Global Profiling and Novel Structure Discovery Using Multiple Neutral

Nov 25, 2015 - In this work, we describe a new approach to globally profile and discover novel compounds from an herbal extract using multiple neutral...
1 downloads 11 Views 1MB Size
Subscriber access provided by The Chinese University of Hong Kong

Article

Global profiling and novel structure discovery using multiple neutral loss / precursor ion scanning combined with substructure recognition and statistical analysis (MNPSS): Characterization of terpene-conjugated curcuminoids in Curcuma longa as a case Xue Qiao, Xiong-hao Lin, Shuai Ji, Zheng-xiang Zhang, Tao Bo, De-an Guo, and Min Ye Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b02729 • Publication Date (Web): 25 Nov 2015 Downloaded from http://pubs.acs.org on November 27, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Global profiling and novel structure discovery using multiple neutral loss / precursor ion scanning combined with substructure recognition and statistical analysis (MNPSS): Characterization of terpene-conjugated curcuminoids in Curcuma longa as a case

Xue Qiao†, Xiong-hao Lin†, Shuai Ji†, Zheng-xiang Zhang§, Tao Bo§, De-an Guo†,‡, and Min Ye †,‡,*

Affiliations: †

State Key Laboratory of Natural and Biomimetic Drugs, School of Pharmaceutical Sciences,

Peking University, 38 Xueyuan Road, Beijing 100191, China ‡

State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese

Academy of Sciences, 555 Zuchongzhi Road, Shanghai 201203, China §

*

Agilent Technologies, 3 Wangjing North Road, Beijing 100102, China

Corresponding author. Tel.: +86 10 82801516. Fax: +86 10 82802024. Email address:

[email protected] (M. Ye).

1

ACS Paragon Plus Environment

Analytical Chemistry

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT To fully understand the chemical diversity of an herbal medicine is challenging. In this work, we describe a new approach to globally profile and discover novel compounds from an herbal extract using multiple neutral loss / precursor ion scanning combined with substructure recognition and statistical analysis (MNPSS). Turmeric (the rhizomes of Curcuma longa L.) was used as an example. This approach consists of three steps: i) multiple neutral loss / precursor ion scanning to obtain substructure information; ii) targeted identification of new compounds by extracted ion current and substructure recognition; and iii) untargeted identification using total ion current and multivariate statistical analysis to discover novel structures. Using this approach, 846 terpecurcumins (terpene-conjugated curcuminoids) were discovered from turmeric, including a number of potentially novel compounds. Furthermore, two unprecedented compounds (terpecurcumins X and Y) were purified, and their structures were identified by NMR spectroscopy. This study extended the application of mass spectrometry to global profiling of natural products in herbal medicines and could help chemists rapidly discover novel compounds from a complex matrix.

2

ACS Paragon Plus Environment

Page 2 of 28

Page 3 of 28

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

INTRODUCTION Plant-derived natural products are a consistent source for drug discoveries.1,2 However, to capture the full chemo-diversity of a medicinal plant is still a big challenge.3 This is partly due to the complexity of herbal extracts, in which chemical constituents are numerous in number and trace in amount, and thus hinders their systematic detection. Furthermore, the discovery of novel compounds is usually hampered by co-existence of known major compounds. To distinguish new structures from known ones has been pursued in recent years,4 but there is still a lack of effective analytical approaches.

Hyphenated analytical techniques are the major tools for novel natural products discovery.4,5 Among the most powerful techniques is mass spectrometry. The hyphenations of different types

of

mass

spectrometry

with

chromatography,

including

LC/IT-TOF-MS,6

LC/qTOF-MS,7-9 UHPLC-FTMS,10 LC/IM-MS11 and HPLC-MS-NMR12,13 allow rapid structural characterization of unknown compounds in complex herbal extracts. However, these techniques suffer from low efficiency because they rely on precise spectral analysis (MS, MS/MS, and MSn) for each of the numerous compounds, which is not easy for structurally complicated natural products. Moreover, the mass spectra of minor compounds may be missing in data acquisition, and thus could not be characterized. 6,7 In this study, we develop a new approach based on neutral loss (NL) and precursor ion (PRE) scanning, which could globally profile compounds containing similar substructures.14-16 The data are then processed by statistical analysis to efficiently discover novel structures. This approach is named MNPSS (multiple neutral loss / precursor ion scanning combined with substructure

3

ACS Paragon Plus Environment

Analytical Chemistry

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

recognition and statistical analysis), and is illustrated by analyzing terpecurcumins in turmeric as an example.

Terpecurcumins are a class of novel terpene-conjugated curcuminoids recently discovered from the herbal medicine turmeric (the rhizomes of Curcuma longa L.).17,18 They are hybrids of curcuminoid and sesquiterpene units connected through C-C or C-O bonds. In our previous phytochemical studies, 26 terpecurcumins have been isolated. Some of them show even more potent cytotoxic activities than curcumin, a popular anticancer agent. More intriguingly, they could induce apoptosis of MCF-7 human breast cancer cells through activation of the ROS-p38-H2AX axis, which is distinct from curcumin.19 Terpecurcumins could be promising cancer preventive or therapeutic agents.

In the present work, we use the MNPSS approach to globally profile and identify novel terpecurcumins in C. longa. The workflow is depicted in Figure 1. First, 12 NL and PRE scans resulted from characteristic MS/MS fragmentations of terpecurcumins were used to analyze turmeric extracts. The data were then processed by targeted and untargeted means with the aid of substructure recognition and multivariate statistical analysis, respectively. Finally, 846 terpecurcumins were discovered, including a number of potentially novel compounds. Among them, two compounds were purified by LC/MS-guided separation, and their structures were unambiguously identified by 1D and 2D NMR spectroscopy.

4

ACS Paragon Plus Environment

Page 4 of 28

Page 5 of 28

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Characteristic MS/MS fragments of terpecurcumins Input

Multiple Neutral loss / Precursor ion scan Output

12 NL/PRE chromatograms Targeted (datapoints)

Untargeted (datasets)

Extracted ion current

Total ion current

Intensive [M-H]- from full MS

Align Filter

Preliminary ID by NL/PRE profiles

Principal Component Analysis Grouped points

HRMS and MS/MS 27.5

28

Trace back and MS/MS

28.5

Rapid global profiling

Discover novel structures

Substructure recognition

Statistical analysis

Figure 1. Workflow of the MNPSS approach.

EXPERIMENTAL Reference standards, plant materials and sample pretreatment Reference compounds (terpecurcumins) R1–R18 were isolated by the authors from Curcuma longa L.17,18 Their structures are given in Figure S1. Dried rhizomes of C. longa L. were extracted by methanol and pretreated by a Toyopearl® column to afford 15 fractions (T01-T15). The details are described in Supporting Information and Figure S2.

Multiple LC-NL/PRE-MS analysis An Agilent 6495 triple quadrupole mass spectrometer was coupled with a 1290 UHPLC instrument via a Jet Stream ESI interface (Agilent Technologies, Waldbronn, Germany). Samples were separated on an Acquity UPLC HSS T3 column (1.8 µm, 2.1 mm i.d. × 150 mm) (Waters, MA, USA). The mobile phase consisted of acetonitrile (A) and water 5

ACS Paragon Plus Environment

Analytical Chemistry

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

containing 0.1% formic acid (B). A linear gradient elution program was used: 0–30 min, 50–90% A; 30–31 min, 90–100% A; 31–35 min, 100% A. The column temperature was 45 °C. The flow rate was 0.3 mL/min. The DAD detector scanned from 190 to 450 nm. The ESI source was operated in the negative ion mode, and the ions were transferred by iFunnel technology.20 High-purity nitrogen (N2) was used as the nebulizing (12 L/min, 200 °C, 30 psig) and sheath gas (10 L/min, 350 °C). The other parameters were as follows: capillary voltage, 3500 V; delta EMV, 50 V; fragment voltage, 380 V; parent ion range, m/z 450–650. Neutral loss scan and precursor ion scan parameters were set according to the characteristic MS/MS fragmentations of reference compounds. Fragments and collision energies are listed in Table 1. Data were analyzed by MassHunter software 06.00 (Agilent). To obtain high-resolution mass spectral data, a Q-Exactive hybrid quadrupole-orbitrap mass spectrometer (Thermo, USA) was used as described in Supporting Information.

Data mining Data mining consisted of two parts as shown in Figure 1. In the targeted identification, data were processed by extracted ion current (EIC): (1) Abundant pseudomolecular ions ([M-H]-, m/z) were selected from a full scan mass spectrum; (2) These ions were extracted from multiple NL/PRE scans to obtain EIC chromatograms; (3) Corresponding profiles for each peak to the 12 NL/PRE scans were recorded; (4) Substructures were recognized by comparing the NL/PRE profiles with those of reference standards; and (5) Structures were confirmed by HRMS and MS/MS analysis.

6

ACS Paragon Plus Environment

Page 6 of 28

Page 7 of 28

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

In the untargeted identification, data were treated by total ion current (TIC) as consecutive datasets and then analyzed by statistical means: (1) Parent ion intensities at each scan point were exported for the 12 NL/PRE chromatograms; (2) Signal intensities of different NL/PRE chromatograms were aligned according to their scan numbers; (3) Filters were added with regard to retention time and signal intensity (optional); (4) Principal component analysis (PCA) was performed to group data points with similar NL/PRE patterns so as to find those with different NL/PRE patterns; and (5) These special data points were traced back to pseudomolecular ions, and their MS/MS spectra were acquired for structural characterization. PCA was performed using SIMCA-P software (v11.5, Umetrics AB, Umeå, Sweden).

Isolation and structural identification of new compounds To confirm the new structures discovered by MNPSS, two compounds (terpecurcumins X and Y) were isolated by phytochemical means. Details are described in Supporting Information.

Terpecurcumin X (600#): orange amorphous powder; [α]25 D +8.4 (c 0.15, CH3OH); UV (MeCN) λmax = 374 nm; IR νmax = 3432, 2921, 2852, 1629, 1515, 1466, 1382, 1264, 1034, 579 cm−1; 1H and 13C NMR data, see Table S1; HRESIMS m/z 571.3064 [M+H]+, calcd for C36H43O6, 571.3060.

Terpecurcumin Y (317#): orange amorphous powder; [α]25 D +108 (c 0.15, CH3OH); UV (MeCN) λmax = 368 nm; IR νmax = 3432, 2920, 2852, 1617, 1594, 1515, 1392, 1267, 1123,

7

ACS Paragon Plus Environment

Analytical Chemistry

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1033, 811 cm−1; 1H and

13

C NMR data, see Table S1; HRESIMS m/z 573.3219 [M+H]+,

calcd for C36H45O6, 573.3216.

RESULTS AND DISCUSSION Experimental design The secondary metabolites of turmeric exhibit significant chemical diversity, mainly including curcuminoids (like curcumin, demethoxycurcumin, and bisdemethoxycurcumin) and sesquiterpenes (like α-zingiberene and α-/β-turmerone).21,22 Terpecurcumins, or terpene-conjugated curcuminoids, are a group of minor compounds reported for the first time from turmeric in 2011. Thus far, our group has isolated 26 terpecurcumins from turmeric,17,18 and many of them showed potent cytotoxic activities.19 LC/MS analysis indicated the presence of a large array of unknown terpecurcumins. However, because of low abundance, large number, and chromatographic co-elution, they could not be comprehensively detected and characterized. The present study aims to rapidly profile these terpecurcumins and effectively discover novel structures. The experimental procedure includes four parts: Sample pretreatment. Owing to the predominance of curcuminoids (> 90% of peak area), the extract was fractionated by Toyopearl® column chromatography to remove the curcuminoids (see Supporting information). Fractions T09–T11 containing terpecurcumins were subject to further analysis. Data acquisition. Based on the MS/MS fragmentation studies of 18 known terpecurcumins, 12 common neutral losses and product ions were used for multiple NL/PRE scan settings.

8

ACS Paragon Plus Environment

Page 8 of 28

Page 9 of 28

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Data mining. The NL/PRE data were processed by targeted (using extracted ion current) and untargeted (using total ion current) analysis for preliminary structural characterization (Figure 1). Structure confirmation. Two novel terpecurcumins were isolated from C. longa by LC/MS-guided separation, and their structures were fully identified by NMR spectroscopy.

MS/MS fragmentations of terpecurcumins and NL/PRE scan settings The MS/MS spectra of 18 reference compounds were analyzed using QqQ-MS and orbitrap-MS instruments. According to hybridization mode of the curcuminoid unit with the terpene unit, the 18 terpecurcumins could be classified into four types: connected through C-O-C bonds at C-8 (COC, R1−R7, R14, R15), through C-C bonds at C-9 (CC, R8, R13, R16), via Diels−Alder cycloaddition (DA, R9−12, R17), and via Michael addition and condensation (MC, R18). Representative structures are shown in Figure 2.

The COC type terpecurcumins yielded m/z 366 (R1, R2, R3, R6), 336 (R4, R7, R15), and 306 (R5, R14) fragments, corresponding to a curcumin, demethoxycurcumin and bisdemethoxycurcumin (CUR, DMC, BDMC) unit, respectively. This fragmentation was due to homolytic cleavage of the C-O bond.23 Because all the nine COC type terpecurcumins exhibited the same MS/MS fragmentation pattern (Figure S1), the characteristic fragment ions m/z 306, 336 and 366 were selected as precursor ions for PRE scanning of unknown COC type structures.

9

ACS Paragon Plus Environment

Analytical Chemistry

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 28

The DA type (R9−12, R17) and CC type (R8, R13, R16) could both cleave at the C-2/C-3 bond (Figure 2, Figure S1), and lose 220 Da and 218 Da, respectively. For the DA type, the rearrangement and dehydrogenation of C-3/C-4 led to the neutral loss of 220 Da. For the CC type, however, the double bond between C-3 and C-4 might suppress the rearrangement. Thus, 220 Da and 218 Da were set for NL scanning of unknown DA and CC type terpecurcumins, respectively.

x10

3

C20 H16O5•m/z 336.1009 (∆ 1.8 ppm) PRE 336

119

1.5 1.25 1 0.75

Ter−COC−Cur 451

336

191

571

0.5

149

0.25 0 50 x10

100

150

200

250

300

350

400

450

500

550

600

4

C12 H12O4 220.0735

149

1.2 1

421

0.8

Ter− DA −Cur 571

(∆ -0.5 ppm)

351

0.6 0.4 175

0.2

NL 220

213

R9

0 50

100

150

200

250

300

4

C10H10O 4 194.0581

2.5

(∆1.2 ppm)

x10

R3 R1=R2=OCH3 R4 R1=H, R2=OCH3 R5 R1=R2=H

350

400

450

353

m/z 351

600

Ter− CC −Cur

(∆1.4 ppm)

NL 194 NL 218

1.5 173

550

C12 H10O4 218.0582

2

1

500

217

0.5

377

571

421

R13

0 50 x10

100

150

200

250

300

350

400

450

500

550

600

4

Ter− MC −Cur

C12H18O 2 194.1309

243

3 2.5

m/z 353

(∆1.0 ppm)

2 1.5

137

1

NL 194

349 203

543

437

0.5

R18

0 50

100

150

200

250

300

350

400

450

500

550

600

m/z 349

Time (min)

Figure 2. (-)-ESI-MS/MS fragmentations of COC, DA, CC and MC type terpecurcumins. Ter, terpene unit; Cur, curcuminoid unit.

The MC type (R18) could lose 194 Da (C12H18O2, ∆=1.0 ppm) via RDA cleavage, as depicted in Figure 2. This fragmentation could also be observed in CC type due to cleavage of C1-C2′ bond and rearrangement of the –OH group, though the elemental composition was C10H10O4.

10

ACS Paragon Plus Environment

Page 11 of 28

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Thus, the NL scanning of 194 Da was set to discover MC or CC type terpecurcumins. Elemental compositions of the above neutral loss or precursor ions were confirmed by orbitrap-MS analysis, and the mass error values were less than 2.0 ppm or 2.0 mDa (Table 1).

In addition to major diagnostic fragments, terpecurcumins also produced common low-mass fragments (MW < 200) corresponding to the curcuminoid part. The cleavage between C-2′ and C-3′ yielded m/z 149 (CUR or DMC) and 119 (DMC or BDMC), respectively (Figure S3).24 These fragments were observed in all the terpecurcumins we analyzed. Furthermore, R9−12 and R17 could cleave at C1-C2′ to produce m/z 175, whereas R8, R13 and R16 cleave at C1-C2 followed by the loss of H2O to produce m/z 173 (Figure S1). The fragments m/z 175 and 173 were produced by the CUR unit. If the curcuminoid unit was replaced by DMC or BDMC, fragments m/z 145 and 143 would be produced. All these small fragments were common for terpecurcumins and showed certain variation among different types. Because of their high intensities and diagnostic structural information for the curcuminoid unit, these fragments were also included in the multiple NL/PRE scans.

The mass spectrometry parameters were then optimized. Collision energies were set at 20–35 eV (Figure S4 and Table 1). Scan ranges for parent ions were set at m/z 450–650, based on the molecular weights of terpecurcumins (Figure S5). Repeatability of the NL/PRE scans was suggested by eight successive injections of sample T11 (Figure S6). Specificity of the method was indicated by analyzing herbal samples that did not contain terpecurcumins (Figure

11

ACS Paragon Plus Environment

Analytical Chemistry

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 28

S7–S9). No terpecurcumins were detected from the negative samples following the MNPSS approach.

Targeted data mining: extracted ion current and substructure recognition Multiple NL/PRE scanning was applied to analyze terpecurcumin-containing samples. Compared to the full scan mode (qTOF-MS), the NL/PRE modes (QqQ-MS) provided remarkably high sensitivities (Figure S10). The chromatograms of samples T09, T10 and T11 are given in Figure 3 and Figure S11, and were processed by extracted ion chromatograms (EIC).

x10 4

x10 4

PRE 143

425#

5

x10 4

438# 444#

512# 522#

460#

x10 4

508# 487#

0

528#

PR E 366

501# 531# 478# 514#

581#

0

PRE 175

x10 5

532#

N L 218

460#

452#

403#

436#

1

508#

0

0

PRE 119

x10 4 5

487# 512# 522#

N L 220

532# 522#

579#

0 x10 5

572#

2

436#

x10 4 5

512# 482#

0

PRE 173

5

5

PR E 336

2

0

x10 4

579# 585#

0

PRE 145

1

x10 4

502# 494#

545#

0 x10 4

PR E 306

2

451#

550#

0

PRE 149

x10 4 2

508#

1

N L 194

460#

532# 550# 519#

0

0 6

8

10

12 14 16 18 20 Counts vs. Acquisition Time (min)

22

24

26

6

8

10

12 14 16 18 20 Counts vs. Acquisition Time (min)

22

24

26

Figure 3. NL/PRE chromatograms of sample T10.

The EIC of m/z 601 was analyzed in detail as an example. A total of 18 compounds could be observed in the m/z 601 chromatograms (Figure S12). Their responses to different NL/PRE scans were remarkably different, and were used for a preliminary structural characterization 12

ACS Paragon Plus Environment

Page 13 of 28

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

(Figure 4A). As shown in Table S2, compounds 406#, 412#, 501# and 514# could be detected by PRE 366, indicating they were COC type terpecurcumins containing a curcumin unit. Compounds 445#, 450#, 460# and 471# were detected by NL 218, and could belong to the CC type. Similarly, 534#, 601# and 614# were observed in the NL 220 chromatogram, and could belong to the DA type. Peaks 398# and 465# exhibited NL 194 signals, indicating they were of the MC/CC type.

To confirm the preliminary characterizations, high-resolution MS and MS/MS spectra were obtained. All the above 18 compounds had the molecular formula of C36H42O8, except for 614# and 619#, whose formula was C37H46O7 (∆ 435, 391, 367, 227, 217, 173 587 > 437, 366, 217, 175, 173, 149

20.89-20.98 22.54-22.63 21.71-21.85 29.56-29.65

572# 585# 579# 627#

557.2913 527.2809 527.2808 571.3069

0.72 1.14 0.95 0.70

PRE 336 PRE 306 PRE 306 NL 218

557 > 437, 336, 173, 143, 119 527 > 407, 306, 145, 143, 119 527 > 407, 306, 210, 145, 119 571 > 421, 353, 227, 217, 173, 149

G1ab 24.11-24.21

600#

569.2913

0.70

PRE 149

569 > 419, 349, 175, 149

No.

tR (min)

A Bb C D E Fb b

MS/MS d

G2

28.23-28.31

622#

571.3070

0.88

PRE 149

571 > 421, 351, 175, 149

a

24.68-24.77

606#

511.2856

0.39

PRE 119

511 > 391, 321, 183, 145, 119

I J ab

22.75-22.94 20.10-20.17

588# 317#

557.2913 571.307

0.72 0.88

PRE 336 NL 194

557 > 407, 336, 149 571 > 377, 362, 251, 149, 134

K1 K2 La

15.38-15.49 18.19-18.39 17.93-18.01

224# 275# 267#

557.2915 587.3019 589.3175

1.08 1.70 0.68

NL 218 NL 218 NL 220

557 > 339, 321, 217, 173, 119 587 > 369, 351, 243, 217, 173 589 > 369, 243, 134

Ma

18.82-18.97

289#

509.2702

0.98

PRE 306

509 > 306, 163, 145, 119

H

a

b

Potentially novel compounds. Identified by comparing with authentic reference standards: B=terpecurcumin B (R2); G2=terpecurcumin N (R11); F=terpecurcumin T (R13); G1= terpecurcumin X (600#); J= terpecurcumin Y (317#); and K2= terpecurcumin X (R8).c major signals for tracing back to pseudomolecular ions. d The most abundant ions are in bold face.

27

ACS Paragon Plus Environment

Analytical Chemistry

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 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 28

For TOC Only Multiple NL/PRE

PCA analysis

Unprecedented skeleton t[3]

New



20.1020.17 min

Known 5

6

7

8

t[2]

317# Terpecurcumin Y

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Counts vs. A cquisition Time (min)

28

ACS Paragon Plus Environment