori R.
        Anal. Biochem. 1982; 123: 320-324
        • Schachter H.
        Methods Enzymol. 1975; 41: 3-10
        • Dale M.P.
        • Kopfler W.P.
        • Chait I.
        • Byers L.D.
        Biochemistry. 1986; 25: 2522-2529
        • Sinnot M.L.
        • Vitarelle O.M.
        Biochem. J. 1973; 133: 81-88
        • Toone E.J.
        • Simon E.S.
        • Bednarsta M.D.
        • Whitesides G.M.
        Tetrahedron. 1989; 45: 5365-5422
        • Bjorkling F.
        • Godtfredsen S.E.
        Tetrahedron. 1988; 44: 2957-2962
        • Ohtsuru M.
        • Tsuruo I.
        • Hata T.
        Agric. Biol. Chem. 1969; 33: 1320-1325
        • Hanley A.B.
        • Kwiatkowska C.A.
        • Fenwick G.R.
        J. Sci. Food Agric. 1990; 51: 417-420
        • Dale M.P.
        • Ensley H.E.
        • Kern K.
        • Sastry K.A.R.
        • Byers L.D.
        Biochemistry. 1985; 24: 3530-3539
        • Gopalan V.
        • Vander Jagt D.J.
        • Libell D.P.
        • Glew R.H.
        J. Biol. Chem. 1992; 267: 9629-9638