Biotransformation of Digitoxigenin by Cochliobolus lunatus

A reação de biotransformação da digitoxigenina (1) por Cochliobolus lunatus foi investigada. Foram realizados experimentos com duração de 4 dias, que resultaram no isolamento de quatro produtos, cujas estruturas químicas foram elucidadas como sendo 1β-hidroxidigitoxigenina (2), 7β-hidroxidigitoxigenina (3), 8β-hidroxidigitoxigenina (4) e digitoxigenona (5). A obtenção desses produtos nas condições empregadas nunca foi anteriormente descrita. A produção da substância 4 em uma reação de biotransformação também é inédita.


Introduction
Enzymes are known to possess a wide substrate tolerance by keeping their exquisite catalytic properties with respect to chemo-, regio-and enantio-selectivity, playing an important role in biotransformations. 1 Biotransformation reactions can be accomplished at room temperature and in aqueous medium, presenting itself as a milder alternative to classical chemical reactions, [2][3][4] being employed for the resolution of racemates and to introduce chiral centers in substrates, among other uses. 5ungi are eukaryotic organisms that possess enzyme systems similar to those of mammalians.They usually present highly flexible metabolism, thus accepting varied sources of carbon and nitrogen.7][8] Such attributes suggest fungi as suitable organisms to perform biotransformation reactions.
The biotransformation of steroidal compounds by fungi has been extensively evaluated, including reactions with cardiac glycosides (Pádua et al. 9 and references herein; Table 3 of the present work).Digoxin, a Digitalis cardenolide, is still the drug of choice for the treatment of congestive heart failure, acting as a selective inhibitor of the Na + ,K + ATPase enzyme.Biotransformation of cardenolides has been investigated either as a strategy to obtain new derivatives or to convert the A-type cardenolides into the corresponding C-type compounds, which have clinical relevance. 9he main reactions obtained so far for cardenolide biotransformation were hydroxylation in different positions of the steroidal skeleton, oxidation, glycosylation, epimerization and esterification of the hydroxyl

Materials and Methods
General 1 H NMR, 13 C NMR, 1 H-1 H COSY and HMQC spectra were recorded on a Bruker DRX-400 spectrometer ( 1 H 400 MHz and 13 C 100 MHz) using TMS as internal standard for both nuclei.Chemical shifts (δ) are given in ppm and J couplings in Hertz (Hz).Optical rotations were measured with Perkin Elmer 341 polarimeter.

Results and Discussion
TLC analysis of the combined residues from digitoxigenin transformation (1) by C. lunatus, employing Kedde as spray reagent, showed spots with distinct Rf values of 1 solely for the biotransformation reaction, i.e., no product was observed for controls C1 and C2, as expected.
Chromatographic separation of the combined residues from biotransformation experiments of 1 resulted in the isolation of four compounds, along with the recovery of part of the starting material digitoxigenin (1).The structures of the products were defined based on spectroscopic analysis, using digitoxigenin (1) as model compound, and also by comparison with spectral data reported for the compounds or structurally related cardenolides.
The 1 H NMR spectrum of compound 2 showed a complex profile and the only signals easily assigned were those of H-3α (δ 4.13), H-21(δ 5.03 and 4.91), H-22 (δ 5.89), H-18 (δ 0.88) and H-19 (δ 1.09), thus confirming the presence of the C-3 hydroxyl group, the integrity of the α,β unsaturated lactone ring at C-17 and the methyl groups in the steroidal structure.The hydroxylation site was suggested by the paramagnetic shift observed for H-19 (δ 1.09) in comparison to digitoxigenin (1) (δ 0.96), pointing out the hydroxylation position close to C-19 (Table 1).The presence of a broad signal at δ 3.76 was also indicative of an additional hydroxyl group.
The HMQC spectrum obtained for 2 allowed determining 1 H/ 13 C one-bond shift correlations of all hydrogen-bearing carbon atoms in the compound.Correlation spots between C-1 / C-2 (δ 29.6 / 27.9) and their corresponding hydrogens H-1α and H-1β (δ 1.49 and 1.49) / H-2α and H-2β (δ 1.53 and 1.53) were clearly observed in the HMQC spectrum of model compound 1.On the other hand, those cross-peaks were absent in the equivalent region of HMQC spectrum of 2, indicating their shift resulting from hydroxylation at C-1 or C-2.
The COSY spectrum was helpful for assigning 1 H NMR chemical shifts of 2 and also disclosed the precise site of hydroxylation.Hence, the sign of a methine hydrogen at Further evidence of the hydroxylation position was given by comparing carbon chemical shifts obtained for compound 2 and those previously reported for 1β-hydroxy-17β-H-digitoxigenin. 14The values showed close correspondence, except for the chemical shifts of C-12, C-16, C-17 and C-18, what was expected result since 2 and 1β-hydroxy-17β-H-digitoxigenin are epimers at C-17.Furthermore, carbon chemical shits obtained for 2 were similar to those of digitoxigenin (1), with the exception of C-1 (δ 74.9), C-2 (δ 33.2), C-3 (δ 69.6), C-5 (δ 32.0) and C-19 (δ 19.6), which presented either paramagnetic or diamagnetic shifts, attributed to α-, βand γ-effects, resulting from the hydroxylation at C-1.
The stereochemistry of the hydroxylation site was indicated by the vicinal coupling constant of H-1 (J = 2.4 Hz), which pointed out the axial position for the C-1 hydroxyl group.7][18][19] NMR data previously reported showed good agreement with those of compound 2. As a result of the small amount of 2 isolated in the present work, 13 C NMR data were obtained indirectly by HMQC experiment.Therefore, it was not possible to attribute the resonances of non-hydrogenated carbons (Table 2). 13C NMR spectrum of compound 3 presented 23 signals, disclosed by DEPT-135 experiment as two methyl, nine methylene, seven methine and five non hydrogenated carbons.Compound 3 presents a methine signal at δ 71.3, not found in the spectrum of digitoxigenin (1), indicating that hydroxylation occurred at a methylene group.Besides, a signal with chemical shift typical of C-7 (δ 21.2) or C-11 (δ 21.3) in digitoxigenin (1) was absent in compound 3. Compounds 1 and 3 present equivalent chemical shift values for C-12 (Table 2); hence, it is very unlikely that hydroxylation had occurred at C-11, since β effect would result in diamagnetic shift of C-12.
Therefore, compound 3 is the hydroxylation product of digitoxigenin at C-7, what can be confirmed by the chemical shift (δ 3.90) and coupling constants of H-7 (J = 10.6, 10.4 and 4.9 Hz), consistent with two trans diaxial and one axialequatorial coupling.Altogether these data allowed identifying compound 3 as 7β-hydroxydigitoxigenin.As expected, 1 H NMR spectrum of 3 was similar to that of digitoxigenin (1), apart from the diamagnetic shifts observed for H-7α (δ 3.90), H-6α (δ 1.49), H-6β (δ 1.90) and H-8β (δ 1.71), resultant from the vicinity of the 7β-hydroxyl group (Table 1).Assignment of these hydrogens was confirmed by data of the HMQC spectrum.
Further confirmation of the hydroxylation site at C-7 was furnished by analysis of the COSY spectrum, which showed cross-peaks between H-7α (δ 3.90) both with 6-CH 2 (δ 1.90 and 1.49) and C-8 methine proton (δ 1.71). 13C and DEPT-135 NMR spectra of 4 revealed two methyl, ten methylene, five methine and six non hydrogenated carbons.The signal at δ 77.7 in the spectrum of 1 disappeared in the DEPT-135 spectrum of 4, suggesting the occurrence of hydroxylation at a methine group.Comparison of 13 C NMR spectrum obtained for compound 4 and digitoxigenin (1) showed that the signal attributed to C-8 (δ 41.8) in the later, was absent in the first.Consequently, compound 4 was identified as 8β-hydroxydigitoxigenin.As expected, 1 H NMR spectra of 4 and digitoxigenin (1)   8β-hydroxyl group.Deshielding induced by this group imposed diamagnetic shifts to H-18 and H-19, due to 1,4 effect. 20he COSY spectrum of 4 did not provide additional information on the hydroxylation site, since it occurred in a methine group.On the other hand, comparison of HMQC spectra recorded for 1 and 4 clearly indicated, in the last, the absence of a cross-peak between C-8 (δ 41.8) and H-8β (δ 1.56) found in digitoxigenin (1).Comparison between NMR data assignments carried out for compound 4 and literature records for its glycoside 21 has confirmed the structure of 4 as 8β-hydroxydigitoxigenin, also named cerdollagenin.
The major differences between 13 C NMR spectra of compound 5 and digitoxigenin (1) were the presence of a carbonyl carbon signal (δ 212.3) and the absence of the C-3 carbinol carbon signal (δ 66.8) in the first.In total, these data strongly suggest that compound 5 is digitoxigenone, a product from digitoxigenin oxidation at C-3.Aside from the diamagnetic shifts observed for C-1 (δ 39.8), C-2 (δ 37.1), C-4 (δ 42.1) and C-5 (δ 43.6) in compound 5, as a result from deshielding effect of the carbonyl at C-3, 13 C NMR chemical shifts observed for 5  and digitoxigenin were similar.Assignments of the above mentioned carbons were attested by HMQC spectrum, whereas COSY data allowed confirming the attribution of some hydrogens.Hence, H-2β (δ 2.34 J = 14.4,14.4 and 5.5 Hz) showed correlation spots with 1-CH 2 (δ 1.59 and 1.46) and H-2α (δ 2.21), while H-4β (δ 2.62 J = 14.2, and 14.2 Hz) showed cross-peaks with 5-CH (δ 1.84) and H-4α (δ 2.13).Definitive confirmation of the structure of 5 was given by comparison with NMR data previously reported for digitoxigenone. 22he biotransformation of digitoxigenin by C. lunatus is here reported for the first time and it afforded hydroxylated products at positions 1β, 7β and 8β.These hydroxylation sites are distinct from those previously described for Δ 4-5 steroids in reactions with the same fungus, that usually occur at the positions 11β, 14α and 7α. 7,23,24uch differences may be explained by distinct enzyme/ substrate interactions, arising from the uncommon configuration of digitoxigenin steroidal frame (cis, trans, cis), in comparison to other steroids, or due to the presence of the α,β-unsaturated lactone ring at C-17.Such hypotheses consider that the same hydroxylases do participate in reactions with digitoxigenin and other steroids.In this sense, Nozaki et al. 25 improved the 7β hydroxylation of digitoxigenin by Absidia coerulea, Rhizopus oryzae and Rhizopus stolonifer after pre-incubation with progesterone and deoxycorticosterone.Such result demonstrates that monooxygenases induced by Δ 4-5 steroids are also capable of hydroxylating digitoxigenin.
Hydroxylation of digitoxigenin at positions 1β and 7β, as well as oxidation of its hydroxyl group at C-3, have been previously reported for plant cell cultures and fungi (Table 3).It should be stressed, however, that this is the first report on compounds 2, 3 and 5 as biotransformation products of digitoxigenin by C. lunatus.Hydroxylation at 1β-position is of special interest considering that some glycosides of 1β-hydroxydigitoxigenin have been reported to exhibit potent in vitro activity against ovarian adenocarcinoma and lung carcinoma. 15,16 ence, such reaction may be employed for the future production of new bioactive cardenolide derivatives.
The 8β-hydroxylation of digitoxigenin employing a cell culture, the fungus C. lunatus, is here described for the first time (Table 3).Some cardiac glycosides hydroxylated at 8β-position have been isolated from the plant species Nerium oleander, Cerbera manghas and Cerbera odalamm. 21The 8β-hydroxydigitoxigenin obtained in the present work is clear evidence that C. lunatus hydroxylases involved in the reaction are not affected by the steric hindrance of the 14β-OH group.Therefore, it is feasible to obtain a product with two vicinal hydroxyls at 8β and 14β-positions.Such reaction may present several synthetic applications in the future and can also lead to new bioactive cardenolides and steroid derivatives.
As a future perspective, the conditions for this biotransformation reaction have to be optimized to overcome the obstacles and to allow its application in large scale: the low aqueous solubility of digitoxigenin, which results in limited substrate accessibility to the biocatalyst, and toxicity of both substrate and product against fungus culture. 26The use of surfactants and water-miscible or immiscible solvents is suggested by several authors as a strategy to diminish these difficulties. 26