Analytica Chimica Acta 820 (2014) 176–186
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Two-dimensional preparative liquid chromatography system for preparative separation of minor amount components from complicated natural products Ying-Kun Qiu ∗ , Fang-Fang Chen, Ling-Ling Zhang, Xia Yan, Lin Chen, Mei-Juan Fang, Zhen Wu ∗ School of Pharmaceutical Sciences and the Key Laboratory for Chemical Biology of Fujian Province, Xiamen University, South Xiang-An Road, Xiamen, 361100, China
h i g h l i g h t s
g r a p h i c a l
a b s t r a c t
• Preparative MDLC system was developed for separation of complicated natural products. • Medium-pressure LC and preparative HPLC were connected by interface of SPE. • Automated multi-step preparative separation of 25 compounds was achieved by using this system.
a r t i c l e
i n f o
Article history: Received 4 January 2014 Received in revised form 12 February 2014 Accepted 16 February 2014 Available online 24 February 2014 Keywords: Two-dimensional preparative liquid chromatography Medium-pressure liquid chromatograph × preparative high-performance liquid chromatography Rheum hotaoense L Anthraquinones Stilbenes Rheumin
a b s t r a c t An on-line comprehensive two-dimensional preparative liquid chromatography system was developed for preparative separation of minor amount components from complicated natural products. Mediumpressure liquid chromatograph (MPLC) was applied as the first dimension and preparative HPLC as the second one, in conjunction with trapping column and makeup pump. The performance of the trapping column was evaluated, in terms of column size, dilution ratio and diameter-height ratio, as well as system pressure from the view of medium pressure liquid chromatograph. Satisfactory trapping efficiency can be achieved using a commercially available 15 mm × 30 mm i.d. ODS pre-column. The instrument operation and the performance of this MPLC × preparative HPLC system were illustrated by gram-scale isolation of crude macro-porous resin enriched water extract of Rheum hotaoense. Automated multi-step preparative separation of 25 compounds, whose structures were identified by MS, 1 H NMR and even by less-sensitive 13 C NMR, could be achieved in a short period of time using this system, exhibiting great advantages in analytical efficiency and sample treatment capacity compared with conventional methods. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Natural products have been a major resource for the investigation of naturally-occurring biologically active substances. The
∗ Corresponding authors. Tel.: +86 592 2189868; fax: +86 592 2189868. E-mail addresses: [email protected] (Y.-K. Qiu), [email protected] (Z. Wu). http://dx.doi.org/10.1016/j.aca.2014.02.023 0003-2670/© 2014 Elsevier B.V. All rights reserved.
traditional way of studying natural products, in most cases, is pretreatment and fractionation of a complex matrix, separation and isolation of the individual components using repeat liquid chromatography, in an off-line mode. The off-line approach is very easy but presents several disadvantages: it is time-consuming, operationally intensive, and difficult to automate and to reproduce. Moreover, sample contamination or formation of artifacts can occur [1]. Techniques developed in recent years for the purpose of
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
analysis and structure elucidation, such as HPLC-SPE-NMR [2] and on-line multidimensional liquid chromatography (MDLC) [3], have shown a significant acceleration in the isolation and identification of natural products. MDLC, usually refers to two-dimensional liquid chromatography (2D-LC), has played an important role in complex sample analysis. 2D-LC combines two independent columns with different separation mechanisms in tandem mode. Depending on whether components can be eluted from the first dimension into the second dimension directly or not, the methods can be classified as on-line [4–6] or off-line [7–9], respectively. Furthermore, depending on whether first dimensional components are eluted into the second dimension completely or not, the methods can be classified as comprehensive [10] or heart cutting [11], respectively. On-line comprehensive 2D-LC is an ideal mode for the analysis of complex samples. Automatic performance in the on-line mode avoids sample contamination, deterioration and personal error, and comprehensive separation enables maximum information to be obtained. Such separation systems have been extensively used to increase the number of compounds separated in complex samples encountered in proteomics, analysis of biopolymers and synthetic polymers, pharmaceutical and environmental analysis and in the analysis of naturally-occurring compounds [12–14]. Nevertheless, several problems occurred when MDLC is used. First, the term “analysis” is added as constraint to the MDLC definition. Preparative separations, with a goal of isolating material, were suggested to be excluded from the definition, because it was considered not providing analytical information at each step [15]. Secondly, substantial amounts of analytes are required in many cases while MDLC system used presently can not meet the demands. For example, 13 C NMR and hyphenated two dimensional (2D) experiments (like heteronuclear multiple bond correlation – HMBC), which are highly important for elucidation of structures of novel compounds, still require at least 100 g of analytes, even if using HPLC-SPE-NMR system or CapNMR probe [16]. The total injected amount of sample on the HPLC column should be in the range of a few decade-milligrams (10–100 mg). Normally an analytical MDLC system cannot handle such amounts of sample. Moreover evaluation of the efficacy and safety of nature products, which are a central issue for the lead compound investigators and pharmaprojects researchers, always demands one to provide adequate amount of individual pure compound in gram-grade.
177
Interface techniques, including trapping columns [17], sample loops [18], parallel columns [4] and vacuum solvent evaporation [19], were developed as the application of on-line comprehensive MDLC. Trapping columns are used in liquid chromatography to trap samples eluted from the first dimension under the first dimensional mobile phase. The samples trapped by trapping columns are eventually eluted by the mobile phases of the second dimension. Trapping columns can delay the elution of samples from the first dimension to the second dimension. Switching the mobile phase on the trapping column can avoid incompatibility between two-dimensional mobile phases. Thus, trapping column interface could be valuable to overcome the above-mentioned problems [20]. We report here an on-line comprehensive medium-pressure liquid chromatography (MPLC) × preparative HPLC system for preparative separation of natural products, where a mediumpressure column was applied as the first dimension and preparative HPLC column as the second one, interfaced by a makeup pump and a 10-port switching valve in conjunction with two reversed-phase trapping columns (Fig. 1). After evaluation in terms of column size, dilution ratio and trapping volume, as well as flow rate and system pressure from the view of medium pressure liquid chromatograph, 15 mm × 30 mm i.d. column packed with ODS was selected as trapping column. In this case, analytes present in the MPLC elution bands are captured online onto the trapping column. The analytes retained on the trapping column are then eluted into the preparative HPLC column process with the corresponding mobile phase. Higher sample processing capability was provided by the 1st dimension MPLC; meanwhile, separation ability was preserved by using a 250 mm × 20 mm i.d. preparative HPLC column as the 2nd dimension. Macro-porous resin enriched water extract of Rheum hotaoense L. was applied as model complex mixture to evaluate the operation and performance of this MPLC × preparative HPLC system. A practical procedure for selection of isolating condition was established, regarding the factors of sample weight, 1st dimensional medium-pressure column size, trapped volume of elution and preparative HPLC column volume. This MPLC × preparative HPLC integrated system was successfully applied in automated multi-step preparative separation of 26 compounds, 25 of which were identified by MS, 1 H NMR and even by less sensitive 13 C NMR, from the macro-porous resin enriched water extract of R. hotaoense, exhibiting great advantages in analytical efficiency
Fig. 1. Scheme for MPLC × preparative HPLC system.
178
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
and sample treatment capacity compared with conventional methods. 2. Experimental 2.1. Chemicals and materials HPLC grade solvents were purchased from Merck, Darmstadt, Germany. The dried roots of Rheum hotaoense L. were purchased from Lu-Yan Pharmaceuticals (Xiamen, China) and identified by Professor Quan-Cheng Chen (Xiamen University, Xiamen, China), by means of routine morphological identification, microscopical identification and physicochemical identification. A voucher specimen (20110501-DH) has been deposited at School of Pharmaceutical Sciences, Xiamen University. 2.2. Instrumentation All instruments used in this study are commercially available. The first dimensional MPLC separation was carried out on a Büchi C601 medium-pressure pump equipped with C-615 pump controller (Büchi Labortechnik AG, Flawil, Switzerland). Another mediumpressure pump was used for the addition of water as makeup fluid through a tee union (VICI, Schenkon, Switzerland) to dilute the elution from 1st dimensional MPLC. The second-dimension LC system using a Varian binary gradient LC system (Varian Inc. Corporate, Santa Clara, USA) containing two solvent deliver modular (PrepStar 218), a photodiode array detector (ProStar 335) and a fraction collector (ProStar 701). Preparative-HPLC control and data acquisition were performed by a Varian Star Workstation 6.41 (Varian Inc. Corporate, Santa Clara, USA). All analytic scale sample analysis was performed on an Agilent Technologies 1100 Series HPLC provided with an automatic injector and a diode array detection system.
bed volume, b.v.: 5.3 mL), 15 mm × 30.0 mm i.d. (Phenomenex Security Guard C18 column, b.v.: 10.6 mL) and 50 mm × 20.0 mm i.d. (Varian Dynamax Microsorb 100-5 C18, b.v.: 15.7 mL), were used for trapping capability evaluation. Five times the bed volume’s aqueous methanol eluent (E30, E50 and E70) was pumped at a flow rate of half bed volume per minute to the tee, in which the eluent was diluted by water into different ratio, and loaded onto the under-evaluated pre-column. Samples trapped on each column were finely eluted with MeOH and weighed for recovery rate (Table 1). 2.6. HPLC analysis of total extract The crude macro-porous resin enriched water extract (ET) were analyzed with a Cosmosil ODS-PAQ column (250 mm × 4.6 mm i.d., 5 m, Nakalai Tesque Co. Ltd., Kyoto Japan). The mobile phase was water (A) and acetonitrile (B) according to the following gradient program (v/v): 0 min 80% A 20% B, 25 min 60% A 40% B, and run back to 80% A 20% B when 30 min. The HPLC analysis performed at flow rate of 1 mL min−1 , and monitored at 254 nm by the Agilent HPLC system (Fig. 2). 2.7. Preparative HPLC isolation In order to compare with the conventional isolation method, the crude macro-porous resin enriched water extract of R. hotaoense (ET) was dissolved in water at concentration of 20 mg mL−1 and subjected to preparative Cosmosil ODS-PAQ column (250 mm × 20.0 mm i.d., 5 m, Cosmosil, Nakalai Tesque Co. Ltd., Kyoto Japan) for isolation, in which 60 mg of extract was loaded at each isolation. Water (A)/acetonitrile (B) were employed as mobile phase, at gradient mode: B from 20% to 40% for 50 min, and run back to 20% when 60 min. After 8 times of reduplicative isolation, 11 compounds were obtained (S0, as shown in Fig. 3).
2.3. Sample preparation In the present study, macro-porous resin enriched water extract from R. hotaoense L. was used as model complex mixture to evaluate the performance of the trapping column and the total system. The root of R. hotaoense (0.5 kg) was extracted with water twice and concentrated under reduced pressure at 40 ◦ C. The concentrate was passed through Diaion HP20 macro-porous resin (Mitsubishi Chem, Tokyo, Japan) column (600 mm × 65 mm i.d.), eluted with H2 O and 95% EtOH 6000 mL respectively. 95% EtOH eluent was collected and concentrated to afford 58.0 g of total extract (ET) for further separation. 10.1 g of the extract was subject to a ODS column (260 mm × 50 mm i.d.), eluted successively with 1500 mL of H2 O, 30% MeOH (E30), 50% MeOH (E50) and 70% MeOH (E70). The aqueous methanol eluent were reserved to evaluate the performance of the trapping column. 2.4. Interface between MPLC and preparative HPLC A medium-pressure pump was used for the addition of water as makeup fluid, which was passing through a tee to dilute the elution from 1st dimensional MPLC. The tandem MPLC and HPLC columns were interfaced by two equivalent trapping columns and an electronically controlled Valco Cheminert 2-position 10-port switching valve (model EDU10UW, VICI, Schenkon, Switzerland) that make up an integrated online column switching MPLC × preparative HPLC system (Fig. 1). 2.5. Trapping column and dilution parameter selection Three commercially available ODS pre-columns in different size, 15 mm × 21.2 mm i.d. (Phenomenex Security Guard C18 column,
2.8. MPLC × HPLC separation procedures The crude macro-porous resin enriched water extract of R. hotaoense (ET) was firstly isolated with high-capacity by MPLC column, after which the eluent was then concentrated online into the solid-phase trapping column and then transferred into HPLC for further high performance separation, through the valve-switching technique. Three 2D MPLC × preparative HPLC separation protocols Table 1 Recovery rate of samples trapped on the trapping columns (X¯ ± RSD, %, n = 3). Sample solutiona
Dilution ratio
Column I
Column II
Column III
E30
1:1 1:2 1:3
55.4 ± 2.5 61.1 ± 3.8 65.7 ± 3.0
53.9 ± 3.9 60.5 ± 4.8 65.7 ± 2.6
58.9 ± 3.2 62.1 ± 5.2 67.7 ± 4.3
E50
1:1 1:2 1:3
66.4 ± 3.5 83.6 ± 2.4 96.8 ± 2.6
63.7 ± 3.3 82.8 ± 1.8 96.0 ± 3.7
67.4 ± 1.9 84.5 ± 3.0 96.9 ± 2.7
E70
1:1 1:2 1:3
91.6 ± 0.8 95.8 ± 2.3 98.4 ± 1.1
87.1 ± 1.4 93.3 ± 0.9 97.8 ± 2.3
91.4 ± 1.7 96.4 ± 2.0 99.2 ± 1.1
a Preparation of sample solution: 10.1 g of the extract was subject to a ODS column (260 mm × 50 mm i.d.), eluted successively with 1500 mL of H2 O, 30% MeOH (E30), 50% MeOH (E50) and 70% MeOH (E70). The aqueous methanol eluent were reserved to evaluate the performance of the trapping column. Three commercially available ODS pre-columns in different size, column I (15 mm × 21.2 mm i.d.), column II (15 mm × 30.0 mm i.d.) and column III (50 mm × 20.0 mm i.d.), were used for trapping capability evaluation. Five times the bed volume’s aqueous methanol eluent (E30, E50 and E70) was pumped at a flow rate of half bed volume per minute to the tee, in which the eluent was diluted by water into different ratio, and loaded onto the under-evaluated pre-column. Samples trapped on each column were finely eluted with MeOH and weighed for recovery rate.
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
179
Fig. 2. HPLC analysis of crude macro-porous resin enriched water extract of R. hotaoense (ET). The analysis was performed on a Cosmosil ODS-PAQ column (250 mm × 4.6 mm i.d., 5 m, Nakalai Tesque Co. Ltd., Kyoto Japan). The mobile phase was water (A) and acetonitrile (B) in a linear gradient mode as follows: B from 20% to 40% for 25 min, and run back to 20% when 30 min. The flow-rate of the mobile phase was 1.0 mL min−1 and the effluents were monitored at 254 nm by a DAD detector. The column temperature was kept at 30 ◦ C.
(S1, S2 and S3) were carried out according to different strategies. All the 1st dimensional separations were performed at a similarity procedure, while the mobile phase composition and elution program used in the 2nd dimensional preparative HPLC separation were different. Firstly, each extract (0.5 g in S1, S2 separation, and 2.0 g in case of S3) dissolved in water (5 mL or 10 mL), was subjected to the first dimensional MPLC column, a medium-pressure column (300 mm × 22.0 mm i.d.) laboratory-packed with Cosmosil ODS gel (75 m, Nakalai Tesque Co. Ltd., Kyoto Japan). Then the 1st dimensional column was eluted by a stepwise elution at a flow rate of 1 mL min−1 . In the separation case of S1 and S2, fractions were eluted with a gradient of water (A) and methanol (B) according to the following gradient program (v/v): 0 min 80% A 20% B, 12 h 20% A 80% B. As for S3, the gradient elution time was extended to 24 h, and then an additional 4 h stepwise elution from 80% B to 100% B was added. Eluent from the 1st dimensional column was diluted 1:2 by water pumped via makeup pump (2 mL min−1 ) and trapped on one of the trapping column. The 2-position 10-port valve was switched every 60 min by an electronic timer, and the sample trapped on the trapping column was then eluted to the preparative Cosmosil
ODS-PAQ column for further second dimension separations, in which different mobile phases were applied. In the separation case of S1 the mobile phase was water (A) and methanol (B), gradient run from 0 to 50 min, 30% methanol increasing to 70% methanol, and run back to 30% methanol when 60 min. As for the separation cases of S2 and S3, water (A)/acetonitrile (B) were employed as mobile phase, at gradient mode: B from 20% to 40% for 50 min, and run back to 20% when 60 min. Accordingly, another gradient elution program was applied during the isolation of the last four fractions of S3, and it was: B from 40% to 100% for 45 min, and run back to 40% when 60 min. Eluent with pure chromatographic peak in the 2nd dimensional chromatography was collected by fraction collector into different fractions. Preparative HPLC chromatograms were recorded and cascaded by Varian Star Workstation v6.41 as shown in Figs. 4 and 5. 2.9. HPLC analysis of the purified fragments Most purified fragments isolated by the MPLC × HPLC system were also analyzed as the crude extract did. Some
180
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
Fig. 3. Conventional preparative HPLC isolation of the crude macro-porous resin enriched water extract of R. hotaoense (ET). The extract was dissolved in water at concentration of 20 mg mL−1 and subjected to preparative Cosmosil ODS-PAQ column (250 mm × 20.0 mm i.d., 5 m, Cosmosil, Nakalai Tesque Co. Ltd., Kyoto Japan), in which 60 mg of extract was loaded at each isolation. Water (A)/acetonitrile (B) were employed as mobile phase, at gradient mode: B from 20% to 40% for 50 min, and run back to 20% when 60 min. Eight reduplicative isolation procedures were performed.
purified fragments with lower polarity were assayed by water (A)/acetonitrile (B) in another linear gradient as: B from 40% to 100% for 25 min, and run back to 40% when 30 min. Those fractions with identical retention time were combined with each other and analyzed again by HPLC before further structural evaluation (Fig. 6). 2.10. MS and NMR experiments Identification of purified compounds was performed by MS and NMR. Samples were dissolved with 0.5 mL of DMSO-d6 in conventional 5 mm NMR tubes, and their 1 H and 13 C NMR spectra were taken on a Brucker Avance 400 FT NMR spectrometer with tetramethylsilane as an internal standard. ESI-MS were measured on an AB MD-SCIEX Advantage spectrometer. 3. Results and discussion 3.1. Evaluation of solid-phase enrichment and water-dilution interface between MPLC and HPLC Since trapping columns are used to collect and concentrate compounds eluted from the first dimension, and then elute them into the second dimension, they should possess two key characteristics. First, samples should be better retained on trapping columns than the corresponding first dimensional columns. Second, the samples trapped on trapping column should be eluted easily by the mobile phases of the second dimension [20]. Although the solid-phase extraction (SPE) has been successfully used for online analyte enrichment in both analytical HPLC analysis [21–25] and preparative HPLC fractions post-collection concentration [26], there must be some uncertainties in the solid-phase trapping step as a novel interface between MPLC and preparative HPLC. On the other hand, unlike common analytical MDLC, both trapping
columns and makeup pump are key components in this system, for the dilution of eluent with water overcomes the influence of solvent effect and make it possible to transfer the primary column eluent onto the secondary column. Therefore, in the present study, different portions of the macro-porous resin enriched water extract from R. hotaoense, representing a range of mixtures with different polarities and chemical properties, were used as model complex mixture to evaluate the performance of the trapping column under the above-mentioned experimental conditions. 3.1.1. Sorbent type and column size selection Since the 2nd dimensional HPLC is the main constraint on separation capacity and efficiency, the trapping column should use the same sorbent type with that of the 2nd column. And in order to preserve separation efficiency, the sorbent bed of pre-column is usually 1/25–1/5 of the corresponding preparative column during conventional HPLC separation. Furthermore, owing to the demand of being connected to the high pressure HPLC flow path, the trapping column should be pressure tolerable, which makes commercially available ODS pre-columns of preparative HPLC a good choice for this 2D MPLC × HPLC system. Then, three pre-column in different size and different diameterheight ratio, 15 mm × 21.2 mm i.d. (b.v.: 5.3 mL), 15 mm × 30.0 mm i.d. (b.v.: 10.6 mL), and 50 mm × 20.0 mm i.d. (b.v.: 15.7 mL), were chosen for evaluating. 3.1.2. Sample retention ability and dilution ratio In order to capture the sample eluted from the 1st dimensional column, another makeup pump is necessary in order to dilute the eluent with water. Results shown in Table 1 indicate that each enrichment column could achieve adequate trapping efficiency (>90% of recovery rate) for sample solution of E50 and E70. And the greater dilution ratio adopted, the lesser sample spillage undergone. As far as the sample with less retention ability (E30) was
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
181
Fig. 4. Two MPLC × HPLC separation protocols, using different mobile phase in the 2nd dimensional chromatography. MPLC condition: Each crude extract (ET, 0.5 g) dissolved in 5 mL water, was subjected to the first dimensional MPLC column, a medium-pressure column (300 mm × 22.0 mm i.d.) laboratory-packed with Cosmosil ODS gel (75 m, Nakalai Tesque Co. Ltd., Kyoto Japan), and eluted by a stepwise elution of water (A) and methanol (B) according to the following gradient program (v/v): 0 min 80% A 20% B, 12 h 20% A 80% B, at a flow rate of 1 mL min−1 . Solid-phase trapping interface: Eluent from the 1st dimensional column was diluted 1:2 by water via makeup pump (2 mL min−1 ) and trapped on one of the trapping column. The 2-position 10-port valve was switched every 60 min by an electronic timer. HPLC condition: The sample trapped on the trapping column was then eluted to the preparative Cosmosil ODS-PAQ column for further second dimension separations, in which different mobile phases were applied. The flow-rate of the mobile phase was 8.0 mL min−1 and the effluents were monitored at 254 nm by a DAD detector. (S1): The mobile phase was water and methanol, gradient run from 0 to 50 min, 30% methanol increasing to 70% methanol, and run back to 30% methanol when 60 min. (S2): Water and acetonitrile were employed as mobile phase, at gradient mode: acetonitrile from 20% to 40% for 50 min, and run back to 20% when 60 min. *Compounds with purity lower than 80% were not numbered in the chromatography.
concerned, much greater sample loss happened when it passed through the enrichment column. This is due to the existence of strong polar components, which can be eluted out of the ODS column by mobile phase comprising lower concentration of organic solvent. The solvent strength of lower organic content solution cannot be reduced efficiently by dilution of water, so that the increase of the dilution ratio contributed little to sample trapping efficiency in E30. Fortunately, the recovery rate of more than 60% (at dilution ratio of 1:2 and 1:3) for E30 is still acceptable in preparative separation. According to the experimental data, the dilution radio was advised to be 1:2. 3.1.3. Diameter-height ratio, trapping efficiency and system pressure As far as the diameter-height ratio was concerned, the data in Table 1 suggested that the column’s trapping efficiency was not mainly dependent on column diameter and height. However, MPLC is a medium-pressure preparative technique. In this study, the system pressure should be taken into account due to
the use of the solid-phase trapping interface, which may cause a rise in column pressure owing to the addition of water at the post-column stage. For preparative purposes, the larger diameter column employed in the solid-phase trapping interface can produce lower back-pressure even with higher postcolumn dilution ratios, which will provide much better system tolerance. For this reason, the 15 mm × 30.0 mm i.d. pre-column was selected for further study. 3.1.4. Desorption Apart from the trapping, the analyte desorption is also an important step affecting the outcome of subsequent HPLC experiments. In the present work, a same sorbent type was used both in the trapping column and the 2nd dimensional HPLC column. Therefore, the trapped analytes were washed out of the solid-phase trapping column in the forward-flush mode (Fig. 1). The analytes desorption in this preparative MDLC system is quite different from that in analytical MDLC. Firstly, solvents with higher strength, that can push the entire slug off the trapping column, is not a
182
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
Fig. 5. MPLC × HPLC separation (S3), in which 2.0 g of R. hotaoense extract (ET) was loaded. MPLC condition: R. hotaoense extract (ET, 2.0 g), dissolved in 10 mL water, was subjected to the first dimensional MPLC column, and eluted by a stepwise elution of water (A)/methanol (B) according to the following gradient program (v/v): 0 min 80% A 20% B, 24 h 20% A 80% B, 28 h 0% A 100% B, at a flow rate of 1 mL min−1 . Solid-phase trapping interface: Eluent from the 1st dimensional column was diluted 1:2 by water via makeup pump (2 mL min−1 ) and trapped on one of the trapping column. The 2-position 10-port valve was switched every 60 min by an electronic timer. HPLC condition: The preparative Cosmosil ODS-PAQ column was used, and water (A)/acetonitrile (B) were employed as mobile phase, at gradient mode: B from 20% to 40% for 50 min, and run back to 20% when 60 min (Fr. 1–Fr. 23); B from 40% to 100% for 45 min, and run back to 40% when 60 min for Fr. 24–Fr. 27.
good choice for desorption. For their solvent effect has a negative impact on the isolation of the 2nd dimensional HPLC. The traps are more suitable to be washed out of the trapping column, by the gradient mobile phase of the 2nd dimensional HPLC. Secondly, unlike back-flush mode commonly used in analytical MDLC, the forward-flush desorption manner is preferred in this system, for it can utilize the trapping column’s separation ability and extend the separation length of the ODS column. 3.2. HPLC analysis of the total extract The crude macro-porous resin enriched water extract of R. hotaoense (ET) was used as model complex mixture to evaluate the performance of the integrated MPLC × HPLC system. According to analysis HPLC chromatogram shown in Fig. 2, (E)-rhaponiticin (12) is the major component present in the crude extract, and (E)-piceatannol 3 -O-ˇ-d-glucopyranoside (8), chrysophanol 8-Oˇ-d-glucopyranoside (20) and rhapontigenin (22) also have a relative high content. Other compositions are in low concentration and poorly separated. 3.3. Conventional preparative HPLC isolation Conventional preparative HPLC isolation of the total extract (ET), lead to the isolation of 11 compounds, whose purity ranged from 78% to 97.5% (Table 2). This result was adequate, for this crude
extract sample is soluble in water, resulting in high efficiency resolution. But in most preparation cases, samples have to be dissolved in organic solvents, whose solvent effect have a negative impact on resolution, resulting in the failure of preparative isolation. 3.4. Integrated reverse phase MPLC × HPLC separation of R. hotaoense macro-porous resin enriched water extract The above studies show that it was feasible to use a 15 mm × 30.0 mm i.d. pre-column packed with ODS materials as trapping columns in the column-switching interface. Then the proposed integrated MPLC × HPLC protocol was applied to the separation of the crude extract of R. hotaoense (ET). 3.4.1. HPLC elution program, governing factor of the MPLC flow rate and interface switching time A gradient elution program up to 60 min was employed to the separation of the MPLC eluent fractions. Since that the trapping column possesses adequate trapping efficiency when the eluent from the 1st dimensional chromatography passing through the column was less than five time of its volume (63.5 mL), the flow rate of the MPLC was adjusted to 1 mL min−1 . Accordingly, the flow rate of the makeup pump should be 2 mL min−1 , at a dilution radio of 1:2. Meanwhile, the 2-position, 10-port valve switching time should be equal to the 2nd dimensional HPLC isolation time.
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
183
Fig. 6. HPLC analysis of isolated compounds. Most of the compounds were obtain from separation of S3, expect for compounds 7 and 10, isolated during S2 separation. The analysis was performed on a Cosmosil ODS-PAQ column (250 mm × 4.6 mm i.d., 5 m, Nakalai Tesque Co. Ltd., Kyoto Japan). The mobile phase was water (A) and acetonitrile (B) in a linear gradient mode as follows: B from 20% to 40% for 25 min, and run back to 20% when 30 min for compounds 1–25, B from 40% to 100% for 25 min, and run back to 40% when 30 min for compounds 26–28. The flow-rate of the mobile phase was 1.0 mL min−1 and the effluents were monitored at 254 nm by a DAD detector. The column temperature was kept at 30 ◦ C.
3.4.2. Sample weight In many reversed-phase medium-pressure liquid chromatograph separation cases, the extract sample is usually partitioned into about 10–30 fractions for further separation. Considering that the 20.0 mm i.d. preparative column can handle up to 50–100 mg of sample with relative high separation efficiency, 0.5–2.0 g of the macro-porous resin enriched water extract of R. hotaoense was subject to the first dimensional MPLC of the 2D system. The MPLC separation was carried out over a laboratory-packed ODS gel column (300 mm × 22.0 mm i.d.), after which the sample was divided into 12–28 fractions by the column-switching interface. Except for the first fraction, which was considered to be dead volume mobile phase and vented, other fractions were subject to the 2nd dimensional HPLC separation. It could be seen from Figs. 4 and 5 and Table 2 that, the greater amount of sample loaded onto the MPLC × HPLC system, the greater sample information can be achieved. For example, the structures of compounds 2, 3, 5, 11, 16, 23 and 24 (Table 3) were clarified by NMR after separation of S3 (sample weight 2.0 g), but their NMR signals were so weak that they could not be identified, although also isolated in S2 (sample weight 0.5 g). Even minor components could be obtained when higher amount of sample was applied in the separation. Compounds 26, 27 and 28, whose solubility in water and concentration in the extract is very low, were also isolated in the S3 separation. However, when the amount of loaded sample was increased, separation efficiency was not decreased substantially, because the sample was divided into more fractions by the 1st dimensional MPLC column before being subjected to HPLC. Only chromatography peaks of minor components that were poorly resolved from a major peak, such as compounds 7 and 10, could not be successfully isolated in high purity.
3.4.3. Mobile phase system used in 1st dimensional MPLC and 2nd dimensional HPLC MPLC × HPLC separation chromatogram (S1) using methanol/ water mobile phase system at both 1st dimensional MPLC and 2nd dimensional HPLC separation, as well as the chromatogram using different mobile phase system (S2 and S3) are shown in Figs. 4 and 5. It is clearly seem that, combination of different mobile systems in the two dimensional chromatography results in a considerable improvement in peak resolution. Only 8 compounds were identified when using a same mobile phase system in both dimensional separations, this result is worse than that of conventional preparative HPLC separation, which resulted in 13 compounds’ isolation. When using a different mobile system, components that were very difficult to separate in conventional preparative HPLC, such as compounds 2, 3, 5, 13, 15, 16, 17, 21, 23 and 24, were all well resolved. Thus, the combination of mobile systems is preferred in this reverse phase MPLC × HPLC system.
3.4.4. Compounds isolated After 12–28 h separation, 22–53 fractions were directly produced according to the MPLC × HPLC elution profile from 0.5 to 2.0 g of crude macro-porous resin enriched water extract of R. hotaoense using the proposed column-switching MPLC × HPLC protocol. The purified fractions isolated were then analyzed by HPLC and combined according to their retention time. As a result, 9–25 identifiable compounds were obtained at a purity of more than 80% in an automated multi-step separation, from 1.1 to 206.3 mg (Fig. 6 and Table 2). For both yields and purity, the MPLC × HPLC protocol is much better when compared with that of conventional preparative HPLC isolation.
184
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
Table 2 Purity and weight of isolated compounds by different separation strategies.a Compd.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
S0
S1
S2
S3
Purity (%)
Weight (mg)
Purity (%)
Weight (mg)
Purity (%)
Weight (mg)
Purity (%)
Weight (mg)
87.1 – – 78.0 – – 81.5 95.0 – 83.3 – 95.7 – 85.9 – – – 91.1 – 82.4 – 97.5 – – 91.3 – – –
2.5 – – 2.4 – – 1.9 12.8 – 6.1 – 50.2 – 3.5 – – – 2.0 – 2.3 – 7.6 – – 1.5 – – –
85.2 – – 67.3 – – – 88.2 – – – 92.4 – – 86.0 – – 82.3 84.6 96.5 – 75.8 – – – – – –
2.1 – – 2.2 – – – 15.6 – – – 48.2 – – 3.6 – – 1.8 8.0 1.4 – 8.0 – – – – – –
87.4 – – 84.5 – – 92.5 94.6 – 93.3 – 95.2 – 87.4 83.7 – 91.6 90.4 88.8 84.6 – 95.4 – – 92.6 – – –
2.7 – – 2.5 – – 1.7 31.1 – 6.7 – 111.5 – 3.6 5.5 – 3.3 4.2 14.7 12.5 – 7.7 – – 1.9 – – –
86.3 86.2 80.2 80.6 98.7 98.1 – 98.2 100.0 – 90.1 97.4 86.6 85.5 94.3 97.3 86.5 99.3 95.5 100.0 93.7 99.2 90.2 90.9 80.2 92.8 96.7 85.1
5.0 4.5 2.0 5.3 1.1 1.2 – 41.7 4.5 – 10.2 206.3 12.3 23.6 15.6 4.8 15.4 7.5 30.3 4.3 3.4 33.4 1.3 3.0 3.4 1.4 1.4 1.2
Those compounds with purity lower than 80% or weighed less than 2.0 mg were not listed here. a The relative contents in percentage were calculated with area normalization method.
Table 3 Identification of isolated compounds. Structure/name
Catechin Rheumin Vanillic acid 1,3-Dihydroxy-6,7-dimethy-aetone-1-O-ˇ-d-glucopyranoside a
Configuration of the alkene double bond.
Compd.
MW
E/Za
R1
R2
R3
1 4 8 10 11 12 13 14 15 17 19 21 22 23
568 406 406 420 390 420 420 572 244 572 404 228 258 242
E E E E E E Z E E E E E E Z
OH OH O-Glc O-Glc H OH OH OH OH OH H H OH H
Glc H H CH3 H CH3 CH3 CH3 H CH3 CH3 H CH3 CH3
Glc Glc H H Glc Glc Glc 6 -O-glloyl-glucopyranosyl H 2 -O-glloyl-glucopyranosyl Glc H H H
7 16 18 20 24 25 26 27 28
432 432 416 416 446 568 270 254 284
CH2 OH OH H CH3 OCH3 CH3 OH CH3 OCH3
H CH3 CH3 H CH3 H CH3 H OH
Glc Glc Glc Glc Glc 6 -O-glloyl-glucopyranosyl H H H
2 3 5 9
290 292 168 418
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
Scheme 1. Keto–enol tautomerism of compound 3.
3.5. Structure evaluation The chemical structures of isolated components were further identified by ESI-MS and 1 H NMR, even by the less-sensitive 13 C NMR (supporting information). The structure of some components with high purity, such as compound 6, were unable to be determined, owing to their minor content and only with 1 H NMR, MS data available. Anyhow, by comparing spectra data with those reported in literature, structures of 27 compounds were identified as shown in Table 3, and they were: (E)-piceatannol 3, 4 -O-ˇ-d-diglucopyranoside (1) [27], catechin (2) [28], rheumin (3) [29], (E)-piceatannol 3-O-ˇ-d-glucopyranoside (4) [30], vanillic acid (5) [31], aloe-emodin 8-O-ˇ-d-glucopyranoside (7) [32], (E)-piceatannol 3 -O-ˇ-d-glucopyranoside (8) [31], 1,3dihydroxy-6,7-dimethy-aetone 1-O-ˇ-d-glucopyranoside (9) [33], (E)-rhapontigenin 3 -O-ˇ-d-glucopyranoside (10) [31], (E)resveratrol 3-O-ˇ-d-glucopyranoside (11) [34], (E)-rhaponiticin (12) [35], (Z)-3,3 ,5-trihydroxy-4 -methoxystilbene 3-O-ˇ-dglucopyranoside (13) [31], (E)-rhaponticin 6 -O-gallate (14) [31], (E)-piceatannol (15) [36], emodin 1-O-ˇ-d-glucopyranoside (16) [27], (E)-rhaponticin 1 -O-gallate (17) [31], chrysophanol 1-Oˇ-d-glucopyranoside (18) [32], (E)-desoxyrhaponticin (19) [35], chrysophanol 8-O-ˇ-d-glucopyranoside (20) [32], (E)-resveratrol (21) [37], (E)-rhapontigenin (22) [35], (Z)-deoxyrhapontigenin (23) [38], physcion 8-O-ˇ-d-glucopyranoside (24) [39], chrysophanol 8-O-ˇ-d-(6 -galloyl)-glucopyranoside (25) [40], emodin (26) [39], chrysophanol (27) [39], and physcion (28) [39]. 3.6. Keto–enol tautomerism of compound 3 By comparing the spectra in separation of Fr. 2 and Fr. 3 in S3, we found that compound 3 presented a longer retention time than that of compound 2. But when the eluate of compound 3 was evaporated and redissolved, its retention time changed, becoming even shorter than that of compound 1. A structural transformation must have occurred in compound 3, and keto–enol tautomerism (Scheme 1) might explain the change of retention time. Keto-form of compound 3, which presents less-polar and higher-retention, was captured by the MPLC × HPLC system even it was in a lower concentration. Enol-form of compound 3, although being much stable than keto-form, was ignored during the separation procedure, for its retention time was so close to that of solvent peak. But when compound 3 was disposed for further HPLC and NMR test, only the stable enol-form, could be observed clearly.
185
demonstrated that the quantity of compound information obtained is higher compared to single-column MPLC or HPLC chromatography. Although the MPLC column, trapping column, and HPLC column all were the same stationary phase (ODS), adequate separation resolution was achieved by using different mobile phase systems in the 1st and 2nd dimensional columns, so that some pairs of peaks that are poorly resolved in one separation mode were easily separated using the MPLC × HPLC approach. If different stationary phases are employed in two different dimensions, higher resolution may be obtained. For example, choosing perfluorophenyl column for the second dimension, or a Sephadex LH 20 column as the first dimension, may afford separations that are not obtained on the ODS dimension. In addition, the isolated compounds were identified by NMR spectra. If more sensitive NMR equipment such as HPLCSPE-NMR system or CapNMR probe were applied, lower abundance compounds could also be identified. On the other hand, reversed-phase preparative HPLC isolation for lipophilic samples is always a difficult problem. Sample dissolution is the major issue particularly for crude extracts and fractions. Dissolution of the sample in the mobile phase is often not possible. To circumvent the solubility problem, samples are dissolved in the mobile phase component of high elution strength. However, this may considerably affect the chromatographic resolution. The integrated MPLC × HPLC system solved the problem. Firstly, solid phase sample introduction, where samples are adsorbed onto stationary phase and loaded to column for elution, can be easily applied in the 1st dimensional MPLC column used in this system [41]. Secondly, the solid-phase interface between the MPLC and HPLC, which overcomes solvent effect by diluting the MPLC eluent with water, enables the hydrophobic compounds’ isolation by preparative HPLC with higher resolution. In conclusion, the proposed protocol is our initial exploration into the preparative and on-line comprehensive MPLC × HPLC technique. The integrated MPLC × HPLC system provided a considerable improvement in peak capacity when compared with HPLC chromatography, giving very efficient separation of the crude extract of R. hotaoense, resulting in 25 compounds’ identification in an automated multi-step preparative separation. It can also be applied to isolate other comprehensive multi-component natural products. The system also captured the keto-enol tautomerism phenomenon of rheumin (3) for the first time, which is a valuable application in exploration of structural transformation in unstable components. Acknowledgments The project was supported by the National Natural Science Foundation of China (No. 81102333 and No. 81072549) and Fundamental Research Funds for the Central Universities (No. 2010121108). The authors would also like to acknowledge the financial supports from Fujian Natural Science Foundation for Distinguished Young Scholars (No. 2012J06020). We acknowledge Prof. D.M.J. Lilley of University of Dundee, for his kindly help on manuscript revise.
4. Conclusions
Appendix A. Supplementary data
In summary, we have shown that the online column-switching MPLC × HPLC coupling is possible in terms of peak concentration prior to further HPLC separation. In conjunction with a solid-phase trapping interface, the system provides simpler and more efficient separation without requiring extra steps to remove organic solvent when compared with conventional off-line MPLC–HPLC separation. It achieved automated high-capacity and high-resolution isolation using only commercially-available equipment. The application of this system to the preparative isolation of a real sample has
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.aca.2013.12.001. References [1] P. Dugo, O. Favoino, R. Luppino, G. Dugo, L. Mondello, Analytical Chemistry 76 (2004) 2525. [2] M. Sandvoss, B. Bardsley, T.L. Beck, Magnetic Resonance in Chemistry 43 (2005) 762. [3] D.R. Stoll, X.P. Li, X.L. Wang, Journal of Chromatography A 1168 (2007) 3.
186
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
[4] K. Wagner, T. Miliotis, G. Marko-Varga, R. Bischoff, K.K. Unger, Analytical Chemistry 74 (2002) 809. [5] G.J. Opiteck, K.C. Lewis, J.W. Jorgenson, R.J. Anderegg, Analytical Chemistry 69 (1997) 1518. [6] P. Dugo, O. Favoino, R. Luppino, G. Dugo, L. Mondello, Analytical Chemistry 76 (2004) 2525. [7] P. Dugo, O. Favoino, P.Q. Tranchida, G. Dugo, L. Mondello, Journal of Chromatography A 1041 (2004) 135. [8] Y. Liu, X. Xue, Z. Guo, Q. Xu, F. Zhang, X. Liang, Journal of Chromatography A 1208 (2008) 133. [9] X. Chen, L. Kong, X. Su, C. Pan, M. Ye, H. Zou, Journal of Chromatography A 1089 (2005) 87. [10] M. Eggink, W. Romero, R.J. Vreuls, H. Lingeman, W.M.A. Niessen, H. Irth, Journal of Chromatography A 1188 (2008) 216. [11] L.M. de Souza, T.R. Cipriani, C.F. Sant’Ana, M. Iacomini, P.A.J. Gorin, G.L. Sassaki, Journal of Chromatography A 1216 (2009) 99. [12] S.P. Dixon, I.D. Pitfield, D. Perrett, Biomedical Chromatography 20 (2006) 508. [13] Q. Wu, H.M. Yuan, L.H. Zhang, Y.K. Zhang, Analytica Chimica Acta 731 (2012) 1. [14] X. Zhang, A.Q. Fang, C.P. Riley, M. Wang, F.E. Regnier, C. Buck, Analytica Chimica Acta 664 (2010) 101. [15] L. Blumberga, M.S. Klee, Journal of Chromatography A 1217 (2010) 99. [16] J.F. Hu, G.R. Eldridge, Y.H. Yu, Progress in Chemistry 20 (2008) 429. [17] K. Im, H.-W. Park, Y. Kim, B. Chung, M. Ree, T. Chang, Analytical Chemistry 79 (2007) 1067. [18] L. Hu, X. Chen, L. Kong, X. Su, M. Ye, H. Zou, Journal of Chromatography A 1092 (2005) 191. [19] H. Tian, J. Xu, Y. Xu, Y. Guan, Journal of Chromatography A 1137 (2006) 42. [20] L.W. Cao, D.H. Yua, X.L. Wang, Y.X. Ke, Y. Jin, X.M. Liang, Analytica Chimica Acta 706 (2011) 184. [21] J.S. Fritz, M. Macka, Journal of Chromatography A 902 (2000) 137.
[22] A. Cappiello, A. Berloni, G. Famiglini, F. Mangani, P. Palma, Analytical Chemistry 73 (2001) 298. [23] C. Seger, M. Godejohann, L.H. Tseng, M. Spraul, A. Girtler, S. Sturm, H. Stuppner, Analytical Chemistry 77 (2005) 878. [24] K. Kato, M.J. Silva, L.L. Needham, A. Calafat, Analytical Chemistry 78 (2006) 6651. [25] C. Clarkson, M. Sibum, R. Mensen, J.W. Jaroszewski, Journal of Chromatography A 1165 (2007) 1. [26] J. Stevens, M. Crawford, G. Robinson, L. Roenneburg, Journal of Chromatography A 1142 (2007) 81. [27] K.K. Sung, Archives of Pharmacal Research 23 (2000) 159. [28] G. Gelles, F. Abdellatif, M. Nadine, K. Manuela, T. Raffaele, P. Roger, P. Vincent, Phytochemistry 52 (1999) 759. [29] J.L. Li, A.Q. Wang, J.S. Li, W.Y. He, M. Kong, Chinese Traditional and Herbal Drugs 29 (1998) 721. [30] Y. Kashiwada, G.I. Nonaka, I. Nishioka, Chemical and Pharmaceutical Bulletin 32 (1984) 3501. [31] Y.K. Qiu, Y.J. Chen, Y.P. Pei, H. Matsuda, M. Yoshikawa, Chemical and Pharmaceutical Bulletin 50 (2002) 1507. [32] J.L. Li, A.Q. Wang, Chinese Traditional and Herbal Drugs 31 (2000) 321. [33] L. Zhou, M. Lin, J.B. Li, Acta Pharmaceutica Sinica 29 (1994) 107. [34] P.W. Teguo, A. Decendit, J. Vercauteren, G. Deffieux, J.M. Mérillon, Phytochemistry 42 (1996) 1591. [35] A.Q. Wang, J.L. Li, Z.Z. Wu, Chinese Traditional and Herbal Drugs 32 (2001) 878. [36] X.X. Li, R.A. Yan, H.Y. Duan, Fine Chemicals 28 (2011) 475. [37] F. Mattvi, F. Reniero, S. Korhammer, Journal of Agricultural and Food Chemistry 43 (1995) 1820. [38] W. Zhang, M.L. Go, European Journal of Medicinal Chemistry 42 (2007) 841. [39] A.Q. Wang, J.L. Li, J.S. Li, Chinese Traditional and Herbal Drugs 41 (2010) 343. [40] H. Matsuda, T. Morikawa, I. Toguchida, J.Y. Park, S. Harima, M. Yoshikawa, Bioorganic and Medicinal Chemistry Letters 9 (2001) 41. [41] P. Webera, M. Hamburgera, N. Schafrothb, O. Potterat, Fitoterapia 82 (2011) 155.
Contents lists available at ScienceDirect
Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca
Two-dimensional preparative liquid chromatography system for preparative separation of minor amount components from complicated natural products Ying-Kun Qiu ∗ , Fang-Fang Chen, Ling-Ling Zhang, Xia Yan, Lin Chen, Mei-Juan Fang, Zhen Wu ∗ School of Pharmaceutical Sciences and the Key Laboratory for Chemical Biology of Fujian Province, Xiamen University, South Xiang-An Road, Xiamen, 361100, China
h i g h l i g h t s
g r a p h i c a l
a b s t r a c t
• Preparative MDLC system was developed for separation of complicated natural products. • Medium-pressure LC and preparative HPLC were connected by interface of SPE. • Automated multi-step preparative separation of 25 compounds was achieved by using this system.
a r t i c l e
i n f o
Article history: Received 4 January 2014 Received in revised form 12 February 2014 Accepted 16 February 2014 Available online 24 February 2014 Keywords: Two-dimensional preparative liquid chromatography Medium-pressure liquid chromatograph × preparative high-performance liquid chromatography Rheum hotaoense L Anthraquinones Stilbenes Rheumin
a b s t r a c t An on-line comprehensive two-dimensional preparative liquid chromatography system was developed for preparative separation of minor amount components from complicated natural products. Mediumpressure liquid chromatograph (MPLC) was applied as the first dimension and preparative HPLC as the second one, in conjunction with trapping column and makeup pump. The performance of the trapping column was evaluated, in terms of column size, dilution ratio and diameter-height ratio, as well as system pressure from the view of medium pressure liquid chromatograph. Satisfactory trapping efficiency can be achieved using a commercially available 15 mm × 30 mm i.d. ODS pre-column. The instrument operation and the performance of this MPLC × preparative HPLC system were illustrated by gram-scale isolation of crude macro-porous resin enriched water extract of Rheum hotaoense. Automated multi-step preparative separation of 25 compounds, whose structures were identified by MS, 1 H NMR and even by less-sensitive 13 C NMR, could be achieved in a short period of time using this system, exhibiting great advantages in analytical efficiency and sample treatment capacity compared with conventional methods. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Natural products have been a major resource for the investigation of naturally-occurring biologically active substances. The
∗ Corresponding authors. Tel.: +86 592 2189868; fax: +86 592 2189868. E-mail addresses: [email protected] (Y.-K. Qiu), [email protected] (Z. Wu). http://dx.doi.org/10.1016/j.aca.2014.02.023 0003-2670/© 2014 Elsevier B.V. All rights reserved.
traditional way of studying natural products, in most cases, is pretreatment and fractionation of a complex matrix, separation and isolation of the individual components using repeat liquid chromatography, in an off-line mode. The off-line approach is very easy but presents several disadvantages: it is time-consuming, operationally intensive, and difficult to automate and to reproduce. Moreover, sample contamination or formation of artifacts can occur [1]. Techniques developed in recent years for the purpose of
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
analysis and structure elucidation, such as HPLC-SPE-NMR [2] and on-line multidimensional liquid chromatography (MDLC) [3], have shown a significant acceleration in the isolation and identification of natural products. MDLC, usually refers to two-dimensional liquid chromatography (2D-LC), has played an important role in complex sample analysis. 2D-LC combines two independent columns with different separation mechanisms in tandem mode. Depending on whether components can be eluted from the first dimension into the second dimension directly or not, the methods can be classified as on-line [4–6] or off-line [7–9], respectively. Furthermore, depending on whether first dimensional components are eluted into the second dimension completely or not, the methods can be classified as comprehensive [10] or heart cutting [11], respectively. On-line comprehensive 2D-LC is an ideal mode for the analysis of complex samples. Automatic performance in the on-line mode avoids sample contamination, deterioration and personal error, and comprehensive separation enables maximum information to be obtained. Such separation systems have been extensively used to increase the number of compounds separated in complex samples encountered in proteomics, analysis of biopolymers and synthetic polymers, pharmaceutical and environmental analysis and in the analysis of naturally-occurring compounds [12–14]. Nevertheless, several problems occurred when MDLC is used. First, the term “analysis” is added as constraint to the MDLC definition. Preparative separations, with a goal of isolating material, were suggested to be excluded from the definition, because it was considered not providing analytical information at each step [15]. Secondly, substantial amounts of analytes are required in many cases while MDLC system used presently can not meet the demands. For example, 13 C NMR and hyphenated two dimensional (2D) experiments (like heteronuclear multiple bond correlation – HMBC), which are highly important for elucidation of structures of novel compounds, still require at least 100 g of analytes, even if using HPLC-SPE-NMR system or CapNMR probe [16]. The total injected amount of sample on the HPLC column should be in the range of a few decade-milligrams (10–100 mg). Normally an analytical MDLC system cannot handle such amounts of sample. Moreover evaluation of the efficacy and safety of nature products, which are a central issue for the lead compound investigators and pharmaprojects researchers, always demands one to provide adequate amount of individual pure compound in gram-grade.
177
Interface techniques, including trapping columns [17], sample loops [18], parallel columns [4] and vacuum solvent evaporation [19], were developed as the application of on-line comprehensive MDLC. Trapping columns are used in liquid chromatography to trap samples eluted from the first dimension under the first dimensional mobile phase. The samples trapped by trapping columns are eventually eluted by the mobile phases of the second dimension. Trapping columns can delay the elution of samples from the first dimension to the second dimension. Switching the mobile phase on the trapping column can avoid incompatibility between two-dimensional mobile phases. Thus, trapping column interface could be valuable to overcome the above-mentioned problems [20]. We report here an on-line comprehensive medium-pressure liquid chromatography (MPLC) × preparative HPLC system for preparative separation of natural products, where a mediumpressure column was applied as the first dimension and preparative HPLC column as the second one, interfaced by a makeup pump and a 10-port switching valve in conjunction with two reversed-phase trapping columns (Fig. 1). After evaluation in terms of column size, dilution ratio and trapping volume, as well as flow rate and system pressure from the view of medium pressure liquid chromatograph, 15 mm × 30 mm i.d. column packed with ODS was selected as trapping column. In this case, analytes present in the MPLC elution bands are captured online onto the trapping column. The analytes retained on the trapping column are then eluted into the preparative HPLC column process with the corresponding mobile phase. Higher sample processing capability was provided by the 1st dimension MPLC; meanwhile, separation ability was preserved by using a 250 mm × 20 mm i.d. preparative HPLC column as the 2nd dimension. Macro-porous resin enriched water extract of Rheum hotaoense L. was applied as model complex mixture to evaluate the operation and performance of this MPLC × preparative HPLC system. A practical procedure for selection of isolating condition was established, regarding the factors of sample weight, 1st dimensional medium-pressure column size, trapped volume of elution and preparative HPLC column volume. This MPLC × preparative HPLC integrated system was successfully applied in automated multi-step preparative separation of 26 compounds, 25 of which were identified by MS, 1 H NMR and even by less sensitive 13 C NMR, from the macro-porous resin enriched water extract of R. hotaoense, exhibiting great advantages in analytical efficiency
Fig. 1. Scheme for MPLC × preparative HPLC system.
178
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
and sample treatment capacity compared with conventional methods. 2. Experimental 2.1. Chemicals and materials HPLC grade solvents were purchased from Merck, Darmstadt, Germany. The dried roots of Rheum hotaoense L. were purchased from Lu-Yan Pharmaceuticals (Xiamen, China) and identified by Professor Quan-Cheng Chen (Xiamen University, Xiamen, China), by means of routine morphological identification, microscopical identification and physicochemical identification. A voucher specimen (20110501-DH) has been deposited at School of Pharmaceutical Sciences, Xiamen University. 2.2. Instrumentation All instruments used in this study are commercially available. The first dimensional MPLC separation was carried out on a Büchi C601 medium-pressure pump equipped with C-615 pump controller (Büchi Labortechnik AG, Flawil, Switzerland). Another mediumpressure pump was used for the addition of water as makeup fluid through a tee union (VICI, Schenkon, Switzerland) to dilute the elution from 1st dimensional MPLC. The second-dimension LC system using a Varian binary gradient LC system (Varian Inc. Corporate, Santa Clara, USA) containing two solvent deliver modular (PrepStar 218), a photodiode array detector (ProStar 335) and a fraction collector (ProStar 701). Preparative-HPLC control and data acquisition were performed by a Varian Star Workstation 6.41 (Varian Inc. Corporate, Santa Clara, USA). All analytic scale sample analysis was performed on an Agilent Technologies 1100 Series HPLC provided with an automatic injector and a diode array detection system.
bed volume, b.v.: 5.3 mL), 15 mm × 30.0 mm i.d. (Phenomenex Security Guard C18 column, b.v.: 10.6 mL) and 50 mm × 20.0 mm i.d. (Varian Dynamax Microsorb 100-5 C18, b.v.: 15.7 mL), were used for trapping capability evaluation. Five times the bed volume’s aqueous methanol eluent (E30, E50 and E70) was pumped at a flow rate of half bed volume per minute to the tee, in which the eluent was diluted by water into different ratio, and loaded onto the under-evaluated pre-column. Samples trapped on each column were finely eluted with MeOH and weighed for recovery rate (Table 1). 2.6. HPLC analysis of total extract The crude macro-porous resin enriched water extract (ET) were analyzed with a Cosmosil ODS-PAQ column (250 mm × 4.6 mm i.d., 5 m, Nakalai Tesque Co. Ltd., Kyoto Japan). The mobile phase was water (A) and acetonitrile (B) according to the following gradient program (v/v): 0 min 80% A 20% B, 25 min 60% A 40% B, and run back to 80% A 20% B when 30 min. The HPLC analysis performed at flow rate of 1 mL min−1 , and monitored at 254 nm by the Agilent HPLC system (Fig. 2). 2.7. Preparative HPLC isolation In order to compare with the conventional isolation method, the crude macro-porous resin enriched water extract of R. hotaoense (ET) was dissolved in water at concentration of 20 mg mL−1 and subjected to preparative Cosmosil ODS-PAQ column (250 mm × 20.0 mm i.d., 5 m, Cosmosil, Nakalai Tesque Co. Ltd., Kyoto Japan) for isolation, in which 60 mg of extract was loaded at each isolation. Water (A)/acetonitrile (B) were employed as mobile phase, at gradient mode: B from 20% to 40% for 50 min, and run back to 20% when 60 min. After 8 times of reduplicative isolation, 11 compounds were obtained (S0, as shown in Fig. 3).
2.3. Sample preparation In the present study, macro-porous resin enriched water extract from R. hotaoense L. was used as model complex mixture to evaluate the performance of the trapping column and the total system. The root of R. hotaoense (0.5 kg) was extracted with water twice and concentrated under reduced pressure at 40 ◦ C. The concentrate was passed through Diaion HP20 macro-porous resin (Mitsubishi Chem, Tokyo, Japan) column (600 mm × 65 mm i.d.), eluted with H2 O and 95% EtOH 6000 mL respectively. 95% EtOH eluent was collected and concentrated to afford 58.0 g of total extract (ET) for further separation. 10.1 g of the extract was subject to a ODS column (260 mm × 50 mm i.d.), eluted successively with 1500 mL of H2 O, 30% MeOH (E30), 50% MeOH (E50) and 70% MeOH (E70). The aqueous methanol eluent were reserved to evaluate the performance of the trapping column. 2.4. Interface between MPLC and preparative HPLC A medium-pressure pump was used for the addition of water as makeup fluid, which was passing through a tee to dilute the elution from 1st dimensional MPLC. The tandem MPLC and HPLC columns were interfaced by two equivalent trapping columns and an electronically controlled Valco Cheminert 2-position 10-port switching valve (model EDU10UW, VICI, Schenkon, Switzerland) that make up an integrated online column switching MPLC × preparative HPLC system (Fig. 1). 2.5. Trapping column and dilution parameter selection Three commercially available ODS pre-columns in different size, 15 mm × 21.2 mm i.d. (Phenomenex Security Guard C18 column,
2.8. MPLC × HPLC separation procedures The crude macro-porous resin enriched water extract of R. hotaoense (ET) was firstly isolated with high-capacity by MPLC column, after which the eluent was then concentrated online into the solid-phase trapping column and then transferred into HPLC for further high performance separation, through the valve-switching technique. Three 2D MPLC × preparative HPLC separation protocols Table 1 Recovery rate of samples trapped on the trapping columns (X¯ ± RSD, %, n = 3). Sample solutiona
Dilution ratio
Column I
Column II
Column III
E30
1:1 1:2 1:3
55.4 ± 2.5 61.1 ± 3.8 65.7 ± 3.0
53.9 ± 3.9 60.5 ± 4.8 65.7 ± 2.6
58.9 ± 3.2 62.1 ± 5.2 67.7 ± 4.3
E50
1:1 1:2 1:3
66.4 ± 3.5 83.6 ± 2.4 96.8 ± 2.6
63.7 ± 3.3 82.8 ± 1.8 96.0 ± 3.7
67.4 ± 1.9 84.5 ± 3.0 96.9 ± 2.7
E70
1:1 1:2 1:3
91.6 ± 0.8 95.8 ± 2.3 98.4 ± 1.1
87.1 ± 1.4 93.3 ± 0.9 97.8 ± 2.3
91.4 ± 1.7 96.4 ± 2.0 99.2 ± 1.1
a Preparation of sample solution: 10.1 g of the extract was subject to a ODS column (260 mm × 50 mm i.d.), eluted successively with 1500 mL of H2 O, 30% MeOH (E30), 50% MeOH (E50) and 70% MeOH (E70). The aqueous methanol eluent were reserved to evaluate the performance of the trapping column. Three commercially available ODS pre-columns in different size, column I (15 mm × 21.2 mm i.d.), column II (15 mm × 30.0 mm i.d.) and column III (50 mm × 20.0 mm i.d.), were used for trapping capability evaluation. Five times the bed volume’s aqueous methanol eluent (E30, E50 and E70) was pumped at a flow rate of half bed volume per minute to the tee, in which the eluent was diluted by water into different ratio, and loaded onto the under-evaluated pre-column. Samples trapped on each column were finely eluted with MeOH and weighed for recovery rate.
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
179
Fig. 2. HPLC analysis of crude macro-porous resin enriched water extract of R. hotaoense (ET). The analysis was performed on a Cosmosil ODS-PAQ column (250 mm × 4.6 mm i.d., 5 m, Nakalai Tesque Co. Ltd., Kyoto Japan). The mobile phase was water (A) and acetonitrile (B) in a linear gradient mode as follows: B from 20% to 40% for 25 min, and run back to 20% when 30 min. The flow-rate of the mobile phase was 1.0 mL min−1 and the effluents were monitored at 254 nm by a DAD detector. The column temperature was kept at 30 ◦ C.
(S1, S2 and S3) were carried out according to different strategies. All the 1st dimensional separations were performed at a similarity procedure, while the mobile phase composition and elution program used in the 2nd dimensional preparative HPLC separation were different. Firstly, each extract (0.5 g in S1, S2 separation, and 2.0 g in case of S3) dissolved in water (5 mL or 10 mL), was subjected to the first dimensional MPLC column, a medium-pressure column (300 mm × 22.0 mm i.d.) laboratory-packed with Cosmosil ODS gel (75 m, Nakalai Tesque Co. Ltd., Kyoto Japan). Then the 1st dimensional column was eluted by a stepwise elution at a flow rate of 1 mL min−1 . In the separation case of S1 and S2, fractions were eluted with a gradient of water (A) and methanol (B) according to the following gradient program (v/v): 0 min 80% A 20% B, 12 h 20% A 80% B. As for S3, the gradient elution time was extended to 24 h, and then an additional 4 h stepwise elution from 80% B to 100% B was added. Eluent from the 1st dimensional column was diluted 1:2 by water pumped via makeup pump (2 mL min−1 ) and trapped on one of the trapping column. The 2-position 10-port valve was switched every 60 min by an electronic timer, and the sample trapped on the trapping column was then eluted to the preparative Cosmosil
ODS-PAQ column for further second dimension separations, in which different mobile phases were applied. In the separation case of S1 the mobile phase was water (A) and methanol (B), gradient run from 0 to 50 min, 30% methanol increasing to 70% methanol, and run back to 30% methanol when 60 min. As for the separation cases of S2 and S3, water (A)/acetonitrile (B) were employed as mobile phase, at gradient mode: B from 20% to 40% for 50 min, and run back to 20% when 60 min. Accordingly, another gradient elution program was applied during the isolation of the last four fractions of S3, and it was: B from 40% to 100% for 45 min, and run back to 40% when 60 min. Eluent with pure chromatographic peak in the 2nd dimensional chromatography was collected by fraction collector into different fractions. Preparative HPLC chromatograms were recorded and cascaded by Varian Star Workstation v6.41 as shown in Figs. 4 and 5. 2.9. HPLC analysis of the purified fragments Most purified fragments isolated by the MPLC × HPLC system were also analyzed as the crude extract did. Some
180
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
Fig. 3. Conventional preparative HPLC isolation of the crude macro-porous resin enriched water extract of R. hotaoense (ET). The extract was dissolved in water at concentration of 20 mg mL−1 and subjected to preparative Cosmosil ODS-PAQ column (250 mm × 20.0 mm i.d., 5 m, Cosmosil, Nakalai Tesque Co. Ltd., Kyoto Japan), in which 60 mg of extract was loaded at each isolation. Water (A)/acetonitrile (B) were employed as mobile phase, at gradient mode: B from 20% to 40% for 50 min, and run back to 20% when 60 min. Eight reduplicative isolation procedures were performed.
purified fragments with lower polarity were assayed by water (A)/acetonitrile (B) in another linear gradient as: B from 40% to 100% for 25 min, and run back to 40% when 30 min. Those fractions with identical retention time were combined with each other and analyzed again by HPLC before further structural evaluation (Fig. 6). 2.10. MS and NMR experiments Identification of purified compounds was performed by MS and NMR. Samples were dissolved with 0.5 mL of DMSO-d6 in conventional 5 mm NMR tubes, and their 1 H and 13 C NMR spectra were taken on a Brucker Avance 400 FT NMR spectrometer with tetramethylsilane as an internal standard. ESI-MS were measured on an AB MD-SCIEX Advantage spectrometer. 3. Results and discussion 3.1. Evaluation of solid-phase enrichment and water-dilution interface between MPLC and HPLC Since trapping columns are used to collect and concentrate compounds eluted from the first dimension, and then elute them into the second dimension, they should possess two key characteristics. First, samples should be better retained on trapping columns than the corresponding first dimensional columns. Second, the samples trapped on trapping column should be eluted easily by the mobile phases of the second dimension [20]. Although the solid-phase extraction (SPE) has been successfully used for online analyte enrichment in both analytical HPLC analysis [21–25] and preparative HPLC fractions post-collection concentration [26], there must be some uncertainties in the solid-phase trapping step as a novel interface between MPLC and preparative HPLC. On the other hand, unlike common analytical MDLC, both trapping
columns and makeup pump are key components in this system, for the dilution of eluent with water overcomes the influence of solvent effect and make it possible to transfer the primary column eluent onto the secondary column. Therefore, in the present study, different portions of the macro-porous resin enriched water extract from R. hotaoense, representing a range of mixtures with different polarities and chemical properties, were used as model complex mixture to evaluate the performance of the trapping column under the above-mentioned experimental conditions. 3.1.1. Sorbent type and column size selection Since the 2nd dimensional HPLC is the main constraint on separation capacity and efficiency, the trapping column should use the same sorbent type with that of the 2nd column. And in order to preserve separation efficiency, the sorbent bed of pre-column is usually 1/25–1/5 of the corresponding preparative column during conventional HPLC separation. Furthermore, owing to the demand of being connected to the high pressure HPLC flow path, the trapping column should be pressure tolerable, which makes commercially available ODS pre-columns of preparative HPLC a good choice for this 2D MPLC × HPLC system. Then, three pre-column in different size and different diameterheight ratio, 15 mm × 21.2 mm i.d. (b.v.: 5.3 mL), 15 mm × 30.0 mm i.d. (b.v.: 10.6 mL), and 50 mm × 20.0 mm i.d. (b.v.: 15.7 mL), were chosen for evaluating. 3.1.2. Sample retention ability and dilution ratio In order to capture the sample eluted from the 1st dimensional column, another makeup pump is necessary in order to dilute the eluent with water. Results shown in Table 1 indicate that each enrichment column could achieve adequate trapping efficiency (>90% of recovery rate) for sample solution of E50 and E70. And the greater dilution ratio adopted, the lesser sample spillage undergone. As far as the sample with less retention ability (E30) was
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
181
Fig. 4. Two MPLC × HPLC separation protocols, using different mobile phase in the 2nd dimensional chromatography. MPLC condition: Each crude extract (ET, 0.5 g) dissolved in 5 mL water, was subjected to the first dimensional MPLC column, a medium-pressure column (300 mm × 22.0 mm i.d.) laboratory-packed with Cosmosil ODS gel (75 m, Nakalai Tesque Co. Ltd., Kyoto Japan), and eluted by a stepwise elution of water (A) and methanol (B) according to the following gradient program (v/v): 0 min 80% A 20% B, 12 h 20% A 80% B, at a flow rate of 1 mL min−1 . Solid-phase trapping interface: Eluent from the 1st dimensional column was diluted 1:2 by water via makeup pump (2 mL min−1 ) and trapped on one of the trapping column. The 2-position 10-port valve was switched every 60 min by an electronic timer. HPLC condition: The sample trapped on the trapping column was then eluted to the preparative Cosmosil ODS-PAQ column for further second dimension separations, in which different mobile phases were applied. The flow-rate of the mobile phase was 8.0 mL min−1 and the effluents were monitored at 254 nm by a DAD detector. (S1): The mobile phase was water and methanol, gradient run from 0 to 50 min, 30% methanol increasing to 70% methanol, and run back to 30% methanol when 60 min. (S2): Water and acetonitrile were employed as mobile phase, at gradient mode: acetonitrile from 20% to 40% for 50 min, and run back to 20% when 60 min. *Compounds with purity lower than 80% were not numbered in the chromatography.
concerned, much greater sample loss happened when it passed through the enrichment column. This is due to the existence of strong polar components, which can be eluted out of the ODS column by mobile phase comprising lower concentration of organic solvent. The solvent strength of lower organic content solution cannot be reduced efficiently by dilution of water, so that the increase of the dilution ratio contributed little to sample trapping efficiency in E30. Fortunately, the recovery rate of more than 60% (at dilution ratio of 1:2 and 1:3) for E30 is still acceptable in preparative separation. According to the experimental data, the dilution radio was advised to be 1:2. 3.1.3. Diameter-height ratio, trapping efficiency and system pressure As far as the diameter-height ratio was concerned, the data in Table 1 suggested that the column’s trapping efficiency was not mainly dependent on column diameter and height. However, MPLC is a medium-pressure preparative technique. In this study, the system pressure should be taken into account due to
the use of the solid-phase trapping interface, which may cause a rise in column pressure owing to the addition of water at the post-column stage. For preparative purposes, the larger diameter column employed in the solid-phase trapping interface can produce lower back-pressure even with higher postcolumn dilution ratios, which will provide much better system tolerance. For this reason, the 15 mm × 30.0 mm i.d. pre-column was selected for further study. 3.1.4. Desorption Apart from the trapping, the analyte desorption is also an important step affecting the outcome of subsequent HPLC experiments. In the present work, a same sorbent type was used both in the trapping column and the 2nd dimensional HPLC column. Therefore, the trapped analytes were washed out of the solid-phase trapping column in the forward-flush mode (Fig. 1). The analytes desorption in this preparative MDLC system is quite different from that in analytical MDLC. Firstly, solvents with higher strength, that can push the entire slug off the trapping column, is not a
182
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
Fig. 5. MPLC × HPLC separation (S3), in which 2.0 g of R. hotaoense extract (ET) was loaded. MPLC condition: R. hotaoense extract (ET, 2.0 g), dissolved in 10 mL water, was subjected to the first dimensional MPLC column, and eluted by a stepwise elution of water (A)/methanol (B) according to the following gradient program (v/v): 0 min 80% A 20% B, 24 h 20% A 80% B, 28 h 0% A 100% B, at a flow rate of 1 mL min−1 . Solid-phase trapping interface: Eluent from the 1st dimensional column was diluted 1:2 by water via makeup pump (2 mL min−1 ) and trapped on one of the trapping column. The 2-position 10-port valve was switched every 60 min by an electronic timer. HPLC condition: The preparative Cosmosil ODS-PAQ column was used, and water (A)/acetonitrile (B) were employed as mobile phase, at gradient mode: B from 20% to 40% for 50 min, and run back to 20% when 60 min (Fr. 1–Fr. 23); B from 40% to 100% for 45 min, and run back to 40% when 60 min for Fr. 24–Fr. 27.
good choice for desorption. For their solvent effect has a negative impact on the isolation of the 2nd dimensional HPLC. The traps are more suitable to be washed out of the trapping column, by the gradient mobile phase of the 2nd dimensional HPLC. Secondly, unlike back-flush mode commonly used in analytical MDLC, the forward-flush desorption manner is preferred in this system, for it can utilize the trapping column’s separation ability and extend the separation length of the ODS column. 3.2. HPLC analysis of the total extract The crude macro-porous resin enriched water extract of R. hotaoense (ET) was used as model complex mixture to evaluate the performance of the integrated MPLC × HPLC system. According to analysis HPLC chromatogram shown in Fig. 2, (E)-rhaponiticin (12) is the major component present in the crude extract, and (E)-piceatannol 3 -O-ˇ-d-glucopyranoside (8), chrysophanol 8-Oˇ-d-glucopyranoside (20) and rhapontigenin (22) also have a relative high content. Other compositions are in low concentration and poorly separated. 3.3. Conventional preparative HPLC isolation Conventional preparative HPLC isolation of the total extract (ET), lead to the isolation of 11 compounds, whose purity ranged from 78% to 97.5% (Table 2). This result was adequate, for this crude
extract sample is soluble in water, resulting in high efficiency resolution. But in most preparation cases, samples have to be dissolved in organic solvents, whose solvent effect have a negative impact on resolution, resulting in the failure of preparative isolation. 3.4. Integrated reverse phase MPLC × HPLC separation of R. hotaoense macro-porous resin enriched water extract The above studies show that it was feasible to use a 15 mm × 30.0 mm i.d. pre-column packed with ODS materials as trapping columns in the column-switching interface. Then the proposed integrated MPLC × HPLC protocol was applied to the separation of the crude extract of R. hotaoense (ET). 3.4.1. HPLC elution program, governing factor of the MPLC flow rate and interface switching time A gradient elution program up to 60 min was employed to the separation of the MPLC eluent fractions. Since that the trapping column possesses adequate trapping efficiency when the eluent from the 1st dimensional chromatography passing through the column was less than five time of its volume (63.5 mL), the flow rate of the MPLC was adjusted to 1 mL min−1 . Accordingly, the flow rate of the makeup pump should be 2 mL min−1 , at a dilution radio of 1:2. Meanwhile, the 2-position, 10-port valve switching time should be equal to the 2nd dimensional HPLC isolation time.
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
183
Fig. 6. HPLC analysis of isolated compounds. Most of the compounds were obtain from separation of S3, expect for compounds 7 and 10, isolated during S2 separation. The analysis was performed on a Cosmosil ODS-PAQ column (250 mm × 4.6 mm i.d., 5 m, Nakalai Tesque Co. Ltd., Kyoto Japan). The mobile phase was water (A) and acetonitrile (B) in a linear gradient mode as follows: B from 20% to 40% for 25 min, and run back to 20% when 30 min for compounds 1–25, B from 40% to 100% for 25 min, and run back to 40% when 30 min for compounds 26–28. The flow-rate of the mobile phase was 1.0 mL min−1 and the effluents were monitored at 254 nm by a DAD detector. The column temperature was kept at 30 ◦ C.
3.4.2. Sample weight In many reversed-phase medium-pressure liquid chromatograph separation cases, the extract sample is usually partitioned into about 10–30 fractions for further separation. Considering that the 20.0 mm i.d. preparative column can handle up to 50–100 mg of sample with relative high separation efficiency, 0.5–2.0 g of the macro-porous resin enriched water extract of R. hotaoense was subject to the first dimensional MPLC of the 2D system. The MPLC separation was carried out over a laboratory-packed ODS gel column (300 mm × 22.0 mm i.d.), after which the sample was divided into 12–28 fractions by the column-switching interface. Except for the first fraction, which was considered to be dead volume mobile phase and vented, other fractions were subject to the 2nd dimensional HPLC separation. It could be seen from Figs. 4 and 5 and Table 2 that, the greater amount of sample loaded onto the MPLC × HPLC system, the greater sample information can be achieved. For example, the structures of compounds 2, 3, 5, 11, 16, 23 and 24 (Table 3) were clarified by NMR after separation of S3 (sample weight 2.0 g), but their NMR signals were so weak that they could not be identified, although also isolated in S2 (sample weight 0.5 g). Even minor components could be obtained when higher amount of sample was applied in the separation. Compounds 26, 27 and 28, whose solubility in water and concentration in the extract is very low, were also isolated in the S3 separation. However, when the amount of loaded sample was increased, separation efficiency was not decreased substantially, because the sample was divided into more fractions by the 1st dimensional MPLC column before being subjected to HPLC. Only chromatography peaks of minor components that were poorly resolved from a major peak, such as compounds 7 and 10, could not be successfully isolated in high purity.
3.4.3. Mobile phase system used in 1st dimensional MPLC and 2nd dimensional HPLC MPLC × HPLC separation chromatogram (S1) using methanol/ water mobile phase system at both 1st dimensional MPLC and 2nd dimensional HPLC separation, as well as the chromatogram using different mobile phase system (S2 and S3) are shown in Figs. 4 and 5. It is clearly seem that, combination of different mobile systems in the two dimensional chromatography results in a considerable improvement in peak resolution. Only 8 compounds were identified when using a same mobile phase system in both dimensional separations, this result is worse than that of conventional preparative HPLC separation, which resulted in 13 compounds’ isolation. When using a different mobile system, components that were very difficult to separate in conventional preparative HPLC, such as compounds 2, 3, 5, 13, 15, 16, 17, 21, 23 and 24, were all well resolved. Thus, the combination of mobile systems is preferred in this reverse phase MPLC × HPLC system.
3.4.4. Compounds isolated After 12–28 h separation, 22–53 fractions were directly produced according to the MPLC × HPLC elution profile from 0.5 to 2.0 g of crude macro-porous resin enriched water extract of R. hotaoense using the proposed column-switching MPLC × HPLC protocol. The purified fractions isolated were then analyzed by HPLC and combined according to their retention time. As a result, 9–25 identifiable compounds were obtained at a purity of more than 80% in an automated multi-step separation, from 1.1 to 206.3 mg (Fig. 6 and Table 2). For both yields and purity, the MPLC × HPLC protocol is much better when compared with that of conventional preparative HPLC isolation.
184
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
Table 2 Purity and weight of isolated compounds by different separation strategies.a Compd.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
S0
S1
S2
S3
Purity (%)
Weight (mg)
Purity (%)
Weight (mg)
Purity (%)
Weight (mg)
Purity (%)
Weight (mg)
87.1 – – 78.0 – – 81.5 95.0 – 83.3 – 95.7 – 85.9 – – – 91.1 – 82.4 – 97.5 – – 91.3 – – –
2.5 – – 2.4 – – 1.9 12.8 – 6.1 – 50.2 – 3.5 – – – 2.0 – 2.3 – 7.6 – – 1.5 – – –
85.2 – – 67.3 – – – 88.2 – – – 92.4 – – 86.0 – – 82.3 84.6 96.5 – 75.8 – – – – – –
2.1 – – 2.2 – – – 15.6 – – – 48.2 – – 3.6 – – 1.8 8.0 1.4 – 8.0 – – – – – –
87.4 – – 84.5 – – 92.5 94.6 – 93.3 – 95.2 – 87.4 83.7 – 91.6 90.4 88.8 84.6 – 95.4 – – 92.6 – – –
2.7 – – 2.5 – – 1.7 31.1 – 6.7 – 111.5 – 3.6 5.5 – 3.3 4.2 14.7 12.5 – 7.7 – – 1.9 – – –
86.3 86.2 80.2 80.6 98.7 98.1 – 98.2 100.0 – 90.1 97.4 86.6 85.5 94.3 97.3 86.5 99.3 95.5 100.0 93.7 99.2 90.2 90.9 80.2 92.8 96.7 85.1
5.0 4.5 2.0 5.3 1.1 1.2 – 41.7 4.5 – 10.2 206.3 12.3 23.6 15.6 4.8 15.4 7.5 30.3 4.3 3.4 33.4 1.3 3.0 3.4 1.4 1.4 1.2
Those compounds with purity lower than 80% or weighed less than 2.0 mg were not listed here. a The relative contents in percentage were calculated with area normalization method.
Table 3 Identification of isolated compounds. Structure/name
Catechin Rheumin Vanillic acid 1,3-Dihydroxy-6,7-dimethy-aetone-1-O-ˇ-d-glucopyranoside a
Configuration of the alkene double bond.
Compd.
MW
E/Za
R1
R2
R3
1 4 8 10 11 12 13 14 15 17 19 21 22 23
568 406 406 420 390 420 420 572 244 572 404 228 258 242
E E E E E E Z E E E E E E Z
OH OH O-Glc O-Glc H OH OH OH OH OH H H OH H
Glc H H CH3 H CH3 CH3 CH3 H CH3 CH3 H CH3 CH3
Glc Glc H H Glc Glc Glc 6 -O-glloyl-glucopyranosyl H 2 -O-glloyl-glucopyranosyl Glc H H H
7 16 18 20 24 25 26 27 28
432 432 416 416 446 568 270 254 284
CH2 OH OH H CH3 OCH3 CH3 OH CH3 OCH3
H CH3 CH3 H CH3 H CH3 H OH
Glc Glc Glc Glc Glc 6 -O-glloyl-glucopyranosyl H H H
2 3 5 9
290 292 168 418
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
Scheme 1. Keto–enol tautomerism of compound 3.
3.5. Structure evaluation The chemical structures of isolated components were further identified by ESI-MS and 1 H NMR, even by the less-sensitive 13 C NMR (supporting information). The structure of some components with high purity, such as compound 6, were unable to be determined, owing to their minor content and only with 1 H NMR, MS data available. Anyhow, by comparing spectra data with those reported in literature, structures of 27 compounds were identified as shown in Table 3, and they were: (E)-piceatannol 3, 4 -O-ˇ-d-diglucopyranoside (1) [27], catechin (2) [28], rheumin (3) [29], (E)-piceatannol 3-O-ˇ-d-glucopyranoside (4) [30], vanillic acid (5) [31], aloe-emodin 8-O-ˇ-d-glucopyranoside (7) [32], (E)-piceatannol 3 -O-ˇ-d-glucopyranoside (8) [31], 1,3dihydroxy-6,7-dimethy-aetone 1-O-ˇ-d-glucopyranoside (9) [33], (E)-rhapontigenin 3 -O-ˇ-d-glucopyranoside (10) [31], (E)resveratrol 3-O-ˇ-d-glucopyranoside (11) [34], (E)-rhaponiticin (12) [35], (Z)-3,3 ,5-trihydroxy-4 -methoxystilbene 3-O-ˇ-dglucopyranoside (13) [31], (E)-rhaponticin 6 -O-gallate (14) [31], (E)-piceatannol (15) [36], emodin 1-O-ˇ-d-glucopyranoside (16) [27], (E)-rhaponticin 1 -O-gallate (17) [31], chrysophanol 1-Oˇ-d-glucopyranoside (18) [32], (E)-desoxyrhaponticin (19) [35], chrysophanol 8-O-ˇ-d-glucopyranoside (20) [32], (E)-resveratrol (21) [37], (E)-rhapontigenin (22) [35], (Z)-deoxyrhapontigenin (23) [38], physcion 8-O-ˇ-d-glucopyranoside (24) [39], chrysophanol 8-O-ˇ-d-(6 -galloyl)-glucopyranoside (25) [40], emodin (26) [39], chrysophanol (27) [39], and physcion (28) [39]. 3.6. Keto–enol tautomerism of compound 3 By comparing the spectra in separation of Fr. 2 and Fr. 3 in S3, we found that compound 3 presented a longer retention time than that of compound 2. But when the eluate of compound 3 was evaporated and redissolved, its retention time changed, becoming even shorter than that of compound 1. A structural transformation must have occurred in compound 3, and keto–enol tautomerism (Scheme 1) might explain the change of retention time. Keto-form of compound 3, which presents less-polar and higher-retention, was captured by the MPLC × HPLC system even it was in a lower concentration. Enol-form of compound 3, although being much stable than keto-form, was ignored during the separation procedure, for its retention time was so close to that of solvent peak. But when compound 3 was disposed for further HPLC and NMR test, only the stable enol-form, could be observed clearly.
185
demonstrated that the quantity of compound information obtained is higher compared to single-column MPLC or HPLC chromatography. Although the MPLC column, trapping column, and HPLC column all were the same stationary phase (ODS), adequate separation resolution was achieved by using different mobile phase systems in the 1st and 2nd dimensional columns, so that some pairs of peaks that are poorly resolved in one separation mode were easily separated using the MPLC × HPLC approach. If different stationary phases are employed in two different dimensions, higher resolution may be obtained. For example, choosing perfluorophenyl column for the second dimension, or a Sephadex LH 20 column as the first dimension, may afford separations that are not obtained on the ODS dimension. In addition, the isolated compounds were identified by NMR spectra. If more sensitive NMR equipment such as HPLCSPE-NMR system or CapNMR probe were applied, lower abundance compounds could also be identified. On the other hand, reversed-phase preparative HPLC isolation for lipophilic samples is always a difficult problem. Sample dissolution is the major issue particularly for crude extracts and fractions. Dissolution of the sample in the mobile phase is often not possible. To circumvent the solubility problem, samples are dissolved in the mobile phase component of high elution strength. However, this may considerably affect the chromatographic resolution. The integrated MPLC × HPLC system solved the problem. Firstly, solid phase sample introduction, where samples are adsorbed onto stationary phase and loaded to column for elution, can be easily applied in the 1st dimensional MPLC column used in this system [41]. Secondly, the solid-phase interface between the MPLC and HPLC, which overcomes solvent effect by diluting the MPLC eluent with water, enables the hydrophobic compounds’ isolation by preparative HPLC with higher resolution. In conclusion, the proposed protocol is our initial exploration into the preparative and on-line comprehensive MPLC × HPLC technique. The integrated MPLC × HPLC system provided a considerable improvement in peak capacity when compared with HPLC chromatography, giving very efficient separation of the crude extract of R. hotaoense, resulting in 25 compounds’ identification in an automated multi-step preparative separation. It can also be applied to isolate other comprehensive multi-component natural products. The system also captured the keto-enol tautomerism phenomenon of rheumin (3) for the first time, which is a valuable application in exploration of structural transformation in unstable components. Acknowledgments The project was supported by the National Natural Science Foundation of China (No. 81102333 and No. 81072549) and Fundamental Research Funds for the Central Universities (No. 2010121108). The authors would also like to acknowledge the financial supports from Fujian Natural Science Foundation for Distinguished Young Scholars (No. 2012J06020). We acknowledge Prof. D.M.J. Lilley of University of Dundee, for his kindly help on manuscript revise.
4. Conclusions
Appendix A. Supplementary data
In summary, we have shown that the online column-switching MPLC × HPLC coupling is possible in terms of peak concentration prior to further HPLC separation. In conjunction with a solid-phase trapping interface, the system provides simpler and more efficient separation without requiring extra steps to remove organic solvent when compared with conventional off-line MPLC–HPLC separation. It achieved automated high-capacity and high-resolution isolation using only commercially-available equipment. The application of this system to the preparative isolation of a real sample has
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.aca.2013.12.001. References [1] P. Dugo, O. Favoino, R. Luppino, G. Dugo, L. Mondello, Analytical Chemistry 76 (2004) 2525. [2] M. Sandvoss, B. Bardsley, T.L. Beck, Magnetic Resonance in Chemistry 43 (2005) 762. [3] D.R. Stoll, X.P. Li, X.L. Wang, Journal of Chromatography A 1168 (2007) 3.
186
Y.-K. Qiu et al. / Analytica Chimica Acta 820 (2014) 176–186
[4] K. Wagner, T. Miliotis, G. Marko-Varga, R. Bischoff, K.K. Unger, Analytical Chemistry 74 (2002) 809. [5] G.J. Opiteck, K.C. Lewis, J.W. Jorgenson, R.J. Anderegg, Analytical Chemistry 69 (1997) 1518. [6] P. Dugo, O. Favoino, R. Luppino, G. Dugo, L. Mondello, Analytical Chemistry 76 (2004) 2525. [7] P. Dugo, O. Favoino, P.Q. Tranchida, G. Dugo, L. Mondello, Journal of Chromatography A 1041 (2004) 135. [8] Y. Liu, X. Xue, Z. Guo, Q. Xu, F. Zhang, X. Liang, Journal of Chromatography A 1208 (2008) 133. [9] X. Chen, L. Kong, X. Su, C. Pan, M. Ye, H. Zou, Journal of Chromatography A 1089 (2005) 87. [10] M. Eggink, W. Romero, R.J. Vreuls, H. Lingeman, W.M.A. Niessen, H. Irth, Journal of Chromatography A 1188 (2008) 216. [11] L.M. de Souza, T.R. Cipriani, C.F. Sant’Ana, M. Iacomini, P.A.J. Gorin, G.L. Sassaki, Journal of Chromatography A 1216 (2009) 99. [12] S.P. Dixon, I.D. Pitfield, D. Perrett, Biomedical Chromatography 20 (2006) 508. [13] Q. Wu, H.M. Yuan, L.H. Zhang, Y.K. Zhang, Analytica Chimica Acta 731 (2012) 1. [14] X. Zhang, A.Q. Fang, C.P. Riley, M. Wang, F.E. Regnier, C. Buck, Analytica Chimica Acta 664 (2010) 101. [15] L. Blumberga, M.S. Klee, Journal of Chromatography A 1217 (2010) 99. [16] J.F. Hu, G.R. Eldridge, Y.H. Yu, Progress in Chemistry 20 (2008) 429. [17] K. Im, H.-W. Park, Y. Kim, B. Chung, M. Ree, T. Chang, Analytical Chemistry 79 (2007) 1067. [18] L. Hu, X. Chen, L. Kong, X. Su, M. Ye, H. Zou, Journal of Chromatography A 1092 (2005) 191. [19] H. Tian, J. Xu, Y. Xu, Y. Guan, Journal of Chromatography A 1137 (2006) 42. [20] L.W. Cao, D.H. Yua, X.L. Wang, Y.X. Ke, Y. Jin, X.M. Liang, Analytica Chimica Acta 706 (2011) 184. [21] J.S. Fritz, M. Macka, Journal of Chromatography A 902 (2000) 137.
[22] A. Cappiello, A. Berloni, G. Famiglini, F. Mangani, P. Palma, Analytical Chemistry 73 (2001) 298. [23] C. Seger, M. Godejohann, L.H. Tseng, M. Spraul, A. Girtler, S. Sturm, H. Stuppner, Analytical Chemistry 77 (2005) 878. [24] K. Kato, M.J. Silva, L.L. Needham, A. Calafat, Analytical Chemistry 78 (2006) 6651. [25] C. Clarkson, M. Sibum, R. Mensen, J.W. Jaroszewski, Journal of Chromatography A 1165 (2007) 1. [26] J. Stevens, M. Crawford, G. Robinson, L. Roenneburg, Journal of Chromatography A 1142 (2007) 81. [27] K.K. Sung, Archives of Pharmacal Research 23 (2000) 159. [28] G. Gelles, F. Abdellatif, M. Nadine, K. Manuela, T. Raffaele, P. Roger, P. Vincent, Phytochemistry 52 (1999) 759. [29] J.L. Li, A.Q. Wang, J.S. Li, W.Y. He, M. Kong, Chinese Traditional and Herbal Drugs 29 (1998) 721. [30] Y. Kashiwada, G.I. Nonaka, I. Nishioka, Chemical and Pharmaceutical Bulletin 32 (1984) 3501. [31] Y.K. Qiu, Y.J. Chen, Y.P. Pei, H. Matsuda, M. Yoshikawa, Chemical and Pharmaceutical Bulletin 50 (2002) 1507. [32] J.L. Li, A.Q. Wang, Chinese Traditional and Herbal Drugs 31 (2000) 321. [33] L. Zhou, M. Lin, J.B. Li, Acta Pharmaceutica Sinica 29 (1994) 107. [34] P.W. Teguo, A. Decendit, J. Vercauteren, G. Deffieux, J.M. Mérillon, Phytochemistry 42 (1996) 1591. [35] A.Q. Wang, J.L. Li, Z.Z. Wu, Chinese Traditional and Herbal Drugs 32 (2001) 878. [36] X.X. Li, R.A. Yan, H.Y. Duan, Fine Chemicals 28 (2011) 475. [37] F. Mattvi, F. Reniero, S. Korhammer, Journal of Agricultural and Food Chemistry 43 (1995) 1820. [38] W. Zhang, M.L. Go, European Journal of Medicinal Chemistry 42 (2007) 841. [39] A.Q. Wang, J.L. Li, J.S. Li, Chinese Traditional and Herbal Drugs 41 (2010) 343. [40] H. Matsuda, T. Morikawa, I. Toguchida, J.Y. Park, S. Harima, M. Yoshikawa, Bioorganic and Medicinal Chemistry Letters 9 (2001) 41. [41] P. Webera, M. Hamburgera, N. Schafrothb, O. Potterat, Fitoterapia 82 (2011) 155.
