Rapid and Non-Destructive Determination of the Echinacoside Content in Echinacea Roots by ATR-IR and NIR Spectroscopy

• Abstract
• Introduction
• Materials and Methods
• Plant material and chemicals
• Sample cleanup and HPLC reference analysis
• HPLC apparatus
• ATR-IR spectroscopic measurements
• NIR spectroscopic measurements
• Results and Discussion
• Acknowledgements
• References

Abstract

NIR reflection and ATR-IR spectroscopy methods are developed to determine the echinacoside content in roots of Echinacea angustifolia and Echinacea pallida. Based on the recorded spectra and the HPLC reference data, chemometrical analyses are performed using a partial least squares (PLS) algorithm. Generally, good calibration statistics are obtained for the prediction of the echinacoside content presenting comparatively high coefficients of determination (R²) and low root mean standard errors of cross validation (RMSECV). It is demonstrated that optimal predictions are possible when using a dispersive spectrometer covering the spectral range from 1100 to 2500 nm. In contrast to the time-consuming HPLC method, the described non-destructive measurements allow us to predict the echinacoside content already after an analysis time of approx. one minute. Both spectroscopic techniques presented in this paper are shown to be useful in agricultural practice as well as in the phytopharmaceutical industry.

Introduction

Echinacea species, primarily E. purpurea (L.) Moench, E. angustifolia DC., and E. pallida (Nutt.) are widely used in traditional medicine both in North America and Europe for various pharmaceutical purposes such as immuno-stimulation and wound healing. Several studies have been performed in order to determine the main active principles of the drug but up to now it is still uncertain as to which extent the individual components contribute to the described pharmacological effects. According to the present knowledge, the immuno-stimulatory properties of Echinacea species are mainly attributed to alkamides, caffeic acid derivatives (chicoric acid and echinacoside), glycoproteins and polysaccharides which have been analysed applying various chromatographic separation techniques [1], [2], [3], [4], [5], [6], [7].

Whereas echinacoside occurs in dried roots of E. pallida and E. angustifolia in relatively high amounts of 1 - 3 g/100 g, no or only low echinacoside contents can be detected in E. purpurea root extracts [8]. Alcoholic extracts of the root of E. pallida and the pressed juice of the above ground plant material of E. purpurea at the time of flowering are the parts approved for use by the German Commission E. However no mention is made of E. angustifolia and also some other species belonging to the same genus such as E. paradoxa, E. atrorubens, E. sanguinea or E. simullata in the Commission E monographs.

Numerous in-vitro and in-vivo studies have been conducted on the pharmacological viability of these herbal preparations [9], [10]. Echinacea products are often sold as a combination of the three species mentioned above, but to date no reliable information exists regarding the individual medicinal value of these remedies. Currently, the US Pharmacopoeia is reviewing Echinacea species for inclusion in its botanical monograph series. At present most relevant root constituents such as echinacoside (Fig. [1]) are determined by HPLC [1]. However, this methodical approach includes time-consuming clean-up steps and requires sophisticated equipment as well as personnel. Therefore this study aimed to develop fast and reliable spectroscopic techniques for the determination of the echinacoside content in roots of various Echinacea species. In this context mid-infrared (IR) and near-infrared (NIR) spectroscopy methods were applied and the analytical results were compared with special reference to the practical use in agriculture (breeding and cultivation) and phytopharmaceutical industry. During the last years NIR spectroscopy became a very useful tool for fast and non-destructive measurements of valuable medicinal and aromatic plants [11], [12] in spite of the fact that interpretation of the broad and weak NIR absorptions can be only performed by suitable statistic algorithms such as partial least squares (PLS). Whereas first applications of ATR-IR were already performed more than 30 years ago [13], [14] the recent availability of diamond attenuated total reflection (ATR) technology for IR spectroscopy offers the advantage to analyse plant material directly without the necessity to perform any clean-up steps. Additionally, the concept was to supply those analytical methods which in principle can be used also for on-site measurements.

Materials and Methods

Plant material and chemicals

One part of the Echinacea plants (roots of E. purpurea (L.) Moench and E. angustifolia DC.) was cultivated in the Institute for Plant Genetics and Crop Plant Research (IPK) at Gatersleben (Germany). The other samples were provided by various plant business firms in Germany (e. g., Martin Bauer, Vestenbergsgreuth, Germany) and in the USA (Strategic Sourcing Inc., Wyomissing, USA). The echinacoside content of the individual root samples (registration numbers: NIR1 - NIR32, NIR34 - NIR40, NIR42 - NIR45, NIR47 - NIR133) covers the range between 0.1 g/100 g to 3.0 g/100 g dried roots. The pure echinacoside standard (art. no.: REC 00 401, HPLC content 98.4 g/100 g) was obtained from Phytolab GmbH, Labor Addipharma, Hamburg (Germany). The other reagents were of analytical grade and used without further purification.

Sample cleanup and HPLC reference analysis

Approximately 1 g of the powdered drug was extracted with 100 mL methanol in a Soxhlet apparatus. The resulting extract was evaporated to dryness and the residue was taken up in 25 mL of the HPLC eluent (mixture of eluents A and B, 95 : 5 v/v). If necessary, the sample solution was diluted by addition of eluent mixture to reach the concentration level of the echinacoside calibration curve. After centrifugation at 3000 rpm an aliquot of the sample was injected. The separation was performed using a Luna 5 μ C18 column (250 × 4.6 mm i. d., Phenomenex). The column was kept at 22 °C using a column oven. Eluents were (A) 85 % o-phosphoric acid/water (1 : 1000 v/v) and (B) 85 % o-phosphoric acid/acetonitrile (1 : 1000 v/v). The gradient used was as follows:[]

Flow rate was 1 mL/min; UV-detection was set at 330 nm. Different calibration samples of echinacoside were prepared in the concentration range between 0.01 and 6 mg/100 mL methanol from a methanolic stock solution by addition of individual amounts of eluent A and B.

HPLC apparatus

Reference analyses were carried out on a Waters HPLC system consisting of two HPLC pumps (model W 515), an automatic liquid chromatographic injector W 717, a photo diode array detector W 996, a UV/Vis detector W 486 and a workstation with Millennium 32 software of Waters.

ATR-IR spectroscopic measurements

The mid-infrared analyses were carried out on a portable diamond ATR/FT-IR spectrometer (Resultec Analytical Equipment, Garbsen, Germany) in a single reflection configuration. Approximately 2 - 5 mg of the finely ground drug material (< 0.5 mm sieve) were placed on the surface of the diamond composite ATR crystal. An intimate contact between the sample and the sensing device was achieved by the use of a load-restraining pressure applicator. An over-pressure restraint and a visual display of the actually existing pressure allows the user to prevent the risk of over-pressurisation and to apply a reproducible amount of pressure from sample to sample. The instrument can be easily operated by a 12 V car battery and is fitted with a Michelson interferometer with fixed and moving corner-cube mirrors and a DLATGS detector. The wavenumber region used for the analysis was 650 - 3500 cm^-1 with a spectral resolution of 2 cm^-1.

Chemometric analyses of the ATR-IR spectra were performed using a commercial software programme (Opus/Quant 2.0, Bruker GmbH, Rheinstetten, Germany).

Statistical accuracy is described by the coefficient of determination (R²) and the overall error between modelled and reference values [root mean square error of cross validation (RMSECV) resulting from cross validation] using the leave-one-out procedure. The spectral data were transformed with first derivative processing and subsequently mean-centred. The whole IR wavenumber range from 400 to 3500 cm^-1 was used for the partial least-squares (PLS) calibration. The optimum number of PLS factors for each component was determined by application of the PRESS (predictive residual error sum of squares) calculation.

NIR spectroscopic measurements

The same sample set used for ATR-IR experiments was also measured both on a dispersive near-infrared NIR System 5000 (Foss Instruments Inc., Hamburg, Germany) and a diode array spectrometer (model ”Corona”) equipped with an InGaAs array (Carl Zeiss Jena GmbH, Jena, Germany). For dispersive measurements a few g of the powdered root sample were transferred into rectangular cups (51 × 64 mm, 11 mm depth), placed in a transport unit, moving the sample up and down at right-angles to the incident radiation. Each sample was measured twice with 32 scans each time. The same sample cup was applied for the NIR diode array studies; here the spectral results were averaged from 4 measurements carried out at different parts of the drug material. Development of appropriate chemometric methods were carried out with the commercial statistic programme WINISI (Infrasoft International Inc., Port Matilda, USA). The spectral data were pre-treated with standard or weighted multiplicative scatter correction (SWSC, WMSC) and were transformed individually with first or second derivative processing. The calibration programme was set up with the whole wavelength range (dispersive spectrometer: 1100 - 2500 nm with a data interval of 2 nm; diode array spectrometer: 950 - 1700 nm with a data interval of 6 nm) using the partial least squares (PLS) algorithm. All data in the calibration set were checked carefully to detect and eliminate outlier samples.

Results and Discussion

A typical IR spectrum of homogenised Echinacea root is shown in Fig. [2] (A) . It presents an intense broad absorption band at approx. 1020 cm^-1 which is superimposed by some narrow signals of lower intensity. These bands are attributed to ν (C-O-C) vibrational modes of various carbohydrates (e. g. mono-, di- and polysaccharides) which are the most abundant structural groups in these compounds. Furthermore, stretching vibrations of OH-groups occur as broad absorption bands centred at about 3 300 cm^-1, which correspond to the same plant substances. Additionally, several smaller absorptions are observed in the fingerprint area between 1 200 and 1 800 cm^-1. Due to the complex composition of the plant material, the spectral bands consist of composite contributions of various functional groups of carbohydrates, amino acids, proteins, phenolic compounds, etc. In spite of the fact that in comparison to Echinacea samples, the IR spectrum of echinacoside shows similar absorption bands there exist some characteristic ”key vibrational modes” for this molecule which can be identified as small superimposed bands of lower intensity (e. g., at 811 and 1 517 cm^-1). It was proved that these small absorption bands do not occur in E. purpurea root known to contain no detectable amounts of echinacoside. However, using only these selected wavenumbers for the chemometrical calculation, very low R² values were obtained. This finding is related to the fact that various spectral data of a broad wavenumber range (mainly the spectrum range between 700 and 1 200 cm^-1) contribute to the whole prediction quality although distinctive absorption bands of the analyte molecule cannot be clearly assigned.

Due to the low mechanical strength of ZnSe, a new composite diamond ATR system was developed for this purpose [15]. In this context the special benefits of diamond are its strength and hardness, its chemical inertness and the transparency throughout most of the IR region.

Based on the measured ATR-IR spectra, a PLS algorithm was applied to develop a chemometric equation for echinacoside. As to be seen from Table [1] and Fig. [3] good correlation statistics were found between reference HPLC and predicted spectral data. But special care must be taken that, prior to measurement, the samples are sufficiently homogenised in order to obtain reliable results.

The spectral data obtained either from the scanning monochromator (Foss) or the diode array spectrometer (Zeiss) also presented very high prediction quality. However, the chemometrical predictions based on the Foss instrument achieved even better chemometrical results in terms of R² and standard error of prediction than those obtained from the portable Zeiss spectrometer. This is mainly explained by the different measurement range and optical resolution of both spectrometer systems. Fig. [4] shows the NIRS calibration plot for echinacoside based on 130 ground root samples (E. angustifolia and E. pallida) combined in one common data set. A PLS algorithm was applied to develop individual chemometric equations using an optimum number of factors, determined by the PRESS calculation, which proved to be a good tool for separating spectral information from background noise. The number of loading vectors used in the PLS analysis was 6 and 10, respectively. The newly developed ATR-IR and NIR methods can be successfully applied not only during breeding (selection of suitable single plants with high echinacoside content) and cultivation (prediction of the optimal harvest time) but also for efficient quality control purposes in the phytopharmaceutical industry.

Acknowledgements

The authors would like to thank the Fachagentur für Nachwachsende Rohstoffe e. V. (FNR) in Gülzow, Germany for the financial support. Thanks are also due to Dr. Manfred Feustel (Resultec Analytic Equipment in Garbsen, Germany) for supplying the Travel-IR instrument.

References

• 1 Bauer R, Remiger P, Wagner H. Echinacea .  Dtsch Apoth Ztg. 1988;  128 174-80
  PubMedSearch in Google Scholar
• 2 Bauer R. Echinacea containing drugs - Effects and active constituents.  Z Ärztliche Fortbild. 1996;  90 111-5
  PubMedSearch in Google Scholar
• 3 Bauer R, Wagner H. Echinacea species as potential immunostimulatory drugs. In: Economic and Medicinal Research, Wagner H, Farnsworth NR, editors.  Academic Press Inc, San Diego, CA (USA),. 1991;  5 253-321
  PubMedSearch in Google Scholar
• 4 Mazza G, Cottrell T. Volatile components of roots, stems, leaves and flowers of Echinacea species.  J Agric Food Chem. 1999;  47 3081-5
  CrossrefPubMedSearch in Google Scholar
• 5 Lienert D, Anklam E, Panne U. Gas chromatography-mass spectral analysis of roots of Echinacea species and classification by multivariate data analysis.  Phytochem Anal. 1998;  9 88-98
  CrossrefPubMedSearch in Google Scholar
• 6 Facino R M, Carini M, Aldini G, Saibene L, Pietta P, Mauri P. Echinacoside and caffeoyl conjugates protect collagen from echinacea extracts in the prevention of skin photodamage.  Planta Med. 1995;  61 510-4
  PubMedSearch in Google Scholar
• 7 Anonymus Monograph-draft ”Pale coneflower root” and ”Pale coneflower herb”. Pharmeuropa 2002 14: No. 1
  Search in Google Scholar
• 8 Hu C, Kitts D D. Studies on the antioxidant activity of Echinacea root extract.  J Agric Food Chem. 2000;  48 1466-72
  CrossrefPubMedSearch in Google Scholar
• 9 Bauer R, Puhlmann J, Wagner H. Immunologische In-vivo und In-vitro-Untersuchungen mit Echinacea-Extrakten.  Arzneimittelforschung (Drug Res.). 1988;  38 276-81
  PubMedSearch in Google Scholar
• 10 Blaschek W, Schütz M, Kraus J, Franz G. In vitro production of specific polysaccharides: isolation and structure of an antitumor active β-glucan from Phytophthora parasitica .  Food Hydrocolloids. 1987;  1 371-80
  PubMedSearch in Google Scholar
• 11 Schulz H, Steuer B, Krüger H, Schütze W. Möglichkeiten und Grenzen NIR-spektroskopischer Qualitätsbestimmung pflanzlicher Drogen.  Z Arzn Gew Pfl. 2001;  6 138-42
  PubMedSearch in Google Scholar
• 12 Schulz H, Steuer B, Krüger H. Rapid near infrared spectroscopic prediction of secondary metabolites in tea drugs and spice plants.  In: Proc. 9^th Int. Conference on NIRS in Verona. Davies AMC, Giangiacomo R, editors NIR Publications Chichester, West Sussex (UK),; 2000: 447-53
  Search in Google Scholar
• 13 Harrick N J. Internal reflection spectroscopy. Interscience Publ New York/London/Sidney,; 1967
  Search in Google Scholar
• 14 Mirabella F M, Harrick N J. Internal Reflection Spectroscopy: Review and Supplement. Harrick Scientific Corp Ossening, New York; 1985
  Search in Google Scholar
• 15 Coates J P, Sanders A. A universal sample handling system for FT-IR spectroscopy.  Spectroscopy Europe. 2000;  5 12-22
  PubMedSearch in Google Scholar

Prof. Dr. Hartwig Schulz

Federal Centre for Breeding Research on Cultivated Plants,

Institute for Plant Analysis, Neuer Weg 22 - 23,

06484 Quedlinburg, Germany

Email: H.Schulz@bafz.de

Fax: +49-3946/47-234

References

• 1 Bauer R, Remiger P, Wagner H. Echinacea .  Dtsch Apoth Ztg. 1988;  128 174-80
  PubMedSearch in Google Scholar
• 2 Bauer R. Echinacea containing drugs - Effects and active constituents.  Z Ärztliche Fortbild. 1996;  90 111-5
  PubMedSearch in Google Scholar
• 3 Bauer R, Wagner H. Echinacea species as potential immunostimulatory drugs. In: Economic and Medicinal Research, Wagner H, Farnsworth NR, editors.  Academic Press Inc, San Diego, CA (USA),. 1991;  5 253-321
  PubMedSearch in Google Scholar
• 4 Mazza G, Cottrell T. Volatile components of roots, stems, leaves and flowers of Echinacea species.  J Agric Food Chem. 1999;  47 3081-5
  CrossrefPubMedSearch in Google Scholar
• 5 Lienert D, Anklam E, Panne U. Gas chromatography-mass spectral analysis of roots of Echinacea species and classification by multivariate data analysis.  Phytochem Anal. 1998;  9 88-98
  CrossrefPubMedSearch in Google Scholar
• 6 Facino R M, Carini M, Aldini G, Saibene L, Pietta P, Mauri P. Echinacoside and caffeoyl conjugates protect collagen from echinacea extracts in the prevention of skin photodamage.  Planta Med. 1995;  61 510-4
  PubMedSearch in Google Scholar
• 7 Anonymus Monograph-draft ”Pale coneflower root” and ”Pale coneflower herb”. Pharmeuropa 2002 14: No. 1
  Search in Google Scholar
• 8 Hu C, Kitts D D. Studies on the antioxidant activity of Echinacea root extract.  J Agric Food Chem. 2000;  48 1466-72
  CrossrefPubMedSearch in Google Scholar
• 9 Bauer R, Puhlmann J, Wagner H. Immunologische In-vivo und In-vitro-Untersuchungen mit Echinacea-Extrakten.  Arzneimittelforschung (Drug Res.). 1988;  38 276-81
  PubMedSearch in Google Scholar
• 10 Blaschek W, Schütz M, Kraus J, Franz G. In vitro production of specific polysaccharides: isolation and structure of an antitumor active β-glucan from Phytophthora parasitica .  Food Hydrocolloids. 1987;  1 371-80
  PubMedSearch in Google Scholar
• 11 Schulz H, Steuer B, Krüger H, Schütze W. Möglichkeiten und Grenzen NIR-spektroskopischer Qualitätsbestimmung pflanzlicher Drogen.  Z Arzn Gew Pfl. 2001;  6 138-42
  PubMedSearch in Google Scholar
• 12 Schulz H, Steuer B, Krüger H. Rapid near infrared spectroscopic prediction of secondary metabolites in tea drugs and spice plants.  In: Proc. 9^th Int. Conference on NIRS in Verona. Davies AMC, Giangiacomo R, editors NIR Publications Chichester, West Sussex (UK),; 2000: 447-53
  Search in Google Scholar
• 13 Harrick N J. Internal reflection spectroscopy. Interscience Publ New York/London/Sidney,; 1967
  Search in Google Scholar
• 14 Mirabella F M, Harrick N J. Internal Reflection Spectroscopy: Review and Supplement. Harrick Scientific Corp Ossening, New York; 1985
  Search in Google Scholar
• 15 Coates J P, Sanders A. A universal sample handling system for FT-IR spectroscopy.  Spectroscopy Europe. 2000;  5 12-22
  PubMedSearch in Google Scholar

Prof. Dr. Hartwig Schulz

Federal Centre for Breeding Research on Cultivated Plants,

Institute for Plant Analysis, Neuer Weg 22 - 23,

06484 Quedlinburg, Germany

Email: H.Schulz@bafz.de

Fax: +49-3946/47-234