Phytochemistry 88 (2013) 43–53 Contents lists available at SciVerse ScienceDirect Phytochemistry journal homepage: www.elsevier.com/locate/phytochem Seasonal accumulation of major alkaloids in organs of pharmaceutical crop Narcissus Carlton Andrea Lubbe a,⇑, Henk Gude b, Robert Verpoorte a, Young Hae Choi a a b Natural Products Laboratory, Institute of Biology Leiden, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands Wageningen University and Research Centre, Applied Plant Research, Flowerbulbs, Nursery Stock and Fruits, P.O. Box 85, 2160 AB Lisse, The Netherlands a r t i c l e i n f o Article history: Received 15 October 2012 Received in revised form 13 December 2012 Available online 11 January 2013 Keywords: Narcissus pseudonarcissus Amaryllidaceae Daffodils Seasonal variation Alkaloids Galanthamine Haemanthamine Narciclasine a b s t r a c t Narcissus pseudonarcissus (L.) cv. Carlton is being cultivated as a main source of galanthamine from the bulbs. After galanthamine, haemanthamine and narciclasine are the next most abundant alkaloids in this cultivar. Both these compounds are promising chemical scaffolds for potential anticancer drugs. For further research and drug development, a reliable supply of these compounds will be needed. In this study a field experiment was conducted to investigate the levels of galanthamine, haemanthamine and narciclasine in plants of N. pseudonarcissus cv. Carlton. In a field experiment alkaloids in the bulbs, leaves and roots were analyzed by quantitative 1H NMR to monitor the variations during the growing season. Major primary and secondary metabolites were identified in the various plant parts. Multivariate data analysis was performed on the 1H NMR spectra to investigate how metabolites changed in the plant organs over time. The results show that the leaves have relatively high concentrations of the alkaloids before flowering. The bulbs had lower concentrations of the compounds of interest but would have a higher total yield of alkaloids due to bigger biomass. Narcissus pseudonarcissus cv. Carlton represents a good source of galanthamine, and can potentially be a source of the other major alkaloids depending on choice of organ and harvest time. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction Daffodils (Narcissus species) are one of the major ornamental flower crops in the Netherlands, with large volumes of bulbs being planted every year. Since recently one of the varieties, Narcissus pseudonarcissus (L.) cv. Carlton, is being cultivated for the extraction of the compound galanthamine from the bulbs (Berkov et al., 2009). This cultivar was chosen because of the relatively high concentration of galanthamine in the bulbs, the large bulb size and the availability of large volumes of planting stocks of the bulbs (Kreh, 2002). Galanthamine is an alkaloid that occurs in several species of the Amaryllidaceae family. Due to its effects on the human brain it is now a registered drug for the symptomatic treatment of early stage Alzheimer’s disease (Sramek et al., 2000; Heinrich and Lee Teoh, 2004). More than 300 alkaloids have been isolated from the genus of Narcissus, and most of them possess some biological activities (Bastida et al., 2006). Galanthamine is usually reported as the major alkaloid in the bulbs of N. pseudonarcissus, followed by haemanthamine as the second-most abundant (Fig. 1), and a number of other minor alkaloids such as homolycorine, lycoramine and O-methyllycorenine (Kreh et al., 1995; Gotti ⇑ Corresponding author. E-mail addresses: (A. Lubbe). a.lubbe@chem.leidenuniv.nl, andrealubbe@gmail.com 0031-9422/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.phytochem.2012.12.008 et al., 2006). More than 20 alkaloids have been isolated from the Narcissus cultivar ‘‘Carlton’’ (Bastida et al., 2006). Narciclasine is an isocarbostyril compound similar in structure to the alkaloid lycorine. In spite of its non-basic character it is usually still included in the group of Amaryllidaceae alkaloids (Bastida et al., 2006). Narciclasine was first identified in N. pseudonarcissus by Piozzi et al. (1969). Although narciclasine is known to occur in N. pseudonarcissus cv. Carlton, this compound is usually not included in studies on the alkaloid profile of this cultivar. This is due to the extraction conditions in these analyses (usually involving acid–base extraction of alkaloids) being selective for basic compounds. Consequently, the studies that report changes in the alkaloid levels in Narcissus during the course of the growing season do not include narciclasine (Kreh, 2002). Haemanthamine also has interesting biological properties, including inhibition of protein synthesis, antiretroviral, antiparasite and antimalarial activity, as well as cytotoxicity against various cancer cells (Bastida and Viladomat, 2002; Sener et al., 2003; Szlavik, 2004; Osorio et al., 2010). The bioactivity of haemanthamine to induce apoptosis in cancer cells has attracted much research attention, and the compound has potential as a chemical scaffold for producing derivatives that could become future cancer drugs (McNulty et al., 2007; Evidente and Kornienko, 2009; Van Goietsenoven et al., 2010). Narciclasine also has the ability to induce apoptosis in various human cancer cell lines (Dumont et al., 44 A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 Fig. 1. Chemical structures of galanthamine, haemanthamine and narciclasine. The protons used for quantitative NMR analysis are indicated on the structures. 2007). This compound and some of its semi-synthetic derivatives are also promising candidates for anticancer drugs, particularly against apoptosis-resistant cancer cells (Ingrassia et al., 2009). For further research, clinical trials and beyond a stable supply of these compounds would be needed. This is often a limitation in drug development with natural products (McChesney et al., 2007). Since N. pseudonarcissus cv. Carlton is already being cultivated for the extraction of galanthamine, it represents an available and well-established source of the other compounds as well. In this study a field experiment was conducted to investigate the levels of galanthamine, haemanthamine and narciclasine in plants of N. pseudonarcissus cv. Carlton. Plants were harvested at different time-points throughout the growing season. The bulbs, leaves and roots were analyzed by quantitative 1H NMR to monitor the variations in the major alkaloids through time. Various primary and secondary metabolites were identified in different parts of the plant. Multivariate data analysis was performed on the 1H NMR spectra to see how detectable metabolites changed in the plant organs over time. 2. Results 2.1. Quantitation of major alkaloids in bulbs, stems and roots The galanthamine concentration in the bulbs was determined using a quantitative 1H NMR method developed previously (Lubbe et al., 2010). The area under the doublet signal at d 6.17 belonging to H-4a of galanthamine was used for quantitation. The doublet signal at d 6.52 (H-1) of haemanthamine and the multiplet signal at d 6.22 (H-1) of narciclasine were used in a similar manner for quantitation of these compounds. In some leaf samples the H-1 doublet of haemanthamine was overlapped with other signals, in which case the broad singlet at d 5.99 (OCH2O) was used for quan- titation. The same signals, where present, were used for quantitation of these metabolites in the leaf and root samples. Galanthamine was the major alkaloid in all the bulb samples, followed by haemanthamine and narciclasine (Fig. 2). The average bulb galanthamine concentration increased from the first timepoint to a maximum before flowering. The concentration decreased over the next two time-points but showed a slight increase again at senescence of the aerial parts. Haemanthamine, the second most abundant alkaloid showed a similar pattern, with the highest average concentration just before flowering. Narciclasine in the bulbs was highest on average before flowering and during flowering, and thereafter steadily decreased until the end of the season. In the leaves haemanthamine was the major alkaloid in the first three time-points (Fig. 2). The concentration varied through the season with a maximum before flowering. Galanthamine and narciclasine were present at roughly the same levels in the leaves, with steady levels until full flowering followed by a decrease after flowering. These three compounds could not be detected in the senescent leaves. Galanthamine and haemanthamine were present in the root, with galanthamine at higher concentrations in the first two time-points and haemanthamine significantly higher in the final time-point (Fig. 2). Narciclasine, if present in the roots, was at levels too low to detect. 2.2. Metabolite identification NMR-based analysis was applied not only to the quantitation of alkaloids but also to other metabolites profiling. Identification of metabolites was done with the aid of two-dimensional NMR experiments (COSY, J-Resolved and HMBC), as well as comparison of signals with an in-house metabolite database and previously reported data (Verpoorte et al., 2007; Kim et al., 2010). Signal assignments in bulb, leaf and root extracts are summarized in Table 1. The metabolites detected in the bulbs in this study were similar to what has previously been reported in studies using this NMR extraction method (Lubbe et al., 2010, 2011), and a GC–MS based analysis of N. pseudonarcissus cv. Carlton bulb metabolites (Berkov et al., 2011). The bulb spectra were mostly similar in terms of the metabolites present, but the levels of some signals fluctuated between time-points. In the leaf samples differences in metabolites were clearly seen between time-points, with more qualitative differences between the different stages (Fig. 3). In all spectra intense signals of the carbohydrates sucrose, fructose and glucose were present. The alka- Fig. 2. Results of quantitative 1H NMR analysis of the major alkaloids in Narcissus pseudonarcissus cv. Carlton (a) bulbs, (b) leaves and (c) roots at different time-points during the growing season. (A) shoots emerge, (B) before flowering, (C) full flowering, (D) after flowering, (E) shoot senescence. 45 A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 Table 1 1 H chemical shifts (d) and coupling constants (Hz) of Narcissus pseudonarcissus bulb metabolites in CH3OH-d4-KH2PO4 in D2O at pH 6.0. Metabolite Chemical shifta (d) and coupling constant (Hz) Organb Trigonelline Fumaric acid Kaempferol analog 1 Kaempferol analog 2 Cytosine analog Phenylpropanoid Quercetin analog 1 Quercetin analog 2 Quercetin analog 3 Rutin 9.14 (s), 8.87 (m), 8.10 (dd) J = 7.5, 6.5, 4.42 (s) 8.47 (s) 8.09 (d) J = 8.9, 7.24 (d) J = 8.9, 6.45 (d) J = 2.0, 6.29 (d) J = 2.0, 5.17 (7.6) 8.02 (d) J = 8.73, 6.98 (d) J = 8.73 7.99 (d) J = 8.2, 5.96 (d) J = 8.2 7.76 (d) J = 16.0, 6.48 (d) J = 16.0 7.86 (d) J = 2.0, 7.68 (dd) J = 2.0, 9.9, 7.26 (d) J = 9.9 7.69 (d) J = 8.8, 7.56 (dd) J = 8.8, 2.0, 6.97 (d) J = 8.8 7.60 (d) J = 1.84, 7.54 (dd) J = 8.55, 1.85 7.72 (d) J = 2.1, 7.67 (dd) J = 2.1, 8.5, 6.99 (d) J = 8.5, 6.51 (d) J = 2.0, 6.31 (d) J = 2.0, 5.07 (d) J = 7.0, 4.53 (d) J = 2.6, 1.1 (d) J = 6.4 7.61 (d) J = 15.9, 6.37 (d) J = 15.6, 7.14 (d) J = 1.0, 7.06 (dd) J = 1.0, 8.9, 6.88 (d) J = 8.9, 5.32 (t of d), 4.10 (d), 3.98, 2.02, 2.03, 1.93 (d of t), 1.88 (dd) 7.61 (d) J = 15.9, 6.41 (d) J = 15.9, 7.15 (d) J = 1.8, 7.05 (dd) J = 8.0, 1.8, 6.88 (d) J = 8.0 7.24 (d) J = 8.8, 7.06 (d) J = 8.8 7.16 (d) J = 8.6, 6.80 (d) J = 8.6 7.12 (d) J = 8.5, 6.76 (d) J = 8.5 7.72 (d) J = 7.9, 7.47 (d) J = 8.13, 7.29 (s), 7.21 (t) J = 7.5 7.42–7.33 (m), 3.09 (dd) J = 8.3, 14.8 7.18 (d) J = 8.4, 6.85 (d) J = 8.4 7.13 (d) J = 8.4, 6.73 (d) J = 8.4, 3.02 (d) J = 13.6, 2.98 (d) J = 13.6 7.06 (s), 7.04 (s), 6.02 (s), 5.73 (brs) 7.03 (s) 6.94 (d) J = 8.4, 6.88 (d) J = 8.4, 6.16 (d) J = 10.5, 6.06 (dd) J = 10.5, 5.0, 2.86 (s) 7.06 (s), 6.71 (s), 6.51 (d) J = 10.3, 6.36 (dd) J = 10.3, 5.0, 5.97 (brs), 6.22 (m), 6.09 (2d) J = 4.5, 6.73 (s), 4.39 (m), 4.32 (m) 6.22 (m), 6.09 (2d) J = 4.5, 6.73 (s), 4.39 (m), 4.32 (m) 5.55 (d) J = 3.8, 5.27 (d) J = 3.8 5.41 (d) J = 3.8, 4.17 (d) J = 8.7, 4.03 (t) J = 8.3, 3.78–3.83 (m), 3.75 (t) J = 9.5, 3.66 (s), 3.51 (dd) J = 9.9, 3.9, 3.43 (t) J = 9.5 5.23 (d) J = 3.7 5.17 (d) J = 3.8, 5.40 (d) J = 3.9 5.14 (d) J = 1.5 5.11 (d) J = 1.5, 4.87 (d) J = 0.7, 1.28 (t) J = 6.5 4.58 (d) J = 7.9 (b-anomer), 5.19 (d) J = 3.8 (a-anomer), 3.20 (dd) J = 8.9, 8.8 4.07 (m), 4.02 (dd) J = 13.8, 1.0, 3.94 (m), 3.85 (dd) J = 9.8, 3.6, 3.79 (m), 3.70 (d) J = 11.8, 3.52 (d) J = 11.8 3.21 (s) 3.12 (t) J = 5.3 3.94 (dd) J = 8.0, 4.0, 2.95 (dd) J = 16.8, 3.8, 2.81 (dd) J = 17.0, 8.2 2.82 (dd) J = 17.0, 8.5, 2.63 (dd) J = 17.0, 9.5 2.71 (d) J = 15.8, 2.56 (d) J = 15.8 2.68 (dd) J = 15.7, 3.4, 2.36 (dd) J = 15.7, 10.4, 4.28 (dd) J = 10.4, 3.2 2.46 (t of d), 2.16–2.10 (m) 2.39 (td) J = 7.1, 2.5, 2.10–2.18 (m) 2.47 (s) 3.24 (t) J = 8.0, 1.92 (m), 1.65–1.78 (m), 3.71 (t) J = 5.8 1.91 (s) 1.49 (d) J = 7.2 1.06 (d) J = 7.0, 1.01 (d) J = 7.0 1.03 (d) J = 7.1, 0.96 (t) J = 7.4 1.19 (t) J = 7.0 1.34 (d) J = 6.6, 4.22 (m) 1.31 (brs), 2.18 (t) J = 7.4, 1.56 (m), 0.89 (t) J = 7.4, 5.40 (m) R B B,L B,L B,R B,L B,L B,L R B,L Chlorogenic acid (5-Ocaffeoylquinic acid) 5-O-Feruloylquinic acid Unidentified Phenolic 1 Unidentified Phenolic 2 Unidentified Phenolic 3 Tryptophan Phenylalanine Tyrosine 4-Hydroxyphenylpyruvate Lycorenine cis-Aconitic acid Galanthamine Haemanthamine Narciclasine Raffinose Sucrose Arabinose Maltose Mannose Rhamnose Glucose Fructose Choline Ethanolamine Asparagine Aspartic acid Citric acid Malic acid Glutamine Glutamic acid Succinic acid Ornithine Acetic acid Alanine Valine Isoleucine 1-O-ethyl glucoside Threonine Fatty acids a b B,L B,L R R R L,R B,L,R B,R B,L B B,L,R B,L,R B,L,R B,L B B,L,R B,R B,R B B B,L,R L,R B B,L,R B,L,R B,L,R B,L B,L,R L,R B,L R B B B,L,R B,L,R B,L,R B,L,R B,L,R B,R Chemical shift was calibrated to TMSP at d 0.0. Metabolite detected in organs: Bulb (B), Leaf (L), Root (R). loid levels were highest at the earlier time-points and decreased over the course of the season. Major signals were seen in the leaf spectra at around d 4.28, d 2.70 and d 2.43 and were assigned to malic acid. These signals are shifted in the later time-points (Fig. 3b), a phenomenon known to occur with malic acid signals due to changes in concentration and pH (Kim et al., 2006). The amino acids alanine, threonine, valine, isoleucine, glutamine, aspartic acid, asparagine, phenylalanine and tryptophan were also identified in the leaf samples. The aromatic region of the leaf spectra contained various signals not seen in the bulb spectra (Fig. 4, Table 1). One of these was identified as chlorogenic acid, a phenolic compound commonly found in photosynthetic tissue (Verpoorte et al., 2007). Signals similar to chlorogenic acid, but shifted slightly downfield, match those reported for 5-O-feruloylquinic acid, which has a methoxy group instead of a hydroxyl group at C-30 (Leiss et al., 2009). The flavonol glycoside rutin was identified by its characteristic doublet and double-doublet signals of the A ring at d 6.51 and d 6.31 and B ring at d 7.72, d 7.67, d 6.99 of the flavonol moiety, as well as the doublet signal of the rhamnosyl moiety at d 1.10 with doublet (J = 6.4 Hz) (Zhi et al., 2012). Signals indicating the presence of two kaempferol analogs and two quercetin analogs were detected. Some signals were seen only in the first time-point, namely those assigned to the alkaloid precursor 4-hydroxyphenylpyruvate (4HPP) and an unidentified phenylpropanoid. A triplet signal at d 1.19, assigned to 1-O-ethyl glucoside was seen only in leaf samples in the after flowering stage. The metabolite profiles of root samples were inspected and several primary and secondary metabolites were assigned (Table 1, Figs. 5 and 6). Haemanthamine and galanthamine were present in all the samples, as well as cis-aconitic acid. A cytosine analog was seen in the aromatic part of the spectrum in all time-points, 46 A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 Fig. 3. 1H NMR spectra from (a) d 6.0–8.2 and (b) d 0.9–3.2 of representative Narcissus pseudonarcissus cv. Carlton leaf extracts from different time-points in the growing season. (A) Shoots emerge, (B) before flowering, (C) full flowering, (D) after flowering, (E) shoot senescence. (1) Kaempferol analog 1, (2) kaempferol analog 2, (3) quercetin analog 1, (4) rutin, (5) chlorogenic acid, (6) 5-O-feruloylquinic acid, (7) tryptophan, (8) phenylalanine, (9) 4-hydroxyphenylpyruvate, (10) cis-aconitic acid, (11) haemanthamine, (12) galanthamine, (13) narciclasine, (14) malic acid, (15) asparagine, (16) citric acid, (17) aspartic acid, (18) glutamine, (19) glutamic acid, (20) alanine, (21) threonine, (22) valine, (23) isoleucine, (24) 1-O-ethyl glucoside. and primary metabolites malic acid, fructose, glucose and sucrose were present in all samples. Signals matching those of trigonelline (Lopez-Gresa et al., 2010) were seen in the first two time-points, together with the amino acids tryptophan and phenylalanine. The amino acids such as threonine, isoleucine and valine were also present in earlier samples. Signals tentatively assigned to a querce- A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 47 Fig. 4. 1H–1H–J-resolved spectrum of a representative Narcissus pseudonarcissus leaf extract in KH2PO4 buffer and CH3OH-d4 (1:1) pH 6.0, harvested around the time of flowering. (1) Kaempferol analog 1, (2) quercetin analog 1, (3) rutin, (4) chlorogenic acid, (5) 5-O-feruloylquinic acid, (6) quercetin analog 2, (7) haemanthamine, (8) galanthamine, (9) narciclasine. tin analog occurred in root samples around the time of flowering. At leaf senescence, many signals not seen at any other time-points were present. In the aromatic region of the spectra, characteristic phenolic signals were observed at relatively high levels. Fatty acid signals were observed in the last three time-points, with the signals the most intense at the senescence time-point. 2.3. Multivariate data analysis The recorded 1H NMR spectra of the bulbs, stems and roots of N. pseudonarcissus harvested at different time-points were submitted to principal component analysis (PCA) to obtain an overview of the variation in metabolite profiles between time-points. For the bulbs the first two Principal components (PCs) of the model together accounted for 79.2% of the variance in the dataset. The first four timepoints were separated along PC1, while PC2 separated the last time-point from the others (Fig. 7a). There were some overlaps between the first four time-points, although most of A and B samples occurred on the negative side of PC1, and most of C and D samples on the positive side of PC1. The variance in the leaf 1H NMR spectra was well described by the PCA model of these samples. The first two PCs accounted for 84.1% of the variance, and clear separation and grouping of the time-points could be seen in the resulting score plot shown in Fig. 7b. PC1 separated time-point A and B from the rest of the groups, and C, D and E were separated along PC2. The results of the PCA of the roots samples are shown in Fig. 7c. The first two principal components accounted for 68.7% of the variance for the original dataset. PC1 separated the senescence timepoint from the other time-points, which were separated along PC2. The first time-point (A) was well separated from the others, but as in the bulb samples there was some overlap between the groups of time-point B, C and D along PC2. The difference between the metabolite profiles of the different organs of N. pseudonarcissus was further investigated to see which metabolites were responsible for the groupings observed in the PCA results. The NMR samples of each organ were grouped according to time-points. A supervised data analysis method, Partial Least Squares Discriminant Analysis (PLS-DA) was used to correlate NMR signals to the different time-point groups. The results of the PLSDA analysis are shown in Fig. 8. Bulb 1H NMR signals (the X variables in this analysis) correlated to the different time-points (Y variables) could be read from the loading scatter plot (Fig. 8b). With the aid of a variable importance plot (not shown) the signals most important for the discrimination of the groups in the model could be identified. For the bulbs, signals important in the first two time-points belonged to phenylalanine, alanine, valine, and ornithine. Signals belonging to unsaturated fatty acids were also identified as important for these time-points. Bulbs samples harvested in the full flowering timepoint or after flowering grouped together in the PLS-DA analysis (Fig. 8a). Signals highly correlated to these time-points belonged to the sugars glucose and sucrose. The last time-point (E: senescence) bulb samples were quite clearly separated from the others in the PLS-DA score scatter plot. Here the signals most important for this discrimination belonged to 4-HPP. Also correlated to this time-point were the double-doublet signals at d 2.81 and d 2.82 of asparagine and aspartic acid, and alkaloids signals of galanthamine, narciclasine, and citric acid. A cytosine analog with doublet signals at d 7.99 and d 5.96 (J = 8.3) was also important for the discrimination of this group. 1 H NMR signals correlated to the first two time-points in the leaves were identified by inspection of the PLS-DA loading scatter plot (Fig. 8) and Variable Importance plot. These signals belonged to galanthamine and narciclasine and the alkaloid precursor 4hydroxyphenylpyruvate (4-HPP). Additional signals correlated to the earlier time-point samples were unsaturated fatty acid signals, aspartic acid, asparagine, alanine, fructose and glucose. Samples of the full flowering and after flowering time-points were quite sim- 48 A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 Fig. 5. 1H NMR spectra from (a) d 8.1–5.9 and (b) d 3.0–0.8 of representative Narcissus pseudonarcissus cv. Carlton root extracts from different time-points in the growing season, measured in KH2PO4 buffer and CH3OH-d4 (1:1) pH 6.0. (A) Shoots emerge, (B) before flowering, (C) full flowering, (D) after flowering, (E) shoot senescence. (1) Cytosine analog, (2) tryptophan, (3) phenylalanine, (4) phenolic 1, (5) tyrosine, (6) phenolic 2, (7) phenolic 3, (8) cis-aconitic acid, (9) haemanthamine, (10) galanthamine, (11) malic acid, (12) aspartic acid, (13) asparagine, (14) glutamine, (15) succinic acid, (16) acetic acid, (17) alanine, (18) threonine, (19) valine, (20) isoleucine, (21) fatty acids. ilar, with the same 1H NMR signals important for their discrimination. These included signals of secondary metabolites, such as chlorogenic acid, 5-O-feruloylquinic acid and rutin. Signals from a flavonoid tentatively assigned as a quercetin analog were also A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 49 Fig. 6. 1H–1H–J-resolved spectrum of a representative Narcissus pseudonarcissus root extract in KH2PO4 buffer and CH3OH-d4 (1:1) pH 6.0, harvested at time of leaf senescence. (1) Phenolic 1, (2) tyrosine, (3) phenolic 2, (4) phenolic 3, (5) cis-aconitic acid, (6) haemanthamine, (7) galanthamine. Fig. 7. Score scatter plots of principal component analysis (PC1 versus PC2) of Narcissus pseudonarcissus bulbs (a), leaf (b) and root (c) samples harvested at different timepoints throughout the growing season. (A) Shoots emerge, (B) before flowering, (C) full flowering, (D) after flowering, (E) shoot senescence. 50 A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 Fig. 8. Score plots (PLS component 1 and 2) of PLS-DA results obtained from 1H NMR spectra of Narcissus pseudonarcissus bulbs (a), leaves (c) and roots (e) and corresponding loading plots for bulbs (b), leaves (d) and roots (f). Samples assigned to classes according to time of harvest: (A) shoots emerge, (B) before flowering, (C) full flowering, (D) after flowering, (E) shoot senescence. In the loading plots the squares (j) indicate the average of the classes; triangles (N) represent 1H NMR signal buckets. Signal buckets important for discrimination of the assigned classes are labeled; Suc: sucrose, Glc: glucose, Fru: fructose, Ara: Arabinose, MA: malic acid, Phe: phenylalanine, Ala: alanine, Orn: ornithine, Val: valine, FA: fatty acids, FA(u): unsaturated fatty acids, 4-HPP: 4-hydroxyphenylpyruvate, CA: citric acid, GAL: galanthamine, NAR: narciclasine, CHLA: chlorogenic acid, 5FQA: 5-O-feruloyquinic acid, QA: Quercetin analog, Cyt: cytosine derivative, RU: rutin, Acon: cis-aconitic acid, Eth: ethanolamine, Asp: aspartic acid, Asn: Asparagine, Tyr: tyrosine, Tr: trigonelline. important for the discrimination. The senescent leaves of the last time-point had metabolite profiles similar to the after flowering samples, except with no alkaloid signals present. Many of the other secondary metabolite signals (flavonoids, chlorogenic acids) were still present but at lower intensities. In the PLS-DA loading scatter plot signals assigned to cis-aconitic acid and malic acid were correlated to the last time-point samples. Several unassigned signals were also seen to be important for the discrimination, such as a doublet at d 5.60 (H-1 of glycoside) and singlets at d 1.15, d 1.05 and d 0.99. A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 In the root samples, the first four time-points were mainly correlated to sugar signals. Signals of sucrose and fructose were important for discrimination of the first two time-points, and those of glucose and arabinose for the flowering and after flowering points. Additional signals correlated to the first two and following two time-points were those assigned to trigonelline and malic acid, respectively. The senescence root samples were characterized by more intense signals in the aromatic region of the spectra, and less intense sugar signals. The PLS-DA analysis revealed signals important for the discrimination of these samples belonging to tyrosine and various unidentified phenolic compounds. Signals of a quercetin analog and cis-aconitic acid were also correlated to the senescence root samples, as well as signals assigned to fatty acids and glutamine. 3. Discussion Seasonal changes in the overall metabolite profiles of N. pseudonarcissus cv. Carlton have to our knowledge not been reported before. In this study 1H NMR-based metabolite profiling allowed changes in major primary and secondary metabolites to be observed in the bulbs, leaves and roots of the plant. The most notable changes in the bulb samples over the growing season were in the level of sugars, fatty acids and amino acids. In the leaves, the profile of aromatic compounds changed substantially throughout the season. Apart from clear changes in the alkaloid levels, there were also qualitative changes in the flavonoid and chlorogenic acids profiles. Rutin, the most abundant flavonol glycoside, and other quercetin analogs increased around the time of flowering to reach maximum levels at or after flowering. In contrast, kaempferol analogs were higher in the early time-points shortly after the shoots emerged from the soil. Chlorogenic acid and the related compound 5-O-feruloylquinic acid also showed the highest concentration in the leaves at the time of flowering. Many roles have been attributed to flavonoids in plant development, such as in defense, allelopathy, hormone transport and regulation, plant architecture and modulation of reactive oxygen species (reviewed in Buer et al., 2010). Similarly, chlorogenic acid and related compounds are believed to be involved in stress responses, photoprotection, cell wall building and organogenesis (Grace et al., 1998; Franklin and Dias, 2011). As yet not much is known about the role of these phenolic compounds in geophytes such as Narcissus, but in this study seasonal changes occurred that seem related to the flowering phase. 1H NMR spectra of the roots revealed the presence of trigonelline in the early time-point samples. Other major changes in the root samples were the appearance of large phenolic signals at leaf senescence, possibly related to structural changes of the root cell walls before dormancy (Wilson and Anderson, 1979). A large increase in fatty acids was also seen in the roots towards the end of the season. Such an increase may be related to increased deposition of the poly-aliphatic domain of suberin at this point in the season, as observed in onion roots by Meyer et al. (2011). The alkaloid precursors tyrosine and 4-HPP were present in all time-points in the bulbs, with 4-HPP being highest in the last time-point. This suggests that the alkaloids are being produced in the bulbs throughout the season, with an increase in 4-HPP leading to an increase in alkaloids at the end of the season. Similarly, higher levels of alkaloids in the leaves (before flowering) were correlated with higher levels of the alkaloid precursor 4-HPP. The root samples had similar levels of the alkaloids at all the time-points, with the exception of the last one where the average concentration of haemanthamine increased. An increase in tyrosine, but not 4-HPP, was also seen in the last time-point. No 4-HPP was detected in any of the root samples, suggesting that the increase in haemanthamine seen at the end of the season was the result of trans- 51 port from other organs as opposed to biosynthesis in the roots. The increased tyrosine was likely related to the up-regulation of the phenylpropanoid pathway for biosynthesis of other phenolic compounds. Quantitation of galanthamine and haemanthamine in the different organs during the growth season yielded results comparable to those of previous reports in cv. Carlton (Kreh, 2002). The method used in this study also allowed quantitative analysis of narciclasine. In the bulbs, the narciclasine concentration reached a maximum before flowering, followed by a decrease towards the end of the season. In the leaves narciclasine was highest at the beginning of the season, with concentrations steadily decreasing during the season, similar to galanthamine and haemanthamine. Narciclasine was not detected in the root samples. Narciclasine was previously reported in Carlton bulbs at 100 mg/kg in fresh bulbs (Piozzi et al., 1969). In this study we analyzed narciclasine content by dry weight, with comparable results. Compared to galanthamine, the haemanthamine and narciclasine concentrations in the bulbs are relatively low throughout the growing season. For extraction of a compound for industrial use, as high a concentration of the compound as possible is desired. However, if the bulbs are already being harvested for extraction of galanthamine, narciclasine can be an additional product obtained from the bulb material. Similarly, haemanthamine could also be obtained as a useful side-product from the bulbs. Making use of the basic properties of galanthamine and haemanthamine versus the acidic properties of narciclasine, these compounds can be separated relatively easily for further purification. Early in the growth season, the leaves had relatively high concentrations of galanthamine, narciclasine and especially haemanthamine. Although higher concentrations of alkaloids could be obtained from the leaves harvested at this time than from the bulbs, the total amount of biomass and thus total yield of alkaloids are not very high at this time (Kreh, 2002). Harvesting the whole plant for extraction of the bulbs and foliage at an early time-point would be possible; however bulbs of a higher mass would be obtained at the end of the season at the usual harvest time. Also, for sustainable production of Narcissus plants for alkaloid extraction, harvesting bulbs at the normal time (as for ornamental plant production) would mean that some bulbs could be replanted for the next season, while the rest are extracted for the target compounds. Narcissus pseudonarcissus cv. Carlton is already a source of galanthamine, and can also be used as a source of haemanthamine and/or narciclasine. 4. Materials and methods 4.1. Chemicals and solvents For the NMR analysis CH3OH-d4 (99.80%) from Cambridge Isotope Laboratories (Andover, MA, USA), and phosphate (KH2PO4) buffer (pH 6.0) in deuterium oxide (CortecNet, VoisinsLe-Bretonneux, France) containing 0.01% trimethylsilylproprionic acid sodium salt-d4 (TMSP, w/w) as an internal standard for quantitation and calibration of chemical shift was used. 4.2. Plant material Bulbs of N. pseudonarcissus L. (Amaryllidaceae) cv. Carlton were planted in November 2009 in sandy soil in Lisse, the Netherlands. Three plots were planted, each consisting of two rows of 11, and two rows of 10 bulbs (total 42 bulbs). The rows were planted 18 cm apart, and each plot was surrounded by an edge of open space of 70 cm. All plots received the standard amount of nitrogen and potassium fertilizers, consisting of 110 kg/hectare of Kalksal- 52 A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 Table 2 Summary of the time when the Narcissus pseudonarcissus cv. Carlton plants were harvested in the field. Time-point Description Date lifted A B C D E Shoots emerge 10 cm above ground Before flowering Full flowering After flowering Shoots senescing, end of season (before lifting) 16/03/2010 06/04/2010 03/05/2010 21/05/2010 08/07/2010 peter (Ca(NO3)2 with 19% Ca and 15.5% N, consisting of 14.4% NNO3, 1.1% N-NH4) and 150 kg/hectare Patentkali (K2SO4 and MgSO4, 30% K as K2O, 10% Mg as MgO and 42% S as SO3). Plants were harvested throughout the growing season. Twentyfour plants of each treatment were harvested at five time-points: when shoots had emerged and were about 10 cm above ground (A), before flowering (B), during full flowering (C), after flowering (D) and at normal harvest time after shoot senescence (D). Bulbs, roots and leaves were processed for extraction on the same day of harvest. A summary of the harvest times and dates are given below in Table 2. 4.3. Sample preparation and 1H NMR measurement The plants were rinsed with water to remove soil particles. The roots, bulbs and leaves were separated by cutting with a sharp blade. The basal plates were removed from the bulbs to aid grinding. The different organ samples from each plant were frozen in liquid nitrogen and individually ground to fine powder in a Waring laboratory blender (Waring Products Inc., Torrington, CT, USA). The ground plant material was freeze-dried and kept at 80 °C until analysis. Extraction of N. pseudonarcissus cv. Carlton bulbs, roots and leaves and 1H NMR measurements were carried out as described in one of our previous paper (Lubbe et al., 2010). For the bulbs, twenty-four biological replicates were extracted, and for the leaves and roots there were eight biological replicates at each time-point. Fifty milligram of freeze-dried plant material was weighed into 2 mL microtubes and extracted with 1.5 mL of a mixture of phosphate buffer (pH 6.0) in deuterium oxide containing 0.01% trimethylsilylproprionic acid sodium salt-d4 (TMSP, w/w) and CH3OH-d4 (1:1). The NMR samples were ultrasonicated for 30 min, followed by centrifugation at 13,000 rpm for 10 min. An aliquot of 1 mL of the supernatant was collected and 800 lL transferred to 5 mm-NMR tubes for 1H NMR measurement. 1H NMR spectra were recorded with a Bruker AV 600 spectrometer (Bruker, Karlsruhe, Germany). For each sample 64 scans were recorded using the following parameters: 0.167 Hz/point, pulse width (PW) 4.0 ls and relaxation delay (RD) = 5.0 s. FIDs were Fourier transformed with LB = 0.3 Hz. Manual phase adjustment and baseline correction were applied as well as calibration with internal standard TMSP to 0.0 ppm. 4.4. Data processing and multivariate data analysis For quantitative NMR analysis of galanthamine, integration of the doublet proton signal at d 6.17 (galanthamine H-4a) was performed. The ratio of this integral to that of the internal standard was used to calculate the amount of galanthamine per milligram material. Quantitative analysis of haemanthamine and narciclasine was performed in the same way by integration of the doublet proton signal at d 6.48 (haemanthamine H-1) and multiplet proton signal at d 6.22 (narciclasine H-1), respectively. For multivariate data analysis, 1H NMR spectra were automatically binned by AMIX software (v.3.7, Biospin, Bruker). Spectral intensities were scaled to total intensity and the region of d 0.32–10.0 was reduced to integrated regions (‘‘buckets’’ or ‘‘bins’’) of 0.04 ppm each. The regions d 4.7–5.0 and d 3.28–3.34 were excluded from the analysis because of the presence of the residual water and methanol signal, respectively. Principal component analysis (PCA) was performed with SIMCA-P software (v. 12.0 Umetrics, Umeå, Sweden) using the Pareto scaling method. Acknowledgments The authors acknowledge Peter Vreeburg of Praktijkonderzoek Plant en Omgewing (PPO) in Lisse, the Netherlands for help with planning and carrying out the field experiment. This study was done as part of the project ‘‘Inhoudstoffen bloembolgewassen, de basis voor een innovatieve keten’’ of the Zuid-Holland Province (project number 3030001607076). The funding source had no role in the preparation of this article. References Bastida, J., Lavilla, R., Viladomat, F., 2006. Chapter 3 chemical and biological aspects of Narcissus. In: Cordell, G.A. (Ed.), The Alkaloids: Chemistry and Biology. Elsevier, Amsterdam, pp. 87–179. Bastida, J., Viladomat, F., 2002. Alkaloids of Narcissus. In: Hanks, G. (Ed.), Narcissus and Daffodil: The Genus Narcissus. Taylor and Francis, New York, pp. 141–214. Berkov, S., Bastida, J., Viladomat, F., Codina, C., 2011. Development and validation of a GC–MS method for rapid determination of galanthamine in Leucojum aestivum and Narcissus ssp.: a metabolomic approach. Talanta 83, 1455–1465. Berkov, S., Georgieva, L., Kondakova, V., Viladomat, F., Bastida, J., Codina, C., 2009. Plant sources of galanthamine: phytochemical and biotechnological aspects. Biotechnol. Biotechnol. Equip. 23, 1170–1176. Buer, C.S., Imin, N., Djordjevic, M.A., 2010. Flavonoids: new roles for old molecules. J. Integr. Plant Biol. 52, 98–111. Dumont, P., Ingrassia, L., Rouzeau, S., Ribaucour, F., Thomas, S., Roland, I., Darro, F., Lefranc, F., Kiss, R., 2007. The Amaryllidaceae isocarbostyril narciclasine induces apoptosis by activation of the death receptor and/or mitochondrial pathways in cancer cells but not in normal fibroblasts. Neoplasia 9, 766–776. Evidente, A., Kornienko, A., 2009. Anticancer evaluation of structurally diverse Amaryllidaceae alkaloids and their synthetic derivatives. Phytochem. Rev. 8, 449–459. Franklin, G., Dias, A.C.P., 2011. Chlorogenic acid participates in the regulation of shoot, root and root hair development in Hypericum perforatum. Plant Physiol. Biochem. 49, 835–842. Gotti, R., Fiori, J., Bartolini, M., Cavrini, V., 2006. Analysis of Amaryllidaceae alkaloids from Narcissus by GC–MS and capillary electrophoresis. J. Pharm. Biomed. 42, 17–24. Grace, S.C., Logan, B.A., Adams, W.W., 1998. Seasonal differences in foliar content of chlorogenic acid, a phenylpropanoid antioxidant, in Mahonia repens. Plant Cell Environ. 21, 513–521. Heinrich, M., Lee Teoh, H., 2004. Galanthamine from snowdrop – the development of a modern drug against Alzheimer’s disease from local Caucasian knowledge. J. Ethnopharmacol. 92, 147–162. Ingrassia, L., Lefranc, F., Dewelle, J., Pottier, L., Mathieu, V., Spiegl-Kreinecker, S., Sauvage, S., El Yazidi, M., Dehoux, M., Berger, W., Van Quaquebeke, E., Kiss, R., 2009. Structure–activity relationship analysis of novel derivatives of narciclasine (an Amaryllidaceae isocarbostyril derivative) as potential anticancer agents. J. Med. Chem. 52, 1100–1114. Kim, H.K., Choi, Y.H., Verpoorte, R., 2010. NMR-based metabolomic analysis of plants. Nat. Protoc. 5, 536–549. Kim, H.K., Choi, Y.H., Verpoorte, R., 2006. Metabolomic analysis of Catharanthus roseus using NMR and principal component analysis. In: Saito, K., Dixon, R.A., Willmitzer, L. (Eds.), Plant Metabolomics. Springer, Berlin, pp. 261–276. Kreh, M., 2002. Studies on galanthamine extraction from Narcissus and other Amaryllidaceae. In: Hanks, G. (Ed.), Narcissus and Daffodil: The Genus Narcissus. Taylor and Francis, New York, pp. 256–272. Kreh, M., Matusch, R., Witte, L., 1995. Capillary gas chromatography-mass spectrometry of Amaryllidaceae alkaloids. Phytochemistry 38, 773–776. Leiss, K.A., Maltese, F., Choi, Y.H., Verpoorte, R., Klinkhamer, P.G.L., 2009. Identification of chlorogenic acid as a resistance factor for thrips in Chrysanthemum. Plant Physiol. 150, 1567–1575. Lopez-Gresa, M.P., Maltese, F., Belles, J.M., Conejero, V., Kim, H.K., Choi, Y.H., Verpoorte, R., 2010. Metabolic response of tomato leaves upon different plant– pathogen interactions. Phytochem. Anal. 21, 89–94. Lubbe, A., Choi, Y.H., Vreeburg, P., Verpoorte, R., 2011. Effect of fertilizers on galanthamine and metabolite profiles in Narcissus bulbs by 1H NMR. J. Agric. Food Chem. 59, 3155–3161. Lubbe, A., Pomahacova, B., Choi, Y.H., Verpoorte, R., 2010. Analysis of metabolic variation and galanthamine content in Narcissus bulbs by 1H NMR. Phytochem. Anal. 21, 66–72. McChesney, J.D., Venkataraman, S.K., Henri, J.T., 2007. Plant natural products: back to the future or into extinction? Phytochemistry 68, 2015–2022. A. Lubbe et al. / Phytochemistry 88 (2013) 43–53 McNulty, J., Nair, J.J., Codina, C., Bastida, J., Pandey, S., Gerasimoff, J., Griffin, C., 2007. Selective apoptosis-inducing activity of crinum-type Amaryllidaceae alkaloids. Phytochemistry 68, 1068–1074. Meyer, C.J., Peterson, C.A., Bernards, M.A., 2011. Spatial and temporal deposition of suberin during maturation of the onion root exodermis. Botany 89, 119–131. Osorio, E.J., Berkov, S., Brun, R., Codina, C., Viladomat, F., Cabezas, F., Bastida, J., 2010. In vitro antiprotozoal activity of alkaloids from Phaedranassa dubia (Amaryllidaceae). Phytochem. Lett. 3, 161–163. Piozzi, F., Marino, M.L., Fuganti, C., Di Martino, A., 1969. Occurrence of non-basic metabolites in Amaryllidaceae. Phytochemistry 8, 1745–1748. Sener, B., Orhan, I., Jutamad, S., 2003. Antimalarial activity screening of some alkaloids and the plant extracts from Amaryllidaceae. Phytother. Res. 17, 1220. Sramek, J.J., Frackiewicz, E.J., Cutler, N.R., 2000. Review of the acetylcholinesterase inhibitor galanthamine. Expert Opin. Invest. Drugs 9, 2393–2402. 53 Szlavik, L., 2004. Alkaloids from Leucojum vernum and antiretroviral activity of Amaryllidaceae alkaloids. Planta Med. 70, 871. Van Goietsenoven, G., Andolfi, A., Lallemand, B., Cimmino, A., Lamoral-Theys, D., Gras, T., Abou-Donia, A., Dubois, J., Lefranc, F., Mathieu, V., Kornienko, A., Kiss, R., Evidente, A., 2010. Amaryllidaceae alkaloids belonging to different structural subgroups display activity against apoptosis-resistant cancer cells. J. Nat. Prod. 73, 1223–1227. Verpoorte, R., Choi, Y.H., Kim, H.K., 2007. NMR-based metabolomics at work in phytochemistry. Phytochem. Rev. 6, 3–14. Wilson, K., Anderson, G.J.H., 1979. Further observations on root contraction. Ann. Bot. 43, 665–675. Zhi, H.-J., Qin, X.-M., Sun, H.-F., Zhang, L.-Z., Guo, X.-Q., Li, Z.-Y., 2012. Metabolic fingerprinting of Tussilago farfara L. Using 1H-NMR spectroscopy and multivariate data analysis. Phytochem. Anal. 23, 492–501.