Vol. 57, No 1/2010

Biotransformation of deuterated-4'-O-methylnorbelladine into alkaloids galanthamine and lycorine in tissue cultures of Leucojum aestivum was demonstrated using HPLC coupled to mass spectrometry. GC-MS screening was also carried to investigate other native and deuterated alkaloids. A total of six labeled alkaloids were identified indicating that 4'-O-methyl-d(3)-norbelladine is incorporated into three different groups of Amaryllidaceae alkaloids that are biosynthesized by three modes of intramolecular oxidative phenol coupling.


INTRODUCTION
Amaryllidaceae alkaloids have important pharmacological properties such as acetylcholinesterase inhibitory activity, cytotoxicity and antitumoral activity (Bastida et al., 2006).Galanthamine, an isoquinoline alkaloid, is obtained on a commercial scale for its pharmacological interest from Narcissus spp.and Leucojum aestivum as well as synthetically (Guillou et al., 2001;Marco-Contelles et al., 2006).Lycorine, a pyrrolophenanthridine alkaloid, displays a strong antiviral effect against poliovirus, measles and Herpes simplex type 1 viruses, as well as high antiretroviral (Szlávik et al., 2004) and strong antimitotic activities (Kukhanova et al., 1983).
Nevertheless, the biosynthetic pathway of Amaryllidaceae alkaloids, particularly in Leucojum aestivum, has not been totally elucidated yet.Eichhorn et al. (1998) have established a revised scheme for the biosynthesis of galanthamine (Fig. 1, para-ortho′ oxidative coupling of 4'-O-methylnorbelladine).
The investigations on the biosynthesis of Amaryllidaceae alkaloids used 14 C-labeled 4′-O-methylnorbelladine injected into organs of daffodil plants or 13 C or 3 H 3 C-labeled 4'-O-methylnorbelladine applied to organs of field grown L. aestivum plants ( Barton et al., 1963;Eichhorn et al., 1998).
Here we report the development of a rapid and convenient method for the study of the biosynthetic pathway of Amaryllidaceae alkaloids using a method based on deuterium labeled precursor fed to in vitro cultures of L. aestivum.This paper reports for the first time the biotransformation of the common precursor 4′-O-methyl-d 3 -norbelladine in shoot cultures of L. aestivum.Mass spectrometry was used for the identification of the labeled alkaloids.HPLCMS was used for the analysis of deuterated galanthamine and lycorine in tissue cultures and in the liquid medium.GC-MS analyses are also reported for the screening and identification of other deuterated alkaloids.

Plant material.
Leaves isolated from Leucojum aestivum L. bulbs (from French local markets) chilled for 12 weeks at 5 °C were surface-sterilized in 70 % ethanol (1 min), then shaken for 15 min in 15 % Domestos (with sodium hypochlorite and sodium hydroxide content below 5 %; Unilever, Hungary) and rinsed three times with sterile water.Sterilized leaves were cut into thin slices (about 2 to 3 mm in length) and plated on culture medium.
LC-MS analysis.The LC consisted of a U3000-Dionex system, an injector with a 1 µl loop and a UV detector at 280 nm.The analytical column used was an Acclain PepMap C18 ID 1 mm column (150 mm × 3 µm × 100 µm) and was eluted at a flow rate of 40 µl/min using a gradient ranging from 0 % solvent B to 100 % solvent B in a time span of 36 min.Solvent A consisted of 97.5 % 10 mM ammoniumbicarbonate pH 7.8 with 2.5 % methanol and solvent B consisted of 97.5 % methanol and 2.5 % 10 mM ammoniumbicarbonate pH 7.8.The ESi-HRMS was a micrOTOF Q TM (Bruker Daltonics) apparatus.
GC-MS identification.Analyses were performed using QP2010 Shimadzu equipment operating in the EI mode at 70 eV.An AT-1 column (25 m × 0.32 mm × 0.30 µm) was employed with a 33 min temperature program of 80-280 °C at 10°C/min followed by a 10 min hold at 280 °C.The injector temperature was 280 °C, the flow rate of the carrier gas (helium) was 0.8 ml/min, the split ratio was 1 : 50.Identification of the alkaloids was performed by comparing the measured data with those of authentic compounds (galanthamine, lycorine) or with literature data as specified in the text.
Then, as previously described (Szewczyk et al., 1988) to get the non-labeled molecule, reductive amination with tyramine and NaBH 4 afforded the target molecule, which was isolated as a hydrochloride in 33 % yield.acetic acid (NAA) (10 µM) and benzylaminopurine (BAP) (5 µM) were subcultured in medium containing labeled precursor 4′-O-methyl-d 3 -norbelladine at various concentrations (0.05, 0.10 and 0.20 g/L) and incubated for various periods of time (15, 30 and 40 days).Whatever the concentration of the deuterated precursor used, the growth kinetics of the shoot cultures were similar, although the growth rates (final shoot culture fresh weight -inoculum fresh weight/inoculum fresh weight) of the treated shoot cultures were lower than those of the control cultures (Table 1).These results indicated that the deuterated precursor added in the culture medium could be toxic for the shoot cultures.The lowest concentration (0.05 g/L) of 4′-O-methyl-d 3 -norbelladine led, unexpectedly, to the slowest growth of the cultures.The medium concentration (0.10 g/L) also inhibited the growth, but the plant cultures showed the same growth rate as the control cultures after 40 days of incubation.This can be due to a correlation between shoot culture growth and the alkaloid synthesis in these tissues incubated with the labeled precursor.

Shoot cultures of
After harvesting the shoot cultures, the precursor and the other alkaloids were extracted from the tissues as well as from the medium and the compounds were identified using mass spectrometry.HPLC coupled with high-resolution mass spectrometry (ESI/QqTOF) was used in order to identify native galanthamine and lycorine, by comparison with authentic compounds (Figs.3a and 4a) as previously described (Ptak et al., 2009), and also deuterated galanthamine and lycorine (Figs.3b and 4b).These labeled alkaloids showed the same retention time as the native alkaloids, i.e. 33.4 min for galanthamine and 25.3 min for lycorine.Native galanthamine and deuterated galanthamine displayed [M + H] + at a m/z in accordance with the calculated values (288.1594 and 291.1783, respectively) (Fig. 3).Native lycorine and deuterated lycorine dis-      mentation patterns characteristic of the Amaryllidaceae alkaloids (Table 3) (Fig. 5).The identification of these alkaloids was performed by comparing the measured data with previously published results (Ptak et al., 2009) and with literature data (Berkov et al., 2005).Similar retention times and similar fragmentation pattern permitted us to detect unambiguously deuterated alkaloids among native ones.This way, six labeled alkaloids were identified: d 3demethylnarwedine, d 3 -demethylgalanthamine, d 3 -galanthamine, d 2 -lycorine, d 2 -crinine and d 3 -demethylmarithidine (Fig. 5, Table 3).As was shown for lycorine, two deuterium atoms were incorporated in the crinine structure, bearing the same methylenedioxy bridge.These results show that 4′-O-methyl-d 3 -norbelladine was incorporated into the three different groups of Amaryllidaceae alkaloids that are biosynthesized by three modes of phenol coupling on the common precursor (Fig. 1).All the labeled alkaloids were observed after 15 days of incubation with the labeled precursor except d 2 -lycorine that was detected only after 40 days of incubation as shown above by the LC-MS analysis.Concerning the alkaloids biosynthesized by the para-ortho′ oxidative coupling of 4′-O-methyl-norbelladine, labeled demethylnarwedine, demethylgalanthamine and galanthamine were detected while no labeled N-formylgalanthamine was observed.However, this alkaloid was detected in the extracts as a native compound.This could be due to a too short incubation time with the labeled precursor.Surprisingly, d 3 -demethylnarwedine was detected only in the liquid culture medium while d 3 -demethylgalanthamine was detected, initially, in the plant tissue only.All these results could be explained by the compartmentation in connection with enzymes that could be part of multienzyme complexes (Verpoorte et al., 1999).It has been reported that compartmentation plays a major role in the regulation of secondary metabolite pathways.For example, the biosynthesis of terpenoid indole alkaloids requires at least three compartments, the plastids, the cytosol and the vacuole (Verpoorte et al., 1997).The absence of native or labeled narwedine was noted.Concerning the alkaloids biosynthesized by the para-para′ oxidative coupling, labeled crinine and demethylmarithidine were detected.Both alkaloids were first observed after 15 days of incubation with the labeled precursor respectively at 0.10 and 0.05 g/L.It is noteworthy that these labeled compounds appeared simultaneously with d 3 -galanthamine suggesting that the competition between the three modes of phenol coupling on the common precursor O-methyl-norbella- played [M + H] + at a m/z in accordance with the calculated values (288.1230 and 290.1356, respectively) (Fig. 4).In control shoot cultures, native galanthamine and lycorine that exhibited only the pseudo molecular ion characteristic of standard compounds were detected both in plant tissues and in liquid medium indicating that these alkaloids were able to diffuse in the culture medium.Isoquinoline alkaloids from Papaver somniferum tissue cultures were also found in liquid-medium culture (Le Flem-Bonhomme et al., 2004).The excretion of a secondary metabolite in the medium is a characteristic leading to an improved productivity in bioreactor cultures (Bourgaud et al., 2001).Regardless of the incubation duration and the concentration of precursor, the latter was detected both in shoot cultures and culture media indicating that it was transported into the plant tissue (Table 2).However, the uptake of this compound by plant tissue was not complete, as 4'-Omethyl-d 3 -norbelladine was still detected in culture media after 15, 30 or 40 days of incubation.
Incorporation of 4′-O-methyl-d 3 -norbelladine into both alkaloids was observed in different conditions tested.The labeled galanthamine and lycorine contained respectively 3 and 2 atoms of deuterium.Incorporation of two deuterium atoms in lycorine might result from an intramolecular cyclisation of the ortho-methoxyphenol yielding the methylenedioxy bridge.This is in accordance with what was previously proposed (Barton et al., 1963) for haemanthamine, another Amaryllidaceae alkaloid, and for benzophenanthridine (Ikezawa et al., 2007) or berberine-like alkaloids (Iwasa & Kim, 1997;Cui et al., 2007).These results appear to constitute a new experimental proof of the proposed cyclisation mechanism.
Labeled galanthamine was observed after 15 days of incubation with 0.10 g/L of the precursor up to 40 days of incubation.Nevertheless, at this time 0.20 g/L concentration of precursor did not lead to galanthamine identification.On the other hand, labeled lycorine was observed only after 40 days of incubation with 0.20 g/L of this precursor.These interesting results demonstrate that the biosynthetic pathways of these two alkaloids are in competition, 4′-O-methylnorbelladine being the common precursor.Nevertheless, a domination of the parapara' oxidative coupling on the norbelladine, leading to marithidine and crinine derivatives, has been observed in various populations of Galanthus elwesii (Berkov et al., 2004).
As for the native alkaloids, also deuterated galanthamine and lycorine diffused in the liquid culture medium.

GC-MS analysis
Capillary GC-MS was used in order to identify the various alkaloids, labeled or not, present in the complex fractions of L. aestivum shoot cultures.Derivatization was not required, since the Amaryllidaceae alkaloids retain their characteristic EI/MS fragmentation patterns when employing GC conditions, as reported by Kreh et al. (1995) andTram et al. (2002).As the mass spectrometer coupled with GC works in low-resolution mode (see experimental part), some alkaloids show the same molecular mass, particularly galanthamine and lycorine [M + = 287], demethylnarwedine and crinine [M + = 271], and demethylgalanthamine and demethylmaritidine [M + = 273].However, their fragmentation patterns and their retention times are different.The same GC-MS protocol as previously reported for L. aestivum alkaloids was used (Ptak et al., 2009).Seven compounds showed MS frag-dine was in favour of the para-ortho′ (galanthamine synthesis) and the para-para' (crinine and demethylmarithidine synthesis) oxidative couplings.No native nor labeled trisphaeridine was detected in the extracts.Further investigations with the labeled precursor are still required to better understand Amaryllidaceae alkaloid metabolism in tissue cultures of L. aestivum, in particular the flux rates between the biosynthetic intermediates.In conclusion, the common precursor 4′-O-methyl-d 3 -norbelladine is biotransformed in shoot cultures of L. aestivum and gives labeled alkaloids previously reported in the literature but using 13 C or 14 C-labeled precursor.

Figure 4 .
Figure 4. LC-MS mass spectra of lycorine.Spectra were acquired in the positive-ion mode.(a) Mass spectrum of authentic lycorine; (b) mass spectrum of native and deuterated lycorine from an extract of shoot cultures of Leucojum aestivum.Error < 4 ppm in both cases, theoretical masses are given in parentheses.

Table 1 . Growth rate of Leucojum aestivum shoot culture in me- dium enriched with 4'-O-methyl-d 3 -norbelladine (D3MN) at vari- ous concentrations
Data represent average of four replications with standard deviation.
Spectra were acquired in the positiveion mode.(a) Mass spectrum of authentic galanthamine; (b) mass spectrum of native and deuterated galanthamine from an extract of shoot cultures of Leucojum aestivum.Error < 4 ppm in both cases, theoretical masses are given in parentheses.Biotransformation of deuterium-labeled precursor in Amaryllidaceae