Screening of Complex Natural Extracts by Countercurrent

Mar 30, 2009 - The rapid screening CCC protocol associated to the preparative ... Citation data is made available by participants in Crossref's Cited-...
0 downloads 0 Views 3MB Size
Download PDF
Anal. Chem. 2009, 81, 4048–4059

Screening of Complex Natural Extracts by Countercurrent Chromatography Using a Parallel Protocol Yanbin Lu,† Alain Berthod,*,‡ Ruilin Hu,† Wenyan Ma,† and Yuanjiang Pan*,† Department of Chemistry, Zhejiang University, Hangzhou 310027, Zhejiang Province, China, and Laboratoire des Sciences Analytiques, CNRS, Universite´ de Lyon, 69622 Villeurbanne, France In countercurrent chromatography (CCC) the choice of the liquid system is the heart of any separation. It corresponds to the selection of the mobile phase and the stationary phase at the same time. Any change in one phase composition induces a change in the other phase composition which renders the choice of the appropriate liquid system difficult and lengthy. A scale of compositions of the heptane-ethyl acetate-methanol-water quaternary liquid system was referred to by letters from A to Z and called the Arizona (AZ) liquid system. Each composition of the AZ system has the same heptane/ethyl acetate and methanol/water volume ratios. It is shown that there is a continuous polarity change from the hydrophilic A composition (ethyl acetate-water) to the hydrophobic Z (heptane-methanol) mixture by measuring the distribution constant KD of a known test mixture. For all compounds, the log KD is linearly increasing with the water content of the lower aqueous phase of the composition used. The slopes of the log KD versus percent H2O have very different values which means that the chromatographic selectivity changes with liquid system compositions. The AZ system was associated to the elution-extrusion method to design a procedure to identify rapidly the appropriate solvent composition able to fractionate correctly a complex natural extract. With the use of an integrated threecoil CCC column (40 mL each coil) able to test three AZ compositions in parallel, it is shown that the optimum AZ composition is found in half a day using less than a liter total volume of solvents. Two natural extracts are rapidly screened using the proposed protocol. An extract of Piper longum L. of intermediate polarity was fractionated in five usable portions using the 3/2/3/2 (Q) composition of the AZ system. A polar extract of Polygonum cuspidatum was also separated in five fractions using the 1/6/1/6 (D) composition. In both cases, a 140 mL CCC column was used for a direct scale-up transfer with the same liquid system. Purified fractions were subjected to an antioxidant activity assay and liquid chromatography with UV and * To whom correspondence should be addressed. E-mail: panyuanjiang@ zju.edu.cn (Y.P.); [email protected] (A.B.). † Zhejiang University. ‡ Universite´ de Lyon.

4048

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

mass spectrometry detection (LC-UV/MS) analysis to determine the molecular weight, number, and quantity of compounds in the active fractions. Four fractions of P. cuspidatum showed excellent antioxidant activity. They were rapidly produced at the milligram level by the 140 mL CCC column and fractionated by semipreparative high-performance liquid chromatography (HPLC) in individual compounds that were each identified by NMR and MS and reevaluated for confirmation of bioactivity. The rapid screening CCC protocol associated to the preparative capability of CCC allows for a fast identification and characterization of active compounds in natural products. Countercurrent chromatography (CCC) is a liquid chromatography working with a support-free liquid stationary phase.1-3 Biphasic liquid systems are used. CCC can work with concentrated samples since the injected solutes have access to the volume of the liquid stationary phase and not only to the solid stationary phase-liquid mobile phase interphase area. It is a preparative technique mainly used for large-scale separation and/or purification.4 The ability of a chromatographic system to separate solutes is called selectivity. In CCC, selectivity is adjusted working on the biphasic liquid system used for the separation. However, it is the critical point for an efficient separation. Indeed, selecting the biphasic liquid system is selecting the mobile phase and the stationary phase of the chromatographic process at the same time. The two liquid phases are not independent. Their respective compositions correspond to a liquid-liquid particular equilibrium. Any change in the composition of one liquid phase induces a composition change in the other phase.1-3,5 Gradient elution is difficult or even not directly possible. (1) Berthod, A., Ed. Countercurrent Chromatography. The Support-Free Liquid Stationary Phase; Comprehensive Analytical Chemistry, Vol. XXXVIII; Elsevier: Amsterdam The Netherlands, 2002. (2) Conway, W. D. Countercurrent Chromatography, Apparatus, Theory & Applications; VCH Publishers: Weinheim, Germany, 1990. (3) Ito, Y., Conway, W. D., Eds. High Speed Countercurrent Chromatography; Chemical Analysis, Vol. 132; Wiley: New York, 1996. (4) Berthod, A.; Billardello, B. In Advances in Chromatography; Brown, P., Grushka, E., Eds.; Marcel Dekker: New York, 2000; Vol. 40, Chapter 10, pp 503-538. (5) Mandava, N. B., Ito, Y., Eds. Countercurrent Chromatography; Chromatographic Science Series, Vol. 44; Marcel Dekker: New York, 1985. 10.1021/ac9002547 CCC: $40.75  2009 American Chemical Society Published on Web 03/30/2009

Different approaches were proposed to simplify the selection of the solvent system in CCC.6 The use of homogeneous scales of solvent compositions was first proposed by Oka et al.7 He designed an approach based on mixtures of five solvents: hexane, ethyl acetate, butanol, methanol, and water. This approach was simplified by Margraff that designed the AZ scale made of compositions referred by a letter from the more polar A composition (ethyl acetate/water) to the less polar Z, heptane/methanol, composition.8 This homogeneous scale was named the Arizona system for AZ, the postal code of the U.S. Arizona state.6 This system proved to be very useful in the fractionation of extracts from vegetable or biological origin.9 Medicinal plants have served as an important source of drugs for combating diseases since ancient times and are playing more and more important roles in clinical therapy because of their high pharmacological activity, low toxicity, and rare complication.10-12 Due to the distinct features, CCC is a method of choice to fractionate plant extracts in the search for active principles. However, the composition of any plant extract is rather complicated, containing an extremely large number of small molecules differing in molecular weight, structural class, and, most important, hydrophobicity. This high hydrophobicity variety presents a significant challenge for the standard CCC technique, where optimizing the most suitable biphasic liquid system is a major problem. Therefore, in this work, the Arizona scale polarity properties are investigated using a recently proposed test mixture covering a wide polarity range.13 Results indicated that the quaternary Arizona scale is very useful in rapid CCC fractionations. These liquid systems are associated with the elution-extrusion (EECCC) way to use a CCC column14-16 as well as with an integrated CCC apparatus equipped with three CCC columns working in parallel.17 The apparatus was modified reducing the CCC column volume. A rapid screening method for natural extracts is proposed. The method combines the advantages of the Arizona liquid system with the throughput of the adapted three-column CCC apparatus and the EECCC method. It was tested with two real natural extract samples, Piper longum L. and Polygonum cuspidatum, showing that the two complex extracts could be efficiently fractionated in a working day. The fractions were further evaluated by antioxidant bioassay and liquid chro(6) Renault, J. H.; Nuzillard, J. M.; Intes, O.; Maciuk, A. In Countercurrent Chromatography. The Support-Free Liquid Stationary Phase; Berthod, A., Ed.; Comprehensive Analytical Chemistry, Vol. XXXVIII; Elsevier: Amsterdam, The Netherlands, 2002; Chapter 3, pp 49-84. (7) Oka, F.; Oka, H.; Ito, Y. J. Chromatogr. 1991, 538, 99–105. (8) Margraff, R. In Centrifugal Partition Chromatography; Foucault, A., Ed.; Chromatographic Science Series, Vol. 68; Marcel Dekker: New York, 1994; pp 331-350. (9) Marston, A.; Hostettmann, K. J. Chromatogr., A 2006, 1112, 181–194. (10) Potterat, O.; Hamburger, M. In Progress Drug Research; Petersen, F., Amstutz, R., Eds.; Birkhauser Verlag: Basel, Switzerland, 2008; pp 46118. (11) Cordell, G. A. Phytochem. Rev. 2002, 1, 261–273. (12) Hostettmann, K.; Marston, A. Phytochem. Rev. 2002, 1, 275–285. (13) Freisen, J. B.; Pauli, G. F. J. Liq. Chromatogr. Relat. Technol. 2005, 28, 2777–2806. (14) Ingkaninan, K.; Hazekamp, A.; Hoek, A. C.; Balconi, S.; Verpoorte, R. J. Liq. Chromatogr. Relat. Technol. 2000, 23, 2195–2208. (15) Berthod, A.; Ruiz-Angel, M. J.; Carda-Broch, S. Anal. Chem. 2003, 75, 5886– 5894. (16) Berthod, A.; Freisen, J. B.; Inui, T.; Pauli, G. F. Anal. Chem. 2007, 79, 3371–3382. (17) Wu, S.; Lang, L.; Gao, Y.; Liu, X.; Liu, F. J. Chromatogr., A 2008, 1180, 99–107.

matography/mass spectrometry (LC/MS) characterization to confirm the active individual components, exhibiting a potential strategy for natural drug discovery program. EXPERIMENTAL SECTION Chemicals. 1,1-Diphenyl-2-picrylhydrazyl (DPPH), 2,6-di-tertbutyl-4-methylphenol (BHT), and resveratrol were purchased from Sigma, St. Louis, MO. The biphasic liquid systems were compositions of the heptane-ethyl acetate-methanol-water system. The organic solvents were obtained from SDS (Peypin, France, a Carlo Erba division) and/or Huadong Chemicals (Hangzhou, China). Water was deionized and distilled with a resistivity passing 18 MΩ cm-1. Table 1 lists the solutes selected as the generally useful estimation of solvent system (GUESS) mixture.13 The letter code previously used is given along with the octanol-water partition coefficient.13,16 They were obtained from the Sigma-Aldrich-Fluka group (L’Isle d’Abeau Chesnes, France) and used as received. Instrumentation. CCC. Three different CCC columns were used. The commercial analytical CCC column, the Spectrum HPCCC (Dynamic Extractions Ltd., Slough, SL1 4LP, U.K.) has a volume of 20 mL with a single coil made by winding 40 m of 0.8 mm i.d. poly(tetrafluoroethylene) (PTFE) tubing, and a counterweight. The β-value is defined as β ) r/R, where r is the coiled tubing radius and R is the revolution radius or the distance between the coil axis and holder frame central axis. β varied from 0.5 for the internal first tubing layer, with the revolution radius R ) 7 cm, to 0.80 for the external layer. The Spectrum column is able to retain high volumes of stationary phase because it can work with very high centrifugal field (up to 2000 rpm producing a gravitational field of about 3000 m s-2 or 300g). A classical Shimadzu high-performance liquid chromatography (HPLC) pump and UV detector (Shimadzu, Lyon, France) were used with a 500 µL Rheodyne injection valve. The Azur chromatographic data processing software was used to record the chromatogram (Datalys, Grenoble, France). An analytical-scale integrated parallel CCC apparatus was designed and manufactured by the Zhejiang University machine shop (Hangzhou, China). The three-column apparatus was recently fully described in the literature.17 The apparatus holds three independent identical multilayer coils in a symmetrical position around a rotary frame at a distance of R ) 6 cm (revolution radius) from the central axis of the centrifuge. The three-coil arrangement allows for a good balance without any rotary seal. Each coil is a 40 mL CCC column made by winding 22.6 m of 1.5 mm i.d. PTFE tube. β varied from 0.34 for the internal first tubing layer to 0.75 for the external layer. The three multilayer coils are connected as three independent CCC columns with parallel connecting flow tubes made of 0.75 mm i.d. PTFE tubing. Each coil undergoes an identical synchronous planetary motion: revolution around the centrifuge axis and rotation about its own axis at the same rotation speed in the same direction. The revolution speed of the apparatus can be regulated with a speed controller in the 0-1000 rpm range able to produce a maximum gravitational field of about 60g. The integrated parallel CCC system was equipped with three model 2W-2B constant-flow pumps (Beijing Xingda Equipment, Beijing, China), three Valco Cheminert injection valves (each with a 2 mL sample loop, model-C22, Valco Inc., Houston, TX), three HD-9704 UV spectrophotometers (Jingke Equipment, Shanghai, China), and three BSZ-100 fraction collectors. The N2000 data analysis systems Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

4049

Table 1. Structures and Physicochemical Properties of Test Solutes Used

(Institute of Automation Engineering, Zhejiang University, Hangzhou, China) was used to record and process the three CCC chromatograms. For large-scale fractionation, a 140 mL Ito scheme IV CCC column was used (Zhejiang University machine shop). The

apparatus has a single coil and a counterweight both held in a 12 cm rotor (revolution radius R ) 12 cm). The multilayer 140 mL coil was prepared by winding a 26.4 m of 2.6 mm i.d. PTFE tube. The β-value varied from 0.33 to 0.60. The rotational speed of the apparatus could be regulated with a speed controller in the range

Figure 1. Integrated three CCC column parallel setup able to work simultaneously with three different liquid systems. P ) pumps; C ) 40 mL CCC coils; D ) UV detectors. All three independent CCC columns rotate at the same rotation speed. 4050

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

of 0-1000 rpm producing a maximum g field of about 120g. Sample injection was accomplished through an injection valve with a 5 mL sample loop. Analytical HPLC-UV/MS Instrumentation. An Agilent 1100 system was used for all HPLC analyses (Agilent, Beijing). It was equipped with a G1311A Quatpump, a G1322 degasser, a G1314A UV detector, a Rheodyne 7725i manual injection valve with a 20 µL loop, and an Agilent Chemstation for data treatment. LC/ESIMS peak identification were performed using the above-described Agilent HPLC system coupled with a Bruker Esquire 3000 plus ion trap mass spectrometer (Bruker-Franzen Analytik, Bremen, Germany) equipped with an electrospray ionization (ESI). Instrument control and data acquisition were performed using Esquire 5.0 software. Preparative HPLC. The isolation of individual compounds was performed using a preparative Zorbax Eclipse XDB-C18 column (250 mm × 20 mm i.d., 5 µm) on a Waters HPLC system (Milford, MA) consisting of a Waters 600E pump, a model 7725i injection valve with a 1 mL loop, an in-line degasser, and a Waters 2996 photodiode array detector. Evaluation and quantification were made on an Empower data system (Waters). NMR. NMR data for the structure elucidation of purified compounds were acquired by utilizing a Bruker Advance DMX 500 MHz NMR system (Bruker, Rheinstetten, Germany) using tetramethylsilane (TMS) as internal standard. Natural Extract Preparation. About 5 kg of long pepper fruits (P. longum L.) were bought from a local Hangzhou open market and dried to constant mass at 55 °C in a vacuum oven and then pulverized. One kilogram of powder was extracted with 4 L of 95% ethanol for 2 h under reflux. The extraction procedure was repeated three times. The combined 12 L were concentrated to dryness producing 80 g of extract that was stored at 4 °C for subsequent CCC separations. P. cuspidatum was obtained from the Guilin Pharmaceuticals Group of China (Zhongshan Road, Guilin City, China). Dry roots of P. cuspidatum were ground, and about 100 g of ground material was extracted with 1 L of 95% ethanol for during 2 h. After evaporating to dryness, 16 g of extract was stored at 4 °C for further CCC fractionations. 2VC Elution-Extrusion Method. The EECCC method was initially proposed by Conway,2 first experimentally used by Ingkaninan et al.,14 and fully developed by Berthod and coworkers15,16 for the rapid screening of the polarity distribution of the analytes contained in a complex mixture. In short, the method is performed in two steps, each step using 1 column vol of each phase of the biphasic liquid system. Step 1 is the classical elution step lasting for exactly 1 column vol. The solute retention volumes are linked to their distribution ratios or partition coefficients, KD, by VR ) VM + KDVS

(1)

with VM and VS being the mobile and stationary phase volumes contained in the CCC column. We have VM + VS ) VC, the column volume. Consequently, for any CCC column configuration, the solute that partitions equally between the two liquid phase (KD ) 1) is eluted with a VC retention volume. After eluting 1 column vol of mobile phase, step 2 is started by simply changing the entering liquid phase for the other phase, which

was the stationary phase during step 1. As described,12 the normal chromatographic process goes on as long as the VM volume of mobile phase is not extruded out of the column. Next, the “stationary” phase is seen at the column outlet, and all compounds contained in the column are extruded. When a column volume of “stationary” phase has been passed in the column, there is no compound that could possibly remain in the CCC column. During step 1, the elution step, solutes with low partition coefficients 0 e KD e 1 + VM /VS VR e VC + VM (2)

are eluted with retention volumes

During step 2, the extrusion step, all remaining solutes with partition coefficients 1 + VM /VS e KD e ∞ VC + VM e VR e

are eluted with retention volumes

2VC (3) The resolution factor between two peaks 1 and 2 is classically expressed by RS ) 2

VR2 - VR1 Wb2 + Wb1

(4)

in which Wb is the peak width at base. The resolution equations are16 RS )

KD2 - KD1 √N 4 VM (KD2 + KD1) + VS 2

during the elution step (5)

and RS )

√N 2√VC

(

1 KD1

 ) 1 KD2

during the extrusion step (ref 16) (6) Figure 2 shows that the 2VC EECCC method allows relating the whole range of KD coefficients with a retention volume (bottom line). The peak separation is very good during the elution step with resolution factors (upper line) higher than 1.5. However, the resolution drops dramatically for high KD values that are rapidly indistinguishable at the end of the extrusion step. By analogy with reversed-phase liquid chromatography (RPLC), in step 1 the mobile phase can be the aqueous denser phase and the stationary phase (and extruding phase in step 2) can be the upper organic phase. In this situation, the polar solutes are eluted during step 1 elution and the less polar and apolar solutes are eluted during step 2 extrusion. Using only 1 column vol of each liquid phase, the 2VC EECCC method allows for a rapid screening of the polarity range of a complex mixture. In the present work, the parallel 2VC EECCC biphasic solvent system screening was performed as follows: Each CCC column was first filled with the corresponding upper phase as stationary phase. Then the coils were rotated at the desired speed, and Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

4051

Figure 2. 2VC EECCC method. Bottom line: KD solute coefficient plotted vs the corresponding retention volume; full blue line and dotted blue bar, elution step (eqs 1 and 2); dotted red line and red hatched bar, rapidly increasing KD during the extrusion step (eq 3). Top green line: resolution factor obtained between two solutes with KD2/KD1 ) 2; full green line, elution step (eq 5); dotted green line, extrusion step (eq 6). The RS ) 1.5 line corresponds to a baseline separation of the two compounds. CCC column volume: 40 mL; VM ) 8 mL; VS ) 32 mL; N ) 800 plates.

the three respective lower phases were pumped at the selected flow rate in the head-to-tail direction (reversed-phase mode) (Figure 1). When the hydrodynamic equilibrium was established in each column, the same sample solution was injected through three injection valves (Figure 1). The effluents were independently monitored at 254 nm and collected by three fraction collectors. After elution of 1 column vol of mobile phase (VCM ) VC ) 40 mL, with CM for classical mode), the stationary phase of each column (instead of the mobile phase) was pumped into the column without changing the flow rate, flow direction, or rotor rotation speed. This phase change marks the beginning of the extrusion step. All solutes are eluted out of each column after pumping another column volume of stationary phase; hence the 2VC EECCC naming of the method. HPLC-UV/MS Analysis. The crude extracts and collected CCC fractions were analyzed by HPLC-UV/MS. For P. longum L. analysis, a gradient elution was performed with methanol-water mobile phases and a Zorbax Eclipse XDB-C8 column (150 mm × 4.6 mm i.d., 5 µm particles, Agilent, Beijing). From 0 to 5 min a 70% methanol solution was used, between 5 and 13 min the methanol content was increased to 90%, between 13 and 25 min it was further increased to 95% and maintained 5 min. The flow rate was 1 mL min-1 monitoring the effluent at 254 nm. For P. cuspidatum analysis, a Dikma C18 column (250 mm × 4.6 mm i.d., 5 µm particles, Dalian) and the 0.5% acetic acid/methanol gradient elution was as follows: 0-24 min, 10-50% methanol; 24-39 min, 50-95% methanol. The flow rate of the mobile phase was 0.8 mL min-1. The effluents were monitored at 254 nm. The MS data were acquired on a Bruker Esquire 3000 ion trap mass spectrometer with an electrospray interface. The instrument was operated in negative mode. The ion source temperature was 250 °C, and needle voltage was always set at 4.0 kV. Nitrogen was used as the drying and nebulizer gases at a flow rate of 10 L min-1 and a backpressure of 0.2 MPa or 30 psi. Antioxidant Activity Assay. In vitro antioxidant activities of CCC purified fractions were screened by DPPH assay as previ4052

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

ously described.18 Briefly, 20 µL of sample solution (different concentrations in ethanol) and 180 µL of DPPH (150 µM) in ethanol were taken in a 96-well microplate and incubated at 37 °C for 30 min. The absorbance was measured at 517 nm by a microplate reader (Molecular Device, U.S.A.). Percent radical scavenging activity was determined by comparison with an ethanol-containing control and calculated by the following equation: I(%) ) 100(Ablank - Asample)/Ablank

(7)

where Ablank is the absorbance of the control reaction mixture excluding the test compounds and Asample is the absorbance of the reaction mixture with the tested compounds. IC50 values represent the concentration of compounds able to scavenge 50% of DPPH radicals and are expressed as means of three separate experiments. BTH and resveratrol were used as potent antioxidant reference controls. Structure Elucidation of Bioactive Compounds. The constituents of active fractions were isolated using a semipreparative HPLC system. The preparative chromatographic conditions applied to the separations were based on the elution profile observed during the HPLC separation. They were optimized for baseline separation of the compounds. The HPLC purification required ∼2 mg per injection, and the yield per compound was typically in the milligram range. The structural identification of isolated compounds was carried out by ESI-MS, 1H NMR, and 13C NMR analyses and compared with the literature data. RESULTS AND DISCUSSION Arizona Liquid System. The main advantages of the four solventssheptane, ethyl acetate, methanol, and watersare (i) UV transparency with UV cutoff shorter than 205 nm except for ethyl acetate (254 nm), (ii) low boiling points, (iii) low viscosity, and (iv) a reduced toxicity and relatively environmentally friendly character. The four-solvent compositions of the AZ scale have all an identical methanol over water ratio and heptane over ethyl (18) He, S.; Wu, B.; Pan, Y.; Jiang, L. J. Org. Chem. 2008, 73, 5233–5241.

Table 2. Compositions of the Arizona Scale Biphasic Liquid Systemsa % v/v in the lower phasesb

v/v/v/v letter

heptane

ethyl acetate

methanol

water

heptane

ethyl acetate

methanol

water

up/low phase ratio

up/low density difference

A B C D F G H J K L M N P Q R S T U V W X Y Z

0 1 1 1 1 1 1 2 1 2 5 1 6 3 2 5 3 4 5 6 9 19 1

1 19 9 6 5 4 3 5 2 3 6 1 5 2 1 2 1 1 1 1 1 1 0

0 1 1 1 1 1 1 2 1 2 5 1 6 3 2 5 3 4 5 6 9 19 1

1 19 9 6 5 4 3 5 2 3 6 1 5 2 1 2 1 1 1 1 1 1 0

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.05 0.06 0.07 0.10 0.11 0.12 0.18 0.40 0.50 1.20 2.20 3.10 5.80 9.90 25.6

7.8 8.3 5.9 8.4 7.1 9.2 8.4 10.5 11.3 14.1 16.3 18.3 19.2 19.1 16.8 16.3 14.7 12.6 11.5 9.6 8.0 3.5 0.0

0.0 4.6 8.4 13.0 13.7 17.9 21.3 25.2 27.8 33.1 36.7 39.9 43.8 48.1 56.3 59.6 63.6 68.8 72.3 74.7 77.6 82.3 74.4

92.2 87.1 85.8 78.6 79.2 72.9 70.3 64.2 60.8 52.8 46.9 41.7 36.9 32.7 26.7 23.6 21.2 17.4 14.0 12.6 8.6 4.3 0.0

0.88 0.92 0.965 0.96 0.95 0.95 0.945 0.91 0.88 0.84 0.80 0.72 0.69 0.68 0.68 0.70 0.735 0.76 0.78 0.775 0.77 0.71 0.45

0.086 0.093 0.101 0.104 0.108 0.110 0.111 0.115 0.120 0.130 0.138 0.146 0.150 0.153 0.153 0.154 0.153 0.150 0.147 0.139 0.127 0.112 0.083

a

A-G is the polar region I; H-M is the less polar region II; N-T is the low polar region III; U-Z is the apolar region IV. b Data from ref 20.

acetate ratio (Table 2). Also the more polar (A-M) and less polar (P-Z) compositions have solvent proportions inversed symmetrically toward the N (1/1/1/1) central liquid system. The chemical compositions of the upper and lower phases were determined by gas chromatography (GC)19 and/or GC and Karl Fisher titration.20 Another property of the AZ compositions is that they produce a similar volume of upper organic and lower aqueous phase. The volume ratio is given in Table 2. The 0.72 value for the N composition means that mixing 0.5 L of all four solvents produces about 0.84 L of upper phase and 1.16 L of lower aqueous phase with 0.84/1.16 ) 0.72. The phase density difference is also listed in Table 2. It was demonstrated that a high density difference allows for good liquid stationary phase retention in a given CCC column.1,3,21 Liquid System Composition and Solute Partition Coefficients. The AZ liquid system forms a scale of compositions with polarity decreasing from the A ethyl acetate-water binary system to the less polar Z methanol-heptane system. The GUESS test mixture was designed to study the capabilities of a given biphasic liquid system used in a CCC column.13 Since the full chemical composition of the two phases of the AZ system are known, the GUESS mixture can be used to follow the changes in partition coefficients associated with the changes in liquid system composition. The 20 mL Spectrum CCC column was used to measure the KD coefficients of the GUESS solutes with five AZ central compositions from L (heptane/ethyl acetate/methanol/water 2/3/2/3 v/v) to Q (3/2/3/2 v/v) in the reversed-phase mode (aqueous lower mobile phase). For accuracy, most solutes were injected one by one and eluted by classical elution at 2 mL/ (19) Garrard, I. J.; Janaway, L.; Fisher, D. J. Liq. Chromatogr. Relat. Technol. 2007, 30, 151–163. (20) Berthod, A.; Hassoun, M.; Ruiz-Angel, M. J. Anal. Bioanal. Chem. 2005, 383, 327–340.

Table 3. Slope, Intercept, and Regression Coefficient of the log KD versus Water Content (% v/v) of the Aqueous Lower Phase code

name

C caffeine V vanillin M coumarin Q quercetin F ferulic acid U umbelliferone Z salicylic acid N narigenin O carvone E estradiol log KD(L) vs log Ko/w log KD(Q) vs log Ko/w

slope

intercept

r2

KD (L)

KD (Q)

0.0253 0.0312 0.0375 0.0835 0.0453 0.0483 0.0228 0.0793 0.0453 0.0545 0.4102 0.2825

-2.124 -1.751 -1.662 -4.318 -2.798 -2.689 -0.752 -3.740 -1.405 -2.252 -0.621 -1.338

0.852 0.988 0.987 0.999 0.979 0.998 0.952 0.997 0.986 0.991 0.743 0.294

0.16 0.79 2.08 1.23 0.39 0.73 2.83 2.60 9.70 4.22

0.05 0.19 0.37 0.03 0.05 0.08 0.99 0.07 1.19 0.34

min up to 3 column vol (60 mL or 30 min) corresponding to a KD of 5. When the measurement was not possible either because the solute was eluting to close to the hold-up volume or when it did not elute within 30 min, the experiment was done again in the normal phase mode (organic upper mobile phase). With no exception, the KD coefficients of all solutes obey the relationship log KD ) a(%H2O) + b

(8)

over the L-Q range of compositions. The %H2O is the volume percentage of water in the lower phase, a value listed in Table 2. a and b are the slopes and intercepts of the regression lines. These values are listed in Table 3 along with the regression coefficients and the KD values obtained for the L and Q compositions. Such correlations between the log of solute distribution constants and liquid system composition have been observed and studied previously.2,3,20,22-24 It justifies the design and use of solvent scales in CCC. Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

4053

Figure 3. Rapid screening strategy of a complex mixture based on three parallel 2VC EECCC experiments with different biphasic liquid compositions of the AZ liquid system.

Figure 4. HPLC analysis of the studied natural extracts. Top chromatogram: column Zorbax Eclipse XDB-C8 150 mm × 4.6 mm, 5 µm particles, water/methanol gradient elution; 0-5 min methanol 70% v/v, increased to 90% between 5 and 13 min and to 95% between 13 and 25 min, flow rate 1 mL min-1; UV detection 254 nm. Bottom chromatogram: column Dikma C18 250 mm × 4.6 mm, 5 µm particles, gradient elution aqueous 0.5% v/v acetic acid solution/methanol with 0-24 min from 10% to 50% methanol, increased to 95% between 24 and 39 min and kept at 95% methanol, flow rate 0.8 mL min-1; UV detection at 254 nm.

The lowest regression coefficient was 0.852 obtained for caffeine, the most polar compound of the set (Table 2). The error is high due to the low retention volume of caffeine with the five solvent composition tested. Large differences in slope values were observed. The slope value of salicylic acid (0.023) is 4 times lower 4054

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

than that of quercetin (0.835, Table 3). Since the slopes are positive, it means that all solute elution volumes will decrease in (21) Berthod, A.; Schmitt, N. Talanta 1993, 40, 1489–1498. (22) Conway, W. D. J. Liq. Chromatogr. Relat. Technol. 2001, 24, 1555–1573.

Figure 5. Rapid screening strategy of a P. longum L. extract. (a-c) The three parallel 2VC EECCC experiments with the J, N, and S compositions (see Table 2). (d-f) Second run with the P, Q, and R compositions showing the desired fractionation in panel e. CCC column volume, VC ) 40 mL, VS g 22 mL; rotor rotation, 650 rpm; lower aqueous mobile phase flow rate, 2 mL/min in the head-to-tail (descending) direction (blue dotted bar) followed by upper organic phase at the same flow rate and direction (extrusion, red hatched bar); detection, UV 254 nm; 10 mg of crude extract injected in 2 mL of an equal amount of upper and lower phases.

the reversed-phase mode (lower aqueous mobile phase) when the lower phase water content decreases, i.e., the reference letter increases from L (52.8% water content) to Q (32.7%, Table 2). Since the slopes differ widely, the solute elution order will change with the solvent system used allowing for selectivity adjustment. Using the Table 1 compound codes and KD values listed in Table 3, the elution order obtained with composition L is C F UV Q M N Z E O; with composition Q it is QFCNU V EM Z O. The joined letters indicate indistinguishable (fused) peaks. As predicted by their high slope values, quercetin (Q) and narigenin (N) move from retained solutes with composition L to fast eluting solutes with composition Q. The measured log KD values were plotted versus the respective log Ko/w coefficient, the reference value in hydrophobicity.25 The slopes and intercepts for the L and Q compositions are listed in Table 3. The more polar composition L is loosely related to the octanol-water system (r2 ) 0.743) when the less polar composition Q containing much less water is not really correlated to this hydrophobicity scale (r2 ) 0.294, Table 3). Screening Strategy. Knowing the exponential change of solute partition coefficients with the liquid system composition, it is possible to design a rapid screening strategy to find the optimum composition for a given complex mixture. With the use of the 2VC EECCC method with the three-column CCC apparatus three AZ compositions can be tested simultaneously. Figure 3 illustrates the proposed strategy. The solvent proportions in the

AZ system are symmetrical toward the central N composition with an equal amount of heptane, ethyl acetate, methanol, and water (Table 2). We arbitrarily selected the N composition as the middle polarity starting point. The A-M polar compositions are further divided in two regions: region I, hydrophilic, from A to H and region II, less hydrophilic, from K to N. Similarly, the N-Z less polar compositions are divided in region III, less hydrophobic, from N to S and region IV, hydrophobic, from T to Z. The integrated three-coil CCC apparatus allows for three experiments running in parallel. A first run will simultaneously try the J, N, and S compositions using the 2VC EECCC method needing 40 mL of upper and 40 mL of lower phases of each three systems. At 2 mL min-1, the experiment duration will be 40 min. According to the results of these three experiments, the second run will also try simultaneously three AZ compositions in the adequate polarity region, e.g., K, L, and M, the intermediate hydrophilic region II, if the correct mixture development was found to be between J and N looking at the first runs. Another 120 mL (3 times 40 mL) of the upper and lower phases of the systems will be needed along with 40 min time to perform the experiments. A third run may be necessary to find the best composition (Figure 3). In about 2 h and using less than a total of 0.75 L of solvents, the correct AZ system for most natural extracts can be found as will be shown with two real examples. A technical drawback may be the significant bench space needed for the three pumps, three detectors, and especially, Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

4055

Figure 6. Rapid screening strategy of a P. cuspidatum extract. (a-c) The three parallel 2VC EECCC experiments with the J, N, and S compositions (see Table 2). (d-f) Second run with the polar A, D, and H compositions showing the desired fractionation in panel e. CCC column volume, VC ) 40 mL, VS g 22 mL; rotor rotation, 650 rpm; lower aqueous mobile phase flow rate, 2 mL/min in the head-to-tail (descending) direction (dotted bar) followed by upper organic phase at the same flow rate and direction (extrusion, hatched bar); detection, UV 254 nm; 5 mg of crude extract injected in 2 mL of an equal amount of upper and lower phases.

Figure 7. LC/MS (total ion current, TIC) chromatogram of the P. cuspidatum extract showing the compounds found in the five P.cus 2VC EECCC fractions (see Table 5). Column Dikma C18 250 mm × 4.6 mm, 5 µm particles, gradient elution aqueous 0.5% v/v acetic acid solution/ methanol with 0-24 min from 10% to 50% methanol, increased to 95% between 24 and 39 min and kept at 95% methanol, flow rate 0.8 mL min-1; ion trap MS detection, ESI, source 250 °C, needle 4 kV, N2 drying and nebulizer gas 10 L min-1, 0.2 MPa.

three fraction collectors associated to the dedicated three-coil HSCCC apparatus. The gain of method is the sure finding of the best AZ composition associated to the studied natural extract. The proposed protocol can be performed sequentially working with a small-volume CCC column. This way will need more time but much less equipment. Rapid Fractionating of Complex Natural Extracts. Two popular traditional Chinese herbal extracts were evaluated using the proposed parallel screening strategy. The HPLC analyses 4056

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

shown in Figure 4 indicates that the two natural extracts are complex and would normally require multiple chromatographic steps for exhaustive bioactive component studies.6 Figure 5 illustrates the solvent screening chromatograms of the crude P. longum extract with the parallel 2VC EECCC protocol. Figure 5a shows that the whole content of the extract is eluted at the end of the extraction step using the J hydrophilic system. Most of the extract is still eluted by composition N (Figure 5b), and a great part of the extract elutes rapidly with composition S. After

Table 5. LC/MS Data of P. cuspidatum 2VC EECCC Fractions II-V

Figure 8. Proposed strategy for parallel 2VC EECCC protocol applied to the process of drug discovery from natural resources. The blue area outlines the steps done with the CCC technique. Table 4. Results of Antioxidant Activity Assay As Determined by DPPH Radical Assay samples

DPPH radical IC50 (µg/mL)a,b

P.lon-I P.lon-II P.lon-III P.lon-IV P.lon-V P.cus-I P.cus-II P.cus-III P.cus-IV P.cus-V resveratrolc BTHc

126 >190 >200 >220 >220 55.0 22. 7 21.8 26.8 88.3 24.3 28.1

a IC50 values were expressed as means of three separate experiments. DPPH stands for 1,1-diphenyl-2-picrylhydrazyl radical. c These are the standard compounds for positive control. b

40 min and 3 × 80 mL, the conclusion of the first run is that the right AZ composition is between N and S. The extract belongs to region III, medium hydrophobic. The second run tests compositions P, Q, and R (Figure 5d-f). The P. longum extract is best fractionated by composition Q (3/2/3/2 heptane-ethyl acetatemethanol-water % v/v) that spread its content in five wellseparated fractions (Figure 5e). (23) Leo, A.; Hansch, C. J. Org. Chem. 1971, 36, 1539–1544.

samples

detected components per sample

retention time (min)

[M - H](Da)

P.cus-II P.cus-II P.cus-II P.cus-II P.cus-II P.cus-III P.cus-III P.cus-III P.cus-III P.cus-IV P.cus-IV P.cus-IV P.cus-IV P.cus-V P.cus-V P.cus-V P.cus-V P.cus-V P.cus-V P.cus-V P.cus-V

5 5 5 5 5 4 4 4 4 4 4 4 4 8 8 8 8 8 8 8 8

24.3 26.6 28.8 29.8 31.1 31.5 33.4 34.5 34.8 37.6 38.8 39.3 39.8 43.1 43.8 45.1 48.0 49.0 49.3 51.6 52.0

389 389 541 389 389 227 227 431 407 431 283 283 285 311 205 269 285 277 289 279 423

Subsequently, the rapid screening procedure was used again to study the fractionation of the crude ethanol extract of P. cuspidatum. Figure 6 shows that the P. cuspidatum sample is much more polar than the P. longum sample. A great portion of the extract is already eluted during the elution step with the more polar J solvent system (Figure 6a). The intermediate system N separates the sample in two parts: one part is weakly retained in the elution step, the other part remains in the extrusion step (Figure 6b). There is nothing remaining for the extrusion step with the hydrophobic system S (Figure 6c). Since the desired part of the P. cuspidatum is known to contain polar phenolic compounds,26 the second run was centered on region I and the hydrophilic AZ compositions. Figure 6e shows an acceptable fractionation and composition D (1/6/1/6 heptane-ethyl acetate-methanol-water % v/v) was selected for the P. cuspidatum extract. Once the appropriate liquid system is known, it is straightforward to scale up the separation using the 140 mL CCC column.6 In the present work, 100 mg of P. longum extract was injected and fractionated with the Q liquid system. Five fractions (P.lonI-V) were collected as indicated in Figure 5e. During each CCC run, the quantities of the collected fractions were ∼20 mg. With the optimized D liquid system, 50 mg of P. cuspidatum extract was fractionated using the 140 mL CCC column. A chromatogram very similar to that of Figure 6e was obtained allowing collecting five fractions (P.cus-I-V) in quantities of ∼10 mg each. Antioxidant Bioassay-Guided Fraction Analysis. The purified 10 fractions from these two complex natural extracts were first screened for their antioxidant activities by the DPPH radical assay, which is widely used for the evaluation of antioxidant (24) Freisen, J. B.; Pauli, G. F. J. Chromatogr., A 2007, 1151, 51–59. (25) Sangster, J., Ed. Octanol-Water Partition Coefficients: Fundamentals and Physical Chemistry; Wiley Series in Solution Chemistry, Vol. 2; John Wiley & Sons: Chichester, U.K., 1997. (26) Pan, Y.; Zhang, X.; Wang, H.; Liang, Y.; Zhu, J.; Li, H.; Zhang, Z.; Wua, Q. J. Food Chem. 2007, 105, 1518–152.

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

4057

Table 6. Structure Elucidation and Antioxidant Activity of Purified Compounds

a

Same conditions as Table 4. b Dissolved in DMSO-d6.

activities of natural products.27-29 The results are summarized in Table 4. Obviously, the P. longum fractions showed relative weak antioxidant activities with IC50 > 200 µg/mL, whereas the P. cuspidatum fractions exhibited potent antioxidant activity against DPPH radical compared to the reference BTH and resveratrol controls. Among them, fractions P.cus-II, -III, and -IV showed comparable or even stronger antioxidant activity than resveratrol and BTH. Therefore, the P. cuspidatum fractions were of great interest and were recognized as active fractions and subjected in a subsequent LC-UV/MS analysis and characterization. The results are shown in Figure 7, and the mass spectra of relevant peaks in each fraction are presented in Table 6. The results revealed that the purified fractions primarily contained 4-8 compounds (P.cus-I mainly consisted of hydrophilic large molecules), demonstrating the high fractionation efficiency of the proposed CCC method. It was necessary to point out that the sample injected for rapid EECCC fractionation in this study was the most crude plant extract without additional sample pretreatment procedures (only solvent extraction of solid plant materials was needed). Plant extracts of this kind always exhibited extensive complexity and contained a large number of molecules differing in molecular weight (from large tannins to small hydrocarbons) and structural classes (terpene, alkaloid, flavone, saponin, etc.).10-12 This high complexity is always a problem in the sampling step of conventional preparative solid-liquid chromatographic separations. Therefore, multiple sample pretreatments and tedious chromatographic steps were always needed. In contrast, combining the advantage of unique liquid stationary

phase and AZ composition-KD correlation, herein we rapidly and successfully fractionated the two studied crude extracts. This demonstrates the potential of the proposed parallel 2VC EECCC approach in the fractionation of complex natural extracts. Purification and Identification of Hit Compounds and Confirmation of Biological Activities. The antioxidant P. cuspidatum CCC fractions (P.cus-II-V) were further purified using simple isocratic semipreparative HPLC separations. The sample loading onto the column was ∼2 mg according to their solubility, and the recovery of each peak was up to 60%. The isolated compounds were identified by NMR and MS. 1H NMR spectra were recorded at 500 MHz, and 13C NMR spectra were recorded at 125 MHz with TMS and solvent signals as internal references. With the use of literature data,30-32 piceid, resveratroloside, resveratrol, antraglycoside B, and emodin were identified in the P.cus active fractions (Table 6). The purified compounds were also tested for DPPH antioxidant activity screening for confirmation. The results are presented in Table 5 and show that piceid, resveratroloside, and antraglycoside B exhibited excellent antioxidant activities. Synergistic effects were unlikely since the activity of the isolated compounds matches that of their original fractions, except for emodin that showed a relative weaker DPPH activity than that of its original P.cus-V fraction (Tables 4 and 6). The rapid parallel 2VC EECCC and individual compound isolation process as described here significantly reduced the time needed to purify active materials from natural sources.

(27) Lannang, A. M.; Komguem, J.; Ngninzeko, F. N.; Tangmouo, J. G.; Lontsi, D.; Ajaz, A.; Choudhary, M. I.; Ranjit, R.; Devkota, K. P.; Sondengam, B. L. Phytochemistry 2005, 66, 2351–2355. (28) Kolak, U.; Ozturk, M.; Ozgokce, F.; Ulubelen, A. Phytochemistry 2006, 67, 170–2175. (29) Luo, X.; Basile, M. J.; Kennelly, E. J. J. Agric. Food Chem. 2002, 50, 1379– 1382.

(30) Jayatilake, G. S.; Jayasuriya, H.; Lee, E.-S.; Koonchanok, N. M.; Geahlen, R. L.; Ashendel, C. L.; McLaughlin, J. L.; Chang, C. J. J. Nat. Prod. 1993, 56, 1805–1810. (31) Steglich, W.; Losel, W. Tetrahedron 1969, 25, 4391–4399. (32) Chen, D. C. Handbook of Reference Substance for Traditional Chinese Herbs; China Pharmaceutical Technology Publishing House: Beijing, 2000; p 4344.

4058

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

The proposed protocol is complementary to the established approaches used with natural products.14,33-35 It optimize the use of the relationship of solute partition coefficients with the AZ system composition associating it with the use of a parallel CCC apparatus. It is possible to perform a rapid screening of the optimum biphasic liquid system for target analytes in half a day using less than 2 L total volume of solvents. Moreover, combining the unique CCC advantage of the liquid stationary phase, the elution-extrusion method can improve effectively the width of the polarity range for CCC separations. Although needing a significant amount of equipment, this protocol is extremely suitable for the dereplication of complex crude plant extracts. Figure 8 summarizes a strategy associating the parallel 2VC EECCC protocol with high-throughput bioassay to optimize and rationalize the quest for active compound in natural products for drug discovery. CONCLUSION The potential of CCC to fractionate complex natural extracts is known. The association of the rational heptane-ethyl acetatemethanol-water Arizona scale with the elution-extrusion operating mode allowed designing the 2VC EECCC rapid screening protocol. Its use with an integrated three 40 mL column apparatus allows us to find the correct AZ composition able to fractionate any given natural extract in a half-day of work using less than a total volume of 1 L of combined solvents. Once the right liquid system is well-established, the scaling-up for preparative fractionation of compounds contained in the extract is straightforward as long as the preparative column is able to retain a correct amount of stationary phase. From the natural drug discovery point of view, the present 2VC EECCC method showed no problem in being associated with bioactivity assays. Such association is very promising to increase the throughput of finding new active molecules from plant extracts. Compari(33) Chang, C. D.; Armstrong, D. W.; Fleischmann, T. J. J. Liq. Chromatogr. 1994, 17, 19–32. (34) Snyder, J. K.; Nakanishi, K.; Hostettmann, K.; Hostettmann, M. J. Liq. Chromatogr. 1984, 7, 243–256. (35) Sutherland, I. A.; Fisher, D. J. Chromatogr., A 2009, 1216, 740–753.

son of the proposed CCC protocol with conventional flash chromatography methods turns to the advantage of CCC in terms of sample pretreatment, solubility problems (an important factor for sample loading), and fraction recovery. In addition, it was recently demonstrated20 that alkanes akin to heptane, such as hexane, iso-octane, or hexanes could be used instead of the more expensive heptane producing minor changes in the chromatograms. The use of the AZ system in CCC has been well-documented,9,36,37 but it is necessary to point out that it has a limitation. The system can be successfully used for separation of highly nonpolar to moderately hydrophilic compounds. However, it cannot be used for many charged and/or extremely polar biomolecules such as amino acids, peptides, sugars, nucleic acids, and proteins. When solubility problems are encountered with the AZ system, it was suggested to replace ethyl acetate by methyl-tert-butyl ether and/or methanol by acetonitrile,7 or even to use completely different polar biphasic liquid systems such as aqueous two-phase solvent systems (ATPS).38,39 Obviously, such changes in the selected liquid system will produce completely different solute partition coefficients and will impose a new rapid screening optimization. There is no objection to use the rapid screening method with other solvent scales. ACKNOWLEDGMENT Y.P. thanks the National Science Foundation of China for Grant No. 20775069, the Ministry of Education of China for Grant NCET06-0520, and the National Science Foundation of Zhejiang Province for Grant Z206510. A.B. thanks the French Centre National de la Recherche Scientifique for financial support through UMR5180 (P. Lanteri). Received for review February 3, 2009. Accepted February 27, 2009. AC9002547 (36) Pan, Y.; Lu, Y. J. Liq. Chromatogr. Relat. Technol. 2007, 30, 649–679. (37) Freisen, J. B.; Pauli, G. F. J. Agric. Food Chem. 2008, 56, 19–28. (38) Armstrong, D. W.; Menges, R.; Wainer, I. W. J. Liq. Chromatogr. 1990, 13, 3571–3581. (39) Foucault, A.; Nakanishi, K. J. Liq. Chromatogr. 1990, 13, 2421–2440.

Analytical Chemistry, Vol. 81, No. 10, May 15, 2009

4059