Soft Corals Biodiversity in the Egyptian Red Sea: a Comparative MS and NMR Metabolomics Approach of Wild and Aquarium Grown Species † ‡ § ⊥ Mohamed A

Article

pubs.acs.org/jpr

Soft Corals Biodiversity in the Egyptian Red Sea: A Comparative MS and NMR Metabolomics Approach of Wild and Aquarium Grown Species † ‡ § ⊥ Mohamed A. Farag,*, Andrea Porzel, Montasser A. Al-Hammady, Mohamed-Elamir F. Hegazy, ¶ ⊥ ¶ ‡ Achim Meyer, Tarik A. Mohamed, Hildegard Westphal, and Ludger A. Wessjohann*, † Pharmacognosy Department, College of Pharmacy, Cairo University, Kasr el Aini st., P.B. 11562, Cairo 12613, Egypt ‡ Department of Bioorganic Chemistry, Leibniz Institute of Plant Biochemistry, Weinberg 3, D-06120 Halle (Saale), Germany § National Institute of Oceanography and Fisheries, Red Sea Branch, Hurghada 84511, Egypt ⊥ Phytochemistry Department, National Research Centre, 33 El Bohouth St, Dokki, Giza 12622, Egypt ¶ Leibniz Institute of Tropical Marine Ecology, Fahrenheit Str.6, D-28359 Bremen, Germany

*S Supporting Information

ABSTRACT: Marine life has developed unique metabolic and physiologic capabilities and advanced symbiotic relationships to survive in the varied and complex marine ecosystems. Herein, metabolite composition of the soft coral genus Sarcophyton was profiled with respect to its species and different habitats along the coastal Egyptian Red Sea via 1H NMR and ultra performance liquid chromatography-mass spectrometry (UPLC−MS) large-scale metabolomics analyses. The current study extends the application of comparative secondary metabolite profiling from plants to corals revealing for metabolite compositional differences among its species via a comparative MS and NMR approach. This was applied for the first time to investigate the metabolism of 16 Sarcophyton species in the context of their genetic diversity or growth habitat. Under optimized conditions, we were able to simultaneously identify 120 metabolites including 65 diterpenes, 8 sesquiterpenes, 18 sterols, and 15 oxylipids. Principal component analysis (PCA) and orthogonal projection to latent structures-discriminant analysis (OPLS) were used to define both similarities and differences among samples. For a compound based classification of coral species, UPLC−MS was found to be more effective than NMR. The main differentiations emanate from cembranoids and oxylipids. The specific metabolites that contribute to discrimination between soft corals of S. ehrenbergi from the three different growing habitats also belonged to cembrane type diterpenes, with aquarium S. ehrenbergi corals being less enriched in cembranoids compared to sea corals. PCA using either NMR or UPLC−MS data sets was found equally effective in predicting the species origin of unknown Sarcophyton. Cyclopropane containing sterols observed in abundance in corals may act as cellular membrane protectant against the action of coral toxins, that is, cembranoids. KEYWORDS: corals, cembranoids, Sarcophyton, metabolomic fingerprinting, nuclear magnetic resonance (NMR), ultra performance liquid chromatography−mass spectrometry (UPLC−MS), cyclopropyl sterols

■ INTRODUCTION communication in symbiotic relationships.4 Coral reef Seas cover over 70% of the earth. The total global biodiversity ecosystems support enormous biological diversity including is estimated to amount to some 500 × 106 species of structurally and functionally complex benthic communities. prokaryote and eukaryote organisms. Of these, marine The Red Sea is an epicenter for marine biodiversity with a high macrofauna comprise an estimated range of 0.5−30 × 106 percentage of endemic biota including the northern most tropical reefs with stony corals and soft corals. Of the 180 soft species with a broader range of taxonomic diversity than that ﬁ found in the traditional sources of natural products, the coral species identi ed worldwide, approximately 40% are 1 native to the Red Sea. Soft corals are marine invertebrates terrestrial macrofauna. Only a few thousand compounds have 5 been reported from marine origin, and hence seas and oceans possessing a vast range of terpenoid metabolites. These are believed to have an enormous potential of providers for new terpenes, mostly cembranoids, represent the main chemical bioactive metabolites.2 Marine natural products display an defense of corals against natural predators. Soft corals of the 3 genus Sarcophyton are particularly rich in cembrane terpenes.6,7 extraordinary chemical and pharmacological scope. This could 8 be attributed to the necessity of marine invertebrates to release Cembranoids contain a 14-membered macrocyclic skeleton, secondary metabolites as their own chemical defense tools to survive in the speciﬁc temperature, salinity, and pressure Received: January 2, 2016 conditions, and to resist their predators or to provide chemical Published: February 19, 2016

Figure 1. Photos of soft corals examined and location map of the collection area along the Red Sea, Egypt. Photograph courtesy of Dr. Montasser A. Al-Hammady and Dr. Mohamed A. Farag. Copyright 2016. biosynthesized by macrocyclization of geranylgeranyl- Nevertheless, 1D-NMR on its own cannot always provide diphosphate, and exhibit a wide range of biological properties unambiguous metabolite identifications and suffers from signal including antitumor, neuro-protective, antimicrobial, calcium- overlap. Application of NMR in the field of marine drugs antagonistic, and anti-inflammatory activity.9,10 includes analysis of spatial variation in soft corals from South In addition, marine organisms show intense symbiosis with China23 and profiling of stony reef corals from Hawaii,24 other organisms such as plants or bacteria. Chemical diversity is although in both studies, limited numbers of metabolites were a critical factor in an organism’s adaptation and fitness and a annotated, mostly targeting primary metabolites. In previous primary reason for the large number of natural products found studies, some of the present authors have successfully applied a in marine organisms including their symbionts.11 Metabolomics comparative metabolomic approach combining both NMR and strategies have recently emerged to help us gain a broader MS based technologies with multivariate data analyses for insight into the biochemical composition of living organisms.12 herbal drug analyses.18,25,26 With the recent developments in plant metabolomics The present work is focused on evaluating the capability of techniques,13,14 it is now possible to detect several hundreds 1D and 2D-NMR for metabolomic fingerprinting and profiling of metabolites simultaneously and to compare samples reliably of soft coral extracts, ideally without any preliminary chromato- for differences and similarities in a semiautomated and graphic step, in parallel to more conventional chromatography- untargeted manner. Metabolomics makes use mostly of coupled MS techniques. Such a comparative metabolomics hyphenated techniques, which rely on chromatographic approach is the first time to be applied in a marine-type separation of metabolites using either gas chromatography metabolomics project by exploring the diversity of soft coral (GC) or liquid chromatography (LC) coupled to mass secondary metabolism in the context of its genotype and spectrometry (MS) or, upcoming, nuclear magnetic resonance growing habitat. The effect of the growing habitat on secondary spectroscopy (NMR), to analyze complex mixtures of extracted metabolite accumulation in soft corals was assessed by metabolites.15 While a NMR metabolite profiling approach collecting soft corals from different locations in the Egyptian provides the complete and quantitative metabolite signature of part of the Red Sea and at different sea levels along with soft a complex extract, LC−MS much better resolves individual corals grown in an experimental aquarium facility under chemical components into separate peaks, enhancing the controlled conditions. Five species of the soft coral genus opportunity to uncover novel metabolites of low abundance.16 Sarcophyton, known to have many secondary metabolites The profiling of secondary metabolites in terrestrial plants including cembranoids, which are well represented along the recently has also been utilized, for example, for the quality Red Sea coast, have been comprehensively studied. To verify − controlofherbaldrugs.17 19 Plants,fungi,and(some) the effectiveness of the developed method, extracts from other microorganisms are much richer in (secondary) metabolites soft corals were analyzed for comparison, including Lobophyton than mammals or most other animals, and thus plant studies are and Sinularia sp. from the Red Sea coast. Owing to the more suitable as templates for studies on marine organisms. complexity of coral extracts, statistical multivariate analyses Although the use of metabolomics in plant analyses has been including principal component analysis (PCA) and partial least- extensive over the past 20 years,20 very little has been done in squares-discriminant analysis (OPLS-DA) were performed for terms of applying it to marine organisms. One of the recent samples classification. applications of LC−MS metabolomics was found to be quite effective for marine microbial strain prioritization to support ■ MATERIALS AND METHODS drug discovery of unique natural products.21 With an increasing interest in utilizing NMR for metabolic Soft Coral Material fingerprinting, it is now possible to record NMR spectra from The soft coral Sarcophyton of the family Alcyoniidae is the most crude extracts, providing its valuable metabolite signature.22 common genus along the Red Sea coast. Soft corals of five

1275 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article species of this genus, namely S. glaucum, S. acutum, S. each specimen, three biological replicates representing three ehrenbergi, S. convolutum, and S. regulare, as well as the two different coral bulbs were collected to assess biological variance. soft coral species Lobophyton pauciliforum and Sinularia Among the collected samples, five were from Northern polydactela, are studied here (Figure 1). Samples were collected Hurgada (S. glaucum, n =1;S. ehrenbergi, n =1;S. acutum, n from two widely geographically separated areas, offshore Al- =1;S. convolutum, n =1;S. regulare, n = 1), nine from Guna (20 km northen Hurghada) and offshore Makadi (55 Km Northern Safaga (S. glaucum, n =1;S. ehrenbergi, n =2;S. northern Safaga) at two locations along the Egyptian Red Sea acutum, n =1;S. convolutum, n =1;S. regulare, Sarcophyton spp., coast, located 55 km apart (Al-Guna and Makadi bay Figure 1). n =1,Sinularia polydactela, n =1;Lobophyton pauciliforum, n = By using SCUBA diving, samples were collected in water 1), and two (S. glaucum, n =1;S. ehrenbergi, n = 1) were depths of 2−5m(Table 1, Table S1). Fragments of 5−7cm2 cultured in the aquarium facility (see Table 1). Chemicals and Reagents Table 1. Origin of Soft Corals Sarcophyton, Lobophyton, and Sinularia Samples Used in Analysis Methanol-D4 (99.80% D), acetone-D6 (99.80% D), and hexamethyldisiloxane (HMDS) were purchased from Deutero accession species original source depth GmbH (Kastellaun, Germany). For calibration of chemical SG1 Sarcophyton glaucum Northern Hurghada (Al- reef flat shifts, HMDS was added to a final concentration of 0.94 mM. Guna) Acetonitrile and acetic acid (LC−MS grade) were obtained SA Sarcophyton acutum Northern Hurghada (Al- 3m from Baker (The Netherlands); Milli-Q water was used for LC Guna) analysis. Sarcophine standard was purchased from AG SR1 Sarcophyton regulare Northern Hurghada (Al- 3m fi Guna) Scienti c, San Diego, CA (St. Louis, MO, USA). Chromoband SC1 Sarcophyton convolutum Northern Hurghada (Al- 2m C18 (500 mg, 3 mL) cartridge was from Macherey and Nagel Guna) (Düren, Germany). All other chemicals and standards were SE1 Sarcophyton ehrenbergi Northern Hurghada (Al- reef flat available from Sigma-Aldrich (St. Louis, MO, USA). Guna) (2R,3E,7R,8S,11E)-7,8-Dihydroxy-1(15),3,11-cembratrien- SR2 Sarcophyton regulare Northern Safaga (Makadi 3m bay) 16,2-olide, 11(S)-hydroperoxylsarcoph-12(20)ene, 12-hydro- SE2 Sarcophyton ehrenbergi Northern Safaga (Makadi 2m peroxylsarcoph-10-ene, 7-hydroxy-8-methoxy-1(15),3,11-cem- bay) bratrien-16,2-olide, crassocolide K, horiolide, hydroxyl dihy- S Unidentified Sarcophyton Northern Safaga (Makadi 5m drobovolide, isosarcophinone, norcembrene, trochelioid A, 16- sp. bay) oxosarcophytonin E, 7b-acetoxy-8a-hydroxydeepoxysarcophine, SG2 Sarcophyton glaucum Northern Safaga (Makadi reef flat and ent-sarcophine were isolated from S. ehrenbergi and S. bay) glaucum. SC2 Sarcophyton convolutum Northern Safaga (Makadi reef flat bay) Soft Coral Extraction Procedure and Sample Preparation SR2 Sarcophyton regulare Northern Safaga (Makadi 3m for NMR and MS Analyses bay) LP Lobophyton pauciliforum Northern Safaga (Makadi 3m Approximately 100 mg of tissue from the umbrella was cut with bay) a clean scalpel and transferred to liquid nitrogen. The SP Sinularia polydactela Northern Safaga (Makadi 2m powdered freeze-dried soft coral tissues were ground with a bay) pestle in a mortar using liquid nitrogen. The powder was then SE3 Sarcophyton ehrenbergi Northern Safaga (Makadi 3m homogenized with 5.0 mL of 100% ethyl acetate containing 5 bay) μg/mL of umbelliferone (as internal standard for UPLC−MS, SE4 Sarcophyton ehrenbergi Aquarium, ZMT, Germany ultra performance LC−MS) using an ultrasonic bath for 20 SG3 Sarcophyton glaucum Aquarium, ZMT, Germany min. Extracts were then vortexed vigorously and centrifuged at 12 000g for 5 min to remove debris; 3 mL was aliquoted for were cut and collected from the disk of large colonies that were NMR evaporated under nitrogen, and the residue was tagged to recur it more than one time. On deck of the vessel, resuspended in 800 μL of acetone-D6 containing HMDS at a the samples were immediately transferred into boxes filled with concentration of 0.94 mM. After centrifugation (13 000g for 1 dry ice for transportation to the laboratories. We are well aware min), the supernatant was transferred to a 5 mm NMR tube. All that these handling procedures are not ideal but were the best 1H NMR spectra for multivariate data analysis were acquired fi adaptation available to eld conditions. Taxonomy and consecutively within a 48 h time-interval, with samples determination of the corals was done by coauthor Dr. Montaser resuspended in deuterated acetone immediately before data A. Al-Hammady (National Institute of Oceanography and acquisition. Repeated control experiments after 48 h showed no fi sheries, Red Sea Branch) on the basis of Fabricious and additional variation. For UPLC−MS analyses, 500 μL was 27 fi Aldersdale. Identi cation of Sarcophyton includes the isolation aliquoted and placed on a (500 mg) C18 cartridge, of sclerites (Figure S1). Specimens of S. glaucum and S. preconditioned with methanol and water. Samples were then ehrenbergi were additionally cultured in the aquarium facility of eluted using 6 mL of methanol; the eluent was evaporated the Leibniz Center for Tropical Marine Ecology, Bremen, under a nitrogen stream, and the obtained dry residue was ° Germany. These corals were kept at 26 C in a 2500 L resuspended in 1 mL of methanol. Three microliters was used recirculation system with 50 cm distance between coral nubbins for UPLC−MS analysis. For each specimen, three biological fl and a blue/white combination of two 39W uorescence light replicates were provided and extracted in parallel under the bulbs. same conditions. A total of 16 samples of various Sarcophyton and two other soft coral species belonging to the genera Sinularia and High Resolution UPLC-PDA-qTOF-MS Analysis Lobophyton collected from different sites along the Red Sea Chromatographic separations were performed on an Acquity coast were compared to the ones grown in aquarium tanks. For UPLC system (Waters) equipped with a HSS T3 column (100

1276 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article

× 1.0 mm2 particle size 1.8 μm; Waters) applying the following variance of metabolites, and OPLS-DA was performed to − elution binary gradient at a flow rate of 150 μL min 1: 0 to 1 obtain information on differences in the metabolite composi- min, isocratic 95% A (water/formic acid, 99.9/0.1 [v/v]), 5% B tion among samples. Distance to the model (DModX) test was (acetonitrile/formic acid, 99.9/0.1 [v/v]); 1−16 min, linear used to verify the presence of outliers and to evaluate whether a − − from 5 to 95% B; 16 18 min, isocratic 95% B; 18 20 min, submitted sample fell within the model applicability domain. μ isocratic 5% B. The injection volume was 3.1 L (full loop OPLS model was evaluated by the two parameters Q2 and R2X, injection). The C18 bonded phase used for the HSS T3 where R2X is used to quantify the goodness-of-fit, whereas Q2Y sorbents is compatible with 100% aqueous mobile phase and is employed to assess the predictability of the model. To rule provides ultralow MS bleed while promoting superior polar out the nonrandomness of separation between groups, an compound retention, which has been successfully used for fi iteration permutation test was performed. In all cases, the pro ling of similar plant extracts. Eluted compounds were 2 2 detected from m/z 100−1000 using a MicroTOF-Q hybrid values of Q and R resulting from the original model were quadrupole time-of-flight mass spectrometer (Bruker Dalton- found higher than the corresponding values from the permutation test confirming the model validity. R2 did not ics) equipped with an Apollo-II electrospray ion source in 2 2 positive ion modes using the following instrument settings: exceed Q by more than one unit and with no negative Q nebulizer gas, nitrogen, 1.6 bar; dry gas, nitrogen, 6 L min−1, values suggesting the validity of the model. The S loading plots 190 °C; capillary, −5500 V (+4000 V); end plate offset, − 500 of the OPLS-DA model was further used to identify the V; funnel 1 RF, 200 Vpp; funnel 2 RF, 200 Vpp; in-source CID variables responsible for sample differentiation on the scores energy, 0 V; hexapole RF, 100 Vpp; quadrupole ion energy, 5 plot. eV; collision gas, argon; collision energy, 10 eV; collision RF NMR Analyses 200/400 Vpp (timing 50/50); transfer time, 70 μs; prepulse storage, 5 μs; pulser frequency, 10 kHz; spectra rate, 3 Hz. All spectra were recorded on an Agilent VNMRS 600 NMR spectrometer operating at a proton NMR frequency of 599.83 UPLC−ESI−MSn Analysis MHz using a 5 mm inverse detection cryoprobe. 1H NMR − n Electrospray ionization (ESI) MS mass spectra were obtained spectra were recorded with the following parameters: digital from a LCQ Deca XP MAX system (ThermoElectron, San resolution, 0.367 Hz/point (32K complex data points); pulse Jose,́ USA) equipped with a ESI source (electrospray voltage, μ ° ° width (pw) = 3 s (45 ); relaxation delay = 23.7 s; acquisition 4.0 kV; sheath gas, nitrogen; capillary temperature, 275 C) in time = 2.7 s; number of transients = 160. Zero filling up to 128 positive ionization mode. The Ion Trap MS system is coupled K and an exponential window function with lb = 0.4 was used with the Exact Waters UPLC setup using the same elution prior to Fourier transformation. The 2D NMR spectra were gradient for the high resolution UPLC−ESI−TOF-MS analysis. The MSn spectra were recorded during the UPLC run using the recorded using standard CHEMPACK 5.1 pulse sequences following conditions: MS/MS analysis with a starting collision- (gDQCOSY, gHSQCAD, gHMBCAD) implemented in Varian induced dissociation energy of 20 eV and an isolation width of VNMRJ 4.0A spectrometer software. The HSQC experiment 1 +2 amu in a data dependent, positive ionization mode. was optimized for JCH = 146 Hz with DEPT-like editing and 13 MS Data Processing for Multivariate Analysis: PCA and C-decoupling during acquisition time. The HMBC experi- OPLS ment was optimized for a long-range coupling of 8 Hz; a two- step 1J filter was used (130−165 Hz). Relative quantification and comparison of Sarcophyton meta- CH bolic profiles after UPLC−MS was performed using XCMS NMR Data Processing and PCA Analysis data analysis software under R 2.9.2 environment, which can be The 1H NMR spectra were automatically Fourier transformed downloaded for free as an R package from the Metlin to ESP files using ACD NMR Manager Lab version 10.0 Metabolite Database. This software approach employs peak software (Toronto, Canada). The spectra were referenced to alignment, matching, and comparison to produce a peak list. internal HMDS at 0.062 ppm for 1H NMR and at 1.96 ppm for The resulting peak list was processed using the Microsoft Excel 13C NMR. Spectral intensities were reduced to integrated software (Microsoft, Redmond, WA), where the ion features regions, referred to as buckets, of equal width (0.04 ppm) were normalized to the total integrated area (1000) per sample within the region of δ 8.0 to −0.4 ppm. The regions between δ and imported into the R 2.9.2 software package for PCA. − δ − Absolute peak area values were autoscaled (the mean area value 2.00 2.15 ppm and 2.75 3.00 ppm corresponding to of each feature throughout all samples was subtracted from residual acetone and water signals, respectively, were removed each individual feature area and the result divided by the prior to multivariate analyses. PCA was performed with R standard deviation) prior to PCA. This provides similar weights package (2.9.2) using custom-written procedures after scaling for all the variables. PCA was then performed on the MS-scaled to HMDS signal and exclusion of solvent regions. data to visualize general clustering and trends, and outliers Sclerite Isolation from Coral Heads and Microscopic among the samples were identified based on the scores plot. Analysis Orthogonal projection to latent structures-discriminant analysis (OPLS-DA) was performed with the program SIMCA-P Sarcophyton has a mushroom-shaped polypary consisting of a Version 13.0 (Umetrics, Umea,̊ Sweden). OPLS-DA is a smooth and marginally folded disc, which projects beyond ff supervised pattern recognition technique that aims to find the clearly di erentiated base or stalk (Figure S1). Most species of 28 − maximum separation between a priori groups that was applied the genus Sarcophyton possess clubs some 0.20 0.30 mm in to discriminate, for example, between species. Biomarkers for length, rarely exceeding 0.60 mm in length, with the clubs species were subsequently identified by analyzing the S-plot, having low, rounded processes. The sclerite morphology was showing the relation of covariance (p) and correlation (pcor). examined with an optical microscope at ×200 magnification. The PCA was run for obtaining a general overview of the Sclerite identification followed protocol.29

1277 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article

Figure 2. Representative UPLC-qTOF-MS base peak chromatograms of Sarcophyton glaucum (SG1) showing sarcophine C23 and gauiacophene C69 as the major terpenoid peaks. The identities, Rt-values, and MS data of all peaks are listed in Table S2. ■ RESULTS AND DISCUSSION positive mode. Soft coral extracts were initially analyzed in both − The major goal of this study was to pave the way for a positive and negative ion ESI MS modes as changes in ESI polarity can often alter competitive ionization and suppression metabolomics-based secondary metabolite investigation of ff corals exemplified with Sarcophyton spp. (Figure 1)inan e ects revealing otherwise suppressed metabolite signals. untargeted, holistic manner in the context of its different Compared to the negative-ion ESI mode, positive-ion MS species, growing habitat, and to some extent water-depth so as spectra revealed better sensitivity and more observable peaks, to enable metabolite pattern-based taxonomy and to pave the with a total of 3341 mass signals extracted in positive mode way to identify species with constituents of medicinal potential using XCMS software versus 2689 mass signals found in − in Sarcophyton (Table 1). negative ionization mode. Representative UPLC MS total ion chromatograms of S. glaucum (SG1) are presented in Figure 2. Development of a Single Pot Extraction Method for UPLC−MS and NMR Analyses The identities, retention times, and observed molecular and fragment ions for individual components are presented in Table To allow for a comparative analysis of the metabolite data S2. Metabolite assignments were made by comparing high ff derived from these di erent technology platforms, a one pot resolution TOF MS data (accurate mass, isotopic distribution extraction method compatible for both NMR and MS and fragmentation pattern) of the compounds detected with metabolomics was developed. Several solvents were initially Sarcophyton compounds reported in the literature. More than tested including acetone, methanol, and ethyl acetate. Solvent 91 terpenoid and sterol peaks were resolved; cembranoids selection for sample preparation was evaluated with respect to accounted for the highest abundance (65 peaks) among species reproducibility, quality, and recovery of Sarcophyton cembra- 1 (Table S2, Figure S3), of which C20-diterpene sarcophine/ent- noids secondary metabolites as revealed by HNMR. sarcophine and the C -sesquiterpene guaiacophine (Figure S3) Compared with MS, 1H NMR can detect metabolites 15 ff were the major constituents in most coral extracts examined. universally, thus providing an unbiased picture of di erences Several oxylipid classes were identified, that is, cerebrosides, among extraction methods. Compared with methanol, acetone fatty acids, and their amides detected at late elution times of the and ethyl acetate show a better recovery rate of cembranoids chromatogram 600−700 s.31 The enhanced resolution afforded based on the NMR signals as is evident from sarcophine δ − by TOF MS measurement assisted to distinguish between resonances at 5.0 5.8 ppm (Figure S2). We are aware that several cembranoid isobaric peaks with same nominal masses. methanol in principal covers more hydrophilic compounds, The example compounds are tortuosene B (peak C15) and which are sometimes relevant in terrestrial plants, while ethyl hydroxydeepoxy sarcophine acetate (peak C90), differing in acetate shifts the focus slightly to more lipophilic metabolites mass by 0.038 amu, a difference that translates to formulas of abundant in corals. Previously, ethyl acetate was shown to + + C21H29O6 and C22H33O5 . In general, the mass spectra possess high efficiency for the extraction of coral diterpenes − 23 obtained by UPLC TOF MS yielded quasi-molecular ions, from Sarcophyton, Sinularia, and Dendronephthya species. for example, [M + H]+ and with a predicted formula showing a Considering our interest in profiling coral cembranoids,30 100% high number of double bond equivalents accounting for 2−3 ethyl acetate was chosen for preparing the bulk extract, further rings, three double bonds, and one carbonyl shared in most aliquoted for NMR and UPLC−MS analyses. The use of cembrane diterpene skeletons (Figure S3). The main observed deuterated solvent, routinely used for NMR sample extraction, product ions of cembranoids were the result of consecutive was ruled out in our case due to possible deuterium exchange eliminations of 18 or 28 units due to losses of H O and CO, with sample components affecting further UPLC−MS acquis- 2 respectively, on the basis of the chemical structure of a ition and metabolites assignment. diterpene (i.e., excision of ether-like oxygen and hydroxy − − fi UPLC ESI MS Peak Identi cation groups). Further fragmentation results in the opening of the γ- LC−MS so far has not been used for large scale metabolomics lactone ring and loss of the corresponding acyloxy group based analyses of coral. Coral cembranoids are relatively (C3H5O2, 71 amu) similar to sesquiterpene lactone fragmenta- nonpolar with commonly a hydroxy or peroxy group in the tions.32 For cembranoid peaks, that is, peak C7 that exhibits molecule and thus can be readily ionized in the ESI source hydroperoxy group, major fragmentations appeared as consec-

1278 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article

Figure 3. UPLC-qTOF-MS (m/z 100−600) PCA of Sarcophyton species based on group average cluster analysis of MS profiles (n = 3). The metabolome clusters are located at the distinct positions described by two vectors of principal component 1 (PC1 = 47%) and principal component 2 (PC2 = 20%). (A) Score plot of PC1 versus PC2 scores. (B) Loading plots for PC1 and PC2. It should be noted that ellipses do not denote statistical significance but are rather for better visibility of clusters as discussed. For sample codes, refer to Table 1. SG1 (○), SA (Δ), SR1 (+), SC1(×), SE1 (◇), SR2 (∇), SE2 (□ with × through it), S (*), SG2 (⊕), SC2 (diamond with + through it), SR3 (Star of David), SE3 (square with + through it), SE4 (□ with v through it), SG3 (⊗). utive losses of (−OH) and (−O) moieties attributed to the ingly, biscembranoids reported in Sarcophyton spp.36,37 from cleavage of the hydroperoxy group. Loss of 42 amu was evident Japanese or Kenyan coasts could not be identified in any of our in cembranoids containing an acetyl group and was used as a examined Sarcophyton species from the Red Sea, except for one + tool to screen for acylated diterpenes as in peaks C31, C51, diterpene dimer peak C73 (701.4603, C41H65O9 ). Whether C57, and C90. Notably, acylated derivatives were not identified such discrepancy is due to ecological differences or rather due in any sesquiterpene and only in one sterol (C68) suggestive to analysis limitations, with failure of biscembranoids to elute or for the presence of acylating enzymes with high substrate ionize under LC−MS conditions, has yet to be proved. In specificity toward diterpenes. Few terpenoids C3, C4, and C9 addition to the sesqui- and diterpenoids, several tetracyclic C- showed the loss of 44 amu corresponding to CO2 group and 28 sterols possessing the ergostan (C45, C46, C52, C99, and thus indicative of carboxylic acid groups. These polar C103) and cholestan skeleton (C103, C108, C112, C115, compounds eluted much earlier at a high solvent polarity rt C118) were identified in coral extracts, all of which contained 300−350 s. Compared with diterpenes, sesquiterpenes were cyclopropane rings in their side chain backbone,38 a sterol much less abundant in all examined coral extracts. A total of pattern common in zooxanthellae.39 Interestingly, gorgostene- eight peaks were detected of a guaiane skeleton, with diol,40 a sterol that bears an unusual C-23 methyl group and a guaiacophine (C69) and calamanene (C95) being the most cyclopropane ring in its side chain (C108) was identified along common forms. It cannot be excluded that more volatile simple with its mono hydroxy (C113, Figure S3) and desmethyl sesqui- and monoterpenes may be lost during solvent derivative (C118)inS. ehrenbergi samples but is likely to be evaporation. Soft corals are recognized for their high diterpene derived from its harbored zooxanthellae. Gorgosterol is known content, of which cembrane diterpenes are the characteristic to be produced by zooxanthellae, intracellular photosynthetic constituent of the genus.33 It should be noted that unless dinoflagellate symbionts living inside corals.41 An acetyl confirmed by standard and the proper spectroscopic technique, derivative of a gorgostane sterol was also tentatively assigned fi + neither the establishment of relative nor absolute con guration in peak C115 (535.3996 [M + H] ,C32H55O6) based on its MS of terpene assigned peaks can be made, for example, with data. several stereogenic centers being present in many of these The last part analyzed in the chromatographic run (600−750 compounds, six different stereoisomers A−F could be assigned s) revealed the presence of several fatty acids (FAs) as major to the sarcosolide peak C14,34 see Figure S3. Detailed MS peaks and expectedly with a low response in positive ionization fragmentation studies for diterpene stereoisomers analyzed mode. For the genus Sarcophyton, the presence of several using chiral columns might help to clarify for that issue but polyunsaturated fatty acids (PUFAs) is reported, with recent currently are more established for sesquiterpenes.35 Interest- evidence for FA transfer to its symbiont.42 The negative ion MS

1279 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article

Figure 4. (A) Score plot of PC1 vs PC2 scores. (B) Loading plots for PC1 and PC2 with contributing mass peaks and their assignments, with each − ff metabolite denoted by its mass/Rt (sec) pair. UPLC MS based score plots for di erentiation of regional S. ehrenbergi samples modeled separately (n = 3). The model explains 71% of total variation. The aquarium coral samples cluster more with reef flat corals on the left side (negative side of PC1). For sample codes, refer to Table 1. SE1 (○), SE2 (Δ), SE3 (+), SE4 (×). spectra of the fatty acids arachidonic acid (20:4 (n−6), C94), for possible heterogeneities among different genotypes and for tetracosapentaenoic acid (24:5 (n−6), C116), octadecatetrae- the chemotaxonomy of Sarcophyton in an untargeted approach noic acid (18:4 (n-3), C78), stearic acid (18:0, C120), and (Figure S4). PCA is an unsupervised clustering method palmitic acids (16:0, C106) were straightforward to interpret requiring no knowledge of the data sets and acts to reduce based on their high-resolution mass at 303.2307, 357.2747, the dimensionality of multivariate data while preserving most of 277.2168, 283.2597, and 255.232 with predicted molecular the variance within.45 PCA was applied to the TIC data from − − − − formulas of C20H32O2 ,C24H37O2 ,C18H29O2 ,C18H35O2 , the UPLC-qTOF-MS analysis retrieving a score plot in which − and C16H31O2 , respectively. Several nitrogen containing lipids, Lobophyton and Sinularia species appeared as most distant from that is, cerebroside (sarcoehrenoside B, C113) and stearamide other Sarcophyton species, suggesting that our profiling method (C117), were also detected in S. ehrenbergi extracts as evident can distinguish Sarcophyton from other coral species. However, from their even molecular ion peaks, albeit their origin (poly or it should be noted that some Sarcophyton sp. are chemically symbiont) cannot be unequivocally determined. Cerebroside more separated from each other than from Sinularia/ derivatives have been previously isolated from various marine Lobophyton. Distant clustering of Lobophyton specimen was invertebrates, such as sea stars, sponges, and soft corals, due to its enrichment in mass signals m/z 397.349 and including sarcoehrenoside B from S. ehrenbergi collected off the 401.3375 corresponding to ergosta-trien-ol (C99) and an Taiwan coast.43 Several different types of biological activities unknown compound (data not shown). In contrast, Sarcoph- have been ascribed to these compounds including antifungal, yton specimens were more enriched in diterpenes (data not antitumor, antiviral, cytotoxic, and immuno-modulatory proper- shown). Considering that our interest was to establish ties.44 This also is the first report for the presence of fatty acid Sarcophyton taxa relatedness based on its species, geographical amides in soft coral extracts. origin, and growing habitat, another multivariate analysis was fi performed excluding Sinularia and Lobophyton. Multivariate Data Analysis of Soft Corals Chemical Pro les The metabolome clusters were located at different points in via UPLC−MS the two-dimensional space prescribed by two vectors, principal Although different metabolite patterns could be observed by component 1 (PC1 = 47%) and principal component 2 (PC2 = visual inspection of UPLC−MS traces from different speci- 20%) (Figure 3A). Generally, all replicates from each sample mens, PCA was employed as a more holistic approach to test clustered together and were separated from other genotypes,

1280 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article confirming the repeatability of the methods used in this study. was established (Figure 4). The main principal components to The PC1/PC2 scores plot (Figure 3A) shows that two distinct differentiate between samples, that is, PC1 and PC2, accounted clusters are formed, corresponding to the five different species for 71% of the variance. The score plots show that the samples studied. It should be noted that within a species, separation could be differentiated without overlap. Most notably, corals based on geographical region or growing habitat could not be grown in aquarium clustered separately from all sea samples observed from PCA as, for example, values for S. ehrenbergi with negative score values along PC1. No clear difference could derived from different geographical sites or grown in tanks are be observed among sea corals from different habitats along all clustering together. Also, the numbers collected do not give PC1. Although along PC2, it should be noted that reef flat enough statistical safety. The same observation regarding corals (shallower water) appeared to be separable from corals geographical origin was observed in S. regulare specimens. collected at 2−3 m sea depth. In general and compared to The tight clustering in S. ehrenbergi and S. regulare individual aquarium specimens, sea corals contained higher levels of specimens suggests that they share similar biosynthetic cembranoids, that is, cembrapentaene (C96), 10-oxocembrene pathways involved in the production of cembranoids. Such a (sarcophytonin A, C80), and cembratetraen-16,2-olide (sar- clustering pattern was not observed in S. glaucum specimens, cophytonin B, C85) as denoted from their high resolution with individuals from different origins (Table 1) failing to masses of 271.2430, 287.2379, and 301.2171 amu, respectively. group together. S. glaucum chemical content was found to vary In contrast, samples collected from aquarium showed less to a large degree, and it was concluded there are at least nine diterpene content, being more enriched in oxylipid peaks with + chemotypes are existing within S. glaucum, which might account high resolution masses of 355.2244 (C23H31O3) and 359.2190 46 + for their dispersal, that is, the species is not suitable for (C23H35O3) . We propose that these masses are for unknown chemotaxonomic identification due to its chemical variability. oxylipids. Note the four extra hydrogen atoms in the predicted + + Most of the species were placed on the right side of the vertical formula C23H35O3 , compared with C23H31O3 common in line representing PC1 (negative score values), whereas S. lipids with a polyunsaturated acyl chain. Cembranoids play an ehrenbergi and S. acutum samples were placed on the left side ecological role in coral life and are likely to be produced at (positive score values). Examination of the loadings plot higher levels where corals are under more stressful marine life suggests that the variables refer to the MS signals of nitrogen conditions compared to corals grown in an aquarium tank. containing lipids and cembranoid diterpenes, that is, sarcophine Supervised OPLS Analyses of S. ehrenbergi, S. glaucum, and contributes to the discrimination of species (Figure 3B). In Lobophyton pauciliforum Analyzed via UPLC−MS detail, cembrapentaen-20,10-olide, hydroxy-cembrapentaen- 20,10-olide, cembratetraen-16,2-olide (sarcophytonin B),sar- In spite of the clear separation observed in PCA for coral samples based on genotype, supervised OPLS-DA was applied cophine (7,8-epoxy-1(15),3,11-cembratrien-16,2-olide) and ent- fi sarcophine were more enriched in S. regulare and S. convolutum, to build a classi cation model that could allow to identify + metabolite markers for coral species and to confirm whether S. whereas oxylipid peaks C35 and C77 of (327.2301, C22H31O2 ) − fi and (355.2244, C H O +)weremoreabundantinS. ehrenbergi and L. pauciliforum UPLC MS ngerprints were 23 31 3 sufficiently unique to be identified as markers for each species. ehrenbergi and S. acutum specimens along with cembrene C. fi Our results are in agreement with findings on qualitative OPLS-DA also has greater potential in the identi cation of ff ff markers by providing the most relevant variables for the di erences in cembranoids in di erent Sarcophyton chemotypes ff from Okinawa, Japan.46 It should be noted that MS signals for di erentiation between two sample groups. Two new models sesquiterpenes and sterols detected via MS (Table S2) did not were constructed with S. ehrenbergi (Figure S5A) and L. contribute for segregation in PCA loading plots along PC1 in pauciliforum (Figure S5C) each modeled separately, one at a the UPLCMS data set, suggestive that sesquiterpenes and time, against all other corals data present in one class group. The S. ehrenbergi model showed one orthogonal component sterols are present at comparable levels in all Sarcophyton 2 2 2 species or at least their extracts. with R = 0.83 and Q = 0.79 and for L. pauciliforum with R = 2 2 fi 2 Our attempt is the first one to derive Sarcophyton species 0.89 and Q = 0.85. R measures the goodness of t, while Q relatedness based on metabolite data, as molecular phylogenetic measures the predictive ability of the model. The S-plot results analyses alone so far have been insufficient to clearly identify for the S. ehrenbergi model (Figure S5B) show that this genus is particularly enriched in cembranoids, namely cembratetraene Sarcophyton species. The lack of understanding in both fi intraspecific variations of diagnostic morphological characters C110, sarcophytonin A C80, and unidenti ed lipid C76, within that genus in addition to a lack of solid taxonomic and though it should be noted that these constituents were found in ecological work on Sarcophyton pose problems to derive a clear all samples and therefore cannot serve as a chemical marker for phylogenetic based analysis.47 With regards to metabolites S. ehrenbergi. In contrast, the S-plot of Lobophyton (Figure S5D) based classification as in the current study, it should be noted showed sterol enrichment, namely ergosta-tetraen-ol, present that results are also not yet conclusive considering that the only at trace levels in other Sarcophyton species. The attempt to build an OPLS-DA model for S. regulare and Sinularia biogenetic source of metabolites whether from corals or its fi symbionts is not definitive. polydactela classi cation revealed no distinct chemical markers ff ff (data not shown). To also distinguish between S. glaucum and Di erentiation of S. ehrenbergi Grown in Di erent Regions S. ehrenbergi, appearing to cluster together in PCA analysis, and and Growing Habitats with no readily visible distinctive differences in their sclerite With the effective differentiation of coral samples of different features (Figure S1), supervised OPLS-DA was used to build a genetic origin, we examined whether multivariate statistical classification model. S. glaucum and S. ehrenbergi were modeled analysis can also differentiate the growing habitat within a single against each other using OPLS-DA with the derived score plot species. Therefore, we used multivariate data analysis to analyze showing a clear separation between both samples (data not S. ehrenbergi samples collected at different sea levels and growth shown). The OPLS score plot explained 78% of the total habitats (reef flat) separately, and a PCA classification model variance (R2 = 0.78) with the prediction goodness parameter

1281 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article

Figure 5. (A) 1H NMR spectrum of S. glaucum (SG1) extract showing characteristic signals for secondary metabolites in the most relevant shift range (δ-0.3−6.5 ppm. Expanded spectral region from (B) 3.25−6.25, (C) 0.7−2.9, and (D) −0.1−0.8 ppm with assigned peaks: N1 (guaiacophine), N2/3 (sarcophine/ent-sarcophine), N4 (7α,8β-dihydroxydeepoxysarcophine), N5 (7β,8β dihydroxydeepoxysarcophine), N6 (gorgosterol). The assignments were established using NMR spectra of standards. Signal numbers correspond to those listed in Table 2 for metabolite identiﬁcation using 1H NMR.

Q2 = 0.69. Compared to S. ehrenbergi, for S. glaucum, the S plot the symbiont or how the host (soft coral polyp) utilizes them to derived from OPLS contains more sarcophine/ent-sarcophine produce its secondary metabolites. (C23) but less cembrene C (C107) and 10-oxcembrene Visual Inspection of 1H NMR Spectra and Assignments of (C80). Correlation analysis was performed and reveals that Metabolites in Coral Crude Extract 2 cembrene C and sarcophine were negatively correlated with R The use of NMR has been previously reported in the study of = 0.65. Such negative correlation is rational considering that marine drug extracts from Hawaii and Japan, although targeting cembrene C is the building unit for the biosynthesis of mostly the more abundant primary metabolites.23,24 We chose sarcophine, and decrease in its pool is likely due to its S. glaucum (SG1) to positively demonstrate that NMR consumption acting as precursor for late step cembranoids spectroscopic characterization of secondary metabolites (Figure formation, that is, sarcophine. Interestingly, another mass of 5) investigated in this study is feasible. A comparative NMR 327.2301, C22H31O2 (C35) enriched in S. ehrenbergi, was spectrum of Sarcophyton species is depicted in Figure S6.In1D annotated as zooxanthellactone, a furanone found solely in and 2D NMR using 1H−1H COSY (correlation spectroscopy), 48 Symbiodinium (Zooxanthellae) species. Soft corals possess HMQC (heteronuclear multiple quantum coherence), and endosymbiotic dinoflagellates of the genus Symbiodinium, with HMBC (heteronuclear multiple bond coherence), peaks were the microalgae providing photosynthates to the host coral assigned whenever possible on the basis of spectra of reference tissue.49 Although there is piling evidence that photosynthetic standards. Chemical shifts of metabolites that were identified products are chemically incorporated into the tissue of the host, are listed in Table 2. The 1H NMR spectrum is characterized by little is known about the quantity of metabolites produced by two main regions: a low field region between 5.0 and 7.0 ppm

1282 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article

Table 2. Resonance Assignments with Chemical Shifts of Constituents Identiﬁed in 1H (at 600 MHz) and 1H/13C 2D-NMR Spectra of Sarcophyton Species (Acetone-D6)

metabolite assignment 1H (multiplicity) 13C HMBC N1 Guaiacophine H-1 2.79 45.9 H-4 2.61 40.8 H-6 6.26 br s 117.7 C-1 (45.9), C-4 (40.8), C-8 (205.8), C-11 (136.9) H-9a/b 2.39 dd (11.7, 7.5); 2.46 dd (11.7, 7.5) 51.7 C-1 (45.9), C-7 (138.5), C-8 (205.8), C-10 (36.7), C-15 (16.7) H-10 2.31 36.7 C-1 (45.9)

H3-12 1.81 s 21.1 C-7 (138.5), C-11 (136.9), C-13 (22.4)

H3-13 1.82 s 22.4 C-7 (138.5), C-11 (136.9), C-12 (21.1)

H3-14 1.13 d (6.9) 19.7 C-3 (34.3), C-4 (40.8), C-5 (151.8)

H3-15 0.89 d (6.0) 16.7 C-1 (45.9), C-9 (51.7), C-10 (36.7) N2/3 Sarcophine/ent-Sarcophine H-2 5.71 d (10.1) 79.2 C-3 (121.8) H-3 5.08 dq (10.1, 1.4) 121.8 C-5 (37.9), C-18 (16.1), H-7 2.63 61.5 H-10a/b 2.25; 1.95 23.9 H-11 5.23 125.6

H3-17 1.79 s 10.0 C-1 (163.5), C-15 (122.9), C-16 (174.7)

H3-18 1.94 s 16.1 C-3 (121.8), C-4 (144.6), C-5 (37.9)

H3-19 1.25 17.4 C-7 (61.5), C-8 (59.8), C-9 (39.8)

H3-20 1.64 s 16.0 C-11 (125.6), C-12 (136.4), C-13 (36.9) N4 7α,8β- H-2 5.71 d (10.1) 79.4 C-3 (122.1) Dihydroxydeepoxysarcophine H-3 5.08 dq (10.1, 1.4) 122.1 C-5 (38.2), C-18 (16.5) H-7 3.55 72.0 C-6 (26.1), C-9 (40.0)

H3-11 5.23 125.9 C-9 (40.0), C-20 (15.5)

H3-17 1.79 s 10.0 C-15 (122.5)

H3-18 1.94 s 16.5 C-3 (122.1), C-4 (144.9), C-5 (38.2)

H3-19 1.25 s 26.5 C-7 (72.0), C-8 (75.5), C-9 (40.0)

H3-20 1.63 s 16.0 C-11(125.9), C-12 (136.0), C-13 (37.0) N5 7β,8β- H-2 5.71 d (10.1) 79.4 C-3 (122.1) Dhydroxydeepoxysarcophine H-3 5.08 dq (10.1, 1.4) 122.1 C-5 (38.2), C-18 (16.5), H-7 3.44a 72.3 C-6 (28.0), C-8 (71.7), C-19 (26.8) H-11 5.23 125.9 C-9 (40.0), C-20 (16.0)

H3-17 1.79 s 10.0 C-15 (122.5)

H3-18 1.94 s 16.5 C-3 (122.1), C-4 (144.9), C-5 (38.2)

H3-19 1.25 s 26.8 C-7 (72.3), C-8 (71.7), C-9 (40.0)

H3-20 1.63 s 16.0 C-11 (125.9), C-12 (136.0), C-13 (37.0) N6 Gorgosteol H-3 3.48b 72.0 C-2 (32.0) H-6 5.32 m 121.5 H-22 0.22 m 32.9 b H3-26 0.90 21.5 C-24 (51.8), C-25 (32.2), C-27 (21.3)

H3-27 0.80 21.3 C-24 (51.8), C-25 (32.2), C-26 (21.5) b H3-29 0.88 14.3 C-22 (32.9), C-23 (26.0), C-30 (21.8) H-30a/b 0.48 dd (9.0, 4.2); −0.08 dd (5.9, 4.2) 21.8 aCorresponds to overlapping NMR signals in N5. bCorresponds to overlapping NMR signals in N6. fi with signals principally due to ole nic protons of cembranoids H3-13, 1.82 s; H3-14, 1.13 d (6.9), and H3-15, 0.89 d (6.0) and a high-field region between 3.2 and 0.8 ppm with high using 1H−1H COSY and 1H−13C correlations observed in the density signals due to terpene methyl, methylene, and methine HMBC spectra (Table 2).50 NMR spectra of all examined signals indicating that corals contain a large amount of Sarcophyton accessions show the presence of diterpenes bearing terpenoids. Of the 73 terpenoids identified using UPLC−MS, a sarcophine skeleton51,52 as major constituents as revealed δ only six terpenoids/sterol (Figure 6) with skeletons commonly from two doublet signals at H 5.71 d (10.1) and 5.08 dq (10.1, isolated from the genus Sarcophyton were detected using NMR. 1.4) for H-2 and H-3, respectively, in the five-membered α,β- Despite such current limitations of NMR, the patterns observed unsaturated-γ-lactone ring.53 The vicinal coupling constant of in the coral metabolite compositions are intriguing as revealed about10.1HzbetweenH-2andH-3suggestsacis from multivariate data analysis. The presence of guaiacophine configuration between the γ-lactone proton (H-2) and the (N1), a sesquiterpene isolated from Sarcophyton, was readily olefinic proton (H-3). The clear downfield singlet signals δ δ assigned based on its characteristic signal H-6 appearing at H integrate for three protons at H 1.79, 1.94, and 1.64 s indicate fi fi 6.26 br. s and the up eld chemical shift signals of H-9 2.39 dd the presence of ole nic methyl signals for H3-17, H3-18, and (11.7, 7.5)/2.46 dd, (11.7, 7.5); H-10, (2.31); H3-12, 1.81 s; H3-20, respectively. Since enantiomers in achiral solvents show

1283 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article

Figure 6. Structures of the major C15,C20-terpenoids and sterols detected in Sarcophyton extract using NMR and discussed in the manuscript. Note the carbon numbering system for each compound is used throughout the manuscript for NMR assignment and thus is based on analogy rather than IUPAC rules. the same NMR spectra, sarcophine/ent-sarcophine (N2/3) absolute configuration for all terpenoids and sterols identified cannot be distinguished. The 1H NMR spectrum of S. (Figure 6) cannot be determined by the methods used in this convolutum (SC1, Table 1) shows characteristic downfield study. It has been based on biogenetic or literature precedence δ signals of H-2, H-3, and H-11 at H 5.71 dq (10.1, 1.4), 5.08 dq or for relative stereochemistry, if no reference compound was (10.1, 1.4), and 5.23, respectively, in 7α,8β-dihydroxydeepoxy- available, and is mostly based on comparison with previously δ 9,54,55 sarcophine (N4). The proton signal at H 3.55 shows an HSQC reported NMR data. δ correlation with C 72.0 (C-7) as well as HMBC correlations 1 − δ Comparison of H NMR and UPLC MS Multivariate Data with C 26.1 (C-6) and 40.0 (C-9). Along with the HMBC Analyses correlations of H -19 (1.25 ppm) with C-7 (72.0 ppm), C-8 3 − (75.5 ppm), and C-9 (40.0 ppm), this shows the presence of We further compared the performance of PCA from UPLC 1 two hydroxy groups at C-7/C-8 instead of an epoxy group, MS and H NMR spectra of Sarcophyton extracts to better fi which led to the identification of (N4)as7α,8β- evaluate the classi cation potential of both technologies. PCA 1 dihydroxydeepoxysarcophine. The presence of the (N4) epimer was performed for all Sarcophyton samples within the H NMR δ − 7β,8β-dihydroxydeepoxysarcophine (N5) was concluded from region of 0.4 until 8.0 ppm with a total of 167 spectral bins δ 1 (variables) describing metabolite profiles. When compared to an oxygenated methine signal at H 3.44, correlated via JCH δ δ the PC plot obtained from UPLC−MS data (Figure 4A), it is with C 72.3 (C-7) and in the HMBC spectrum with C C-19 (26.8), C-6 (28.0), and C-8 (71.7). Because of too much signal apparent that the NMR results (Figure S7A) were in general overlap in N4/5, structural assignments are rather tentative agreement. In both score plots, S. ehrenbergi and S. acutum were compared to other structures (N1, N2/3, and N6), and plotted on the left side (negative score values). In contrast, and assignments are equivocal. Characteristic signals for a cyclo- contrary to UPLC−MS, S. convolutum specimens failed to propane bearing gorgostane type side chain in a sterol were cluster together and were spread along PC1 as for S. glaucum.It δ − observed at H 0.08 dd (5.9, 4.2) (H-30a), 0.48 dd (9.0, 4.2) should be noted that several Sarcophyton samples showed (H-30b), and 0.22 m (H-22) in addition to an olefinic signal at considerable overlap in NMR scoring plots that even of δ 1 −13 H 5.32 (H-6), commonly found in cholestan sterols. H C biological replicas within the same sample failed to group correlations observed in the HMBC spectrum assigned these together, as revealed from the PCA (Figure S7A). This finding signals to gorgosterol (N6), identified from UPLC−MS suggests that for such soft coral samples, UPLC−MS data sets analysis. The HMBC spectrum of (N6) shows key correlations provide more efficient sample classifications than NMR. Similar δ δ for both H3-26 at H 0.90 and H3-27 at H 0.80 cross- results were observed by our group in case of some plant correlating to their respective carbon signals in addition to sample classifications.26 The separation observed in PCA can be δ fi signals at C 32.2 (C-25) and 51.8 (C-24). Furthermore, a explained in terms of the identi ed compounds using the δ methyl group signal at H 0.88 (H3-29) showed HMBC loading plots for PC1 (Figure S7B) exposing the most δ correlations to C 32.9 (C-22), 26.0 (C-23), and 21.8 (C-30). discriminatory NMR signals. Two major groups stand out in The high-field 13C chemical shifts of C-22, C-23, and C-30, this plot. The first corresponds to the chemical shifts of δ δ δ caused by the ring strain of the three-membered ring, sarcophine (H3-17, 1.79; H3-18, 1.94; H3-19, 1.25; H3-20, confirmed the structure of N6. It should be noted that the δ 1.64) contributing positively to PC1. The second group is

1284 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article characterized by resonances appearing at δ-0.1, 0.1, and 5.31 for cembranoids could provide further understanding of our gorgosterol. The remaining unidentified compounds show metabolomics results in Sarcophyton. Indeed, the correlative proton signals at δ 0.84 and 1.36 ppm matching spectral analysis of differential metabolic profiling and gene expression locations of lipids. Examination of the 1H NMR loading plots profiling in planta has proven a powerful approach for the suggests that fatty acid signals and cembranoids contribute the identification of candidate genes and enzymes, particularly for 1 most in coral samples discrimination. NMR can detect all H- those of secondary metabolism.57 Differences among examined containing species in a sample, that is, including fatty acids Sarcophyton species mainly emanate from their content of posing NMR as a powerful technique in metabolomics if cembranoid diterpenes and lipids. Compared to Sarcophyton looking at more abundant metabolites. However, contents of species, the genus Lobophyton appeared more enriched such (primary) metabolites are usually dependent on growth (abundance) in sterols versus diterpenes and sesquiterpenes conditions and are often not indicative for the pharmaceutically in Sarcophyton. The cyclopropane containing sterols in coral relevant compounds. Considering that biological and technical specimens examined might be linked with coral adaptation to replicates in several S. ehrenbergi specimens failed to cluster the membranolytic activities of their own toxins, that is, tightly together as observed in UPLC−MS data set derived cembranoids. Sarcophine-diol has shown a cytotoxic effect on a score PCA plots (Figure 3A), no further attempt was made to melanoma cancer cell line mediated via inhibition of cell use multivariate data analysis for assessing growing habitat membrane permeability.58 Such phenomenon, which involves effect on Sarcophyton chemical composition from NMR derived the interdependent presence of two different types of secondary data sets. Aside from HMBC spectra providing structural “ assignment of terpenoids in extracts, careful examination of metabolites in an organism, that is, a biochemical coordina- tion” of the type “membranolytic toxinsunusual sterols”, are correlation pattern of coral extracts HMBC spectra (Figure S8) 59 clearly showed distinct fingerprint spectra in Sarcophyton evidenced for several marine sponges. For example, acutum (SA) (A)andSarcophyton ehrenbergi (SE1) (B) preincubation of carcinoma cells with a cyclopropane- differentfromthatofSarcophyton glaucum (SG1) (C), containing sterol from the marine sponge Rhizochalina ff Sarcophyton convolutum (SC1) (D), and Sarcophyton regulare incrustata decreased the cytotoxic e ects of rhizochalin, a 60 (SR1) and concurring PCA results derived from UPLC−MS toxic constituent from the same sponge. Whether such data set (Figure 4). Such observations need to be further scenario is also true for the corals here, that is, cytotoxic confirmed by subjecting HMBC spectra to multivariate data cembranoids and gorgosterol to counteract their toxic effect, is analyses. Our previous work on hop resin revealed better still speculative and has yet to be examined. The same workflow classification results derived from 2D-NMR compared to 1D- of sample preparation, measurement, and processing can be NMR,56 which has yet to be investigated in case of corals. easily transferred to other coral metabolome studies, pertaining coral ecological monitoring and enabling researchers to ■ CONCLUSIONS monitor shifts in coral metabolic composition in response to To the best of our knowledge, this study provides the first environmental conditions such as global warming or water comparative metabolomics approach to reveal for composi- pollution. We are aware that this first, method oriented tional differences in secondary metabolites among various soft metabolic fingerprinting study of selected coral species will coral species. NMR and UPLC−MS techniques coupled with require further in depth studies with many more accessions to multivariate data analyses were used and compared to obtain provide reliable hierarchical clusters. However, even in this first the experimental results, and interesting and meaningful study, significant information could be obtained, that is, that differences between the various species and detection methods metabolic profiles of aquarium specimen clearly differ from were identified. Our comparative metabolomics approach those of their wild brothers and that the profiles may be able to demonstrates that it can be a powerful tool for capturing the aid species determination. (secondary) metabolome status and peculiarities of soft corals and is likely to have value in monitoring corals and their ■ ASSOCIATED CONTENT genetically and environmentally caused metabolic differences. In this context, the presented study provides the first holistic *S Supporting Information ff approach to reveal for cembranoid compositional di erences The Supporting Information is available free of charge on the among Sarcophyton via metabolomics. Such strategy would ACS Publications website at DOI: 10.1021/acs.jproteo- prioritize coral species that ought to be subjected for further me.6b00002. detailed chemical analysis and speed up natural product research in corals, thereby helping to reduce the gap that has Physico-chemical parameters at diving sites and metabolites annotated in Sarcophyton species via UPLC-qTOF- opened to modern drug discovery demands. In particular, 1 corals from natural seawater environments contained signifi- MS; sclerite photos from soft corals; H NMR spectra of ff cantly higher levels of cembranoids than aquarium samples. It S. glaucum coral tissue extracted using di erent solvents; remains to be examined whether differential metabolite structure of the major terpenoids, sterols, and fatty acids accumulation patterns are due to ecological differences such detected in Sarcophyton extract via UPLC−MS and as organismic interactions, to physical parameters (ion strength, discussed in the manuscript; UPLC-qTOF-MS PCA temperature, light, etc.) or to precursor limitations such as the analysis of Sarcophyton, Sinularia, and L. pauciliforum; isoprene pool, or to genetic differences in expression, UPLC-qTOF-MS OPLS-DA score plot; 1HNMR regulation, and enzymatic activity as no clones were available. spectra of Sarcophyton species; 1H NMR (−0.8−8.0) Coupling these differential metabolite profile data with gene peak based PCA of different Sarcophyton samples with transcript levels may assist in probing involved genes and the methylene region of fatty acids excluded from biosynthetic pathway analyses. Probing enzymatic activity or analysis; partial HMBC spectrum of Sarcophyton species gene expression levels involved in the biosynthesis of (PDF)

1285 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article ■ AUTHOR INFORMATION (14) Sumner, L. W.; Lei, Z.; Nikolau, B. J.; Saito, K. Modern plant Corresponding Authors metabolomics: advanced natural product gene discoveries, improved technologies, and future prospects. Nat. Prod. Rep. 2015, 32 (2), 212− *E-mail: [email protected]:+20- 229. 1004142567. (15) Zhang, A.; Sun, H.; Wang, P.; Han, Y.; Wang, X. Modern *E-mail: [email protected]. analytical techniques in metabolomics analysis. Analyst 2012, 137 (2), 293−300. Notes − fi (16) Gika, H. G.; Wilson, I. D.; Theodoridis, G. A. LC MS-based The authors declare no competing nancial interest. holistic metabolic profiling. Problems, limitations, advantages, and future perspectives. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. ■ ACKNOWLEDGMENTS 2014, 966,1−6. M.A.F. thanks the Hanse-Wissenschaftskolleg (HWK), Ger- (17) Farag, M. A.; Mohsen, M.; Heinke, R.; Wessjohann, L. A. many, for financial support and the funding received by Science Metabolomic fingerprints of 21 date palm fruit varieties from Egypt using UPLC/PDA/ESI-qTOF-MS and GC-MS analyzed by chemo- and Technology Development fund STDF, Egypt (Grant No. − 12594). We also thank Dr. Christoph Böttcher for assistance metrics. Food Res. Int. 2014, 64, 218 226. (18) Farag, M. A.; Porzel, A.; Mahrous, E. A.; El-Massry, M. m. M.; with the UPLC−MS. We are grateful to Dr. Steffen Neumann ̈ − Wessjohann, L. A. Integrated comparative metabolite profiling via MS and Dr. Tilo Lubken for providing R scripts for UPLC MS and NMR techniques for Senna drug quality control analysis. Anal. data analysis. Bioanal. Chem. 2015, 407 (7), 1937−1949. (19)Mahrous,E.A.;Farag,M.A.TwodimensionalNMR ■ REFERENCES spectroscopic approaches for exploring plant metabolome: A review. (1) Danovaro, R.; Company, J. B.; Corinaldesi, C.; D’Onghia, G.; Journal of advanced research 2015, 6 (1), 3−15. Galil, B.; Gambi, C.; Gooday, A. J.; Lampadariou, N.; Luna, G. M.; (20) Yuliana, N. D.; Jahangir, M.; Verpoorte, R.; Choi, Y. H. Morigi, C.; Olu, K.; Polymenakou, P.; Ramirez-Llodra, E.; Sabbatini, Metabolomics for the rapid dereplication of bioactive compounds from A.; Sarda, F.; Sibuet, M.; Tselepides, A. Deep-sea biodiversity in the natural sources. Phytochem. Rev. 2013, 12 (2), 293−304. Mediterranean Sea: the known, the unknown, and the unknowable. (21) Hou, Y.; Braun, D. R.; Michel, C. R.; Klassen, J. L.; Adnani, N.; PLoS One 2010, 5 (8), e11832. Wyche, T. P.; Bugni, T. S. Microbial Strain Prioritization Using (2) Hamed, I.; Ozogul, F.; Ozogul, Y.; Regenstein, J. M. Marine Metabolomics Tools for the Discovery of Natural Products. Anal. Bioactive Compounds and Their Health Benefits: A Review. Compr. Chem. 2012, 84 (10), 4277−4283. Rev. Food Sci. Food Saf. 2015, 14 (4), 446−465. (22) Wolfender, J. L.; Marti, G.; Thomas, A.; Bertrand, S. Current (3) Sawadogo, W. R.; Boly, R.; Cerella, C.; Teiten, M. H.; Dicato, M.; approaches and challenges for the metabolite profiling of complex Diederich, M. A Survey of Marine Natural Compounds and Their natural extracts. Journal of chromatography. A 2015, 1382, 136−64. Derivatives with Anti-cancer Activity Reported in 2012. Molecules (23) He, Q.; Sun, R.; Liu, H.; Geng, Z.; Chen, D.; Li, Y.; Han, J.; Lin, 2015, 20 (4), 7097−142. W.; Du, S.; Deng, Z. NMR-Based Metabolomic Analysis of Spatial (4) Abraham, I.; El Sayed, K.; Chen, Z.-S.; Guo, H. Current Status on Variation in Soft Corals. Mar. Drugs 2014, 12 (4), 1876−1890. Marine Products with Reversal Effect on Cancer Multidrug Resistance. (24) Sogin, E. M.; Anderson, P.; Williams, P.; Chen, C.-S.; Gates, R. − Mar. Drugs 2012, 10 (10), 2312 2321. D. Application of H-1-NMR Metabolomic Profiling for Reef-Building (5) Hegazy, M. E.; Mohamed, T. A.; Alhammady, M. A.; Shaheen, A. Corals. PLoS One 2014, 9 (10), e111274. M.; Reda, E. H.; Elshamy, A. I.; Aziz, M.; Pare, P. W. Molecular (25) Farag, M. A.; Porzel, A.; Schmidt, J.; Wessjohann, L. A. architecture and biomedical leads of terpenes from red sea marine − Metabolite profiling and fingerprinting of commercial cultivars of invertebrates. Mar. Drugs 2015, 13 (5), 3154 81. Humulus lupulus L. (hops): a comparison of MS and NMR methods in (6) Blunt, J. W.; Copp, B. R.; Keyzers, R. A.; Munro, M. H. G.; metabolomics. Metabolomics 2012, 8 (3), 492−507. Prinsep, M. R. Marine natural products. Nat. Prod. Rep. 2012, 29 (2), − (26) Farag, M. A.; Porzel, A.; Wessjohann, L. A. Comparative 144 222. metabolite profiling and fingerprinting of medicinal licorice roots using (7) Blunt, J. W.; Copp, B. R.; Keyzers, R. A.; Munro, M. H. G.; a multiplex approach of GC-MS, LC−MS and 1D NMR techniques. Prinsep, M. R. Marine natural products. Nat. Prod. Rep. 2014, 31 (2), Phytochemistry 2012, 76,60−72. 160−258. (27) Fabricius, K.; Alderslade, P. Soft Corals and Sea Fans: A (8) Wessjohann, L. A.; Ruijter, E.; Garcia-Rivera, D.; Brandt, W. Comprehensive Guide to the Tropical Shallow-Water Genera of the What can a chemist learn from nature’smacrocycles?–abrief, Central-West Pacific, the Indian Ocean and the Red Sea; Australian conceptual view. Mol. Diversity 2005, 9 (1−3), 171−86. (9) Hegazy, M.-E. F.; Eldeen, A. M. G.; Shahat, A. A.; Abdel-Latif, F. Institute of Marine Science: Townsville, Australia, 2001. F.; Mohamed, T. A.; Whittlesey, B. R.; Pare, P. W. Bioactive (28) Bylesjo, M.; Rantalainen, M.; Cloarec, O.; Nicholson, J. K.; Hydroperoxyl Cembranoids from the Red Sea Soft Coral Sarcophyton Holmes, E.; Trygg, J. OPLS discriminant analysis: combining the glaucum. Mar. Drugs 2012, 10 (1), 209−222. strengths of PLS-DA and SIMCA classification. J. Chemom. 2006, 20 (8−10), 341−351. (10) El Sayed, K. A.; Hamann, M. T.; Waddling, C. A.; Jensen, C.; fi Lee, S. K.; Dunstan, C. A.; Pezzuto, J. M. Structurally novel (29) Janes, M. P. Laboratory Methods for the Identi cation of Soft bioconversion products of the marine natural product sarcophine Corals (Octocorallia: Alcyonacea); 1st International Symposium of effectively inhibit JB6 cell transformation. J. Org. Chem. 1998, 63 (21), Coral Husbandry in Public Aquaria: Arnhem, The Netherlands, 2008. 7449−7455. (30) Yang, B.; Zhou, X.-F.; Lin, X.-P.; Liu, J.; Peng, Y.; Yang, X.-W.; (11) Grosso, C.; Valentao, P.; Ferreres, F.; Andrade, P. B. Alternative Liu, Y. Cembrane Diterpenes Chemistry and Biological Properties. and efficient extraction methods for marine-derived compounds. Mar. Curr. Org. Chem. 2012, 16 (12), 1512−1539. Drugs 2015, 13 (5), 3182−230. (31) Kobayashi, J.; Ishibashi, M.; Nakamura, H.; Hirata, Y.; Yamasu, (12) Farag, M. A. Comparative mass spectrometry & nuclear T.; Sasaki, T.; Ohizumi, Y. Symbioramide, a novel Ca2+-ATPase magnetic resonance metabolomic approaches for nutraceuticals quality activator from the cultured dinoflagellate Symbiodinium sp. Experientia control analysis: a brief review. Recent Pat. Biotechnol. 2014, 8 (1), 17− 1988, 44 (9), 800−2. 24. (32) Crotti, A. E.; Lopes, J. L.; Lopes, N. P. Triple quadrupole (13) Kim, H. K.; Choi, Y. H.; Verpoorte, R. NMR-based plant tandem mass spectrometry of sesquiterpene lactones: a study of metabolomics: where do we stand, where do we go? Trends Biotechnol. goyazensolide and its congeners. J. Mass Spectrom. 2005, 40 (8), 2011, 29 (6), 267−275. 1030−4.

1286 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287 Journal of Proteome Research Article

(33) Hanson, J. R. Diterpenoids. Nat. Prod. Rep. 2009, 26 (9), 1156− hydroperoxyl cembranoids from the Red Sea soft coral Sarcophyton 1171. glaucum. Mar. Drugs 2012, 10 (12), 209−22. (34) Cheng, Y.-B.; Shen, Y.-C.; Kuo, Y.-H.; Khalil, A. T. Cembrane (52) Hegazy, M.-E. F.; Mohamed, T. A.; Abdel-Latif, F. F.; Alsaid, M. diterpenoids from the Taiwanese soft coral Sarcophyton stolidotum. J. S.; Shahat, A. A.; Pare, P. W. Trochelioid A and B, new cembranoid Nat. Prod. 2008, 71 (7), 1141−1145. diterpenes from the Red Sea soft coral Sarcophyton trocheliophorum. (35) Crotti, A. E. M.; Lopes, J. L. C.; Lopes, N. P. Triple quadrupole Phytochem. Lett. 2013, 6 (3), 383−386. tandem mass spectrometry of sesquiterpene lactones: a study of (53) Hegazy, M.-E.; El-Beih, A. A.; Moustafa, A. Y.; Hamdy, A. A.; goyazensolide and its congeners. J. Mass Spectrom. 2005, 40 (8), Alhammady, M. A.; Selim, R. M.; Abdel-Rehim, M.; Pare,P.W.́ 1030−1034. Cytotoxic cembranoids from the Red Sea soft coral Sarcophyton (36) Iwagawa, T.; Hashimoto, K.; Yokogawa, Y.; Okamura, H.; glaucum. Nat. Prod. Commun. 2011, 6 (12), 1809−1812. Nakatani, M.; Doe, M.; Morimoto, Y.; Takemura, K. Cytotoxic (54) Finer, J.; Clardy, J.; Kobayashi, A.; Alam, M.; Shimizu, Y. Biscembranes from the Soft Coral Sarcophyton glaucum. J. Nat. Prod. Identity of stereochemistry of dinosterol and gorgosterol side-chain. J. 2009, 72 (5), 946−949. Org. Chem. 1978, 43 (10), 1990−1992. (37) Feller, M.; Rudi, A.; Berer, N.; Goldberg, I.; Stein, Z.; Benayahu, (55) Yao, L.-G.; Liu, H.-L.; Guo, Y.-W.; Mollo, E. New Cembranoids Y.; Schleyer, M.; Kashman, Y. Isoprenoids of the soft coral from the Hainan Soft Coral Sarcophyton glaucum. Helv. Chim. Acta Sarcophyton glaucum: Nyalolide, a new biscembranoid, and other 2009, 92 (6), 1085−1091. terpenoids. J. Nat. Prod. 2004, 67 (8), 1303−1308. (56) Farag, M.; Mahrous, E.; Lübken, T.; Porzel, A.; Wessjohann, L. (38) Wessjohann, L. A.; Brandt, W.; Thiemann, T. Biosynthesis and Classification of commercial cultivars of Humulus lupulus L. (hop) by metabolism of cyclopropane rings in natural compounds. Chem. Rev. chemometric pixel analysis of two dimensional nuclear magnetic 2003, 103 (4), 1625−48. resonance spectra. Metabolomics 2014, 10 (1), 21−32. (39) Kobayashi, M.; Mitsuhashi, H. Marine sterols. XII. Glaucasterol, (57) Farag, M. A.; Deavours, B. E.; de Fatima, A.; Naoumkina, M.; a novel C27 sterol with a unique side chain, from the soft coral Dixon, R. A.; Sumner, L. W. Integrated metabolite and transcript Sarcophyton glaucum. Steroids 1982, 40 (6), 665−72. profiling identify a biosynthetic mechanism for hispidol in Medicago (40) Anjaneyulu, A. S. R.; Krishna Murthy, M. V. R.; Gowri, P. M.; truncatula cell cultures. Plant Physiol. 2009, 151 (3), 1096−113. Venugopal, M.; Laatsch, H. A rare prostaglandin from the soft coral (58) Szymanski, P. T.; Muley, P.; Ahmed, S. A.; Khalifa, S.; Fahmy, Sarcophyton crassocaule of the Indian Ocean. J. Nat. Prod. 2000, 63 H. Sarcophine-Diol Inhibits Expression of COX-2, Inhibits Activity of (10), 1425−1426. cPLA(2), Enhances Degradation of PLA(2) and PLC(gamma)1 and (41) Withers, N. W.; Kokke, W.; Fenical, W.; Djerassi, C. Sterol Inhibits Cell Membrane Permeability in Mouse Melanoma B16F10 − patterns of cultured zooxanthellae isolated from marine-invertebrates - Cells. Mar. Drugs 2012, 10 (10), 2166 2180. synthesis of gorgosterol and 23-desmethylgorgosterol by aposymbiotic (59) Santalova, E. A.; Makarieva, T. N.; Gorshkova, I. A.; Dmitrenok, algae. Proc. Natl. Acad. Sci. U. S. A. 1982, 79 (12), 3764−3768. A. S.; Krasokhin, V. B.; Stonik, V. A. Sterols from six marine sponges. − (42) Imbs, A. B.; Yakovleva, I. M.; Dautova, T. N.; Bui, L. H.; Jones, Biochem. Syst. Ecol. 2004, 32 (2), 153 167. P. Diversity of fatty acid composition of symbiotic dinoflagellates in (60) Makarieva, T. N.; Stonik, V. A.; Ponomarenko, L. P.; Aminin, D. corals: evidence for the transfer of host PUFAs to the symbionts. L. Unusual marine sterols may protect cellular membranes against action − Phytochemistry 2014, 101,76−82. of some marine toxins 1998,37 40. (43) Cheng, S. Y.; Wen, Z. H.; Chiou, S. F.; Tsai, C. W.; Wang, S. K.; Hsu, C. H.; Dai, C. F.; Chiang, M. Y.; Wang, W. H.; Duh, C. Y. Ceramide and cerebrosides from the octocoral Sarcophyton ehrenbergi. J. Nat. Prod. 2009, 72 (3), 465−8. (44) Kolter, T.; Doering, T.; Wilkening, G.; Werth, N.; Sandhoff, K. Recent advances in the biochemistry of glycosphingolipid metabolism. Biochem. Soc. Trans. 1999, 27 (4), 409−415. (45) Goodacre, R.; Shann, B.; Gilbert, R. J.; Timmins, E. M.; McGovern, A. C.; Alsberg, B. K.; Kell, D. B.; Logan, N. A. Detection of the dipicolinic acid biomarker in Bacillus spores using Curie-point pyrolysis mass spectrometry and fourier transform infrared spectroscopy. Anal. Chem. 2000, 72 (1), 119−127. (46) Aratake, S.; Tomura, T.; Saitoh, S.; Yokokura, R.; Kawanishi, Y.; Shinjo, R.; Reimer, J. D.; Tanaka, J.; Maekawa, H. Soft Coral Sarcophyton (Cnidaria: Anthozoa: Octocorallia) Species Diversity and Chemotypes. PLoS One 2012, 7 (1), e30410. (47) McFadden, C. S.; Alderslade, P.; Van Ofwegen, L. P.; Johnsen, H.; Rusmevichientong, A. Phylogenetic relationships within the tropical soft coral genera Sarcophyton and Lobophytum (Anthozoa, Octocorallia). Invertebrate Biology 2006, 125 (4), 288−305. (48) Onodera, K.; Fukatsu, T.; Kawai, N.; Yoshioka, Y.; Okamoto, T.; Nakamura, H.; Ojika, M. Zooxanthellactone, a novel gamma-lactone- type oxylipine from dinoflagellates of Symbiodinium sp.: structure, distribution, and biological activity. Biosci., Biotechnol., Biochem. 2004, 68 (4), 848−52. (49) Sammarco, P. W.; Strychar, K. B. Responses to high seawater temperatures in zooxanthellate octocorals. PLoS One 2013, 8 (2), e54989. (50) Feller, M.; Rudi, A.; Berer, N.; Goldberg, I.; Stein, Z.; Benayahu, Y.; Schleyer, M.; Kashman, Y. Isoprenoids of the soft coral Sarcophyton glaucum: Nyalolide, a new biscembranoid, and other terpenoids. J. Nat. Prod. 2004, 67 (8), 1303−8. (51) Hegazy, M. E.; Eldeen, A. M. G.; Shahat, A. A.; Abdel-Latif, F. F.; Mohamed, T. A.; Whittlesey, B. R.; Pare, P. W. Bioactive

1287 DOI: 10.1021/acs.jproteome.6b00002 J. Proteome Res. 2016, 15, 1274−1287