- Email: [email protected]
Identification and quantification of active alkaloids in Catharanthus roseus by liquid chromatography–ion trap mass spectrometry
Food Chemistry 139 (2013) 845–852
Contents lists available at SciVerse ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
Identification and quantification of active alkaloids in Catharanthus roseus by liquid chromatography–ion trap mass spectrometry Qinhua Chen a,b, Wenpeng Zhang a, Yulin Zhang a, Jing Chen a, Zilin Chen a,⇑ a Key Laboratory of Combinatorial Biosynthesis and Drug Discovery (Wuhan University), Ministry of Education, and Wuhan University School of Pharmaceutical Sciences, Wuhan 430071, China b Affiliated Dongfeng Hospital, Hubei University of Medicine, Hubei 442008, China
a r t i c l e
i n f o
Article history: Received 16 January 2012 Received in revised form 31 October 2012 Accepted 28 January 2013 Available online 9 February 2013 Keywords: Alkaloids Identification Quantification Liquid chromatography–ion trap mass spectrometry
a b s t r a c t Catharanthus roseus is an important dicotyledonous medicinal plant that produces anticancer compounds. The active alkaloids vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were identified by direct-injection ion trap-mass spectrometry (IT-MS) for collecting MS1–2 spectra. The determinations of five alkaloids were accomplished by liquid chromatography (LC) with UV and MS detections. The analytes provided good signals corresponding to the protonated molecular ions [M+H]+ and product ions. The precursor ions and product ions for quantification of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were m/z 825 ? 807, 457 ? 397, 353 ? 144, 337 ? 144 and 809 ? 748 by LC–IT-MS, respectively. Two methods were used to evaluate a number of validation characteristics (repeatability, LOD, calibration range, and recovery). MS provided a high selectivity and sensitivity for determination of five alkaloids in positive mode. After optimisation of the methods, separation, identification and quantification of the five components in C. roseus were comprehensively accomplished by HPLC with UV and MS detection. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Catharanthus roseus L (G.) Don, commonly known as Madagascar periwinkle, belongs to the Apocynaceae family. It is known world-wide as a medicinal plant used for production of antitumor, hypotensive, and antiarrhythmic agents (Chen, Li, Zhang, Chen, & Chen, 2011b; Pereira et al., 2010). The major anticancer alkaloids in C. roseus are vinblastine, vincristine, vinleurosine, and vinorelbine. Vinblastine and vinleurosine are indole alkaloids that mainly exist in the aerial part of the plant. Vinblastine has been applied in the treatment of several kinds of cancer, such as Hodgkin’s disease, lymphosarcoma, choriocarcinoma, neuroblastoma, carcinoma of the breast, and chronic leukaemia (Aslam, Mujib, & Sharma, 2011; Javanovics, Bittner, Dezseri, Eles, & Szasz, 1971; Robert, Denise, Wim, Didier, & Robert, 2004). Creasey reported that vinleurosine can be used for the treatment of ascitic tumor (Creasey, 1969). Vinblastine and vinleurosine are produced in vivo by the combination of vindoline and catharanthine, both of which are derived from the terpenoid indole alkaloid biosynthetic intermediate strictosidine through multiple steps. Effective synthetic approaches have been developed by directly coupling catharanthine and vindoline to produce vinblastine (Ferreres
⇑ Corresponding author. Tel.: +86 27 68759893; fax: +86 27 68759850. E-mail address: [email protected] (Z. Chen). 0308-8146/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2013.01.088
et al., 2010a; Ishikawa, Colby, & Boger, 2008). The anti-hypertensive ingredient is ajmalicine, which is used to combat heart arrhythmias and to improve blood circulation in the brain (Almagro, Perez, & Pedreno, 2011). As we know, the chemical complexity and unique bisindole alkaloid structure of the aforementioned molecules have so far hindered their laboratorial synthesis. The isolation of vinblastine and vinleurosine is laborious and costly, mainly due to their low concentrations in the plant and co-existence of a large number of other alkaloids (Bashir, Hafeez, Perveen, Fatima, & Mistry, 1983). Several papers have reported the separation, identification and quantification of active compounds in C. roseus (Barthe et al., 2002; Tikhomiroff & Jolicoeur, 2002; Verma, Hartonen, & Riekkola, 2008; Zhao et al., 2006; Zhu, Huang, Li, & Yin, 2010). The methods of isolation and purification were focused on liquid–liquid extraction (Tikhomiroff & Jolicoeur, 2002), solid phase extraction (Zhao et al., 2006), supercritical fluid extraction (SFE) (Verma et al., 2008), molecularly imprinted polymers (MIPs) extraction (Zhu et al., 2010), and so on. Gas chromatography (GC) is suitable for volatile compounds but requires laborious derivatization. However, GC is unsuitable to analyse the bisindole alkaloids due to their high melting point. Liquid chromatography (LC) and capillary electrophoresis (CE) allow direct analysis of non-volatile metabolites. Non-aqueous capillary electrophoresis (NACE) (Barthe et al., 2002) and CE–MS (Chen et al., 2011b) have been developed for the analysis of Vinca alkaloids. In previous work, we reported
846
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
simultaneous determination of vinblastine and its monomeric precursors by CE–IT-MS. As the injected volume was 2.7 nL, the limits of detection (LOD) by CE–MS were as low as 0.8, 1.0 and 1.0 g/mL for vinblastine, vindoline and catharanthine, respectively (Chen et al., 2011b). Compared with CE, HPLC can provide better analytical precision and higher sample loading capacity. HPLC coupled to UV (Singh, Maithy, Verma, Gupta, & Sushil, 2000; Tikhomiroff & Jolicoeur, 2002) and MS detection (Corona, Casetta, Sandron, Vaccher, & Toffoli, 2008; Ferreres, Pereira, Valentao, & Andrade, 2010b) has been applied in the determination of the active compounds in C. roseus. But, HPLC in combination with tandem mass spectrometry (MS) appears to be a more suitable technique for the screening of the active compounds in plant samples in terms of sensitivity and selectivity. In recent years, analytical techniques have made great advances. LC–MS and LC-tandem MS (LC–MS/MS) such as quadrupole time of flight (Q-TOF) (Klausen et al., 2010) and quadrupole-linear ion trap (Q-LIT) (Ferreres et al., 2010b) have played a crucial role in plant analysis. The application of ion trap (IT) mass spectrometers in the structure elucidation and in the identification of unknowns in complex matrices has been well established (Izumi, Okazawa, Bamba, Kobayashi, & Fukusaki, 2009). Moreover, several reports show that IT mass spectrometers can also be very useful for quantitative analysis (Huang, Chen, Ye, & Chen, 2007; Izumi et al., 2009). The five alkaloids can be studied in positive ion modes by IT-MS, due to their weak basicity from the azacyclo-group in the structure. More importantly, MS2 (MS MS) can afford much more confidence for compound identification and much lower noise level for quantification. The fragmentation map of a target analyte is useful in performing MS2 for either qualitative or quantitative analysis. In previous work, we developed a LC–IT-MS for identification and quantification of oleanolic acid and ursolic acid in nine Chinese herbs (Chen et al., 2011a). In the present work, we have identified the active alkaloids, ajmalicine (A), catharanthine (B), vindoline (C), vinblastine (D) and vinleurosine (E) (Fig. 1). Direct infusion was used to collect MS1 profile and targeted MS2 data of selected ions (five alkaloids) in positive-ion modes. Meanwhile, we developed the HPLC–UV and HPLC–MS/MS methods for simultaneous determination and
quantification of five alkaloids. Quantification of five components was assigned in positive-ion mode. The precursor ions and product ions for quantification of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were m/z 825 ? 807, 457 ? 397, 353 ? 144, 337 ? 144 and 809 ? 748 by LC–IT-MS, respectively. The application of the method on profiling of five alkaloids in C. roseus is also presented. 2. Experimental 2.1. Chemicals and materials C. roseus L (G.) Don was collected from Hainan Yueyang Biotech Co., Ltd (Hainan, China). The samples were crushed and dried in an oven at 50 °C for 24 h before being subjected to extraction procedures. Vinblastine, vindoline, catharanthine, and vinleurosine were also obtained from Hainan Yueyang Biotech Co., Ltd (Hainan, China). Ajmalicine were bought from the National Institute for the Pharmaceutical and Biological Products of China. HPLC grade methanol, acetonitrile and ammonium acetate were obtained from Wuhan Analytical Reagent Company (Wuhan, China). Deionised water was purified using a Milli-Q system (Millipore, Bedford, MA, USA); Helium (purity, 99.999%) and liquid nitrogen were obtained from Wuhan Analytical Instrument Factory (Wuhan, China). Other reagents used in the experiment were of analytical grade and from commercial sources. 2.2. Standard and sample preparation The standard stock solutions of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were prepared by dissolving 10 mg of each in 5 mL methanol to achieve a concentration of 2 mg/mL, and kept at 4 °C. The vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine stock solution were diluted with methanol to obtain calibration solutions ranging from 1 to 100, 4 to 400, 2 to 200, 2 to 200 and 1 to 100 lg/mL, respectively. The sample of C. roseus L (G.) Don (10 g of dried stems) was ground to a fine powder, and then 5 g ultrasonicated with 50 mL 3% sulphuric acid for 30 min. The mixture was filtered and the
Fig. 1. Chemical structures of ajmalicine (A), catharanthine (B), vindoline (C), vinblastine (D) and vinleurosine (E).
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
pH adjusted with 4.0 mL 3% ammonia water. Afterwards, the aqueous solution was partitioned with 20 mL dichloromethane, the organic layer removed (sometimes centrifugation was required to separate both phases), and the same procedure was repeated three times. The organic layers were combined and evaporated under vacuum, and the residue was dissolved with 5.0 mL methanol. Prior to injection, all samples were filtered with a 0.45 lm nylon membrane filter; each sample solution was analysed in triplicate. The sample solutions were stable for at least three weeks if stored at 4 °C (confirmed by re-assay). 2.3. MSn analysis of the reference compounds solution A LC/MSD Trap SL Plus spectrometer (Agilent Technologies, Waldbronn, Germany) equipped with electrospray ionization (ESI) interface and an ion trap mass analyser was applied to the MS and multistage mass spectrometry (MSn) analysis. System control and data analysis was provided by the Agilent LC Chemstation and by Bruker Daltonics Trap Control and QuantAnalysis. A syringe pump was used for the direct loop injections of reference compound working solutions, and the flow rate was set at 0.5 mL/h. The ESI source was used and operated in both positive and negative ion mode. Typical operating conditions were as follows: drying gas (N2) temperature of 350 °C, 5 L/min drying gas flow, 10 psi nebulizer gas (N2) pressure, and 4500 V of capillary voltage. Data were acquired with a smart target of 30,000 and a max accumulation time of 200 ms. First, full-scan MS spectra were obtained by scanning from 100 to 1000 m/z. MS2 acquisition of the most abundant ions in the full-scan MS mode was carried out. 2.4. Chromatography conditions and instrumentation HPLC analyses were performed on an 1100 HPLC instrument (Agilent Technologies, California, USA) equipped with a binary pump, a UV detector, an autosampler, and a column thermostat. Chromatographic separations were carried out on a DL-Cl8 column (5.0 lm, 250 4.6 mm, Japan) column at 25 °C. Elution was performed with a flow rate of 0.5 mL/min. Acetonitrile (A) and 10 mmoL ammonium acetate in water (B) were used as a mobile phase. Gradient elution was used as follows: 10% A (2 min), 10– 30% B (10 min), 30–45% B (30 min), 45–70% B (50 min), 70–80% B (65 min), 80–95% (70 min). After the running, the gradient was set back to 10% A and the system was allowed to equilibrate. The injection volume was 10 lL and the detection wavelength was 280 nm. For LC–MS analysis, the Agilent 1100 HPLC system was coupled on-line to a LC/MSD Trap SL Plus spectrometer (Agilent Corp, Waldbronn, Germany) equipped with electrospray ionization (ESI) source. The Auto MS operation parameters are described as follows: positive-ion mode (ESI+); nitrogen drying gas, 10 L/min; nebulizer, 40 psi; gas temperature, 350 °C; Compound stability, 80%; mass range, 100–1000 m/z. The precursor ions and product ions for quantification of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were m/z 825 ? 807, 457 ? 397, 353 ? 144, 337 ? 144 and 809 ? 748 by LC–IT-MS, respectively. 2.5. Validation and statistical analysis The limit of detection (LOD) and the limit of quantification (LOQ) were calculated using the S/N ratio. LOD and LOQ were calculated as the concentrations producing a recognisable peak with S/N were 3 and 10, respectively. All the determinations were carried out in triplicate, and the experimental results were expressed as means; standard deviations were also given whenever needed. Statistical analysis was
847
performed by Agilent Chemstation Data Analysis software (Agilent Corp, Waldbronn, Germany). 3. Results and discussion 3.1. Identification and MSn behaviour of the reference compounds Electrospray ionization (ESI) in positive mode was chosen for the analysis of five alkaloids. It has been reported that ESI shows the greatest sensitivity for alkaloids when compared with electron impact (EI) and atmospheric pressure chemical ionization (APCI) (Favretto, Piovan, & Filippini, 2001). Because all components have azacyclo-group in their structure, they are weak basic compounds. The weakly basic alkaloids can be ionised readily to obtain cations, which are necessary for the positive ESI responses due to the carried positive charge in a acidic solution with pHs lower above the pKas (Nesbitt, Zhang, & Yeung, 2008). Thus, the positive ion mode and ESI source were adopted for the analysis of the five alkaloids. Five reference compounds were dissolved in methanol and analysed by ESI–MSn in positive-ion mode. The protonated molecule ions of [M+H]+ and major fragment ions were observed in the full scan MS and MS2 spectra in positive ion mode. Fig. 2 shows the protonated molecular ions and major fragment ions spectra for active alkaloids, vinleurosine, vinblastine, catharanthine, vindoline and ajmalicine by the ESI–MS and ESI–MS2. As shown in Fig. 2A and F, the protonated molecular ion [M+H]+ m/z 809 exhibited the high intensity in MS spectra, and major fragment ions were [M H2O+H]+ m/z 791, [M C3H8O+H]+ m/z 748, 624, 527, 469, 353, 224 in MS2 spectra for vinleurosine. The product ion [M C3H8O+H]+ m/z 748 exhibited the highest intensity. Fig. 2B and I showed representative fragment ions were [MH2SO4+H]+ m/z 923, [M+H]+ m/z 825, [M H2O+H]+ m/z 807, [M C3H8O+H]+ m/z 765, 704 in MS and MS2 spectra. The product ion [M H2O+H]+ m/z 825 ? 807 exhibited the highest intensity, and was therefore selected as the quantification ion. The protonated molecular ions and major fragment ions spectra for vindoline, ajmalicine and catharanthine were shown in Fig. 2C–E. All protonated molecular ions exhibited the highest intensity, and the highest intensity of product ions were m/z 144, [M C3H8O+H]+ m/z 397 and m/z 144 for catharanthine, vindoline and ajmalicine, respectively. Thus, quantification of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were m/z 825 ? 807, 457 ? 397, 353 ? 144, 337 ? 144 and 809 ? 748 were chosen for LC–IT-MS, respectively. 3.2. Identification of alkaloids in C. roseus As ESI–IT-MS has proved to be a suitable tool for the identification of the active compounds, the separation and determination of constituents in the extract of C. roseus was performed by LC–ESI– IT-MS in positive mode. A typical total ion chromatogram (TIC) of the identified compounds with MS detection is displayed in Fig. 3. It was observed that several peaks were composed of two or three components, but most peaks had symmetric peak shape and good resolution. The identification of components in the extracts was obtained by mass spectra, retention times, molecular weights (MW), and relative contents reported in literatures (Bashir et al., 1983; Ferreres et al., 2010a; Ishikawa et al., 2008; Javanovics et al., 1971; Tikhomiroff & Jolicoeur, 2002; Zhou, Tai, Sun, & Pan, 2005). The peak identification of major active compounds in the extracts of C. roseus in Fig. 3 is summarised in Table 1. In general, the identified components corresponded to about half the total peak area in TIC chromatograms under positive mode. Compounds 12, 13, 14, 17, 18 were vinblastine, vindoline, catharanthine, ajmalicine and vinleurosine, and identified by the
848
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
Fig. 2. Full-scan ESI–MS spectra of vinleurosine (A), vinblastine (B), catharanthine (C), vindoline (D), and ajmalicine (E), and ESI–MS2 spectra of vinleurosine (F), vinblastine (G), catharanthine (H), vindoline (I) and ajmalicine (J) in positive ion mode by direct loop injecting in ion trap mass.
849
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
Fig. 3. A typical total ion chromatogram (TIC) of active alkaloids in C. roseus (A) and typical extraction ion chromatograms (EIC) of five standard analytes (B) and sample (C) by LC–MS in the positive ion mode. Peak identification: (1) vinblastine, (2) vindoline, (3) catharanthine, (4) ajmalicine, (5) vinleurosine.
Table 1 LC–MS data for the C. roseus. Compound
Ret. time
Name
Protonated molecule ion[M+H]
Major fragment ions (m/z)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
25.5 26.2 27.5 32.3 41.6 44.2 44.9 45.1 48.5 50.2 54.0 54.7 57.7 58.2 58.5 59.2 59.4 60.6 63.1
Sitsirikine Isositsirikine Lochrovicine Lochnerine 19-s-vindolinine Vindolinine Serpentine Serpentine isomer Perimivine Tabersonine Catharosine Vinblastine Vindoline Catharanthine Vincristine Catharine Ajmalicine Vinleurosine Vindolicine
355 355 339 325 337 337 349 349 368 337 385 825 457 337 825 823 353 809 925
340, 297, 274, 227 340, 321, 274, 227 679 [2M+H]+, 322, 300, 283, 227, 185, 163 274, 249, 227, 174, 149 311, 223, 194, 144 320, 308, 276 697 [2M+H]+, 337, 320, 296, 263, 227 697 [2M+H]+, 337, 320, 296, 263, 227 735, 340, 322, 249, 227 323, 305, 276 791[2M+H]+, 366, 353, 335, 324, 275 807, 765, 705 397, 362, 341, 188, 162 320, 173, 144 807, 765, 747, 723, 705, 687 767, 588, 337, 144 319, 291, 265, 249, 222, 210, 178,144 791, 749, 624,527, 469, 353, 224 805, 749, 588, 463, 383
retention times and mass spectra of reference compounds. Four compounds (5, 6, 10, 14) with identical protonated molecule ions of [M+H]+ m/z 337 are shown in Table 1. The major fragment ions of tabersonine were m/z 323, 305, 276, which were described by Ferreres et al. (2010a), and thus the compound 10 was identified as tabersonin and other isomers (19-s-vindolinine and vindolinine). The peak for 19-s-vindolinin occurred more rapidly than
vindolinine, so it was assumed compounds 12 and 13 are 19-s-vindolinine and vindolinine, respectively. Compounds 7 and 8 with identical protonated molecule ions of [M+H]+ m/z 349, dimer-molecule ions of [2M+H]+ m/z 697 were identified as serpentine and its isomer (Ferreres et al., 2010a; Zhou et al., 2005). Compounds 12 and 15 with identical [M+H]+ 231 m/z 825, compound 12 was identified as vinblastine and the other vincristine (Zhou et al., 2005). In
850
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
order to identify other compounds, their characteristics, retention time, protonated molecule ion, main fragment ion and relative content reported in the literature (Barthe et al., 2002; Tikhomiroff & Jolicoeur, 2002; Verma et al., 2008; Zhao et al., 2006; Zhu et al., 2010) were taken into account. We came to a conclusion that compounds 1, 2, 3, 4, 9, 11, 16, 19 were sitsirikine, lsositsirikine, lochrovicine, lochnerine, perimivine, catharosine, catharine and vindolicine, respectively. 3.3. Quantitative analysis by LC–UV and LC–MS 3.3.1. Optimisation of LC conditions For LC–UV–MS, elution was performed with a flow rate of 0.5 mL/min. Also, Flow rates of about 1 ml/min are often used with LC–UV systems to achieve good separation efficiency. By LC–MS, especially MS with ESI source, the flow rate is limited by the spray and evaporation capability of the instrument. A flow rate of 0.5 mL/ min was used in our instrument for the consideration of separation efficiency and ionisation capability. Acetonitrile (A) and 10 mmoL ammonium acetate in water (B) were used as a mobile phase. Gradient elution was used as follows: 10% A (2 min), 10–30% B (10 min), 30–45% B (30 min), 45–70% B (50 min), 70–80% B (65 min), 80–95% (70 min). With a sharp gradient, the base-line separation of five alkaloids could be realised; however, for analysis of real samples, a flat gradient was required for separation of most of compounds in the real sample. 3.3.2. Validation of the method Detector response (relative peak area) was linearly dependent on sample concentration over the described above range for
analytzes. The typical chromatograms of analytes with different detection methods are shown in Figs. 3 and 4. Linear calibration curves were obtained for standard analytes at different concentration levels. The characteristics of the calibration plots are summarised in Table 2. As it can be seen, the five compounds showed excellent correlation and sensitivity in all analytical methods – define excellent. The limits of detection (LODs) were determined by the signal-to-noise (S/N) ratio = 3. The LODs of vinblastine, vindoline, catharanthine, ajmalicine and vinleurosine were determined to be 100, 200, 50, 50, 200 ng/mL for LC–UV, and 5.0, 1.0, 1.0, 10.0, 20.0 ng/mL for LC–MS, respectively. In general, LODs of analytzes by LC–MS were lower than those obtained using LC–UV. The results are shown in Table 2. LC–MS provides a higher selectivity and sensitivity for determination of five alkaloids in positive mode. As compared with our previous work (Chen et al., 2011b), the amount of LOD by CE–IT/MS were 2.16, 0.27and 0.27 for vinblastine, vindoline and catharanthine, respectively. The LOD for vinblastine, vindoline and catharanthineby CE–IT/MS were 2.16, 0.27and 0.27, respectively. The LOD using CE–MS was 23, 37 and 37 times higher than that obtained with LC–IT/MS. Thus, the extent of limits of detection using CE–IT/MS increased sensitivity compared with both LC–IT/MS and LC–UV. Repeatability was investigated by sequentially injecting a series of six standards. Precision was studied by injecting six freshly prepared mixtures. RSD values for relative peak areas are shown in Table 2 using three analytical methods. As shown in Table 2, intra-day precisions were from 0.53% to 2.03% and from 2.55 to 7.80 for analytes by LC–UV and LC–MS, respectively. And inter-day precisions were from 1.03% to 2.95%, and from 4.21% to 8.08% for
Fig. 4. Typical chromatograms of five standard analytes (A) and sample (B) by LC–UV. Peak identification: (1) vinblastine, (2) vindoline, (3) catharanthine, (4) ajmalicine, (5) vinleurosine.
851
Q. Chen et al. / Food Chemistry 139 (2013) 845–852 Table 2 Method validation data and the contents of vinblastine, vindoline, catharanthine, ajmalicine and vinleurosine in C. roseus by LC–UV and LC–MS. Method
Compound
Calibration curve
UV
Vinblastine Vindoline Catharanthine Ajmalicine Vinleurosine
y = 11.6x y = 3.638x y = 23.61x y = 20.79x y = 3.740 x
18.14 7.111 18,82 17.65 4.533
MS
Vinblastine Vindoline Catharanthine Ajmalicine Vinleurosine
y = 2 107x + 9 107 y = 4 107x + 7 108 y = 2 107x + 3 108 y = 2 107x + 2 108 y = 4 106x + 3 107
R2
Line arrange (lg/mL)
Limits of detection (ng/mL)
Precision Intra-day (RSD%)
Inter-day (RSD%)
0.999 0.999 1.000 1.000 0.999
1.0–100 4.0–400 2.0–200 2.0–200 1.0–100
100 200 50 50 200
0.53 2.03 0.55 0.87 1.97
0.978 0.990 0.991 0.992 0.989
1.0–100 4.0–400 2.0–200 2.0–200 1.0–100
5.0 1.0 1.0 10.0 20.0
2.55 6.70 4.05 3.61 7.80
Recovery
Content (lg/g)
Mean (%)
RSD%
1.03 2.95 1.34 1.09 2.76
98.6 99.3 98.5 96.3 97.9
1.04 2.08 2.28 2.57 2.89
7.52 125 45.2 76.3 2.10
4.21 7.01 6.18 5.14 8.08
97.3 97.4 103.9 107.5 95.7
4.88 6.36 7.29 7.11 4.32
6.02 131 40.7 69.2 1.85
analytes by LC–UV and LC–MS, respectively. Although there was better repeatability by LC–UV than that LC–MS, techniques were comparable as estimated by the Student’s t-test at 95% confidence level, which indicated no significant differences between two methods. In order to evaluate the accuracy of proposed method, recovery was tested. Accurate amounts of mixed standards were added to powder sample, and then extracted as described in the preparation of sample and analysed. The recovery values were obtained by their peak areas from the calibration curves under the same conditions by LC–UV and LC–MS. As shown in Table 2, the recoveries of analytes were from 96.3% to 99.3% with RSD from 1.04% to 2.89%, and from 95.7% to 107.5% with RSD from 4.32% to 7.29% by LC– UV and LC–MS, respectively. The results indicate that the three methods are suitable for the sample analysis.
for determination of the analytzes in positive mode and UV can obtain an excellent repeatability for analysis of them in C. roseus. In conclusion, the two methods have a good linear, reproducibility, precision, accuracy and recovery, and could be used for quantitative analysis of the five active compounds in C. roseus.
3.3.3. Determination of alkaloids in C. roseus The content of five alkaloids in C. roseus was determined by triplicate injections of three samples. Typical LC–UV and LC–MS chromatograms are shown in Figs. 3B, and 4B. The contents are listed in Table 2. The results obtained by the two techniques were comparable as estimated by the Student’s t-test at 95% confidence level, which indicated no significant differences between two methods. Otherwise, the results obtained by the LC–MS, HPLC/UV and CE–MS in the literature (Chen et al., 2011b; Tikhomiroff & Jolicoeur, 2002; Zhao et al., 2006 and Favretto et al., 2001) were comparable as it was also estimated by the Student’s t-test at 95% confidence level, which indicated no significant difference between the methods. LC–MS provided better sensitivity than CE–MS, as well as better analytical precision and higher sample loading capacity. When comparing with LC–UV, LC–MS has superiority on the screening of the active compounds in plant samples in terms of sensitivity and structure elucidation. However, LC–UV often possessed better precision. Thus, the developed LC–UV and LC–MS methods are suitable for the determination of five active alkaloids in C. roseus.
Almagro, L., Perez, A. J. L., & Pedreno, M. A. (2011). New method to enhance ajmalicine production in Catharanthus roseus cell cultures based on the use of cyclodextrins. Biotechnology Letters, 33, 381–385. Aslam, J., Mujib, A., & Sharma, M. P. (2011). Influence of freezing and non-freezing temperature on somatic embryogenesis and vinblastine production in Catharanthus roseus (L.) G. Don. Acta Physiologiae Plantarum, 33, 473–480. Barthe, L., Ribet, J. P., Pélissou, M., Degude, M. J., Fahy, J., & Duflos, A. (2002). Optimisation of the separation of Vinca alkaloids by nonaqueous capillary electrophoresis. Journal of Chromatography A, 968, 241–250. Bashir, M., Hafeez, M., Perveen, N., Fatima, J., & Mistry, A. N. (1983). A rapid procedure for the isolation of catharanthine, vindoline and vinblastine. Planta Medica, 47, 246–247. Chen, Q., Zhang, Y., Zhang, W., & Chen, Z. (2011a). Identification and quantification of oleanolic acid and ursolic acid in Chinese herbs by liquid chromatography– ion trap mass spectrometry. Biomedical Chromatography, 25, 1381–1388. Chen, Q., Li, N., Zhang, W., Chen, J., & Chen, Z. (2011b). Simultaneous determination of vinblastine and its monomeric precursors vindoline and catharanthine in Catharanthus roseus by capillary electrophoresis–mass spectrometry. Journal of Separation Science, 34, 2885–2892. Corona, G., Casetta, B., Sandron, S., Vaccher, E., & Toffoli, G. (2008). Rapid and sensitive analysis of vincristine in human plasma using on-line extraction combined with liquid chromatography/tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 22, 519–525. Creasey, W. A. (1969). Biochemical effects of the Vinca alkaloids. IV. Vinleurosine. Biochemical Pharmacology, 18, 227–232. Favretto, D., Piovan, A., & Filippini, R. (2001). Monitoring the production yields of vincristine and vinblastine in Catharanthus roseus from somatic embryogenesis. Semiquantitative determination by flow-injection electrospray ionization mass spectrometry. Rapid Communications in Mass Spectrometry, 15, 364–369. Ferreres, F., Pereira, D. M., Valent, P., Oliveira, J. M. A., Faria, J., Gaspar, L., et al. (2010a). Simple and reproducible HPLC–DAD–ESI–MS/MS analysis of alkaloids in Catharanthus roseus roots. Journal of Pharmaceutical and Biomedical Analysis, 51, 65–69. Ferreres, F., Pereira, D. M., Valentao, P., & Andrade, P. B. (2010b). First report of noncoloured flavonoids in Echium plantagineum bee pollen: Differentiation of isomers by liquid chromatography/ion trap mass spectrometry. Rapid Communications in Mass Spectrometry, 24, 801–806. Huang, L. Y., Chen, T. W., Ye, Z., & Chen, G. N. (2007). Use of liquid chromatography– atmospheric pressure chemical ionization–ion trap mass spectrometry for identification of oleanolic acid and ursolic acid in Anoectochilus roxburghii (wall.) Lindl. Journal of Mass Spectrometry, 42, 910–917. Ishikawa, H., Colby, D. A., & Boger, D. L. (2008). Direct coupling of catharanthine and vindoline to provide vinblastine: Total synthesis of (+)- and ent-( )-vinblastine. Journal of the American Chemical Society, 130, 420–423.
4. Conclusions This method provides simple, clean and reliable analysis of alkaloids from extracts in C. roseus by LC–IT-MS. Another benefit is that LC–IT-MS data permit unambiguous identification of the 19 alkaloids of extracts in C. roseus. Five reference alkaloids were directly-injected and analysed by ESI–MS1–2 in positive-ion mode. According to the ion peak intensity, we can choose precursor ions and product ions for quantification for analytzes. In addition, two analytical methods (LC–UV and LC–MS) have been developed for the determination of the five alkaloids in C. roseus. Comparing with the two methods, MS can supply a high selectivity and sensitivity
Acknowledgements This work was supported by the National Scientific Foundation of China (Nos. 90817103, 20775055, 30973672), the Important National Science and Technology Specific Projects (No. 2009ZX0930114) and the Fundamental Research Funds for the Central Universities. References
852
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
Izumi, Y., Okazawa, A., Bamba, T., Kobayashi, A., & Fukusaki, E. (2009). Development of a method for comprehensive and quantitative analysis of plant hormones by highly sensitive nanoflow liquid chromatography–electrospray ionization-ion trap mass spectrometry. Analytica Chimica Acta, 648, 215–225. Javanovics, K., Bittner, E., Dezseri, E., Eles, J., & Szasz, K. (1971). Selective manufacture of antitumorous vinblastine, vinleurosine, and vincristine from Vinca rosea. Ger. Offen. 18. Klausen, K., Mortensen, A. G., Laursen, B., Haselmann, K. F., Jespersen, B. M., & Fomsgaard, I. S. (2010). Phenolic compounds in different barley varieties: Identification by tandem mass spectrometry (Q-Star) and NMR; quantification by liquid chromatography triple quadrupole-linear ion trap mass spectrometry (Q-Trap). Natural Product Communications, 5, 407–414. Nesbitt, C. A., Zhang, H. X., & Yeung, K. K. C. (2008). Recent applications of capillary electrophoresis–mass spectrometry (CE–MS): CE performing functions beyond separation. Analytica Chimica Acta, 627, 3–24. Pereira, D. M., Ferreres, F., Oliveira, J. M. A., Gaspar, L., Faria, J., Valentao, P., et al. (2010). Pharmacological effects of Catharanthus roseus root alkaloids in acetylcholinesterase inhibition and cholinergic neurotransmission. Phytomedicine, 17, 646–652. Robert, H., Denise, I. J., Wim, S., Didier, H., & Robert, V. (2004). The catharanthus alkaloids: Pharmacognosy and biotechnology. Current Medicinal Chemistry, 11, 607–628.
Singh, D. V., Maithy, A., Verma, R. K., Gupta, M. M., & Sushil, K. (2000). Simultaneous determination of catharanthus alkaloids using reversed phase high performance liquid chromatography. Journal Liquid Chromatography & Relate Technology, 23, 601–607. Tikhomiroff, C., & Jolicoeur, M. (2002). Screening of Catharanthus roseus secondary metabolites by high-performance liquid chromatography. Journal of Chromatography, A, 955, 87–93. Verma, A., Hartonen, K., & Riekkola, M. L. (2008). Optimisation of supercritical fluid extraction of indole alkaloids from Catharanthus roseus using experimental design methodology – Comparison with other extraction techniques. Phytochemical Analysis, 19, 52–63. Zhao, X. J., Tang, Z. H., Guo, X. R., Zu, Y. G., Jiao, Y., Sun, Y. F., et al. (2006). Determination of 3 endogenous hormones by HPLC using solid phase extraction. Chinese Journal of Chromatography, 24, 534. Zhou, H., Tai, Y., Sun, C. R., & Pan, Y. J. (2005). Rapid identification of vinca alkaloids by direct-injection electrospray ionisation tandem mass spectrometry and confirmation by high-performance liquid chromatography–mass spectrometry. Phytochemical Analysis, 16, 328–333. Zhu, Q. H., Huang, D. D., Li, L., & Yin, Y. G. (2010). Synthesis of molecularly imprinted polymers for the application of selective clean-up vinblastine from Catharanthus roseus extract. Science China Chemistry, 53, 2587–2592.
Contents lists available at SciVerse ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
Identification and quantification of active alkaloids in Catharanthus roseus by liquid chromatography–ion trap mass spectrometry Qinhua Chen a,b, Wenpeng Zhang a, Yulin Zhang a, Jing Chen a, Zilin Chen a,⇑ a Key Laboratory of Combinatorial Biosynthesis and Drug Discovery (Wuhan University), Ministry of Education, and Wuhan University School of Pharmaceutical Sciences, Wuhan 430071, China b Affiliated Dongfeng Hospital, Hubei University of Medicine, Hubei 442008, China
a r t i c l e
i n f o
Article history: Received 16 January 2012 Received in revised form 31 October 2012 Accepted 28 January 2013 Available online 9 February 2013 Keywords: Alkaloids Identification Quantification Liquid chromatography–ion trap mass spectrometry
a b s t r a c t Catharanthus roseus is an important dicotyledonous medicinal plant that produces anticancer compounds. The active alkaloids vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were identified by direct-injection ion trap-mass spectrometry (IT-MS) for collecting MS1–2 spectra. The determinations of five alkaloids were accomplished by liquid chromatography (LC) with UV and MS detections. The analytes provided good signals corresponding to the protonated molecular ions [M+H]+ and product ions. The precursor ions and product ions for quantification of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were m/z 825 ? 807, 457 ? 397, 353 ? 144, 337 ? 144 and 809 ? 748 by LC–IT-MS, respectively. Two methods were used to evaluate a number of validation characteristics (repeatability, LOD, calibration range, and recovery). MS provided a high selectivity and sensitivity for determination of five alkaloids in positive mode. After optimisation of the methods, separation, identification and quantification of the five components in C. roseus were comprehensively accomplished by HPLC with UV and MS detection. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Catharanthus roseus L (G.) Don, commonly known as Madagascar periwinkle, belongs to the Apocynaceae family. It is known world-wide as a medicinal plant used for production of antitumor, hypotensive, and antiarrhythmic agents (Chen, Li, Zhang, Chen, & Chen, 2011b; Pereira et al., 2010). The major anticancer alkaloids in C. roseus are vinblastine, vincristine, vinleurosine, and vinorelbine. Vinblastine and vinleurosine are indole alkaloids that mainly exist in the aerial part of the plant. Vinblastine has been applied in the treatment of several kinds of cancer, such as Hodgkin’s disease, lymphosarcoma, choriocarcinoma, neuroblastoma, carcinoma of the breast, and chronic leukaemia (Aslam, Mujib, & Sharma, 2011; Javanovics, Bittner, Dezseri, Eles, & Szasz, 1971; Robert, Denise, Wim, Didier, & Robert, 2004). Creasey reported that vinleurosine can be used for the treatment of ascitic tumor (Creasey, 1969). Vinblastine and vinleurosine are produced in vivo by the combination of vindoline and catharanthine, both of which are derived from the terpenoid indole alkaloid biosynthetic intermediate strictosidine through multiple steps. Effective synthetic approaches have been developed by directly coupling catharanthine and vindoline to produce vinblastine (Ferreres
⇑ Corresponding author. Tel.: +86 27 68759893; fax: +86 27 68759850. E-mail address: [email protected] (Z. Chen). 0308-8146/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2013.01.088
et al., 2010a; Ishikawa, Colby, & Boger, 2008). The anti-hypertensive ingredient is ajmalicine, which is used to combat heart arrhythmias and to improve blood circulation in the brain (Almagro, Perez, & Pedreno, 2011). As we know, the chemical complexity and unique bisindole alkaloid structure of the aforementioned molecules have so far hindered their laboratorial synthesis. The isolation of vinblastine and vinleurosine is laborious and costly, mainly due to their low concentrations in the plant and co-existence of a large number of other alkaloids (Bashir, Hafeez, Perveen, Fatima, & Mistry, 1983). Several papers have reported the separation, identification and quantification of active compounds in C. roseus (Barthe et al., 2002; Tikhomiroff & Jolicoeur, 2002; Verma, Hartonen, & Riekkola, 2008; Zhao et al., 2006; Zhu, Huang, Li, & Yin, 2010). The methods of isolation and purification were focused on liquid–liquid extraction (Tikhomiroff & Jolicoeur, 2002), solid phase extraction (Zhao et al., 2006), supercritical fluid extraction (SFE) (Verma et al., 2008), molecularly imprinted polymers (MIPs) extraction (Zhu et al., 2010), and so on. Gas chromatography (GC) is suitable for volatile compounds but requires laborious derivatization. However, GC is unsuitable to analyse the bisindole alkaloids due to their high melting point. Liquid chromatography (LC) and capillary electrophoresis (CE) allow direct analysis of non-volatile metabolites. Non-aqueous capillary electrophoresis (NACE) (Barthe et al., 2002) and CE–MS (Chen et al., 2011b) have been developed for the analysis of Vinca alkaloids. In previous work, we reported
846
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
simultaneous determination of vinblastine and its monomeric precursors by CE–IT-MS. As the injected volume was 2.7 nL, the limits of detection (LOD) by CE–MS were as low as 0.8, 1.0 and 1.0 g/mL for vinblastine, vindoline and catharanthine, respectively (Chen et al., 2011b). Compared with CE, HPLC can provide better analytical precision and higher sample loading capacity. HPLC coupled to UV (Singh, Maithy, Verma, Gupta, & Sushil, 2000; Tikhomiroff & Jolicoeur, 2002) and MS detection (Corona, Casetta, Sandron, Vaccher, & Toffoli, 2008; Ferreres, Pereira, Valentao, & Andrade, 2010b) has been applied in the determination of the active compounds in C. roseus. But, HPLC in combination with tandem mass spectrometry (MS) appears to be a more suitable technique for the screening of the active compounds in plant samples in terms of sensitivity and selectivity. In recent years, analytical techniques have made great advances. LC–MS and LC-tandem MS (LC–MS/MS) such as quadrupole time of flight (Q-TOF) (Klausen et al., 2010) and quadrupole-linear ion trap (Q-LIT) (Ferreres et al., 2010b) have played a crucial role in plant analysis. The application of ion trap (IT) mass spectrometers in the structure elucidation and in the identification of unknowns in complex matrices has been well established (Izumi, Okazawa, Bamba, Kobayashi, & Fukusaki, 2009). Moreover, several reports show that IT mass spectrometers can also be very useful for quantitative analysis (Huang, Chen, Ye, & Chen, 2007; Izumi et al., 2009). The five alkaloids can be studied in positive ion modes by IT-MS, due to their weak basicity from the azacyclo-group in the structure. More importantly, MS2 (MS MS) can afford much more confidence for compound identification and much lower noise level for quantification. The fragmentation map of a target analyte is useful in performing MS2 for either qualitative or quantitative analysis. In previous work, we developed a LC–IT-MS for identification and quantification of oleanolic acid and ursolic acid in nine Chinese herbs (Chen et al., 2011a). In the present work, we have identified the active alkaloids, ajmalicine (A), catharanthine (B), vindoline (C), vinblastine (D) and vinleurosine (E) (Fig. 1). Direct infusion was used to collect MS1 profile and targeted MS2 data of selected ions (five alkaloids) in positive-ion modes. Meanwhile, we developed the HPLC–UV and HPLC–MS/MS methods for simultaneous determination and
quantification of five alkaloids. Quantification of five components was assigned in positive-ion mode. The precursor ions and product ions for quantification of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were m/z 825 ? 807, 457 ? 397, 353 ? 144, 337 ? 144 and 809 ? 748 by LC–IT-MS, respectively. The application of the method on profiling of five alkaloids in C. roseus is also presented. 2. Experimental 2.1. Chemicals and materials C. roseus L (G.) Don was collected from Hainan Yueyang Biotech Co., Ltd (Hainan, China). The samples were crushed and dried in an oven at 50 °C for 24 h before being subjected to extraction procedures. Vinblastine, vindoline, catharanthine, and vinleurosine were also obtained from Hainan Yueyang Biotech Co., Ltd (Hainan, China). Ajmalicine were bought from the National Institute for the Pharmaceutical and Biological Products of China. HPLC grade methanol, acetonitrile and ammonium acetate were obtained from Wuhan Analytical Reagent Company (Wuhan, China). Deionised water was purified using a Milli-Q system (Millipore, Bedford, MA, USA); Helium (purity, 99.999%) and liquid nitrogen were obtained from Wuhan Analytical Instrument Factory (Wuhan, China). Other reagents used in the experiment were of analytical grade and from commercial sources. 2.2. Standard and sample preparation The standard stock solutions of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were prepared by dissolving 10 mg of each in 5 mL methanol to achieve a concentration of 2 mg/mL, and kept at 4 °C. The vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine stock solution were diluted with methanol to obtain calibration solutions ranging from 1 to 100, 4 to 400, 2 to 200, 2 to 200 and 1 to 100 lg/mL, respectively. The sample of C. roseus L (G.) Don (10 g of dried stems) was ground to a fine powder, and then 5 g ultrasonicated with 50 mL 3% sulphuric acid for 30 min. The mixture was filtered and the
Fig. 1. Chemical structures of ajmalicine (A), catharanthine (B), vindoline (C), vinblastine (D) and vinleurosine (E).
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
pH adjusted with 4.0 mL 3% ammonia water. Afterwards, the aqueous solution was partitioned with 20 mL dichloromethane, the organic layer removed (sometimes centrifugation was required to separate both phases), and the same procedure was repeated three times. The organic layers were combined and evaporated under vacuum, and the residue was dissolved with 5.0 mL methanol. Prior to injection, all samples were filtered with a 0.45 lm nylon membrane filter; each sample solution was analysed in triplicate. The sample solutions were stable for at least three weeks if stored at 4 °C (confirmed by re-assay). 2.3. MSn analysis of the reference compounds solution A LC/MSD Trap SL Plus spectrometer (Agilent Technologies, Waldbronn, Germany) equipped with electrospray ionization (ESI) interface and an ion trap mass analyser was applied to the MS and multistage mass spectrometry (MSn) analysis. System control and data analysis was provided by the Agilent LC Chemstation and by Bruker Daltonics Trap Control and QuantAnalysis. A syringe pump was used for the direct loop injections of reference compound working solutions, and the flow rate was set at 0.5 mL/h. The ESI source was used and operated in both positive and negative ion mode. Typical operating conditions were as follows: drying gas (N2) temperature of 350 °C, 5 L/min drying gas flow, 10 psi nebulizer gas (N2) pressure, and 4500 V of capillary voltage. Data were acquired with a smart target of 30,000 and a max accumulation time of 200 ms. First, full-scan MS spectra were obtained by scanning from 100 to 1000 m/z. MS2 acquisition of the most abundant ions in the full-scan MS mode was carried out. 2.4. Chromatography conditions and instrumentation HPLC analyses were performed on an 1100 HPLC instrument (Agilent Technologies, California, USA) equipped with a binary pump, a UV detector, an autosampler, and a column thermostat. Chromatographic separations were carried out on a DL-Cl8 column (5.0 lm, 250 4.6 mm, Japan) column at 25 °C. Elution was performed with a flow rate of 0.5 mL/min. Acetonitrile (A) and 10 mmoL ammonium acetate in water (B) were used as a mobile phase. Gradient elution was used as follows: 10% A (2 min), 10– 30% B (10 min), 30–45% B (30 min), 45–70% B (50 min), 70–80% B (65 min), 80–95% (70 min). After the running, the gradient was set back to 10% A and the system was allowed to equilibrate. The injection volume was 10 lL and the detection wavelength was 280 nm. For LC–MS analysis, the Agilent 1100 HPLC system was coupled on-line to a LC/MSD Trap SL Plus spectrometer (Agilent Corp, Waldbronn, Germany) equipped with electrospray ionization (ESI) source. The Auto MS operation parameters are described as follows: positive-ion mode (ESI+); nitrogen drying gas, 10 L/min; nebulizer, 40 psi; gas temperature, 350 °C; Compound stability, 80%; mass range, 100–1000 m/z. The precursor ions and product ions for quantification of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were m/z 825 ? 807, 457 ? 397, 353 ? 144, 337 ? 144 and 809 ? 748 by LC–IT-MS, respectively. 2.5. Validation and statistical analysis The limit of detection (LOD) and the limit of quantification (LOQ) were calculated using the S/N ratio. LOD and LOQ were calculated as the concentrations producing a recognisable peak with S/N were 3 and 10, respectively. All the determinations were carried out in triplicate, and the experimental results were expressed as means; standard deviations were also given whenever needed. Statistical analysis was
847
performed by Agilent Chemstation Data Analysis software (Agilent Corp, Waldbronn, Germany). 3. Results and discussion 3.1. Identification and MSn behaviour of the reference compounds Electrospray ionization (ESI) in positive mode was chosen for the analysis of five alkaloids. It has been reported that ESI shows the greatest sensitivity for alkaloids when compared with electron impact (EI) and atmospheric pressure chemical ionization (APCI) (Favretto, Piovan, & Filippini, 2001). Because all components have azacyclo-group in their structure, they are weak basic compounds. The weakly basic alkaloids can be ionised readily to obtain cations, which are necessary for the positive ESI responses due to the carried positive charge in a acidic solution with pHs lower above the pKas (Nesbitt, Zhang, & Yeung, 2008). Thus, the positive ion mode and ESI source were adopted for the analysis of the five alkaloids. Five reference compounds were dissolved in methanol and analysed by ESI–MSn in positive-ion mode. The protonated molecule ions of [M+H]+ and major fragment ions were observed in the full scan MS and MS2 spectra in positive ion mode. Fig. 2 shows the protonated molecular ions and major fragment ions spectra for active alkaloids, vinleurosine, vinblastine, catharanthine, vindoline and ajmalicine by the ESI–MS and ESI–MS2. As shown in Fig. 2A and F, the protonated molecular ion [M+H]+ m/z 809 exhibited the high intensity in MS spectra, and major fragment ions were [M H2O+H]+ m/z 791, [M C3H8O+H]+ m/z 748, 624, 527, 469, 353, 224 in MS2 spectra for vinleurosine. The product ion [M C3H8O+H]+ m/z 748 exhibited the highest intensity. Fig. 2B and I showed representative fragment ions were [MH2SO4+H]+ m/z 923, [M+H]+ m/z 825, [M H2O+H]+ m/z 807, [M C3H8O+H]+ m/z 765, 704 in MS and MS2 spectra. The product ion [M H2O+H]+ m/z 825 ? 807 exhibited the highest intensity, and was therefore selected as the quantification ion. The protonated molecular ions and major fragment ions spectra for vindoline, ajmalicine and catharanthine were shown in Fig. 2C–E. All protonated molecular ions exhibited the highest intensity, and the highest intensity of product ions were m/z 144, [M C3H8O+H]+ m/z 397 and m/z 144 for catharanthine, vindoline and ajmalicine, respectively. Thus, quantification of vinblastine, vindoline, ajmalicine, catharanthine, and vinleurosine were m/z 825 ? 807, 457 ? 397, 353 ? 144, 337 ? 144 and 809 ? 748 were chosen for LC–IT-MS, respectively. 3.2. Identification of alkaloids in C. roseus As ESI–IT-MS has proved to be a suitable tool for the identification of the active compounds, the separation and determination of constituents in the extract of C. roseus was performed by LC–ESI– IT-MS in positive mode. A typical total ion chromatogram (TIC) of the identified compounds with MS detection is displayed in Fig. 3. It was observed that several peaks were composed of two or three components, but most peaks had symmetric peak shape and good resolution. The identification of components in the extracts was obtained by mass spectra, retention times, molecular weights (MW), and relative contents reported in literatures (Bashir et al., 1983; Ferreres et al., 2010a; Ishikawa et al., 2008; Javanovics et al., 1971; Tikhomiroff & Jolicoeur, 2002; Zhou, Tai, Sun, & Pan, 2005). The peak identification of major active compounds in the extracts of C. roseus in Fig. 3 is summarised in Table 1. In general, the identified components corresponded to about half the total peak area in TIC chromatograms under positive mode. Compounds 12, 13, 14, 17, 18 were vinblastine, vindoline, catharanthine, ajmalicine and vinleurosine, and identified by the
848
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
Fig. 2. Full-scan ESI–MS spectra of vinleurosine (A), vinblastine (B), catharanthine (C), vindoline (D), and ajmalicine (E), and ESI–MS2 spectra of vinleurosine (F), vinblastine (G), catharanthine (H), vindoline (I) and ajmalicine (J) in positive ion mode by direct loop injecting in ion trap mass.
849
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
Fig. 3. A typical total ion chromatogram (TIC) of active alkaloids in C. roseus (A) and typical extraction ion chromatograms (EIC) of five standard analytes (B) and sample (C) by LC–MS in the positive ion mode. Peak identification: (1) vinblastine, (2) vindoline, (3) catharanthine, (4) ajmalicine, (5) vinleurosine.
Table 1 LC–MS data for the C. roseus. Compound
Ret. time
Name
Protonated molecule ion[M+H]
Major fragment ions (m/z)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
25.5 26.2 27.5 32.3 41.6 44.2 44.9 45.1 48.5 50.2 54.0 54.7 57.7 58.2 58.5 59.2 59.4 60.6 63.1
Sitsirikine Isositsirikine Lochrovicine Lochnerine 19-s-vindolinine Vindolinine Serpentine Serpentine isomer Perimivine Tabersonine Catharosine Vinblastine Vindoline Catharanthine Vincristine Catharine Ajmalicine Vinleurosine Vindolicine
355 355 339 325 337 337 349 349 368 337 385 825 457 337 825 823 353 809 925
340, 297, 274, 227 340, 321, 274, 227 679 [2M+H]+, 322, 300, 283, 227, 185, 163 274, 249, 227, 174, 149 311, 223, 194, 144 320, 308, 276 697 [2M+H]+, 337, 320, 296, 263, 227 697 [2M+H]+, 337, 320, 296, 263, 227 735, 340, 322, 249, 227 323, 305, 276 791[2M+H]+, 366, 353, 335, 324, 275 807, 765, 705 397, 362, 341, 188, 162 320, 173, 144 807, 765, 747, 723, 705, 687 767, 588, 337, 144 319, 291, 265, 249, 222, 210, 178,144 791, 749, 624,527, 469, 353, 224 805, 749, 588, 463, 383
retention times and mass spectra of reference compounds. Four compounds (5, 6, 10, 14) with identical protonated molecule ions of [M+H]+ m/z 337 are shown in Table 1. The major fragment ions of tabersonine were m/z 323, 305, 276, which were described by Ferreres et al. (2010a), and thus the compound 10 was identified as tabersonin and other isomers (19-s-vindolinine and vindolinine). The peak for 19-s-vindolinin occurred more rapidly than
vindolinine, so it was assumed compounds 12 and 13 are 19-s-vindolinine and vindolinine, respectively. Compounds 7 and 8 with identical protonated molecule ions of [M+H]+ m/z 349, dimer-molecule ions of [2M+H]+ m/z 697 were identified as serpentine and its isomer (Ferreres et al., 2010a; Zhou et al., 2005). Compounds 12 and 15 with identical [M+H]+ 231 m/z 825, compound 12 was identified as vinblastine and the other vincristine (Zhou et al., 2005). In
850
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
order to identify other compounds, their characteristics, retention time, protonated molecule ion, main fragment ion and relative content reported in the literature (Barthe et al., 2002; Tikhomiroff & Jolicoeur, 2002; Verma et al., 2008; Zhao et al., 2006; Zhu et al., 2010) were taken into account. We came to a conclusion that compounds 1, 2, 3, 4, 9, 11, 16, 19 were sitsirikine, lsositsirikine, lochrovicine, lochnerine, perimivine, catharosine, catharine and vindolicine, respectively. 3.3. Quantitative analysis by LC–UV and LC–MS 3.3.1. Optimisation of LC conditions For LC–UV–MS, elution was performed with a flow rate of 0.5 mL/min. Also, Flow rates of about 1 ml/min are often used with LC–UV systems to achieve good separation efficiency. By LC–MS, especially MS with ESI source, the flow rate is limited by the spray and evaporation capability of the instrument. A flow rate of 0.5 mL/ min was used in our instrument for the consideration of separation efficiency and ionisation capability. Acetonitrile (A) and 10 mmoL ammonium acetate in water (B) were used as a mobile phase. Gradient elution was used as follows: 10% A (2 min), 10–30% B (10 min), 30–45% B (30 min), 45–70% B (50 min), 70–80% B (65 min), 80–95% (70 min). With a sharp gradient, the base-line separation of five alkaloids could be realised; however, for analysis of real samples, a flat gradient was required for separation of most of compounds in the real sample. 3.3.2. Validation of the method Detector response (relative peak area) was linearly dependent on sample concentration over the described above range for
analytzes. The typical chromatograms of analytes with different detection methods are shown in Figs. 3 and 4. Linear calibration curves were obtained for standard analytes at different concentration levels. The characteristics of the calibration plots are summarised in Table 2. As it can be seen, the five compounds showed excellent correlation and sensitivity in all analytical methods – define excellent. The limits of detection (LODs) were determined by the signal-to-noise (S/N) ratio = 3. The LODs of vinblastine, vindoline, catharanthine, ajmalicine and vinleurosine were determined to be 100, 200, 50, 50, 200 ng/mL for LC–UV, and 5.0, 1.0, 1.0, 10.0, 20.0 ng/mL for LC–MS, respectively. In general, LODs of analytzes by LC–MS were lower than those obtained using LC–UV. The results are shown in Table 2. LC–MS provides a higher selectivity and sensitivity for determination of five alkaloids in positive mode. As compared with our previous work (Chen et al., 2011b), the amount of LOD by CE–IT/MS were 2.16, 0.27and 0.27 for vinblastine, vindoline and catharanthine, respectively. The LOD for vinblastine, vindoline and catharanthineby CE–IT/MS were 2.16, 0.27and 0.27, respectively. The LOD using CE–MS was 23, 37 and 37 times higher than that obtained with LC–IT/MS. Thus, the extent of limits of detection using CE–IT/MS increased sensitivity compared with both LC–IT/MS and LC–UV. Repeatability was investigated by sequentially injecting a series of six standards. Precision was studied by injecting six freshly prepared mixtures. RSD values for relative peak areas are shown in Table 2 using three analytical methods. As shown in Table 2, intra-day precisions were from 0.53% to 2.03% and from 2.55 to 7.80 for analytes by LC–UV and LC–MS, respectively. And inter-day precisions were from 1.03% to 2.95%, and from 4.21% to 8.08% for
Fig. 4. Typical chromatograms of five standard analytes (A) and sample (B) by LC–UV. Peak identification: (1) vinblastine, (2) vindoline, (3) catharanthine, (4) ajmalicine, (5) vinleurosine.
851
Q. Chen et al. / Food Chemistry 139 (2013) 845–852 Table 2 Method validation data and the contents of vinblastine, vindoline, catharanthine, ajmalicine and vinleurosine in C. roseus by LC–UV and LC–MS. Method
Compound
Calibration curve
UV
Vinblastine Vindoline Catharanthine Ajmalicine Vinleurosine
y = 11.6x y = 3.638x y = 23.61x y = 20.79x y = 3.740 x
18.14 7.111 18,82 17.65 4.533
MS
Vinblastine Vindoline Catharanthine Ajmalicine Vinleurosine
y = 2 107x + 9 107 y = 4 107x + 7 108 y = 2 107x + 3 108 y = 2 107x + 2 108 y = 4 106x + 3 107
R2
Line arrange (lg/mL)
Limits of detection (ng/mL)
Precision Intra-day (RSD%)
Inter-day (RSD%)
0.999 0.999 1.000 1.000 0.999
1.0–100 4.0–400 2.0–200 2.0–200 1.0–100
100 200 50 50 200
0.53 2.03 0.55 0.87 1.97
0.978 0.990 0.991 0.992 0.989
1.0–100 4.0–400 2.0–200 2.0–200 1.0–100
5.0 1.0 1.0 10.0 20.0
2.55 6.70 4.05 3.61 7.80
Recovery
Content (lg/g)
Mean (%)
RSD%
1.03 2.95 1.34 1.09 2.76
98.6 99.3 98.5 96.3 97.9
1.04 2.08 2.28 2.57 2.89
7.52 125 45.2 76.3 2.10
4.21 7.01 6.18 5.14 8.08
97.3 97.4 103.9 107.5 95.7
4.88 6.36 7.29 7.11 4.32
6.02 131 40.7 69.2 1.85
analytes by LC–UV and LC–MS, respectively. Although there was better repeatability by LC–UV than that LC–MS, techniques were comparable as estimated by the Student’s t-test at 95% confidence level, which indicated no significant differences between two methods. In order to evaluate the accuracy of proposed method, recovery was tested. Accurate amounts of mixed standards were added to powder sample, and then extracted as described in the preparation of sample and analysed. The recovery values were obtained by their peak areas from the calibration curves under the same conditions by LC–UV and LC–MS. As shown in Table 2, the recoveries of analytes were from 96.3% to 99.3% with RSD from 1.04% to 2.89%, and from 95.7% to 107.5% with RSD from 4.32% to 7.29% by LC– UV and LC–MS, respectively. The results indicate that the three methods are suitable for the sample analysis.
for determination of the analytzes in positive mode and UV can obtain an excellent repeatability for analysis of them in C. roseus. In conclusion, the two methods have a good linear, reproducibility, precision, accuracy and recovery, and could be used for quantitative analysis of the five active compounds in C. roseus.
3.3.3. Determination of alkaloids in C. roseus The content of five alkaloids in C. roseus was determined by triplicate injections of three samples. Typical LC–UV and LC–MS chromatograms are shown in Figs. 3B, and 4B. The contents are listed in Table 2. The results obtained by the two techniques were comparable as estimated by the Student’s t-test at 95% confidence level, which indicated no significant differences between two methods. Otherwise, the results obtained by the LC–MS, HPLC/UV and CE–MS in the literature (Chen et al., 2011b; Tikhomiroff & Jolicoeur, 2002; Zhao et al., 2006 and Favretto et al., 2001) were comparable as it was also estimated by the Student’s t-test at 95% confidence level, which indicated no significant difference between the methods. LC–MS provided better sensitivity than CE–MS, as well as better analytical precision and higher sample loading capacity. When comparing with LC–UV, LC–MS has superiority on the screening of the active compounds in plant samples in terms of sensitivity and structure elucidation. However, LC–UV often possessed better precision. Thus, the developed LC–UV and LC–MS methods are suitable for the determination of five active alkaloids in C. roseus.
Almagro, L., Perez, A. J. L., & Pedreno, M. A. (2011). New method to enhance ajmalicine production in Catharanthus roseus cell cultures based on the use of cyclodextrins. Biotechnology Letters, 33, 381–385. Aslam, J., Mujib, A., & Sharma, M. P. (2011). Influence of freezing and non-freezing temperature on somatic embryogenesis and vinblastine production in Catharanthus roseus (L.) G. Don. Acta Physiologiae Plantarum, 33, 473–480. Barthe, L., Ribet, J. P., Pélissou, M., Degude, M. J., Fahy, J., & Duflos, A. (2002). Optimisation of the separation of Vinca alkaloids by nonaqueous capillary electrophoresis. Journal of Chromatography A, 968, 241–250. Bashir, M., Hafeez, M., Perveen, N., Fatima, J., & Mistry, A. N. (1983). A rapid procedure for the isolation of catharanthine, vindoline and vinblastine. Planta Medica, 47, 246–247. Chen, Q., Zhang, Y., Zhang, W., & Chen, Z. (2011a). Identification and quantification of oleanolic acid and ursolic acid in Chinese herbs by liquid chromatography– ion trap mass spectrometry. Biomedical Chromatography, 25, 1381–1388. Chen, Q., Li, N., Zhang, W., Chen, J., & Chen, Z. (2011b). Simultaneous determination of vinblastine and its monomeric precursors vindoline and catharanthine in Catharanthus roseus by capillary electrophoresis–mass spectrometry. Journal of Separation Science, 34, 2885–2892. Corona, G., Casetta, B., Sandron, S., Vaccher, E., & Toffoli, G. (2008). Rapid and sensitive analysis of vincristine in human plasma using on-line extraction combined with liquid chromatography/tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 22, 519–525. Creasey, W. A. (1969). Biochemical effects of the Vinca alkaloids. IV. Vinleurosine. Biochemical Pharmacology, 18, 227–232. Favretto, D., Piovan, A., & Filippini, R. (2001). Monitoring the production yields of vincristine and vinblastine in Catharanthus roseus from somatic embryogenesis. Semiquantitative determination by flow-injection electrospray ionization mass spectrometry. Rapid Communications in Mass Spectrometry, 15, 364–369. Ferreres, F., Pereira, D. M., Valent, P., Oliveira, J. M. A., Faria, J., Gaspar, L., et al. (2010a). Simple and reproducible HPLC–DAD–ESI–MS/MS analysis of alkaloids in Catharanthus roseus roots. Journal of Pharmaceutical and Biomedical Analysis, 51, 65–69. Ferreres, F., Pereira, D. M., Valentao, P., & Andrade, P. B. (2010b). First report of noncoloured flavonoids in Echium plantagineum bee pollen: Differentiation of isomers by liquid chromatography/ion trap mass spectrometry. Rapid Communications in Mass Spectrometry, 24, 801–806. Huang, L. Y., Chen, T. W., Ye, Z., & Chen, G. N. (2007). Use of liquid chromatography– atmospheric pressure chemical ionization–ion trap mass spectrometry for identification of oleanolic acid and ursolic acid in Anoectochilus roxburghii (wall.) Lindl. Journal of Mass Spectrometry, 42, 910–917. Ishikawa, H., Colby, D. A., & Boger, D. L. (2008). Direct coupling of catharanthine and vindoline to provide vinblastine: Total synthesis of (+)- and ent-( )-vinblastine. Journal of the American Chemical Society, 130, 420–423.
4. Conclusions This method provides simple, clean and reliable analysis of alkaloids from extracts in C. roseus by LC–IT-MS. Another benefit is that LC–IT-MS data permit unambiguous identification of the 19 alkaloids of extracts in C. roseus. Five reference alkaloids were directly-injected and analysed by ESI–MS1–2 in positive-ion mode. According to the ion peak intensity, we can choose precursor ions and product ions for quantification for analytzes. In addition, two analytical methods (LC–UV and LC–MS) have been developed for the determination of the five alkaloids in C. roseus. Comparing with the two methods, MS can supply a high selectivity and sensitivity
Acknowledgements This work was supported by the National Scientific Foundation of China (Nos. 90817103, 20775055, 30973672), the Important National Science and Technology Specific Projects (No. 2009ZX0930114) and the Fundamental Research Funds for the Central Universities. References
852
Q. Chen et al. / Food Chemistry 139 (2013) 845–852
Izumi, Y., Okazawa, A., Bamba, T., Kobayashi, A., & Fukusaki, E. (2009). Development of a method for comprehensive and quantitative analysis of plant hormones by highly sensitive nanoflow liquid chromatography–electrospray ionization-ion trap mass spectrometry. Analytica Chimica Acta, 648, 215–225. Javanovics, K., Bittner, E., Dezseri, E., Eles, J., & Szasz, K. (1971). Selective manufacture of antitumorous vinblastine, vinleurosine, and vincristine from Vinca rosea. Ger. Offen. 18. Klausen, K., Mortensen, A. G., Laursen, B., Haselmann, K. F., Jespersen, B. M., & Fomsgaard, I. S. (2010). Phenolic compounds in different barley varieties: Identification by tandem mass spectrometry (Q-Star) and NMR; quantification by liquid chromatography triple quadrupole-linear ion trap mass spectrometry (Q-Trap). Natural Product Communications, 5, 407–414. Nesbitt, C. A., Zhang, H. X., & Yeung, K. K. C. (2008). Recent applications of capillary electrophoresis–mass spectrometry (CE–MS): CE performing functions beyond separation. Analytica Chimica Acta, 627, 3–24. Pereira, D. M., Ferreres, F., Oliveira, J. M. A., Gaspar, L., Faria, J., Valentao, P., et al. (2010). Pharmacological effects of Catharanthus roseus root alkaloids in acetylcholinesterase inhibition and cholinergic neurotransmission. Phytomedicine, 17, 646–652. Robert, H., Denise, I. J., Wim, S., Didier, H., & Robert, V. (2004). The catharanthus alkaloids: Pharmacognosy and biotechnology. Current Medicinal Chemistry, 11, 607–628.
Singh, D. V., Maithy, A., Verma, R. K., Gupta, M. M., & Sushil, K. (2000). Simultaneous determination of catharanthus alkaloids using reversed phase high performance liquid chromatography. Journal Liquid Chromatography & Relate Technology, 23, 601–607. Tikhomiroff, C., & Jolicoeur, M. (2002). Screening of Catharanthus roseus secondary metabolites by high-performance liquid chromatography. Journal of Chromatography, A, 955, 87–93. Verma, A., Hartonen, K., & Riekkola, M. L. (2008). Optimisation of supercritical fluid extraction of indole alkaloids from Catharanthus roseus using experimental design methodology – Comparison with other extraction techniques. Phytochemical Analysis, 19, 52–63. Zhao, X. J., Tang, Z. H., Guo, X. R., Zu, Y. G., Jiao, Y., Sun, Y. F., et al. (2006). Determination of 3 endogenous hormones by HPLC using solid phase extraction. Chinese Journal of Chromatography, 24, 534. Zhou, H., Tai, Y., Sun, C. R., & Pan, Y. J. (2005). Rapid identification of vinca alkaloids by direct-injection electrospray ionisation tandem mass spectrometry and confirmation by high-performance liquid chromatography–mass spectrometry. Phytochemical Analysis, 16, 328–333. Zhu, Q. H., Huang, D. D., Li, L., & Yin, Y. G. (2010). Synthesis of molecularly imprinted polymers for the application of selective clean-up vinblastine from Catharanthus roseus extract. Science China Chemistry, 53, 2587–2592.