Selective Chlorophyll Removal Method to “Degreen” Botanical Extracts

Abstract

Chlorophylls are present in all extracts from the aerial parts of green plant materials. Chlorophylls may act as in vitro bioassay nuisance compounds, possibly preventing the reproducibility and accurate measurement of readouts due to their UV/vis absorbance, fluorescence properties, and tendency to precipitate in aqueous media. Despite the diversity of methods used traditionally to remove chlorophylls, details about their mode of operation, specificity, and reproducibility are scarce. Herein, we report a selective and efficient 45 min liquid-liquid/countercurrent chlorophyll clean-up method using Centrifugal Partition Chromatography (CPC) with a solvent system composed of hexanes-EtOAc-MeOH-water (5:5:5:5, v/v) in elution-extrusion mode. The broader utility of the method was assessed with four different extracts prepared from three well-characterized plant materials: Epimedium sagittatum (leaves), Senna alexandrina (leaves), and Trifolium pratense (aerial parts). The reproducibility of the method, the selectivity of the chlorophyll removal, as well as the preservation of the phytochemical integrity of the resulting chlorophyll-free (“degreened”) extracts were evaluated using HPTLC, UHPLC-UV, ¹H NMR spectroscopy, and LC-MS as orthogonal phytochemical methods. The clean-up process adequately preserves the metabolomic diversity as well as the integrity of the original extracts and is sufficiently rapid for the “degreening” of botanical extracts in higher-throughput sample preparation utilized for biological screening.

Graphical Abstract

Chlorophylls are present in the cells of nearly all photosynthetic organisms such as cyanobacteria, algae, and plants. Chlorophylls are cyclic structures consisting of tetrapyrroles linked via methine bridges, thereby forming chlorin, a tetrapyrrole macrocycle that is not to be confused with the atomic element, chlorine. In chlorophylls, the chlorin macrocycle chelates a magnesium ion (Mg²⁺) in its center.¹ Due to the high degree of electronic conjugation within the chlorin macrocycle, chlorophylls participate actively in photon capturing, excitation energy transfer, and storage that occurs within the light harvesting complex (i.e., antennae complex) of chloroplasts. This energy transfer enables cellular biosynthesis in photosynthesis.^2,3 The light energy captured by the chlorophyll pigments is used for photosynthesis and is re-emitted as a red fluorescence,^4–6 which can be used to study the photosynthesis process. Likewise, phytochemists utilize this red fluorescence to detect the presence of chlorophylls in botanical crude extracts (CEs).⁷ The structural characteristics of the chlorophylls are also responsible for their strong absorption in the blue and red spectrum of white light. Paired with poor absorption in the green spectrum, this explains the green color of chlorophyll containing organisms.⁸

Another structural characteristic of chlorophylls is the presence of a phytyl (C₂₀) side chain at position 17 (Figure 1). The different types of substituents on the chlorin backbone distinguish the five chlorophylls known to date: while chlorophyll a (1) is present in all oxygenic photosynthetic organisms, chlorophyll b (2) is found exclusively in algae and plants. Chlorophyll c (3) does not have the phytyl side chain, and together with chlorophyll d (4) and f (5) exists only in photosynthetic microorganisms.^9,10 Terrestrial plants utilize exclusively 1, which bears a methyl substituent at position 7 (B-ring), and 2 with a formyl group in the same position.¹¹ In plant cells, 1 and 2 are anchored into the thylakoid membranes of chloroplasts through their hydrocarbon phytic side chains. In their non-chelated form, 1 and 2 are called pheophytin a and b, respectively.

Figure 1.

Structures of the chlorophylls and their major degradation products.

Chlorophyll pigments are readily extracted from botanical matrices by a wide polarity range of organic solvents. They dissolve readily in lipophilic solvents such as petroleum ether, alkanes, chloroform, and to a lesser extent ethyl acetate (EtOAc). They are also soluble in alcohols such as MeOH and EtOH. Traditional methods for the preparation of botanical CEs such as aqueous infusion, decoction, and hydro-alcoholic maceration are associated inevitably with the co-extraction of chlorophylls. Therefore, green pigments commonly are present in botanical preparations at various concentrations, depending on the plant part(s) and solvents used.

Due to their enhanced reactivity in the presence of light and oxygen, chlorophylls have a notorious susceptibility to auto-degradation when handled in visible light and with aqueous media such as when under cell-based in vitro bioassay conditions.^12–16 As both the chlorophylls and their degradation products [e.g., pheophorbide a (6), Figure 1] naturally fluoresce, typically they interfere with in vitro bioassay readings that rely on fluorescence measurements. This is all the more important as the impact of fluorescence quenching cannot always be accounted for by controls. Chlorophyll pigments and their degradation products have a strong absorption between 400 and 700 nm, which falls frequently within the measurement range of dyes used to indicate biological endpoints and, thus, may potentially lead to false positive or negative results.^17,18 Hence, the accuracy, validity, and reproducibility of biochemical investigations with botanical samples is challenged generally by the near-ubiquitous presence of these interfering green phytoconstituents.^19–21 Therefore, reliable, rapid, and reproducible methods for the specific removal of chlorophylls in botanical CEs are highly desirable. Establishing a standardized, broadly applicable method for chlorophyll clean-up (“degreening”) of CEs should enhance the cell-based biological evaluation of botanical materials and, thereby, help advance the pharmacological evaluation of plant extracts.

Over the years, various methods to remove chlorophylls from botanical CEs have been reported. Most of these rely on solid-phase extraction (SPE)²² using different stationary phases to remove chlorophylls during initial fractionation. Diaion HP-20^22–25 is a widely used stationary phase for that purpose. When the goal is to remove chlorophylls or other coloring pigments entirely from the CE without fractionation of the other constituents, the use of charcoal has been reported.^26–32 Collectively, all these methods are known to remove the target pigments, but also eliminate other unspecified, potentially bioactive, compounds along with the chlorophylls.^27,30,32 Detailed protocols of the chlorophyll clean-up methods mentioned in the literature are generally left to the discretion of authors. The identity of compounds being removed together with the chlorophyll pigments and/or information concerning the preservation of the original phytochemical profiles are generally not assessed.

Liquid-liquid or countercurrent separation (CCS) methods have demonstrated utility for the isolation of larger quantities of target compounds from complex matrices.^33–38 The relative ease of use, affordable cost, and reduced environmental footprint (solvent usage), as well as the performance characteristics of contemporary equipment make CCS methods powerful tools in natural products research. Several CCS methods have been developed for the isolation of plant pigments such as 1, 2, anthocyanins, and carotenoids from chlorophyll-enriched organic extracts obtained from grass, spinach, and other plant materials.^39–44 However, the suitability for the removal of chlorophylls has not been assessed. In CCS, both the stationary and the mobile phases are liquids that represent a mixture of different solvents, constituting a solvent system. The latter has to be chosen carefully to form a biphasic mixture that, ideally, solubilizes the extract to be fractionated entirely, while still allowing for a balanced partitioning of the target compounds between the mobile and the stationary phase.^6,7,12,13 All phytochemicals distributed in both phases (notably the stationary phase, which is often aqueous) can be readily recovered by solvent evaporation. Hence, CCS is considered a loss-free technique as long as the analytes themselves are not volatile, requiring special, low-temperature recovery.

In contrast to the general use of CCS, which is dedicated typically to the purification of certain target compounds, the present work sought to implement such a technique for the selective removal of chlorophylls from CEs, so as to produce chlorophyll-free, “degreened” Knock-Out Extracts (chlorophyll-KOEs). Such a chlorophyll removal method should enable the selective subtraction of these assay interference compounds, along with an optimal recovery of all other phytochemicals present in the original CE. As described recently, CCS has enabled the production of DESIGNER (Deplete and Enrich Select Ingredients to Generate Normalized Extract Resources), which can serve as distinctive pharmacological tools to better understand the role of target phytochemicals in the biological properties of a complex botanical extract.^33,45 In line with the DESIGNER concept, this study details a CCS method for the production of chlorophyll-KOEs from botanical CEs obtained from three mainstream botanicals: Epimedium sagittatum, Senna alexandrina L. and Trifolium pratense L. This study also aims at evaluating the reproducibility and selectivity of the proposed CCS clean-up method towards the removal of chlorophylls in different plant matrices. For this purpose, orthogonal analytical methods such as HPTLC, UHPLC-UV/MS, LC-MS and ¹H NMR spectroscopy were used to compare the phytochemical profiles of the “degreened” extracts with those of their original CE counterparts.

RESULTS AND DISCUSSION

Preparation of Botanical CEs.

The three study plants, E. sagittatum (leaves), S. alexandrina (leaves), and T. pratense (aerial parts) were each selected according to three criteria: (i) their relatively high content of chlorophylls; (ii) their mainstream usage as botanical dietary supplement by the U.S public,⁴⁶ and (iii) the diversity in their phytochemical composition.^47–51 Two extraction methods were considered, in order to generate phytochemically complex crude extracts with different levels of chlorophyll pigments.

Powders of E. sagittatum and S. alexandrina were extracted successively by three solvents of increasing polarity: hexanes, 50% CHCl₃/MeOH, and 50% MeOH/water, using accelerated, pressurized solvent extraction. The three extracts were then combined to yield a single CE that covered the whole polarity range and all structural classes of metabolites, including the chlorophylls. The extraction process was repeated three times in order to evaluate its reproducibility (Figure S1, Supporting Information) and to generate identical CEs of the same plant material that could be used in the subsequent clean-up methods. Solutions of all six CEs displayed a dark brown color. To obtain extracts highly enriched in chlorophylls, the leaves of T. pratense (50.0 g) were extracted by sonication and maceration using two different solvent mixtures: 80% EtOH/water (CE#1) and 50% CH₂Cl₂/MeOH (CE#2). The sonication triggered the disruption of plant cellular membranes, which together with the use of non-polar solvents, favored the extraction of chlorophylls. Hence, both extracts were expected to be highly enriched in chlorophyll derivatives, as reflected in their dark green color. Both extracts were further used to evaluate the efficiency of the proposed chlorophyll clean-up method.

As references for the extraction reproductivity and chlorophyll clean-up evaluation, the overall phytochemical composition of each extract and reproducibility of the extraction process were evaluated. First, the extraction yield was calculated to be 23.0 ± 0.5% (w/w) for E. sagittatum, and 26.7 ± 1.9% for S. alexandrina. The relatively low standard deviations obtained for the three extracts/clean-up plant material indicated the good reproducibility of the extraction process. For T. pratense, CE#1 and CE#2 were prepared once. Second, the reproducibility of the phytochemical fingerprints of the CEs was assessed using orthogonal analytical methods: HPTLC, (U)HPLC-UV, and quantitative ¹H NMR spectroscopy. The three techniques complement each other to detect and identify a wide range of phytochemicals, including the chlorophyll pigments (Figures S2 and S3, Supporting Information).

(HP)TLC is one of the most accessible methods available to the natural product researcher for the chemical analysis of mixtures. HPTLC fingerprints were performed with two different solvent systems on normal-phase silica gel to observe (i) the chlorophyll derivatives through their red fluorescence at 365 nm, and (ii) all medium polar to non-polar phytochemicals present in the samples. (HP)TLC was, thus, used to rapidly visualize the phytochemical similarity of the CE replicates, and later to readily observe the removal of chlorophyll pigments.

Reversed-phase C₁₈ UHPLC provides a complementary chromatographic separation compared to silica gel TLC, with the option of utilizing UV detection with wavelengths adapted to compounds of interest, such as chlorophylls and/or species-specific markers. The chromatographic profiles were analyzed at 254 nm to obtain a general fingerprint of each CE. Figure 2A shows representative chromatographic comparisons of E. sagittatum extract at 254 nm. Chlorophylls and their derivatives were detected at 400 nm. The reproducibility of the phytochemical fingerprints for each replicate was assessed quantitatively by measuring and comparing the peak integrals of the chromatographic profiles at 254 nm and 400 nm (Figure 2A and Figure S4A, Supporting Information). Many phytochemicals, such as terpenoids, are not detectable at 254 nm. Therefore, other wavelengths, especially 205 and 360 nm, were also used to compare the overall similarity between the CE replicates.

Figure 2.

Stacked UHPLC-UV chromatograms (A) and ¹H NMR spectra (B; 400 MHz, DMSO-d₆) of the E. sagittatum extract replicates, n1-n3. (A) The reproducibility of the extraction process and, thus, of the phytochemical profiles were evaluated at 254 nm and 400 nm by measuring and comparing the peak area of each of the four regions of interest (ROIs) to the peak area of the entire chromatogram at 254 nm (and 400 nm for chlorophylls). Results are expressed as peak area percentages. (B) The reproducibility of ¹H NMR spectroscopy phytochemical fingerprints were evaluated by measuring and comparing the absolute integrals of each of the three ROIs to those of the full integral measured for the entire spectrum (2.6 – 10.0, 0.0 – 2.4 ppm, excluding solvent range). Results are also expressed as percentages relative to all quantifiable ¹H NMR signals.

NMR spectroscopy is nearly universal when detecting hydrogen, and truly orthogonal to the two chromatographic methods, TLC and UHPLC. Despite the common signal overlap in ¹H NMR spectra of a complex mixture, structurally distinct compounds often have characteristic ¹H signals that may reveal their presence even in a complex spectrum of a botanical CE. This is the case for the cyclic tetrapyrrole olefinic resonances of chlorophyll in the downfield region (8.0–10.0 ppm), as well as aliphatic side chain of 7 in the upfield region (0.5–2.0 ppm) of each CE ¹H NMR spectrum as shown in the case of E. sagittatum, Figure 2B. In addition to a qualitative comparison, simultaneous quantitative ¹H NMR (qHNMR) spectroscopy evaluation enabled determination of the relative proportions of the phytochemicals in the CEs. The reproducibility of qHNMR phytochemical fingerprints was evaluated by measuring and comparing the absolute integrals for each of the three regions of interest, relative to the total integral measured for the entire spectrum (10.0–4.0, 4.0–2.6, 0.0–2.4 ppm, excluding solvent range) of each CE (Figure 2B, and Figure S4B, Supporting Information).

These initial comparative analyses confirmed the reproducibility of the extraction process for E. sagittatum and S. alexandrina, and enabled the subsequent phytochemical characterization of “degreened” extracts. The reproducibility of the T. pratense extraction process was not assessed as these extracts were prepared only to evaluate the capacity of the proposed clean-up method to remove efficiently chlorophylls in highly enriched samples. Species-specific compounds were identified for the three species as follows: icariin for E. sagittatum, sennosides A and B for S. alexandrina, and finally biochanin A and formononetin for T. pratense (Figures S2, S3, and S14, Supporting Information). Generally, E. sagittatum and S. alexandrina CEs contained proportionally more polar to medium polar phytochemicals than non-polar compounds, as opposed to T. pratense CEs. Due to the relative abundance of more polar phytochemicals in the E. sagittatum and S. alexandrina CEs, the extracted chlorophyll pigments were barely observable on HPTLC at 365 nm. On the other hand, strong 365 nm red fluorescent bands were observed in the HPTLC phytochemical fingerprints of T. pratense extracts.

Selection of the Biphasic CCS Solvent System.

CCS relies on the use of a solvent mixture that forms two immiscible phases, herein called biphasic solvent systems. Various solvent systems were evaluated according to their capacity to (i) contain chlorophylls in mainly one phase; (ii) dissolve the CE entirely, with no visible particles; and (iii) prevent formation of emulsions. The solubility and distribution of chlorophyll derivatives in all prepared botanical CEs were evaluated in solvent systems belonging to the “ChMWat” solvent system family composed of CHCl₃-MeOH-water, and the HEMWat family composed of hexanes-EtOAc-MeOH-water. These two solvent system families were chosen as they generally cover a wide range of polarity.⁴⁵ When evaluating the ChMWat solvent systems, the attempted dissolution of the CEs resulted in a turbid biphasic solution with visible particles. Additionally, when performing the partitioning experiments with the ChMWat solvent systems, the solubility of CEs was not optimal: with CHCl₃-MeOH-water ratios at 10:3:7 v/v (ChMWat 0) and at 10:1:9 v/v ratios (ChMWat −2) relatively stable emulsions were formed. Therefore, the ChMWat solvent system family was not considered further. The partitioning studies performed with the hexanes-EtOAc-MeOH-water or HEMWat solvent system family (Figure 3 and Figure S5, Supporting Information) revealed an excellent capacity to completely dissolve all the investigated botanical CEs. Visual observation of the organic upper phase and the aqueous lower phase, shown in Figure 3, revealed the general distribution of pigments in the E. sagittatum samples. Comparative HPTLC analysis indicated that the more polar HEMWat +6 to +8 solvent systems led to the solubilization of chlorophyll pigments exclusively in the aqueous upper phase. However, some compounds of interest were noticed in the chlorophyll-enriched upper phase of the polar hexanes-EtOAc-MeOH-water solvent system (Figure S5 with S. alexandrina CE, Supporting Information). The non-polar HEMWat −4 to −6 solvent systems led to a distribution of chlorophylls in both phases. Hence, these solvent systems were not optimal for the selective removal of all chlorophylls. The solvent system composed of hexanes-EtOAc-MeOH-water (5:5:5:5, v/v) was found to be most suitable: all detected chlorophylls were concentrated in the organic upper phase, which also contained only small amounts of other phytochemicals (Figure 3). This solvent system enables the separation of medium polarity compounds such as flavonoid aglycones, and, therefore, is used for the general fractionation of botanical CEs.^52–57 Hence, the solvent system composed of hexanes-EtOAc-MeOH-water (5:5:5:5, v/v) was chosen for the subsequent liquid/liquid clean-up investigation.

Figure 3.

(A) Partitioning studies comparing the solubility of E. sagittatum. CE (at 12.5 mg/mL) and the distribution of the chlorophyll pigments in key solvent systems of the HEMWat family. Equal volumes (10 μL) of the upper and lower phase of each solution were spotted on TLC and observed under 365 nm after elution. (B) Solvent composition in v/v ratio of the tested solvent systems in the HEMWat family (H = hexanes, E = ethyl Acetate, M = methanol and Wat = water). Each solvent system is numbered from −8 to +8 according to its overall polarity, reflecting its solvent composition.

Removal of Chlorophylls by Simple Liquid/Liquid Partitioning.

The next step was to evaluate whether or not simple liquid/liquid (L/L) partitioning with the chosen solvent system would be apt to efficiently remove chlorophyll pigments from botanical CEs. Traditionally, L/L partitioning with solvents such as hexanes or petroleum ether is the first fractionation step of botanical CEs, mainly to remove waxes and other non-polar constituents including chlorophyll derivatives.⁵⁸ The ability to remove chlorophyll in a single step via L/L partition with the chosen solvent system would be attractively straightforward as L/L partitioning does not rely on instrumentation and, thus, applicable even with limited laboratory resources. Two different L/L partitioning tests were performed with the very chlorophyll-rich T. pratense CE#2. The first experiment used a proportion of CE to the total volume of solvent that was similar to what is generally used when performing a partitioning study to guide the selection of solvent systems for CCS (~12.5 mg/mL). For the second experiment, the distribution of chlorophyll relative to the other phytochemicals was assessed by using the same proportion of CE to the total volume of solvents that would be utilized in a CCS instrument (typically 1 mg/mL). After dissolving the extract in hexanes-EtOAc-MeOH-water (5:5:5:5, v/v; HEMWat 0), equal volumes of both upper phase and lower phase were collected separately in order to compare their phytochemical compositions by HPTLC, UHPLC-UV, and ¹H NMR spectroscopy (Figure 4 and Figures S6–S8 with the other plant materials, Supporting Information). The upper phase was highly enriched in chlorophyll derivatives as exemplified by its deep green color and the intense red fluorescent bands on the TLC (Figure 4C). The lower phase displayed UHPLC-UV and ¹H NMR spectroscopy phytochemical fingerprints that were very similar to those of the original CE. However, as shown by the comparative UHPLC-UV chromatograms and the ¹H NMR spectra (Figure 4A–B), the organic upper phase contained key T. pratense isoflavones that were also detected in the lower phase. Hence, the simple L/L partitioning approach was found to be insufficiently selective for the removal of chlorophylls. In fact, this method likely removes other potentially bioactive phytochemicals together with the chlorophylls. Nevertheless, the results demonstrate that a single-step L/L partition with the chosen solvent system leads to the concentration of chlorophyll pigments in the upper phase. Ultimately, for laboratories that do not have access to a CCS instrument, or in cases where, e.g., bioactivity results provide justification, this simplistic approach may still be used to remove a large part of the chlorophyll pigments non-specifically from polar to medium polar botanical CEs.

Figure 4.

Phytochemical fingerprints of T. pratense CE#2, obtained by (A) UHPLC-UV at 254 nm, (B) ¹H NMR (DMSO-d₆, 400 MHz) spectroscopy, and (C) HPTLC of the upper phase and lower phase generated from hexanes-EtOAc-MeOH-water 5:5:5:5 liquid/liquid partitioning (12.5 mg/mL). The results demonstrate that single-step liquid/liquid partition with hexanes-EtOAc-MeOH-water 5:5:5:5 can remove substantial amounts of the chlorophyll pigments from a botanical crude extract (CE). However, in such a simple procedure, some medium polar isoflavones (marked *) were also removed together with the chlorophylls as observed in the upper phase. Chlorophyll ¹H NMR spectroscopy signals are marked with dashed boxes and show as red bands at UV 365 nm on the HPTLC plates. The liquid/liquid partition results obtained with CEs from the other plant materials are available as Supporting Information (Figures S6–S8). Chlorophyll(s) are abbreviated as Cphyl(s) in this figure.

Removal of Chlorophylls by Countercurrent Separation (CCS).

In CCS, the use of the hydrostatic centrifugal partition chromatography (CPC) instruments is particularly attractive in terms of CE loading capacity (up to 100g scale depending on instrument) and flow rate operation (~25–350 mL/min), leading to faster fractionations of relatively large amounts of sample.^59,60 Taking advantage of the liquid nature of the stationary phase, it is possible to recover all of the compounds retained n this phase by extruding them from the CPC column. This enables a high recovery of all metabolites present in the original CE, ideally, without chemical alteration (assuming stability in the solvents and processing). For the CPC experiments, the CEs of E. sagittatum. (1.22 g ± 0.15), S. alexandrina (1.74 g ± 0.06), and T. pratense (CE#1: 1.01 g, and CE#2: 1.09 g) were prepared in equal volumes upper phase and lower phase (20 mL each) of the chosen solvent system prior to injection. The CCS runs were performed in the descending mode (upper phase as stationary phase) at a flow rate of 25 mL/min and a rotation speed of 2500 rpm. The overall CPC process was monitored by UV at different wavelengths, including 407 nm in order to observe the extrusion of chlorophyll derivatives. After CE injection, three column volumes of aqueous mobile phase were eluted for at least 30 min or until a partition coefficient of 3.8 was reached. At this stage, all polar and medium-polar phytochemicals had been eluted, whereas chlorophylls and other low-polarity metabolites remained in the stationary phase. During this ca. 30 min elution period, the samples collected led to the production of the chlorophyll knock-out extract (chlorophyll-KOE). Subsequently, the organic upper phase was reintroduced into the CPC column for a duration of 15 min (equivalent to 1.5 column volumes) in order to extrude the stationary phase and, thus, recover the chlorophyll pigments. Samples collected during the extrusion produced the chlorophyll-enriched fraction, chlorophyll-F. Hence, each processed botanical CE yielded a chlorophyll-KOE and a corresponding chlorophyll-F (Figure 5 and Figures S9–S12, Supporting Information).

Figure 5.

CPC chromatograms and HPTLC profiles of the produced chlorophyll knock-out materials and chlorophyll-containing fractions of E. sagittatum CE (A), and T. pratense CE#1 (B). The CPC chromatograms were monitored at UV 407 nm. The chlorophyll pigments of each CE remained in the stationary phase during the 30 min elution of the mobile phase (elution mode) and were recovered as a separate fraction once the stationary phase was extruded from the CPC column (extrusion mode). The developed 45-min CPC method leads to the production of corresponding pairs of chlorophyll-KOE and chlorophyll-F materials. The chlorophyll-KOE is formed after concentrating all collected mobile phase eluent. Comparative analysis of the HPTLC fingerprints revealed that most of the chlorophylls (red bands at 365 nm) were no longer present in the chlorophyll-KOE and effectively concentrated in the chlorophyll-F. Chlorophyll(s) are abbreviated as Cphyl(s) in this figure.

Figures 5A and B show the CPC chromatograms obtained with E. sagittatum and T. pratense CEs. The similarities of the HPTLC fingerprints between the CEs and their corresponding chlorophyll-KOEs suggested the preservation of the overall phytochemical composition after the CPC clean-up procedure. Table 1 shows the mass recovery of all runs, calculated as the ratio of the total dry weight of chlorophyll-KOE plus chlorophyll-F over the injected amount of CE. The average mass recovery was 95.9 ± 0.63% w/w, suggesting that the process produced negligible sample loss. Interestingly, the chlorophyll-F represented approximately 12.8% w/w of E. sagittatum and S. alexandrina CEs. This demonstrated that, even when botanical CEs exhibit no visually striking green color, and/or when chlorophyll pigments are barely detectable on (HP)TLC, these chlorophyll pigments may still be present in substantial quantities. As expected, the chlorophyll-F obtained from T. pratense CE#2 represented 21.1% w/w, the highest proportion of the original CE. The mass balance recovery results obtained with E. sagittatum and S. alexandrina extracts also suggest that the chlorophyll clean-up method leads to reproducible KOE from the same botanical material. (Table 1) Since the organic liquid stationary phase is replaced entirely during the extrusion step, the CCS “degreening” method may be performed sequentially with the same sample if small amounts of cross contamination may be acceptable. This continuous re-injection potential reduces the need for column-filling and increases throughput.

Table 1.

Mass Balance Recovery after a Single-step CCS Chlorophyll Clean-up

	chlorophyll-KOE	chlorophyll-F	total
	in % w/w of the original CE (means ± SD)
E. sagittatum	83.2 ± 0.76 (1.22 g ±0.15)	13.2 ± 0.78 (1.02 g ± 0.13)	96.5 ± 0.05 (0.16 g ± 0.02)
S. alexandrina	83.1 ± 0.72 (1.74 g ± 0.06)	12.4 ± 0.45 (1.44 g ± 0.04)	95.5 ± 0.45 (0.22 g ± 0.01)
T. pratense CE#1a	90.0 (1.01 g)	5.8 (0.89 g)	95.9 (0.01 g)
T. pratense CE#2a	76.2 (1.09 g)	21.1 (0.83 g)	97.3 (0.23 g)

^aT. pratense CE#1 and CE#2 were extracted and processed once.

Preservation of the Phytochemical Integrity.

This step of the study assessed (i) whether the proposed method left the chemical composition of the original CE unaltered, except for the unwanted chlorophyll pigments, and (ii) the reproducibility of the process. Therefore, a series of orthogonal phytochemical analyses were performed, including UHPLC-UV and qHNMR fingerprints, as exemplified in Figure 6 for E. sagittatum (Figures S10B–S12B, Supporting Information). The phytochemical similarities between the original CE and the chlorophyll-KOE were evident from visual comparison but also determined quantitatively (Figures 7 and 8).

Figure 6.

Comparative (A) UHPLC-UV chromatograms (254 nm) and (B) ¹H NMR spectra (400 MHz, DMSO-d₆) of the produced E. sagittatum KOE compared to its original CE. The similarities between chlorophyll-KOE and CE phytochemical profiles indicate that the proposed clean-up process preserved the metabolomic diversity and integrity of the original CE. The few compounds removed from the original CE (marked *) were recovered in the chlorophyll-F. This fraction represented 13.2% w/w of the original CE, whereas the KOE accounted for 83.2% w/w. (B) Characteristic ¹H NMR resonances spectroscopy (also marked *) related to the presence of chlorophylls and derivatives were observed in the ¹H NMR spectra of both chlorophyll-F and CE, but were absent in the chlorophyll-KOE. Chlorophyll(s) are abbreviated as Cphyl(s) in this figure.

Figure 7.

Orthogonal phytochemical analyses of the “degreening” of E. sagittatum extract, giving the chlorophyll fraction (F in green), analyzed by (A) HPTLC and (B) UHPLC-UV. The presence and identity of phytol (7), several fatty acids [oleic acid (8), linoleic acid (9), and palmitic acid (10)] and β-sitosterol (11) were confirmed by comparison with commercial reference compounds. Palmitic acid (10) was not detectable under our UHPLC-UV conditions. Pheophorbide a (6) was also identified in F by LC-MS analyses (Figure S17, Supporting Information) and against a commercial reference standard. As suggested by the red tailing on the HPTLC fingerprint of the chlorophyll-F, several other chlorophyll degradation products are present in this fraction. Mix represents chlorophyll-F spiked with 6–11.

The comparative UHPLC-UV chromatograms in Figure 6 showed near identical chemical profiles for the chlorophyll-KOE and the CE, with the exception of some signals of low abundance between 15 and 20 min. This retention time window corresponds to the chlorophyll pigments (Figures S10–S12 for other botanical samples, Supporting Information). Next, the peak areas of the four chromatographic regions of interest (ROIs) at 254 nm were measured and compared to that of the entire chromatogram (Figure S15, Supporting Information). The results showed that the chlorophyll-KOEs maintained their metabolomic composition based on the UHPLC patterns between 0 and 15 min compared to the CEs. The non-polar metabolites, depleted in the chlorophyll-KOEs, were recovered in the chlorophyll-F.

Likewise, comparison of the ¹H NMR spectra (Figure 6B) exhibited the preservation of most of the ¹H resonances between the CE and the chlorophyll-KOE, except for the resonances corresponding to the aliphatic side chain (0.5 – 2.0 ppm) and the residual ¹H NMR signals of the tetrapyrrole ring of chlorophyll (8.5 – 10.0 ppm). To evaluate further the similarity between the chlorophyll-KOE and the CE, integral values of four ROIs of each KOE replicate and each original CE were measured and then compared to the full integral of the entire spectrum (Figure S16, Supporting Information). Overall, the comparative ¹H NMR spectroscopy results confirmed (1) the similarity of the chlorophyll-KOE replicates that displayed similar fingerprints qualitatively and quantitatively; and (2) the similarity of the fingerprints between CE and chlorophyll-KOE, except for the intended loss of ¹H resonances corresponding predominantly to the aliphatic ¹H NMR signals of chlorophyll derivatives.

Phytochemical Analyses of the Enriched Chlorophyll Fraction.

In a complementary manner, the chlorophyll-F was analyzed for the presence of compounds other than the targeted chlorophylls. From a general point of view, one confounding factor results from the fact that chlorophylls are naturally unstable: they are prone to degradation under visible light, oxygen, heat, and in the presence of weak acids, to cite a few conditions.^61,62 During leaf senescence and fruit ripening, chlorophyll degradation is a dynamic oxidative process^63–66 that involves the breakdown of the aliphatic chain (dephytylation) and loss of the central Mg ion by the chlorophyllase enzyme.^16,67 Subsequently, 1 and 2 produce pheophorbide and pheophytin derivatives by additional oxidative reactions.⁶³ Therefore, it was expected that any chlorophyll-F would contain various chlorophyll degradation products, such as pheophorbide a (6) and phytol (7). Hence, the chlorophyll-Fs were analyzed by UHPLC-UV-MS and ¹H NMR spectroscopy in comparison with reference standards of 6 and 7. Interestingly, neither 1 nor 2 were detected in the chlorophyll-F suggesting their degradation into 6 and 7 as observed on the HPTLC profiles of Figure 7A. The presence of 6 in the chlorophyll-F was confirmed by UHPLC-UV and LC-MS analyses (Figure S17, Supporting Information). In addition to the chlorophyll degradation products, the chlorophyll-F could contain other non-polar metabolites that would be more concentrated, and, thus, easier to detect than in the CE. In fact, other fatty acids, such as oleic (8) and linoleic acid (9), and potentially palmitic acid (10), as well as the ubiquitous β-sitosterol (11) were identified in the fractions by UHPLC-UV, and by HPTLC after spiking with commercial reference standards of 6–11 (“Mix” in Figure 7A). The analyses of the different ¹H NMR chlorophyll-F spectra (Figure S17, Supporting Information) also confirmed the presence of ¹H NMR resonances spectroscopy corresponding to fatty acids: 5.30 ppm (olefinic proton), 2.17 ppm (α-methylene proton), 1.0 – 1.5 ppm (aliphatic chain), and 0.84 – 0.92 ppm (methyl). While representing a limitation of the approach, the depletion of fatty acids and/or phytosterols from the original CE is only of minor concern for the intended use of chlorophyll-KOEs as pharmacological tools: due to their nature as pan-assay interference (PAINS) compound and invalid/improbable/interfering metabolic panaceas (IMPs), their collateral removal could even be considered a positive side-effect of the clean-up procedure with regard to enhancing the probability of identifying new bioactives from the chlorophyll-KOEs.²¹

Loading Capacity of the CCS Chlorophyll Clean-up Method.

An important CCS method development step is to determine the maximum sample loading capacity that could be processed without losing resolution and, therein, the ability to remove the chlorophylls. First, the solubility of the CEs in the chosen solvent system was explored to optimize the injection volume and sample mass that could be cleaned-up in one single CCS step. For this purpose, a loading capacity investigation was undertaken with T. pratense CE#2, a CE highly enriched in chlorophylls. Increasing quantities of CE#2 were dissolved progressively in 40 mL of the selected solvent system (1:1 ratio of upper phase and lower phase) until precipitation or insoluble particles were observed even after thorough mixing and prolonged sonication. As a result, a maximum of 4.6 g of T. pratense CE#2 could be dissolved entirely in a 1:1 mixture of upper phase/lower phase of hexanes-EtOAc-MeOH-water (5:5:5:5, v/v; HEMWat 0). Upon injection of this maximum sample load, the weight recovery was calculated to be 96.7% w/w. Phytochemical fingerprinting of the produced chlorophyll-KOE confirmed the preservation of the metabolomic features of the original CE (Figure S14, Supporting Information). In addition to chlorophyll derivatives, the chlorophyll-F produced contained the same fatty acids and phytosterols as observed when lower quantities of CEs were cleaned up. In conclusion, the developed method is capable of “degreening” substantial quantities (~2 g per 100 mL CPC instrument volume) of botanical CE in a single step.

Concluding Remarks.

The removal of unwanted chlorophylls and derivatives has been a long-term challenge in the research of plant-based natural products, affecting their testing with in vitro cell-based and in vivo bioassays. The present study demonstrates that crude botanical materials (including crude extracts, CEs) can be “degreened” efficiently, and in a target compound-specific fashion, using CCS methodology with a solvent system composed of hexanes, ethyl acetate, MeOH, and water (5/5/5/5, v/v) in elusion-extrusion mode. The chlorophyll cleaned-up material, also termed a chlorophyll-Knock-Out Extract (chlorophyll-KOE), was obtained within 30 min during the elution stage of the CCS run, whereas the chlorophyll pigments were recovered during the subsequent 15 min of stationary phase extrusion. As elution-extrusion CCS furnished a ready-to-use CCS column after each run, the method can be used to “degreen” botanical CEs in a continuous series of batch injections.

The method has the demonstrated capability of removing the chlorophylls from up to 4.6 g of a CE highly enriched in chlorophyll, in a single run and using a 250-mL CPC instrument. This makes scale-up “degreening” of kilogram quantities of crude botanicals feasible, as 5-L and higher volume hydrostatic CCS instruments are available. The use of UV monitoring during the CCS process also makes it possible to adapt the exact time necessary for each elution or extrusion step to the different types (specific chlorophyll pattern and content) of botanical CEs. The general suitability of the standardized CCS method proposed herein was demonstrated for four representative botanical CEs, which had significantly different phytochemical profiles and contained different proportions of chlorophyll derivatives.

Finally, the liquid/liquid-based nature of this chlorophyll clean-up method has led to reproducible phytochemical profiles of the chlorophyll-KOEs and was found to preserve the metabolomic diversity and integrity of the original CEs with acceptable confidence; observed differences in the phytochemical fingerprints were only very minor. In addition to the targeted chlorophyll derivatives and certain degradation products [pheophorbide a 6 and phytol 7], the chlorophyll-enriched fraction, chlorophyll-F, contained free (phytol-like) fatty acids as well as some phytosterols. Considering the known PAINS/IMP nature of these non-chlorophyll metabolites, this imperfection in the analytical specificity of the “degreening” method is of minor concern for many of the intended uses of chlorophyll-KOE extracts, especially when they are employed as pharmacological tools in botanical research. The demonstrated ability to clean-up (“degreen”) chlorophyll from botanical CEs with high specificity, and via a standardized method, opens new opportunities for undertaking advanced, more accurate cell-based bioassays, which are needed to enhance the discovery of bioactive principles in botanical preparations.

EXPERIMENTAL SECTION

General Experimental Procedures.

HPLC grade solvents (Thermo Fisher Scientific, Waltham, MA, USA) were used for experiments. Plant materials were extracted using the Dionex ASE 350 Accelerated Solvent Extractor (Thermo Fisher Scientific, Waltham, MA, USA). HPTLC was performed on HPTLC Silica gel F₂₅₄ (EMD Chemistry Inc., Germany) and HPTLC Nano-SIL HD/UV₂₅₄ plates (Macherey-Nagel, Germany). Each CE was spotted onto the HPTLC plates using the CAMAG Automatic TLC Sampler 4. HPTLC pictures were taken with a UVP MultiDoc-It Digital Imaging System.

A centrifugal partition chromatography (CPC) extractor, SCPC-250B (Gilson-Armen Instrument Inc. SAS, France), was used for the production of the chlorophyll-KOE. UHPLC analyses were performed on a Kinetex 1.7 μm XB-C18 100Å column (50 × 2.1 mm, Phenomenex, Torrance, CA, USA), using a Shimadzu UFLC (Shimadzu Corp., Kyoto, Japan) Nexera UHPLC system equipped with a diode array detector (DAD, Shimadzu SPD-M20-A, Kyoto, Japan), and fluorescence detector (Shimadzu RF-20A/20Axs, Kyoto, Japan). UHPLC data analyses were processed with the Shimadzu LabSolutions software package.

The ¹H NMR spectroscopy analyses (including quantitation by qHNMR) were performed on a Jeol ECZ 400 MHz (Jeol, Tokyo, Japan) equipped with a Super COOL probe (NM-Z161331TH5SC), and z-axis pulse field gradient. The acquired spectra were processed using the Mnova NMR software package (v.12.0.4, MestreLab Research S.L., A Coruña, Spain). Mass spectrometry analyses were performed on an ESI-Q-TOF IMPACT II mass spectrometer (Bruker QtofControl Version 4.1) coupled to a Shimadzu Nexera X2 UHPLC system. Data analyses were performed on the Compass Data Analysis software (Bruker Version 4.4). Icariin (part# 06–018/ lot 141526), sennoside A (part# 020456S/ lot 17061644), sennoside B (part# 020457S/lot 11120230), biochanin A (part# B-106/lot 23347), and formononetin (part# F-103/lot 0303049) reference compounds were purchased from Indofine Chemicals Company, Inc. (Hillsborough, NJ, USA). Further commercial reference compounds used for the analysis of the chlorophyll-F were pheophorbide a (part# 15664–29-6, Cayman Chemical, Ann Arbor, MI, USA), phytol (part#7541–49-3, St. Louis, Sigma-Aldrich, MO, USA), oleic acid (part# 01008–1G, Sigma-Aldrich), linoleic acid (part# 463–40-1, Sigma-Aldrich), palmitic acid (part# P0500–10G, Sigma-Aldrich), and β-sitosterol (part# 1612947, USP, Rockville, MD, USA).

Plant Materials and Authentication.

Three different botanical materials were used in this study. Powdered aerial parts from Epimedium sagittatum (part# 209365–51/ lot 72840) and powdered Senna alexandrina leaves (part# 209560–51/ lot 72794) were both purchased from Starwest Botanicals (Sacramento, CA, USA). Crushed aerial parts of Trifolium pratense (packed on April 2013/Lot 20042) were purchased from Mountain Rose Herbs (Eugene, OR, USA). All plant materials were identified through DNA barcoding,⁶⁸ macroscopic analyses, microscopic analyses for E. sagittatum and S. alexandrina powders.^69–71 The protocol used for botanical identification by DNA barcoding, the DNA sequences obtained, as well as the combined microscopic and phytochemical fingerprints are made available at DOI: https://doi.org/10.7910/DVN/M8CW8Z. Phytochemical analyses of the CEs were performed by means of HPTLC, UHPLC-UV, and ¹H NMR spectroscopy fingerprinting (see S2, S3 and S14, Supporting Information) in comparison with their respective reference compounds: icariin for E. sagittatum, sennosides A and B for S. alexandrina; biochanin A and formononetin for T. pratense.

Crude Extract Preparation.

Authenticated plant materials (10.0 g each) of E. sagittatum and S. alexandrina were extracted successively with hexanes, CHCl₃-MeOH (50:50), and MeOH-water (50:50) using the a Dionex ASE 350 instrument (Figure S1, Supporting Information). The solutions from all three extraction steps were combined to produce the final CE. Each extraction was repeated three times for each plant material. Parameters of extraction on the Dionex ASE 350 instrument were set as follows: temperature of extraction 60 °C; static time (duration of extraction) 15 min; 1 cycle of extraction; plant powder rinsing after extraction with 30% of initial extraction volume; the pressure during extraction for each sample was approximately 1500 psi; the ratio of plant material to volume of solvent was 1:10 (g/mL). The extraction yields were determined as 23 ± 0.5 % (w/w) for E. sagittatum and 26.7 ± 1.9 % for S. alexandrina. Crushed T. pratense aerial parts were extracted with ultrasonic treatment to obtain an extract highly enriched in chlorophylls. The extraction was performed using an ultrasound probe (QSONICA Sonicator Q500) to lyse plant cells. The plant material-to-solvent ratio was 1/14 (g/mL). A 30-min ultrasonication was performed (80% amplitude, with a pulse on/off: 20/5 sec), followed by an overnight maceration without heat. Two different T. pratense extracts were prepared: one with a solvent mixture composed of EtOH-water (80:20, v/v; CE#1), and another with CH₂Cl₂-MeOH (50:50, v/v; CE#2). The extraction yields were calculated to be 24.2% w/w for CE#1 and 5.9% w/w for CE#2.

Phytochemical Analysis.

HPTLC, UHPLC-UV, and ¹H NMR spectroscopy analyses were performed for the initial CEs and the prepared chlorophyll-KOEs to evaluate the preservation of phytochemical integrity and the efficiency of chlorophyll removal. HPTLC fingerprints were performed with two different elution solvents for each extract to observe (i) the chlorophyll derivatives and (ii) all medium polar to non-polar phytochemicals. For (i), a solvent composed of hexanes, EtOAc, acetone, and MeOH (6:1:1:0.4, v/v) was used for development, and the HPTLC plates were evaluated at 254 and 365 nm. For (ii), a solvent composed of EtOAc, HOAc, HOForm, and water (100:11:11:26, v/v) was used for development, then the HPTLC plates were evaluated at 254 and 365 nm.

UHPLC-UV analyses were performed with a phytochemical screening gradient on a Kinetex UHPLC column with a solvent mixed from (A) water and (B) CH₃CN both with 0.1 % formic acid, increasing B from 5% to 100% in 25 min at 0.7 mL/min. Specific elution conditions were created for E. sagittatum CE as follows: from 5% up to 35% B in 12 min, with a 35% B isocratic step during 5 min, then from 35% to 100% B in 3 min. Under these conditions, icariin eluted at 8.9 min, as confirmed with a reference standard (Figure S2, Supporting Information). Specific elution conditions for S. alexandrina CE were as follows: from 5% to 21% B in 12 min, with a 21% B isocratic elution during 5 min, then from 21% to 100% B in 3 min. Under these conditions, sennosides A and B were detected at 9.3 min and 7.8 min, respectively and in agreement with the retention times of the reference standards (Figure S3, Supporting Information). Specific elution conditions for the T. pratense CE involved increasing B from 5% up to 100% in 25 min. Under these conditions, the characteristic compounds, biochanin A and formononetin, were detected at 7.1 min and 5.9 min, respectively and in agreement with the retention time of the reference standards (Figure S14, Supporting Information). Elution conditions for the evaluation of the maximum loading capacity of T. pratense CE#2 involved increasing B from 5% to 100% in 30 min, which led to the elution of the chlorophyll derivatives around 20 min as detected at 407 nm (Figure S13, Supporting Information).

For both HPTLC and UHPLC analyses, all extracts were prepared at 10 mg/mL in MeOH/water for E. sagittatum and S. alexandrina (50:50, v/v) and T. pratense (100:0, v/v). The chlorophyll-KOE and chlorophyll-F were also analyzed at 10.0 mg/mL, and the reference compounds at 1.0 mg/mL. Aliquots of 5 μL of the solutions were spotted on the HPTLC plates, whereas 2 μL was injected for UHPLC-UV analyses. The reproducibility of the phytochemical fingerprints for each replicate was assessed by measuring and comparing the peak area of the chromatographic profiles at 254 nm (for all phenolic compounds) and 400 nm (for the analysis of chlorophylls). For this purpose, each chromatogram was divided in four regions of interest (ROI), for which integration of peak area values were measured and reported to the integration of peak area) obtained for the entire chromatogram (Figures 2 and S4, Supporting Information). Each specific ROI was, thus, characterized by a percent of the full chromatogram.

The ¹H NMR (qHNMR) spectroscopy analyses were performed at 25 °C under quantitative conditions using a 90° single pulse experiment (relaxation delay: 60 sec, receiver gain: 46, number of scans: 64). For qHNMR analyses, approximately 5.0 mg of CE, chlorophyll-KOE and chlorophyll-F were exactly weighed (0.01 mg) and dissolved in 200.0 μL of DMSO-d₆ (to give a final concentration of ~ 25 mg/mL), then transferred into 3 mm NMR tubes. Each sample was weighed and prepared in triplicate (#A/B/C) as follows: for E. sagittatum CE (#A: 5.07 mg, #B: 5.12 mg, and #C: 5.08 mg), chlorophyll-KOE (#A: 5.02 mg, #B: 5.06 mg and #C: 5.04 mg), chlorophyll-F (#A: 5.00 mg, #B: 5.00 mg, #C: 5.00 mg), for S. alexandrina CE (#A: 5.06 mg, #B: 5.06 mg, #C: 5.07 mg), chlorophyll-KOE (#A: 5.05 mg, #B: 5.05 mg, #C: 5.09 mg) and chlorophyll-F (#A: 5.00 mg, #B: 5.00 mg, #C: 5.00 mg), icariin (0.86 mg), sennosides A and B (A: 0.94 mg and B: 0.96 mg), T. pratense CE#1 (6.35 mg) and CE#2 (7.25 mg), biochanin A (3.87 mg), and formononetin (3.15 mg). Comparative analyses between CEs and KOEs were performed by dividing each ¹H NMR spectrum into three ROIs: 4.0–10.0 ppm, 2.6–4.0 ppm, and 0.0–2.4 ppm. The absolute integrals were measured for each ROI without considering the chemical shift range surrounding the residual DMSO-d₅ (2.40–2.60 ppm) and HDO (3.25–3.45 ppm) signals for the quantitative analysis. Integral values obtained for each ROI were related to the integral value of the entire spectrum (0.0–10.0 ppm; excluding the solvent ranges), thereby leading to the determination of a percentage for each ROI. Each measured value was then normalized to the exact weight of the sample.

Selection of CCS Solvent System.

Approximately 20 mg of each of the CE were weighed into a 4 mL vial, to which 1 mL of each phase of the selected solvent system was added. The mixtures were then sonicated. Next, the vials were shaken vigorously to equilibrate the sample thoroughly between the two phases. For HPTLC analysis, 5 μL aliquots of each phase were spotted onto the plate, which was then developed in the solvents described above to visualize the distribution of the chlorophylls and other phytochemicals between both upper and lower phases.

Chlorophyll Clean-up by One Step Liquid/Liquid Partition.

To evaluate the efficiency of a one-step L/L partition in a separation funnel, 25 and 250 mg of each of the CEs were diluted in 20 mL of lower phase and upper phase (1:1 v/v) of the selected solvent system composed of hexanes-EtOAc-MeOH-water (5:5:5:5, v/v; HEMWat 0). Full dissolution of each CE was sonicated until fully dissolved. The solutions were transferred into separation funnels, shaken several times, and after a min settling time, the two phases were collected. The phytochemical composition of each phase was then analyzed by HPTLC, UHPLC, and ¹H NMR spectroscopy.

Chlorophyll Clean-up by One Step CCS.

CCS was performed on an ARMEN Spot-Prep CPC system. Initially, the column, with a total capacity of 250 mL, was filled with the organic upper phase of the selected solvent system at a flow rate of 25 mL/min at 2500 rpm. The aqueous lower phase was then introduced as mobile phase at a flow rate of 25 mL/min, with the rotation speed adjusted to 2500 rpm, thereby leading to a stationary phase retention with an S_f value of 0.72 (V_{stationary phase} = 180 mL and V_{mobile phase} = 70 mL). CEs of E. sagittatum (#A: 1.386 g, #B: 1.185 g, #C: 1.096 g), S. alexandrina (#A: 1.678 g, #B: 1.797 g, #C: 1.738 g), and T. pratense (CE#1: 1.010 g and CE#2: 1.091 g) were dissolved in 20 mL of upper phase and 20 mL of lower phase from the selected solvent system and injected into the CPC column. The separation was performed in the reversed-phase (descending) mode with EECCC (Elution Extrusion CounterCurrent Chromatography) applied.^72,73 During the elution time, the extract was cleansed of the chlorophylls, which remained dissolved in the stationary phase. During the subsequent 15-min extrusion stage, the effluent was collected into test tubes. The chlorophylls were recovered at 37 min, after ~175 mL of extrusion. The total duration of the CPC clean-up process was approximately 45 minutes per injected CE. The total volume of solvents used per sample clean-up was three column volumes, equivalent to 750 mL of the solvent system lower phase during the elution, and 2 column volumes of upper phase during the extrusion, for a total of 1.25 L of solvent mixture.