Quantitative Determination of Fatty Acids in Marine Fish and Shellfish from Warm Water of Straits of Malacca for Nutraceutical Purposes

This study was conducted to quantitatively determine the fatty acid contents of 20 species of marine fish and four species of shellfish from Straits of Malacca. Most samples contained fairly high amounts of polyunsaturated fatty acids (PUFAs), especially alpha-linolenic acid (ALA, C18:3 n3), eicosapentaenoic acid (EPA, C20:5 n3), and docosahexaenoic acid (DHA, C22:6 n3). Longtail shad, yellowstripe scad, and moonfish contained significantly higher (P < 0.05) amounts of eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), and alpha-linolenic acid (ALA), respectively. Meanwhile, fringescale sardinella, malabar red snapper, black pomfret, Japanese threadfin bream, giant seaperch, and sixbar grouper showed considerably high content (537.2–944.1 mg/100g wet sample) of desirable omega-3 fatty acids. The polyunsaturated-fatty-acids/saturated-fatty-acids (P/S) ratios for most samples were higher than that of Menhaden oil (P/S = 0.58), a recommended PUFA supplement which may help to lower blood pressure. Yellowstripe scad (highest DHA, ω − 3/ω − 6 = 6.4, P/S = 1.7), moonfish (highest ALA, ω − 3/ω − 6 = 1.9, P/S = 1.0), and longtail shad (highest EPA, ω − 3/ω − 6 = 0.8, P/S = 0.4) were the samples with an outstandingly desirable overall composition of fatty acids. Overall, the marine fish and shellfish from the area contained good composition of fatty acids which offer health benefits and may be used for nutraceutical purposes in the future.


Introduction
Fish and shell�sh are widely accepted as highly nutritious and healthy foods. However, people usually think that different types of �sh are of similar nutritional value, and �sh selections are made only based on availability, freshness, �avor, and other physical factors [1]. Based on the Malaysian Adult Nutrition Survey (2002)(2003), the prevalence of daily consumption of marine �sh among rural and urban adults were quite high at 51% and 34%, respectively [2]. erefore, it is crucial to increase the awareness of different nutrient contents of �sh and shell�sh species by providing complete nutritional value information, especially for fatty acid content, which is associated with various health-related effects.
Nutritional analysis of fatty acids can be classi�ed as qualitative and quantitative. Qualitative analysis of fatty acid produces data regarding the fatty acid composition in the form of percentages of total fatty acids (% of total fatty acids). Meanwhile, quantitative analysis is able to quantify the actual amount (weight) of each fatty acid that is present in the food. Quantitative data are oen presented in the form of weight of the fatty acid per weight of food or fat (e.g., mg/g oil).
Currently, there are limited qualitative fatty acid data on marine �sh and shell�sh from warm water area; meanwhile, no quantitative data are available especially as a mean to utilize source of natural product for nutraceutical purposes.
Reliable analytical data are prerequisite for correct interpretation of �ndings in nutritional content analysis, since unreliable results may lead to over-or underestimations, false interpretations, and unwarranted conclusions [3]. us, validation procedures of analytical methods were also performed to provide reliable quantitative data on the fatty acid contents of various species of local marine �sh and shell�sh.

Chemicals and Reagents.
All chemicals and reagents used for analysis were of analytical and gas chromatography (GC) grade. Menhaden oil, 37 components FAME mix 47885-U (Supelco, Germany), and tridecanoic acid (C13, internal standard, Sigma, USA) were used as standards in the fatty acid content analysis.

Preparation of Sample.
A strati�ed random sampling procedure was used as it was the most suitable method in database work [4]. To ensure representativeness, ten �sh landing areas along the Straits of Malacca were identi�ed with the help of Malaysian Fisheries Development Authority (LKIM). e locations are marked as L1 through L10, respectively (Figure 1).
At each of the collection sites, available samples were collected randomly according to species. All samples were from free-roaming �sh and shell�sh; and they were collected fresh (caught within a period of 0 to 36 hours). All samples were immediately placed on ice, kept cold, and transported in polystyrene boxes to maintain freshness. Upon arrival at Universiti Putra Malaysia, the temperature of ice boxes was checked to ensure samples were still within the range of −4 ∘ C to 0 ∘ C. en, samples for nutrient determination were individually measured for total body weight and length.
Only samples of a weight within the narrow range for each species were included as primary samples (Table 1). en, the samples were beheaded, gutted, washed, and �lleted. ese primary samples were placed in sealed plastic bags and frozen at −20 ∘ C. e actual degree of freshness of samples during transportation and time of analysis may not be assured as there was no analysis of freshness or quality index of samples was done. However, appropriate precautions were performed to sustain freshness of samples and minimize oxidation throughout the study by performing procedures in chillers' room (4 ∘ C) and under minimal light exposures.
A small-scale experiment performed independently showed insigni�cant di�erences in nutrient contents of samples from di�erent locations. is observation justi�ed that units of primary samples can be combined or composited by geographical locations to minimize the number of analytical measurements and yet represent the contribution of the unit to the estimate of central tendency [4]. us, before analysis, three composite samples were prepared by mixing individual samples ( �12, whole �llets) of the same weight for each species. Individual samples from L1, L2, L3, and L4 constituted as Composite 1; individual samples from L5, L6, and L7 constituted as Composite 2; while individual samples from L8, L9, and L10 constituted as Composite 3 ( Figure 1). All composite samples were analyzed separately and data presented are the mean values of each of the species.

Extraction of Fat.
Extraction of fat was done following Bligh and Dyer [6], with slight modi�cations by Kinsella et al. [7]. Representative samples of �sh �lets (30 g) were homogenized in a Waring blender for 2 min with a mixture of methanol (60 mL) and chloroform (30 mL). One volume of chloroform (30 mL) was added to the mixture and aer blending for additional 30 seconds, distilled water (30 mL) was added. e homogenate was stirred with a glass rod and �ltered through Whatman No. 1 �lter paper on a Buchner funnel with slight suction. e �ltrate was transferred to a separatory funnel. e lower clear phase was drained into a 250 mL round-bottom �ask and concentrated with a rotary evaporator at 40 ∘ C. To minimize oxidation, the extracted lipids were kept in solvents containing 0.05% butylated hydroxytoluene (B.H.T.), in glass bottles, �ushed with nitrogen, and wrapped with aluminium foils to avoid any light exposure. en, the bottles were stored immediately at −20 ∘ C and only being taken out from freezer just before analysis.

Analysis of Fatty Acid Content.
Lipid samples were converted to their constituent FAMEs following the method used in previous study [8]. Approximately 25 mg (±0.1 mg) of oil was weighed and added with 1.5 mL of NaOH 0.50 M in methanol in a 15 mL capped centrifuge tube. e mixture was heated in a water bath at 100 ∘ C for 5 min and then cooled at room temperature. e mixture was added with 2.0 mL of boron tri�uoride (BF 3 , 12%) in methanol and heated again in a water bath at 100 ∘ C for 30 minutes. Next, the tube was cooled in running water at room temperature before adding 1 mL of isooctane. It was vigorously stirred for 30 seconds Categories of �sh samples were based on the fat content [5]. 2 Different superscript lowercase letters ( a-c ) in the same column show signi�cant differences at < 0.05 (Tukey post-hoc test) [5]. before adding 5.0 mL of a saturated sodium chloride solution to facilitate the phase separation. e esteri�ed sample was placed in a refrigerator and le to rest for better phase separation. Aer collecting the supernatant, another 1.0 mL of isooctane (containing 0.05% B.H.T. as antioxidant) was added into the tube and stirred. e supernatant was collected and added to the previous fraction. e sample was concentrated to a �nal volume of 1.0 mL for later injection into the gas chromatograph. As precautions, amber vials were used in order to minimize oxidation during analysis.
Analysis of Gas Chromatography (GC). Analysis of methyl esters was performed by a capillary gas chromatograph model Agilent 6890 (USA Agilent Technology) equipped with a split-splitless injector, �ame ionization detection (FID) system used to separate and quantify each FAME component. FAMEs were separated using a highly polar HP88 column (Agilent, USA) column (100 m length × 0.25 mm I.D. × 0.2 m D.F.). Carrier gas was helium at a linear velocity of 30.0 mL/min. Split injection with a split ratio (volume of gas passing to waste: volume of gas passing down the capillary column) of 10 : 1 and 99.9 mL/min split �ow was used. e operating conditions were 250 ∘ C injection port, 250 ∘ C �ame ionization detector, and 200 ∘ C column temperature. Compounds were identi�ed by comparison of the retention times of 37 components FAME mix 47885-U (Supelco, Germany) and Menhaden oil standards (Supelco, Germany).
��anti�cation of �atty Acids. e concentration of fatty acids in mg/g of total lipids was measured against tridecanoic acid methyl ester (C13:0, Sigma, USA) as an internal standard. is followed procedures described in a previous study [6] with slight modi�cations; in which the Empirical Response Factor ( ) of FID (�ame ionization detector) was calculated and used instead of theoretical response factor ( FX ). Calculation of the empirical response factor ( ) was done as described in the literature [9]. e following formulae were used in the quanti�cation of fatty acids: where Psi: peak area of individual fatty acids in mixed FAME standard solution, PsIS: peak area of fatty acid internal standard in mixed FAME standard solution, WISm: weight of fatty acid internal standard in mixed FAME standard solution, and Wi: weight of individual FAME in mixed FAME standard solution; (ii) fatty acid mg/g total lipid = (AX) (WIS) where AX: peak area of fatty acid, AIS: peak area of internal standard (IS) WIS: weight (mg) of IS added to the sample (mg), WX: sample weight (mg), Ri: empirical response factor, and 1.04: conversion factor necessary to express results in mg fatty acid/g oil (rather than as methyl ester).

Analysis of Method Validation
Linearity Test. ree fatty acid standards, myristic acid, heptadecanoic acid, and linoleic acid were prepared at different concentrations: myristic acid (0.4, 2.0, and 4.0 mg/mL), heptadecanoicacid (0.4, 2.0, and 4.0 mg/mL), linoleic acid (0.5, 1.0, and 5.0 mg/mL). ese fatty acid standards were used as they represent both saturated and unsaturated fatty acids present in samples. Calibration curves were formed for each of the compounds. e linearity parameters, which included linear regression ( = ) and the correlation coefficient ( 2 ) were obtained from the linear relationship between the peak area and the concentrations of the fatty acid standards.
Precision Test. Two precision tests of repeatability (withinday) and reproducibility (between-day) were performed. Two fatty acid standards were quanti�ed in three randomly selected samples in both tests. For repeatability (within-day precision), four replicates of each sample were analyzed in a single day using the same procedures as in the fatty acid analysis. Data are reported as relative standard deviations (RSD) of four replicates, with minimum and maximum values of both palmitic and linolenic acids quanti�ed in each sample. For reproducibility (between-day precision), samples were analyzed using the same procedures as in the fatty acid analysis on three different days, representing three replicates of each sample. Palmitic acid and linolenic acids were quanti�ed and presented as relative standard deviations (RSD) of three replicates, with minimum and maximum values of each fatty acid.
Recovery Test. e recovery of the method was determined using three different concentrations (0.4, 2.0, 4.0 mg/mL) of myristic acid, heptadecanoic acid, and stearic acid standards. Two samples were randomly selected and used for the recovery analysis of each of the fatty acid standards. During the sample preparation, each standard was added together with lipid extract and internal standards and prepared following the same procedures used in normal sample preparation. Recoveries of different standards were performed separately to avoid overlapping of peaks in chromatograms which could lead to biased results. Recoveries of fatty acid standards at different concentrations were determined by comparing the content of the fatty acids in samples with and without the addition of the standard. Data are presented as the percentage of recovery.
Statistical Analysis. Data were analyzed using SPSS (Scienti�c Package of Social Science) version 17.0. e mean, standard deviation (SD), and one-way ANOVA test followed by Tukey post-hoc analysis were performed to compare differences in the mean fatty acid contents of different species of �sh and shell�sh. Bivariate correlation (Pearson�s ) was performed to explore the relationships between different fatty acid classes in samples.

Results and Discussion
3.1. Fatty Acid Content in Samples. Data from this study are reported in the form of milligrams per 100 grams of wet muscle. Tables 2, 3, and 4 show the sample content of saturated, monounsaturated, and polyunsaturated fatty acids, respectively.
Among the low-fat �sh, yellowstripe scad contained a very high PUFA content, with signi�cantly higher ( ) levels of DHA (782.1 mg/100 g wet sample) compared to other samples (Table 4). e DHA amount was about 2.7 times higher than �apanese thread�n bream (2nd highest DHA, 293.0 mg/100 g wet sample); and about 86.6 times higher than long-tailed butter�y ray (lowest DHA, 9.0 mg/ 100 g wet sample). Additionally, yellowstripe scad also had a high ALA (C18:3n3) level at 338.1 mg/100 g wet sample.

3.�. Fatty Acids o� Fis� and S�ell�s� �rom �ocal and �t�er
Countries. Most of previous local studies have focused on qualitative aspects of fat content in various samples of marine and freshwater origin samples [10][11][12][13][14][15]. A previous study found very high percentage of PUFA; with omega-3 PUFA (29.7-48.4%), other PUFA (27.7-40.0%), and omega-6 PUFA (11.0-20.0%) in ten common �sh species with the current study [9]. Meanwhile, percentages of SFA and MUFA were quite low; at 3.63-11.4% and 1.37-9.12%, respectively. is study found hexadecadienoic acid as the most prominent fatty acid (18.1-24.9%) in six samples, hexadecatrienoic acid (C16:3n4) in two samples, and hexadecadienoic acid (C16:2) in other two samples [10]. Another local study reported PUFA as the dominant group of fatty acid in seven species, and SFA as the dominant fatty acid group in another six species of their �sh samples [11]. Palmitic acid (C16:0) was the most prominent fatty acid (17.6-32.1%) in eight samples; meanwhile, DHA (C22:6n3) was the highest fatty acid (19.3-24.0%) in another �ve samples [11]. Same �ndings were reported by other study of 16 species of local pelagic �sh; with palmitic acid (C16:0) as the most prominent fatty acid (20.9-34.5%) in nine species, meanwhile, DHA (C22:6n3) was the highest fatty acid (26.0-29.8%) in another seven species [12]. e current study, however, focuses on the quantitative aspect of samples fat thus allow limited comparison be made with these previous local data. However,    there were a few studies focused on the quantitative aspects of fatty acids in other cold water species of marine and freshwater �sh and shell�sh performed in other countries. e comparisons made with the previous �ndings would be useful in giving better overview of the content of fatty acids in local marine �sh and shell�sh.
Most samples in current study contained lower EPA (C20:5n3, 2.7-343.0 mg/100 g wet sample) compared to wild salmon (414 mg/100 g), farmed salmon (1079 mg/100 g), supermarket salmon (969 mg/100 g), and salmon feed (2638 mg/100 g) [16]. Only longtail shad was found to contain comparable amount of EPA with salmon feed; at 2638 mg/100 g wet sample [16]. It is really interesting to �nd such a high level of EPA in longtail shad, as high intake of this fatty acid had been related with protective effects to the occurrence of asthma, coronary heart problems and many other diseases [17][18][19]. Meanwhile, compared to DHA values (629-2633 mg/100 g) reported by the previous study [16], the content of this fatty acid in most samples of the   Categories of �sh samples were based on the fat content [5].
current study were lower (9.0-277.1 mg/100 g wet sample). However, yellowstripe scad (782.1 mg/100 g wet sample) was found to contain slightly higher DHA content compared to wild salmon (629 mg/100 g) [16]. Data from National Nutrient Data by United States Department of Agriculture (USDA) showed Greenland halibut, farmed cat�sh, wild cat�sh, farmed salmon, and wild salmon contained EPA + DHA at 1177.6, 177.6, 236.5, 2147.1, and 1840 mg/100 g muscles [20]. Generally, most samples in the current study contained EPA + DHA amounts (11.8-551.7 mg/100 g wet sample); which were lower compared to Greenland halibut, farmed salmon, and wild salmon; but comparable with farmed and wild cat�sh. However, longtail shad (2210 mg/100 g wet sample) was found to contain EPA + DHA at comparable amount with farmed and wild salmon [20]. is could be due to the high fat content of the �sh (23.2% fat) [5].
e current study is novel as it provides new quantitative �ndings for warm water species of �sh and shell�sh. Moreover, the �ndings are of necessary representativeness, which resulted from systematic sampling procedures performed. Data from the current study is also highly important as it represents samples from the Straits of Malacca, which aligns the west coast of Peninsular Malaysia, the main contributor of marine landings production in the country which produced 50.16% of the total marine landings production of Malaysia and 67.34% of the marine landings production of peninsular areas (Department of Fisheries Malaysia, 2007).

Ratios of Polyunsaturated/Saturated (P/F) and -3/ -6
Fatty Acids. Overall, seventeen species of samples contained SFA as the dominant group of fatty acid (Table 5), with palmitic acid (C16:0) as the highest fatty acid quanti�ed in most of samples (Table 2). e high amount of palmitic acid is due to its function as a key metabolite in �sh, and the level is not in�uenced by the diet [21]. Meanwhile, seven of the samples (golden snapper, Indian thread�n, malabar red snapper, long-tailed butter�y ray, large-scale tongue sole, yellowstripe scad, and prawn) contained PUFA as the dominant group of fatty acid. Generally, the PUFA to SFA (P/S) ratios of most of the samples of this study (Table 5) were above the value for Menhaden oil (0.58); as suggested by Food and Drug Association (FDA) as PUFA supplement [22]. Besides, ratios exceeding 0.50 have also been shown to lower blood cholesterol level [23].
Among the lean �sh, hardtail scad and Indian mackerel contained fairly high SFA but fairly low PUFA; which resulted in low P/S ratio of 0.6 and 0.3, respectively. Meanwhile, other lean �sh showed higher P/S ratios, between 0.6 and 1.4. For low-fat �sh, four-�nger thread�n showed low P/S ratio of 0.5. Other low-fat �sh showed P/S ratio between 0.6 and 1.7. �hile for the medium fat �sh, moon�sh contained quite high P/S ratio of 1.0. e high-fat �sh, longtail shad contained higher SFA content compared to PUFA that resulted in low P/S ratio of 0.4. Overall, most of samples contained favorable ratio of fatty acids, which was bene�cial for PUFA intake and lowering blood cholesterol [22,23]. ere is no speci�c trend in levels of saturation or unsaturation relative to the fat content of samples. On the other hand, a very high positive correlation was found between fat content of samples with all three classes of fatty acids; SFA (Pearson's , ), MUFA (Pearson's , ), and PUFA (Pearson's , ). is suggested that increment in fat content is followed by increment in the content of all forms of fatty acids: SFA, MUFA, and PUFA in the samples. e −3 : − ratio has been suggested as a useful indicator for comparing relative nutritional values of �sh oils. e ratio is also expressed in the form of − / − . It has been suggested that an − : − ratio of 1 : 1 to 1 : 5 (or − / − between 0.2 and 1.0) would constitute a healthy human diet [22]. High − : − ratio is preferred, as excess − fatty acids can counteract the health bene�ts of − fatty acid intake [24].
Overall, long-tailed butter�y ray and cuttle�sh had very high ratios of − / − , at 13.3 and 15.3, respectively (Table 5). However, the high ratio was not contributed to by high − fatty acids, but by the low content of − fatty acids. ese �sh contained 349.3 and 472.9 mg/100 g wet samples of − fatty acids, respectively, which were within the common range of other samples. Meanwhile, their − fatty acids contents were only 26.3 mg/100 g wet sample for long-tailed butter�y ray, and 30.9 mg/100 g wet sample for cuttle�sh. e same reason applied to the fairly high ratio of − / − at 9.8, 8.7, 7.8, and 7.3 for another four samples; namely, Indian thread�n, prawn, large-scale tongue sole, and Spanish mackerel. ese samples contained − fatty acids between 235.9 and 414.3 mg/100 g wet samples; meanwhile their − fatty acids were below than 40 mg/100 g wet sample. While for yellowstripe scad, the high ratio of − / − was contributed by its high content of − fatty acids (1224.8 mg/100 g wet sample) compared to − fatty acids (192.2 mg/100 g wet sample). In contrast, longtail shad contained high content of − fatty, resulted in fairly low ratio of −3 : − (0.8), which was the lowest among all samples.

Validation Procedures.
A few validation analyses were performed to ensure the reliability of the method used for fatty acid analysis in this study; these measures included linearity, precision (repeatability and reproducibility), and recovery tests. Linear relationships were observed for myristic acid (C14:0) standard (linearity parameters: − ; ), heptadecanoic acid (C17:0) standard (linearity parameters: − ; ), and linoleic acid (C18:2n6) standard (linearity parameters: ; ). e precision test assessed both repeatability (withinday) and reproducibility (between-day) tests. Table 6 shows the RSD values for the repeatability test of fatty acids; these values ranged between 1.4-3.6% for palmitic acid and 0.2-1.9% for linolenic acid. Meanwhile, Table 7 shows the RSD values for the reproducibility test, which ranged between 0.9-2.4% for palmitic acid, and 0.16-1.86% for linolenic acid. e RSD values in this study were lower compared to most RSD values (3.1-13.3%) as reported previously [25], but were higher when compared with excellent RSD values (<2.0%) reported by another study [26]. However, from overall, the RSD values in all repeatability and reproducibility tests were satisfactory and showed that the methods used were reliable to produce precise data for multiple determinations of fatty acids (saturated and unsaturated) in �sh and shell�sh samples, both in a single-day and multiple-day determinations. Meanwhile, the percentages of recoveries for different concentrations of myristic acid were between 90.3-105.6% in golden snapper and 106.8-119.4% in cockles (Table 8). Meanwhile for heptadecanoic acid, the percentages of recoveries were between 87.1-111.9% in golden snapper and 112.8-115.9% in cockles. For stearic acid, the range of 93.4− 108.8% in cockles. Overall, the recovery percentages were satisfactory, at about 90-120%, which is comparable to the values (most showed 80-120%) reported by previous studies [25,27]. is shows that the methods used in the analysis are highly accurate for determinations of fatty acid contents at various concentrations in �sh and shell�sh samples.

Conclusion
Overall, most of the marine �sh and shell�sh samples contained desirable compositions of fatty acids with a fairly high amount of − fatty acids, a suitable ratio of − / − fatty acids, and preferable P/S ratios which were higher than the level in the recommended PUFA supplement, Menhaden oil. ree samples were identi�ed as being outstanding in their desirable overall composition of fatty acids: yellowstripe scad (highest DHA, − − , P/S = 1.7), moon�sh (highest ALA, − − , P/S = 1.0), and longtail shad (highest EPA, − − , P/S = 0.4). ese �ndings showed that marine �sh and shell�sh from warm water area contain a good composition of fatty acids and could provide many health bene�ts if consumed regularly. ese reliable and representative data are very useful to develop a nutritional database of marine �sh and shell�sh from warm water area and as reference to people for intake locally and globally. Additionally, the �ndings also showed that some identi�ed species of �sh and shell�sh from this area may have possible value in terms of future manipulation for various nutraceutical purposes. However, further studies should be developed on this matter.