Functionalized magnetic nanoparticles coupled with mass spectrometry for screening and identification of cyclooxygenase-1 inhibitors from natural products

Journal of Chromatography B, 960 (2014) 126–132

Contents lists available at ScienceDirect

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

Functionalized magnetic nanoparticles coupled with mass spectrometry for screening and identification of cyclooxygenase-1 inhibitors from natural products Yuping Zhang a , Shuyun Shi a,b,∗ , Xiaoqin Chen a , Mijun Peng c a

College of Chemistry and Chemical Engineering, Central South University, Changsha 410083, China Key Laboratory of Resources Chemistry of Nonferrous Metals, Central South University, Changsha 410083, China c Key Laboratory of Hunan Forest Products and Chemical Industry Engineering, Jishou University, Zhangjiajie 427000, China b

a r t i c l e

i n f o

Article history: Received 27 December 2013 Accepted 16 April 2014 Available online 26 April 2014 Keywords: Cyclooxygenase-1 Ligand fishing MS Turmeric Curcuminoid

a b s t r a c t Development of simple and effective methods for high-throughput, high-fidelity screening and identification of cyclooxygenase-1 (COX-1) inhibitors from natural products are important for drug discovery to treat inflammation and carcinogenesis. Here, we developed a new screening assay based on cyclooxygenase-1 (COX-1) functionalized magnetic nanoparticles (i.e. Fe3 O4 @SiO2 –COX1) for solid phase ligand fishing, and then mass spectrometry (MS) was applied for structural identification. Incubation conditions were optimized. High specificity for isolating COX-1 inhibitors was achieved by testing positive control, indomethacin, with active and inactive COX-1. Moreover, high stability of immobilized COX-1 (remained 95.3% after ten consecutive cycles) allows the analysis reproducible. When applied to turmeric extract, four curcuminoids (i.e. curcumin, demethoxycurcumin, bisdemethoxycurcumin, and 1-(4-hydroxy-3,5-dimethoxyphenyl)-7-(4-hydroxy3-methoxyphenyl)-(1E,6E)-1,6-heptadiene-3,5-dione), difficult to be distinguished from original MS spectrum of turmeric extract, were isolated as main COX-1 inhibitors. Their structures were characterized based on their accurate molecular weight and diagnostic fragment ions. The results indicated that the proposed method was a simple, robust and reproducible approach for the discovery of COX-1 inhibitors from complex matrixes. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Enzymes are attractive drug targets, and enzyme inhibitors represent almost half the drugs used in clinical today [1]. Therefore, discovery of new enzyme inhibitors has been one of the major interests and challenges in drug discovery and development process [2]. Historically, majority of new drugs were generated from natural products or compounds derived from natural products [3]. From 1981 to 2010, 50% of all the marketed-new chemical entities were shown to be of natural origin [4]. Undoubtedly, natural products continue to play a highly significant role in the discovery of drug leads.

∗ Corresponding author at: School of Chemistry and Chemical Engineering, Central South University, Changsha 410083, China. Tel.: +86 731 88830833; fax: +86 731 88830833. E-mail address: [email protected] (S. Shi). http://dx.doi.org/10.1016/j.jchromb.2014.04.032 1570-0232/© 2014 Elsevier B.V. All rights reserved.

Natural products embody hundreds or even thousands of secondary metabolites, but a few of them are responsible for the pharmacological activity. The key step in natural products research is to develop high throughput, high fidelity methods for discovery of bioactive components. Moreover, the process of modernization and globalization of natural products necessitate the biochemical profiles analysis for quality control purpose. To address these requirements, biofingerprint chromatogram based on ligand fishing has been proposed to provide unique information on the multiple bioactive compounds direct interaction with drug targets (i.e. protein, enzyme and receptor). Ligand fishing assay has been considered as the most convenient and efficient technology with high-selectivity and high-throughput in early stage bioactive components discovery, which include, but are not limited to, centrifugation [6], ultrafiltration [7–9], equilibrium dialysis [10], microdialysis [11], magnetic solid phase fishing [7,12], and surface plasmon resonance [13,14]. Immobilized drug targets on solid surface are more robust and resistant to environmental changes than solution phase drug targets [7,15]. At the same time,

Y. Zhang et al. / J. Chromatogr. B 960 (2014) 126–132

O

127

O

R1

R3

HO

OH R2

R1 = R3 = OCH3, R2 = H; Curcumin (1) R1 = R2 = H, R3 = OCH3; Demethoxycurcumin (2) R1 = R2 = R3 = H; Bisdemethoxycurcumin (3) R1 = R2 = R3 = OCH3; 1-(4-Hydroxy-3,5-dimethoxyphenyl)-7(4-hydroxy-3-methoxyphenyl)-(1E,6E)-1,6-heptadiene-3,5-dione (4) Fig. 1. Chemical structures of four investigated curcuminoids.

magnetic nanoparticles could be separated from solution conveniently. Therefore, ligand fishing based on functionalized magnetic nanoparticles has been proved to be an exciting method to screen ligands from natural products [7,12]. COX-1 is known as a housekeeping enzyme constitutively expressed in almost all the mammalian tissues. The prostaglandins produced from arachidonic acid transformation through the participation of COX-1 play an important role in inflammation and carcinogenesis [16]. Importantly, platelet COX-1 is the target of one of the most efficacious antithrombotic agents used for prevention of vascular occlusive events (i.e. aspirin) [17], which thereby provides the rationale for the development of COX-1 inhibitors. Many reports have focused on the detection of COX-1 inhibitory activity of commercial isolated compounds from natural products or synthesized compounds [17–20]. However, much less attention has been paid to develop facile screening assay to discover COX-1inhibitors from natural products. Turmeric is dried powder from rhizomes of Curcuma longa L. (Zingiberaceae), which has been used as a traditional medicine for its various pharmacological activities such as anti-inflammatory, antiviral, antioxidant, anti-infectious activities, anti-parasitic infection, anti-mutagenic effect, and anticancer [21,22]. The underlying mechanisms of these effects involve the recognition of various molecular targets, such as enzymes and protein kinases. Recent studies showed that curcuminoids, the major yellow pigment and active components of turmeric [23], had significantly higher inhibitory effects on the peroxidase activity of COX-1 than COX-2 [24]. However, no reports systematically analyzed COX-1 inhibitors in turmeric. Herein, we report proof of principle for the first time of integration of COX-1 functionalized magnetic nanoparticles and direct infusion MS for facile, specific screening and identification of COX-1 inhibitors from complex natural products. Four curcuminoids (i.e. curcumin [1], demethoxycurcumin [2], bisdemethoxycurcumin [3], and 1-(4-hydroxy-3,5-dimethoxyphenyl)-7-(4-hydroxy-3methoxyphenyl)-(1E,6E)-1,6-heptadiene-3,5-dione [4], Fig. 1) with COX-1inhibitory activity were isolated from turmeric. 2. Experimental 2.1. Chemicals and reagents Ovine COX-1, arachidonic acid, prostaglandin E2 (PGE2 ), [d4 ]-PGE2 , co-factors l-epinephrine and hematin, 3aminopropyltrimethoxysilane (APTMS), tetraethyl orthosilicate, and glutaraldehyde (25%, w/v aqueous solution) were acquired from Sigma–Aldrich Chemicals (St. Louis, MO, USA). The HPLC grade acetonitrile and methanol were bought from Tedia Company,

Fig. 2. TEM image of the synthesized Fe3 O4 @SiO2 –COX-1 nanoparticles.

Inc. (Ohio, USA). Ultrapure water (18.2 M) was obtained from a Milli-Q water purification system (Millipore, Bedford, MA, USA). Buffer solution used all over the experiments was 10 mM ammonium acetate buffer solution with pH at 7.4. All of other chemicals were analytical grade and purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Reference compounds, curcumin (1), demethoxycurcumin (2), bisdemethoxycurcumin (3), and positive control sample, indomethacin, with purities over 99% were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). 2.2. Preparation of turmeric extraction Turmeric was purchased from local drugstore in Changsha, which was identified as C. longa L. by one of the authors, Prof. Mijun Peng. Approximate 20 g of turmeric was powdered and extracted with 300 ml of 75% (v/v) ethanol three times, each for 3 h, and the filtrates were concentrated on a rotary evaporator (Shanghai Yarong Biochemical Instrument Factory, Shanghai, China) under reduced pressure at 40 ◦ C to yield dried residue (3.1 g). A stock solution (3 mg/ml) of the residue was then stored at 4 ◦ C for further experiments. 2.3. COX-1 inhibition assay COX-1 inhibitory assay was performed according to a previously described PGE2 (a stable oxidation product resulting from COX-1 oxidation of arachidonic acid) detection method [25]. In brief, 2 ␮l of hematin (100 ␮M) mixed with 10 ␮l of l-epinephrine (40 mM) and made up with buffer solution to a final volume of 140 ␮l. Then 20 ␮l of COX-1 (0.1 ␮g) was added and incubated at 25 ◦ C for 2 min. After that, 20 ␮l of different concentrations of COX-1 ligands were added and preincubated at 37 ◦ C for 10 min. The COX-1 inhibition reaction was initiated by adding 20 ␮l of arachidonic acid (50 ␮M) and terminated by adding 20 ␮l of HCl (2.0 M). The concentration of product PGE2 was detected by HPLC–MS/MS method by using [d4 ]-PGE2 as surrogate standard. The inhibitory activity was determined by comparing the amount of PGE2 produced with that of

128

Y. Zhang et al. / J. Chromatogr. B 960 (2014) 126–132

Fe3O4

indomethacin

mAU

800 580

Fe3O4@SiO2

10 a b c

470

1631

20

1090

Transmittance (%)

(A)

0

4000 3500 3000 2500 2000 1500 1000

500

0

2

Wavelength ( cm-1)

a negative control (buffer solution). The COX-1 inhibitory activity was expressed as the half maximal inhibitory concentration (IC50 ). 2.4. COX-1 functionalized magnetic nanoparticles for ligand fishing 2.4.1. Preparation of Fe3 O4 @SiO2 –COX-1 nanoparticles The APTMS-modified Fe3 O4 @SiO2 nanoparticles were synthesized according to our previous work [26]. COX-1 was then immobilized onto APTMS-modified Fe3 O4 @SiO2 nanoparticles by a typical glutaraldehyde activation procedure. Briefly, APTMSmodified Fe3 O4 @SiO2 nanoparticles (5 mg) were suspended in glutaraldehyde solution (2 ml, 5%) for 1 h. After activation, the nanoparticles were collected by magnetic separation and then a solution of COX-1 (2.0 ml, containing 1000 ␮g of COX-1) was added and shaken for 2 h, after which the magnetic Fe3 O4 @SiO2 –COX-1 nanoparticles were collected by magnet and washed three times with buffer solution, and then dispersed in buffer solution and stored at 4 ◦ C for further experiments. The supernatant and elution after COX-1 immobilization process were combined to determine the amount of COX-1 immobilized onto the Fe3 O4 @SiO2 surface by measuring the concentration of PGE2 formed by COX-1 [25]. Fe3 O4 @SiO2 –inactive COX-1 was synthesized with the same procedures as mentioned above, while inactive COX-1 was prepared by boiling it in water for 10 min. 2.4.2. Characterizations of Fe3 O4 @SiO2 –COX-1 nanoparticles Transmission electron microscopy (TEM) (JEM-2100F, JEOL, Japan) was used to investigate the morphology of functionalized magnetic nanoparticles. The infrared spectrum was recorded on a Fourier transform infrared spectrometer (FT-IR) (Nicolet 6700, Thermo Nicolet Corp., Waltham, MA, USA) in the wavelength range of 4000–400 cm−1 . Magnetization was measured at room temperature in a vibration sample magnetometer (VSM7307, Lake Shore, USA).

100

Relative intensity

Fig. 3. IR spectra of the Fe3 O4 and Fe3 O4 @SiO2 magnetic nanoparticles.

4

T/min

6

8

10

358.0849

(B)

COOH H3CO CH3

50

N O

Cl

indomethacin

0 200

400

m/z

600

800

Fig. 5. (A) HPLC chromatograms of indomethacin (8.0 ␮g/ml) (a) and eluent after ligand fishing assay by Fe3 O4 @SiO2 –active COX-1 (b) and Fe3 O4 @SiO2 –inactive COX1 (c); (B) ESI-MS spectrum of the bound indomethacin eluted by Fe3 O4 @SiO2 –active COX-1 ([M+H]+ at m/z 358.0849, 0.8 ppm error).

2.4.3. Assay verification COX-1 inhibitor, indomethacin, was selected as positive control to assess specificity of the method, optimize incubation temperature and time, and reusability of immobilized COX-1. 2.4.4. Ligand fishing assay The mixtures of Fe3 O4 @SiO2 –COX-1 nanoparticles (5 mg) and turmeric extract (3 mg/ml, 2 ml) were incubated and shaken at 30 ◦ C for 30 min and then the Fe3 O4 @SiO2 –COX-1-ligand complexes were separated by magnet after being washed three times with 2 ml of buffer solution to remove non-specific absorbents. The Fe3 O4 @SiO2 –COX-1-ligand complexes were then incubated with 2 ml of methanol to disrupt ligands. The supernatant with ligands was carefully collected and filtered with 0.45 ␮m membrane for direct infusion MS analysis. Identical incubation using Fe3 O4 @SiO2 –inactive COX-1 was carried out for control experiments.

60

a

2.5. HPLC-DAD and ESI-MS analysis

b

HPLC-DAD analysis for positive control was conducted on a Thermo Accela LC system (ThermoFisher, San Jose, CA, USA). Isocratic chromatography was conducted on a SunFireTM C18 column (150 mm × 4.6 mm i.d., 5 ␮m, Waters, Milford, MA, USA) in tandem with a Phenomenex C18 guard cartridge (4.0 mm × 3.0 mm, Phenomenex, Torrance, CA, USA). Mobile phase was a 0.4% acetic acid solution/acetonitrile (V/V = 45/55) mixtures. Flow rate was set at 1.0 ml/min while temperature was controlled at 30 ◦ C. Sample injection volume was 20 ␮l. Spectra were recorded from 190 to 400 nm (peak width 0.2 min and data rate 1.25 s−1 ) while chromatogram was acquired at 228 nm.

M(emu/g)

40 20 0 -20 -40 -60 -12000 -8000 -4000

0

4000

8000 12000

H(Oe) Fig. 4. Magnetization curves of the Fe3 O4 (a) and Fe3 O4 @SiO2 –COX-1 (b) magnetic nanoparticles.

Y. Zhang et al. / J. Chromatogr. B 960 (2014) 126–132

100

339.1236

100

(A)

129

255.1023 (C)

200

300

400 500 m/z

600

700

0 100

800

150

285.1125

100 (B)

250 m/z

269.1175

219.0654 229.0868 245.0819

200

+

[M+H] 339.1236 321.1132

0 100

177.0554 147.0443

189.0551

309.1131

50

175.0757

50

145.0650

Relative intensity

Relative intensity

369.1341

300

350

400

225.0913

100 (D)

150

200

250 m/z

300

350

400

0 100

150

200

250 m/z

291.1017 309.1130

0 100

351.1231

219.0660

299.1285

175.0756

50

239.1074

+

[M+H] 369.1342

189.0549 199.0762 215.0711

50

145.0655

245.0817

Relative intensity

147.0442

259.0968

Relative intensity

177.0549

+

[M+H]

300

350

400

Fig. 6. ESI-MS spectrum of the mixtures of compounds 1–3 (A), and MS/MS spectra of compounds 1 (B), 2 (C) and 3 (D) obtained at collision energy of 30 V.

Bioactive components were analyzed and identified by a highresolution ion trap mass spectrometer (ThermoFisher Exactive orbitrap, San Jose, CA, USA) equipped with an ESI source in positive ion detection mode, which was calibrated using the calibration mixtures that manufacture provided, corresponding to 50,000 resolution and 2 Hz scan speed. Initial instrument parameters were tuned by infusing curcumin, demethoxycurcumin, and bisdemethoxycurcumin, or their mixtures, each at 1 ␮mol/l with flow rate of 20 ␮l/min using a fusion 100 syringe pump (Chemyx, Inc., Stafford, TX, USA). The final parameters were as follows: nitrogen sheath gas flow rate, 28 (arbitrary units); spray voltage, 4.5 kV; capillary temperature, 275 ◦ C; capillary voltage, 42 V; tube lens voltage, 95 V; skimmer voltage, 29 V. The instrument was operated in full scan mode from m/z 100 to 800 with a resolution of 50,000, and fullscan MS mode from m/z 100 to 800 with a resolution of 25,000 at high energy collision induced dissociation (HCD) set at 30 V, while the high-purity nitrogen gas was used as collision gas in the ion trap. 3. Results and discussion 3.1. Characterization of Fe3 O4 @SiO2 –COX-1 nanoparticles Magnetic nanoparticles (e.g. Fe3 O4 ) were widely used for enzyme immobilization and separation due to its higher surface area, lower mass transfer resistance and easier solid–liquid separation. Fe3 O4 based ligand fishing has been successfully applied for identifying bioactive components from natural products, such as binders of serum albumin [12,7,26], tyrosinase [7], ␣-glucosidase [27,28], maltase [29], and xanthine oxidase [30]. In addition, Fe3 O4 @SiO2 could promote enzyme stability because of the

prevention of surface oxidation and agglomeration of the silica shell [31]. Our previous investigation also indicated that the stability of enzyme immobilized on Fe3 O4 @SiO2 was mostly improved compared with that on Fe3 O4 [30]. Therefore, Fe3 O4 @SiO2 was selected as the material for enzyme immobilization. Figs. 2–4 show that COX-1 was successfully immobilized onto the surface of Fe3 O4 @SiO2 nanoparticles [26]. No remanence was detected after removing applied magnetic field, revealing that Fe3 O4 @SiO2 –COX-1 was superparamagnetic. Moreover, Fe3 O4 @SiO2 –COX-1 accumulated within 5 s in solution under magnetic field and dispersed quickly with a slight shake once the magnetic field was removed (Fig. S1), which indicated that Fe3 O4 @SiO2 –COX-1 had high magnetic responsivity. The amount of immobilized COX-1 on Fe3 O4 @SiO2 surface was about 177 ␮g/mg. 3.2. Assay verification Indomethacin was selected as positive control sample to optimize the incubation conditions and evaluate the feasibility and specificity of Fe3 O4 @SiO2 –COX-1 based ligand fishing assay. Working pH and temperature would affect the activity of COX-1, while incubation time affected the quantity of bioactive components bound to COX-1. Therefore, some essential incubation conditions including working pH (from 5 to 9), incubation temperature (from 25 ◦ C to 65 ◦ C), and incubation time (from 0 min to 60 min) were optimized during ligand fishing process. The results revealed that the highest COX-1 activity and binding capacity of indomethacin could be achieved in the following conditions, pH at 7.4, temperature at 37 ◦ C, and time for 30 min. Washing steps were also optimized to minimize nonspecific binding and dissociate bound components from Fe3 O4 @SiO2 –COX-1. Finally, working

Y. Zhang et al. / J. Chromatogr. B 960 (2014) 126–132

Seldom changed amount/activity of immobilized COX-1 is a crucial prerequisite to obtain highly reproducible and accurate data by capturing exactly the same amount of bioactive components in a series of binding/dissociation cycles. The reusability of Fe3 O4 @SiO2 –COX-1 was evaluated by calculating the amount of indomethacin bound to COX-1 under the same conditions for ten consecutive cycles. In avoid of particle loss, Fe3 O4 @SiO2 –COX1 was separated by prolonged exposure to magnetic field. The amount of indomethacin bound to Fe3 O4 @SiO2 –COX-1 retained 95.3% after ten consecutive cycles. The stability of the immobilized COX-1 was tested through inter-day assays. No significant change for bound indomethacin appeared after 5 days storage in buffer solution at 4 ◦ C, and RSD was less than 8%. Stable immobilized COX-1 did not need to consume additional COX-1, which can decrease test costs and enhance experimental efficiency and sample throughput. Excellent reproducibility was attained among three batches of Fe3 O4 @SiO2 –COX-1 (RSD < 8%). Encapsulation with SiO2 could form the hydrophilic surface to prevent Fe3 O4 nanoparticles from oxidizing and aggregating [30], which would effectively protect the expression of COX-1 on the Fe3 O4 @SiO2 surface from activity loss during repeated usage. Hu et al. [27] and Liu et al. [30] have found that the reusability and reproducibility of immobilized enzyme were superior to those of the free one.

251.1085 (A)

50 267.2029

0 100 100

200

300

400

500 m/z

600

700

600

700

800

369.1342 1

(B)

50 339.1236 2 309.1133 3

0 100

200

300

399.1449

3.3. Reusability and reproducibility of immobilized COX-1

100

Relative intensity

buffer was used as washing solution and methanol was selected to dissociate bound ligands. Fig. 5(A and B) shows that indomethacin bound to active COX1 but not to inactive COX-1, which indicated immobilized COX-1 was enzymatically active. No non-specific binding of indomethacin occurred, which showed the assay was specific and feasible.

Relative intensity

130

4

400 500 m/z

800

Fig. 7. ESI-MS spectra of turmeric extract (A) and eluent after Fe3 O4 @SiO2 –COX-1 based ligand fishing (B).

3.4. Analysis of curcuminoids by ESI-MS Curcuminoids, the major phytochemicals in turmeric [23], had significantly high inhibitory effects on peroxidase activity of COX-1 [24]. Then, three main curcuminoids [curcumin (1), demethoxycurcumin (2), and bisdemethoxycurcumin (3)] have been mixed and infused for optimization of initial ESI-MS parameters and investigation of diagnostic fragment ions for structural identification. Positive and negative ions were applied in ESI-MS experiments, and the parameters were optimized automatically. Results indicated that both positive and negative ions could give clean ion spectrum, good resolution and low background noise (Fig. 6A shows the ESI-MS spectrum of the mixture of compounds 1–3 in positive ion mode). However, the fragmentation ions were dramatically different between positive and negative ion modes. Simple MS spectrum with few fragmentation ions occurred in negative ion mode (data not shown), which was not beneficial for structural identification. Fig. 6(C and D) shows the ESI-MS spectra of compounds 1–3 in positive ion mode, which indicated that three components had similar fragmentation patterns and fragmentation behavior always occurred on the heptanoid moiety for their rearrangement and new bond formation. Compounds 1–3 had quasimolecular ions [M+H]+ m/z at 369.1341 (C21 H21 O6 , 0.8 ppm error), 339.1236 (C20 H19 O5 , 1.2 ppm error) and 309.1131 (C19 H17 O4 , 1.3 ppm error). Low abundance product ions m/z at 351.1232, 321.1132, and 291.1017 ([M+H−H2 O]+ ) were a neutral loss of H2 O from precursor ions. The rearrangement of diketone group and then loss of diketocyclopropane moiety could form fragment ions [M+H−C3 H2 O2 ]+ with m/z at 299.1285, 269.1175 and 239.1074, while loss of 1hydroxy-3-ketocyclobutene moiety (C4 H4 O2 ) could give fragment ions m/z at 285.1125, 255.1023, and 225.0913 with a high intensity.

Similarly, fragment ions m/z at 245.0817 [M+H−C7 H7 O2 −H]+ for 1, 245.0817 [M+H−C6 H6 O−H]+ for 2 and 215.1 [M+H−C6 H6 O−H]+ for 3 arose from rearrangement of diketone group and then loss of an aryl group and a hydrogen atom. In addition, fragment ions m/z at 259.0968, 229.0868, and 199.0762 corresponded to rearrangement of diketone group and then neutral loss of 1-hydroxy-5-ketocyclo1,3-hexadiene group (C6 H6 O2 ). Fragment ions m/z at about 219.1 and 189.1 were observed probably for hydrogen migration on diketone group followed by a neutral loss of arylethene moiety. At the same time, an oxo migration on diketone group and then a neutral loss of 3-arylcarboxyprop-2-ene moiety can produce fragment ions with m/z about 175.1 and 145.1. Moreover, fragment ions m/z at about 177.1 and 147.0 corresponded to 3,4-bond cleavage of precursor ions followed by neutral loss of a 1-aryl-3-hydroxy-1,3butadiene moiety. The results coincided well with the published data [32,33]. 3.5. COX-1 inhibitors in turmeric extract Turmeric extract showed potent COX-1 inhibitory activity with IC50 value at 653.8 ␮g/ml, which indicated that turmeric extract was rich of COX-1 inhibitors. Natural products are normally complex and, more importantly, contain many non-active components, and then the identification of bioactive components has been a challenge. Direct MS detection with/without chromatographic separation cannot distinguish the bioactive components from complex system, moreover, it encounters significant problems of ion suppression due to matrix effects, therefore, some components, especially low-content components, sometimes cannot be identified. The usual used procedure for discovery of COX-1 inhibitors was bioassay-guided fractionation

Y. Zhang et al. / J. Chromatogr. B 960 (2014) 126–132

177.0550

50

0 100

150

200

+

[M+H] 399.1449

250 300 m/z

381.1342

329.1394

205.0862 219.0661 245.0814 249.0766 275.0916 289.1071

207.0653 175.0763

Relative intensity

analysis times (1–2 min MS analysis time) over previously reported HPLC–ESI-MS technology [6–14], which therefore permitted a high sample throughput. ESI-MS spectra of turmeric extract and eluent after Fe3 O4 @SiO2 –COX-1 based ligand fishing were shown in Fig. 7. Numerous compounds present in turmeric extract resulted in a complicated ESI-MS spectrum, and the main curcuminoids (i.e. compounds 1–3) were difficult to be distinguished from the total ion chromatogram in Fig. 7A because of the existence of interferences of many non-active compounds. However, the MS signal of bioactive curcuminoids can be easily found and highly detected after Fe3 O4 @SiO2 –COX-1 based ligand fishing experiment (Fig. 7B) because of the preliminary removal of interference and inactive compounds. The higher intensity peaks in Fig. 7A could not be detected in Fig. 7B, which suggested that these compounds had no COX-1 inhibitory activities. To our best knowledge, curcuminoids identified in turmeric showed no isomers [35–37]. Peaks 1–3 gave the same [M+H]+ ions with those of curcumin, demethoxycurcumin and bisdemethoxycurcumin, respectively. Peak 4 exhibited [M+H]+ ion at m/z 399.1449 (C22 H23 O7 , 1.3 ppm error, 30 Da greater than that of curcumin), indicating that peak 4 contained one more methoxyl group than curcumin. Further, the structural verification was confirmed by fragmentation ions. Fig. 8 shows the ESI-MS spectrum of purified peak 4. The fragmentation patterns were similar with those of curcuminoids displayed in Fig. 6B–D, while some fragment ions were 30 Da greater than those of curcuminoids, which displayed that peak 4 had one more methoxyl group in an aryl ring. Then four detected ions in ESI-MS spectrum were identified as curcumin

315.1234

100

350

400

131

450

Fig. 8. ESI-MS spectrum of peak 4 in Fig. 7B with collision energy at 30 V.

[33,34], during which the repeated column separation procedures were time-consuming and labor intensive and sometimes bioactive components cannot be isolated due to dilution effects or decomposition [5]. Ramirez-Cisneros et al. screened ligands to COX-1 using a simple and rapid method, pulsed ultrafiltrationHPLC [25], but free COX-1 used in pulsed ultrafiltration was difficult to be reused. Here, Fe3 O4 @SiO2 –COX-1 based ligand fishing combined with ESI-MS analysis may offer a convenient and effective approach to tackle above mentioned problems. It is noted that the direct ESI-MS analysis has advantageous including short

OH H3CO H3 CO

OCH3

OCH3

HO HO OCH3 m/z 315.1234

O H3CO

OH

OCH3

HO

m/z 329.1394

OCH3

O

O

O

H 3CO

OH

H 3CO

OH OH m/z 289.1071

O

O

-ArH

OCH3

OCH3

HO

OCH 3

HO

OH

m/z 177.0550

H 3CO

OH OCH 3

O

HO

m/z 399.1449 [M+H]+

H3 CO

OH

H3 CO HO

OH

O

OH

H

H

OCH 3 m/z 207.0653

OCH3

H3 CO

O OH

H 3CO

OCH 3

HO

O H3 CO

OH OCH3

HO

OH OCH3

O H3 CO

O

OH

OH

OH OCH 3

m/z 381.1342

H3CO

H3CO

H

H HO

OCH 3

OCH3

HO

O

OH

HO

HO

-H 2O

OCH 3

OCH 3

O

m/z 245.0812

HO

O

O

HO

m/z 275.0916

H3CO

O

OH

H3CO

O

OCH3

m/z 249.0766

HO

m/z 219.0661

H 3CO HO

m/z 175.0763

Fig. 9. Main MS fragmentation pathway of 1-(4-hydroxy-3,5-dimethoxyphenyl)-7-(4-hydroxy-3-methoxyphenyl)-(1E,6E)-1,6-heptadiene-3,5-dione.

132

Y. Zhang et al. / J. Chromatogr. B 960 (2014) 126–132

(peak 1), demethoxycurcumin (peak 2), bisdemethoxycurcumin (peak 3) and 1-(4-hydroxy-3,5-dimethoxyphenyl)-7-(4-hydroxy3-methoxyphenyl)-(1E,6E)-1,6-heptadiene-3,5-dione (peak 4). The proposed fragmentation pathways of peak 4 were given in Fig. 9. Peaks 1–4 have ever been isolated from Curcuma genus, and the in vitro assay showed that curcumin, demethoxycurcumin, and bisdemethoxycurcumin had significant COX-1 inhibiting activity at 125 ␮g/ml [38]. This is the first time to report that 1-(4-hydroxy-3,5-dimethoxyphenyl)-7-(4-hydroxy3-methoxyphenyl)-(1E,6E)-1,6-heptadiene-3,5-dione had potent COX-1 inhibitory activity. 4. Conclusion We have proposed a new assay based on Fe3 O4 @SiO2 –COX-2 ligand fishing for facile and rapid screening COX-1 inhibitors from complex natural products, and then direct infusion MS was developed for structural elucidation. The proposed method was verified by COX-1 inhibitor, indomethacin, which indicated the method could successfully isolate COX-1 inhibitors in rid of nonspecific binding. Another particularly noteworthy advantage of the developed method was that immobilized COX-1 was stable, which could continuously perform multiple assays, and then enhanced sample throughput and reduced assay cost, moreover, high reproducibility could be attained. When applied for turmeric, four bioactive curcuminoids were screened and characterized. The results indicated that the proposed method could definitely accelerate the discovery of bioactive compounds from natural products. In addition, the high efficiency and specificity of the proposed method and the wide range of immobilization biomacromolecules (enzyme, protein, receptor, etc.) provided convenient conditions for screening of a broad range of bioactive components from complex natural products. Acknowledgements Partial support of this work by a grant from the National Natural Science Foundation of China (21275163), the Science and Technology Program of Hunan Province, China (2012FJ2006), the Shenghua Yuying project of Central South University, China, aid program for Science and Technology Innovative Research Team (Chemicals of Forestry Resources and Development of Forest Products) in Higher Educational Institutions of Hunan Province, China, and the Postdoctoral Science Foundation of Central South University, China. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jchromb. 2014.04.032.

References [1] R.A. Copeland, Evaluation of Enzyme Inhibitors in Drug Discovery: A Guide for Medicinal Chemists and Pharmacologists, John Wiley & Sons, New York, 2013. [2] D.J. Newman, G.M. Cragg, J. Nat. Prod. 70 (2007) 461–477. [3] J.W.H. Li, J.C. Vederas, Science 325 (2009) 161–185. [4] D.J. Newman, G.M. Cragg, J. Nat. Prod. 75 (2012) 311–335. [5] S.Y. Shi, H.H. Zhou, Y.P. Zhang, X.Y. Jiang, X.Q. Chen, K.L. Huang, Trends Anal. Chem. 28 (2009) 865–877. [6] S.L. Su, L. Yu, Y.Q. Hua, J.A. Duan, H.S. Deng, Y.P. Tang, Y. Lu, A.W. Ding, Biomed. Chromatogr. 22 (2008) 1385–1392. [7] L.L. Liu, S.Y. Shi, X.Q. Chen, M.J. Peng, J. Chromatogr. B 932 (2013) 19–25. [8] X. Chen, L.X. Li, S. Chen, Y.T. Xu, Q. Xia, Y. Guo, X. Liu, Y.T. Tang, T.J. Zhang, Y. Chen, C. Yang, W.Q. Shui, Anal. Chem. 85 (2013) 7957–7965. [9] S.Y. Shi, M.J. Peng, Y.P. Zhang, S. Peng, Anal. Bioanal. Chem. 405 (2013) 4213–4223. [10] G.Y. Hou, J. Niu, F.R. Song, Z.Q. Liu, S.Y. Liu, J. Chromatogr. B 923 (2013) 1–7. [11] B.V. de Araujo, J.V. Laureano, L.D. Grunspan, T. Dalla Costa, L. Tasso, J. Chromatogr. B 919 (2013) 62–66. [12] L.S. Qing, Y. Xue, Y. Zheng, J. Xiong, X. Liao, L.S. Ding, B.G. Li, Y.M. Liu, J. Chromatogr. A 1217 (2010) 4663–4668. [13] M.J. Peng, Y.P. Zhang, S.Y. Shi, S. Peng, J. Chromatogr. B 940 (2013) 86–93. [14] Y.P. Zhang, S.Y. Shi, J.F. Guo, Q.P. You, D.S. Feng, J. Chromatogr. A 1293 (2013) 92–99. [15] K. Ashtari, K. Khajeh, J. Fasihi, P. Ashtari, A. Ramazani, H. Vali, Int. J. Biol. Macromol. 50 (2012) 1063–1069. [16] L.J. Crofford, M.D. Breyer, C.V. Strand, F. Rushitzka, K. Brune, M.E. Farkouh, L.S. Simon, J. Rheumatol. 22 (2006) 1403–1408. [17] P. Vitale, S. Tacconelli, M.G. Perrone, P. Malerba, L. Simone, A. Scilimati, A. Lavecchia, M. Dovizio, E. Marcantoni, A. Bruno, P. Patrignani, J. Med. Chem. 56 (2013) 4277–4299. [18] R. Gautam, A. Srivastava, S.M. Jachak, A. Saklani, Fitoterapia 81 (2010) 45–49. [19] X.P. Chu, Q.F. Zhou, S. Zhao, F.F. Ge, M. Fu, J.P. Chen, T. Lu, Chin. Chem. Lett. 24 (2013) 120–122. [20] R. Gautam, S.M. Jachak, V. Kumar, C.G. Mohan, Bioorg. Med. Chem. Lett. 21 (2011) 1612–1616. [21] M.K. Kim, G.J. Choi, H.S. Lee, J. Agric. Food Chem. 51 (2003) 1578–1581. [22] K.Y. Yang, L.C. Lin, T.Y. Tseng, S.C. Wang, T.H. Tsai, J. Chromatogr. B 853 (2007) 183–189. [23] V. Kakkar, S. Singh, D. Singla, S. Sahwney, A.S. Chauhan, G. Singh, I.P. Kaur, J. Chromatogr. B 878 (2010) 3427–3431. [24] J. Hong, M. Bose, J. Ju, J.H. Ryu, X.X. Chen, S.M. Sang, M.J. Lee, C.S. Yang, Carcinogenesis 25 (2004) 1671–1679. [25] H.M. Cao, R. Yu, Y. Choi, Z.Z. Ma, H.J. Zhang, W. Xiang, D.Y.W. Lee, B.M. Berman, K.D. Moudgil, H.H.S. Fong, R.B. van Breemen, Pharmacol. Res. 61 (2010) 519–524. [26] Y.P. Zhang, M.K. Nie, S.Y. Shi, Q.P. You, J.F. Guo, L.L. Liu, Food Chem. 146 (2014) 56–64. [27] F.L. Hu, C.H. Deng, X.M. Zhang, J. Chromatogr. B 871 (2008) 67–71. [28] L.S. Qing, N. Tang, Y. Xue, J. Liang, Y.M. Liu, X. Liao, Anal. Methods 4 (2012) 1612–1615. [29] Y. Tao, Z. Chen, Y.F. Zhang, Y. Wang, Y.Y. Cheng, J. Pharm. Biomed. Anal. 78–79 (2013) 190–201. [30] L.L. Liu, S.Y. Shi, H.D. Zhao, J.G. Yu, X.Y. Jiang, X.Q. Chen, J. Chromatogr. B 945–946 (2014) 163–170. [31] F.X. Hu, S.H. Chen, R. Yuan, Sens. Actuators B: Chem. 176 (2013) 713–722. [32] H.L. Jiang, Á. Somogyi, N.E. Jacobsen, B.N. Timmermann, D.R. Gang, Rapid Commun. Mass Spectrom. 20 (2006) 1001–1012. [33] M. Resch, J. Heilmann, A. Steigel, R. Bauer, Planta Med. 67 (2001) 437–442. [34] M.A. Ramirez-Cisneros, M.Y. Rios, R. Rios-Gomez, A.B. Aguilar-Guadarrama, Planta Med. 78 (2012) 1942–1948. [35] W. Li, H.B. Xiao, L.X. Wang, X.M. Liang, Chin. J. Chromatogr. 27 (2009) 264–269. [36] T. Masuda, A. Jitoe, J. Isobe, N. Nakatani, S. Yonemori, Phytochemistry 32 (1993) 1557–1560. [37] A. Suksamrarn, S. Eiamong, P. Piyachaturawat, J. Charoenpiboonsin, Phytochemistry 36 (1994) 1505–1508. [38] R.S. Ramsewak, D.L. DeWitt, M.G. Nair, Phytomedicine 7 (2000) 303–308.