Quantification of selected furocoumarins by high-performance liquid chromatography and UV-detection: Capabilities and limits

Abstract

The performance of HPLC–UV as a means of quantifying selected furocoumarins in essential oils has been evaluated, based on a ring test validation approach. Accuracy profiles were generated, to determine bias and statistical confidence associated with determination at different concentrations, along with lower limits of quantification (LOQ). From these findings, it can be concluded that the method described may only be used in simple cases (essential oils), to measure individual furocoumarin compounds at concentrations greater than 10 mg/l; the non compound-specific nature of detection by absorption in the UV range is unable to overcome the effect of interferences arising from chromatographic coelutions, such as those encountered in the analysis of complex commercial fragrance mixtures. The use of an algorithmically calculated ‘spectral similarity’ function, with reference to authentic standards, may be used to improve reliability in assignment and quantification.

Introduction

Furocoumarins (FC) are bioactive constituents occurring naturally in various plants; some have been found to exhibit significant phototoxicity [1]. Among commonly used fragrance raw materials, are essential oils produced from such plants, notably those from citrus peels. Many citrus-derived oils are used in large quantities in the manufacture of finished fragrances.

While the FC content of steam-distilled oils is typically tiny (<0.02%), owing to the low volatility of FC, their concentration in cold-pressed oils may rise to significant levels (bergamottin: 10 to >20 g/l) [2]. In response to this, the fragrance industry, via its self-regulatory system of IFRA Standards [3] and other regulatory schemes (EU Cosmetics Directive [4], [5], ASEAN Cosmetics Directive [6] has followed restrictions on FC content in cosmetic products for a number of years. The current limitation of FC in the EU and ASEAN Cosmetic Regulation (entry 358 in Annex II – List of Substances which must not form part of the composition of cosmetic products) states that FC (e.g. xanthotoxin and bergapten) are prohibited, except when they derive directly from the natural essences used (the use of “essence” reflects the terminology used for “essential oil” in that part of the ASEAN directive, which refers to limitations on the content of furocoumarins in Cosmetic Products). In sun-protection and in bronzing products, those FC resulting from natural essences shall be below 1 mg/kg.

The Scientific Committee on Consumer Products (SCCP), an advisory body of the European Commission, published in 2005, as an outcome of another review of the data, an opinion expressing that a concentration greater than 1 ppm of any one of the FC in any finished cosmetic product would be of concern [1]. The European Commission has not yet taken a decision on whether there is a need to extend legislation and, if so, what is the most suitable way to further regulate these compounds beyond the existing directives. Aware of the SCCP opinion, several groups have been engaged in the development of quantitative analytical methods to detect and measure FC in essential oils.

Earlier literature references will not be considered hereafter, since they propose the analysis of only a few FC, whereas the objective of recent work has been the development of multi-analyte methods, allowing the quantification, in a single chromatographic run, of up to 15 FC. Although GC–MS has been proposed [7], it is not appropriate for all analytes under study: some are of low volatility and bergamottin is known to be unstable under typical GC inlet conditions [8]. Published methods typically involve analysis by HPLC followed by UV detection, this is a corollary of the presence, in FC, of a highly chromophoric polycyclic moiety exhibiting extended conjugation. FC separation is achieved either via normal phase [9] or, more commonly, reversed-phase chromatography [8], [10], [11], [12], [13], [14], [15], [16], [17]. The quantification wavelength of 310 nm seems to be the best compromise between sensitivity, linearity, and reproducibility [9], [12], [14], [16], although some papers report detection at other wavelengths: 248 nm [8], 250 nm [10], 254 nm [11], 315 nm [17], and 370 nm [15]. The limits of detection and quantification (LOD and LOQ) varied significantly in the literature. Two distinct LOD ranges were reported: 0.01–0.2 mg/l [8], [12], [14], [15], [17], and 1.7–15 mg/l [13], [16], Dugo et al. [15], and Russo et al. [17] reported LOQs in the range 0.019–0.062 mg/l; Govindarajan et al. [13] 3–5 mg/l and Lin et al. [14] and Vogl et al. [16] LOQs of 31–61 mg/l. It is noteworthy that values were obtained from the signal-to-noise ratio calculated from a blank injection, except the figures of Dugo's team, which were obtained from the injection of a sweet orange essential oil.

Given the variable and complex nature of natural products, it is prudent to consider possible interference from other matrix components, as reported by Peroutka et al. [8]. To detect if a coelution occurred in analysis of Heracleum candicans in one study, and in celery and parsnip in another, during the quantification of 4 and 5 FC respectively, two separate teams proposed the use of a peak purity test, in addition to the peak identification using the spectra of authentic analytes recorded under the same elution conditions [11], [13], but the occurrence frequency of such coelutions was not reported.

Although most papers report FC analyses without any pre-treatment of plant extracts prior to HPLC analysis, some authors subjected their extracts to a clean-up step using solid-phase extraction cartridges (SPE) [9], [11], or even to a combination of TLC, followed by SPE of selected TLC gel zones [10]. Interestingly, Prosen and Kocar [9] reports the recoveries obtained from different SPE sorbents, which do not exceed 65% for bergamottin. However, these pretreatment were applied to plant or fruit juice extracts, and are unnecessary for the analysis of essential oils [15].

Furocoumarins can be reliably separated by comprehensive liquid chromatography as demonstrated by Mondello's group, using a combination of a normal phase and a reverse phase as the first and the second dimensions, respectively [18]. Although the same group also showed that the quantification is feasible by comprehensive liquid chromatography (LC × LC) [19], no paper has been specifically dedicated to FC quantification, to our knowledge. The separation of FC from a lemon residue was also achieved by supercritical fluid chromatography, but, again, without quantification [20].

Several papers report the quantification of FC employing HPLC with subsequent detection by mass spectrometer: either with an ion trap [8], [12], [16], or a triple quadrupole and multiple-reaction monitoring [21]. The HPLC was interfaced to the MS via an electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) source. The LOD were in the range of 0.01–0.23 mg/l with ion trap MS and APCI interface [8], [12] and 0.0005–0.147 mg/l using multiple reaction monitoring with a triple-quadrupole MS and ESI or APCI interface [9], [21].

HPLC–UV remains the most popular quantification method for FC. However, the discrepancies between reported LOQs and the possible risk of interferences with other matrix constituents question its suitability. Therefore, the present work, based on the HLPC-UV method developed by Frérot and Decorzant [12], aims to determine its practical applicability, using a collaborative ring test.