Comparison of Antioxidant Capability after Isopropanol Salting-Out Pretreatment and n-Butanol Partition Extraction, and Identification and Evaluation of Antioxidants of Sedum formosanum N.E.Br.

Crude extracts of Sedum formosanum N.E.Br. obtained from n-butanol partition (BP) and isopropanol salting-out pretreatment (ISP) were analyzed using antioxidation assays. The results indicated that the extract from ISP contained more potent antioxidants and thus exhibited more antioxidant activity in all the assays. The superoxide radical-scavenging activity and inhibition of nitric oxide radicals achieved after ISP were 3.65 and 2.18 times higher than those achieved through BP, respectively. Eight bioactive natural products were isolated and identified according to an analysis of antioxidation activity in different fractions of the ISP crude extract, namely three cyanophoric glycosides 1–3, three flavonoids 4–6 and two phenolic compounds (7 and a new compound 8). Among them, compounds 5 and 6 exhibit the highest antioxidation capability, and the ISP is suitable for obtaining compounds 5 and 6 using HPLC chromatograms. Therefore, ISP is an excellent extraction technology that can be used to extract antioxidant compounds in the nutraceutical and pharmaceutical industries.


Introduction
In the human body, the normal oxidative metabolism constantly produces reactive oxygen species (ROSs), such as hydrogen peroxide, superoxide (O 2 .´) , the hydrogen radical ( . OH), singlet oxygen, Table 1 displays the antioxidant capability results after use of the ISP and BP extraction technology, which were obtained using assays on superoxide radical-scavenging activity, oxygen radical absorbance capacity (ORAC) radical-scavenging activity, chelation of ferrous ions, and inhibition of NO radical activity. The antioxidant activity of the crude extract is expressed in micromoles of Trolox equivalent (TE) per gram of dried materials. After statistical calculation, the t-test values are 2.442, 43.268, 15.08 and 22.624 for FRAP, SOD, NO and ORAC, respectively. Therefore, the antioxidant capability comparison of the ISP and BP extractions possesses statistical significance. The results indicated that the antioxidant capability of the crude extract from the ISP extraction technology was between 3.65 and 1.38 times higher than that of the crude extract obtained by the BP extraction technology. The superoxide radical-scavenging activity and inhibition of the NO radical obtained using ISP extraction technology were respectively 3.65 and 2.18 times higher than those obtained using BP extraction technology (0.21 vs. 0.057 micromoles of TE/g; 0.118 vs. 0.054 micromoles of TE/g, respectively). These data suggest that ISP extraction technology is preferable to BP extraction technology for extracting antioxidants. Consequently, isolating and evaluating the antioxidant capability of the pure compounds of S. formosanum N.E.Br. by using ISP extraction technology is significant.

Antioxidant Activity Potentials of Identified Compounds
Compounds 1-3 are cyanogenic glycosides, Compounds 4-6 are flavonoid glycosides, and compounds 7 and 8 are other compounds. Compounds 4 and 5 are kaemferol glycosides, whereas compound 6 is a flavone glycoside. The radical-scavenging activity of these compounds was ranked as 7 > 6 > 5 > 4 > 2 > 8 > 1 > 3 in the superoxide radical-scavenging activity assay; 6 > 5 > 4 > 8 in the ORAC radical-scavenging activity assay; 6 > 5 > 4 > 7 > 8 > 2 > 3 > 1 in the assay on the chelation of ferrous ions; and 5 > 6 > 4 > 8 > 7 > 1 > 2 « 3 in the NO activity inhibition assay. Among these compounds, compounds 6 and 5 exhibited the highest antioxidant capability, as determined using the afore-mentioned methods (superoxide radical-scavenging activity: 1.03 vs. 0.95 mmol of TE; ORAC radical-scavenging activity: 1.06 vs. 1.03 mmol of TE; chelation of ferrous ions: 1.23 vs. 1.14 mmol of TE; and inhibition of NO activity: 0.83 vs. 1.01 mmole of TE, respectively). As mentioned previously, the structural characteristics of the analytes are mainly crucial to determining antioxidant capability. Compounds 4-6 contain similar aglycon structures (the C-ring 2,3-double bond does not link the OH in position 3 of compound 6), but at different positions and exhibit the conjugation of one or two sugar moieties. Thus, it is presumed that the solubility of these compounds [19], steric effects, and the degree of facilitation of the delocalization of electrons from the B-ring to the C-ring may determine their antioxidant capability [20]. Compounds 4 and 5 have identical structures (one sugar linked at the C3 position); however, compound 5 has one more sugar link at the A7 position. Compound 5 has improved solubility in an aqueous solution, facilitating the delocalization of the electrons from the B-ring to the C-ring. Consequently, compound 5 has higher antioxidant capability than compound 4 does. Although compounds 5 and 6 have equal numbers of sugars (linked at different positions of the aglycon structure), in compound 5, the one sugar link at the C3 position of the C-ring at the 2,3-double bond causes steric hindrance and may interrupt the delocalization of the electrons from the B-ring to the C-ring. Hence, compound 6 was more efficient than compound 5 in the antioxidant capability assays, except for NO inhibiting activity. Compound 5 is traditionally used for treating kidney diseases among Mexican natives and has been isolated and identified from certain ferns [21,22]. The linking of cyanide at different positions of the double bond may cause differences in antioxidant capability among compounds 1-3. Cyanogenic glycosides has little toxicity and are crucial compounds for pharmaceutical use. Compound 1 is the main natural medicine constituent of R. quadrifida (Pall.) Fisch. et Mey which inhibits histamine release with anti-2,4-dinitrophenyl IgE from sensitized rat peritoneal exudate cells, and exhibits antiallergic activity in rats, according to a passive cutaneous anaphylaxis test [23]. Compound 3 is the main active constituent of S. sarmentosum and is used to treat chronic viral hepatitis in Asia [15].
Compound 8 is a p-coumaric acid derivative, and compound 7 is a p-hydroxybenzyl alcohol derivative. The antioxidant capability of compound 7 is higher than that of compound 8, except for inhibition of NO activity. Compound 7 is the main active constituent of Rhizoma gastrodiae, and is considered a traditional Chinese medicine proven to be an effective and safe drug for clinical use to prevent neurocognitive decline following cardiopulmonary bypass, and benefitting older refractory hypertension patients [24]. This compound can improve the association between endothelin and NO in plasma [25].

Comparison of the Chromatograms after Isopropanol Salting-Out Pretreatment and n-Butanol Partition Extraction Technology Using HPLC Separation
Figure 3a reveals that the retention times (t R , min) of compounds 5 and 6, which exhibited the highest antioxidant activity in all of the antioxidant assays, were 19.0 and 20.0 min, respectively. The HPLC chromatograms of the separation of the compounds that were obtained using the ISP and BP extraction technology differ widely (Figure 3b,c). The ISP extraction technology is suitable for obtaining compounds 5 and 6. This indicates that the antioxidant capability, achieved using ISP extraction technology was 1.38-3.65 times higher than that achieved using BP extraction technology.
IR spectra were obtained using a Spectrum 100 FT-IR spectrometer (Perkin-Elmer, Wellesley, MA, USA). UV spectra were obtained using a U-300 spectrophotometer (Perkin-Elmer) with spectroscopy-grade methanol (Merck). 1 H-NMR and 13 C-NMR spectra were measured using a Innova 400 spectrometer (Varian, Palo Alto, CA, USA). The chemical shift values of the 1 H-and 13 C-NMR spectra are presented as δ (ppm) with TMS as the internal standard. ESI-MS data were recorded on a LCQ instrument (Thermo-Finnigan, San Jose, CA, USA), and HR-FT-MS data were measured using a JMS-SX/SX 102A tandem mass spectrometer (JEOL Ltd., Tokyo, Japan). HPLC data were obtained using the Varian ProStar 240 Solvent Delivery Module.

Sources of Sedum formosanum
Taichung Veterans General Hospital in Taichung, Taiwan, provided the S. formosanum N.E.Br. samples. These samples were dried in an oven at 40˝C and stored overnight before extraction.

Isopropanol Salting-Out Pretreatment Extraction Technology
The S. formosanum N.E.Br. powder (1 g) in aqueous 60% methanol (10 mL) was sonicated for 1 h at room temperature. Filter paper (Toyo Roshi Kaisha, Ltd., Tokyo, Japan) was used to filter the solution, which was then poured into a 50-mL round-bottomed flask. The residues were extracted twice by using 10 mL of aqueous 60% methanol (2ˆ10 mL). Under reduced pressure, the extract was evaporated, the methanol was removed, and water was added to yield a 50 mL aqueous solution. Subsequently, isopropanol (50 mL) and NaCl (12 g) were added to this aqueous solution, and this solution was separated to yield the isopropanol fraction. Under reduced pressure and at a temperature of 40˝C, this fraction was evaporated to yield a dry residue [26,27]. Using water, this residue was dissolved to prepare a working solution (1000 mg/mL) for determining the antioxidant capability.

n-Butanol Partition Extraction Technology
The S. formosanum N.E.Br. powder (1 g) in aqueous 60% methanol (10 mL) was sonicated for 1 h at room temperature. Using filter paper (Toyo Roshi Kaisha, Ltd.), the solution was filtered and then poured into a 50-mL round-bottomed flask. The residues were extracted twice by using aqueous 60% methanol (2ˆ10 mL). Under reduced pressure, the extract was evaporated, the methanol was removed, and water was added to yield a 50 mL aqueous solution. Next, n-butanol (50 mL) was added to this aqueous solution to yield the n-butanol fraction. Under reduced pressure and at a temperature of 40˝C, the n-butanol fraction was evaporated to yield a dry residue. This residue was dissolved in water, and a working solution (1000 mg/mL) was prepared to determine the antioxidant capability.

Extraction and Purification
The S. formosanum N.E.Br. powder (238 g) was extracted four-times at room temperature for 24 h using aqueous 60% methanol (400 mL). The extract was filtered, the methanol was removed through reduced pressure evaporation, and water was added to obtain a 1 L aqueous solution. An aliquot (200 mL) of the aqueous solution, isopropanol (200 mL) and sodium chloride (40 g) were added, and the solution was then separated to obtain isopropanol layers and aqueous layers. The isopropanol layers were collected, and a crude extract (33.6 g) was yielded through reduced pressure evaporation. The isopropanol extract (33.6 g) was chromatographed on silica gel and eluted using a gradient of CH 2 Cl 2 -CH 3 OH-H 2 O (from 89:10:1 to 59:40:1) to yield three fractions. The antioxidant capability of fractions 2 and 3 was higher than that of fraction 1, as determined using antioxidant assays (data not show).
Fraction 3 was chromatographed using a Sephadex-LH-20 column and eluted using a H 2 O-CH 3 OH gradient (from 100:0 to 0:100) to obtain four fractions. Fraction 2 was purified using semi-preparative HPLC with H 2 O and ACN as an eluted solvent at a flow rate of 2 mL/min to obtain kaempferol-3,7-di-O-β-D-glucopyranoside (5, 3.1 mg) and vicenin-2 (6) (5.2 mg). Figure 4 is a flow chart that displays the isolation and analytical sequences. To examine the antioxidant capability of the compounds, methanol of each compound was dissolved to prepare a work solution (1000 mg/mL). Fraction 3 was chromatographed using a Sephadex-LH-20 column and eluted using a H2O-CH3OH gradient (from 100:0 to 0:100) to obtain four fractions. Fraction 2 was purified using semi-preparative HPLC with H2O and ACN as an eluted solvent at a flow rate of 2 mL/min to obtain kaempferol-3,7-di-O-β-D-glucopyranoside (5, 3.1 mg) and vicenin-2 (6) (5.2 mg). Figure 4 is a flow chart that displays the isolation and analytical sequences. To examine the antioxidant capability of the compounds, methanol of each compound was dissolved to prepare a work solution (1000 mg/mL).   118.1, 112.6, 147.0, 71.4, 71.4, 20.2, β-D-glucopyranosyl moiety: 104.0, 75.0, 78.0, 78.0, 71.4, 62.6, ppm. Moreover, the 1 H-NMR and 13 C-NMR data of compound 1 are consistent with existing literature [12]. Therefore, on the basis of these data, compound 1 was determined as rhodiocyanoside A.   13 C-NMR data of compound 1 are consistent with existing literature [12]. Therefore, on the basis of these data, compound 1 was determined as rhodiocyanoside A.  [23]. Therefore, on the basis of these data, compound 2 was determined as rhodiocyanoside D.  [28]. Therefore, on the basis of these data, compound 3 was determined as sarmentosin.  [29]. Therefore, on the basis of these data, compound 4 was elucidated to be kaempferol-3-O-β-D-glucopyranoside.  [30]. Therefore, on the basis of these data, compound 5 was elucidated to be kaempferol-3,7-di-O-β-D-glucopyranoside.  13 13 C-NMR data of compound 6 are consistent with existing literature [31]. Therefore, on the basis of these data, compound 6 was identified as vicenin-2.  [32]. Therefore, on the basis of these data, compound 7 was determined as gastrodin.  Table 2 lists the 1 H-NMR and 13 C-NMR data.

Superoxide Radical Scavenging Activity Assay
Using the method of Lee et al., the superoxide anion scavenging activity of the sample was measured [33]. Through NADH oxidation in a non-enzymatic PMS/NADH system, superoxide anions were generated and assayed through the reduction of NBT. The reagents were prepared in a 100 mM phosphate buffer (pH 7.4). The reaction mixture contained 10 µL of the test sample (1000 ppm), 100 µL of NBT (100 µM), and 100 µL of NADH (468 µM). To this reaction mixture, 10 µL of PMS (60 µM) was added, and the mixture was incubated at room temperature for 15 min. The peak of the UV spectrophotometer changed at 560 nm and recorded the color reaction between the superoxide anion radical and NBT. Trolox was used as a standard for comparative analysis. The reaction mixture without the test sample and without PMS was used as the control and a blank, respectively. Various concentrations of the Trolox solution (20,80,121,181,222, and 242 µM) were used for plotting a calibration curve. The assay results were expressed as the mean moles of TE per moles of the compounds˘SD, and all analyses were performed in triplicate: Scavenging activity p%q " r1´pAbs sample q{pAbs control qsˆ100 (1)

Oxygen Radical Absorbance Capacity Assay
The peroxyl radical scavenging efficacy of the samples was measured using the ORAC assay [34]. A stock solution and dilutions of the test samples were prepared in potassium phosphate solution buffer (75 mmol/L), pH 7.4. Trolox and AAPH were adopted as the standard and peroxyl generator, respectively. Each fluorescein solution (150 µL) (40 nM), 25 µL of AAPH (153 mM), and 25 µL of the sample (1000 ppm) were well mixed. The temperature of the incubator was set at 37˝C for 30 min before measurement, and the fluorescence reading time was recorded every 2 min for 2 h. A fluorescence microplate reader was implemented using an excitation wavelength of 485 nm and an emission wavelength of 528 nm. The areas of the samples under the time and fluorescence intensity were determined by subtracting the area of the blank, and these areas were then compared with those of the standard curve (20, 40, 60, 80, 100, 150, and 220 µM). The assay results were expressed as the mean moles of TE per moles of the compounds˘SD, and all the analyses were performed in triplicate.

Assay on Chelation of Ferrous Ions
The chelation of the ferrous ions of the sample was estimated using the method of Lim et al. [35]. The tested sample solutions (10 µL) were added to a solution of 2.0 mM ferrous chloride (10 µL) and methanol (370 µL). The reaction was initiated by adding 5 mM ferrozine (20 µL), and this mixture was then vigorously shaken and maintained at room temperature for 10 min. The absorbance of the resulting solution was recorded at 562 nm. Various concentrations of the Trolox solution (12.5, 50, 100, 125, 175, and 220 µM) were used for plotting a calibration curve. The assay results were expressed as the mean moles of TE per moles of the compounds˘SD, and all the analyses were performed in triplicate.
Ferrous ion´chelating ability p%q " r1´pAbs sample q{pAbs control qsˆ100 3.6.4. Assay on Inhibition of Nitric Oxide Radical NO generated from aqueous SNP at physiological pH interacted with oxygen to produce nitrite ions, which were measured according to the Griess reaction [36]. NO scavengers compete with oxygen, possibly reducing the production of NO [37]. Nitric oxide radical scavenging p%q " r1´pAbs sample q{pAbs control qsˆ100 (3)

Comparison of Chromatograms after Isopropanol Salting-Out Pretreatment and n-Butanol Partition Extractions Using HPLC Separation
An appropriate volume of methanol was used to individually dissolve compounds 5 and 6 to prepare 1000 µg/mL stock solutions. The mixture of each stock solution (100 µL) was used to prepare a standard working solution that contained 500 µg/mL of each compound. Reversed-phase chromatography was performed on compounds 5 and 6 (Merck C18 column [LiChriCART 5 µm 250-4 RPC18e] and on a Phenomenex C18 guard column [AJ0-4287 4.0ˆ3.0]). The crude extracts were separated under the following experimental conditions: eluent flow rate of 0.4 mL/min; injection volume of 20 µL; detection wavelength of 210 nm; ambient temperature; and an eluent of water (A) and ACN (B) mixtures. Using a linear gradient, the elution program was optimized as follows: 0 min, 0% B; 120 min, 100% B; 125 min, 100% B; and 135 min, 0% B. The retention times (t R , min) of compounds 5 and 6 were 19.0 and 20.0 min, respectively. An appropriate volume of methanol was used to individually dissolve the crude extracts of ISP and BP extraction to obtain 1000 µg/mL work solutions. Chromatograms of these individual work solutions were obtained under identical experimental conditions for HPLC.

Statistical Analysis
Values are represented as the mean˘SD of three parallel experiments and were analyzed through a t-test and analysis of variance.

Conclusions
This investigation demonstrated that the antioxidant capability of compounds obtained through ISP extraction was 1.38-3.65 times higher than that of compounds obtained through BP extraction in antioxidant assays. Eight compounds, namely three flavonoid glycosides, three cyanogenic glycosides, and two phenolic compounds, were extracted and isolated from S. formosanum N.E.Br. through the ISP method. Compound 8 is a new compound. Except for the inhibition of NO activity, the antioxidant capability of compound 6 was 1.03-1.23 times higher than that of the standard compound (Trolox).
According to the HPLC chromatograms, ISP extraction had the highest efficiency to extract compounds 5 and 6. Furthermore, hydrophilic ISP extraction technology is superior to BP extraction technology regarding antioxidant capability and cost and reduces the risk to human health and the environment. Therefore, ISP extraction technology must replace BP extraction technology for extracting and isolating higher polar compounds of natural products. This study devised an effective method of extracting and identifying active compounds for S. formosanum N.E.Br. that can be used as nutraceuticals and pharmaceuticals.