Introduction

Ethanolamine glycerophospholipid (EtnGpl) is a major class of glycerophospholipids found in biological membranes. Moreover, these subclasses exist in three forms with alkyl, alkenyl, or acyl linkages at the sn-1 position of the glycerol moiety (1-O-alkyl-2-acyl-sn-glycero-3-phosphoethanolamine, PakEtn; 1-O-alkenyl-2-acyl-sn-glycero-3-phosphoethanolamine, PlsEtn; and 1,2-diacyl-sn-glycero-3-phosphoethanolamine, PtdEtn, respectively). The alkenylacyl form is called “plasmalogen” [1]. While plasmalogen is restricted in most tissues and cells of animals, the nervous system shows high plasmalogen levels [2]. Moreover, plasmalogen levels were reported to be specifically decreased in postmortem brains from Alzheimer’s disease patients [3–5]. Plasmalogen can also prevent cell death by scavenging singlet oxygen (1O2) and superoxide (O2 −) at its alkenyl (vinyl ether) linkages, as demonstrated by mutant cells that are deficient in an enzyme for plasmalogen biosynthesis [6]. While these findings suggest that plasmalogen has important physiological roles, its absorption and metabolism are not fully understood, mainly due to a lack of plasmalogen resources [7, 8].

Following bovine spongiform encephalopathy outbreaks, use of bovine brain as a plasmalogen resource became difficult. On the other hand, most EtnGpl exists as plasmalogen in some marine invertebrates [9, 10], such as Crassostrea gigas, Mytilus edulis [11], Strongylocentrotus intermedius, and Halocynthia roretzi [12]. These marine invertebrates are readily available around island nations, such as Japan, and can, therefore, be used as new plasmalogen resources.

There are several methods for quantifying plasmalogen that measure either plasmalogen aldehydes with 2,4-dinitrophenylhydrazines [13, 14], iodine caught in a vinyl ether bond [15], lyso-EtnGpl with thin layer chromatography (TLC) [1, 2, 16], or dimethyl acetal with gas chromatography (GC) [17, 18]. However, these methods have some drawbacks in terms of sensitivity and selectively. Other conventional methods, e.g., derivatizing diradylglycerols after phospholipase C treatment, enable complete separation of EtnGpl subclasses [19, 20]. Generally, they are identified by ultraviolet (UV) detection. However, UV detection is affected by double bonds and is, therefore, not suitable for analyzing marine products that are rich in polyunsaturated fatty acids (PUFA). On the other hand, high-performance liquid chromatography (HPLC) with evaporative light-scattering detection (ELSD), which is insensitive to the mobile phase solvents and thus allows direct quantification, has been applied to lipid analyses [21–23]. Unlike UV detection, ELSD is relatively unaffected by double bonds and length of fatty acids [24].

In this study, a method was developed for determining EtnGpl subclasses consisting of derivatization by phospholipase C treatment and acetylation, and separation and detection by HPLC–ELSD. This method was applied to 24 marine invertebrates available in Japan.

Materials and Methods

Materials

Crassostrea gigas, Patinopecten yessoensis, Todarodes pacificus, Paroctopus dofleini, Marsupenaeus japonicas, and Halocynthia roretzi were purchased from local supermarkets in Sendai, Japan. Other marine invertebrates were collected at Onagawa Bay in northwest Japan. 1-O-hexadecenyl-2-octadecenoyl-sn-glycero-3-phosphoethanolamine (16:0/18:1-PlsEtn), 1-O-hexadecenyl-2-docosahexaenoyl-sn-glycero-3-phosphoethanolamine (16:0/22:6-PlsEtn), 1,2-di-octadecenoyl-sn-glycero-3-phosphoethanolamine (18:1/18:1-PtdEtn), and 1-O-hexadecyl-2-octadecenoyl-sn-glycero-3-phosphoethanolamine (16:0/22:6-PakCho) were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Phospholipids (Phospholipid Kit) and 60 wt% plasmalogen of EtnGpl from bovine brain were purchased from Doosan Serdary Research Laboratories (Toronto, ON, Canada). Phospholipase C from Bacillus cereus type III was purchased from Sigma Chemical Co. (St. Louis, MO, USA). Methanol, n-hexane, isopropanol, and chloroform were of HPLC grade, and other reagents were of extra-pure grade.

Lipid Extraction

EtnGpl subclass analysis was performed as described in Fig. 1. Samples were freeze-dried, crushed, and kept at a temperature below −30 °C for further processing. The freeze-dried samples (about 20 mg as total lipid) were added to 20 volumes of chloroform/methanol (2:1, v/v) and vortexed for 1 min. The total lipids were prepared according to the method of Folch et al. [25]. The total lipids were redissolved in 500 μL of chloroform/methanol (2:1, v/v).

Fig. 1
figure 1

Determination for EtnGpl subclasses. The measurements of each process were as follows: 1 wet weight by weighing, 2 dry weight by weighing, 3 total lipid weight by weighing, phospholipid content by Rouser method and content of phospholipid subclasses by HPLC–ELSD, 4 EtnGpl content by HPLC–ELSD, and 6 content of acetyl diradylglycerol subclasses by this developed method (the values were corrected by the recovery ratio of derivatized 18:1/18:1-PtdEtn)

Full size image

Phospholipid Analysis

Phospholipid contents in the total lipids were determined according to the method described by Rouser et al. [26]. Phospholipid classes were analyzed by HPLC–ELSD [27]. The silica column was LiChrosorb SI100 (4.6 × 250 mm, \( \varphi \) 10 μm; Waters Corporation, Milford, MA, USA) with a binary gradient consisting of solvent A [chloroform/methanol/30 % ammonium hydroxide (80:19.5:0.5, by vol)] and solvent B [chloroform/methanol/water/30 % ammonium hydroxide (60:34:5.5:0.5, by vol)]. The gradient profile was as follows: 0–14 min, 100 % B linear gradient; 14–24 min, 100 % B. The flow rate was 1.0 mL/min, and the column was maintained at a temperature of 35 °C. The post-column ELSD was a SEDEX model 55 (Sedere, Vitry sur Seine, France), kept at an evaporation temperature of 60 °C and pressure of 2.0 bar (2.7 L/min) for nebulization gas (nitrogen). The photomultiplier sensitivity was adjusted to a gain of 8.

Phospholipid Class Separation

On TLC (Silica 60, Merck, Darmstadt, Germany), phospholipid classes were separated by chloroform/methanol/acetone/water (80:25:10:4, by vol) and EtnGpl was detected by primulin. The EtnGpl fraction was extracted by the method of Bligh and Dyer [28].

Derivatization of EtnGpl Subclasses

To subfractionate EtnGpl into the diacyl, alkenylacyl, and alkylacyl subclasses, EtnGpl was derivatized by a modification of the method of Guan et al. [29]. About 100 μg of EtnGpl was redissolved in 1.5 mL of diethyl ether saturated with water. Phospholipase C was dissolved in 100 mM Tris–HCl (pH 7.4), and 500 μL of this solution (containing 1.0 unit) was added to the dissolved EtnGpl. Incubation was performed at 37 °C for 30 min. After incubation, 500 μL of distilled water and 1.0 mL of diethyl ether were added and centrifuged at 2,500 rpm for 5 min. After separation of the layers, 2.5 mL of diethyl ether was added to the lower layer and centrifuged. The separated upper layers were mixed and dried under nitrogen. The residue was dissolved in 15 μL of benzene, 15 μL of acetate anhydride, and 30 μL of 0.68 M 4-dimethylaminopyridine in pyridine. The reaction was mixed at 37 °C for 3 h. After the reaction, 1.8 mL of methanol, 100 μL of 100 mM Tris–HCl (pH 8.5) and 1.2 mL of petroleum ether were added and centrifuged. Then, 2.4 mL of petroleum ether was added to the lower layer. The upper layers were collected, dried under nitrogen, and redissolved in 100 μL of methanol.

Liquid Chromatography–Mass Spectrometry (LC–MS) Analysis for Derivatized EtnGpl

Diradylglycerols and acetyl diradylglycerols were confirmed by HPLC with electrospray ionization-time of flight mass spectrometry (ESI-TOF MS). The C18 column was TSK-GEL ODS-80Ts (4.6 × 250 mm, \( \varphi \) 5 μm; TOSOH Co., Tokyo, Japan), and the mobile phase consisted of methanol/0.5 M ammonium acetate (99:1, v/v). The flow rate was 1.0 mL/min, and the column was maintained at a temperature of 28 °C. The ESI-TOF MS was a Mariner (Applied Biosystems, Carlsbad, CA, USA). Electrospray ionization was performed in positive ion mode. Spray voltage, nozzle potential, nozzle temperature, and nebulizer gas flow rate were set to 2,000 V, 150 V, 150 °C, and 3.0 mL/min, respectively.

HPLC–ELSD for Acetyl Diradylglycerols

Acetyl diradylglycerols were analyzed by HPLC–ELSD. The silica column was a Zorbax Rx-SIL (4.6 × 250 mm, \( \varphi \) 5 μm; DuPont, Wilmington, DE, USA), and the mobile phase was n-hexane–isopropanol (100:0.25, v/v). The flow rate was 1.5 mL/min, and the column was maintained at a temperature of 36 °C. The ELSD conditions were as described above. Standards were made from 16:0/18:1-PlsEtn and 18:1/18:1-PtdEtn and the purities were determined by GC [30]. The value of each sample was corrected by the ratio of derivatized 18:1/18:1-PtdEtn under the same conditions.

Spike and Recovery Experiment

To confirm the accuracy of this method, a known amount of plasmalogen was spiked into a biological sample and its response was recovered. We used 60 wt% plasmalogen from bovine brain as the additional plasmalogen, because the single species may have the inherent recovery by the polarity and structure. Briefly, 1, 2, and 4 mg (i.e., 0.6, 1.2, and 2.4 mg for plasmalogen) of 60 wt% plasmalogen were added to 180 mg of rat brain (1.2 mg for plasmalogen) before lipid extraction, and the recoveries were determined after this analytical process. This experiment was conducted in compliance with the policies and procedures detailed in the Animal Experiment Guidelines of Tohoku University.

Results

EtnGpl Derivatization

Reaction conditions of phospholipase C and acetylation were determined by pretesting EtnGpl standards (60 wt% plasmalogen of EtnGpl, 16:0/18:1-PlsEtn and 18:1/18:1-PtdEtn) to confirm diradylglycerols and acetyl diradylglycerols with TLC (data not shown) and total ion current chromatogram (TIC) of LC–MS (Fig. 2). Subsequently, under the same conditions, EtnGpl extracted from rat brains as biological samples was also confirmed to be completely reacted.

Fig. 2
figure 2

Total ionization current chromatogram (TIC) of derivatized EtnGpl by LC–MS. A Derivatives from 16:0/18:1-PlsEtn, B derivatives from 18:1/18:1-PtdEtn; (a), (b), and (c) indicate original EtnGpl, diradylglycerol, and acetyl diradylglycerol, respectively

Full size image

Separation and Detection of Acetyl Diradylglycerols by HPLC–ELSD

Three types of acetyl diradylglycerol, i.e., alkenylacyl, alkylacyl, and diacyl subclasses, could be completely separated from each other on a silica column with an isocratic system consisting of n-hexane–isopropanol (100:0.25, v/v) (Fig. 3). Acetyl diradylglycerol standards, alkenylacyl, alkylacyl, and diacyl subclasses, were derivatized from 16:0/18:1-PlsEtn, 16:0/22:6-PakCho and 18:1/18:1-PtdEtn, respectively.

Fig. 3
figure 3

HPLC–ELSD chromatograms of acetyl diradylglycerols. a Alkenylacyl subclass from 16:0/18:1-PlsEtn, b alkylacyl subclass from 16:0/18:1-PakCho, c diacyl subclass from 18:1/18:1-PtdEtn, d derivatives from rat brain EtnGpl, e derivatives from rat liver EtnGpl. Rat tissues were obtained from male Sprague–Dawley rats (7 weeks old) after fasting for 24 h

Full size image

The calibration curves were constructed with two acetyl diradylglycerol standards, alkenylacyl and diacyl derivatives, ranging in concentration from 1 to 10 μg. The calibration curves were expressed by the equation y = ax b, which is a common feature of ELSD [21, 23, 31, 32], and the regression correlation coefficients were above 0.996 (Fig. 4), with detection limits of 200 ng/injection at a signal-to-noise ratio of 3. The calibration curves for alkenylacyl and diacyl derivatives were nearly the same. The values of the alkenyl and diacyl derivatives were converted to those of plasmalogen and PtdEtn, respectively. The concentrations of PakEtn were determined by subtracting those of the plasmalogen and PtdEtn from the total EtnGpl because the spike and recovery test could not be performed without the PakEtn standard. The reaction recovery of the PakCho was different from those of plasmalogen and PtdEtn.

Fig. 4
figure 4

Calibration curves of acetyl diradylglycerol standards in HPLC–ELSD. The calibration curves were constructed with acetyl diradylglycerol standards of different concentrations (1–10 μg/injection). Three determinations for each standard were given at different concentrations

Full size image

Subsequently, the accuracy of this method was confirmed by a spike and recovery experiment (Table 1).

Table 1 Spike and recovery test for HPLC–ELSD method
Full size table

Comparison of Methods for Determining Plasmalogen

The developed method using HPLC–ELSD was compared with conventional methods for plasmalogen (Table 2), i.e., the 2,4-dinitrophenylhydrazine method (NPH) [14], and separation and detection of acetyl diradylglycerols by TLC [33]. In rat brain, a higher level of plasmalogen was obtained by HPLC–ELSD than by the two other methods. Plasmalogen values obtained by TLC varied widely. In rat liver, plasmalogen was not detected by TLC.

Table 2 Comparison of methods for determining plasmalogen contents in rat tissues
Full size table

Lipid Contents in Marine Invertebrates

The moisture and lipid contents of the marine invertebrates shown in Table 3 were investigated (Table 4). Most of the samples examined had moisture and total lipid contents of 70–90 % and 5–20 % (dry wt), respectively. The liver of Halocynthia roretzi had the highest moisture content, while the viscera of Todarodes pacificus had the lowest moisture content and the highest total lipid content.

Table 3 Taxonomic classification of the marine invertebrates examined
Full size table
Table 4 Moisture and lipid contents in the marine invertebrates examined
Full size table

The phospholipid contents were very high in Asterina pectinifera and muscles of Todarodes pacificus. In Asterias amurensis and the viscera of the invertebrates examined, total lipids consisted mostly of non-phospholipid.

As shown in Table 5, EtnGpl contents of phospholipid were as follows: >35 mol%, Acanthopleura japonica, Cellana grata, Chlorostoma argyrostoma, Septifer virgatus, Stichopus japonicus (red sea cucumber), and the muscle of Paroctopus dofleini; 34–35 mol%, Thais bronni, Nucella heyseana, Mytilus galloprovincialis, Asterias amurensis, and the liver of Halocynthia roretzi. In Stichopus japonicus (green sea cucumber), the gonad of Strongylocentrotus nudus, and the muscle of Todarodes pacificus, ChoGpl constituted about 70 mol% of the phospholipid. The contents of others were high in Halichondria panacea and Halichondria japonica.

Table 5 Phospholipid contents in the marine invertebrates examined
Full size table
Table 6 Comparison of amounts of EtnGpl subclasses in the marine invertebrates examined
Full size table

EtnGpl subclass contents in marine invertebrates

Table 5 shows the EtnGpl subclasses present in marine invertebrates. Plasmalogen levels of EtnGpl were as follows: >90 mol%, Anthopleura midori (Fig. 5a), Anthopleura japonica, Thais bronni, Mytilus galloprovincialis, Crassostrea gigas, Asterias amurensis (Fig. 5b), Asterina pectinifera, and the gonad of Strongylocentrotus nudus; 80–90 mol%, Acanthopleura japonica, Nucella heyseana, Chlorostoma argyrostoma, Septifer virgatus, Stichopus japonicus (red and green sea cucumbers), and the muscle and liver of Halocynthia roretzi (Fig. 5c, d, respectively). In Todarodes pacificus, PtdEtn constituted about 60 mol% of EtnGpl. PakEtn content was high in the muscle of Paroctopus dofleini. In Halichondria panacea and Halichondria japonica, EtnGpl subclasses were not analyzed because of a lack of separation of EtnGpl by TLC.

Fig. 5
figure 5

HPLC–ELSD chromatograms of acetyl diradylglycerols from EtnGpl of marine invertebrates. a Anthopleura midori, b Asterias amurensis, c Halocynthia roretzi muscle, d Halocynthia roretzi liver. a, b, c, and d were injected with 15, 10, 20, and 20 μL as step 6 in Fig. 1, respectively

Full size image

Plasmalogen contents of fresh and dried samples are shown in Table 6. Plasmalogen contents in 100 g of fresh samples were as follows: >400 μmol, Septifer virgatus, Asterias amurensis, and Asterina pectinifera; 300–400 μmol, Thais bronni and Nucella heyseana,; 200–300 μmol, Anthopleura midori, Anthopleura japonica, Cellana grata, Tugali gigas, the gonad of Strongylocentrotus nudus, and the gonad of Hemicentrotus pulcherrimus. Plasmalogen contents of dried samples tended to be the same, except for samples with high moisture and low phospholipid contents, such as Chlorostoma argyrostoma and the liver of Halocynthia roretzi.

Table 7 Contents of EtnGpl subclasses in the marine invertebrates examined
Full size table

Discussion

This study demonstrated a new HPLC–ELSD procedure using enzyme treatment and acetylation for the separation and quantitative analyses of plasmalogen in marine invertebrates. The procedure has several advantages. Cyclopentane, which consists mainly of conventional mobile phase, is expensive [20, 29]. This new mobile phase consists of n-hexane and isopropanol. Therefore, this method can be performed to separate acetyl diradylglycerols at relatively low cost. In addition, ELSD is little affected by double bonds and length of fatty acids, in contrast to UV. The calibration curves for alkenylacyl forms from 16:0/18:1-PlsEtn and 16:0/22:6-PlsEtn were also nearly the same (y = 1.1 × 105 x 1.6 and y = 1.2 × 105 x 1.6, respectively).

In comparison with other methods, the HPLC–ELSD method yielded higher values of plasmalogen. In previous reports for the brain of adult rats with various methods, plasmalogen accounted for 55–57 mol% of EtnGpl [34–38]. In contrast, a lower value was reported using the aldehyde derivatives of alkenylacyl chains (41.7 mol%) [39], and use of HPLC for diradylglycerol derivatives yielded a higher value (65.9 mol%) [29]. It is known that diradylglycerol derivatives are stable, whereas aldehyde and the derivatives of alkenylacyl chains are unstable [40]. The background is very much higher by TLC than by HPLC–ELSD. In contrast, the plasmalogen level (61.4 mol% of EtnGpl) of the rat brain analyzed with HPLC–ELSD was about the same as that reported previously by a number of methods including a reliable method using radioactive isotopes [29]. In addition, this method was confirmed by the spike and recovery procedure.

Recently, new methods with mass spectrometry (MS) were reported for determination of plasmalogen, such as GC–MS [41] and LC–MS/MS [42, 43], and were shown to be applicable to measurement of plasmalogen with high sensitivity. Although these methods enable analysis of molecular species and fatty acids, it is difficult to obtain total plasmalogen because of the requirement for standards of each molecular species. Moreover, the equipment required for MS is very expensive. On the other hand, ELSD is relatively inexpensive and enables analysis of total plasmalogen regardless of the kind of fatty acid. Thus, this method is suitable for screening a wide range of samples.

Using the HPLC–ELSD method, we searched for plasmalogen resources in marine invertebrates. Although limited numbers of samples were collected from local markets and Onagawa Bay, the EtnGpl subclass compositions of different species are shown in Table 7. With the exception of the class Cephalopoda and Crustacea, the marine invertebrates examined had high plasmalogen contents (>60 mol% of EtnGpl). Species with especially high levels of EtnGpl were Asterias amurensis, Asterina pectinifera, Strongylocentrotus nudus, Anthopleura midori, and Anthopleura japonica. The levels of plasmalogen in fresh samples of Asterias amurensis, Asterina pectinifera, and Septifer virgatus were very high.

Although plasmalogen levels of EtnGpl in some marine invertebrates were higher in the present study than those reported previously [10, 12, 44], the differences were thought to be due to the instability of derivatives and sensitivity as described above. Moreover, Joseph previously reported that there are seasonal changes in lipid content of Mollusca because of the reproductive cycle [45]. In the muscle of Halocynthia roretzi, the amounts of the total lipid, phospholipid, and plasmalogen were lower in the present study than those reported previously, whereas the plasmalogen level of EtnGpl was higher [12]. Thus, there may also be seasonal changes in plasmalogen levels.

In H. panacea and Halichondria japonica, plasmalogen was not analyzed because EtnGpl could not be separated by TLC. Ceramide galactoside had an Rf value near EtnGpl and was detected by primulin. In Halichondria japonica, it was reported that the main glycerosphingolipid was ceramide digalactoside [46]. Because of the increased hydrophilicity, it was inferred that ceramide digalactoside was consistent with the position of EtnGpl. Although this separation for screening was performed by TLC, it is expected that separation would be improved by HPLC.

The class Asteroidea was reported previously to have high plasmalogen levels of EtnGpl [47]. In addition, two starfish examined were also rich in plasmalogen. However, starfish are too small and require a great deal of effort for collection and dissection; therefore, it is difficult to prepare large amounts of plasmalogen from starfish. On the other hand, the class Bivalvia was also reported to have high plasmalogen levels [11]. In the bivalves examined except Patinopecten yessoensis, plasmalogen accounted for about 90 mol% of EtnGpl. Septifer virgatus was also especially rich in plasmalogen, comparable to the starfish. Septifer virgatus and Mytilus galloprovincialis cause severe damage to the marine industry, and it is, therefore, possible to obtain these species in large amounts. Moreover, it is easy to separate the meat from the shell of bivalves; therefore, these species are expected to be good resources for plasmalogen.

With regard to the relation between plasmalogen levels of EtnGpl and taxonomic classification, Deuterostomia, including Echinodermata and Protochordata, contained high levels. Platyhelminthes [48] and Annelida [49], not examined in this study, were reported to contain high levels of plasmalogen. Thus, in Protostomia, lower organisms are thought to contain high levels of plasmalogen, whereas higher organisms, such as Arthropoda and the class Cephalopoda, contain low levels. In Deuterostomia, it was also reported that higher organisms, such as vertebrates, contain low plasmalogen levels except in the nervous system and testis [50–52].

In conclusion, a method for determining EtnGpl subclasses with ELSD was developed and was shown to be a powerful tool for searching for plasmalogen resources.