Journal of Chromatography B, 941 (2013) 54–61

Contents lists available at ScienceDirect

Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb

Versatile solvent systems for the separation of betalains from processed Beta vulgaris L. juice using counter-current chromatography Aneta Spórna-Kucab a,∗ , Svetlana Ignatova b , Ian Garrard b , Sławomir Wybraniec a a Department of Analytical Chemistry, Institute C-1, Faculty of Chemical Engineering and Technology, Cracow University of Technology, ul. Warszawska 24, Cracow 31-155, Poland b Brunel Institute for Bioengineering, Brunel University, Uxbridge, Middlesex, United Kingdom

a r t i c l e

i n f o

Article history: Received 13 May 2013 Received in revised form 27 September 2013 Accepted 1 October 2013 Available online 10 October 2013 Keywords: Betanin Betalains Betacyanins Counter-current chromatography Beta vulgaris L

a b s t r a c t Two mixtures of decarboxylated and dehydrogenated betacyanins from processed red beet roots (Beta vulgaris L.) juice were fractionated by high performance counter-current chromatography (HPCCC) producing a range of isolated components. Mixture 1 contained mainly betacyanins, 14,15-dehydro-betanin (neobetanin) and their decarboxylated derivatives while mixture 2 consisted of decarboxy- and dehydrobetacyanins. The products of mixture 1 arose during thermal degradation of betanin/isobetanin in mild conditions while the dehydro-betacyanins of mixture 2 appeared after longer heating of the juice from B. vulgaris L. Two solvent systems were found to be effective for the HPCCC. A highly polar, high salt concentration system of 1-PrOH–ACN–(NH4 )2 SO4 (satd. soln)–water (v/v/v/v, 1:0.5:1.2:1) (tail-to-head mode) enabled the purification of 2-decarboxy-betanin/-isobetanin, 2,17-bidecarboxy-betanin/-isobetanin and neobetanin (all from mixture 1) plus 17-decarboxy-neobetanin, 2,15,17-tridecarboxy-2,3-dehydroneobetanin, 2-decarboxy-neobetanin and 2,15,17-tridecarboxy-neobetanin (from mixture 2). The other solvent system included heptafluorobutyric acid (HFBA) as ion-pair reagent and consisted of tert-butyl methyl ether (TBME)–1-BuOH–ACN–water (acidified with 0.7% HFBA) (2:2:1:5, v/v/v/v) (head-totail mode). This system enabled the HPCCC purification of 2,17-bidecarboxy-betanin/-isobetanin and neobetanin (from mixture 1) plus 2,15,17-tridecarboxy-2,3-dehydro-neobetanin, 2,17-bidecarboxy-2,3dehydro-neobetanin and 2,15,17-tridecarboxy-neobetanin (mixture 2). The results of this research are crucial in finding effective isolation methods of betacyanins and their derivatives which are meaningful compounds due their colorant properties and potential health benefits regarding antioxidant and cancer prevention. The pigments were detected by LC-DAD and LC–MS/MS techniques. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Counter-current chromatography (CCC) is a liquid–liquid chromatography technique which was invented in the early 1960s [1,2]. In high-speed counter-current chromatography (HSCCC), high speed coil rotation around its own axis and a central axis (planetary motion) generates a centrifugal field to retain the liquid stationary phase in the coil. The mobile phase is pushed through

Abbreviations: ACN, acetonitrile; BuOH, butanol; CCC, counter-current chromatography; CID, collision induced dissociation; EtOH, ethanol; HFBA, heptafluorobutyric acid; HPCCC, high-performance counter-current chromatography; HSCCC, high-speed counter-current chromatography; KD , partition coefficient; MeOH, methanol; PFCA, perfluorocarboxylic acid; RP-HPLC, reversed phase highperformance liquid chromatography; PrOH, propanol; TBME, tert-butyl methyl ether; TFA, trifluoroacetic acid. ∗ Corresponding author: Tel.: +48 12 628 30 74. E-mail addresses: [email protected], [email protected] (A. Spórna-Kucab). 1570-0232/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jchromb.2013.10.001

with a pump. The g-level produced is an effect from the coil rotation and for a typical HSCCC machine, it is between 55 and 80 g-level [1–3]. High-performance counter-current chromatography (HPCCC) is the name given to a high g-level machine (240 g) and was introduced by the Brunel Institute for Bioengineering [4]. The application of CCC to the fractionation and purification of natural plant pigments has been shown in numerous publications [5–11]. Beta vulgaris L. is increasingly utilized as a source of natural food dyes due to a growing interest of consumers in its potential health benefits (antioxidant, anticarcinogenic) and the non-toxic features of betalains. Since some synthetic pigments are considered as toxic and harmful [12] there is a demand for natural equivalents. Choosing a suitable solvent system for betalains purification is challenging due to their low stability in some physicochemical conditions [12–14]. A few pathways of betalain degradation and transformation are known, such as decarboxylation, dehydrogenation, hydrolysis and deglycosylation. Decarboxylation of betalains can occur at either

A. Spórna-Kucab et al. / J. Chromatogr. B 941 (2013) 54–61

55

The novelty of this contribution is a fractionation and isolation of decarboxylated and dehydrogenated derivatives of betanin from processed B. vulgaris L. juice using HPCCC. These mixtures of betalains have never been separated by CCC, which would be a useful technique for the separation as its liquid stationary phase does not catalyze degradation or cause irreversible adsorption and loss of the components, in the way solid stationary phases may do. The differences in elution profiles traced in the HPCCC and HPLC separations were of special interest and were indicated by recent betalain separations [6,8,18]. The HPCCC process was accomplished using two different types of solvent systems: an ion-pair system and a high salt concentration system. The high salt solvent systems were used for the first time in order to separate betalains. Whilst ionpair solvent systems have been reported before for the separation of non-decarboxylated and non-dehydrogenated betalains, nothing is known about their efficiency in the separation of decarboxylated and dehydrogenated betacyanins. Furthermore, the presence of toxic ion-pair agents makes these systems less attractive for use in the food industry [6–9]. 2. Experimental 2.1. Reagents HPLC-grade acetonitrile (ACN), 1-propanol (1-PrOH), ethanol (EtOH), 1-butanol (1-BuOH), ammonium sulphate, tert-butyl methyl ether (TBME), TFA and HFBA were obtained from Fisher Chemicals (Loughborough, UK). Water was deionized (Purite, Thames, Oxon, UK). HPLC-grade formic acid, methanol (MeOH) were obtained from POCH (Gliwice, Poland). 2.2. The preparation of the crude pigment extracts

Fig. 1. Steps of betanin and its diastereomer isobetanin thermal degradation pathways.

C-2, C-15 or C-17 carbon positions, however, usually occurs at C-2 and C-17. The dehydrogenation is observed at C-2,3 and C-14,15. The products of betanin degradation are usually more stable, which makes them interesting material for further application in the pharmaceutical and food industries (Fig. 1) [12,15–17]. Preparative isolation of unstable betalains by HPLC is often problematic due to the catalytic action of the solid stationary phase causing pigment degradation, therefore, new separation methods such as counter-current chromatography create an important possibility of obtaining pure pigments. CCC enables the use of different stationary phases through the application of different solvent systems without the need to buy a new column. In addition, modern CCC technology is as easy as HPLC to scale up to preparative and pilot levels. Hitherto, the first successful isolation and purification of more hydrophobic betalains by HSCCC was carried out in a solvent system consisting of TBME–BuOH–ACN–water (acidified with ion-pair reagents TFA or HFBA) [6–9]. The addition of ion-pair reagents results in a different chromatographic behavior of betalains e.g. longer retention time of betalains in RP-HPLC [6,18]. The addition of ion-pair additives to the CCC solvent systems changes the partition coefficient (KD ) of betalains and efficiently shifts them to the organic phase, creating a new possibility for separation of these highly polar plant pigments [6,8,18].

Two groups of betacyanins with different decarboxylation and dehydrogenation levels were obtained by thermal treatment of B. vulgaris L. juice and then analyzed by LC-DAD and LC–MS/MS. The juice was obtained from red beet roots (purchased as whole beet roots from the local market, Kraków, Poland) which were washed, hand-peeled, cut into small pieces and squeezed in a juice extractor (Zelmer, Rzeszów, Poland) (Table 1). The heating of betalain mixtures in the juice was performed at 85 ◦ C for 30 min (mixture 1) and 60 min (mixture 2), both acidified with 0.2% (v/v) formic acid according to a previous procedure [12]. The mixtures were separately purified on a preparative solid-phase extraction (SPE) column packed with C-18 reversed phase material (Merck, Darmstadt, Germany) [12]. The eluates in aqueous-acetonitrile solution were then concentrated by rotary evaporator and then freeze-dried for the HPLC analysis and the HPCCC experiments. 2.3. Apparatus A semi-preparative Spectrum HPCCC J-type modern hydrodynamic CCC instrument was used (Dynamic Extractions, Slough, UK) for the separation of betanin/isobetanin and their decarboxyand dehydro-derivatives (mixtures 1 and 2). The Spectrum HPCCC had a maximum rotation speed of ca. 1600 rpm (R = 75 mm, 240 g field). The instrument was equipped with two columns of 143.5 ml total capacity, 71 m long and 1.6 mm i.d. The mobile phase was pumped in the ‘tail-to-head’ direction (system AI) and ‘head-to-tail’ direction for system BIV (Table 2). The initial scouting runs were performed on the analytical size Mini HPCCC instrument (systems AI–AIII, and BI–BIV) supplied by Dynamic Extractions (Slough, UK). The Mini HPCCC was equipped with a single 7 cm diameter column made with 0.8 mm i.d. polytetrafluorethylene (PTFE) tubing: 18.2 ml capacity, column length

56

A. Spórna-Kucab et al. / J. Chromatogr. B 941 (2013) 54–61

Table 1 Chromatographic, spectrophotometric and mass spectrometric data of the pigments identified in the crude mixtures submitted for the HPCCC separations. Peak no. 1 2 1 2 3 3 4 4 5 6 7 8 9 10 a b

Compound a

Betanin 17-Decarboxy-betanina Isobetanina 17-Decarboxy-isobetanina 2-Decarboxy-isobetanina 2-Decarboxy-betanina 2,17-Bidecarboxy-betanina 2,17-Bidecarboxy-isobetanina 17-Decarboxy-neobetaninb 2,15,17-Tridecarboxy-2,3-dehydro-neobetaninb 14,15-Dehydro-betanin (neobetanin)a 2,17-bidecarboxy-2,3-dehydro-neobetanin b 2,15,17-Tridecarboxy-neobetaninb 2-Decarboxy-neobetaninb

Symbol

Rt [min]

max [nm]

m/z [M+H]+

m/z from MS/MS of [M+H]+

Bt 17-dBt IBt 17-dIBt 2-dIBt 2-dBt 2,17-dBt 2,17-dIBt 17-dNBt 2,15,17-dec-2,3-dHNBt NBt 2,17-dec-2,3-dHNBt 2,15,17-dNBt 2-dNBt

14.3 15.1 15.6 16.5 18.7 18.7 20.2 20.2 20.5 21.3 22.0 22.8 23.5 26.0

538 505 538 505 533 533 507 507 446 394 468 409 451 483

551 507 551 507 507 507 463 463 505 415 549 459 417 505

389 345 389 345 345 345 301 301 343; 299; 255 253 387; 343 297 255 343; 299; 255

Pigments from the mixture 1. Pigments from the mixture 2 (tentatively identified).

36 m. The column was mounted in a cantilever rotor containing a counterweight for balance when rotating. The distance between the holder axis of the coil and the central axis of the instrument was 50 mm (revolution radius – R). The maximum rotation speed was 2049 rpm (240 g field). The CCC machines were connected to a thermostat, which enabled maintaining a constant temperature during the separation process (20 ◦ C). During all CCC runs a K-501 Knauer (Berlin, Germany) pump, UV-ViS detector Shimadzu (Lyon, France) and fraction collector Foxy Jr. from Knauer company (Berlin, Germany) were used. The positive ion electrospray mass spectra were recorded on a ThermoFinnigan LCQ Advantage (electrospray voltage 4.5 kV; capillary 250 ◦ C; sheath gas: N2 ) coupled to a ThermoFinnigan LC Surveyor pump utilizing the HPLC systems. The MS was controlled and total ion chromatograms and mass spectra were recorded using ThermoFinnigan Xcalibur software (San Jose, CA, USA). Helium was used to improve trapping efficiency and as the collision gas for CID experiments. The relative collision energies for MS/MS analyses were set at 30% (according to a relative energy scale). For the LC–MS/MS analyses, a 25 cm × 3.0 mm, 5 ␮m Luna C18 (2) Phenomenex chromatographic column was used. HPLC analyses were carried out using a Gynkotek HPLC system with UVD340U Gynkotek HPLC Pump Series LPG-3400A and thermostat (Gynkotek Separations, H. I. Ambacht, The Netherlands). The software package Chromeleon 6.0 (Gynkotek Separations) was applied for the data acquisition. For the CCC fraction analysis by HPLC, a 10 cm × 2.1 mm, 2.7 ␮m Supelco (C-18) column was used. 2.4. Solvent systems The solvent systems initially investigated for the HPCCC separation are listed in Table 2 and were divided into two groups: A – highly polar solvent systems containing ammonium sulphate salt (AI–AIII) and B – ion-pair solvent systems containing an ion-pair

agent (BI–BIV). The ion-pair solvent systems were prepared in a separator funnel by mixing appropriate solvents then, after equilibration, the phases were separated and sonicated before HPCCC separations. The biphasic highly polar solvent systems, containing saturated ammonium sulphate solution, were prepared as described by Fahey et al. [19]. Saturated ammonium sulphate was made by dissolving the salt in boiling water, letting this cool down to 78 ◦ C and decanting the supernatant. The saturated ammonium sulphate was then mixed with the remaining solvents in the ratio described in Table 2. The solvent systems were equilibrated at 20 ◦ C and then the phases were separated and sonicated in order to remove dissolved gases.

2.5. Separation of betalains by HPCCC Determination of stationary phase retention for each solvent system and preliminary separation studies were performed on the analytical Mini HPCCC instrument. The solvent systems were prepared according to Section 2.4. The Mini HPCCC instrument was run at a flow rate of 0.25 ml/min in both normal phase (systems AI–AIII) and reversed phase (systems BI–BIV) modes. The sample (15 mg) was dissolved in 1.5 ml of stationary phase (systems AI–AIII) or mobile phase (systems BI–BIV). The choice of the injection solvent was primarily a result of betalain solubility. The chromatographic column was first entirely filled with the stationary phase and the mobile phase was pumped while the coil was rotating at 2049 rpm at constant temperature of 20 ◦ C. The retention of the stationary phase measured for each solvent systems was as follows: 64.3% (system AI), 42.3% (system AII) and 50.0% (system AIII), 52.6% (system BI), 49.0% (system BII), 69.8% (system BIII), and 60.5% (system BIV). The semi-preparative separation of the betalain mixtures was performed on the Spectrum HPCCC. Using either solvent system AI or BIV, the centrifuge was run at a flow rate of 1.0 ml/min.

Table 2 Composition of the solvent systems tested for betalains separation by HPCCC. System no.

Composition

A I II III

Highly polar solvent systems with salt 1-PrOH–ACN–saturated (NH4 )2 SO4 –H2 O (v/v/v/v, 1:0.5:1.2:1 EtOH–ACN-1–PrOH–saturated (NH4 )2 SO4 –H2 O (v/v/v/v/v, 0.5:0.5:0.5:1.2:1) EtOH-1–BuOH–ACN–saturated (NH4 )2 SO4 –H2 O (v/v/v/v/v, 0.5:0.5:0.5:1.2:1)

B I II III IV

Ion-pair solvent systems TBME-1–BuOH–ACN–H2 O (0.7% TFA) (v/v/v/v, 2:2:1:5) TBME-1–BuOH–ACN–H2 O (1.0% TFA) (v/v/v/v, 2:2:1:5) TBME-1–BuOH–ACN–H2 O (0.4% HFBA) (v/v/v/v, 2:2:1:5) TBME-1–BuOH–ACN–H2 O (0.7% HFBA) (v/v/v/v, 2:2:1:5)

A. Spórna-Kucab et al. / J. Chromatogr. B 941 (2013) 54–61

The sample (15 mg) was dissolved in 1.5 ml of stationary phase (system AI) or mobile phase (system BIV). As with the Mini CCC runs, the chromatographic coil was first entirely filled with the stationary phase and the mobile phase was pumped while the coil was rotating at 1600 rpm at constant temperature of 20 ◦ C. The retention of the stationary phase was measured on the Spectrum instrument as follows: 80.5% (system AI) and 81.2% (system BIV). The system AI enabled purification of 2-decarboxy-betanin/-isobetanin (3.24 mg), 2,17-bidecarboxybetanin/-isobetanin (3.42 mg) and neobetanin (0.47 mg) (mixture 1) plus 17-decarboxy-neobetanin (1.65 mg), 2,15,17-tridecarboxy2,3-dehydro-neobetanin (1.88 mg), 2-decarboxy-neobetanin (6.39 mg) and 2,15,17-tridecarboxy-neobetanin (0.87 mg) (mixture 2) and the system BIV was effective for 2,17-bidecarboxybetanin/-isobetanin (3.74 mg) and neobetanin (0.44 mg) (mixture 1) plus 2,15,17-tridecarboxy-2,3-dehydro-neobetanin (2.4 mg), 2,17-bidecarboxy-2,3-dehydro-neobetanin (1.5 mg) and 2,15,17tridecarboxy-neobetanin (1.2 mg) (mixture 2). The effluent from the outlet of the HPCCC was monitored using a UV-ViS detector (Gilson, Middleton, WI, USA) and collected into test tubes with a fraction collector at 6 min intervals (flow rate 0.25 ml/min and 1 ml/min). The elution-mode was stopped when all the pigments had been eluted as shown by the UV-ViS detector. Where necessary, this was followed by the extrusion-mode with the pumping of the stationary phase at a flow rate 0.5 ml/min (analytical scale) and 2.0 ml/min (semi-prep scale) during the coil rotation. 2.6. HPLC analysis (LC-DAD–ESI–MS/MS) To prevent dissolved salt in the fractions from adversely affecting the HPLC analysis, a precipitation of the salt bulk from the samples was accomplished with methanol. LC-DAD analyses of mixtures 1 and 2 and HPCCC fractions were carried out using a gradient elution mode at 40 ◦ C with methanol (A) and 2% aqueous formic acid (B) system: 5% A in B at 0 min, a gradient to 7% A in B at 2 min and 20% A in B at 8 min then 40% A in B at 10 min and 80% A in B at 12 min, returning to the start conditions in 0.6 min. For the LC–MS/MS analyses, a solvent system: 7% A in B at 0 min a gradient to 30% A in B at 35 min (A, methanol; B, 2% formic acid in water) was used. The injection volume was 70 ␮l and the flow rate was 0.5 ml/min (LC-DAD and LC–MS/MS systems). 2.7. Freeze drying The HPCCC fractions were diluted with deionised water because they contained large amounts of solvents. The diluted fractions were then frozen and lyophilized. The fractions containing higher amounts of solvents were partially evaporated by speed vacuum centrifuge at room temperature to minimize compound degradation and then freeze dried. 3. Results and discussion 3.1. Analysis of decarboxylated and dehydrogenated derivatives of betanin/isobetanin For the experiments, two different mixtures of betanin and its derivatives, differing in decarboxylation and dehydrogenation products, were obtained as a result of the different heating times of the acidified betanin extract (Table 1). Mixture 1 contained mainly betanin, isobetanin, neobetanin and decarboxy-betanins while mixture 2 consisted of decarboxy- and dehydro-betanins. The mixtures differed in the pigment polarities and physicochemical properties determining their chromatographic behavior.

57

For example, the compounds in mixture 1 were more unstable than in mixture 2 and the HPLC retention times of the decarboxylated and dehydrogenated derivatives were longer in comparison to their corresponding betacyanins, due to their lower polarity. Betanin, as well as 2-, 17-, and 2,17-bidecarboxybetanins detected by HPLC and LC–MS/MS were identified according to the standards isolated in previous studies [20], and were monitored according to their retention times, and ViS absorption maxima max (538, 533, 505, 507 nm for betanin, 2-monodecarboxy-, 17-monodecarboxy- and 2,17-bidecarboxybetanins, respectively). The other compounds were mostly tentatively identified based on their max (446, 394, 468, 409, 451, 483 nm for 17-decarboxy-neobetanin, 2,15,17-tridecarboxy-2,3dehydro-neobetanin, neobetanin, 2,17-bidecarboxy-2,3-dehydroneobetanin, 2,15,17-tridecarboxy-neobetanin and 2-decarboxyneobetanin, respectively) as well as their protonated molecular and fragmentation ions (Table 1) according to a previous discussion [21,22]. 3.2. HPCCC separations of betacyanins and their derivatives This study is a first attempt of a complete HPCCC separation of betacyanins and their decarboxylated/dehydrogenated derivatives obtained during thermal treatment of red beet juice. Finding an appropriate phase system for the successful CCC separation of polar betacyanins is problematic [6–9]. However, studies on betalains from Phytolacca americana [6] and Bougainvillea glabra [9] suggested that an effective separation of the less polar compounds (e.g. acylated-betacyanins) could be achieved in solvent systems with ion-pair reagents. The use of hydrophilic solvent systems containing ammonium sulphate for the purification of anionic glucosinolates from crude plant homogenates [19] suggested that they were appropriate systems for a purification of polar compounds, however a similar polar system consisting of EtOH–ACN–(NH4 )2 SO4 (satd. soln)–water (1:0.5:1.2:1, v/v/v/v) was unsuccessfully used for separation of betanin and isobetanin [11] as these compounds were co-eluted. In order to investigate the separation of new betacyanin groups (the decarboxylated and dehydrogenated derivatives) by HPCCC, both types of solvent systems were tested, i.e.: (A) Highly polar solvent systems containing a high concentration of ammonium sulphate to enable the formation of two phases in solvent systems containing water in both phases. (B) Ion-pair, aqueous-organic solvent systems including ion-pair reagents (TFA, HFBA). 3.2.1. Highly polar solvent systems containing ammonium sulphate 3.2.1.1. Analytical scale separation of decarboxylated and dehydrogenated betanins. The initial experiments were carried out in an analytical machine with highly polar solvent systems containing ammonium sulphate (Table 2, systems AI–AIII). The pH of these solvent systems is ca. 5.5 and at this pH betalains are more stable. The results of the separation of decarboxylated (Fig. 2) and dehydrogenated (Fig. 3) betanins in the three solvent systems are compared. In this mode of separation (tail-to-head), the mobile phase is the upper phase (organic phase) and the stationary phase is the lower phase (aqueous phase), therefore, the more hydrophobic compounds are eluted first as expected. The high concentration of ammonium sulphate in the aqueous phase enhances the retention of the stationary phase in CCC by increasing the difference in density between the two phases. The stationary phase retention in CCC influences peak resolution; the larger amount of the stationary phase in the coil the higher resolution. In system AI, the highly polar Bt/IBt (1/1 ) are retained longer in the HPCCC coil and are

58

A. Spórna-Kucab et al. / J. Chromatogr. B 941 (2013) 54–61

Fig. 2. Reconstructed HPLC chromatograms of betalains (mixture 1) in three high salt-solvent systems separated by analytical HPCCC (composition of the solvent systems, see Table 2).

eluted with 17-dBt/-dIBt (2/2 ) during “elution extrusion” process (Fig. 2). The separation of NBt (7) is very successful in the applied conditions. 2,17-dBt/-dIBt (4/4 ) are eluted as the first decarboxybetacyanins, partially resolved from 2-dBt/-dIBt (3/3 ). In the group of decarboxylated betacyanins (mixture 1) separated in solvent system AII, 2,17-dBt/-dIBt (4/4 ) are eluted as first, virtually coincident with 2-dBt/-dIBt (3/3 ). Neobetanin (7) is eluted next and is relatively pure. Yet the final four compounds 17-dBt/-dIBt (2/2 ) and Bt/IBt (1/1 ) show a considerable peak overlap, with none being pure. System AIII gives good results for 2,17-dBt/-dIBt (4/4 ) and 2dBt/-dIBt (3/3 ) separation, however, 17-dBt/-dIBt (2/2 ) and Bt/IBt (1/1 ) are eluted during “elution extrusion” process less separated than in system AI (Fig. 2). The elution order of betalains is the same for all solvent systems and mainly depends on their polarity. The more hydrophobic compounds are eluted first, followed by more polar pigments. The applied solvent systems have different polarity, the most polar being system AII, then system AI and system AIII. In system AII, betalains are eluted too fast and therefore with poor resolution. Neobetanin (7) is eluted significantly earlier in the most polar system AII than in the remaining systems. In the case of 2,17-dBt/-dIBt (4/4 ) and 2-dBt/-dIBt (3/3 ), the situation is similar. The separation of these pigments clearly depends on the polarity of the solvent systems and is more effective in less polar solvent systems AI and AIII. The efficiency of separation is also associated with retention of the stationary phase, which is the highest in systems

Fig. 3. Reconstructed HPLC chromatograms of betalains (mixture 2) in three high salt-solvent systems separated by analytical HPCCC (composition of the solvent systems, see Table 2).

AI and AIII. The retention of the stationary phase and polarity of the solvents presumably influence separation of 17-dBt/-dIBt (2/2 ) from Bt/IBt (1/1 ) which is the most effective in solvent system AI. In this case, the retention of the stationary phase is the highest (64.3%). The separation of dehydrogenated betacyanins (mixture 2) by analytical HPCCC is presented in Fig. 3. The best results are obtained for systems AI and AIII where the separation of the majority of dehydro-derivatives is quite effective, with a very good separation of 17-dNBt (5) from the rest of the compounds, as a result of the highest stationary phase retention in these solvent systems and due to lower polarity of 5. In solvent system AI, only a partial overlap in the group of 2,15,17-dec-2,3-dHNBt (6), 2,17-dec-2,3-dHNBt (8), 2,15,17-dNBt (9), and 2-dNBt (10) is observed. Solvent system AII, except for 17-dNBt (5), gives no pure fractions, with the other four peaks substantially overlapped. In this group of compounds tested in system AIII, only a relatively good separation of 2,15,17-dec2,3-dHNBt (6) and 2,15,17-dNBt (9) is observed. In system AII, the pigments are mostly co-eluted except of 17-dNBt (5) in spite of its early elution with the other compounds. The best results are obtained for system I where the retention of the stationary is the highest (Fig. 3). 3.2.1.2. Semi-preparative scale separation of decarboxylated and dehydrogenated betanins. The system AI was used to separate mixtures 1 and 2 using a semi-preparative machine. The applied solvent

A. Spórna-Kucab et al. / J. Chromatogr. B 941 (2013) 54–61

59

Fig. 4. HPCCC chromatogram of betalains (mixture 1) after the separation in the high salt-solvent system by semi-prep HPCCC (system AI, see Table 2). Peak numbers refer to compounds shown in Table 1.

system enables a significantly better separation of betalains due to a higher retention of the stationary phase (80.5%). The HPCCC instrument used has almost twice the coil length of the analytical instrument, plus a wide bore (1.6 mm) to reduce any plug flow effects. The HPCCC chromatogram monitored at 500 nm (system AI) for mixture 1 is shown in Fig. 4. The first two peaks (2,17-dBt/-dIBt (4/4 ) and 2-dBt/-dIBt (3/3 )) are partially resolved with a resolution of 0.67 and the third compound (NBt (7)) is eluted as a single peak. The most polar compound pairs (17-dBt/-dIBt (2/2 ) and Bt/IBt (1/1 )) appear to co-elute in fractions 4 and 5 (“elution extrusion” mode). However, the reconstructed HPCCC chromatogram (Fig. 5A) of mixture 1 shows a tendency of a separation of the pairs 1/1 and 2/2 . The extrusion of the column content results in fractions highly rich in either Bt/IBt (1/1 ) or 17-dBt/-dIBt (2/2 ) (Fig. 5A). Comparison of Fig. 2 and Fig. 5A and B demonstrates that polarity of the solvent systems is the main factor determining the resolution of the compounds. The separation of mixture 1 is not much effective despite considerably higher retention of the stationary phase on Spectrum CCC. The HPLC chromatograms of the crude mixtures and the purified fractions are depicted in Fig. 6. Fig. 7 demonstrates the HPCCC chromatogram of mixture 2 (monitored at 500 nm ) separated in system AI with only partial

Fig. 6. HPLC chromatograms of betalains (mixture 1) before separation by semiprep HPCCC (a) and selected fractions after the separation in the high salt-solvent system (AI, see Table 2) (b–f).

Fig. 7. HPCCC chromatogram of betalains (mixture 2) after the separation in the high salt-solvent system by semi-prep HPCCC (system AI, see Table 2). Peak numbers refer to compounds shown in Table 1.

Fig. 5. Reconstructed HPLC chromatograms of betalains after the separation in the high salt-solvent system (a – mixture 1, solvent system AI, b – mixture 2, solvent system AI) and ion-pair solvent system with 0.7% HFBA (c – mixture 1, solvent system BIV, d – mixture 2, solvent system BIV, see Table 2) by semi-prep HPCCC.

60

A. Spórna-Kucab et al. / J. Chromatogr. B 941 (2013) 54–61

Fig. 9. HPLC chromatogram of betalains (mixture 1) before separation by semi-prep HPCCC (a) and selected fractions after the separation in the ion-pair solvent system with 0.7% HFBA (BIV, see Table 2) (b–e). Fig. 8. HPLC chromatogram of betalains (mixture 2) before separation by semi-prep HPCCC (a) and selected fractions after the separation in the high salt-solvent system (AI, see Table 2) (b–f).

overlap of the first four components and a complete separation of 17-dNBt (5) eluted as a single irregular broad peak. Two partially resolved dehydrogenated betacyanins (2-dNBt (10) and 2,17-dec-2,3-dHNBt (8)) are eluted in the first and second peak, respectively, and are well separated from 2,15,17-dec-2,3-dHNBt (6) and 2,15,17-dNBt (9). Fig. 8 shows the HPLC chromatogram of the crude mixture 2 plus selected fractions from the HPCCC purification run. 3.2.2. HPCCC solvent systems containing ion-pair reagents (TFA, HFBA) In this study, four solvent systems (Table 2, systems BI–BIV) containing TFA or HFBA were compared. The presence of TFA or HFBA in the solvent systems at different concentrations influences the stationary phase retention (which is higher at a lower concentration of the acids). The best separation results for mixture 1 (Fig. 5C) were obtained in the system with 0.7% HFBA (system BIV) (despite a smaller stationary phase retention than in the system with 0.4% HFBA (system BIII)). Comparison of systems BIII and BIV (data not shown) leads to a conclusion that the amount of acid is a more significant factor determining the resolution of the compounds than the retention of the stationary phase. Most of betalains from mixture 1 are not separated with solvent systems BI, BII nor BIII at all (data not shown). TFA forms less lipophilic ion-pairs than HFBA with betalains which are eluted too early from the coil which makes their effective separation impossible. For further experiments on semipreparative scale, only system BIV was taken. The separation of the dehydrogenated pigments (mixture 2) on analytical scale is not successful in the solvent systems containing TFA (systems BI-BII) and all compounds are co-eluted (data not shown). HFBA creates more lipophilic ion-pairs than TFA, therefore, the presence of HFBA significantly shifts the analytes to the organic phase, improving KD values. However, comparing the systems BIII and BIV reveals that only system BIV could be useful for the dehydrogenated betacyanins separation (data not shown), considering that the amount of acid influences betalains separation. Based on the initial results obtained for the analytical systems, system BIV was used on a semi-preparative scale for the separation of the two groups of betacyanins (the retention of the stationary phase is 81.2% in the operated apparatus). For the mixture 1, almost pure fractions are obtained for the less polar neobetanin (7) and for

the more polar 2,17-dBt/-dIBt (4/4 ) (Fig. 5C). It can be noticed that neobetanin (7) is eluted much faster than betanin/isobetanin (1/1 ) and decarboxy-betacyanins (2/2 , 3/3 , 4/4 ). The faster elution of neobetanin (7) results from the weaker formation of ion-pairs with the anions due to a lower protonation of its structure. Bt/IBt (1/1 ) is eluted with just a minor contamination from NBt (7), however, its tailing peak co-elutes with unresolved pairs of 17dBt/-dIBt (2/2 ) and 2-dBt/-dIBt (3/3 ) (Fig. 5C). For the purification of 17-dBt/-dIBt (2/2 ) and 2-dBt/-dIBt (3/3 ) the high salt system AI (Fig. 5A) is recommended instead. The HPLC chromatograms of the crude injection material and selected purified fractions from mixture 1 can be seen in Fig. 9. Comparing Figs. 5C and 9, the elution profiles of betalains obtained from the CCC with ion-pair solvent systems (reversed mode) are completely different from the profiles observed in the HPLC system (working also in the reversed mode). For mixture 1, the following elution order in the HPLC system (Fig. 9) is usually observed: Bt (1), 17-dBt (2), IBt (1 ), 17-dIBt (2 ), 2-dBt/-dIBt (3/3 ), 2,17-dBt/-dIBt (4/4 ) and NBt (7), whereas in the CCC system it is: 7, 1/1 , 2/2 , 3/3 , and 4/4 (Fig. 5C), indicating that 1/1 and 2/2 are eluted as pairs in contrast to HPLC elution. The studied differences result from different effectiveness of the interactions between selected betalains and the ion-pair reagents, which influences their separation and elution order. In particular, the differences in ionization properties are observed between the decarboxylated and dehydrogenated betacyanins (e.g. very fast elution of NBt (7)). Elution of betalains in HPLC is based on their polarity, the more polar pigments are eluted as first. In CCC the situation is similar, more polar betanins and decarboxy-betanins are eluted depending on their polarity except NBt (7). Neobetanin (less positively charged pigment) (7) presumably does not create stabile ion-pairs with HFBA, therefore, its polarity is not significantly changed during the CCC separation. For mixture 2, a good separation of dehydrogenated betacyanins is observed in system BIV (Fig. 5D) on the semi-preparative scale. Interestingly, the compounds are eluted in different order than in the high salt system AI (Fig. 5B). The different elution order is observed due to different separation modes (tail-to-head versus head-to-tail) but also because of a formation of ion-pairs with betalains. Especially a completely different relative retention is observed for 17-dNBt (5) and 2,15,17-dec-2,3-dHNBt (6) in both the systems. Only a slightly higher overlap is observed for 17-dNBt (5) and 2-dNBt (10) in system BIV and a good separation of 2,15,17dec-2,3-dHNBt (6), 2,17-dec-2,3-dHNBt (8) and 2,15,17-dNBt (9) is accomplished.

A. Spórna-Kucab et al. / J. Chromatogr. B 941 (2013) 54–61

61

2,15,17-tridecarboxy-2,3-dehydro-neobetanin, 2,17-bidecarboxy2,3-dehydro-neobetanin, 2-decarboxy-neobetanin and 2,15,17tridecarboxy-neobetanin (mixture 2). In conclusion, the two solvent systems presented (highly polar solvent systems containing ammonium sulphate salt and ion-pair solvent systems) are capable of producing a number of pure components with HPCCC of betalains, opening up the possibility of utilizing these compounds commercially. Moreover, a combination of the two CCC solvent systems together with RP-HPLC, results in completely different elution orders and makes them a very versatile tool for the isolation of a pigment on demand. The elution profiles for the CCC and HPLC runs are significantly different, indicating their different modes of separation. In addition, the application of highly polar solvent systems with ammonium sulphate salt containing no toxic perfluorinated acids is a first step in the search for food-grade solvent systems, which would be applied in the food, cosmetic and pharmaceutical industries, taking advantage of betalains colorant, antioxidant and possible chemopreventive properties. Fig. 10. HPLC chromatogram of betalains (mixture 2) before separation by semiprep HPCCC (a) and selected fractions after the separation in ion-pair solvent system with 0.7% HFBA (BIV, see Table 2) (b–f).

Fig. 10 shows the HPLC chromatograms of the crude mixture 2 together with selected fractions of the pure components. As in the case of mixture 1, the elution order of dehydrogenated betacyanins in the CCC system (Fig. 5D) is different from their elution order in the RP-HPLC system (Fig. 10). In the CCC system, early elution of 2-dNBt (10) is observed, which in principle is eluted very late in the RP-HPLC systems. In addition, the elution of 2,15,17-dec-2,3dHNBt (6) is very late in the CCC system, contrasting with a fast elution in RP-HPLC. 4. Conclusions In this study, we have shown that the separation of betalains using highly polar solvent systems is possible and very effective for selected structures of betalains: 2-decarboxy-betanin/-isobetanin, 2,17-bidecarboxy-betanin/-isobetanin and neobetanin (mixture 1), and 17-decarboxy-neobetanin, 2,15,17-tridecarboxy-2,3-dehydroneobetanin, 2,17-bidecarboxy-2,3-dehydro-neobetanin, 2,15,17tridecarboxy-neobetanin and 2-decarboxy-neobetanin (mixture 2). This is the first report on preparative isolation of the mentioned compounds, using HPCCC with highly polar solvent systems. The study confirms that ion-pair solvent systems with HFBA are much more effective than those with TFA because betalains create more hydrophobic structures which are shifted to the organic phase improving their KD values. Moreover, the acid concentration influences the stationary phase retention. Increasing concentration of the acid can decrease differences of the density of the upper and lower phases as well as can enhance creation of emulsions. In spite of the lower stationary phase retention, the system with 0.7% HFBA is more effective than the system with 0.4% HFBA for the separation of 2,17-bidecarboxy-betanin/-isobetanin, betanin/-isobetanin and neobetanin (mixture 1) and 17-decarboxy-neobetanin,

Acknowledgements The research was supported by the European Union through the European Social Fund within “Cracow University of Technology development program – top quality teaching for the prospective Polish engineers; University of the 21st century” project (contract no.UDA-POKL.04.01.01-00-029/10-00). The financial support by the UK Higher Education Infrastructure Fund (HEIF4) is also gratefully acknowledged. References [1] Y. Ito, W.D. Conway (Eds.), High-Speed Countercurrent Chromatography (Chemical Analysis, vol. 132), Wiley-Interscience, New York, 1996. [2] Y. Ito, J. Chromatogr. A 1065 (2005) 145. [3] A. Berthod, Countercurrent Chromatography, Elsevier, Amsterdam, 2002. [4] D. Fisher, I.J. Garrard, R. van den Heuvel, J.A. Sutherland, F.E. Chou, J.W. Fahey, J. Liq. Chromatogr. Relat. Technol. 28 (2005) 1913. [5] P. Winterhalter, Trends Food Sci. Technol. 18 (2007) 508. [6] G. Jerz, T. Skotzki, K. Fiege, P. Winterhalter, S. Wybraniec, J. Chromatogr. A 1190 (2008) 63. [7] G. Jerz, S. Wybraniec, N. Gebers, P. Winterhalter, J. Chromatogr. A 1217 (2010) 4544. [8] S. Wybraniec, P. Stalica, G. Jerz, B. Klose, N. Gebers, P. Winterhalter, A. Spórna, M. Szaleniec, Y. Mizrahi, J. Chromatogr. A 1216 (2009) 6890. [9] S. Wybraniec, G. Jerz, N. Gebers, P. Winterhalter, J. Chromatogr. B 878 (2010) 538. [10] F. das Neves Costa, G. Guimarães, J. Sep. Sci. 33 (2010) 336. [11] A. Dagenhardt, P. Winterhalter, J. Liq. Chromatogr. Relat. Technol. 24 (2001) 1745. [12] S. Wybraniec, J. Agric. Food Sci. 53 (2005) 3483. [13] F. Stintzing, R. Carle, Trends Food Sci. Technol. 18 (2007) 514. [14] D. Strack, W. Steglich, W. Wray, Methods in Plant Biochemistry, 8, Academic Press, London, 1993, pp. 421. [15] S. Wybraniec, Anal. Bioanal. Chem. 389 (2007) 1611. [16] K.M. Herbach, F.C. Stintzing, R. Carle, J. Food Sci. 71 (2006) 41. [17] K.M. Herbach, F.C. Stintzing, R. Carle, Eur. Food Res. Technol. 219 (2004) 377. [18] S. Wybraniec, Y. Mizrahi, J. Chromatogr. A 1029 (2004) 97. [19] J.W. Fahey, K.L. Wade, K.K. Stephenson, F.E. Chou, J. Chromatogr. A 996 (2003) 85. [20] S. Wybraniec, B. Nowak-Wydra, Y. Mizrahi, Tetrahedron Lett. 47 (2006) 1725. [21] S. Wybraniec, T. Michałowski, J. Agric. Food Chem. 59 (2011) 9612. [22] S. Wybraniec, K. Starzak, A. Skopin ska, B. Nemzer, Z. Pietrzkowski, T. Michałowski, J. Agric. Food Chem. 61 (2013) 6465–6476.