J Sep Sci 2008, 31, 3683 – 3687 Cristina Moiteiro1 Helena Gaspar1 Ana I Rodrigues1 Jo¼o F Lopes1 Valdemar Carnide2 C Moiteiro et al 3683 Original Paper HPLC quantification of dye flavonoids in Reseda luteola L from Portugal INETI – Instituto Nacional de Engenharia, Tecnologia e Inova¼o I.P., DTIQ, Lisboa, Portugal Centro de Gentica e Biotecnologia, Instituto de Biotecnologia e Bioengenharia, UTAD – Universidade de Trs-osMontes e Alto Douro, Vila Real, Portugal A HPLC method was developed for the simultaneous identification of Reseda luteola L (weld) flavonoids and quantification of the main compounds responsible for the yellow color This method was applied to a large number of wild Portuguese weld to evaluate its potential application as dyestuff for textile factories, as a substitute for the synthetic dyes currently used Portuguese weld dyestuff content ranged between 1.04 and 5.87%, corresponding to a wide variation of the flavonoids amount (1.39 – 9.04%) Luteolin 49-O-glucoside was found for the first time in R luteola, but kaempferol, isorhamnetin, and their glycosides were not detected in the Portuguese specimens Keywords: Flavonoids / HPLC quantification / Luteolin / Natural yellow dye / Reseda luteola / Received: July 8, 2008; revised: August 10, 2008; accepted: August 10, 2008 DOI 10.1002/jssc.200800383 Introduction Until the 19th century, dyeing plants played an important role in the textile industry as natural sources of dyes, but with the development of synthetic dyes the use of these plants almost disappeared An increasing awareness of environmental issues, the production of some synthetic dyes from nonrenewable resources, as well as the toxicity of some synthetic dyes, led to an increasing demand for natural dyes [1] Reseda luteola L (weld or dyer's rocket) is a natural dyeing species from the Mediterranean and Macaronesian flora [2] This species is an example of a dyeing plant that was cultivated in Portugal, particularly in the regions of Douro (northern region) and Alentejo (southern region), as a natural yellow dye source for the textile industries, but that is no longer used Flavonoids are the chromophores responsible for the yellow dye in R luteola Since the isolation of luteolin (1) from weld in 1833 by Chevreul (cited in ref [3]), other flavonoids such as luteolin 39-O-glucoside (2), luteolin 7-Oglucoside (3), luteolin 39,7-O-diglucoside (4), apigenin (5), apigenin 7-O-glucoside (6), kaempferol 3-O-glucoside-7-Orhamnoside (7), and brassidin (isorhamnetin 3-O-glucoside-7-O-rhamnoside) (8) have been reported in this species (Fig 1) [3–5] Angelini et al [1] evaluated the agronomic characteristics of weld accessions from different countries, includCorrespondence: Dr Helena Gaspar, INETI – Instituto Nacional de Engenharia, Tecnologia e Inova¼o I.P., DTIQ, Estrada Pao Lumiar, Ed F, 1649-038 Lisboa, Portugal Email: helena.gaspar@ineti.pt Fax: +351-217168100 i 2008 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim ing Portugal (Botanic Garden, Porto), cultivated during years in Central Italy fields, in order to maximize dye yield through quantification of luteolin by HPLC There are also some other works concerning the flavonoid contents and their dyeing capacities [3–7], but there is no information on the productive potential of wild weld in Portugal This work comprises an estimate of dye content by wild R luteola specimens from different regions of Portugal (Fig 2) An HPLC method was developed to quantify the main flavonoids responsible for the yellow dye in weld: luteolin (1), luteolin 7-O-glucoside (3), luteolin 39,7O-diglucoside (4), and luteolin 49-O-glucoside (9) Additionally in the course of this study luteolin 39-O-glucoside (2) was identified Experimental 2.1 Plant materials R luteola L (107 samples) were collected during the flowering period (May/June 2005) at 38 different locations (in general three plants per site) from Portugal (mainly in the Northern region) Seeds were later collected and preserved at the genebank of UTAD The aerials parts of each plant and 10 randomly chosen roots of the 107 samples were dried protected from light at room temperature, ground in an IKA Werke MF10 Basic mill through a 0.5 mm screen width, weighed, and maintained at –158C until the extraction Each sample (500 mg) of powdered R luteola was extracted with 15 mL of methanol/water (8:2) by sonication at room temperature for 10 The extracts were www.jss-journal.com 3684 C Moiteiro et al J Sep Sci 2008, 31, 3683 – 3687 Figure Structures of flavonoids from weld (R luteola L.) tion of standard solutions and mobile phases were methanol (HPLC gradient grade, Riedel-de-Han); acetic acid (98% p.a., Merck); and ultrapure water (Milli-Q Academic equipment, Millipore) 2.3 Calibrations curves Methanol stock solutions of each flavonoid were prepared and diluted to an appropriate concentration range for calibration curves [luteolin 39,7-O-diglucoside (6.4– 200 lg/mL), luteolin 7-O-glucoside (36–600 lg/mL), luteolin 49-O-glucoside (5.6–200 lg/mL), and luteolin (6.4– 300 lg/mL)] The linearity of each calibration curve was determined in duplicate using five different concentrations The LODs for each flavonoid were determined by successive dilutions of standards solutions until the magnitude of the test peak was three times greater than the noise level [luteolin 39,7-O-diglucoside (1.2 lg/mL), luteolin 7-O-glucoside (1.4 lg/mL), luteolin 49-O-glucoside (0.2 lg/mL), and luteolin (0.4 lg/mL)] The quantitative analyses were based on peak area calibration using each flavonoid as external standard 2.4 Instruments and chromatographic conditions 2.4.1 HPLC-DAD Figure Collecting sites of R luteola samples then filtered (Minisart RC15 0.45 mm, Sartorius), diluted (1/15) with methanol/water (8:2), and analyzed by HPLCUV/DAD (diode array detector) (Dionex) For assays each sample was prepared in triplicate and injected twice 2.2 Solvents and chemicals The flavonoid standards, luteolin 39,7-O-diglucoside, luteolin 7-O-glucoside, luteolin 49-O-glucoside, and luteolin, were purchased from Extrasynthse (France) and had a purity greater than 98% Solvents used for the prepara- i 2008 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim The flavonoids analysis was carried out on a Dionex HPLC system equipped with a quaternary pump (P580), an auto-sampler (GINA50), and a photo-DAD UVD170S/ 340S, set to a wavelength range of 200–600 nm The flavonoids separation was obtained using an analytical RP18e column Purospher STAR (250 mm64.5 mm id, lm; Merck) fitted with a guard column and with an injecting loop of 250 lL The column temperature was maintained at 308C using an HPLC oven (Cluzeau Infolabo) The mobile phase consisted of a gradient elution of methanol (solvent A) and water with 0.1% acetic acid (solvent B), with a flow rate of mL/min All mobile phase solutions were filtered and degassed before use The UV detection was set at 350 nm The gradient program was used as follows: the methanol concentration started at 35% and was ramped to 38% over 10 and held constant for 27 min; ramped to 60% during followed www.jss-journal.com J Sep Sci 2008, 31, 3683 – 3687 Liquid Chromatography 3685 by an isocratic for min; ramped to 100% over and maintained for A postrun program was used, with an injection of 250 lL of methanol for cleaning and reconditioning the column The gradient postrun program was as follows: 100% of methanol during then ramped to 35% over and maintained for These two programs resulted in a total run time of 72 The injection volume for all plant samples and flavonoid standards solutions was 20 lL 2.4.2 HPLC-MSn The flavonoid identifications were performed on a Waters Alliance 2695 LC Separation Module equipped with a photo-DAD (Waters 996), and 3-D IT mass spectrometer (Bruker Esquire 3000) with an orthogonal ESI source The LC-MS flavonoids separation conditions used in these analyses were analogous to those for HPLC-DAD described above The ESI-MS operated in negative ion mode under the following conditions: dry temperature 3658C, dry gas (N2) flow 10 L/min, capillary voltage 60 V, cone voltage –109.1 V, gas (N2) nebuliser flow 50 psi, scan range m/z 100–1000 MS2 acquisition was done in auto mode with isolation amplitude of 6.0 m/z in the range of m/z 15–1000, selecting the two most abundant ions as precursors with fragment amplitude of 2.0 V 2.5 Isolation and characterization of luteolin 39-O-glucoside (2) The R luteola methanol/water extract, prepared according to the method described in Section 2.1, was purified by RP-HPLC on a C-18 Luna-Phenomenex column (150 mm64.6 mm id, lm; Merck; flow rate 0.8 mL/ min) with the same gradient described in Section 2.4.1 to yield pure flavonoid luteolin 39-O-glucoside (1 mg; tR = 42.7 min), which was characterized by UV, NMR, and ESI-MS The NMR spectra were recorded on a Bruker AMX 300 MHz NMR spectrometer in CDCl3 TMS was used as an internal standard and chemical shifts are given as d (ppm), and the coupling constants (J) are reported as Hz Luteolin 39-O-glucoside: yellow powder; UV kmax (MeOH) 268, 345 nm; 1H NMR (MeOD) d 7.91 (1H, s, H-29), 7.63 (1H, d, J = 8.4 Hz, H-69), 7.01 (1H, d, J = 8.4 Hz, H-59), 6.67 (1H, s, H-3), 6.53 (1H, s, H-8), 6.22 (1H, s, H-6), 4.65 (1H, s, H-199), 4.02 (1H, d, J = 12.0 Hz, H-699a), 3.77 (1H, dd, J = 12.0 and 6.5 Hz, H-699b), 3.57 (2H, m, H-299 and H-399), 3.42 (1H, d, J = 9.0 Hz, H-499); 13C NMR (MeOD) d 164.4 (C-2), 103.8 (C-3), 182.5 (C-4), 161.8 (C-5), 99.4 (C-6), 165.0 (C-7), 93.3 (C-8), 158.0 (C-9), 103.3 (C-10), 123.0 (C-19), 115.9 (C-29), 146.0 (C39), 151.6 (C-49), 117.0 (C-59), 122.2 (C-69), 102.1 (C-199), 73.0 (C-299), 74.0 (C-399), 70.6 (C-499), 75.7 (C-599), 61.3 (C-699); ESI-MS (negative mode) m/z 447 [M – H]–; 284 [M – H – 162]– i 2008 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Figure HPLC-UV/DAD chromatogram (k = 350 nm) of R luteola extract Column: RP-18e Purospher STAR Methanol/water gradient used: see Section 2.4.1 Flow rate: 0.8 mL/min Oven temperature: 308C Flavonoid retention times (tr): (4) 11.9 min, (3) 16.8 min, (9) 26.5 min, (6) 31.5 min, (2) 42.5 min, (1) 45.3 min, and (5) 49.6 min; t0 = 2.2 Results and discussion A large number of R luteola L samples (107) were collected during the flowering period at 38 specific sites from Portugal (Fig 2) to evaluate their dyestuff content by HPLC The HPLC method developed was a modification of a published method [3], which allowed the simultaneous identification of flavonoids (1–6 and 9) present in aerial parts of weld and the quantification of the main compounds responsible for yellow dye (1, 3, 6, 9) (Fig 3) Furthermore, luteolin 4'-O-glucoside (9) was found for the first time in R luteola Kaempferol and isorhamnetin flavonoids, as well as their glycosides and 8, previously found in R luteola, were not detected in all Portuguese samples analyzed Additionally, the HPLC analysis confirmed the absence of flavonoids in weld roots The flavonoids identifications were performed by comparison of HPLC retention times, UV and ESI-MSn spectra obtained for the compounds found in weld extracts, with those of standard flavonoids, under the same chromatographic conditions The luteolin 39-O-glucoside (2) was isolated by preparative HPLC, and its identification was performed by NMR and ESI-MSn Compound was obtained as a yellow solid and its ESI-MS spectrum showed an ion peak at m/z 447, and its fragment ion at m/z 284 (lost of 162 units), suggesting a monoglycoside with luteolin as aglycone and glucose as sugar Its 1H NMR spectrum confirms the presence of the luteolin skeleton with a glycoside bond C–O at the 39 position by showing the proton signals (H-59, H-69, H-6, H-8, and H-3) similar to luteolin together with the H-29 singlet signal with a downfield shift of 0.43 ppm, compared to that of the luteolin The presence of a sugar at the 49 position www.jss-journal.com 3686 C Moiteiro et al J Sep Sci 2008, 31, 3683 – 3687 Table Content of flavonoids and dyestuff in R luteola L from Portugal Flavonoid Range (%) € SD Mean (%) € SD Luteolin 39,7-O-diglucoside (4) Luteolin 7-O-glucoside (3) Luteolin 49-O-glucoside (9) Luteolin (1) Total flavonoid Dyestuffa) [traces–0.84] € 0.03 [0.80–6.64] € 0.23 [traces–1.36] € 0.04 [0.14–1.99] € 0.05 [1.39–9.04] € 0.12 [1.04–5.87] € 0.08 0.34 € 0.17 2.71 € 1.18 0.46 € 0.25 0.67 € 0.34 4.19 € 1.56 2.86 € 1.01 a) Dyestuff content was calculated by determining the total amount of free and glycosylated luteolin shows a similar downfield shift of the H-59 These results are in agreement with the NMR data described in the literature [8] The chromatographic profiles of all the Portuguese weld samples analyzed were similar, showing a good resolution of the flavonoids (Fig 3) There were no qualitative differences in samples: the same flavonoids were found in each one Quantitatively, the overall amount (mg per 100 mg of dry plant) of flavonoids in weld samples analyzed varied considerably (1.39–9.09%), implying different dye capacities For the four main flavonoids and the total flavonoid statistically significant differences (p a0.01) were observed between all the samples The content of R luteola main flavonoids and its dyestuff is depicted in Table Luteolin 7-O-glucoside was usually the major flavonoid present in the analyzed samples Among the 107 samples analyzed there is a significant variation in both dyestuff content (1.04–5.87%) and plant weight (dry matter, 3.70– 42.83 g) (Fig 4) The highest dyestuff content (5.87%) corresponds to a low plant weight (4.02 g), the highest dry matter (42.83 g) to a low dyestuff content (1.31%), and there is a negative correlation (r = –0.322) between the two parameters As R2 is very low (0.1038), it can be concluded that dry matter is not a good predictor of the dyestuff content In the literature R luteola dyestuff content ranges between 0.4 and 4.0% [3, 6, 7], in particularly the dye content of a Portuguese accession cultivated in Italy has a mean of 1.7% of luteolin [1] Since flavonoids, yellow weld dyestuff, are secondary metabolites their relative amounts depend upon the origin and phenolic stage of the plant, as well as climatic and soil conditions [1] However, the extraction methodology employed could explain the higher dye contents found in Portuguese weld samples The extraction procedure used in this work was selected in order to reduce harsh conditions and, consequently, minimize hydrolysis processes Therefore, sonication was used instead of the typical boiling extraction procedure: effective, but aggressive for the extract This modification allowed obtaining the real flavonoids profile in weld, with free flavonoids and its glycosylated aglycones, instead of the more common and i 2008 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Figure Dyestuff content (mg per 100 mg of dry plant) versus plant weight (dry matter) less authentic chromatographic profile, with mainly free flavonoids (often luteolin) and minor glycosylated aglycones, originated by the hydrolysis during the extraction The extraction procedure was also fast, energy saving, and effective Concluding remarks This improved methodology allowed the identification of all main flavonoids present in weld, together with its minor ones, with good resolution and sensibility The developed procedure also allowed attaining the authentic flavonoids profile, using a smoother extraction method that minimized the hydrolysis processes, leaving the glycosylated flavonoids intact The application of this methodology allowed the quantification of the total amount of luteolins (luteolin and its glycosides) in weld, by combining a fast and efficient extraction procedure with an improved HPLC analysis The method also made possible the evaluation of R luteola full dye capacity It will be interesting to evaluate the relationship between the dyestuff content and genotypes in order to increase the production of the plants which genotypes had high dyestuff content that will have potential interest for textile factories The market of natural dyes and their use by industry is increasing, in particular the one www.jss-journal.com J Sep Sci 2008, 31, 3683 – 3687 obtained from R luteola, and can be important in a near future, due in part to the EU agricultural and environmental policies The authors acknowledge Funda¼o para a CiÞncia e Tecnologia for financial support (POCI/AGR/56087/2004) The authors have declared no conflict of interests References Liquid Chromatography 3687 [2] Mabberley, D J., The Plant-Book: A Portable Dictionary of the Higher Plants, Cambridge University Press, Cambridge 1993, p 497 [3] Cristea, D., Bareuau, I., Vilarem, G., Dyes Pigments 2003, 57, 267 – 272 [4] Yualdashev, M P., Batirov, E., Malikov, V M., Yaldasheva, N P., Chem Nat Compd 1996, 32, 923 – 924 [5] Batirov, E K., Tadzhibaev, M M., Malikov, V M., Chem Nat Compd 1979, 15, 643 – 644 [6] Cerrato, A., De Santis, D., Moresi, M., J Sci Food Agric 2002, 82, 1189 – 1199 [7] Hartl, A., Vogl, C R., J Sust Agric 2003, 23, 17 – 39 [8] Bertrand, A., Morel, S., Lefoulon, F., Rolland, Y., Monsan, P., Remaud-Simeon, M., Carbohyd Res 2006, 341, 855 – 863 [1] Angelini, L., Bertoli, A., Rolandelli, S., Pistelli, L., Ind Crops Prod 2003, 17, 199 – 207 i 2008 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim www.jss-journal.com ... allowed attaining the authentic flavonoids profile, using a smoother extraction method that minimized the hydrolysis processes, leaving the glycosylated flavonoids intact The application of this... 11.9 min, (3) 16.8 min, (9) 26.5 min, (6) 31.5 min, (2) 42.5 min, (1) 45.3 min, and (5) 49.6 min; t0 = 2.2 Results and discussion A large number of R luteola L samples (107) were collected during... identification of flavonoids (1–6 and 9) present in aerial parts of weld and the quantification of the main compounds responsible for yellow dye (1, 3, 6, 9) (Fig 3) Furthermore, luteolin 4'-O-glucoside