Introduction

Since the 1980s, reverse-phase high performance liquid chromatography (RP-HPLC) has been commonly used for the separation, identification and quantification of large groups of polar and non-polar pigments. Normal-phase HPLC systems are best suited for separating only a few groups of pigments with similar polarities and structures (Gieskes and Kraay 1983). The idea to employ the RP-HPLC in environmental studies was introduced by Mantoura (Mantoura and Llewellyn 1983) and has been developed by many researchers (Bidigare et al. 1985; Wright et al. 1991; Wright and Jeffrey 1997; Heukelem and Thomas 2001; Marty et al. 2002; Rodriguez et al. 2002; Brotas and Plante-Cuny 2003). The resolution of compounds is based on adsorption and desorption processes on the coating of chromatographic columns (stationary phase) in the presence of the suitable solvents’ composition (mobile phase).

There are many disciplines of environmental studies in which qualitative and quantitative characteristics of pigments obtained by the use of chromatographic techniques have found applications and are indispensable. The use of taxonomically important chlorophylls and carotenoids in the direct identification of phytoplankton classes, their biomass, condition and stage of development in seawater samples has increased due to advancements in these techniques (Wright and Enden 2000; Mackey et al. 2002). The knowledge of photosynthetic and photoprotectant pigment concentrations facilitates the determination of the absorption properties of phytoplankton (Woźniak et al. 2000) and of the rate of photosynthesis in natural waters (MacIntyre et al. 2000) approximated from satellites (Woźniak et al. 1997; Claustre et al. 2004). Additionally, the radioactively labeled taxon-specific pigments can be used to distinguish primary producers in the field (Gieskes and Kraay 1989) and for the mathematical analysis of pigment fingerprints (Gieskes et al. 1988; Rodriguez et al. 2002).

Solvents and their time-variable composition, temperature-controlled systems (ensuring stable retention times) and different columns were tested during years of investigations. Octadecylo silane (ODS) C18, C8 and even C30 chain columns with different diameters, lengths and particle sizes have, with different degrees of effectiveness, found application in the analysis of pigments (Schmid and Stich 1995; Zapata et al. 2000; Barlow et al. 2002; Rodriguez et al. 2002). Octadecyl silica C18 columns (with polymeric, as well as monomeric coating) have been commonly used for the determination of algal pigments. At present, over 50 phytoplankton pigments and their degradation products can be separated in a single analysis (Wright et al. 1991; Wright and Jeffrey 1997). Polymeric C18 columns were used to separate carotenoid pairs, such as lutein and zeaxanthin, 19′hexanoyloxyfucoxanthin and 9′cis-neoxanthin, and α- and β-carotene (Heukelem et al. 1992). The employment of C8 columns allowed separation of divinyl and monovinyl chlorophyll a and b (Goericke and Repeta 1992). In addition, the incorporation of a pyridine-containing mobile phase to the HPLC pigment analysis, either in combination with polymeric C18 columns or monomeric C8 columns, allows separation of chlorophyll c pigments (Zapata et al. 2000, 2001). The pigment resolution may also be enhanced by adjusting the temperature of the column (Heukelem et al. 1994). Also the solvent composition can be modified, for example, by the use of highly ionic solvents with chlorophylls c1, c2 and c3 separated on the C18 column (Kraay et al. 1992).

The RP-HPLC technique is highly accurate (±6%, Latasa et al. 1996; Claustre et al. 2004) in routine pigment measurements, even for trace amounts of <0.1 μg dm−3 (Roy et al. 1996; Stoń et al. 2002). The qualification and quantification of individual chlorophylls and carotenoids is mainly based on their properties in the visible light range. The identification relies on comparisons of the observed diode array spectroscopy during elution, or by transferring HPLC fractions to standard solvents and comparing their visible absorption spectra with reference standards or literature values, or by co-chromatography with authentic standards (Wright et al. 1991). Quantitative characterization of eluted pigments is based on peak area and excitation coefficients published in the literature (Mantoura and Llewellyn 1983; Latasa et al. 1996; Jeffrey 1997a; Bjørland et al. 2003) and on Lambert–Beer’s law. Concentrations of pigments can also be calculated using internal (e.g. canthaxanthin) and external standards (Andersen et al. 1996; Mantoura and Repeta 1997; Obayashi and Tanoue 2002; Qian et al. 2003).

A limited number of high-purity pigment standards exists for the calibration of spectrometers and HPLC detectors. Due to difficulties resulting from a very labile structure of pigment molecules, susceptible to degradation by light, heat, oxygen, acids and bases, the preparation of pigment standards is very complicated. Some synthetic chlorophyll and carotenoid standards are available from commercial sources, e.g. chlorophyll a and b, canthaxanthin, lutein, lycopene, zeaxanthin, α-carotene and β-carotene. In most cases, laboratories prepare and purify their own standards from reference phytoplankton monocultures with well-known pigment composition, or from other sources, such as tomatoes, leaves of maize, or other higher plants. Detailed information, full description and guidelines for the preparation of 17 chlorophyll and 29 carotenoid standards, chemotaxonomically significant in phytoplankton ecology, from reference strains of monocultures were given by Jeffrey (1997b) and Repeta and Bjørland (1997). In general, the approach is the same: pigments are extracted from suitable biological sources by organic solvents, purified by thin-layer or semi-preparative liquid chromatography, crystallized and transferred to an appropriate solvent (90% acetone in the case of chlorophylls; ethanol, diethyl ether, hexane, or petroleum ether for carotenoids) and characterized by spectrophotometry.

The aim of this paper is to present the results of comparative analysis of chromatographic resolution by the use of two types of columns: LiChroCART Hypersil ODS and LiChroCART LiChrospher 100 RP18e. Two groups of high-purity pigment standards from The International Agency for 14C Determination, DHI Institute for Water and Environment, Denmark, were subjected to chromatographic analysis in order to calibrate these two HPLC systems. The external calibration results obtained (such as spectrum shapes in appropriate solvent compositions, retention times, slopes of calibration curves, sensitivity, or detection thresholds) provide possibilities to use interchangeable analytical columns for separation, isolation and quantification of the pigments present in environmental samples.

Materials and methods

Pigment standards

Two groups of high-purity certified chlorophyll and carotenoid standards, isolated from reference strains of phytoplankton monocultures, were obtained from The International Agency for 14C Determination, DHI Institute for Water and Environment, Denmark. The first group of 23 pigments consisted of chlorophylls (a, b, c, c2, c3) and carotenoids (alloxanthin, antheraxanthin, aphanizophyll, 19′butanoyloxyfucoxanthin, canthaxanthin, α- and β-carotenes, diadinoxanthin, echinenone, fucoxanthin, 19′haxanoyloxyfucoxanthin, lutein, myxoxanthophyll, neoxanthin, peridinin, prasinoxanthin, violaxanthin, zeaxanthin). The second group consisted of 25 pigments, chlorophylls a, b, c2 and c3, chlorophyllide a, divinyl chlorophyll a, pheophythin a, alloxanthin, antheraxanthin, 19′butanoyloxyfucoxanthin, canthaxanthin, α- and β-carotenes, diadinoxanthin, diatoxanthin, echinenone, fucoxanthin, 19′haxanoyloxyfucoxanthin, lutein, myxoxanthophyll, neoxanthin, peridinin, prasinoxanthin, violaxanthin and zeaxanthin. Pigments were supplied in 2.5 cm3 sealed vials: the extracts were in 90% acetone for chlorophylls and in 100% ethanol for carotenoids. The parameters of reference pigments are given in Table 1.

Table 1 Parameters [symbol, solvent, extinction coefficient (dm3 g−1 cm−1), concentration (mg dm−3) and source of origin] of reference pigment standards obtained from The International Agency for 14C Determination, DHI Institute for Water and Environment, Denmark (I, II the first and second group of chlorophylls and carotenoids, respectively)
Full size table

These compounds were subjected to chromatographic analysis. Pigment standards were injected onto a chromatographic system, calibration curves were prepared and many factors (such as the response factor fp and detection thresholds) were established.

Apparatus

Pigments were isolated using the RP-HPLC technique. The chromatographic system was equipped with a Hewlett-Packard HP 1050 pump, diode array detector (model HP 1100), an HP 1046 fluorescence detector and a Rheodyne injector with a 100 μl sample loop. Two types of C18 analytical columns: LiChroCART Hypersil ODS (dimension: 250×4 mm, particle size: 5 μm; Merck) and LiChroCART LiChrospher 100 RP18e (dimension: 250×4 mm, particle size: 5 μm; Merck) were used for comparative analysis of analytical column responses and efficiency.

The diode array absorbance detector “dad” was set at λ=440 nm. During a single analysis, the full spectral range of 350–700 nm, with steps of 1 nm and a 0.4 s time resolution, was recorded. The fluorescence detector with the excitation wavelength λex=431 nm and emission λem=660 nm was only used to confirm the presence of chloropigments in the extract.

Solvents and chromatographic analysis

The solvents used for chromatography were filtered and degassed with helium before use. Mobile phases used in the gradient elution were composed of the primary eluant (A), consisting of methanol and 1 M ammonium acetate (80:20 v/v), and the secondary eluant (B), prepared from methanol and acetone (60:40 v/v). In addition, 1 M ammonium acetate was used as the ion-pairing reagent, and it is recommended that it be present in both the sample and the mobile phase, to improve pigment separation and suppress the dissociation of the isolated compounds (Mantoura and Llewellyn 1983). The separation was achieved by changing the solvent mixture composition (Mantoura and Llewellyn 1983; Barlow et al. 1993; Stoń and Kosakowska 2002). The solvent composition was changed linearly from 100% of solvent A to 100% of solvent B in 10 min after injection, and was isocratically held at a constant flow rate of 0.8 ml min−1 till the end of the analysis (25 and 32 min for first and second type of column, respectively). The equilibrium state was attained after 10 min, when the solvent composition returned to the initial conditions.

Statistical analysis

The repeatability of the chromatograms and the method’s precision were also checked. The analyses were based on monocultures grown in laboratory conditions and using pigment standards. The statistical analysis was based on 36 repetitions of the chromatographic resolution of the extracts. The margin of error of the adopted method ranged from 1.28% to 5.55%, with a mean value of ±2.9%.

Results

Chromatographic resolution

The separation of chlorophylls and carotenoid pigment standards by the use of two types of columns yielded satisfactory results. Sufficient resolution of all reference pigments in a single run was achieved. Representative absorbance chromatograms of the pigment standard mixture are shown in Fig. 1. The results obtained indicate that chlorophylls c1 and c2 elute as a single peak. Likewise, lutein and zeaxanthin were not completely resolved in either type of analytical column. A comparison of the results obtained with pigments separated on different columns showed some differences in the elution times of individual components of the mixtures. Although the sequence of retained and eluted pigments remained mostly the same in the two systems, in the case of LiChroCART Hypersil ODS, the retention times of pigments were shorter than those found with LiChroCART LiChrospher 100 RP18e. The greatest differences in elution times were observed for α-carotene and β-carotene (~7 to 8 min). This delay may have resulted from the molecular structure of these two non-oxygenated compounds. As a consequence, the time of a single analysis was prolonged from 25 min for LiChroCART Hypersil ODS to 32 min for LiChroCART LiChrospher 100 RP18e. The values of the chromatographic parameters obtained for each pigment in the two columns are given in Table 2.

Fig. 1
figure 1

Chromatograms of resolved mixtures of phytoplankton pigment standards obtained using an RP-HPLC system equipped with two types of analytical columns: a LiChroCART Hypersil ODS and b LiChroCART LiChrospher 100 RP18e (1 chlorophyll c3; 2 chlorophyllide a; 3 chlorophyll c1+c2; 4 peridinin; 5 19′butanoyloxyfucoxanthin; 6 fucoxanthin; 7 19′haxanoyloxyfucoxanthin; 8 neoxanthin; 9 prasinoxanthin; 9a aphanizophyll; 10 violaxanthin; 11 diadinoxanthin; 12 antheraxanthin; 13 alloxanthin; 14 myxoxanthophyll; 15 diatoxanthin; 16 lutein; 17 zeaxanthin; 18 canthaxanthin; 19 chlorophyll b; 20 divinyl chlorophyll a; 21 chlorophyll a; 22 echinenone; 23 α-carotene; 24 β-carotene)

Full size image
Table 2 Combination of calibration parameters: retention time (min), detection thresholds (ng per 100 μl), response factors fp (ng mAU−1 s−1), correlation coefficients and wavelength of maximum absorption λmax (nm)—of individual standard pigments obtained by RP-HPLC equipped with different analytical columns (I LiChroCART Hypersil ODS; II LiChroCART LiChrospher 100 RP18e)
Full size table

Purity of pigments

The purity of chlorophyll and carotenoid standards (>95%) was confirmed by an analytical chem-station procedure. This procedure allows assessment of whether the eluting peak is spectrally pure, or contains impurities due to simultaneous retention of several compounds, or incomplete separation of peaks, and/or background absorption The peak shape and purity curves are shown in Fig. 2 for fucoxanthin.

Fig. 2
figure 2

Confirmation of the peak purity on the example of fucoxanthin: a peak shape and b peak signals window showing similarity and threshold curves (1 peak signal plots; 2 peak start; 3 initial reference spectrum; 4 similarity curve; 5 position of selected peak spectra; 6 peak end; 7 final reference spectrum; 8 threshold curve; 9 spectral tick marks)

Full size image

Spectral properties of chlorophylls and carotenoids

Identification and quantification of pigments by HPLC is based on the characteristic absorption properties of these compounds in the visible spectral range. Pigments differ from each other in the shape and position of their absorption maxima. Carotenoids absorb light in the range 400–500 nm, displaying two or three distinct absorption maxima in this region. Chlorophylls additionally absorb in the red part of the light spectrum. The ability of carotenoids to absorb light is due to their system of regularly alternating single and double bonds (a chromophore). The chromophore structure in algal carotenoids consists of 9–11 aliphatic double bonds in conjunction with no, one, or two endocyclic double bonds. Spectral properties of carotenoids are not influenced by functional groups. Under given conditions (solvent, temperature) the spectrum is a characteristic feature of the absorbing molecules, and is therefore used not only in quantitative measurements but also to confirm the identity of the pigment.

Results of the comparative analysis of the spectra of two groups of reference pigments are presented in Fig. 3. The wavelengths of the absorption maxima differed between pigments and were generally shifted by about 1–5 nm relative to the same pigments separated on the other type of column. The spectra overlap only in the cases of neoxanthin and alloxanthin (Fig. 3, 9 and 15). Regarding other pigments, the shifts in the absorption maxima result from different retention times and, hence, different compositions of solvents during elution. The respective values of λmax of individual pigments are listed in Table 2.

Fig. 3
figure 3

Comparison of on-line spectrum shapes of pigments obtained during elution by diode-array detector. Continuous lines represent the first group of reference pigments isolated by LiChroCART Hypersil ODS; broken lines represent the second group of standards isolated by the use of the LiChroCART LiChrospher 100 RP18e column (1 chlorophyll c3; 2 chlorophyllide a; 3 chlorophyll c1; 4 chlorophyll c2; 5 peridinin; 6 19′but-fucoxanthin; 7 fucoxanthin; 8 19′hex-fucoxanthin; 9 neoxanthin; 10 prasinoxanthin; 11 violaxanthin; 12 aphanizophyll; 13 diadinoxanthin; 14 antheraxanthin; 15 alloxanthin; 16 myxoxanthophyll; 17 diatoxanthin; 18 lutein; 19 zeaxanthin; 20 canthaxanthin; 21 chlorophyll b; 22 divinyl chlorophyll a; 23 chlorophyll a; 24 echinenone; 25 pheophythin a; 26 α-carotene; 27 β-carotene)

Full size image

Calibration parameters

Calibration curves were prepared based on a series of injections of known concentrations of pigment standards on the chromatographic system. Six to ten different concentrations were injected for each pigment. Such a method of external standard calibration allows establishment of the relationships between the weight of the pigment injected on the chromatographic column and its peak area.

The resultant chromatographic peak areas are related to pigment masses by the response factor (or calibration curve). Calibration factors for each pigment, designated “response factors” (fp, ng mAU−1 s−1) are obtained. The value of parameter fp is representative of an individual pigment isolated from the pigment mixture. The fp values and correlation coefficients obtained by the use of two types of chromatographic columns are given in Table 2. The calibration curves for two groups of standards are shown in Fig. 4.

Fig. 4
figure 4

Comparison of calibration curves of individual pigments obtained by resolution on two different columns: LiChroCART Hypersil ODS (continuous lines) and LiChroCART LiChrospher 100 RP18e (broken lines) (1 chlorophyll c3; 2 chlorophyllide a; 3 chlorophyll c1; 4 chlorophyll c2; 5 peridinin; 6 19′but-fucoxanthin; 7 fucoxanthin; 8 19′hex-fucoxanthin; 9 neoxanthin; 10 prasinoxanthin; 11 violaxanthin; 12 aphanizophyll; 13 diadinoxanthin; 14 antheraxanthin; 15 alloxanthin; 16 myxoxanthophyll; 17 diatoxanthin; 18 lutein; 19 zeaxanthin; 20 canthaxanthin; 21 chlorophyll b; 22 divinyl chlorophyll a; 23 chlorophyll a; 24 echinenone; 25 pheophythin a; 26 α-carotene; 27 β-carotene)

Full size image

Discussion

Within the framework of the experiment presented, the separation ability of two C18 chromatographic columns (LiChroCART Hypersil ODS and LiChroCART LiChrospher 100 RP18e) was compared. Different pigment standards were subjected to chromatography in order to obtain calibration parameters. The standard group encompassed 8 chlorophylls and 19 different carotenoids, with different concentrations (Table 1). For example, the concentration of lutein was equal to 0.659 mg dm−3 in the first group, but in the second group was almost twice as high (1.314 mg dm−3). Pigments originated from reference phytoplankton classes. The role of marker pigments in identifying and quantifying individual species has been well recognized. For example, fucoxanthin is considered to be a marker of diatoms; zeaxanthin-cyanobacteria; 19′-hexanoyloxyfucoxanthin-prymnesiophytes; 19′butanoyloxyfucoxanthin-pelagophytes; alloxanthin-cryptophytes; prasinoxanthin-prasinophytes; peridinin-dinoflagellates; and chlorophyll b and lutein-chlorophytes (Andersen et al. 1996; Jeffrey and Vesk 1997).

The chromatographic system and procedure presented herein can be used to separate chlorophylls and their derivatives, as well as carotenes and their oxygenated forms, called xanthophylls. Only in the case of zeaxanthin and lutein were the resolution results unsatisfactory. Complete resolution of the above pigment pair was accomplished only by the use of a polymeric-packed chromatographic column (Heukelem et al. 1992; Schmid and Stick 1995). A single analysis allowed isolation and resolution of a mixture of 25 pigments present in the extract. Only the duration of the analysis depends on the type of column used (25 min for LiChroCART Hypersil ODS and 32 min for LiChroCART LiChrospher 100 RP18e); the sequence of eluted pigments remains unchanged.

The duration of analysis depends on the isolation system applied (e.g. solvents, appropriate temperature) and on the type of column, its diameter, particle size, pore size and number of plates. The analysis may last as long as 60 min (Heukelem and Thomas 2001); 40 min when a C8 column (150×4.6 mm, 3.5 μm particle size) thermostated at 25°C is employed (Zapata et al. 2000), or 20 min when using 3 μm Hypersil MOS (100×4.6 mm, C8; Shandon) for separation of chlorophyll a from divinyl chlorophyll a and zeaxanthin from lutein (Vidussi et al. 1996).

The calibration parameters obtained during the analysis were different for two kinds of stationary phases. Spectral characteristics allowing qualitative identification of pigments were very close only in the case of alloxanthin (λmax=426, 454, 483 nm) and neoxanthin (λmax=414, 438, 466 nm). In other pigments, the absorption spectrum was shifted by 1–5 nm (all details are presented in Table 2). Spectral characteristics vary not only between individual pigments but also with temperature and solvent within the pigment itself. The data published by Jeffrey and Humphrey (1975) and Bjørland (1997) present data on how solvent effects the algal carotenoid chromofores. The different λmax values and different shapes of spectra presented in Fig. 3 may have resulted from not the same solvent composition during retention.

The slopes of the calibration curves differed not only between different pigments but also for the same pigment when different columns were used. Different response factors (fp, ng mAU−1 s−1) were observed in most cases. The values of factor fp are representative of individual pigments isolated from the pigment mixture. The linear approximation of fp, which resulted from a sequence of injections of fixed pigment amounts onto chromatographic systems, yielded satisfactory correlation (correlation coefficients ranged from 0.92 to 0.99 in the case of the first group and from 0.6 to 0.99 for the second group of pigments). Poor correlation (0.59, wide scatter of points) was noted for pheophythin a. This pigment also has a higher detection threshold (36.3 ng per 100 μl), which means that pheophythin a can only be determined in samples with a high content of degradation products. Other chlorophyll a derivatives, chlorophyllide a and divinyl chlorophyll a, revealed little scatter of points, and, thus, the values of fp were close to 1 (0.98 and 0.94, respectively).

Slopes of the calibration curves of chlorophyll c3, antheraxanthin, echinenone and chlorophyll a were similar for both column types. Slopes of the calibration curves of chlorophyll c2, peridinin, fucoxanthin, neoxanthin, prasinoxanthin, diadinoxanthin, alloxanthin, myxoxanthophyll, lutein, zeaxanthin, chlorophyll b and β-carotene were steeper for the second group of pigments, separated by LiChroCART LiChrospher 100 RP18e. In the other cases the situation was reversed.

The detection thresholds of isolated pigments were very low and in most of the cases did not exceed 1 ng per 100 μl. The comparison of the detection thresholds of chlorophylls and carotenoids in both columns indicated that the second type of columns was more sensitive: from 1.02 times for myxoxanthophyll up to 3.2, 4.06 and 4.34 times for 19′haxanoyloxyfucoxanthin, α-carotene and neoxanthin, respectively. Only prasinoxanthin, antheraxanthin, chlorophyll a and β-carotene could be detected at lower concentrations by the use of the LiChroCART Hypersil ODS analytical column. The minimum delectable concentration of β-carotene was similar in both systems ~3.3 and 3.35 ng per 100 μl. The results obtained also indicate that pheophythin a is difficult to detect because of its very high threshold concentration, 36.7 ng per 100 μl.

In conclusion, employment of the chromatographic systems discussed here allows precise determination of the pigments present in natural samples, even if only present in low amounts (<0.1 μg dm−3). Our experiments have provided unique calibration parameters for RP-HPLC systems equipped with two types of columns: LiChroCART Hypersil ODS and LiChroCART LiChrospher 100 RP18e. Sufficient resolution of all pigments present in the mixture was achieved. Both columns used provided similar results in the isolation of pigments. Except for two pigment pairs, the chlorophylls c1 and c2 and zeaxanthin and lutein, all other chlorophylls and carotenoids injected onto the chromatographic columns were separated to the baseline. Even though zeaxanthin and lutein are hard to separate, the absorption spectrum of combined peaks could allow a statement on which one dominated. Although both pigment groups analyzed differed in amount and type, it was possible to establish calibration parameters for the two HPLC systems. Unfortunately, only the group of compounds tested on LiChroCART LiChrospher 100 RP18e possessed chlorophyll a derivatives (chlorophyllide a and pheophythin a), which may exclude the use of the LiChroCART Hypersil ODS analytical column in certain environmental analyses, for example, of sediment samples. Pigment concentrations in samples of naturally living phytoplankton can be determined by the use of both types of columns. The pigments analyzed are characteristic biomarkers of different phytoplankton groups. The LiChroCART LiChrospher 100 RP18e column was tested on a wide group of pigments and in most chlorophylls and carotenoids yielded better detection results. However, in the case of prasinoxanthin, antheraxanthin, chlorophyll a and β-carotene, the use of LiChroCART Hypersil ODS allows detection of these pigments at lower concentrations. In conclusion, the columns used in the analysis are interchangeable.