Metabolite Profiling of Triterpene Glycosides of the Far Eastern Sea Cucumber Eupentacta fraudatrix and Their Distribution in Various Body Components Using LC-ESI QTOF-MS

The Far Eastern sea cucumber Eupentacta fraudatrix is an inhabitant of shallow waters of the south part of the Sea of Japan. This animal is an interesting and rich source of triterpene glycosides with unique chemical structures and various biological activities. The objective of this study was to investigate composition and distribution in various body components of triterpene glycosides of the sea cucumber E. fraudatrix. We applied LC-ESI MS (liquid chromatography–electrospray mass spectrometry) of whole body extract and extracts of various body components for metabolic profiling and structure elucidation of triterpene glycosides from the E. fraudatrix. Totally, 54 compounds, including 26 sulfated, 18 non-sulfated and 10 disulfated glycosides were detected and described. Triterpene glycosides from the body walls, gonads, aquapharyngeal bulbs, guts and respiratory trees were extracted separately and the distributions of the detected compounds in various body components were analyzed. Series of new glycosides with unusual structural features were described in E. fraudatrix, which allow clarifying the biosynthesis of these compounds. Comparison of the triterpene glycosides contents from the five different body components revealed that the profiles of triterpene glycosides were qualitatively similar, and only some quantitative variabilities for minor compounds were observed.


Introduction
Sea cucumbers (Class Holothuroidea, Phylum Echinodermata) are widespread slow-moving marine animals. Metabolome of these animals is characterized by the high content of triterpene glycosides of a great structural diversity. Triterpene glycosides of sea cucumbers have unique chemical structures, significantly differing from those of terrestrial plants. These compounds possess a variety of biological and pharmacological effects including cytotoxic [1,2], antifungal [3,4], bactericidal, hemolytic, antiviral and antiparasitic properties [5]. Some glycosides are capable to induce apoptosis, inhibit the E. fraudatrix are characterized by the presence of 3-O-MeXyl residue as terminal unit in carbohydrate chain, which is considered a chemotaxonomic marker of the genus Eupentacta. Herein, we describe the application of LC-ESI MS for metabolic profiling, evaluation of the structural variability, structure elucidation and further refinement of known biosynthetic patterns of triterpene glycosides from the sea cucumber E. fraudatrix. In addition, triterpene glycosides from the body walls, gonads, aquapharyngeal bulbs, guts and respiratory trees were extracted separately and distribution of the detected compounds in various body components were analyzed.

Profiling and Structural Identification of the Triterpene Glycosides from E. fraudatrix
Profiling of triterpene glycosides from the whole body extract of the sea cucumber E. fraudatrix using LC-MS approach allowed numerous new as well as previously isolated triterpene glycosides to be characterized. The HPLC profile revealed at least 54 compounds, including 26 sulfated, 18 non-sulfated and 10 disulfated glycosides (Figures 1 and 2; Table 1; Figure S1; the numbers of the compounds correspond to the peak numbers on (−)LC-MS chromatogram). we describe the application of LC-ESI MS for metabolic profiling, evaluation of the structural variability, structure elucidation and further refinement of known biosynthetic patterns of triterpene glycosides from the sea cucumber E. fraudatrix. In addition, triterpene glycosides from the body walls, gonads, aquapharyngeal bulbs, guts and respiratory trees were extracted separately and distribution of the detected compounds in various body components were analyzed.

Profiling and Structural Identification of the Triterpene Glycosides from E. fraudatrix
Profiling of triterpene glycosides from the whole body extract of the sea cucumber E. fraudatrix using LC-MS approach allowed numerous new as well as previously isolated triterpene glycosides to be characterized. The HPLC profile revealed at least 54 compounds, including 26 sulfated, 18 nonsulfated and 10 disulfated glycosides (Figures 1 and 2; Table 1; Figure S1; the numbers of the compounds correspond to the peak numbers on (−)LC-MS chromatogram). In positive ion mode, the majority of triterpene glycosides were detected within m/z range from 1000 to 1400 a.m.u. as [M + Na] + ions. However, stable peaks of [M + Na] + ions for disulfated glycosides were not observed in these conditions. In negative ion mode, sulfated and disulfated glycosides were detected as [M − Na] − and [M − 2Na] 2− peaks, respectively, whereas non-sulfated compounds were detected as [M − H] − peaks.
The assignments of these compounds in all samples analyzed were based on the data of high resolution LC-MS and LC-MS/MS performed in both negative and positive ion modes. Elemental composition, determined on the base of high resolution data (mass accuracy tolerance < 2 ppm), fragmentation patterns of MS and MS/MS spectra, as well as chromatographic behavior of the corresponding compounds, allowed their structures to be proposed. It is known that triterpene glycosides are characterized by large structural variability. Generally, mass spectrometry does not permit to determine configuration of the unknown compounds. Besides, different epimeric monosaccharides as well as types of bonds between sugars cannot be strictly distinguished only by MS. However, the combination of the obtained data and biosynthetic considerations allows tentative structural assignments for the detected compounds. In positive ion mode, the majority of triterpene glycosides were detected within m/z range from 1000 to 1400 a.m.u. as [M + Na] + ions. However, stable peaks of [M + Na] + ions for disulfated glycosides were not observed in these conditions. In negative ion mode, sulfated and disulfated glycosides were detected as [M − Na] − and [M − 2Na] 2− peaks, respectively, whereas non-sulfated compounds were detected as [M − H] − peaks.
The assignments of these compounds in all samples analyzed were based on the data of high resolution LC-MS and LC-MS/MS performed in both negative and positive ion modes. Elemental composition, determined on the base of high resolution data (mass accuracy tolerance < 2 ppm), fragmentation patterns of MS and MS/MS spectra, as well as chromatographic behavior of the corresponding compounds, allowed their structures to be proposed. It is known that triterpene glycosides are characterized by large structural variability. Generally, mass spectrometry does not permit to determine configuration of the unknown compounds. Besides, different epimeric monosaccharides as well as types of bonds between sugars cannot be strictly distinguished only by MS. However, the combination of the obtained data and biosynthetic considerations allows tentative structural assignments for the detected compounds.  (10,11,13,14,15,16,21,23,25,26,28,29,31,32,37,38,41,43,47, and 48) and proposed (12,17,18,22,27,30,33,34,35,36,39,40,42,44,45,49,50,51, and 53) from the sea cucumber E. fraudatrix by LC-MS/MS method.   It is known that the majority of triterpene glycosides have a xylose at C-3 of the aglycone as the first monosaccharide unit, quinovose as the second monosaccharide unit, and glucose (or xylose) as the third monosaccharide unit in the main chain [1,4]. Methylated monosaccharides are always terminal units. Considering that the oligosaccharide chains of earlier isolated triterpene glycosides from this sea cucumber are closely related to each other and have general architecture xylopyranosyl in the linear part of carbohydrate chains and may have β-D-xylopyranosyl-(1→2) in the branching at the second monosaccharide, some suggestions concerning structures of oligosaccharide chains of unknown glycosides may be proposed. In addition, most of the previously isolated glycosides of E. fraudatrix have a 16β-acetoxyholosta-7-ene aglycone. These common structural features give a possibility to propose structures for a series of newly identified glycosides.
The structures of detected glycosides were characterized by tandem MS. The (−)MS/MS provided many product ion series arising from the cleavages of both glycosidic bonds and bond of aglycone side chain. (+)MS/MS provided an intense B-and C-type product ion series (nomenclature according to Domon and Costello [39]) arising from the cleavages of glycosidic bonds with charge located on saccharide fragment (Table S1). These product ion series are characteristic and provided information about the sequence of monosaccharide units in carbohydrate chains. For example, the positive product ion spectrum of [ Figure 3). This fragmentation pattern corresponded to a branched non-sulfated oligosaccharide chain consisting of five monosaccharide units and compound 41 was identified as cucumarioside C 2 . It is known that the majority of triterpene glycosides have a xylose at C-3 of the aglycone as the first monosaccharide unit, quinovose as the second monosaccharide unit, and glucose (or xylose) as the third monosaccharide unit in the main chain [1,4]. Methylated monosaccharides are always terminal units. Considering that the oligosaccharide chains of earlier isolated triterpene glycosides from this sea cucumber are closely related to each other and have general architecture 3-O-methyl-β- in the linear part of carbohydrate chains and may have β-D-xylopyranosyl-(1→2) in the branching at the second monosaccharide, some suggestions concerning structures of oligosaccharide chains of unknown glycosides may be proposed. In addition, most of the previously isolated glycosides of E. fraudatrix have a 16β-acetoxyholosta-7-ene aglycone. These common structural features give a possibility to propose structures for a series of newly identified glycosides.
The structures of detected glycosides were characterized by tandem MS. The (−)MS/MS provided many product ion series arising from the cleavages of both glycosidic bonds and bond of aglycone side chain. (+)MS/MS provided an intense B-and C-type product ion series (nomenclature according to Domon and Costello [39]) arising from the cleavages of glycosidic bonds with charge located on saccharide fragment (Table S1). These product ion series are characteristic and provided information about the sequence of monosaccharide units in carbohydrate chains. For example, the positive product ion spectrum of [ (Figure 3). This fragmentation pattern corresponded to a branched non-sulfated oligosaccharide chain consisting of five monosaccharide units and compound 41 was identified as cucumarioside C2. In some cases, several typical mass losses between the precursor and the fragment ions were detected in product ion spectra. These typical mass losses are related to the aglycone and provide information about the structure of nucleus and side chain. In MS/MS spectra of the majority of the glycosides, a mass loss of 60 Da between the precursor and the intense fragment ion was detected (for example, Figure S2). This corresponds to the loss of C2H4O2 molecule (acetic acid) and is a characteristic of the glycosides containing an acetoxy group [40]. Next intense fragment ion with a mass loss 104 Da from the precursor corresponds to the loss of a [C2H4O2 + CO2] fragment and is characteristic for the glycosides containing an acetoxy group and a 18(20)-lactone cycle. Analysis of In some cases, several typical mass losses between the precursor and the fragment ions were detected in product ion spectra. These typical mass losses are related to the aglycone and provide information about the structure of nucleus and side chain. In MS/MS spectra of the majority of the glycosides, a mass loss of 60 Da between the precursor and the intense fragment ion was detected (for example, Figure S2). This corresponds to the loss of C 2 H 4 O 2 molecule (acetic acid) and is a characteristic of the glycosides containing an acetoxy group [40]. Next intense fragment ion with a mass loss 104 Da from the precursor corresponds to the loss of a [C 2 H 4 O 2 + CO 2 ] fragment and is characteristic for the glycosides containing an acetoxy group and a 18 (20) Fragmentation of oligosaccharide chain of glycosides of the group III (34 and 39) was similar to fragmentation of oligosaccharide chain of glycosides of the group I, but all fragment peaks in MS/MS spectra were shifted by 30 Da (Table S1). This may be due to the replacement of the terminal methylated xylose with a methylated glucose residue. Thus, group III includes compounds having non-sulfated main oligosaccharide chain with methylated glucose, glucose, quinovose and xylose monosaccharide units and xylose as branching unit. Thus, all glycosides of E. fraudatrix can be divided into eleven groups in accordance with structures of oligosaccharide chains. Five types of oligosaccharide chains were not found in E. fraudatrix previously. According to literature data, the majority of triterpene glycosides of E. fraudatrix have 3-O-MeXyl as terminal monosaccharide unit. We revealed new glycosides with terminal 3-O-MeGlc residue (22, 27, 34, and 39); two of them (34, and 39) were isolated, and their tentative structures are further supported by 1D NMR (Figures S6-S8).
The fact that not all previously isolated glycosides have been found by LC-MS approach in this study of E. fraudatrix could be explained by the changes in the quantitative and qualitative composition of different components of the glycosidic fraction in the samples of one species collected in different places and seasons. A representative example of such changes has been reported for the components of the glycosidic fraction of Psolus fabricii, where two different glycosides were predominant or minor in the samples collected near Onekotan Island or Ushishir Islands (Kuril Islands) [41,42]. Another example is given by the glycosides of Massinium (=Neothynidium) magnum where the structures of glycosides from various places or various times of collection were strongly different [43]. In addition, some earlier isolated cucumariosides of A-group may be artifacts formed during the isolation process [24].
Obtained data allowed us to propose a biosynthetic pathway for oligosaccharide chains in E. fraudatrix (Figure 4), in agreement with the biosynthetic pathway of oligosaccharide chains proposed earlier [27]. The elongation of the oligosaccharide chain occurs by the addition of monosaccharide residues to various positions of the forming oligosaccharide chain. This leads to the formation of glycosides with different oligosaccharide chains. Sulfatation of triterpene glycosides may occur at different stages of the forming of carbohydrate chains resulting in the appearance of sulfated oligosaccharide moieties comprised from two to six sugar units. From this viewpoint, cucumarioside B 2 [27]

Distribution of Detected Glycosides in the Different Body Components
Quantitative and qualitative analysis of detected triterpene glycosides in various body components of E. fraudatrix was also performed. We separately extracted triterpene glycoside mixtures from respiratory trees (RT), body walls (BW), gonad tubules (GN), guts (G) and aquapharyngeal bulbs (AB) and analyzed them by LC-ESI QTOF-MS. The profiling revealed that all the triterpene glycosides detected in the whole body extract were also present in all analyzed body parts.
The maximal content of overwhelming majority of the analyzed glycosides was observed in the body walls when compared with other body components of sea cucumber. This observation is a very good corroboration of a defensive role of triterpene glycosides. Since E. fraudatrix does not contain Cuvierian tubules, it accumulates the defensive molecules of triterpene glycosides in the body walls in order to indicate to predator its unpalatability. The main components of glycosidic fractioncucumariosides C1 (37) and C2 (41), compounds 39 and 45, as well as cucumariosides F1 (38) and G1 (32)-predominate in the body walls. These compounds contain pentasaccharide non-sulfated or tetrasaccharide mono-or disulfated carbohydrate chains making them highly hydrophilic substances and accelerating their diffusion to the surrounding water. Moreover, such compounds usually demonstrate significant membranolytic activities [1]. All these data also confirm the main external function of glycosides as chemical defense system. The profiling of extracts from different body components revealed that relative amounts (normalized by sum and scaling) of most compounds were approximately the same ( Figure 5; Figure  S9). However, several minor compounds were more typical for certain body components. Relative amounts of compounds 12, 15, 17 and some others ( Figure 5; Figures S9 and S10) are significantly higher in gonads than in body walls or other organs. The analysis of their structures revealed that compounds 15 and 17 have non-holostane aglycones with shortened side chains thus having structural similarity with steroidal hormones of vertebrates. It is known that sex hormones of vertebrates are biosynthesized from cholesterol via the cleavage of its side chain through the oxidation of C-20 and C-22 positions [44]. Actually, compound 12 contains a oxygen-bearing

Distribution of Detected Glycosides in the Different Body Components
Quantitative and qualitative analysis of detected triterpene glycosides in various body components of E. fraudatrix was also performed. We separately extracted triterpene glycoside mixtures from respiratory trees (RT), body walls (BW), gonad tubules (GN), guts (G) and aquapharyngeal bulbs (AB) and analyzed them by LC-ESI QTOF-MS. The profiling revealed that all the triterpene glycosides detected in the whole body extract were also present in all analyzed body parts.
The maximal content of overwhelming majority of the analyzed glycosides was observed in the body walls when compared with other body components of sea cucumber. This observation is a very good corroboration of a defensive role of triterpene glycosides. Since E. fraudatrix does not contain Cuvierian tubules, it accumulates the defensive molecules of triterpene glycosides in the body walls in order to indicate to predator its unpalatability. The main components of glycosidic fraction-cucumariosides C 1 (37) and C 2 (41), compounds 39 and 45, as well as cucumariosides F 1 (38) and G 1 (32)-predominate in the body walls. These compounds contain pentasaccharide non-sulfated or tetrasaccharide mono-or disulfated carbohydrate chains making them highly hydrophilic substances and accelerating their diffusion to the surrounding water. Moreover, such compounds usually demonstrate significant membranolytic activities [1]. All these data also confirm the main external function of glycosides as chemical defense system. The profiling of extracts from different body components revealed that relative amounts (normalized by sum and scaling) of most compounds were approximately the same ( Figure 5; Figure S9). However, several minor compounds were more typical for certain body components. Relative amounts of compounds 12, 15, 17 and some others ( Figure 5; Figures S9 and S10) are significantly higher in gonads than in body walls or other organs. The analysis of their structures revealed that compounds 15 and 17 have non-holostane aglycones with shortened side chains thus having structural similarity with steroidal hormones of vertebrates. It is known that sex hormones of vertebrates are biosynthesized from cholesterol via the cleavage of its side chain through the oxidation of C-20 and C-22 positions [44]. Actually, compound 12 contains a oxygen-bearing substituent in C-22 position as well as a oxidized C-20 position that makes it a putative biosynthetic precursor of the aglycones with shortened side chains such as 15 and 17. All these data are in good agreement with the earlier suggested internal biological function of the glycosides-the regulation of oocytes maturation in the sea cucumbers.  Interestingly, the glycosides of group IV (36 and 40) having peculiar "undeveloped" branched tetrasaccharide chains without methylated terminal sugar unit were characteristic for guts ( Figure  S11). There is a correlation between the contents of some glycosides in aquapharyngeal bulbs and respiratory trees (compounds 1, 2, 4, 5, 7, 8, and 9; Figure S9) that may indicate additional biological functions of triterpene glycosides in the organism-producer, which have to be investigated.

Sample Preparation and Solid-Phase Extraction (SPE)
Seven fresh animals (m = 3.9 ± 1.3 g) were chopped and extracted thrice with ethanol (totally 0.5 L, for 10 h). Other seven animals were dissected and separated into respiratory trees (RT, m = 1.7 g), body walls (BW, m = 10.2 g), gonad tubules (GN, m = 3.9 g), guts (G, m = 4.3 g) and aquapharyngeal bulbs (AB, m = 3.9 g). Metabolites from five different body components were extracted thrice with ethanol (totally 100 mL, for 10 h). Extracts were filtered, dried and reconstituted in 15 mL ethanol.
Two hundred microliters of the extract was centrifuged, and the supernatant was subjected solid-phase extraction (SPE). SPE cartridges (BondElut C18, 100 mg/1 mL, Agilent Technologies, Santa Clara, CA, USA) were fitted into stopcocks and connected to a vacuum manifold. The sorbent was conditioned with 3 mL of ethanol followed by 3 mL water. Care was taken that the sorbent did not become dry during conditioning. With the stopcocks opened and the vacuum turned on, the samples were loaded onto the cartridge. The 100 µL of extract were loaded into the SPE cartridge by drops. After sample addition, the SPE cartridge was washed with 1 mL of water. Glycosides were eluted with 1 mL of 100% ethanol. These extracts were dried and dissolved in 500 µL 80% MeOH in water (v/v) and subjected to LC-MS analyses.

LC-MS Analysis
Analysis was performed using an Agilent 1200 series chromatograph (Agilent Technologies, Santa Clara, CA, USA) connected to a Bruker Impact II Q-TOF mass spectrometer (Bruker Daltonics, Bremen, Germany). Zorbax Eclipse XDB-C18 column (1.0 × 150 mm, 3.5 µm, Agilent Technologies, Santa Clara, CA, USA) with Zorbax SB-C8 guard-column (2.1 × 12.5 mm, 5 µm, Agilent Technologies, Santa Clara, CA, USA) were used for chromatographic separation. The mobile phases were 0.1% formic acid in H 2 O (eluent A) and 0.1% formic acid in MeOH (eluent B). The gradient program was as follows: isocratic at 60% of eluent B from start to 3 min, from 60% to 90% eluent B from 3 to 29 min, from 90% to 100% eluent B from 29 to 30 min, isocratic at 100% of eluent B to 35 min, from 100% to 60% eluent B from 35 to 38 min. After returning to the initial conditions, the equilibration was achieved after 15 min. Chromatographic separation was performed at a 0.1 mL/min flow rate at 40 • C. Injection volume was 1 µL.
The mass spectrometry detection has been performed using ESI ionization source. Optimized ionization parameters for ESI were as follows: a capillary voltage of ±4.0 kV, nebulization with nitrogen at 0.8 bar, dry gas flow of 7 L/min at a temperature of 200 • C. Metabolite profiles in positive ion mode were registered using post-column addition of 5 × 10 −4 M sodium iodide at 60 µL/h flow rate for obtaining stabilized sodium adduct ions. Post-column infusion was performed with syringe pump via T-mixing tee. Based on the results of preliminary experiments, the mass spectra were recorded within m/z mass range of 100-1500 and 70-1500 for MS/MS spectra (scan time 1 s).
Collision induced dissociation (CID) product ion mass spectra were recorded in auto-MS/MS mode with a collision energy ranging from 75 to 125 eV (an exact collision energy setting depended on the molecular masses of precursor ions). The precursor ions were isolated with an isolation width of 4 Th.
All tests were performed at least in triplicate. The results are expressed as the mean ± standard deviation (SD).

Quantitative Analysis of Detected Triterpene Glycosides in Various Body Component
For quantitative analysis, we used cucumarioside A 1 from E. fraudatrix [24] as a reference standard for non-sulfated glycosides (R 2 = 0.988), typicoside A 2 from Actinocucumis typica [44] as a reference standard for monosulfated glycosides (R 2 = 0.999) and cucumarioside I 2 from E. fraudatrix [37] as a reference standard for disulfated glycosides (R 2 = 0.995). Standards at concentrations of 1.0, 2.5, 5.0, 10.0 and 50 µg/mL were used for building calibration curves ( Figures S3-S5). All the experiments were carried out at least three times, LC-MS conditions are identical to those described above. As a result, the amounts of detected compounds were calculated through calibration curves. Results are shown in Table 1 as mean concentration in µg/g animal organs.
For comparative analysis of the content of triterpene glycosides in different organs of E. fraudatrix data pretreatment was performed. With the aim to make feature compounds more comparable to each other, the datasets were adjusted using scaling by 100% (for 100%-the total concentration of the compound in all organs). Results are shown in Figure 5 and Figures S9-S11 as the mean ± standard deviation (SD).

Conclusions
Profiling of E. fraudatrix was performed by LC-ESI MS. Analysis of chromatographic behavior, MS and MS/MS data allowed the structural identification of the triterpene glycosides to be performed. Totally, 54 triterpene glycosides were found and structures of 44 constituents were proposed based on LC-ESI MS, chromatographic behavior and biogenetic hypotheses. In accordance with the structures of oligosaccharide chains, all analyzed glycosides of E. fraudatrix can be divided into eleven groups, and five of them were found in E. fraudatrix for the first time. A theoretical scheme of biogenesis of oligosaccharide chains in the studied species was given. The comparison of the qualitative and quantitative contents from the five different body components revealed that the profiles of some triterpene glycosides differed in the body walls and gonads indicating different external and internal biological functions of these compounds. The predominance of the main highly hydrophilic and membranolytic glycosides of E. fraudatrix in the body walls confirms their defensive role. The presence of glycosides in all body components of E. fraudatrix indicates their multifunctionality.