Biosynthesis of the trichothecene 3-acetyldeoxynivalenol. Is isotrichodermin a biosynthetic precursor?

3-Acetyldeoxynivalenol is the major trichothecene produced by the fungus Fusarium culmorum. The first proven tricyclic intermediate in the biosynthesis of 3-acetyldeoxynivalenol has been shown by in vivo studies to be isotrichodermin, a natural metabolite of F. culmorum. Indeed, the feeding of ring-deuterated isotrichodermin resulted in ring-deuterated 3-acetyldeoxynivalenol as shown by NMR studies. In this work, we have shown that the 3-acetyl group of isotrichodermin is mostly lost in its metabolism to 3-acetyldeoxynivalenol. We have shown by two different approaches that the deacetylation occurs at an early step after the first oxygenation step at C-15. Derivatives of isotrichodermin lacking the 3-acetyl such as 3-deacetyl isotrichodermin or 3-oxo-12,13-epoxytrichothec-9-ene are not precursors to 3-acetyldeoxynivalenol. The role of this acetyl exchange mechanism is not clear presently.

3-Acetyldeoxynivalenol is the major trichothecene produced by the fungus Fusarium culmorum. The first proven tricyclic intermediate in the biosynthesis of 3-acetyldeoxynivalenol has been shown by in vivo studies to be isotrichodermin, a natural metabolite of F. culmorum. Indeed, the feeding of ring-deuterated isotrichodermin resulted in ring-deuterated 3-acetyldeoxynivalenol as shown by NMR studies.
In this work, we have shown that the 3-acetyl group of isotrichodermin is mostly lost in its metabolism to 3-acetyldeoxynivalenol. We have shown by two different approaches that the deacetylation occurs at an early step after the first oxygenation step at C-15. Derivatives of isotrichodermin lacking the 3-acetyl such as 3-deacetyl isotrichodermin or 3-oxo-12,13-epoxytrichothec-9-ene are not precursors to 3-acetyldeoxynivalenol. The role of this acetyl exchange mechanism is not clear presently.
We have recently investigated the fate of the 3-acetyl group in isotrichodermin conversion to 3-acetyldeoxynivalenol. It has always been assumed without any proof that the 3-acetyl in 3-acetyldeoxynivalenol originates from the 3-acetyl in isotrichodermin. This was based on the seemingly textbook definition of a biosynthetic precursor: it was produced by F. culmorum cultures (4), it was 27% incorporated into 3-acetyldeoxynivalenol, and the incorporation site was rigorously determined by 2 H NMR of ring-deuterated isotrichodermin feedings (3). In order to ensure that we could use the easily synthesized [1Ј-14 C]acetyl isotrichodermin as a marker to isolate enzymes, we decided to confirm that indeed the 3-acetyl in isotrichodermin is retained in its conversion to 3-acetyldeoxynivalenol. We obtained surprising results, which will be discussed here.

EXPERIMENTAL PROCEDURES
Instrumentation-High performance liquid chromatography (HPLC) was performed on a Perkin-Elmer series 3B instrument coupled to a LC-75 variable wavelength detector (Perkin-Elmer) set at 204 nm and a Berthold LB 505 HPLC radioactivity monitor (Labserco, Oakville, Ontario). Thin layer chromatography separations using LHP-KF thin layer chromatography plates (Whatman) were analyzed with a Bioscan Imaging Scanner System 200 (Bioscan, Inc., Washington, D.C.). A Tracor Analytic Delta 300 instrument was used for liquid scintillation counting. Homogenization of F. culmorum cells was accomplished with a Bead-Beater (Biospec Products, Bartlesville, OK). Centrifugation was done using a Dupont-Sorvall 5C centrifuge with an SS-34 rotor and a Beckman Instruments L5-75 ultracentrifuge with a 60 Ti rotor. A Perkin-Elmer model 559A uv-visible spectrophotometer equipped with a digital temperature controller was used for all spectrophotometric measurements. Proton NMR spectra were run on JEOL CPF-270 (270 MHz), while the deuterium NMR spectra were run on XL-300 operating at 46 MHz. The samples (2-5 mg) were dissolved in CDCl 3 (in CHCl 3 for deuterium NMR), and their spectra were recorded at ambient temperature (22°C). For routine 1 H NMR, a 45°pulse was used with an acquisition time of 4 s (no relaxation delay). The data were zero-filled, and resolution enhancement techniques were used to help resolve small couplings. The two-dimensional matrix was 1024 ϫ 1024 data points after processing. The data were pseudoecho-shaped and symmetrized prior to plotting. The 2 H NMR spectra were acquired using 80°pulse and 1-s acquisition times. The data were zero-filled and processed with 1-Hz line broadening. The solvent (CHCl 3 ), used as internal reference, was set at 7.26 ppm.
Strain and Cultivation Conditions-F. culmorum strain HLX 1503 was grown as described previously (9). The seed cultures (50 ml of seed medium in a 250-ml Erlenmeyer flask) were incubated for 3 days. The production cultures (25 ml of production medium in a 125-ml Erlenmeyer flask) were incubated for 2 days.
High Performance Liquid Chromatography-Analytical HPLC was performed with two Whatman partisil 10 ODS-2 analytical columns in series (4.6 ϫ 500 mm). Semipreparative HPLC was performed with two Whatman partisil 10 ODS-2 Mag 9 semipreparative columns in series (9.4 ϫ 500 mm). Preparative HPLC was performed with one partisil 10 ODS-2 MAG-20 preparative column (22 ϫ 500 mm). Program 1 consisted of a linear gradient that lasted for 50 min with an initial concentration of 15% methanol, 85% water and a final concentration of 75% methanol, 25% water, which was then held for 40 min. All ultraviolet detectors were set at 204 nm. Program 2 consisted of a linear gradient lasting 50 min with an initial concentration of 15% methanol, 85% water and a final concentration of 75% methanol, 25% water, which was held for 30 min and then increased linearly for 10 min to 100% methanol. The concentration was held for 20 min at 100% methanol before equilibration to 15% methanol, 85% water. Program 3 consisted of a linear gradient from 15% methanol, 85% water to 75% methanol, 25% water over 25 min. This concentration of methanol was held for 20 min and then increased linearly for 5 min to 100% methanol and maintained for 20 min.
Purification of 3-ADN from the [1Ј-14 C]Acetyl ITD and [4,8, C]ITD Feeding (Fig. 2)-The extract was fractionated by analytical HPLC on program 1 at 1 ml/min, and the peak corresponding to 3-ADN (t R ϭ 44.6 min) was collected and further purified by analytical HPLC using 35% methanol, 65% water as solvent (t R ϭ 45.8 min) (1.33 ϫ 10 4 dpm). NMR analysis confirmed the structure to be that of 3-ADN. Two portions (5% each) of the 3-acetyldeoxynivalenol were subjected to HPLC using program 1, and the 3-ADN collected was analyzed for radioactivity. The average radioactivity obtained of the two injections of 3-ADN was 524 Ϯ 13 dpm. In addition, two portions (5% each) were each incubated in 50 l of methanol and 100 l of 0.1 N NaOH for 17 h at 25°C to hydrolyze the ester at the 3-position of 3-ADN. Each of these samples was subjected to analytical HPLC using program 1. All of the radioactivity corresponded to the deoxynivalenol (DON) peak produced by hydrolysis (t R ϭ 29.7 min). The average radioactivity obtained of the two injections of DON was 376 Ϯ 14 dpm. The ratio of the radioactivity of 3-ADN before hydrolysis to the radioactivity of 3-ADN after hydrolysis was 1.39:1. Hydrolysis
a The letters s, d, t, q, and m following the signals stand for singlet, doublet, triplet, quartet, and multiplet signals. br and o show broad or overlapping lines.
b Two deuterium signals appear in the deuterium spectrum, one at 1.99 ppm or 2.21 ppm for D-4 (integrating for one deuterium) and one signal at 0.74 ppm or 0.78 ppm for D-15A and -B protons (integrating for one deuterium together) for 1 and 3, respectively.

FIG. 1. Tricyclic biosynthetic precursors to 3-acetyldeoxynivalenol and to sambucinol, the major metabolites produced by F. culmorum.
dpm for one injection and 597 dpm for the second injection. Two aliquots of the mixture were hydrolyzed. The radioactivity collected under the deacyl ITD region was 567 dpm for the first injection and 164 dpm for the second injection. From these data the average ratio of 3.5:1 for radioactivity in [ 14 C]acetyl ITD to radioactivity in [4,8,14- (2) and its unlabeled counterpart, ITD The shaded areas represent the deuterated positions.
a The letters s, d, t, q, and m following the signals stand for singlet, doublet, triplet, quartet, and multiplet signals. br and o show broad or overlapping lines b Five deuterium signals appear in the deuterium spectrum, one at 2.11 ppm for D-4 (integrating for one deuterium), one signal at 0.83 ppm for D-15A and -B protons (integrating for one deuterium together), and one at 2.11 ppm for D-2Ј (integrating for three deuteriums).
The samples from each individual feeding experiment were pooled, filtered, saturated with NaCl, and extracted with ethyl acetate. The extracts were dried over magnesium sulfate and then evaporated to dryness in vacuo. The extracts were then dissolved in methanol for fractionation by semipreparative HPLC using program 2 at 3 ml/min.
The fractions corresponding to 3-ADN from the individual feeding experiments were further purified using two ODS-2 analytical columns at 1 ml/min with 35% methanol, 65% water (t R ϭ 45.6 min). The NMR of 3-ADN (4, Fig. 4) was then recorded (Table III).
The fractions corresponding to SOL were further purified using two ODS-2 analytical columns at 1 ml/min with 50% methanol, 50% water (t R ϭ 38.3 min). The purified SOL was acetylated by incubating at 25°C for 16 h with 90 l of acetic anhydride and 60 l of pyridine. The diacetylated SOL was purified using two ODS-2 analytical columns at 1 ml/min with 65% methanol, 35% water (t R ϭ 40.3 min). The 2 H NMR of the diacetylated SOL was taken and showed no deuterium incorporation.
From the feeding of a large amount of 2 (10 mg) (Fig. 4), no peaks corresponding to 3-ADN or SOL were found. On the other hand, two peaks with t R in the range of calonectrin and isotrichodermin were detected. The peak corresponding to calonectrin was isolated and purified on analytical HPLC (t R ϭ 58.4 min) at 1 ml/min with 55% methanol, 45% water. The conditions for the purification of isotrichodermin (t R ϭ 47.5 min) were as follows: analytical HPLC using 65% methanol, 35% water at 1 ml/min. Their NMR agreed with calonectrin (5) and iso-trichodermin (6) (Fig. 4, Table IV).
From the feeding of a large quantity of 1 (5 mg), no peaks corresponding to 3-ADN or SOL were found. The only compound recovered from that feeding was isotrichodermin, which was purified using the same conditions outlined before (t R ϭ 47.2 min; analytical HPLC using 65% methanol, 35% water at 1 ml/min). Its NMR corresponds to compound 7, [4␤,15-2 H]ITD (Table IV).
Preparation of Subcellular Fractions of F. culmorum-Production cultures of F. culmorum, which were 48 h old were harvested by gentle filtration through Miracloth (9). The preparation of microsomes (5) was done by an adaptation of a published procedure (11). The microsomal fraction was either used immediately or stored at Ϫ80°C in 1-ml aliquots.
Conversion of [2Ј-2 H 3 , 4␤,15-2 H]ITD (2) in Microsomes-To each of two 25-ml Erlenmeyer flasks were added 8 mg of [2Ј-2 H 3 , 4␤,15-2 H]ITD (2) dissolved in methanol. The methanol was evaporated with a stream of nitrogen. To each flask was added 0.160 ml of 1% Brij 35, 1.54 ml of 50 mM potassium phosphate buffer at pH 7.5 containing 0.1 M KCl and 30% glycerol, 1.15 ml of 20 mM NADPH, 1.15 ml of 20 mM NADH. The mixture was incubated with 4 ml of microsomes, at 220 rpm, 25°C for 18 h. The contents of each flask were added to two CE 1020 tubes and were extracted with ethyl acetate (7 ϫ 15 ml). The ethyl acetate extracts were pooled and rotavaporated. The residue was fractionated by semipreparative HPLC using program 3 at 3 ml/min. A peak with a retention time of 26.0 min was characterized by NMR and was identified as 15-deacetylcalonectrin (8, Fig. 5 and Table V).

Syntheses of Doubly Radiolabeled
Isotrichodermin-In order to take advantage of the great sensitivity of radiolabeled compounds we first decided to prepare isotrichodermin ring-labeled with 14 C and mix it with isotrichodermin with 14 C on the 1Ј-position of the acetyl at C-3 (Fig. 2). Since the radiolabeled compounds are used as tracers, the feeding of this mixture is equivalent to feeding doubly radiolabeled isotrichodermin (at the ring at C-4, C-8, and C-14 and at the 1Ј-acetyl). The ringlabeled compound at positions 4, 8, and 14 was derived from the feeding of (3R)-[2-14 C]mevalonate to production cultures of F. culmorum. The [1Ј-14 C]acetyl isotrichodermin was prepared by acetylating unlabeled deacetyl-ITD (3) with 14 C-acetic anhydride. A mixture of these two labeled isotrichodermins (made by mixing approximately 3 times the amount of radioactivity of the [1Ј-14 C]acetyl-ITD for one [4,8, C]ITD to account for the number of 14 Cs) leads to virtually doubly radiolabeled isotrichodermin at the ring (at positions 4, 8, and 14) and on the 1Ј-acetyl position. The accurate ratio of radioactivity in the doubly labeled isotrichodermin (ring-radiolabeled ϩ 1Ј-14 Cacetyl) versus the ring-radiolabeled isotrichodermin was rigorously determined to be 3.5:1. This value was obtained by averaging the radioactivity values obtained by injecting different aliquots of the mixture onto the HPLC before and after chemical hydrolysis of the ester at the 3-position. Before chemical hydrolysis, the counts are equal to counts from the doubly labeled ITD. After chemical hydrolysis, all of the 1Ј-14 C-acetyl radioactivity has been lost, and the radioactivity detected in deacetyl-ITD is only due to [4,8, Feeding of Doubly Radiolabeled Isotrichodermin to F. culmorum Cultures-The doubly radiolabeled isotrichodermin was fed to F. culmorum cultures, and the resulting 3-acetyldeoxynivalenol and sambucinol were isolated and purified as described for other substrates (5). The sambucinol isolated and purified showed no radioactivity, confirming our previous finding that isotrichodermin is not a biosynthetic precursor to sambucinol (3) (Fig. 2). The 3-ADN isolated and purified was radioactive as expected. In addition, the incorporation of isotrichodermin was also on the order of 30% as previously found (3,5). In order to understand the fate of the 3-acetyl of isotrichodermin in the conversion to 3-ADN, the following experiments were done: (i) the radioactivity of an aliquot of the pure 3-ADN derived from the doubly radiolabeled isotrichodermin feeding was measured and recorded; and (ii) an equivalent aliquot of that pure 3-ADN derived from the doubly radiolabeled isotrichodermin feeding was hydrolyzed. The ester at the 3-position of 3-ADN was hydrolyzed, the DON obtained was purified, and the DON radioactivity was counted. The ratio of radioactivity in 3-ADN before hydrolysis to the radioactivity in DON after hydrolysis was calculated and found to be 1.39:1. If the 3-acetyl of isotrichodermin had been converted to the 3-acetyl of 3-ADN with no physiological hydrolysis and reacylation at the 3-position, the ratio would have remained 3.5:1. Complete physiological hydrolysis would have given a ratio of 1:1. However, the result obtained indicates that ϳ89% of the acetyl groups initially at the 3-position of ITD were lost during its conversion to 3-ADN by F. culmorum cells (Fig. 2). This result was very reproducible with various aliquots of doubly radiolabeled ITD-derived 3-ADN, with the values varying from 84 to 89% loss of acetate ( Fig. 2).
Synthesis of Deuterated Isotrichodermin Derivatives (Fig.  3)-The result obtained with the doubly radiolabeled isotrichodermin was unexpected. We therefore decided to confirm it with isotrichodermin labeled with stable isotopes in the ring as well as the acetate in the 3-position. The method is less sensitive than radiolabeled experiments, but the locus of incorporation can be easily and accurately determined by 2 H NMR. In addition, if indeed the 3-acetyl in isotrichodermin is not retained in its conversion to 3-ADN, then two other isotrichodermin derivatives, 3-deacetyl ITD and 3-oxo-EPT, could be substrates. We therefore had to prepare them with deuterium labeling. Their syntheses were accomplished in the following manner: the [4␤,15-2 H]3-deacetyl ITD (1, Fig. 3) was prepared as described previously (3) and was subsequently used for the synthesis of [2Ј-2 H 3 , 4␤,15-2 H]ITD (2) and of [4␤,15-2 H]3-oxo-EPT (3, Fig. 3). In the first case, this necessitated an acetylation with [ 2 H 6 ]acetic anhydride and in the second case an oxidation of the 3-hydroxyl to a keto-group. After extensive purification and analyses by 1 H and 2 H NMR, the three deuterated compounds were ready to be fed to F. culmorum cultures (Tables I and II). Fig. 4)--We have previously shown that feeding an excess of substrate enables us to  (1), respectively The shaded areas represent the deuterated positions.
a The letters s, d, t, q, and m following the signals stand for singlet, doublet, triplet, quartet, and multiplet signals. br and o show broad or overlapping lines. b In the deuterium NMR of both calonectrin (5) and isotrichodermin (6) the acetyl (CD 3 -CO) (2.08 ppm or 2.10 ppm) and D-4 (2.10 ppm) overlap. However, the integration of the peaks reveals that some acetyl is still partially deuterated. Indeed, this is obtained from the relative height of D-4/CD 3 with respect to the Me-15.

FIG. 5. In vitro feeding of doubly deuterated isotrichodermin.
The 15-deacetylcalonectrin derived from that feeding has retained all the deuteriums. The shaded areas represent the deuterated positions. accumulate biosynthetic intermediates that are otherwise undetected or present in minute quantities (5). We therefore fed in duplicate three different amounts of [2Ј-2 H 3 , 4␤,15-2 H]ITD (2) to F. culmorum production cultures. The results are shown in Fig. 4. The size of the structures in Fig. 4 emphasizes the amounts fed. When small amounts of [2Ј-2 H 3 , 4␤,15-2 H]ITD (2) were fed, 3-ADN was obtained with deuteriums only in the ring-positions at C-4 and C-15 and absolutely no deuteriums in the 3-acetyl group (4, Fig. 4 and Table III). This confirms our preliminary results with the radiolabeled compounds. The discrepancy between 100% loss of the 3-acetyl in the deuterium study versus 89% loss in the radiolabeled investigation emphasizes the relative sensitivity of the methods. The feeding with radiolabeled substrates uses trace amounts, and the detection via radioactivity is so sensitive that very small incorporations (11%) can be detected. On the other hand, the inferior sensitivity of deuterium labeling is compensated with definite determination of the site of incorporation, with no danger of easy contamination.
When larger amounts of [2Ј-2 H 3 , 4␤,15-2 H]ITD (2) were fed, the production of the end product 3-ADN was inhibited, and one of its known (5) biosynthetic intermediates, calonectrin (5, Fig. 4 and Table IV), accumulated. It is interesting to note that in this case two-thirds of the calonectrin obtained from the feeding had no deuteriums in the 3-acetyl group, whereas onethird retained the three deuteriums ( Fig. 4 and Table IV). In addition, the isotrichodermin recovered from the feedings of excess substrate had the same distribution as the calonectrin; two-thirds had lost the deuteriums at C-3 (6, Fig. 4 and Table  IV). This last result suggests that the loss of deuteriums in the 3-acetyl of isotrichodermin occurs at the initial stages of its metabolism to 3-acetyldeoxynivalenol (the first oxidation step at C-15) ( Fig. 4 and Table IV (Fig. 4)-The almost complete loss of the 3-acetyl group in isotrichodermin in its in vivo conversion to 3-acetyldeoxynivalenol led us to consider two other possible substrates: 3-deacetyl isotrichodermin (with a hydroxyl at C-3) or the isotrichodermin derivative with a ketone at C-3 (3-oxo-EPT). They were synthesized with deuteriums at positions 4␤ and 15 (1 and 3, Fig. 3) and fed to F. culmorum. The first observation on these two feedings was that the compounds inhibited considerably the production of the end products: 3-acetyldeoxynivalenol and sambucinol. For example, when 5 mg of [4␤,15-2 H]deacetyl isotrichodermin (1, Fig. 3) were fed, the only metabolites detected were the recuperated starting material and [4␤,15-2 H]isotrichodermin (7 , Table IV). Nevertheless, enough 3-ADN and SOL were isolated from the 1-and 2.5-mg feedings. As expected in the feeding of both substrates, the sambucinol isolated contained no deuteriums. On the other hand, these deuterated isotrichodermin derivatives were incorporated into 3-ADN (4, Fig. 4), as could be easily observed from the deuterium NMR (Table III).
In Vitro Metabolism of [2Ј-2 H 3 , 4␤,15-2 H]ITD (2, Fig. 5)--In order to determine the fate of the 3-acetyl group of isotrichodermin in its cell-free metabolism, doubly deuterated isotrichodermin (in the ring and in the 3-acetyl group), [2Ј-2 H 3 , 4␤,15-2 H]ITD (2) was fed to a microsomal fraction of F. culmorum. After the incubation period, the extract was fractionated by HPLC (using program 3), and a major single peak was obtained (t R ϭ 26.0 min). This new peak was purified by analytical HPLC. Proton and deuterium NMR established rigorously this compound to be 15-deacetylcalonectrin deuterated at positions 4 and 15 and at the C-3-O-acetyl position (8, Fig. 5 and Table  V). Therefore, in the microsomes where isotrichodermin is only metabolized to 15-deacetylcalonectrin (8), the acetyl moiety is retained.

Is Isotrichodermin a Precursor to 3-Acetyldeoxynivalenol?-
Isotrichodermin was first detected as a metabolite of various Fusarium species including F. culmorum (4). Investigating the sequence of appearance of metabolites in F. culmorum by the kinetic pulse-labeling method (9) revealed that the disappearance of isotrichodermin with time coincided with the formation of 3-acetyldeoxynivalenol and had no effect on the biosynthesis of sambucinol (3). In addition, radiolabeled and deuterated isotrichodermin (ring-deuterated) were very good precursors (27-30% total incorporations) (3). The nonincorporation of the 3-acetyl group of isotrichodermin into the 3-acetyl of the product 3-acetyldeoxynivalenol (Figs. 2 and 4) might suggest that 3-deacetyl isotrichodermin is in fact the real biosynthetic intermediate. This compound was therefore synthesized with two deuteriums at C-4 and C-15 (1, Fig. 3). Our feeding experiments in this work showed that 1 inhibited the growth of the fungi and was incorporated into the product 3-acetyldeoxynivalenol but at a much lower extent than isotrichodermin. In addition, 3-deacetyl isotrichodermin was never isolated as a  (2) The shaded areas represent the deuterated positions. This table emphasizes the retention of all of the deuteriums that were present in the starting material. a The symbols s, d, t, and m following the signals stand for singlet, doublet, triplet, and multiplet signals. br and o show broad or overlapping lines.
b Five deuterium signals appear in the deuterium spectrum, one at 2.20 ppm for D-4 (integrating for one deuterium), one signal at 3.69/3.50 ppm for D-15A and -B protons (integrating for one deuterium together), and one at 2.10 ppm for D-2Ј (integrating for three deuteriums). natural product, whereas isotrichodermin is produced by F. culmorum (4). A second possibility for a biosynthetic intermediate close to isotrichodermin but with no acetyl at C-3 was 3-oxo-12,13-epoxytrichothec-9-ene. Similarly, the labeled 3-oxo-12,13-epoxytrichothec-9-ene (also not a natural metabolite) (3, Fig. 3), inhibited the growth of F. culmorum and was incorporated into 3-acetyldeoxynivalenol to a lower extent than isotrichodermin. Therefore, isotrichodermin must be the biosynthetic precursor, but on its metabolism to 3-acetyldeoxynivalenol it loses its acetate and is reacetylated de novo by an acetylase. The biosynthetic step where the initial deacetylation occurs has been determined by the feeding of excess doubly deuterated isotrichodermin (two deuteriums positioned on the ring and three on the 3-acetyl). As we have seen previously (5), when an excess of substrate is fed, the metabolic turnover is repressed, causing the enzymes to be saturated. This results in the accumulation of intermediates generally found in trace amounts. Therefore, when we fed an excess of doubly deuterated isotrichodermin (Fig. 4), the end product 3-acetyldeoxynivalenol was not detected. The starting material (6, Fig. 4) was recovered with two deuteriums on the ring and partly deuterated on the 3-acetyl. In addition, calonectrin, a known precursor of 3-acetyldeoxynivalenol (5), was also obtained with the same label distribution (5, Fig. 4): the deuteriums in the ring are retained, but two-thirds of the deuteriums on the 3-acetyl have been lost. We therefore may conclude that the deacetylation occurs at an early step, either just prior to the first oxygenation step at C-15 or thereafter. We therefore decided to reinvestigate this result with an in vitro system.
In Vivo and in Vitro Metabolism of ITD (Figs. 2, 4, and 5)-In vivo experiments have demonstrated that during metabolism the 3-acetyl of ITD is mostly lost in its metabolism to 3-acetyldeoxynivalenol (Figs. 2 and 4). Feeding with large amounts of precursor (Fig. 4) suggested that the deacetylation occurs after the first hydroxylation at C-15. Due to the importance of this result, we wanted to confirm it with a different experiment. We know that with microsomal preparations, the main product obtained is 15-deacetyl isotrichodermin (15-DAC) (6) with no traces of the end product, 3-acetyldeoxynivalenol. This is probably due to the extensive washing of the microsomes during their preparation, hence disappearance of the enzymes required for the subsequent steps between 15-DAC and 3-ADN. Since 15-DAC is the end product, we could investigate with this system the fate of the 3-acetyl in the metabolism of isotrichodermin. Feeding the doubly labeled isotrichodermin ([2Ј-2 H 3 , 4␤,15-2 H]ITD, indicated as 2 in Fig. 3) to a cell free preparation could demonstrate if the 3-deacetylation occurs prior to the 15-hydroxylation or after. The result shown in Fig. 5 demonstrates that in the 15-DAC derived from that feeding, the 3-acetyl of ITD is fully retained. Thus, this experiment with microsomal preparation confirms what was suggested by the in vivo experiments, namely that the loss of acetate occurs after the first oxygenation step of isotrichodermin. The loss of acetate is probably linked to the acetate turnover, i.e. it is hydrolyzed off and then is put back on again using ATP and coenzyme A to form acetyl-CoA, which can acetylate deacetylated intermediates in the cell. We have shown earlier that DON can be acetylated to 3-ADN by F. culmorum cells (9). We have seen that ITD is a better precursor than deacetylated ITD. Therefore, despite the existence of an acetyl exchange mechanism, the acetyl seems to have an essential role that is not clear presently.