2-Nitro-3-(p-hydroxyphenyl) - Europe PMC

2 downloads 0 Views 1MB Size Report
Sep 4, 1990 - We thank Hanne Linde Nielsen and Inga Olsen for excellent technical assistance, Dr. J0rgen 0gaard Madsen for recording the mass spectra ...
Proc. Nati. Acad. Sci. USA Vol. 88, pp. 487-491, January 1991 Biochemistry

2-Nitro-3-(p-hydroxyphenyl)propionate and aci-1-nitro-2-(phydroxyphenyl)ethane, two intermediates in the biosynthesis of the cyanogenic glucoside dhurrin in Sorghum bicolor (L.) Moench (N-hydroxylation/N-hydroxyTosine/10 incorporation/glucosinolates)

BARBARA ANN HALKIER, JENS LYKKESFELDT, AND BIRGER LINDBERG M0LLER* Plant Biochemistry Laboratory, Department of Plant Biology, Royal Veterinary and Agricultural University, 40 Thorvaldsensvej, DK-1871 Frederiksberg C, Copenhagen, Denmark

Communicated by Eric E. Conn, October 1, 1990 (received for review September 4, 1990)

ABSTRACT

The biosynthetic pathway for the cyanogenic

glucoside dhurrin derived from tyrosine has been studied in vitro

by using ['80]oxygen and a microsomal enzyme system obtained from etiolated sorghum seedlings. The products formed were purified by HPLC and TLC, and the incorporation of ['8O]oxygen was monitored by mass spectrometry. In the presence of NADPH and ['8'Odioxygen, L-tyroSine is converted to (E)- and (Z)-p-hydroxyphenylacetaldehyde oxime with qmantitative incorporation of an ['0]oxygen atom into the oxime function. The first step in this conversion is the N-hydroxylation of L-tyrosine to N-hydroxytyrosine. Administration of N-hydroxytyrosine as a substrate results in the production of 1-nitro-2-(phydroxyphenyl)ethane in addition to (E)- and (Z)-p-hydroxyphenylacetaldehyde oxime, with quantitative incorporation of a single [O81oxygen atom in all three products. These data demonstrate that the conversion of N-hydroxytyrosine to p-hydroxyphenylacetaldehyde oxime involves additional N-hydroxylation and N-oxidation reactions giving rise to the formation of 2-nitro-3-(p-hydroxyphenyl)propionate, which by decarboxylation produces aci-l-nitro-2-(p-hydroxyphenyl)ethane. Both compounds are additional intermediates in the pathway. The two [180oxygen atoms introduced by the N-hydroxylations are enzymatically dthable as demonstrated by the specfic loss of the oxygen atom introduced by the first N-hydroxylation reaction in the subsequent conversion of aci-1-nitro-2-(phydroxyphenyl)ethane to (E)-p-hydroxyphenylactaldehyde oxime. A high flux of intermediates through the microsomal enzyme system is obtained with N-hydroxytyrosine as a substrate. This renders the conversion of the ad-nitro compound rate limiting and results in its release from the active site of the enzyme system and accumulation of the tautomeric nitro compound.

Sorghum [Sorghum bicolor (L.) Moench] seedlings synthesize the cyanogenic glucoside dhurrin (fi-D-glucopyranosyloxy(S)-p-hydroxymandelonitrile) (1). A microsomal enzyme system obtained from etiolated sorghum seedlings catalyzes the in vitro conversion of the parent amino acid tyrosine to the cyanohydrin p-hydroxymandelonitrile (2, 3). Biosynthetic studies using this experimental system have identified N-hydroxytyrosine, (E)-p-hydroxyphenylacetaldehyde oxime, (Z)p-hydroxyphenylacetaldehyde oxime, and p-hydroxyphenylacetonitrile as intermediates (2-5). In vivo, a soluble UDPglucose glucosyltransferase converts the cyanohydrin to the cyanogenic glucoside (6) (Fig. 1). Subsequent studies using microsomal preparations from a number of other plant species have revealed the same biosynthetic pathway (7-11). Simultaneous measurements of oxygen consumption and tyrosine metabolism using the microsomal enzyme system

isolated from sorghum have shown that three molecules of oxygen are consumed for each molecule of p-hydroxymandelonitrile produced and that the conversion of tyrosine to p-hydroxyphenylacetaldehyde oxime proceeds with the consumption of two oxygen molecules (12). N-Hydroxytyrosine has been identified as an intermediate in the latter conversion (3), and its formation consumes one of the two oxygen molecules. The conversion of N-hydroxytyrosine to p-hydroxyphenylacetaldehyde oxime represents a two-electron oxidative decarboxylation. The number of possible intermediates between N-hydroxytyrosine and p-hydroxyphenylacetaldehyde oxime is restricted by the fact that isotope experiments using L-[a-2HJtyrosine as substrate have demonstrated quantitative retainment of the a-hydrogen atom of tyrosine in the p-hydroxyphenylacetaldehyde oxime produced (4). Biosynthetic studies have demonstrated that the microsomal enzyme system is able to produce and metabolize 1-nitro-2(p-hydroxyphenyl)ethane and indicate that the nitro compound is positioned between N-hydroxytyrosine and p-hydroxyphenylacetaldehyde oxime (12). However, the amount of 1-nitro-2-(p-hydroxyphenyl)ethane accumulated and metabolized is low compared with the amounts observed with the previously identified intermediates (2, 3, 5), indicating the possible involvement of secondary metabolic transformations not directly related to the biosynthetic pathway. Hosel et al. (13) have observed the production in vitro of 1-nitro-2-(p-hydroxyphenyl)ethane from tyrosine by microsomal preparations obtained from osmotically stressed cell suspension cultures ofCalifornia poppy (Eschscholtzia California Cham.). California poppy contains the two tyrosine-derived cyanogenic glucosides dhurrin and triglochinin. The cellsuspension cultures produce the phenolic glucoside of the nitro compound, whereas they do not produce the cyanogenic glucosides. Conversely, microsomes prepared from 6-day-old seedlings of California poppy did not produce 1-nitro-2-(phydroxyphenyl)ethane (13). The production of the nitro compound and its glucoside was concluded to be elicited by the osmotic stress conditions (13). In this paper, we report the results of incorporation experiments with [180dioxygen, which demonstrate that 2-nitro-3-(p-hydroxyphenyl)propionate and aci-1-nitro-2-(phydroxyphenyl)ethane are obligatory intermediates positioned between N-hydroxytyrosine and p-hydroxyphenylacetaldehyde oxime in the biosynthesis of the cyanogenic glucoside dhurrin.

MATERIALS AND METHODS Isotope. [180]Dioxygen (99.2 atom % excess) was purchased from Amersham and stored in a gas burette over a surface of ethylene glycol.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

*To whom reprint requests should be addressed.

487

488

HO

Biochemistry: Halkier et al.

Proc. Natl. Acad. Sci. USA 88 (1991)

O7H)rHCHH-coOH HOO -Q)HcI4CH-COOH -HOJ)O

CHtCH-COOH

0'

OH

L-tyrosine

Nhydroxytyrosine

CHiCN

*-

p-hydroxyphenylacetonitrile

HO-§ CHiCH2-NOI

C

2-nitro-3-(p-hydroxy-

1-aci-nitro-2-(phydroxy-

phenyl)propionic acid

HO-f

HO CHf

HO-jCHgCsH

phenyl)ethane

HO

-

6H (Z)-p-hydroxyphenyl-

1 -nitro-2-(p-hydroxy-

phenyl)ethane

jC*.C$H

qN-OH

(E)-p-hydroxyphenyl-

acetaldehyde oxime

acetaldehyde oxime

1rNADPH + Og

WKAPG HO-FCHO + HCN p-hydroxybenzaldehyde

in vitro -*

hydrogen cyanide

HO

CH-CN 6=H

in vivo

HO-

p-hydroxymandelonitrile

o9CCN

\=

6-gluome

dhurrin

FIG. 1. The biosynthetic pathway for the tyrosine-derived cyanogenic glucoside dhurrin from S. bicolor. The intermediates positioned between N-hydroxytyrosine and (E)-p-hydroxyphenylacetaldehyde oxime are identified in this study.

Chemical Syntheses. (R,S)-N-Hydroxytyrosine was synthesized by chemical reduction of p-hydroxyphenylpyruvic acid oxime with sodium cyanoborohydride (14). (E)- and (Z)-p-hydroxyphenylacetaldehyde oxime were synthesized in an =1:1 ratio by oxidative decarboxylation of N-hydroxytyrosine in 1 M NH3 (15). Each of the isomers was isolated from the reaction mixture by reverse-phase HPLC (4). 1-Nitro-2-(p-hydroxyphenyl)ethane was synthesized by sodium borohydride reduction of 1-nitro-2-(p-hydroxyphenyl)ethene obtained by condensation of p-hydroxybenzaldehyde and nitromethane (16). Preparation of the Microsomal Enzyme System. The microsomal enzyme system was prepared from etiolated seedlings of S. bicolor (L.) Moench (hybrid Hybridum) (Seedtec International, Hereford, TX) as earlier described (1, 3). When prepared in the presence of dithiothreitol, the microsomal enzyme system catalyzes the conversion of tyrosine to p-hydroxymandelonitrile (2, 3). When prepared in the absence of dithiothreitol and subjected to dialysis for 2 days in the absence of dithiothreitol, the microsomal enzyme system selectively catalyzes the conversion of tyrosine to the stereoisomeric p-hydroxyphenylacetaldehyde oximes (2, 17). The latter type of preparation was used in the present study. ['80]Oxygen Labeling Experiments. The biosynthetic reactions were carried out in septum-covered glass reaction vessels (5 ml) fitted with an inlet for a vacuum line and one for the administration of ['8O]dioxygen. The pH and type of buffer used in the biosynthetic reaction mixtures were varied to optimize the metabolic conversion rate for each specific substrate used (3). N-Hydroxytyrosine (1.50 ,umol) was administered to reaction mixtures containing 200 ,ul of microsomal enzyme system (2.0 mg of protein) and 30 ,mol of potassium phosphate (pH 6.5) in a total volume of 530 ,ul. Tyrosine, 1-nitro-2-(p-hydroxyphenyl)ethane, or p-hydroxyphenylacetaldehyde oxime (0.55 umol) was administered to reaction mixtures containing 240 ,ul of microsomal enzyme system (2.4 mg of protein) and 5 ,umol of N-tris(hydroxymethyl)methylglycine (Tricine) (pH 7.9) in a total volume of 530 ,l. After flushing with argon for 5 min and subsequent evacuation (13 mm Hg; 1 mm Hg = 133 Pa), ['80]dioxygen

was admitted from the gas burette until atmospheric pressure was obtained in the reaction vessels. NADPH (0.90 ,tmol) was injected, and the reaction mixture was incubated 1 hr at 30°C. At the end of the incubation period, accumulated intermediates were separated by direct injection of the reaction mixture onto a reverse-phase HPLC column (Fig. 2) (4). The HPLC effluents containing 1-nitro-2-(p-hydroxyphenyl)ethane and (E)- and (Z)-p-hydroxyphenylacetaldehyde oximes were collected and lyophilized, and the intermediates were purified by TLC [Merck F254 aluminum plates; hexane/ ether, 1:3 (vol/vol)]. Mass spectra were recorded on a VG Masslab TRIO-2 mass spectrometer.

RESULTS The mass spectra of 1-nitro-2-(p-hydroxyphenyl)ethane and p-hydroxyphenylacetaldehyde oxime display prominent molecular ions [M]+ at m/z 167 and 151, respectively. The incorporation of [180]oxygen into the nitro compound and the £

TYR

CHO

C~~~~C

|o

' -I

0

5

E)OXa

( 10

20 15 Retention time (min)

25

I

30

FIG. 2. Separation by reverse-phase HPLC of reference compounds representing those involved in dhurrin biosynthesis. The separation was carried out using a Nucleosil C18 (10-Am) column and isocratic elution with 2% 2-propanol in 25 mM Hepes (pH 7.9) (4). TYR, tyrosine; CHO, p-hydroxybenzaldehyde; CN, p-hydroxyphenylacetonitrile; OX, p-hydroxyphenylacetaldehyde oxime; NO2, 1-nitro-2-(p-hydroxyphenyl)ethane. Neither CHO nor CN is produced under the experimental conditions used in this study.

Biochemistry: Halkier et al. oxime can therefore be quantitatively assessed from the shift in intensities of the molecular ions [M]+ from m/z 167 to 169 and from 151 to 153, respectively. The diagnostic fragment ions [HO-Ph-CH=CH21' at m/z 120 and [HO-Ph-CH2]+ at m/z 107 constitute the base peak in the spectrum of the nitro compound and the oxime, respectively. The masses of these fragment ions remained at m/z 120 and 107, regardless of whether the isolation of the nitro compound and the oxime from biosynthetic experiments was carried out in an 1802 or 1602 atmosphere. This demonstrates that with the substrates used exchange reactions resulting in incorporation of [18O]oxygen into the aromatic hydroxyl group do not occur. In the biosynthetic incorporation experiments, all relative intensities are corrected for the 0.2% natural abundance of the 180 isotope. Tyrosine and N-hydroxytyrosine were used as substrates in the biosynthetic experiments together with [18O]dioxygen (Table 1). The [18O]dioxygen enrichment of the atmosphere in the reaction vessels was 90.5 ± 0.1 atom % excess as measured by gas mass spectrometry. When tyrosine is used as substrate in the presence of NADPH and 1802, HPLC analyses show the accumulation of the (E)- and (Z)-phydroxyphenylacetaldehyde oximes in the reaction mixture, whereas no 1-nitro-2-(p-hydroxyphenyl)ethane is detectable. The mass spectra of the (E)- and (Z)-p-hydroxyphenylacetaldehyde oximes isolated from the reaction mixtures are identical to those of the authentic compounds except for shifts related to incorporation of an [180]oxygen atom into the oxime group. For the (E)- and (Z)-p-hydroxyphenylacetaldehyde oximes, the enrichment based on the intensities of the m/z 153 and m/z 151 molecular ions is 89.1% and 88.1%, respectively. These two molecular ions exhibit identical evaporation profiles. Thus incorporation of [18O]oxygen into the oxime function takes place without dilution, demonstrating that dioxygen serves as the sole source for this oxygen atom.

Administration of N-hydroxytyrosine to the microsomal enzyme system in the presence of NADPH and [180]dioxygen results in the accumulation of 1-nitro-2-(p-hydroxyphenyl)ethane and (E)- and (Z)-p-hydroxyphenylacetaldehyde oxime. In some experiments, the amount of the nitro compound produced equals that of the oxime isomers, whereas in other experiments it is somewhat lower. The fragmentation pattern of the isolated nitro compound is identical to that of the chemically synthesized standard, except for differences attributed to the incorporation of an [180]oxygen atom into the nitro group. Thus the molecular ion was shifted to [M]+ m/z 169 with an enrichment of 88.0% based on the intensities of the m/z 169 and m/z 167 ions. These ions also have identical evaporation profiles. The incorporation of [180]oxTable 1. Incorporation of [18O]oxygen into components produced by the microsomal enzyme system in incubation atmospheres containing [180]dioxygen

[180]Oxygen enrichment of isolated intermediate, Dilution of atom % excess [180]oxygen Substrate NO2 (E)-OX (Z)-OX NO2 (E)-OX (Z)-OX L-Tyrosine 89.1 88.1 1.02 1.03 N-Hydroxytyrosine 88.0 29.8 26.9 1.03 3.04 3.36 N-Hydroxytyrosine 1.22* 1.20* Dilution of [180]oxygen was calculated as the ratio of the [180]oxygen content of the incubation atmosphere to that of the intermediates isolated. The [180]dioxygen enrichment of the incubation atmosphere was 90.5 atom % excess. NO2, 1-nitro-2-(p-hydroxyphenyl)ethane; OX, p-hydroxyphenylacetaldehyde oxime. *Corrected for p-hydroxyphenylacetaldehyde oxime produced due to chemical decomposition of N-hydroxytyrosine.

Proc. NaMl. Acad. Sci. USA 88 (1991)

489

ygen without any dilution demonstrates that the conversion of N-hydroxytyrosine to 1-nitro-2-(p-hydroxyphenyl)ethane proceeds with quantitative incorporation of an oxygen atom derived from dioxygen. As in the experiments with tyrosine as a substrate, the mass spectra of the (E)- and (Z)-p-hydroxyphenylacetaldehyde oximes isolated from the reaction mixtures with N-hydroxytyrosine as substrate are identical to those of the authentic standards except for the shifts related to incorporation of an [180]oxygen atom into the oxime group. The enrichment based on the intensities of the m/z 153 and m/z 151 molecular ions is 29.8% and 26.9% for the E and Z isomers, respectively, corresponding to dilution factors of 3.04 and 3.36. The dilution reflects the previously reported chemical decomposition of N-hydroxytyrosine into p-hydroxyphenylacetaldehyde oxime, which occurs along with the enzymatic conversion (15). The amount of (E)- and (Z)-p-hydroxyphenylacetaldehyde oximes formed by chemical decomposition was determined in control experiments in which NADPH was omitted from the incubation mixture and was found to constitute 60% and 64%, respectively, of the amount of E and Z isomers formed in the biosynthetic experiments carried out in the presence of NADPH. We have earlier reported that chemical decomposition of N-hydroxytyrosine produces the E and Z isomers in an -1:1 ratio (4). In the present experiment the ratio determined by the HPLC analysis was 1.57:1. The microsomal enzyme system initially produces the E isomer, which subsequently is enzymatically converted to the Z isomer (4). Generally, the ratio between the enzymatically produced E and Z isomers varies between 2.6 and 3.3 (4). This ratio explains why the calculated dilution factor is higher for the Z isomer compared with the E isomer. When the dilution factors are corrected for the contribution of the chemically formed oxime isomers, dilution factors of 1.22 and 1.20 are obtained. These dilution factors demonstrate an almost quantitative incorporation of an [180]oxygen atom derived from dioxygen into each of the two oxime isomers when N-hydroxytyrosine is used as substrate and unequivocally establishes that the enzymatic conversion of N-hydroxytyrosine to p-hydroxyphenylacetaldehyde oxime requires an N-oxidation reaction. It also demonstrates that the oxygen atom of the hydroxyamino group of N-hydroxytyrosine is lost in the overall conversion of N-hydroxytyrosine to p-hydroxyphenylacetaldehyde oxime. Except for the chemical decomposition of N-hydroxytyrosine into the stereoisomeric p-hydroxyphenylacetaldehyde oximes, the production of 1-nitro-2-(p-hydroxyphenyl)ethane and (E)-p-hydroxyphenylacetaldehyde oxime in the biosynthetic reaction mixtures is strictly dependent on the presence of NADPH. Incubation of N-hydroxytyrosine, 1-nitro-2-(p-hydroxyphenyl)ethane, and p-hydroxyphenylacetaldehyde oxime with the microsomal system in an 1802 atmosphere in the presence or absence of NADPH and subsequent reisolation of the administered substrate revealed that exchange reactions leading to the incorporation of [180joxygen in the substrates do not occur. Similarly, chemical decomposition of N-hydroxytyrosine to p-hydroxyphenylacetaldehyde oxime proceeds without any incorporation of [180]oxygen. In no case did the administration ofthe oxime isomers to the microsomal enzyme system result in the production of the nitro compound.

DISCUSSION In the present study, we have used incorporation experiments with the stable isotope [180]oxygen to conclusively demonstrate the obligatory involvement of an N-oxidation reaction in the enzymatic conversion of N-hydroxytyrosine to p-hydroxyphenylacetaldehyde oxime. Taken together with the initial N-hydroxylation reaction of tyrosine to N-hy-

490

Biochemistry: Halkier et al.

Proc. Natl. Acad. Sci. USA 88 (1991)

droxytyrosine (3), these data explain the previously reported consumption of two oxygen molecules in the conversion of tyrosine to (E)- and (Z)-p-hydroxyphenylacetaldehyde oxime (12). The interpretation of the previously obtained biosynthetic data on the nitro compound was ambiguous due to the low metabolic conversion rate of 1-nitro-2-(p-hydroxyphenyl)ethane when compared with the established intermediates. The Vmax value for the nitro compound is 14 nmol per hr per mg of protein (12) compared with values of 145, 345, and 400 nmol per hr per mg of protein for tyrosine, N-hydroxytyrosine, and p-hydroxyphenylacetaldehyde oxime, respectively (3). Only minute amounts of the nitro compound accumulated in biosynthetic reaction mixtures with tyrosine as substrate (12). In the present study, the use of N-hydroxytyrosine as a substrate resulted in the accumulation of significant amounts of the nitro compound. This indicates that the initial N-hydroxylation of tyrosine to produce N-hydroxytyrosine is the overall limiting reaction in the pathway. At the increased flux of intermediates obtained with N-hydroxytyrosine as substrate, a later step in the conversion presumably becomes rate limiting as indicated by the accumulation of 1-nitro-2-(p-hydroxyphenyl)ethane. When produced from N-hydroxytyrosine in an [18O]dioxygen atmosphere, one of the oxygen atoms of the nitro group is labeled without any dilution (Table 1). In the same reaction mixtures, an [18O]oxygen atom is almost quantitatively incorporated into the oxime produced. The latter result demonstrates that the two oxygen atoms introduced by the two successive N-oxidation reactions are enzymatically distinguishable throughout the subsequent enzymatic transformations. This excludes 1-nitro-2-(p-hydroxyphenyl)ethane as a freely diffusible intermediate in the pathway. Only one biosynthetic route accounts for the accumulation of the nitro compound in the conversion of N-hydroxytyrosine to p-hydroxyphenylacetaldehyde oxime and at the same time is consistent with the previously obtained biosynthetic data. This pathway involves the N-oxidation of N-hydroxytyrosine giving rise to the formation of 2-nitro-3-(phydroxyphenyl)propionate. This conversion represents a four-electron oxidation composed of the dioxygen and NADPH-dependent hydroxylation reaction and a twoelectron oxidation reaction. If the two-electron oxidation reaction precedes the N-hydroxylation, then the substrate for the N-hydroxylation reaction is 3-(p-hydroxyphenyl)-2nitrosopropionate. Due to their chemical lability, it is not possible to directly test the a-nitrosocarboxylate and the a-nitrocarboxylate as substrates for the microsomal enzyme system. a-Nitrosocarboxylates decarboxylate to oximes. a-Nitrocarboxylates are labile in aqueous solutions at physiological pH and decarboxylate to aci-nitro compounds (18, 19). According to this reaction, the aci-tautomer of the nitro compound is the enzymatically active species. aci-Nitro

compounds are in equilibrium with their parent nitro compounds. The equilibrium favors the nitro compound (19, 20), and the observed accumulation of the nitro compound with N-hydroxytyrosine as a substrate may reflect tautomerization of the aci-nitro compound released from the active site on the microsomal enzyme system. The involvement of the aci-nitro compound as the enzymatically active species would explain the low metabolic activity observed upon administration of the parent nitro compound (12). The almost quantitative incorporation of ['8O]oxygen into the oxime with N-hydroxytyrosine as substrate (Table 1) precludes the possibility of free rotation around the C-N bond of the a-nitrocarboxylate ion. An equilibrium between 1-nitro-2-(p-hydroxyphenyl)ethane with free rotation around the C-N bond and the aci-nitro tautomer during the enzymatical conversion to the oxime is similarly excluded. These conclusions are in agreement with the earlier reported strong channeling of the conversion of tyrosine to p-hydroxyphenylacetaldehyde oxime as demonstrated by double-labeling experiments (17). The fixed orientation of the oxygen atoms introduced by the successive N-oxidation reactions may be accomplished by hydrogen bonds, metal chelation, or the presence of a positively charged amino acid residue provided by the active site of the enzyme. Alternative routes for the conversion of N-hydroxytyrosine to p-hydroxyphenylacetaldehyde oxime are less likely. The involvement of an a-aci-nitrocarboxylic acid can be excluded since this compound lacks the a-hydrogen atom of tyrosine. Experiments have shown that this hydrogen atom is retained when tyrosine is converted to p-hydroxyphenylacetaldehyde oxime (4). An alternative route would involve initial decarboxylation of N-hydroxytyrosine to produce N-hydroxytyramine, which by an N-oxidation reaction may be converted into the nitro compound. Since the sorghum microsomal enzyme system does not utilize N-hydroxytyramine as a substrate and since N-hydroxytyramine is not labeled in trapping experiments (3), this route is considered unlikely. A pathway involving dehydrogenationofN-hydroxytyrosine to 3-(p-hydroxyphenyl)-2-nitrosopropionate and subsequent decarboxylation to produce the oxime has earlier been proposed (4, 21). This pathway is not consistent either with the dioxygen stoichiometry data (12) or with the quantitative incorporation of an [18O]oxygen atom from dioxygen into p-hydroxyphenylacetaldehyde oxime when N-hydroxytyrosine is used as substrate as reported in this study. The conversion of tyrosine to p-hydroxyphenylacetaldehyde oxime is therefore concluded to involve two N-hydroxylations and a two-electron oxidation with an N-hydroxyamino acid and an a-nitrocarboxylic acid as intermediates. The a-nitrocarboxylic acid decarboxylates to produce the aci-nitro compound, which is stereoselectively reduced to an oxime. The data presented exclude the previously suggested positioning of the nitro compound as an intermediate between the oxime R-cs H-- Cyanogenic glucosides

|77,H R-CH-cOOi

NH2

-

R-CH-COOH

NHOH

NO

R-CH-COOH

NO2

,H

NOH

R-CH-COOH

~~4H2

NR-CH-OOH NHOH

R-CH-COOH

NO2

-

9 -

Glucosinolates FIG. 3. Comparison of the biosynthetic pathways for cyanogenic glucosides and glucosinolates with aci-nitro compounds as the branching

point.

Biochemistry: Halkier et aL and the nitrile in the biosynthesis of cyanogenic glucosides (22-24). The biosynthetic studies indicate that the conversion of the nitro compound to the oxime is irreversible (12). a-Nitrocarboxylic acids have never been isolated from biological material. In 1968, Ettlinger and Kjaer (25) hypothesized that these might be intermediates in the biosynthesis of glucosinolates. Recently, L-nitrosuccinate has been reported as an intermediate in the biosynthesis of 3-nitropropionate from L-aspartate in Penicillium atrovenetum (26). Due to the lability of the a-nitrocarboxylic acid, the biosynthetic experiments were carried out by administration of the stable diethyl ester, reasoning that hydrolysis in vivo would release the free a-nitrocarboxylic acid within the fungal cells (26). The conversion of L-aspartate to 3-nitropropionate proceeds with both oxygens of the nitro group being derived from dioxygen (27) and with retention of the a-hydrogen atom (26). These data would support a pathway analogous to that here reported for the biosynthesis of dhurrin. In the biosynthesis of glucosinolates (28), an aci-nitro compound has been suggested as an intermediate between the oxime and the S-alkylthiohydroximic acid (29). The involvement of an aci-nitro compound in this pathway is attractive due to the possibility of forming the S-alkylthiohydroximic acid by a nucleophilic attack on the a-carbon atom of the aci-nitro compound and a subsequent elimination (25). The suggestion was supported by the demonstration of 1-nitro-2-phenylethane as a precursor of benzylglucosinolate in Tropaeolum majus (29). Trapping experiments where '4C-labeled phenylacetaldehyde oxime was fed to T. majus in the presence of unlabeled nitro compound resulted in the production of '4C-labeled 1-nitro-2-phenylethane (29). Other experiments have shown that oximes are precursors for glucosinolates (30, 31) and that amino acids are precursors for the oximes in glucosinolate-containing plants (32). These results have led to the proposal that the oxime and the aci-nitro compound are intermediates in the biosynthesis of glucosinolates (25) and that the aci-nitro compound is positioned after the oxime (29). If these conclusions are correct, the position of the aci-nitro compound and the oxime as intermediates in the biosynthesis of cyanogenic glucosides and glucosinolates is inverted. We do not favor this hypothesis. Criteria for the establishment of a compound as a true intermediate include the demonstration of enzymatic production and utilization of the compound (32, 33), but conclusions based solely on biosynthetic experiments become ambiguous if the compound tested is in an enzyme-catalyzed or chemical equilibrium with a true intermediate (32, 33). On the basis of the results presented here on the biosynthesis of cyanogenic glucosides, we speculate that the observed production of 1-nitro-2-phenylethane from phenylacetaldehyde oxime and the demonstrated trapping of oximes in feeding studies with amino acids represents side reactions in relation to the biosynthetic pathway for glucosinolates. We propose that the aci-nitro compound is the branching point between the two biosynthetic pathways (Fig. 3). Thus in plants producing cyanogenic glucosides, the aci-nitro compound is enzymatically reduced to the E-oxime, whereas in plants producing glucosinolates, the aci-nitro compound is converted to an S-alkylthiohydroximic acid by nucleophilic attack from a

sulfhydryl compound.

We thank Hanne Linde Nielsen and Inga Olsen for excellent technical assistance, Dr. J0rgen 0gaard Madsen for recording the mass spectra, Profs. Anders Kjaer and Peder Olesen Larsen for

Proc. NatL. Acad. Sci. USA 88 (1991)

491

helpful comments about the manuscript, and Dr. Jim Foster (Seedtec International Inc., Hereford, TX) for continuously supplying the sorghum seeds. The work was partially supported by the Danish Agricultural and Veterinary Research Council, the Danish International Development Agency, the Danish Government Program for Biotechnology Research, the Carlsberg Foundation, the Rockefeller Foundation, and the Commission of the European Communities Science and Technology for Development. 1. Halkier, B. A. & M0ller, B. L. (1989) Plant. Physiol. 90, 1552-1559. 2. McFarlane, I. J., Lees, E. M. & Conn, E. E. (1975) J. Biol. Chem. 250, 4708-4713. 3. Moller, B. L. & Conn, E. E. (1979) J. Biol. Chem. 254, 8575-8583. 4. Halkier, B. A., Olsen, C. E. & Moller, B. L. (1989) J. Biol. Chem. 264, 19487-19494. 5. Shimada, M. & Conn, E. E. (1977) Arch. Biochem. Biophys. 180, 199-207. 6. Reay, P. F. & Conn, E. E. (1974) J. Biol. Chem. 249, 58265830. 7. Collinge, D. B. & Hughes, M. A. (1984) Plant Sci. Lett. 34, 119-125. 8. Cutler, A. J., Sternberg, M. & Conn, E. E. (1985) Arch. Biochem. Biophys. 238, 272-279. 9. Hosel, W. & Nahrstedt, A. (1980) Arch. Biochem. Biophys. 203, 753-757. 10. Conn, E. E. (1980) Annu. Rev. Plant Physiol. 31, 433-451. 11. Conn, E. E., McFarlane, I. J., Moller, B. L. & Shimada, M. (1979) FEBS Fed. Eur. Biochem. Soc. Meet. 55, 63-71. 12. Halkier, B. A. & Moller, B. L. (1991) J. Biol. Chem., in press. 13. Hosel, W., Berlin, J., Hanzlik, T. N. & Conn, E. E. (1985) Planta 166, 176-181. 14. Moller, B. L., McFarlane, I. J. & Conn, E. E. (1977) Acta Chem. Scand. Ser. B 31, 343-344. 15. Moller, B. L. (1978) J. Labelled Compd. Radiopharm. 14, 663-671. 16. Halkier, B. A., Olsen, C. E. & Marcussen, J. (1991) J. Labelled Compd. Radiopharm., in press. 17. M0oler, B. L. & Conn, E. E. (1980) J. Biol. Chem. 255, 3049-3056. 18. Pedersen, K. J. (1934) J. Phys. Chem. 38, 559-571. 19. Finkbeiner, H. L. & Stiles, M. (1963) J. Am. Chem. Soc. 85, 616-622. 20. Nielsen, A. T. (1969) in The Chemistry of the Nitro and Nitroso Groups, ed. Feuer, H. (Wiley, New York), pp. 349-486. 21. Larsen, P. 0. (1981) in The Biochemistry of Plants, A Comprehensive Treatise, eds. Stumpf, P. K. & Conn, E. E. (Academic, New York), Vol. 7, pp. 501-525. 22. Moller, B. L. (1981) in Cyanide in Biology, eds. Vennesland, B., Conn, E. E., Knowles, C. J., Westley, J. & Wissing, F. (Academic, New York), pp. 197-215. 23. Halkier, B. A., Scheller, H. V. & Moller, B. L. (1988) CIBA Found. Symp. 140, 49-66. 24. Conn, E. E. (1988) Am. Chem. Soc. Symp. Ser. 380,143-154. 25. Ettlinger, M. & Kj r, A. (1968) Recent Adv. Phytochem. 1, 59-144. 26. Baxter, R. L., Hamley, A. B. & Chan, H. W.-S. (1988) J. Chem. Soc. Chem. Commun., 757-758. 27. Baxter, R. L. & Greenwood, S. L. (1986) J. Chem. Soc. Chem. Commun., 175-176. 28. Kjor, A. (1981) Food Chem. 6, 223-234. 29. Matsuo, M., Kirkland, D. F. & Underhill, E. W. (1972) Phytochemistry 11, 697-701. 30. Tapper, B. A. & Butler, G. W. (1967) Arch. Biochem. Biophys. 120, 719-721. 31. Underhill, E. W. (1967) Eur. J. Biochem. 2, 61-63. 32. Underhill, E. W., Wetter, L. R. & Chisholm, M. D. (1973) Biochem. Soc. Symp. 38, 303-326. 33. Tapper, B. A. & Butler, G. W. (1972) Phytochemistry 11, 1041-1046.