The biosynthesis of cyanogenic glucosides in higher plants. Identification of three hydroxylation steps in the biosynthesis of dhurrin in Sorghum bicolor (L.) Moench and the involvement of 1-ACI-nitro-2-(p-hydroxyphenyl)ethane as an intermediate.

N-Hydroxytyrosine, (E)- and (Z)-p-hydroxyphenyl-acetaldehyde oxime, p-hydroxyphenylacetonitrile, and p-hydroxymandelonitrile are established intermediates in the biosynthesis of the tyrosine-derived cyanogenic glucoside dhurrin (Halkier, B. A., Olsen, C. E., and Møller, B. L. (1989) J. Biol. Chem. 264, 19487-19494. Simultaneous measurements of oxygen consumption and biosynthetic activity using a microsomal enzyme system isolated from etiolated sorghum seedlings demonstrate a requirement for three oxygen molecules in the conversion of tyrosine to p-hydroxymandelonitrile. Two oxygen molecules are consumed in the conversion of tyrosine to (E)-p-hydroxyphenylacetaldehyde oxime, indicating the existence of a previously undetected hydroxylation step in addition to that resulting in the formation of N-hydroxytyrosine. Radioactively labeled 1-nitro-2-(p-hydroxyphenyl)ethane was chemically synthesized and tested as a possible intermediate. Biosynthetic experiments demonstrate that the microsomal enzyme system metabolizes the nitro compound to the subsequent intermediates in dhurrin synthesis (Km = 0.05 mM; Vmax = 14 nmol/mg of protein/h). Low amounts of 1-nitro-2-(p-hydroxyphenyl)ethane are produced in the microsomal reaction mixtures when tyrosine is used as substrate. These data support the involvement of 1-nitro-2-(p-hydroxyphenyl)ethane or more likely its aci-nitro tautomer as an intermediate between N-hydroxytyrosine and p-hydroxyphenylacetaldehyde oxime. The conversion of (E)-p-hydroxyphenylacetaldehydeoxime to p-hydroxymandelonitrile requires a single oxygen molecule. The oxygen molecule is utilized for hydroxylation of p-hydroxyphenylacetonitrile into p-hydroxymandelonitrile. This indicates that the conversion of p-hydroxyphenylacetaldehyde oxime into p-hydroxyphenylacetonitrile proceeds by a simple dehydration reaction.

Simultaneous measurements of oxygen consumption and biosynthetic activity using a microsomal enzyme system isolated from etiolated sorghum seedlings demonstrate a requirement for three oxygen molecules in the conversion of tyrosine to p-hydroxymandelonitrile.
Two oxygen molecules are consumed in the conversion of tyrosine to (E)-p-hydroxyphenylacetaldehyde oxime, indicating the existence of a previously undetected hydroxylation step in addition to that resulting in the formation of N-hydroxytyrosine. -0265-DK. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed.
somal enzyme system from etiolated sorghum seedlings catalyzes the conversion of tyrosine to (S)-p-hydroxymandelonitrile (2)(3)(4)(5). In uivo, (S)-p-hydroxymandelonitrile is glucosylated into dhurrin by a soluble UDP-glucose:glucosyltransferase (6). In the absence of the UDP-glucose:glucosyltransferase, (S)-p-hydroxymandelonitrile dissociates into p-hydroxybenzaldehyde and cyanide, which are therefore the end products obtained in vitro using the microsomal enzyme system. The biosynthetic pathway involves N-hydroxytyrosine, (E)-and (Z)-p-hydroxyphenylacetaldehyde oxime, p-hydroxyphenylacetonitrile, and (S)-p-hydroxymandelonitrile as intermediates and thus includes two hydroxylation steps ( Fig.  1) (2,4). The participation of each of these intermediates is well established. However, as indicated in Fig. 1, two of the biosynthetic steps have been suggested to represent multistep conversions. The conversion of tyrosine to p-hydroxyphenylacetaldehyde oxime represents a four-electron oxidative decarboxylation.
N-Hydroxytyrosine has been identified as an intermediate in this conversion (4). The retention of the (Yhydrogen atom of tyrosine in the conversion of tyrosine to phydroxyphenylacetaldehyde oxime excludes p-hydroxyphenylpyruvic acid oxime as intermediate between N-hydroxytyrosine and p-hydroxyphenylacetaldehyde oxime (2). Instead, the participation of 3-(p-hydroxyphenyl)-2-nitrosopropionic acid has been suggested (2,4). This compound, which can be envisualized formed by dehydrogenation of N-hydroxytyrosine, is labile and produces the oxime upon decarboxylation. The lability of the cu-nitrosocarboxylic acid could explain why this intermediate has not been isolated. The conversion of phydroxyphenylacetaldehyde oxime to p-hydroxyphenylacetonitrile has been reported to require NADPH as a cofactor (3)(4)(5), suggesting that this represents a multistep conversion embodying a redox reaction. In the biosynthesis of glucosinolates, an aci-nitro compound has been proposed as the intermediate following the oxime (7,8). Based on the apparent similarities between the biosynthetic pathways for glucosinolates and cyanogenic glucosides, an aci-nitro compound has subsequently been suggested as intermediate between the oxime and the nitrile in the biosynthetic pathway for cyanogenie glucosides (9,10). Chemical conversion of nitro compounds into nitriles is well established (11).
In  (12). Unlabeled (E)-and (Z)-p-hydroxyphenylacetaldehyde oxime were produced by oxidative decarboxylation of N-hydroxytyrosine in 1 N NH3 (13). [U-"C](E)-and (Z)-p-hydroxyphenylacetaldehyde oxime were produced enzymatically from [U-"Cltyrosine using the -SH microsomal enzyme system' as described earlier (2). 1-Nitro-2-@-hydroxyphenyl)ethane was chemically synthesized by sodium borohydride reduction of I-nitro-2-@-hydroxyphenyl)ethene obtained from condensation of benzaldehyde and nitromethane (14). A racemate of (R)-and (S)-  is administered to the +SH microsomal enzyme system in the presence of NADPH and 0, (4). The second and third methods for measuring biosynthetic activity involve the use of radioactively labeled substrates and separation of intermediates and end products by TLC (4) and HPLC (2). These methods are more informative since they permit the monitoring of individual steps in the pathway and the detection of intermediates accumulating in the reaction mixtures.
Since a racemate of (R)-and (S)-[2-3H1]1-nitro-2-(phydroxyphenyl)ethane was administered to the microsomal enzyme system, corrections were made for the loss of 50% of the radioactivity in the production of (S)-p-hydroxymandelonitrile. Using 14C-labeled substrates, corrections were made for the loss of one or two carbon atoms as products were formed.

Determination of Number of Hydroxylation
Steps Involved in Biosynthesis of Cyanogenic Glucosides-Molecular oxygen is consumed as a substrate in the conversion of tyrosine to N-hydroxytyrosine (4)3 and in the conversion of p-hydroxyphenylacetonitrile to p-hydroxymandelonitrile ( Fig. 1) (14). The validity of this pathway was tested by preparing the reaction mixtures in the chamber of an oxygen electrode permitting simultaneous quantitative measurements of biosynthetic activity and oxygen consumption (Fig. 2). Based on the experimentally determined amounts of intermediates and end products accumulated in the reaction mixture at the end of the incubation period, an expected oxygen consumption could be calculated and compared to the oxygen consumption measured. A typical set of data is shown in Table I. When tyrosine is used as substrate in combination with the +SH microsomal enzyme system, the oxygen consumption measured exceeds that calculated according to the generally accepted pathway ( that the conversion of oxime to p-hydroxybenzaldehyde involves a single hydroxylation reaction independent of the use of the (E)-or (Z)-oxime as substrate. These data are in accord with Pathway B, but not with Pathway C, and indicate that two molecules of oxygen are consumed in the conversion of tyrosine to p-hydroxyphenylacetaldehyde oxime. This was investigated experimentally by administration of tyrosine to the -SH microsomal enzyme system, which converts tyrosine to p-hydroxyphenylacetaldehyde oxime. The oxygen consumption compared to the oxime production demonstrates that two oxygen molecules are required in the conversion of tyrosine top-hydroxyphenylacetaldehyde oxime ( phenyl)ethane to the +SH microsomal enzyme system results in the production of [l-3H]p-hydroxybenzaldehyde, the end product of the +SH microsomal enzyme system (Fig. 3). Usually, only insignificant amounts of the intermediates positioned between the administered substrate and p-hydroxybenzaldehyde accumulate in the +SH microsomal reaction mixtures (2,4,15). However, using 1-nitro-2-(p-hydroxyphenyl)ethane as substrate, a considerable amount of p-hydroxyphenylacetaldehyde oxime accumulates in the reaction mixture (Fig. 3). When (2-3Hl]l-nitro-2-(p-hydroxyphenyl)ethane is incubated in reaction mixtures devoid of NADPH or microsomes, no production of p-hydroxyphenylacetaldehyde oxime, p-hydroxyphenylacetonitrile, or p-hydroxybenzaldehyde can be detected. Although the metabolic rates are low, the data demonstrate that the microsomal enzyme system is capable of catalyzing an NADPH-dependent conversion of 1-nitro-2-(p-hydroxyphenyl)ethane to phydroxyphenylacetaldehyde oxime and p-hydroxybenzaldehyde. The pH optimum for the production of p-hydroxybenzaldehyde is 7.5. When the pH of the reaction mixture is lowered from 7.5 to 7.0, the aldehyde production is reduced, whereas the accumulation of oxime is increased, indicating that the enzyme converting the oxime to p-hydroxybenzaldehyde is more sensitive to lowering of pH than the enzyme catalyzing the reduction of l-nitro-2-(p-hydroxyphenyl)ethane to p-hydroxyphenylacetaldehyde oxime. The apparent K, for 1-nitro-2-(p-hydroxyphenyl)ethane is 0.05 mM, and the V,,,,, for the combined rate of oxime and aldehyde production is 14 nmol/mg of protein/h. The values for K, and V,,, were calculated from Lineweaver-Burk plots since even at the very high substrate concentrations used, a constant maximum velocity was not reached.
When [2-3H1]l-nitro-2-(p-hydroxyphenyl)ethane is administered to the -SH microsomal enzyme system, [2-3HlJphydroxyphenylacetaldehyde oxime accumulates in the reaction mixture (Fig. 3). The nitro compound is not produced when '%-labeled oxime is administered to the -SH microso-ma1 enzyme system. 4 In agreement with the data on the oxygen consumption (Table I, Pathway B), this indicates that

Production of 1 -Nitro-2-(p-hydroxyphenyl)ethane by Microsomal Enzyme System-When
[U-'4C]tyrosine is administered to the +SH or -SH microsomal enzyme system, phydroxybenzaldehyde is the main product accumulating in the +SH microsomal reaction mixtures, whereas p-hydroxyphenylacetaldehyde oxime accumulates in the -SH microso-ma1 reaction mixtures (4). Accumulation of 14C-labeled lnitro-2-(p-hydroxyphenyl)ethane in both types of reaction mixtures is demonstrated by a specific blackening of the lnitro-2-(p-hydroxyphenyl)ethane band on the autoradiographs of the TLC plates after prolonged exposure.4 The blackening at this position is not observed in control experi-ments devoid of NADPH or microsomal enzyme system. The amount of 14C-labeled nitro compound accumulated is very low (Fig. 4). The +SH and -SH microsomal enzyme systems differ with respect to their ability to accumulate the nitro compound. At pH 7.9, the percentage of radioactivity present as nitro compound in the -SH microsomal reaction mixture is typically 0.9%, whereas the level observed with the +SH microsomal enzyme system only reaches 0.1% (Fig. 4) attempt to increase the accumulation of the nitro compound in the reaction mixture, the production of 1-nitro-2-(p-hydroxyphenyl)ethane from tyrosine was analyzed at different pH values (Fig. 4). The maximum amount of l-nitro-2-(phydroxyphenyl)ethane was found to accumulate at pH 7.9. However, at low pH, the nitro compound constitutes a larger proportion relative to the other intermediates produced due to the strong inactivation of the aldehyde-, nitrile-, and oximemetabolizing enzymes at low pH (4).
Effect of Exogenously Added I-Nitro-2-(p-hydroxyphenyl)ethane on Metabolic Activity of Microsomal Enzyme System-Trapping experiments, in which [U-'4C]tyrosine was administered to the +SH microsomal enzyme system in the presence of exogenously added unlabeled 1-nitro-2-(p-hydroxyphenyl)ethane, were designed in an attempt to increase the amount of radioactivity accumulating in the nitro compound. The presence of a trap of the nitro compound, however, does not result in an increase in the amount of radioactivity recovered in the nitro compound (Fig. 5). Using the +SH microsomal system, the amount of radioactivity in the nitro compound accounts for 0.04% of the total radioactivity when no trap is added. Upon the addition of 100 nmol of unlabeled nitro compound, no radioactivity is recovered in the nitro compound. The corresponding percentages using the -SH microsomal enzyme system are 0.6 and 0.2%, respectively (Fig. 5). The addition of a trap of the unlabeled nitro compound results in accumulation of the oxime in the reaction mixtures prepared with the +SH microsomal enzyme system (Fig. 5). Accumulation of the oxime was also observed when the 'H-labeled nitro compound was administered to the +SH microsomal enzyme system (Fig. 3). Administration of increasing amounts of unlabeled nitro compound inhibits the metabolism of tyrosine (Fig. 5) as well as of oxime and nitrile (Fig. 6). This general inhibitory effect of the nitro compound on the metabolic activities of the microsomal enzyme system probably explains why the trapping experiments with the nitro compound do not result in the expected accumulation of radioactivity in the nitro compound.

DISCUSSION
The biosynthetic pathway for cyanogenic glucosides is generally accepted to involve two hydroxylation reactions (2)(3)(4)(5). This study on the in vitro biosynthesis of the cyanogenic glucoside dhurrin in sorghum demonstrates the involvement of a third hydroxylation reaction. This conclusion is based on simultaneous quantitative measurements of oxygen consumption and biosynthetic activity which demonstrate that three oxygen molecules are required for the conversion of one molecule of tyrosine into p-hydroxybenzaldehyde. Two of these oxygen molecules are required for the conversion of tyrosine to p-hydroxyphenylacetaldehyde oxime and one molecule for the conversion of oxime to p-hydroxybenzaldehyde (Table I, Pathway B). Attempts to quantify the corresponding amount of NADPH being oxidized were unsuccessful due to the formation of varying amounts of an addition product between the enzymatically produced NADP' and the cyanide formed in the reaction mixture during incubation (16). The addition product has an absorption maximum at 325 nm and interferes with spectrophotometric NADPH determination at 340 nm (10).
The oxygen measurements demonstrate that the conversion of tyrosine to p-hydroxyphenylacetaldehyde oxime involves a hitherto undetected hydroxylation reaction in addition to that resulting in the production of N-hydroxytyrosine. The possible chemical structure of such putative, hitherto undetected intermediates is limited by the experimentally observed quantitative retention of the a-hydrogen of tyrosine in p-hydroxyphenylacetaldehyde oxime (2). In earlier studies (2, 4), 3-(phydroxyphenyl)-2-nitrosopropionic acid was suggested as an intermediate. However, this compound is derived from Nhydroxytyrosine by a dehydrogenation reaction, and its presence in the pathway would not explain the consumption of the second oxygen molecule. Another previously suggested intermediate in the pathway is 1-nitro-2-(p-hydroxyphenyl)ethane (9,lO). This compound was envisioned as being formed by oxidation of the oxime (10). Whereas formation of 1-nitro-2-(p-hydroxyphenyl)ethane would be associated with the consumption of an additional molecule of oxygen, the previously suggested position of the nitro compound as intermediate between the oxime and the aldehyde is not in accordance with the measured oxygen requirements of the partial reactions. The single oxygen molecule consumed in the conversion of p-hydroxyphenylacetaldehyde oxime to p-hydroxybenzaldehyde serves to hydroxylate p-hydroxyphenylacetonitrile top-hydroxymandelonitrile. This excludes the involvement of a hydroxylation step between the oxime and the nitrile and strongly supports that the conversion of the oxime to the nitrile proceeds as a simple dehydration. These data indicate that the nitro compound is formed by an alternative mechanism involving N-oxidation of N-hydroxytyrosine to produce 2-nitro-3-p-hydroxyphenylpropionic acid, which by decarboxylation gives rise to the formation of the l-aci-nitro-2-(p-hydroxyphenyl)ethane (Fig. 7). The a&form of the nitro compound is in tautomeric equilibrium with the parent nitro compound. Since the aci-nitro compound is the direct product of decarboxylation of the Lu-nitrocarboxylic acid, the acitautomer is most likely the biosynthetically active tautomer. The positioning of an aci-nitro compound between N-hydroxytyrosine andp-hydroxyphenylacetaldehyde oxime in the pathway is in agreement with the oxygen stoichiometry data. The presence of the aci-nitro compound as an intermediate in the pathway is justified by the biosynthetic data which demonstrate that the microsomal enzyme system catalyzes the conversion of tyrosine to 1-nitro-2-(p-hydroxyphenyl)ethane as well as the conversion of 1-nitro-2-(p-hydroxyphenyl)ethane to p-hydroxybenzaldehyde (Figs. 3 and 4). When tyrosine is administered to the -SH microsomal enzyme system, phydroxyphenylacetaldehyde oxime and 1-nitro-2-(p-hydroxyphenyl)ethane accumulate. Administration of the nitro compound to the -SH microsomal enzyme system results in production of the oxime, whereas no nitro compound is produced when the oxime is used as substrate.4 Although the incorporation percentages obtained are low, the biosynthetic data indicate that the nitro compound is an intermediate between N-hydroxytyrosine and p-hydroxyphenylacetaldehyde oxime. In a previous study, Mailer and Conn (4) stated that the nitro compound is an unlikely intermediate in the biosynthesis of cyanogenic glucosides since administration of 1-nitro-2-(p-hydroxyphenyl)ethane to the sorghum microso-ma1 enzyme system did not result in any detectable production of hydrogen cyanide. This conclusion was based on a spectrophotometric assay, which is less sensitive than the assay used in this study based on radioactively labeled substrates.
The pH optimum for the conversion of 1-nitro-2-(p-hydroxyphenyl)ethane to p-hydroxybenzaldehyde is 7.5 (Fig. 3). Even at the pH optimum, the conversion rate is low as evidenced by the V,,,.. value of 14 nmol/mg of protein/h for the nitro compound compared to values of 145 and 400 nmol/ mg of protein/h for tyrosine and p-hydroxyphenylacetaldehyde oxime, respectively (4). One possibility is that the nitro (1) At pH values around 7, the equilibrium between the tautomers favors the stable parent nitro compound. aci-Nitro compounds are weak acids formed by protonization of the nitronate anion (17). Attempts to increase the rate of the enzymatic reduction of the nitro compound to the oxime by lowering the pH were unsuccessful probably because the pH values necessary to generate significantly elevated amounts of the oci-nitro tautomer are too low to maintain enzymatic activity (Fig. 3). The microenvironment of the microsomal enzyme system catalyzing the coordinated formation and utilization of the nitro compound may serve to stabilize the aci-tautomer, thus maintaining the nitro compound on the uci-tautomer form suitable for metabolism. Exogenously added nitro compound exerts an inhibitory effect on the metabolic activity of the microsomal enzyme system as measured with tyrosine as well as with phydroxyphenylacetaldehyde oxime and p-hydroxyphenylacetonitrile as substrates (Figs. 5 and 6). Based on the structural similarities between the nitro compound and the oxime, the inhibitory effect on the metabolism of the oxime may be caused by a specific interaction between the nitro compound and the active site on the oxime-metabolizing enzyme. The inhibitory effect of the nitro compound on the metabolism of tyrosine and p-hydroxyphenylacetonitrile indicates that the nitro compound has an unspecific effect on the microsomal enzyme system. The inhibition of the metabolism of the oxime in the presence of the nitro compound may explain why phydroxyphenylacetaldehyde oxime accumulates together with p-hydroxybenzaldehyde in the +SH microsomal reaction mixtures where 1-nitro-2-(p-hydroxyphenyl)ethane is administered as substrate (Fig. 3) and why incubation of radioactively labeled tyrosine with the +SH or -SH microsomal enzyme system in the presence of a trap of unlabeled nitro compound results in the accumulation of radioactivity in the oxime rather than in the nitro compound (Fig. 5). The inability to trap '%-labeled 1-nitro-2-(p-hydroxyphenyl)ethane could indicate that the enzymatically produced nitro compound does not exchange freely with the exogenously added nitro compound. This observation may be explained by permeability barriers imposed by the microsomal vesicles. It could also Ho ' ' cHfct&--No, -0 -indicate that the nitro compound is metabolically channeled. In previous studies (2,15), it was shown that the sorghum microsomal enzyme system exhibits catalytic facilitation with respect to N-hydroxytyrosine, (E)-p-hydroxyphenylacetaldehyde oxime, andp-hydroxyphenylacetonitrile, i.e. the enzyme system metabolizes more efficiently the N-hydroxytyrosine, (E)-p-hydroxyphenylacetaldehyde oxime, and p-hydroxyphenylacetonitrile produced in situ compared with the same intermediates when exogenously added. The position of the nitro compound between N-hydroxytyrosine and (E)-p-hydroxyphenylacetaldehyde oxime, which are both channeled intermediates in the pathway, combined with the difficulties in trapping the nitro compound suggest that also 1-nitro-2-(p-hydroxyphenyl)ethane may be a metabolically channeled intermediate.
The +SH and -SH microsomal enzyme systems differ greatly in their ability to accumulate 1-nitro-2-(p-hydroxyphenyl)ethane (Fig. 4). Generally, the level of accumulated nitro compound is 10 times higher in reaction mixtures prepared with the -SH microsomal enzyme system than in those prepared with the +SH microsomal system. The -SH microsomes are manipulated to lose the ability to convert p-hydroxyphenylacetaldehyde oxime to p-hydroxybenzaldehyde by overnight dialysis against a buffer not containing dithiothreitol.
The higher percentage of l-nitro-2-(p-hydroxyphenyl)ethane accumulating in the -SH reaction mixtures may reflect a partial inactivation of the nitro compoundmetabolizing enzyme activity during the preparation of the -SH microsomal enzyme system.
Glucosinolates and cyanogenic glucosides are synthesized from amino acids. Oximes have been reported as common intermediates (9). From these similarities, the initial steps in the biosynthesis of these two groups of secondary plant products have been speculated to proceed by a common pathway (9). Matsuo  ). This plant contains the two tyrosine-derived cyanogenic glucosides triglochinin and dhurrin. The cyanogenic glucosides were not produced in the cell suspension cultures, which, however, were shown to accumulate the glucoside of 1-nitro-2-(p-hydroxyphenyl)ethane.
Whether accumulation of 1-nitro-2-(p-hydroxyphenyl)ethane occurs in the intact plant under osmotic stress or only in osmotically stressed cell suspension cultures was not reported. The production of the glucoside of the nitro compound in the osmotically stressed cell suspension cultures may represent a secondary unspecific oxidation of the oxime induced as a result of the inability of the cultures to convert the oxime to the cyanogenic glucosides which are the natural constituents of the plant. The occurrence of nitro compounds in biological material is rare. I-Nitro-2-(p-hydroxyphenyl)ethane has been found in the plant Thalictrum aquilegifolium, which also contains the tyrosine-derived cyanogenic glucosides dhurrin and taxiphyllin (19). 3-Nitropropionic acid and its alcohol are toxicants mainly found in legume families (e.g. Astragalus) (20). Many examples of loss of livestock from grazing on Astragalus with poisonous nitro compounds have been reported (21). Several nitro compounds isolated from bacteria have an antibiotic effect, the most important being chloramphenicol. It appears that nitro compounds are potential inhibitors of many reactions.
In conclusion, we have demonstrated the requirement of three oxygen molecules in the biosynthesis of cyanogenic glucosides. The two oxygen molecules consumed in the conversion of tyrosine to p-hydroxyphenylacetaldehyde oxime demonstrate the involvement of a hitherto unknown hydroxylation step. The one oxygen molecule consumed between phydroxyphenylacetaldehyde oxime and p-hydroxybenzaldehyde indicates that the conversion of p-hydroxyphenylacetaldehyde oxime to p-hydroxyphenylacetonitrile proceeds by a simple dehydration.
The occurrence of the additional hydroxylation step compared with the biosynthetic data presented shows the involvement of 1-aci-nitro-2-(p-hydroxyphenyl)ethane as an additional intermediate between N-hydroxytyrosine andp-hydroxyphenylacetaldehyde oxime in the biosynthesis of the cyanogenic glucoside dhurrin (Fig. 7).

B A Halkier and B L Møller
intermediate.

Moench and the involvement of 1-ACI-nitro-2-(p-hydroxyphenyl)ethane as an hydroxylation steps in the biosynthesis of dhurrin in Sorghum bicolor (L.)
The biosynthesis of cyanogenic glucosides in higher plants. Identification of three