Archives of Insect Biochemistry and Physiology 10:199-213 (1 989) Synthesis and Nuclear Magnetic Resonance Study of 3 - Dehydroecdysteroids Jean-Pierre Girault, Catherine Blais, Philippe Beydon, Christian Rolando, and Rene Lafont Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, Uniuersite' Rent? Descartes, CNRS UA 400 (J.-P.G.), Dkpartement de Biologie, Laboratoire de Biochimie et Physiologie de Dkueloppement, Ecole Normale Supe'rieure, CNRS UA 686, (C.B., P.B., R.L.), and Ecole Normale Supkrieure, Dkpartement de Chimie, CNRS UA 1110 (C.R.), Paris Several 3-dehydro- (or ~ - o x o - ecdysteroids ) have been prepared by enzymatic and/or chemical means. Methods for their purification using various highperformance liquid chromatography systems are described. Proton and carbon nuclear magnetic resonance analyses show that 3-dehydroecdysteroids when dissolved i n water or methanol (but not i n chloroform) present a temperature-dependent equilibrium between two forms. The possible structure of these two forms is discussed. Key words: ecdysteroids, proton NMR, carbon NMR INTRODUCTION The formation of 3-0x0- from 3P-hydroxy-ecdysteroidswas first demonstrated in insects by Karlson et al. [l](see Fig. 1 for structures of ecdysteroids). The structure of 3-dehydroecdysone was established from its IR* spectrum (presence of an additional saturated keto group) and from mass spectrometric and Acknowledgments: The authors wish to thank Dr. D.H.S. Horn (Melbourne, Australia) for a generous gift of 2-deoxy-20-hydroxyecdysone and for his continuous interest and valuable discussions and Dr. U. Kerb for a gift of 14-deoxyecdysone. They are grateful to Dr. F. Riera who performed the synthesis of 2-0x0-20-hydroxyecdysone. They are indebted to M. Andrianjafintrimo and D. Rude1 for their contribution to HPLC analyses, to Dr. N. Morin for MS analyses, and to V. Michon and K. du Penhoat for400 MHz NMR facilities. Received July29,1988; accepted December 2,1988. Address reprint requests to Rene Lafont, Ecole Normale Superieure, Departement de Biologie, Laboratoire de Biochime et Physiologiedu Developpement, CNRS UA 686,46 Rue d'Ulm, 75230 Paris, Cedex 05, France. *Abbreviations used: CllD = chemical ionisation/desorption; COSY = correlation spectroscopy; DEPT = distorsionless enhancement by polarization transfer; HPLC = high performance liquid chromatography; INEPT = insensitive nuclei enhancement by polarization transfer; IR = infra-red; MS = mass spectrometry; NMR = nuclear magnetic resonance; NP = normal phase; RP = reverse-phase; TFA = trifluoroacetic acid; TLC = thin-layer chromatography; 3D20E = 3-dehydro-20-hydroxyecdysone;20E = 20 hydroxyecdysone. 0 1989 Alan R. Liss, Inc. 200 Girault et al. Ecdysone 3-Dehydroecdysone 20- Hydroxyecdysone 2-Dehydro-20-hydroxyecdysone 3-Dehydro-20-hydroxyecdysone 2-Deoxy-20-hydroxyecdysone 3-Dehydro-2-deoxy-20-hydroxyecdysone 2-Deoxy-20-hydroxyecdysone 22-Acetate 20-Hydroxyecdysone 2,ZZ-diacetate 20-Hydroxyecdysone 322-diacetate 14-Deoxyecdysone 3-Dehydro-14-deoxyecdysone OH OH OH H H H =O OH H H H H H H H OAc H OH H OH H OH H OH H =O OH OH H H =O OH H =O OH H OH H OAc H OH H =O OH OH OH OH OH OH OH OH OH OH H H H H OH OH OH OH OH OH OH OH H H H H H H H H H Ac Ac Ac H H Fig. 1. Structures of the major compounds mentioned in this report. proton NMR (in C5D5N)analyses. A few years later, chemical methods were designed for the preparation of 3-0x0-ecdysteroids [2-41. In insects, 3-0x0-ecdysteroid formation is catalyzed by an ecdysone oxidase: the reaction products H202and is irreversible .Reduction back to the parent compounds is achieved through a 3-dehydroecdysonereductase 3P-forming [6,7]. Some tissues contain an additional 3-dehydroecdysone reductase 3aforming 171 which produces 3-epimers, a family of compounds firstly characterized in Munducu sextu . 3-Epimers are clearly inactivation products [8-lo], but the physiological significance of 3-0x0-ecdysteroids is still a matter of debate. Early experiments showed a reduced activity in several bioassays [3,11]. However, using a new bioassay, Somme-Martin et al.  found that 3-dehydro-20hydroxyecdysone was far more potent than 20-hydroxyecdysone to induce specific mRNA synthesis in Drosopkih larval fat body. In addition, the interest for 3-0x0-ecdysteroidshas suddenly increased with the recent demonstration that 3-dehydroecdysone (and not ecdysone as earlier believed) represents the major secretory product of M . sextu prothoracic glands cultured in vitro [13,14]. 3-0x0-ecdysteroids display a particular behavior during reverse-phase HPLC as they often elute as broad peaks, although they may elute as sharper peaks under acidic conditions .Moreover, proton NMR analyses provide evidence that they would be in fact mixtures of two compounds [13-151. It has been suggested (141 that these forms would correspond respectively to 3-0x0- and 2-0x0-ecdysteroids. Our data do not support this interpretation, and we present here an alternative explanation for the two forms in equilibrium. Studies on 3-Dehydroecdysteroids 201 MATERIALS AND METHODS Chemicals Ecdysone and 20-hydroxyecdysone were from Simes spa. (Milan, Italy). 2-Deoxy-20-hydroxyecdysone (from Blechnum minus) was a generous gift from D.H.S. Horn (Melbourne, Australia). 14-Deoxyecdysone was a generous gift from Dr. U. Kerb (ScheringA.G., Berlin, F.R.G.). All the reagents used were from Aldrich (gold label quality). Solvents (HPLC grade) were from BDH (Poole, England), Carlo Erba (Milan, Italy), or Prolabo (Paris, France). Chromatographic Procedures Ecdysteroids were separated and purified using HPLC. Normal-phase separations used Zorbax*-Sil (DuPont) columns (25 cm long, 4.6,9.4 or 20.2 mm i. d. according with sample size) eluted with methylene chloride/propan2-oYwater (125:25:2; 125:20:1.5, or 125:15:1) . Reverse-phase separationsused either a ZorbaxB-ODs, a Spherisorb-SODS2 (Phase Sep., Norwalk, CT) column (25 cm long, 4.6 or 9.4 mm i.d.), or a RCM C18 cartridge (10 cm long, 5 mm i.d.) (Waters Assoc., Milford, MA) eluted with one of the following solvent systems: methanol/water, acetonitrile/water, or acetonitrile/O.1%TFA (see "Results" section for further details). Spectrometric Analyses MS analyses used a Riber 10-10Bapparatus (Nermag S.A., Rueil-Malmaison France) equipped with a direct inlet probe. Spectra were recorded using the chemical ionization/ desorption mode with ammonia as the reagent gas. NMR analyses were made with a Bruker WM-250 (250MHz for 'H and 62.9 MHz for I3C)or AM-400 (400 MHz for 'H) apparatus (Bruker, F.R.G.). Samples were dissolved in CD30D, D20, or CDC13according to their solubility in these different solvents. Classical 1D and 2D experiments were performed as previously described 1171. Long-range 2D-COSY experiment at 400 MHz was performed using 0.08 s delay to enhance long-range couplings . SelectiveINEPT experiments 1191 were performed for the assignment of some carbon signals of 3-dehydro-20-hydroxyecdysone with a delay optimized for long-range 'H-I3C coupling constants (values ca. 5Hz) (accumulation time ca. 18 or 36 h corresponding to 25,000 scans for 1-He of the major component or 50,000 scans for 1-He of the minor component). Preparation of Cytosolic Extracts Cytosols from Pieris brassicae pupae were prepared as previously described  and stored at - 80°C until use. They were used to prepare enzymatically small amounts (ca. 100-200 kg) of 3-dehydroecdysone, 3-dehydro-20-hydroxyThe substrates (1-2 ecdysone, and 3-dehydro-2-deoxy-20-hydroxyecdysone. mg) were dissolved in 80 ml of cytosolic extract corresponding to ca. 40 g fresh weight of P. brassicae pupae. After incubation under constant agitation for 3 h at 30"C, samples were extracted with 1/2 volume chloroform, adsorbed onto C-18 Sep-pak@cartridges , then eluted with methanol and purified by normal phase HPLC (ZorbaxB-Sil, solvent dichloromethane/ propan-2-oIlwater 125:25:2). 202 Girault et al. Chemical Preparation of 3-Dehydroecdysone, 3-Dehydro-20-Hydroxyecdysone, and 3-Dehydro-14-Deoxyecdysone These were prepared according to Dinan and Rees . Ecdysone (or 20hydroxyecdysone) (20 mg) dissolved in hot distilled water (100 ml) was oxidized overnight at room temperature with oxygen in the presence of 20 mg reduced platinum as catalyst. After the reaction, the catalyst was removed by filtration, and the solution was evaporated to dryness; then the ecdysteroid mixture was purified by preparative NP-HPLC on a ZorbaxB-Sil column (20.2 mm i.d., 25 cm long) using a mixture of dichloromethane/propan-2-oYwater (125:40:3) as eluent at a flow-rate of 10 ml/min. The overall yield was ca. 50% for 3-dehydro-20-hydroxyecdysonebut it was significantly lower (ca. 20%) for 3-dehydroecdysone, because of the formation of larger amounts of the 3,22dioxo-derivative under our conditions. The same experiment was performed with 14-deoxyecdysone (0.5 mg) in 20 ml water with 1mg platinum. Chemical Preparation of 3-Dehydro-2-Deoxy-20-Hydroxyecdysone 22-Acetate 2-Deoxy-20-hydroxyecdysone(40mg) was first converted to its 22-acetate (yield ca. 25%) upon treatment with 1.5 ml acetic anhydride/pyridine (1:2) for 1.5 h at 20°C. The 22-acetate (10 mg) was purified by preparative NP-HPLC [column ZorbaxB-Sil: 25 cm long, 20.2 mm i.d., solvent dichloromethane/propan2-oYwater (125:40:3), flow-rate 10ml.min-’, retention volume 160ml] and thereafter oxidized with Jones’reagent according to Galbraith et al. . The resulting 3-oxo-2-deoxy-20-hydroxyecdysone 22-acetate (4.5 mg) was purified by preparative HPLC using the same system as above (retention volume 96 ml). Chemical Preparation of 3/2-Oxo-Diacetates Both 2-oxo-3,22-diacetate and 3-oxo-2,22-diacetate of 20-hydroxyecdysone were prepared according to Greenwood et al.  from the 3,22- and 2,22-diacetates. The latter were prepared from 20-hydroxyecdysone 2,3,22-triacetate upon partial deacetylation (5 min in 0.6% potassium carbonate in methanol/ water 90:lO) followed by preparative normal phase HPLC on a ZorbaxB-Sil column (9.4 mm i.d., 25 cm long) eluted at 4 mumin with dichloromethane/ propan-2-oYwater 125:20:1.5). Retention time was 11.8 min for the 3,22 diacetate and 15.7min for the 2,22 diacetate. Each diacetate was oxidized as described , and the corresponding 0x0-derivative was purified by reverse-phase HPLC on a ZorbaxB-ODS column (9.4 mm i.d., 25 cm long) eluted with a linear gradient of 16% to 70% acetonitrile in 0.1% TFA over 60 min. The 3-oxo-2,22diacetate eluted as a very broad peak at ca. 29.3 min, and the 2-oxo-3,22-diacetate eluted as a sharp peak at 34.0 min (3,22-diacetate24.0 min; 2,22-diacetate 27.2 min). Some of the 2-oxo-3,22-diacetate was further treated with potassium carbonate (see above) with the aim to obtain 2-oxo-20-hydroxyecdysone, which was purified by normal-phase HPLC; in fact this compound was obtained mainly in its 5a form (see “Results” section). RESULTS HPLC Behavior of 3-Dehydroecdysteroids 3-Dehydroecdysteroids can be separated from their parent compounds by using various HPLC systems (Table 1).Normal-phase systems (and also TLC Studies on 3-Dehydroecdysteroids 203 TABLE 1. HPLC Data for Ecdysteroid Analyses* Compound Ecdysone 3-Dehydroecd ysone 20-H ydrox yecd ysone 3-Dehydro-20-H ydrox yecd ysone System 1 16.5 6.4 25.2 8.5 Retention volume (ml) System 2 System 3 10.9 11.4 15.2 12.4 5.2 6.4 6.9 6.8 System 4 7.3 8.5 3.3 4.0 ‘System 1:Zorbax-SIL (250 mm long, 4.6 mm i.d.), dichloromethane/propan-Z-ouwater (125:25:2). System 2: Spherisorb 50DS2 (250 mm long, 4.6 mm i.d.), acetonitrile/O.l% trifluoroacetic acid in water (23:77). System 3: Spherisorb 50DS2 (250 mm long, 4.6 mm i.d.), methanovwater (50:50). System 4: RCM Novapak C18 cartridge (100mm long, 5 mm i.d.), methanouwater (45:55). plates) are very efficient to resolve these pairs of compounds, in fact in a much better way than does reversed phase . With RP-HPLC, it seems more advisable to use acetonitrile/O.l% TFA rather than methanollwater (see Fig. 2). Methanol-water mixtures provide very poor resolutions, and in some systems compound pairs comigrate. Spectroscopic Properties of 3-Dehydroecdysteroids:Evidence for Two Forms in Equilibrium The products obtained enzymatically or by synthetic methods are identical. This was established from the comparison of their HPLC and spectroscopic properties. NMR analyses of HPLC-purified 3D20E in D20or in CD30D provided an unexpected result (Table2). It consisted of a mixture of two different compounds. When freshly prepared 3D20E was quickly dissolved in CDC13, two components were also found (ratio 60:40), but within a few minutes only one compound remained. From its chemical shifts, this component could be related to the more abundant one in D 2 0 or in CD30D. 1 acetonitrile/O.l% TFA (23:77) 1 methanoliwater (50:SO) 2 Fig. 2. Separation of 3P-OH/3-oxo pairs using two different solvent systems based on acetonitrile or methanol. Column: Spherisorb-50DS2,25 cm long, 4.6 mm i.d., flow-rate 1 ml.min-’. 1:20-Hydroxyecdysone.2: 3-Dehydro-20-hydroxyecdysone. 3: Ecdysone. 4: 3-Dehydroecdysone. 204 Girault et al. 80°C 20°C 3 I‘ A! ”A” Fig. 3. Evidence for the presence of an equilibrium between two forms (“‘A’’) and (““’7 of 3-dehydro-20-hydroxyecdysonein D20from the two signals corresponding to the 7-H proton, respectively, at 6.04 ppm (‘‘A‘’) and 5.99 ppm (“6”). In D 2 0 or in CD30D these two components are in a thermodynamic equilibrium. This can be clearly established by ‘H- and I3C-NMR temperaturedependent experiments (Fig. 3, 4a,b). In D 2 0 at +20”C, the two components (hereafter called “A” for the major one and “B’ for the minor one) are present in the ratio “A“/“B‘ = 62:38, and this ratio becomes “A”/”B’ = 84:16 at + 80°C. the ratio values are, respectively, For 3-dehydro-2-deoxy-20-hydroxyecdysone, 84:16 ( +20”C)and 95:5 ( +80°C). Lowering the temperature back to 20°C led to the initial values in both cases. Changing the pH from 1.5 to 7 at 20°C in D 2 0 did not elicit variations in this ratio. To our mind, this equilibrium could be explained in different ways: 1. Equilibrium between 2-keto- and 3-keto-compounds, as recently proposed for 3-dehydroecdysone  2. Equilibrium between 3-0x0 5p- and 5a-compounds 3. Monomer-dimer equilibrium Studies on 3-Dehydroecdysteroids l 2lla ~ l !GI 1611 ~ 140 ' l ~ "I l 120 100 "I 88 '~ I 60 l' ' I 40 ' 'I 20 ' 205 "I B ppn Fig. 4. 13CNMRspectra of 3-dehydro-20-hydroxyecdysonein D20. a: Spectrum at +32"C (with 'H broad band decoupling). b: Selective INEPT experiment with 'H selective pulse at 2.0 ppm (I-He of "B" component). c: Same experiment as in b, but at 2.5 pprn (I-He of "A" form) d: Spectrum at +4"C (with 'H broad band decoupling). 4. Conformational equilibrium of 3-dehydro-compounds in hydrophilic solvents 5. Equilibrium between 3-keto and gem-diol in water (or hemiketal in methanol) Evidence Against 2-Dehydro f 3-Dehydro Equilibrium The first hypothesis (1)could be ruled out by a careful analysis of the 250 and 400 MHz 2D-COSY and 2D-COSY long-range NMR experiments on the mixture. Analysis of the A ring proton signals showed the following pattern for the two components present in the mixture: I-Hax 4-Hax The synthesis of 2-dehydro-20-hydroxyecdysone was performed according to . Analysis of its NMR spectrum in D 2 0 showed that only one compo- 206 Cirault et a[. TABLE 2. Proton NMR Data for Ecdysteroidst 20E D20 3-OX0 20E; D20,2O"C, pH7 "A" form "B' form 3-OX0 2D20E; D20r 20"c,pH7 2-oxo-20E ( 5 ~ ) ; "A" form "B' form D20, 2o"C, pH 7 1-Ha 1.38 1.55 1.3 1.6 1-He 1.88 2.5 2.0 2.15 2-Ha 3-He 4-Ha 4-Hb 5-H 7-H 9-Ha 11-Ha 11-He 17-H 22-H 23-Ha 23-Hb 18-Me 19-Me 21-Me 26-Me 27-Me 2.76 2.65,2.45 3.99 4.70 3.92 (m,wl12=22) (d,d 12.6,6.5) (d,d 12.1,4.6) ( ~ 1 1= 2 22) ( ~ 1 1= 2 18) 4.07 ( ~ ~ w= I 8) /z 1.75 2.75 1.7 (t, 14.1) 1.75 2.39 1.8 2.36 2.56 2.34 2.46 (t") (d,d14.1,4.1) (d,d-9.5,5) 5.97 6.04" 5.99" 5.99" (d2.5) 3.11 3.44" 3.14, 3.31" ( m , w 2= 22) 1.8 1.73 1.85 1.75 1.9 1.86 1.95 1.85 2.34(t) 2.33 2.33 3.43" 3.43 3.43" 3.43" (b,d,lO,wl,? = 14) 1.32 1.31 1.30 1.65 1.65 1.63 0.89" 0.87(s) 0.90" 0.87" 1.11" 1.OO( s) 1.08" 0.99* 1.22" 1.26" 1.26" 1.22(s) 1.23* 1.24(s) 1.24" 1.24" i.24isj 1.24* 1.24" 1.23" - 5.99" 3.2 " 4.55 (d,d I2,7,~1/2 = 20) 2.6 ( ~ 1 1= 2 28) 1.8 3.21 (d,d 12,3.8) 6.04 (d2) 3.07" 3.43" 1.75 1.85 2.31 3.43" 1.32 1.65 0.86" 0.96" 1.22" 1.23" 1.23" 1.3 1.6 0.86" 0.81" 1.22" 1.23" 1.23" tMultiplicity of signals: s = singlet; d = doublet; t = triplet; m = multiplet; b = broad signal. t" = triplet-like (4-Ha and 4-He isochronous); wII2= width at half-height in Hertz; 6 in ppm. "Same value as for 20-hydroxyecdysone. nent was present, and 2D-COSY spectra showed a pattern expected for a 2-dehydro-compound: 1 1 -Hax 1-Heq 1 4-Hax / " I 3-H\ 5-H 4-Heq In this case the compound was in 5a configuration: this was established from the large width of half-height of the 3-H signal ( w ~-20 , ~ Hz), which is t y p l for an axial position, a very large downfield shift of the 5-H signal, and a J long-range coupling of 19-Mewith 1-Hax and 5-Hax (only observed in the 5a configuration). Studies on 3-Dehydroecdysteroids 207 The mechanistic pathway for a 2-dehydro 2 3-dehydro equilibrium in vitro would follow the scheme: 2 3 -c-CH- 11 0 I OH ' .- zz C,C', / no OH d --C H-COH In D 2 0 this equilibrium would lead to a quick exchange by deuterium atoms of the 2-H in the 3-dehydro isomer and of the 3-H in the 2-dehydro compound. No such exchange could be observed at 20°C (pH 7) even after an extended period (>2 months) or at 80°C after several hours. This observation ruled out at the same time the possiblity of an epimerization of 2p-OH into 2a-OH of the 3-dehydrocompound, where we would observe an exchange of 2-H by deuterium as in 2-dehydro-/3-dehydroisomerization. Finally, this isomerization would not be possible in the case of 2-deoxycompounds, although NMR data provide evidence for the existence of an analogous equilibrium between two forms of D20(or in CD30D)for 3-dehydro20-hydroxyecdysone. Evidence Against 5a f Equilibrium This second hypothesis (2) could be eliminated on the following evidence: a. We did not observe any exchange of the 5-H proton in D20, as would be expected from the enolization of the 6-ketone needed to change from 5p to 5a forms; b. 5a compounds would show (in D20) a large upfield shift of the 9-H signal (ca. - 0.4 ppm) and a downfield shift of the 5-H signal (ca. 0.3 pprn); c. "J long-range couplings would be observed for 5a compounds in the 2D-COSY or in 2D-COSY long-range experiments between 19-Me, 1-Hax, and 5-Hax (see above); d. The I3C-NMR spectra of 5a compounds with respect to 5p ones would show large differences in the chemical shifts of the 19-Me signals: these are, respectively, at 12.3 and 24.3 pprn for 5a-and 5P-androstane, 11.4 and 22.6 ppm for 501- and 5p- pregran-3,20-dione [24,25]. The two carbon signals (in D20) for the 19-Me of compounds "A" and "B' are found at 24.06 and 25.05 ppm, respectively, and in the same ratio as obtained in 'H-NMR spectra. No signals are found below 19.52 ppm, a unique signal assigned to 18-Me of "A"and "B' (the multiplicity of signals was determined from DEPT 135experiments 1261). + Evidence Against a Monomer-Dimer Equilibrium The NMR experiments were performed with 3-oxo-20E from either biological or chemical origin, which were available in very different amounts. As a consequence its concentration in the NMR tube was, respectively, ca 0.4 and 40 mM. Both experiments gave the same "A"/"B' ratio, which would not be the case if it had been a monomer-dimer equilibrium. Nature and Conformation of "A" Form in Solution Examination of coupling constants of 2-H with 1-H (12.6, 6.5 Hz), of 5-H with 4-H (14.1, 4.1 Hz), of the long-range coupling between 2-Hax and 4- 208 Cirault et al. Hax , and of the variations of chemical shifts in components "A" show that all these data are in good agreement with a 3-keto-compoundsin a pseudochair conformation of ring A as has been established by Warren et al.  with NOE of their major component. This conformation is the only one observed in CDC13. The 2-H signal obtained with a freshly prepared sample appears as a d,d,d, and a new doublet signal is observed at 6 = 3.44 ppm, nearly in the middle of the 22-H signal. It can be assigned to the hydrogen of 2-OH. It could be suppressed by the addition of a few drops of D20 and shaking of the sample (or upon aging of the solution), and the 2-H signals thereafter appeared as a d,d. Decoupling experiments of 2-H showed the relationship of these two signals in the same way. These observations are in agreement with a hydroxyl group engaged in a tight intramolecular hydrogen bonding with the 3-carbonyl group. This strong hydrogen bonding and the low dielectric constant of CHC13with respect to D20 allows form "A' to be stabilized. In this form, the carbonyl groups at C-3 and C-6 are located relatively far away from each other. Nature and Conformation of "B" Form in Solution The structure of "B' form is more difficult to establish. Several features are found: 1.The 2-H signal is a doublet of doublet with a large coupling constant (12.1 Hz) and a smaller one (4.6 Hz); 2. The position of 2-H signal of "B' (6 = 3.92 ppm) is not very different from that of the 2-H of 20-hydroxyecdysone (6 = 3.99 pprn), but it is very different from the 2-H signal of "A" form (6 = 4.70 pprn). 3. The 5-H signal of "B' is a doublet of doublet (ca. 9.1, 5 Hz), and it is not shifted downfield as in the "A" form ( + 0.2 ppm with respect to 20-hydroxyecdysone); 4. The 4-H signals (6 = 1.7 and 1.8 ppm) in "B' component do not strongly differ from the ones of 20-hydroxyecdysone (6 = 1.75pprn), whereas those of "A' component are found much downfield (6 = 2.75 and 2.39 pprn); 5. One notes major differences in 13C chemical shifts (Table 3) between "A" and "B' forms for the A ring signals and smaller ones for B ring signals; no significant differences are observed for the other signals. 13Cchemical shift variations (Table 3) are indicative of modifications of the conformation of A ring. 'J coupling constants of 2-H with the two l-H and of 5-H with the two 4-H would be in agreement with a twist-boat conformation of A ring (Fig. 5) where the 3-carbonyl group and the 4-methylene group have moved down and up, respectively, with regard to the "A" pseudo-chair conformation. The chemical shift difference between 2-H in "A' and "B' is in agreement with the above position of 2-H ("A") with respect to the adjacent carbonyl group and for a position of 2-H ("B') nearly in the plane of the carbonyl. All these points are also in good agreement with the NOE effect observed by Warren et al.  for their "2-dehydro-compound." Inspection of spacefilling models (C.P.K.)shows that in "B' form 2-H is close to 9-H proton as in "A' conformer, but 4-Ha is far away from 2-H (Fig. 5) (in the "A" form, 4-Ha is Studies on 3-Dehydroecdysteroids 209 TABLE 3. 13CNMR Data of 3-Dehydro-20-Hydroxyecdysonein D20 Carbon no. Multiplicity$ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 CH2 CH C CH2 CH C CH C CH C CH2 CH2 C C CH2 CH2 CH CH3 CH3 C CH3 CH CH2 CH2 C CH3 CH3 Shift value (ppm) AS (A-B) (ppm) 40.92 (A) 38.61 (B) 73.89 (A) 72.89 (B) 96.38 (B) 213.67 (A) 46.34 (A) 41.51 (B) 56.60 (B) 59.60 (A) 208.99 (B) 205.97 (A) 123.53(B) 123.07 (A) 170.94 (B) 171.14 (A) 37.16 (A) 36.57 (B) 40.92 (A)* 40.62 (B) 22.76 33.40 49.90 (A) 49.82 (B) 87.49 32.77 22.58 51.63 19.52 24.06 (A) 25.06 (B) 80.38 22.12 79.75 28.46 43.01 74.07 29.97 30.57 + 2.31 + 1.00 + 117.29 + 4.83 + 3.00 - + + + - 3.02 0.46 0.20 0.59 0.30 0 0 0.08 0 0 0 0 0 1.00 0 0 0 0 0 0 0 0 $Multiplicityfrom DEPT 135experiments . *Signals overlap. A form B form - scheme 2 H I H 0 \ H H Fig. 5. Proposed three-dimensional representation of “A” and “6” forms of 3-dehydro-20hydroxyecdysone in water or methanol. A: Psuedo-chair form. 6: Scheme I-twist-boat form; scheme 2-gem-diol form. 210 Girault et al. close to 9-H as is 2-H). In the "B' conformation space-filling models show also a close proximity of 14-OH and 3-0x0 group. This led us to consider whether hydrogen bonding between those groups might be a driving force for "B' conformation. This hypothesis was tested by preparing 3-0x0-14-deoxyecydsone: accordingly, this compound might present only one form in D20, or at least a reduced percentage of "B" form. This was, however, not the case: the same temperature-dependent equilibrium between "A" and "B' form was also found with 3-0x0-14-deoxyecdysone and 3-0x0-ecdysone (data not shown), thus excluding the involvement of such a hydrogen bondin in the stabilization of form "B." Moreover, a more careful examination of C spectra with higher signal-to-noise ratios allowed us to exclude this twist-boat structure and to propose another explanation, which seems to fit better with our spectral data. F Evidence for a Gem-Diol Structure of "B" Form The I3C spectra showed the presence of a quaternary signal at 6 96.38 ppm, which was difficult to assign because of its unusual position: this chemical shift value is found for instance for ketals >C(OR)2 . The intensity and temperature-dependent variations of the intensity of this signal allow us to conclude that it belongs to form "B." In order to assign this signal, we performed Selective INEPT experiments  with 'H selective pulses at values corresponding to 1-He and 2-Ha: this method was thought to provide useful information concerning neighbouring relationships between a given proton and J' and 'J-coupled carbons. Experiments performed with 2-Ha of "A" or "B' forms failed to give any information, perhaps because of null or very small value of J' constant between 2Ha and 3-C. By contrast, experiments performed on 1-He (6 = 2.50 ppm) and 1-He (6 = 2.00 pprn), respectively, for "Arand "B' forms led to selective polarization transfers of magnetization through long-range 'H-I3C coupling constants C3J) to quaternary I3C signals at, respectively, 6 = 213.67 ppm (3-Cof ''At' form) and 6 = 96.38 pprn (Fig. 4b,c). This result was in agreement with a very favourable dihedral angle 1He-lC-2C-3C of MOO, where J' coupling is maximal. It allowed us to assign the 96.38 ppm signal to 3-C of "B' form. Therefore, "B' form would correspond to the hydrated form = gem-diol (Fig. 5) of the 3-keto group in water (or hemiketal in methanol). It is not possible to have a full selectivity toward 1-He: 1-He signal overlaps with 5-H for "A," and with 12-Ha, 16-Ha, and 15-Ha for "B." As a consequence, several carbons were observed (Fig. 4b,c): for instance with the "A" form we see mainly (Fig. 4c) from 1-He carbons 2,3, and 10, and from 5-H carbons, we see mainly 1,3,6, and 9. In the gem-diol form, the presence of two 3-OH groups (one on each side of the steroid nucleus) would lead us to observe chemical shifts for protons of the p side of ring A (1-Ha, 4-He and 5-H) near of the values observed for 20-hydroxyecdysone. In the same way, for protons of the a side (1-He, 2-Ha, and 4-Ha) we would find chemical shift values close to the values observed for 3-epi-20-hydroxyecdysone . The observed data are in full agreement with this hypothesis. The upfield position of 2-Ha, 4-Ha, and 4-He in form "B' with respect to form '/A' are now perfectly understood because these protons are in the vicinity of a >C(OH)' and not of a >C = 0 group. Moreover, the chemical shift of the 9-H in form "B' is in good agreement with the one observed Studies on 3-Dehydroecdysteroids 21 1 for 3-epi-20-hydroxyecdysone, where no effect was due to the 3a-OH group. The same observation applies in the case of 2-deoxy- series (data not shown). In conclusion, the gem-diol hypothesis with a chair conformation of ring A (Fig. 5) seems to be the only one consistent with all spectrometric data. DISCUSSION Formation and Purification of 3-Dehydroecdysteroids Reference ecdysteroids can be easily prepared by either enzymatic or chemical means. The former allows preparation of amounts ranging from 0.1 to 1 mg, whereas chemical means allow the preparation of much larger amounts. These compounds can be easily purified by NP HPLC. MiUigram amounts of 3-0x0-ecdysteroidscan be kept for several months without degradation as methanolic solutions at - 20°C. On the other hand, the minute amounts of tritiated compounds prepared by enzymatic means appear unstable. The reasons for this instability are not yet clear, and the chemical structures of degradation products have not been established. Usual HPLC procedures display a very different ability to separate 3p - OH and 3-0x0 compounds. Normal phase systems (as does TLC-see [l])are very efficient, as oxidation at C-3 strongly modifies the interaction with silica (or with aminopropyl bonded silica ) and results in a much decreased retention volumes. On the other hand, this difference of polarity is no longer evident when using reverse-phase systems. Thus methanollwater systems are inefficient, and they allowed only limited separation (Fig. 2). With such a solvent system and another column, Dinan et al.  were unable to obtain any difference of retention time for these compounds. However, acetonitrile/O.1% TFA mixtures are more efficient and allow baseline separation (Fig. 2). This seems to be linked to both the use of acetonitrile and TFA, as a methanol/O.l% TFA mixture gives the same result as methanoywater. Acetonitrile/water gives poor peak shapes, often for both 3-OX0 and 3P-OH compounds. In a previous study, we obtained satisfactory results with acetonitrile/0.4% acetic acid mixtures at + 50°C 161, but it appears more convenient to use TFA at room temperature. In addition, TFA is volatile (b.p. = 72"C), and it can be removed easily with a rotary evaporator. It is recommended to evaporate at temperatures below + 35"C, because the use of higher temperatures might lead to some degradation of ecdysteroids (either by dehydration as a result of acidic conditions or even the formation of trifluoroacetates). Structure of 3-0x0-Ecdysteroids in Solution 3-Dehydro-20-hydroxyecdysone (in CDC13,C5D$J,and CD2C12)on one hand, 3-dehydro-2-deoxy-20-hydroxyecdysone, 3-dehydro-20-hydroxyecdysone 2,22diacetate, and 3-dehydro-2-deoxy-20-hydroxyecdysone 22-acetate (all three in CDC13) on the other hand are only observed in the "A" pseudo-chair conformation. In hydrophilic solvents (CD30D, D20), the two forms "A" and "B" are observed in thermodynamic equilibrium for 3-dehydro-20-hydroxyecdysone, 3-dehydro-2-deoxy-20-hydroxyecdysone, and 3-dehydro-2-deoxy-20-hydroxyecdysone 22-acetate (CD30D"A"/"B' = 60:40 at 20°C for the latter compound). 21 2 Cirault et al. The "A" form is the major one in all the cases we have checked, but the percentage of "B' increases at low temperature. Thus, for 3-0x0-20-hydroxyecdysone, the "A"/"B' ratio is of 5.25 at +8O"C, 1.63 at +20°C, and close to 1 at +4°C.This would correspond to a AH" value near - 17.7kJ/molfor the "A" - > "B' reaction. "B' form was interpreted thanks to NMR experiments as a hydrated form of the %OX0 group. Hydratation (or hemiketal formation) of the 3-0x0 group could of course only be observed in water (or methanol), in accordance with mass effect of the solvent. An -OH group at C-2 leads to increase the electrophilic character of the 3-OX0 group thanks to hydrogen bonding with this 2-OH group. Moreover, this 2-OH group would stabilize (again)by hydrogen bonding the gem-diol form. Increase of temperature leads to a decrease of the strength of this hydrogen bonding and therefore a decrease of "B' form as observed. By the same way, the lack of this 2-OH in 2-deoxy compounds explains why the "B'/"A'' ratios are lower in that case. The above equilibrium may also help to understand the HLPC data, which show a low polarity of 3-0x0 compounds by NP-HPLC and a polarity close to that of 3-OH compounds by RP-HPLC (Table 1):in organic solvents (NP)intramolecular hydrogen bonding would prevent interaction of solutes with the stationary phase, whereas the presence of hydrated forms would account for the polarity in aqueous solvents (RP). Finally it would be of interest to determine whether this equilibrium has any connection with the very high biological activity of this compound in the Drosophila bioassay . This would require another approach, e.g., receptorbinding experiments using tritiated %OX0 compounds. LITERATURE CITED 1. Karlson P, Bugany H, Dopp H, Hoyer G-A: 3-Dehydroecdysone,a metabolite of ecdysone in the blowfly Calliphoru eythrocephala Meigen. Hoppe Seylers Z Physiol Chem 353,1610 (1972). 2. Spindler K-D, Koolman J, Mosora F, Emmerich H: Catalytical oxidation of ecdysteroids to 3-dehydro products and their biological activities. J Insect Physiol23,441 (1977). 3. Koolman J, Spindler, K-D Enzymatic and chemical synthesis of 3-dehydroecdysterone, a metabolite of moulting hormone of insects. Hoppe Seylers Z Physiol Chem 358,1339 (1977). 4. Dinan L, Rees HH: Preparation of 3-epi-ecdysone and 3-epi-20-hydroxyecdysone. Steroids 32,629 (1978). 5. Koolman J: Ecdysone oxidase. Methods Enzymol212, 419 (1985). 6. Blais C, Lafont R: Ecdysteroid metabolism by soluble enzymes from an insect: Metabolic relationship between 3p-hydroxy-, 3ci-hydroxy- and 3-0x0-ecdysteroids. Hoppe Seylers Z Physiol Chem 365, 809 (1984). 7. Milner N, Rees HH: Involvement of 3-dehydroecdysone in the 3-epimerization of ecdysone. Biochem J 231,369 (1985). 8. Nigg HN, SvobodaJA, Thompson, MJ, Kaplanis JN, Dutky SR, Robbins WE: Ecdysone metabolism: Ecdysone dehydrogenase-isomerase. Lipids 9,971 (1974). 9. Thompson MJ, Kaplanis JN, Robbins WE, Dutky SR, Nigg HN: 3-epi-20-hydroxyecdysone from meconium of the tobacco hornworm. Steroids 24,359 (1974). 10. Kaplanis JN, Thompson MJ, Dutky SR, Robbins WE: The ecdysteroids from the tobacco hornworm during pupal-adult development five days after peak titer of molting hormone activity. Steroids 34,333 (1979). 11. Richards G: The relative biological activities of (Y and p-ecdysone and their 3-dehydro derivatives in the chromosome puffing assay. J Insect Physiol24,329 (1978). 12. Somm6-Martin G, Colardeau J, Lafont R: Induction of PI gene in Drosophila melanogaster:Reg- Studies on 3-Dehydroecdysteroids 21 3 ulation by various ecdysteroids. In: Abstracts of the VIII Ecdysone Workshop. Philipps Univ., Marburg, FRG, (1987). 13. Warren JT, Sakurai S, Rountree DB, Gilbert LI: Synthesis and secretion in uitro of ecdysteroids by the prothoracic glands of Munducu sexfu. In: Communication presented at the VIII Ecdysone Workshop. Philipps Univ., Marburg, FRG, (1987). 14. Warren JT, Sakurai S, Rountree DB, Gilbert LI, Lee SS, Nakanishi K: Regulation of the ecdysteroid titer of Munducu sextu: Reappraisal of the role of prothoracic glands. Proc Natl Acad Sci USA 85,958 (1988). 15. Girault JP, Lafont R: Identification of ecdysteroid metabolites isolated in the 50 pg-100 pg range with 1D and 2D 'H-NMR. In: Communication presented at the VIII Ecdysone Workshop. Philipps Univ., Marburg, FRG, (1987). 16. Lafont R, Somme-Martin G, Chambet J-C: Separation of ecdysteroids using high-pressure liquid chromatography on microparticulate supports. J Chromatogr 170,185 (1979). 17. Girault JC Lafont R: The complete 'H-NMR assignment of ecdysone and 20-hydroxyecdysone. J Insect Physiol34,701(1988). 18. Bax A, Freeman R: Investigation of complex networks of spin-spin coupling by two-dimensional NMR. J Magnetic Resonance 44,542 (1981). 19. Bax A: Structure determination and spectral assignment by pulsed polarization transfer via long-range 'H-I3C couplings. J Magnetic Resonance 57,314 (1984). 20. Lafont R, Pennetier J-L, Andrianjafintrimo M, Claret J, Modde J-F, Blais C: Sample processing for high-performance liquid chromatography of ecdysteroids. J Chromatogr236,529 (1982). 21. Galbraith MN, Horn DHS, Middleton EJ: Insect moulting hormones: The synthesis of possible precursors for metabolic studies. Aust J Chem 27, 1087 (1974). 22. Greenwood DR, Dinan LN, Rees HH: Mechanism of hydroxylation of C-2 during the biosynthesis of ecdysone in ovaries of the desert locust, Schistocerca greguria. Biochem J 227, 783 (19%). 23. Dinan LN, Donnahey PL, Rees HH, Goodwin TW: High-performance liquid chromatography of ecdysteroids and their 3-epi, 3-dehydro and 26-hydroxy derivatives. J Chromatogr 205,139 (1981). 24. Breitmaier E, Voelter W Carbon-13NMR Spectroscopy-High Resolution Methods and Applications in Organic Chemistry and Biochemistry, 3rd edition. VCH, New York, p. 338 (1987). 25. Blunt JW, Stothers JB: I3C N.m.r. spectra of steroids-a survey and commentary. Org Magn Res 9,439 (1977). 26. Dodderell DM, Pegg DT, Bendall MR: Distorsionless enhancement NMR signals by polarization transfer. J Magnetic Resonance 48,323 (1982). 27. Jackmann LM, Stemhell S Applications of Nuclear Resonance Spectroscopyin Organic Chemistry, 2nd edition. Pergamon Press, Oxford p 338 (1969). 28. Jackmann LM, Stemhell S Applications of Nuclear Resonance Spectroscopyin Organic Chemistry, 2nd edition. Pergamon Press, Oxford p 241 (1969). 29. Breitmaier E, Voelter W: Carbon-13NMR Spectroscopy-High Resolution Methods and Applications in Organic Chemistry and Biochemistry, 3rd edition. VCH, New York, p 220 (1987). 30. Beydon P, Girault JP, Lafont R: Ecdysone metabolism in Pieris brussicue during the feeding last larval instar. Arch Insect Biochem Physiol4, 139 (1987).