HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: D300017-JLR200v1 44/12/2406    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Siddiqui, N. Articles by Grootveld, M. Search for Related Content PubMed PubMed Citation Articles by Siddiqui, N. Articles by Grootveld, M. Social Bookmarking                      What's this?

Journal of Lipid Research, Vol. 44, 2406-2427, December 2003
Copyright © 2003 by American Society for Biochemistry and Molecular Biology

Methods

Multicomponent analysis of encapsulated marine oil supplements using high-resolution ¹H and ¹³C NMR techniques

Navaede Siddiqui^*, Julia Sim^*, Christopher J. L. Silwood^*, Harold Toms, Richard A. Iles^* and Martin Grootveld^1,

^* Department of Diabetes and Metabolic Medicine, Barts and the London, Queen Mary's School of Medicine and Dentistry, Whitechapel, London E1 1BB, United Kingdom
Department of Chemistry, Queen Mary University of London, Mile End Road, London E1 4NS, United Kingdom
School of Dentistry, The Queen's University of Belfast, Royal Victoria Hospital, Grosvenor Road, Belfast BT12 6BP, Northern Ireland

Published, JLR Papers in Press, September 1, 2003. DOI 10.1194/jlr.D300017-JLR200

¹ To whom correspondence should be addressed at Room AW518, Medical Unit, 5th Floor, Alexandra Wing, Royal London Hospital, London E1 1BB, United Kingdom. e-mail: m.grootveld{at}qmul.ac.uk

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Multicomponent high-resolution ¹H and ¹³C NMR analysis has been employed for the purpose of detecting and quantifying a wide range of fatty acids (as triacylglycerols or otherwise) in encapsulated marine cod liver oil supplements. The ¹H NMR technique provided quantitative data regarding the docosahexaenoic acid content of these products, which serves as a valuable index of fish oil quality, and a combination of both ¹H and ¹³C spectroscopies permitted the analysis of many further components therein, including sn-1 monoacylglycerols, sn-1,2 and -1,3 diacylglycerol adducts, together with a range of minor components, such as trans-fatty acids, free glycerol and cholesterol, and added vitamins A and E. The identities of each of the above agents were confirmed by the application of two-dimensional ¹H-¹H spectroscopies. The NMR techniques employed also uniquely permitted determinations of the content of nonacylglycerol forms of highly unsaturated (or other) fatty acids in these products (i.e., ethyl esters), and therefore served as a means of distinguishing "natural" sources of cod liver oils from those subjected to chemical modification to and/or supplementation with synthetic derivatives such as ethyl docosahexaenoate or eicosopentaenoate.

The analytical significance and putative health effects of the results acquired are discussed.

Supplementary key words docosahexaenoic acid • eicosopentaenoic acid • n-3 fatty acids • fatty acid ethyl esters • -tocopherol acetate • all-trans-retinol palmitate • regiospecific analysis

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
There are currently many reports available suggesting that the consumption of dietary n-3 (Omega-3) fatty acids are beneficial to human health, in particular their abilities to reduce cardiovascular mortality (1, 2) and putatively control the progression of a variety of inflammatory diseases (e.g., rheumatoid arthritis) (3). Marine oils rich in the highly unsaturated n-3 fatty acids, eicosopentaenoic acid [EPA, 20:5(n-3)], and docasahexaneoic acid [DHA, 22:6(n-3)] are an integral component of the traditional, predominantly carnivorous diet of Greenland Inuit (Eskimos) who display an unusually low incidence of ischemic heart disease and diabetes (4).

With respect to coronary heart disease, the protective actions of these highly unsaturated fatty acids (HUFAs) were originally believed to be ascribable to their roles in impairing platelet function (5), and two in vivo investigations have shown that the feeding of fish oils to animals markedly alleviated platelet accumulation at sites of arterial injury (6). Moreover, these n-3 fatty acids have been demonstrated to exert many striking effects on cellular function that indicate their anti-atherosclerotic, anti-inflammatory and antithrombotic properties (7–9), specifically the inhibition of leukotriene synthesis, protein kinase C activity, and cellular adhesion to surfaces, the stimulation of nitric oxide synthesis (a phenomenon causing vasodilation), and impairment of the production of inflammatory cytokines (e.g., interleukin-1 and tumor necrosis factor) by monocytes. Notwithstanding, further studies have shown that n-3 fatty acids also have a powerful antiarrhythmic property, which probably arises from the modulation of cellular calcium ion (Ca²⁺) flux during episodes of arrhythymogenic stress (10, 11). Indeed, the presence of these fatty acids in the plasma membrane of monocytes readily influences Ca²⁺ channel structure and/or function (12). It should also be noted that the antithrombotic and vasodilatory actions exerted by n-3 fatty acids may enhance local myocardial oxygen supply, a process that would synchronously diminish the possibility of arrhythmia.

More recent investigations have indicated the potential therapeutic applications of EPA and DHA in various chronic conditions, e.g., AIDS, Alzheimer's disease, and diabetes (13–15).

In view of the above possible health benefits offered by marine oils and diets containing high EPA and DHA contents, the market for these products has substantially and rapidly expanded in North America, Europe, and Japan. Such products include encapsulated fish oil supplements, many of which are enriched with both EPA and DHA and fortified with essential nutrients such as vitamins A, D, and E. Consequently, there has been a corresponding demand for analytical methods with the capacity to routinely monitor the nature and levels of n-3 fatty acids (as glycerides or alternative forms) in these products in order to ensure their quality control and quality maintenance. Methods for monitoring the contents of added vitamins and for detecting and quantifying additional low-level components such as cholesterol are also essential requirements for quality assurance purposes.

Although gas chromatographic (GC) methods have previously been extensively employed for the purpose of determining these HUFAs in both fish and fish oils (16, 17), these methods are labor intensive and time consuming and also involve a complex series of chemical manipulation stages that, together with high GC injector and oven temperatures, may give rise to their artifactual oxidation; in view of their high degree of unsaturation, both EPA and DHA are particularly susceptible to oxidative deterioration, a process generating conjugated hydroperoxydiene and hydroxydiene and, subsequently, aldehydic products, among others (18). More recently, the analysis of these HUFA components in marine oils has been augmented by the employment of high-resolution ¹H and ¹³C NMR spectroscopy (19).

High-resolution NMR spectroscopy offers many advantages over alternative analytical methods, because it allows the rapid, simultaneous, noninvasive, and nondestructive study of a plethora of components present in complex, multicomponent samples such as foods, pharmaceutical formulations and biofluids (20–22). The principles of this spectroscopic technique involve the subjection of samples to an intense magnetic field. In view of their inherent magnetic moment, ¹H or other nuclei in organic (or other) molecules align themselves either with (lower energy state) or against the field (higher energy state), the former state being slightly predominant. However, in the presence of energy from the radiofrequency (RF) region of the electromagnetic spectrum, the nuclei may orient themselves in the higher energy state; because the absorption frequencies of such nuclei are critically dependent on their chemical, and therefore magnetic, microenvironment, varying levels of radiofrequency energy must be applied to excite the different classes of sample nuclei to the higher energy state. Data acquired are presented as a spectrum consisting of a plot of signal intensity versus resonance frequency in parts per million (ppm, a dimensionless unit). The majority of ¹H nuclei present in organic and biological molecules resonate (i.e., absorb RF energy) within the narrow spectral range of 0–10 ppm. The lower and higher frequency regions of the spectrum (smaller and larger ppm respectively) contain signals for ¹H nuclei that require lesser and higher amounts of RF energy, respectively, for resonance. The reference signal located at 0.00 ppm arises from an organic solvent- or water-soluble tetra- or trimethylsilane derivative, respectively (either internal or external). Samples examined are placed in a 5-mm outer diameter NMR tube which is then inserted into the magnetic field and irradiated with radiofrequency energy encompassing the entire 10 ppm spectral range. An appropriate deuterated [²H] species [a deuterochloroform (C²HCl₃) solvent system in the case of culinary oil sample analysis] is added to provide a field frequency lock, i.e., prevent any modifications or drift in the magnetic field during spectral acquisition. The appearance (multiplicity) of an ¹H signal is influenced by adjacent ¹H nuclei in a well-characterized manner, and the intensity of each resonance detectable is directly proportional to the product of the number of magnetically equivalent nuclei in the structural/functional group giving rise to it and the concentration of the molecule containing that group in the sample examined.

High-resolution NMR analysis has been previously successfully employed for determinations of the DHA and total n-3 fatty acid contents of fish oils and further fish products (23–28). However, with the exception of a ¹³C NMR study focused on the positional distribution of DHA and EPA between the 1(3)- and 2-glycerol backbone positions of triacylglycerols (29), this technique has not been applied to the multicomponent analysis of encapsulated marine oil supplements fortified with essential vitamin nutrients. In the present study, we have employed high-field (9.4 and 14.1 T), high-resolution ¹H and ¹³C NMR spectroscopy to detect and/or determine the concentrations of a wide range of fatty acid and alternative components present in such products (specifically those based on cod liver oil), the latter including cholesterol, added dl--tocopherol acetate (vitamin E derivative) and all-trans-retinol palmitate (vitamin A). Included are methods for determining the DHA content (mg·g^-1 and mol·kg^-1), its molar proportion expressed relative to all fatty acids therein (mol%), together with the molar proportion of the total n-3 fatty acids (as glycerides or alternative esters) in the samples investigated. These NMR techniques have also been employed, either individually or in tandem, to provide valuable molecular information regarding the substitutional status and positional distribution of individual fatty acids on the glycerol backbone (both involving the identification of mono- and diacylglycerol adducts), and also to monitor any free and trans-fatty acids detectable in the products investigated. Furthermore, many of the ¹H NMR assignments were confirmed by the application of two-dimensional (2D) ¹H-¹H NMR techniques. We also demonstrate, for the first time, a novel ¹H NMR method for determining the relative proportions of naturally derived and synthetic (chemically modified) forms of these HUFAs in such marine oil supplements, the latter presumably generated via chemical manipulation in one of the products examined. The analytical significance and health aspects of the results obtained are discussed in detail.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Materials
Encapsulated marine (cod liver) oil supplements were purchased from local retail outlets. The first of these (product I) purportedly contained cod liver oil enriched with fish oil n-3 fatty acids (1,050 mg), the n-3 fatty acid content being 400 mg, of which EPA and DHA accounted for 360 mg, dl--tocopherol acetate (10 mg), vitamin A as all-trans-retinol palmitate (800 µg), and vitamin D as a D₃ preparation (5 µg); the content of the second product (product II) was described as cod liver oil (1,000 mg) containing EPA (120 mg), DHA (110 mg), a ‘natural’ source of vitamin E antioxidant compound (unspecified), vitamin A (800 µg), and vitamin D (5 µg).

-Tocopherol, -tocopherol acetate, and 1,3,5-trichlorobenzene (TCB) were obtained from Sigma-Aldrich Chemical Co. (Poole, Dorset, UK), and deuterated chloroform (C²HCl₃) was purchased from Goss Scientific Ltd. (Great Baddow, Essex, UK).

Proton ¹H NMR measurements and sample preparation
Proton ¹H NMR measurements on the the above cod liver oil samples (encapsulated) were conducted on Bruker AMX-400 and AMX-600 spectrometers (Queen Mary University of London facilities, London, UK) operating at frequencies of 400.13 and 600.13 MHz, respectively, and a probe temperature of 298 K. Typically, 200–250 mg quantities of the marine oil preparations were removed from their capsules, accurately weighed, directly dissolved in 0.50 ml of C²HCl₃ containing 5.00 x 10^-3 mol·dm^-3 TCB, thoroughly rotamixed, and then transferred to 5 mm-diameter NMR tubes. The TCB served as a chemical shift reference and an internal quantitative NMR standard (singlet resonance located at 7.227 ppm in C²HCl₃) solution. This TCB solution was standardized spectrophotometrically against independently prepared solutions in the same medium ( = 1.71 x 10² M^-1·cm^-1 at a wavelength of 273 nm).

Deuterated methanol (d₄-MeOH) extracts of these marine oil products were obtained by adding 0.75 ml of this NMR solvent (containing 1.153 x 10^-2 mol·dm^-3 ethanol added as an internal chemical shift reference and quantitative NMR standard) to accurately weighed samples of the oils (250 mg), thoroughly rotamixing the mixture, centrifugation (30 min, 500 g), and removal of the upper, d₄-MeOH phase for ¹H NMR analysis.

Typical pulsing conditions for ¹H NMR experiments were: 128 or 256 free induction decays (FIDs) using 32,768 or 65,536 data points, 72° pulses, a relaxation delay of 2.00 s, and an acquisition time of 1.28 s. The spectral width was 4,831 or 7,246 Hz. Exponential line-broadening functions of 0.20 Hz were routinely employed for purposes of processing. 2D shift-correlated (COSY) spectra of marine oil formulations were acquired on the above 400 or 600 MHz spectrometers using the standard sequence of Au, Bartholdi, and Ernst (30), with 2,048 data points in the t₂ dimension, 512 increments of t₁, a relaxation delay of 2.00 s, and 48 transients. ¹H-¹H total correlation (TOCSY) spectra were recorded using the RD-(90°-t₁-spin lock)-ACQ pulse sequence (31), 90° on the AMX-600 spectrometer being equivalent to 8.4 µs for these experiments. The spin lock employed the MLEV-17 sequence (32), with a typical mixing time of 70 ms. Acquisition parameters were: 128 t₁ increments, each of magnitude 2,048 data points; spectral width 12,019 Hz in each dimension; 64 transients in each case; four dummy scans; relaxation delay 3.0 s; acquisition time 0.17 s. Sine-bell-squared window functions shifted by /2 (i.e., a cosine² function) were employed in each dimension, with zero-filling twice in f₁, prior to transformation of the 2,048 x 1,024 data point-sized matrix. The standard phase-sensitive time-proportional phase incrementation method (33) for optimizing detection along the second dimension was utilized.

Chemical shifts were referenced to TCB (7.227 ppm) and/or residual chloroform (7.262 ppm). Resonances present in the ¹H NMR spectra of marine oil products were assigned by a consideration of chemical shift values, coupling patterns, and coupling constants, with special reference to previously conducted ¹H NMR studies of n-3 fatty acids present in fish oils (24, 25, 28). The relative intensities of signals were determined by electronic integration, and the concentrations of selected components were computed by comparing their resonance areas with that of the added TCB. The DHA content in mg·g^-1 (glycerol-bound or otherwise) was determined by measurement of the combined intensities of its C2,3 (F,F_ß) -CH₂- group multiplet signals at 2.38 ppm and comparison of these with that of the TCB internal standard [Equation 1, where A_DHA and A_TCB represent the signal areas of DHA's F,F_ß methylene group protons and TCB, respectively, H_DHA and H_TCB the number of protons giving rise to the combined DHA C2,C3 methylene resonances (4) and that of TCB (3), respectively, M_DHA and M_TCB the molecular masses of DHA (328.50) and TCB (181.45), respectively, and x and y the weights of internal standard (in mg) and oil sample (in g), respectively, present in the NMR tube]. Comparisons of the intensities of these isolated DHA resonances (4 ¹H nuclei) with those of the C2 (F) methylene group signals arising from all other fatty acid species in the samples examined (2.28 ppm, 2 ¹H nuclei) (23) yielded the DHA content as a percentage of all fatty acid components present (Equation 2, where A_FA represents the signal area of the F methylene group protons of all fatty acids other than DHA). The total n-3 fatty acid molar percentages of the samples examined were concomitantly determined via direct comparisons of the intensities of their terminal-CH₃ (F) group resonance(s) in the 0.90–0.98 ppm region (triplet) with those of the combined n-6, n-9, and saturated fatty acid triplet signal(s) located in the 0.81–0.89 ppm region (Equation 3, where A_n-3 denotes the signal area of the F-CH₃ group protons of n-3 fatty acids, and A_non-n-3 that of the F-CH₃ group of all non-n-3 fatty acids).

The -tocopherol acetate content of fish oil supplements (in mg·g^-1 units) was determined by the method of standard additions. Briefly, 1.00 g quantities of oil were accurately weighed and then treated with accurately weighed amounts of this antioxidant (10 mg, approximately equivalent to the content specified by the manufacturers of product I). Subsequently, accurately weighed quantities (0.25 g) of these -tocopherol acetate-"spiked" oil samples were added to 0.50 ml aliquots of C²HCl₃ containing 5.00 x 10^-3 mol·dm^-3 TCB and then transferred to 5 mm-diameter NMR tubes after thorough rotamixing as described above. The increase in the intensity of its heterocyclic ring 4 position -CH₂- group signal (t, = 2.555 ppm) expressed relative to that of the TCB internal standard was monitored via electronic integration. Each determination was conducted in duplicate.

The free (unesterified) cholesterol content of the fish oil formulations was estimated by integration of its C18-position -CH₃ group resonance (s, = 0.655 ppm) and comparison of its intensity with that of the TCB internal standard.

¹³C NMR measurements
¹³C NMR spectra were obtained on the Bruker AMX-600 MHz facility (at an operating frequency of 150.93 MHz). The 10,240 FIDs were acquired, each of magnitude 65,536 data points, a sweep width of 26,316 Hz, 8.00 µs pulses, a 10.0 s spin relaxation delay between pulses, and an acquisition time of 1.245 s. Sensitivity enhancement was achieved via the employment of a 5.00 Hz line-broadening function, followed by the zero-filling of FIDs to the above number of data points.

Application of computational methods to confirm ¹H NMR assignments
¹H NMR assignments and appropriate chemical shift values and coupling constants of resonances of selected marine oil components with complex spin systems (e.g., all-trans-retinol palmitate) were confirmed with the HNMR prediction software (version 3.0) from the Advanced Chemistry Development (ACD Inc., Toronto, Canada) software suite. This software was utilized to generate a theoretical spectrum that did not incorporate an allowance for solvent-mediated shielding effects, and the ¹H chemical shift and ¹H-¹H spin-spin coupling constant values obtained therefrom were then adjusted to reflect those derived from the experimental spectrum via a first-order analysis. The theoretical spectrum was then recalculated; a correct first-order analysis gave rise to an exact match between theoretically and experimentally derived spectra. The calculations conducted also made an allowance for the concentrations of each spectrally simulated component in fish oil samples.

Typically, the ¹H simulations were based on an internal database containing NMR data for >81K experimental ¹H spectra, the associated algorithms employing intramolecular interaction parameters for >300 structural fragments, and the associated subalgorithms estimating initial values for unique structural fragments. The compilation of fragment data were handled with a modified Hierarchical Organization of Shells Expert code, which allowed for explicit substituent charge and stereo bond conventions, optimizing to the maximal number of spheres. Subsequent quantum mechanical shielding calculations allowed for the number of these codes found in the internal database search, in addition to the number of those sought. Calculational errors were determined as the standard deviations of the experimental values found within the database, and typically ¹H chemical shift and ¹H-¹H coupling constant values were accurate to within 0.05 ppm and 0.2 Hz, respectively.

Statistical analysis
All data (concentrations in mg·g^-1, mmol·kg^-1 or mol%) are expressed as mean value ± standard error (SE). Suspected extreme values ("outliers") in each group of repeated determinations made were tested for by computation of Grubbs' (extreme value - overall mean)/overall standard deviation ratio (34) (for each data set) and the value discarded if the probability of observing this value was <0.05.

Nomenclature
Fatty acid chains (F) are represented by the conventional notation, e.g., 22:6(n-3), which denotes a 22-carbon chain with six carbon-carbon double bonds, the first of these located three atoms distant from the the terminal-CH₃ group (i.e., at carbon number 20). The subsequent double bonds (abbreviated as ) are each separated from those preceding by one -CH₂- group (bis-allylic-CH₂'s, abbreviated as -CH₂-). F and F_ß correspond to the C2 and C3 methylene group protons, while F and F_-1 represent those of the terminal-CH₃ and directly adjacent -CH₂- groups, respectively (35).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
1D . ¹H NMR and 2D ¹H-¹H NMR analysis of fatty acid components present in marine oil supplements
A typical 600 MHz ¹H NMR spectrum of a commercially available cod liver oil supplement (product I), accompanied by expansions of its 0.80–1.00 and 2.20–2.50 ppm regions, is shown in Fig. 1 . The F and F_ß methylene (-CH₂-) groups (C2 and C3 positions, respectively) located between the deshielding carboxylic acid function {esterified to glycerol [1(3)- or 2-positions] or otherwise} and the C4–C5 position double bond of DHA exclusively give rise to ¹H NMR signals downfield of and well resolved from those arising from the C2 (F) -CH₂- groups of all the other (non-DHA) fatty acids (Fig. 1), and this phenomenon readily permits the quantification of DHA in the fish oil supplements examined here, both as a molar proportion (mol%) of all fatty acids present and as mol·kg^-1 or mg·g^-1 units as outlined in the Materials and Methods section.

View larger version (16K): [in this window] [in a new window]   Fig. 1. Partial 600 MHz 1H NMR spectrum of an encapsulated cod liver oil supplement (product I, CDCl3 solution). The spectra exhibit clear resolution between 1) C2 methylene group resonances (expanded 2.20–2.45 ppm region) of docasahexaenoic acid (DHA), C2,C3, and all other fatty acids; and 2) those of the terminal-CH3 groups (expanded 0.80–1.05 ppm region) of n-3 and all other fatty acids. A typical spectrum is shown. TCB, 1,3,5-trichlorobenzene.

Visual inspection of further fatty acid chain resonances in the spectra acquired permitted the complete or partial distinction between different classes of these species. Figure 2 displays the expanded 1.50–1.85 (F_ß), 1.90–2.20 (-1), and 2.70–3.00 ppm (-CH₂-) regions of a typical ¹H marine oil supplement sample spectrum, and each of these contained two or more classes of resonances. The F_ß -CH₂- group region has clearly resolved multiplet resonances centered at 1.59 and 1.69 ppm, the former predominantly arising from 16:0 and 18:1(n-9) and the latter from 20:5(n-3), because only this fatty acid and 20:4 (n-6) (arachidonate) have a signal at this chemical shift value and there are only low levels of the latter in marine oils [i.e., 1.3–2.7 mol% (28)]. For the pattern of resonances in the -1 region, that at 2.00 ppm corresponds to 18:1(n-9) and that at 2.08 ppm to all n-3 fatty acids, predominantly 20:5(n-3), 22:6(n-3), and, to a much lesser extent, 20:4(n-3); contributions from 18:2(n-6) (2.06 ppm), 18:3(n-3) (2.07 ppm), and the C16 (-1) -CH₂- group signal of 20:4(n-6) (2.06 ppm) (37) are much less discernable in view of their markedly lower concentrations [0.6–2.3, 0.4–0.9, and 1.3–2.7 mol%, respectively, in a range of marine oils (28)]. The C4 (-1) resonance of 20:4(n-6) ( = 2.13 ppm), which distinguishes this species from all other unsaturated fatty acids, was not detectable in the spectra acquired, an observation also consistent with its low content in the fish oil supplements examined here. As expected, all linoleate [18:2(n-6)] signals were of a very low intensity when expressed relative to those arising from 16:0, 18:1(n-9), 20:5(n-3), and 22:6(n-3) fatty acid components that are reported to be present at levels of 11–26, 9–22, 3–9, and 20–48 mol%, respectively, in various fish oils (28). Indeed, particularly notable is the very low intensity of its -CH₂- group resonance at 2.738 ppm.

View larger version (14K): [in this window] [in a new window]   Fig. 2. Expanded Fß, -1, and -CH2- regions of a 600 MHz 1H NMR spectrum of a cod liver oil supplement (product I). Expanded Fß (1.50–1.85 ppm) (A), -1 (1.90–2.20 ppm) (B), and -CH2- (2.70–3.00 ppm) (C) regions. A typical spectrum is shown.

For one of the products investigated (product I), a further, intense connectivity between resonances located at 1.21 (apparent triplet) and 4.09 ppm (apparent quartet) were visible. These linked signals are clearly attributable to an ethyl group in an ester derivative (38), i.e., the manufacturers of this particular product have either chemically modified the cod liver oil supplement or added DHA and EPA in the forms of ethyl docosahexaenoate and ethyl eicosopentenoate, respectively. Closer examination of the 1D ¹H NMR spectra of this product in the 1.15–1.25 ppm region (Fig. 3) revealed that the signal at 1.21 ppm consisted of two separate triplets of very similar but nevertheless distinguishable chemical shift values ( = 1.211 and 1.215 ppm, j = 7.36 Hz in each case). In view of the molecular structures of the ethyl ester derivatives of DHA and EPA [i.e., a closer proximity of nearest >C=C< double bond in DHA (C4 position) to the -CH₃ moiety of the ethyl group, together with a larger number of double bonds in this fatty acid when considered as a whole], the separate 1.211 and 1.215 ppm -CH₃ group resonances are tentatively assigned to the ethyl esters of EPA and DHA, respectively, although, as outlined in detail below, it is important to note that ¹H NMR analysis of d₄-MeOH extracts of this product revealed that 30% of these ethyl ester-CH₃ group resonances are accountable for by non-n-3 fatty acid species. The intensities of these two signals are very similar, suggesting that similar molar levels of each HUFA ester have been generated from or added to the product, although the line-width of that assigned to DHA ethyl ester appeared to be somewhat broader than that assigned to the corresponding EPA ester. The DHA and EPA ethyl ester resonances were not detectable in the 600 MHz ¹H NMR spectra acquired on product II, demonstrating that the manufacturers have not chemically modified their cod liver oil supplement or added synthetic DHA or EPA derivatives to it; i.e., all of the HUFAs therein are present as natural cod liver oil acylglycerol species (and/or those arising from any refinement process involved in their production). Electronic integration of these combined signals [the middle and higher field peaks of the triplets only and then multiplication by the factor 4/3, in view of the significant overlap of the lower field triplet peaks with the bulk chain (-CH₂-)_n resonance centered at 1.23 ppm], together with that representing the total fatty acids in spectra of product I [the combined intensities of the n-3 (t, = 0.940 ppm) and non-n-3 (t, = 0.846 ppm) signals], allows an estimation of the molar ratio (and hence percentage) of the synthetic ethyl ester derivatives of these highly unsaturated and other fatty acids to the total fatty acid content of this formulation. Using this method, the mean ± SE percentage of added ethyl esters (predominantly those of DHA and EPA) generated in or added to the cod liver oil in product I was found to be 62.57 ± 1.07 mol% of the total fatty acid content (n = 10), a figure higher than the total n-3 fatty acid content in this product (Table 1), demonstrating that non-n-3 fatty acids are also present as ethyl esters in this formulation. As expected, the ethyl group -CH₂OCOR signal centered at 4.09 ppm was also found to consist of two separate quartets, although, in view of its overlap with one of the triacylglycerol glycerol backbone-CH₂OCOR resonances located at 4.10 ppm, electronic integration of these was not possible. The differences between the chemical shift values of the F,F_ß ¹H NMR resonances of the ethyl ester and glycerol-bound forms of DHA are expected to be extremely small and hence to exert no influence on the estimations of this HUFA made here. Indeed, the chemical shift range for these signals in methyl docosahexaenoate (2.31–2.38 ppm) (38) are very similar to those of the docosahexaenoylglycerol species present in the ¹H spectra of product II, which does not contain the ethyl ester.

View larger version (9K): [in this window] [in a new window]   Fig. 3. The 0.89–1.27 ppm region of the 600 MHz 1H NMR spectrum of product I (CDCl3 solution). The spectrum contains resonances arising from the -CH3 moiety of the ethyl group of fatty acid ethyl esters (predominantly n-3 HUFAs). A typical spectrum is shown. -CO.O.CH2CH3, fatty acid ethyl esters' -CH3 groups.

View larger version (22K): [in this window] [in a new window]   Fig. 4. The 600 MHz NMR spectra of d4-methanol (d4-MeOH) extracts of encapsulated marine oil products. A, B: 0.80–5.45 ppm region of the 600 MHz NMR spectra of a d4-MeOH extract of products I and II, respectively; expanded 0.61–1.30 ppm regions of the partial spectra (A) and (B) are shown in C and D, respectively. EtOH-CH3, ethanol-CH3.

View larger version (43K): [in this window] [in a new window]   Fig. 5. The 600 MHz 1H-1H TOCSY spectrum of a d4-methanol extract of product II. A typical spectrum is shown.

¹H-¹H TOCSY spectra of d₄-MeOH extracts of the products examined showed strong connectivities between the lipid chain resonances. Indeed, the vinylic ¹H signals (5.4 ppm) were, as expected, linked to the -CH₂-, F, -1, and F_ß signals, and, with respect to n-3 and non-n-3 fatty acids, the F triplets were located at 0.92 and 0.78 ppm, respectively, in this medium. These spectra also showed clear connectivities between the -CH₂-, F_, and F_ß resonances of both DHA and EPA (the latter two signals centered at 2.378 ppm for DHA in this solvent system) and also links between these and the C4 (-1) resonance of EPA ( = 2.13 ppm). The identity and relative levels of many further, methanol-extractable components detectable in these spectra are outlined in detail below.

¹H NMR analysis also permitted the direct and rapid detection of diacylglycerols, together with the acquisition of data regarding the relative quantities of the sn-1,2- and sn-1,3-forms of these adducts, which arise from the incomplete biosynthesis of triacylglycerols, a limited activity of the lipase enzyme system, and/or lipolysis of triacylglycerols during processing, refinement, and/or storage of the products investigated. Indeed, sn-1,2-diacylglycerols were readily monitorable in the two marine oil formulations tested here by observation of a clearly resolved -CH₂OH, sn-3 doublet resonance located at 3.66 ppm (Fig. 6) . Further sn-1,2-diacylglycerol signals detectable included those at 4.17, 4.31 (both -CH₂OCOR, sn-1), and 5.07 ppm (-CHOCOR, sn-2). Although some overlap between triacylglycerols and sn-1,3-diacylglycerol resonances is apparent in their glyceridic spectral regions, the latter's 4.03 ppm -CH₂OCOR, sn-1,3 multiplet signal was observable at an operating frequency of 600 MHz. 1D ¹H NMR and 2D ¹H-¹H NMR spectra of the d₄-MeOH extracts acquired on both products demonstrated that relatively high levels of sn-1,2 and -1,3 diacylglycerols were present therein (Figs. 4, 5). Indeed, the TOCSY spectra showed clear connectivities between the sn-1,2 signals located at 3.62 (-CH₂OH, sn-3), 4.08, 4.29 (-CH₂OCOR, sn-1), and 5.01 ppm (-CHOCOR, sn-2) and the sn-1,3 resonances at 3.96 (-CH₂OCOR, sn-1,3) and 4.02 ppm (-CHOH, sn-2) in this medium. Extraction of these cod liver oil products with methanol therefore serves as a partially selective means of isolating sn-1,2- and sn-1,3-diacylglycerols from triacylglycerols therein, insofar as these partial acylglycerols are much more soluble in this solvent system.

View larger version (15K): [in this window] [in a new window]   Fig. 6. The 600 MHz 1H NMR spectra of encapsulated marine oil products in CDCl3 solution. The 3.50–5.40 ppm (glyceryl proton) regions of the 600 MHz 1H NMR spectra of (A) product I and (B) product II (typical spectra are shown). The -CHOH sn-2, sn-1,3-diacylglycerol signal ( = 4.07 ppm) is obscured by the relatively intense triacylglycerol resonance located at 4.10 ppm.

1D . ¹H NMR and 2D ¹H-¹H NMR analysis of minor (low-level) components in marine oil supplements
The ¹H NMR technique also allowed the direct, noninvasive detection of components present at only low concentrations in the marine oil supplements examined. Figures 7 and 8 display the expanded 0.00–7.00, 0.64–0.70, 2.30–2.70, and 5.60–6.80 ppm regions of the 600 MHz 1D ¹H spectra of the products investigated, which highlight selected minor component resonances that are either clearly or partially resolved from those of acylglycerol and/or alternative lipidic components present. These signals consist of those attributable to: 1) -tocopherol acetate [C7- and C8-position aromatic methyl substituents and the heterocyclic ring C3- and C4-position methylene groups (the latter -CH₂- group triplet being completely resolved)]; 2) retinol palmitate (C9-, C10-, C12-, C13-, C14,- and C16-position chain vinylic protons in the 5.70–6.65 ppm chemical shift range, together with those arising from its C8- and C7-position methyl groups at 0.978 and 0.994 ppm, respectively); and 3) free (non-esterified) cholesterol (very clearly resolved C18-methyl group singlet at 0.65 ppm, C26,27- and C19-methyl substituents, together with its H₃ proton multiplet). The identity of the form of -tocopherol as its acetate ester in product II (unspecified by the manufacturers) was confirmed by comparisons of accurately determined ¹H chemical shift values of visible (resolved) resonances with those of an authentic standard in C²HCl₃ solution [as expected, the phenolic-OH group resonance of -tocopherol itself ( = 4.63 ppm in C²HCl₃ solution) was absent from spectra of either product].

View larger version (30K): [in this window] [in a new window]   Fig. 7. 1H NMR detection of -tocopherol acetate, all-trans-retinol palmitate, and free cholesterol in product I (CDCl3 solution). A–D: The 0.00–7.00, and expanded 0.64–0.70, 2.30–2.70, and 5.60–6.80 ppm regions of the 600 MHz 1D 1H NMR spectrum of product I. E: Computer-simulated spectrum of all-trans-retinol palmitate generated from the software suite described in Materials and Methods. CD, conjugated diene region signals ascribable to one or more lipid oxidation products (LOPs), conceivably conjugated hydroxydienes; Ret, all-trans-retinol palmitate; -TocA, -tocopherol acetate; Chol, cholesterol.

View larger version (53K): [in this window] [in a new window]   Fig. 8. The 400 MHz 1H-1H COSY spectrum of product I. A typical spectrum is shown. C12-H, C13-H, and C14-H, all-trans-retinol palmitate C12-, C13-, and C14-position protons.

Our spectral prediction software (ACD, Inc.) served as a valuable means of facilitating and confirming our preliminary assignments for resonances of the retinol moiety of the added retinol palmitate; Fig. 7 shows the excellent agreement observed between such simulated and experimentally acquired spectra. Of course, resonances assignable to the palmitate moiety of this ester were masked by those of the high-level acylglycerol components in the products tested. 2D ¹H-¹H COSY spectra of the products confirmed clear connectivities between the chain vinylic proton resonances of the retinol portion of retinol palmitate; as an example, Fig. 8 displays a typical partial COSY spectrum that demonstrates linkages between the C13- ( = 6.526 ppm), C12- ( = 6.046 ppm), and C14-position ( = 6.167 ppm) ¹H signals.

After consideration of the number of protons giving rise to the ¹H NMR signals of -tocopherol acetate and retinol palmitate, the former were of a greater intensity than those of the latter, as expected from the relative amounts of these vitamin adducts added to each supplement by their manufacturers (see Materials and Methods). Using the standard addition method outlined in Materials and Methods, we found that adding 10 mg·g^-1 of -tocopherol acetate to product I doubled the intensity of its clearly visible 4-position -CH₂- group resonance, (subsequent to making allowances for very small differences in the between-replicate weights of -tocopherol acetate added to this product and minor differences between those of the fish oil itself) a result consistent with its specified content (10 mg·1.050 g, or 9.52 mg·g^-1). However, for product II, these increases were much greater, and the estimated -tocopherol acetate content of this formulation was 3.33 ± 0.07 mg·g^-1, i.e., approximately one-third that of product I.

In view of the very low content of vitamin A in products I and II (800 µg per 1.05 and 1.00 g respectively, and specified as all-trans-retinol palmitate in the former), and also of some overlap between its visible resonances and those of alternative agents [e.g., that involving its C13-position vinylic proton (dd, = 6.58 ppm)], only semiquantitative content data were provided by determinations of their intensities. As expected, vitamin D (specified as a D₃ preparation in product I) was not detectable by ¹H NMR spectroscopy.

Intriguingly, strong connectivities between two further resonances located at 6.38 (apparent dd) and 5.95 ppm (multiplicity unclear) were also discernable in these COSY spectra. As previously reported by our laboratory (39), these signals are in the spectral region that is characteristic of conjugated diene lipid oxidation products (LOPs), and careful consideration of their chemical shift values, and the coupling pattern of that at 6.38 ppm, indicates that they are attributable to one or more conjugated hydroxydiene species [e.g., 13-hydroxy-9-cis,11-trans-octadecadienoate derived from the peroxidation of the low levels of 18:2(n-6) present], the determination of which is known to be of practical value in the monitoring of fish and fish oil quality (40). Experiments to determine the nature and content of such LOPs in cod liver oil supplements are currently in progress in our laboratory, the results of which will be presented elsewhere.

1D spectra acquired on deuterated methanol (d₄-MeOH) extracts of the fish oil products also confirmed the presence of cholesterol in these materials (Figs. 4, 5). Because the C19-position methyl group resonances for free and esterified cholesterol are readily distinguishable in this medium (41), it was possible to estimate the relative amounts of each component. Consistent with the ¹H spectra obtained in the C²HCl₃ solution and also the ¹³C NMR spectra detailed below, cholesterol was predominantly in the form of its free (unesterified) agent, with little or no cholesterol esters present.

¹³C NMR analysis of marine oil supplements
A characteristic of the ¹³C NMR spectra of DHA, EPA, their corresponding glycerides and other esters is the large number of vinylic and allylic carbon resonances present [e.g., DHA has 22 carbon atoms, of which 12 are vinylic and 7 are allylic, with only C1, C2, and C22 (the latter being F) remaining], and the spectra acquired on the cod liver oil products examined reflected the high levels of these HUFAs present.

Reference to the extensive chemical shift data tabulated by Gunstone (43, 44), together with those of Sacchi et al. (45) (for both free fatty acids and their glyceridic and methyl esters) enabled the complete or partial distinction between individual fatty acids, the chemical nature of their esters (glycerides and ethyl esters of DHA and EPA), and the substitutional status of the glycerol backbone, together with the positional distribution of fatty acids on this moiety. The C1 (carboxyl) and vinylic carbon positions of triacylglycerols display small but significant chemical shift differences in relation to both the position of the fatty acid moiety on the glycerol backbone [i.e., sn-1(3) or -2 positions] and the positions and number of carbon-carbon double bonds in the fatty acid chain. The expanded carboxylate and vinylic carbon regions of typical spectra acquired on products I and II are displayed in Figs. 9 and 10 , respectively.

View larger version (15K): [in this window] [in a new window]   Fig. 9. Acylglycerol and ethyl ester C1 (carboxylate) region of the 150.93 MHz 13C NMR spectra of products I and II (CDCl3 solution) are shown in A and B, respectively. (Chemical shift assignments for the labeled resonances are given in Table 2.)

View larger version (24K): [in this window] [in a new window]   Fig. 10. Olefinic region of the 150.93 MHz 13C NMR spectra of products I and II (CDCl3 solution) are shown in A and B, respectively. (Chemical shift assignments for the labeled signals are given in Table 3.)

The vinylic regions of these spectra also provided valuable information concerning the nature of fatty acids present in the products tested, notably DHA and EPA (Table 3). In view of the large number of >C=C< bonds in these HUFAs, both DHA and EPA have a very large number of ¹³C signals in this region. Acylglycerol oleoyl, DHA, EPA, and further fatty acids were readily assignable via comparisons with previously recorded data, and product I also contained two signals representative of simple alkyl esters of two (or more) of these fatty acids [C4 position resonances located at = 129.19 and 129.33 ppm, cf. that of methyl docosahexaenoate, 129.26 ppm (38)]. Indeed, well resolved C4,C5 and C5,C6 resonances of triacylglycerol DHA and EPA, respectively, were particularly notable, together with those arising from the C9 and C10 nuclei of oleoyl- [18:1(n-9)], and the C11 and C12 atoms of eicosenoylglycerols. Regiospecific (glycerol backbone position) chemical shift differences were detectable for selected DHA resonances (e.g., that of C4), but there was an incomplete resolution of these for the two EPA signals observed. Comparisons of the relative intensities of these sn-1(3)- and sn-(2)-glycerol backbone position DHA resonances confirmed that the latter contained a disproportionately high level of this HUFA (consistent with results attained from the carboxylate spectral region). The signal located at 131.92 ppm serves as a criterion for monitoring all the n-3 fatty acids present. Together with all of these resonances, ¹³C NMR spectra acquired on product I also contained a DHA C4-position signal located at 129.14 ppm, which, by reference to previously published ¹³C NMR data on the methyl esters of oleic acid, DHA, EPA, and further n-3 fatty acids in this spectral region, serves as a clear marker of synthetic ethyl esters present in this formulation, but nevertheless clearly distinguishes this product from the unadulterated product II.

Hence, the high level of distinction between the highly unsaturated and all other fatty acids in the products tested demonstrates the usefulness of this ¹³C spectral region for the study of natural, refined, or chemically transformed or adulterated marine oil preparations.

The allylic (-1 and -CH₂-) methylene group ¹³C signals, which also offer a diagnostic capacity, were also examined in detail (Fig. 11A) . The chemical shift values of these carbon atoms are critically dependent on whether they are 1) mid-chain or near to the C1-position carboxyl or terminal-CH₃ group (F), and 2) adjacent to one or two double bonds. Moreover, the configuration of the adjacent vinylic groups (i.e., cis- or trans-) also exerts a major influence on the allylic carbon atoms' chemical shift value. These structure-specific allylic resonances served to detect either EPA alone (C4 group signal at 26.65 ppm) or a combination of DHA and EPA (mid-chain -CH₂- group carbons in the 25.5–26.1 chemical shift range). Although three or more F resonances were present in the spectra acquired on product I (14.2 ppm), one of these is clearly attributable to the -OCH₂CH₃ group of the ethyl esters of DHA, EPA, and other fatty acids, including those of the n-6 and n-9 classes (Fig. 2). The F_-1 (-1) resonances of DHA, EPA, and stearidonate [20:4(n-3)] (20.7 ppm) are clearly distinguishable from those of all other fatty acids present. Contributions toward this signal from linolenoylglycerols and all other n-3 fatty acids present are not considered problematic, in view of their very low contents in marine oils, [e.g., 0.4–0.9 mol% for linolenate (28)]. For product I, the intensity of this signal was approximately two-thirds that of the F_-1 resonances of all other fatty acids (at 22.7 ppm), consistent with our ¹H NMR determination of the mol% of n-3 fatty acids in this supplement (40%, Table 1).

View larger version (41K): [in this window] [in a new window]   Fig. 11. 150.93 MHz 13C NMR spectrum of product I (CDCl3 solution). A: The 10–40 ppm region of the 13C NMR spectra acquired on product I. B: Expanded 28.80–35.30 ppm region of the former spectrum. Typical spectra are shown. AA, arachadonic acid [20:4(n-6)]; trans-FA, trans-fatty acid; OA, oleic acid; SA, saturated fatty acid; SDA, stearidonic acid; EPA, eicosapentaenoic acid; LA, linoleic acid.

The C3 position signals were able to distinguish EPA from stearidonic acid (Fig. 11A), and two further (unassigned) resonances were located at 24.73 and 24.85 ppm, in agreement with the ¹³C NMR data of Gunstone and Seth (29).

Chemical shift values of and assignments for resonances present in the glyceryl carbon region of the ¹³C spectra acquired are given in Table 4. Consistent with our ¹H NMR data (Fig. 6), ¹³C NMR spectroscopy confirmed that these particular marine oil supplements contain several glyceridic ester species, in addition to the predominant triacylglycerols, specifically sn-1-monoacylglycerols, and sn-1,2- and sn-1,3-diacylglycerols. As expected, the intensities of the two resolved sn-1,2-diacylglycerol signals were equivalent, as were those of the three separate sn-1-monoacylglycerol resonances; the intensity of the sn-1,3-diacylglycerol glyceryl-CH₂ signal was twice that of its glyceryl-CH resonance. The sn-1-mono-, sn-1,2-di- and sn-1,3-diacylglycerol content of one of the marine oil supplements (product I) was found to be 41%, 11%, and 48%, respectively, of the total nontriacylglycerol glycerides, and 2.7%, 0.7%, and 3.2% of the total glycerides present.

View this table: [in this window] [in a new window]   TABLE 4. 13C chemical shift values () of glycerol carbons in different classes of glycerides detectable in marine oil supplements

In view of the large amounts of the ethyl esters of DHA and EPA detectable in product I (65% of the total fatty acid content), its ¹³C NMR spectrum contained intense resonances at 60.06 and 60.25 ppm assignable to the ethyl group CH₃CH₂OCOR atom of each species (together with that at 14.27 ppm for their terminal-CH₃ group carbons).

The ¹³C NMR spectroscopic technique also had the ability to detect the low-level (minor) components -tocopherol acetate and free cholesterol in the cod liver oil products examined here (Fig. 12) . Indeed, 17 out of a possible 26 free cholesterol resonances were detectable (i.e., those assignable to C3, C4/13, C5, C6, C9, C10, C11, C12, C14, C17, C18, C20, C21, C22, C23, and C24); the remainder overlapped with those of acylglycerols and, when present, ethyl esters of various fatty acids (including DHA and EPA). The spectra obtained also confirmed that little or no cholesterol ester species were present in either product. For -tocopherol acetate, however, we found that all of the aromatic ring ¹³C resonances were visible, as were those arising from the C2- heterocyclic ring and the C11'-chain position -CH₂- groups.

View larger version (28K): [in this window] [in a new window]   Fig. 12. Expansion of selected regions of the 150.93 MHz 13C NMR spectrum of product I (CDCl3 solution). A, B: Expanded 34.50–83.50 and 112.0–170.50 ppm regions, respectively, showing signals ascribable to low-level (minor) components. A typical spectrum is shown. C3–C24 and T2–T8a, selected cholesterol and -tocopherol acetate resonances, respectively, detectable in the spectra acquired, and corresponding to the carbon nucleus numbering system in the structures depicted. TA represents the acetate group carboxylate resonance of -tocopherol acetate; 1–9 represent resonances arising from glyceryl carbons in different classes of glycerides and correspond to those listed in Table 4.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Application of both 1D ¹H NMR and 2D ¹H-¹H NMR spectroscopic techniques to the analysis of commercially available fish oil supplements provided a high level of data regarding the identification and, where appropriate, quantification of a wide range of key fatty acid components (as triacylglycerols or otherwise) in encapsulated marine oil supplements. Further components detectable included sn-1-monoacylglycerols, and sn-1,2- and sn-1,3-diacylglycerol adducts, together with a range of minor (low content) agents such as free glycerol and cholesterol, and vitamins E and A (as -tocopherol acetate and all-trans-retinol palmitate, respectively). The above techniques were very useful for monitoring the DHA content of the products examined as mg·g^-1, mol·kg^-1, and mol% units (a consequence of its unique, clearly resolved F,F_ß-position ¹H resonances), together with the total n-3 fatty acid content expressed as a percentage of the total fatty acids present. Because the DHA content of marine oils serves as a valuable criterion for determining their quality (47), its rapid and nondestructive analysis by ¹H NMR spectroscopy clearly offers many advantages. Moreover, in 1992, the British Nutrition Foundation recommended that DHA and EPA (combined) should represent 0.5% of the total energy of an average population's dietary intake (48), and hence, specific and selective methods for their determination are required. EPAs as acylglycerol species were also distinguishable from other glycerol-bound fatty acids in the 1D ¹H NMR spectra acquired, and the 2D ¹H-¹H COSY technique facilitated its detection by virtue of strong connectivities between its F,F_ß and C4(-1) signals, and also the very low content of potentially interfering fatty acids.

The 2D ¹H-¹H techniques employed here also facilitated the detection of synthetic ethyl esters of both DHA and EPA, together with lower levels of those of non-n-3 fatty acids, in one of the products tested (product I), a very important observation in view of the current demand for organic, unadulterated food products. Hence, high-field NMR spectroscopy serves as an extremely useful and rapid means of distinguishing natural sources of cod liver and other fish oils (i.e., DHA and EPA in acylglycerol form) from those that have been fortified with synthetic derivatives such as the ethyl esters identified here. It is understood that these ethyl esters are generated in or supplemented to selected encapsulated cod liver oil products as a means of overcoming adverse taste and aftertaste effects associated with the ingestion of these formulations. (However, the manufacturers of product I specify that their capsules contain "Cod liver oil enriched with fish oil Omega-3'".) Because a significant (albeit relatively small) fraction of these ethyl esters are non-n-3 fatty acids (Fig. 4), it appears that the manufacturers have subjected a cod liver oil preparation to a direct esterification process rather than conducting this step subsequent to separation and isolation of the individual fatty acids from the matrix.

Presumably, the manufacturers of product I have added the ethyl rather than methyl esters of these HUFAs, because hydrolysis of the latter in the gut or, alternatively, via the actions of esterase enzymes in vivo will liberate methanol, the toxicological properties of which are well known and extensively documented (49).

As observed here, d₄-MeOH proved to be an effective extraction medium for the ethyl esters of both n-3 and non-n-3 fatty acids present in product I that were acquired by ¹H NMR spectra, demonstrating that little or no triacylglycerols [as revealed by the low intensity of their glyceridic ¹H resonances, together with that of the bulk chain (-CH₂-)_n groups, which partially overlaps that of the ethyl esters' -CH₂OCOR group in spectra of the untreated oil in C²HCl₃ solution] were extractable from this product under these conditions.

In the areas of lipid chemistry and technology, a high level of valuable NMR applications are provided by ¹³C NMR spectroscopy, a phenomenon generally attributable to the much wider chemical shift range of these nuclei. Indeed, resonances assignable to the highly variable carbon environments and also to the substitutional status on the glycerol backbone have much diagnostic value. Such chemical shift data have been extensively reviewed and tabulated by Gunstone and Seth (29, 43, 44) for a range of acylglycerol and free fatty acid species. As demonstrated here, the chemical shift values and relative intensities of methylenic, vinylic, and ester carboxyl resonances are suitable for the direct, rapid, and virtually noninvasive analysis of fatty acid components in commercially available fish oil supplements.

The ¹³C NMR technique is particularly useful in distinguishing between mono-, di- and triacylglycerols, and also the particular substitutional isomers (e.g., sn-1,2- and sn-1,3-diacylglycerols) therein. Indeed, mono-, di- and triacylglycerols have characteristic chemical shift values for not only the three glycerol backbone carbons but also the C1 and C2 carbons in each acyl chain, and such data can, in theory, give rise to both quantitative and semiquantitative information concerning a whole series of these glycerol esters present in fish oil supplements such as those tested here.

Such information is likely to be of much value in distinguishing natural fish oils from those that have been subjected to industrial refining processes. Moreover, these data may also be valuable in determining the shelf life of fish oil products, because in virgin olive oils, sn-1,3-diacylglycerol levels increase after prolonged storage, a consequence of either lipolysis or 1,2/1,3-diacylglycerol isomerism (46). Similarly, information acquired regarding the positional distribution of individual fatty acids on the glycerol backbone of acylglycerols may also serve to provide a discriminatory index concerning the nature, refinement, and/or adulteration of these products.

Although free fatty acids (which provide a valuable measure of the acidity of culinary oil products) were not detectable in the preparations investigated here, NMR-detectable levels of trans-fatty acids were found in one of the fish oil supplements by virtue of a highly specific, albeit narrow, ¹³C chemical shift "window." Consumption of these agents by the human population purportedly gives rise to a wide range of adverse health effects, e.g., the risk of coronary heart disease (50) and low birth weight (51). Interestingly, partially hydrogenated marine oils represent a major source of dietary trans-fatty acids in some countries (51).

Because fish oil supplements have been reported to elevate the concentrations of LDL cholesterol in some individuals (52), the presence of this agent in the products examined here could be considered a contributory factor, although the levels detected were low (5 mg·g^-1).

With regard to LOPs, Igarashi et al. (28) found that the peroxide values of bonito, salmon, and tuna fish oils ranged from 0.6–10.6 meq·kg^-1, and therefore the ¹H NMR detection of conjugated diene signals tentatively assignable to one or more PUFA-derived hydroxydiene species is not unexpected.

A combination of high-resolution NMR data acquired on fish oils with multivariate statistical analysis techniques, such as principal component analysis, may have the ability to provide evidence for the source (i.e., species of fish) and possibly geographic origin of the species from which these products were collected.

Although it is generally accepted that DHA and EPA (together with further n-3 fatty acids), either individually or in concert, are responsible for the health benefits offered by marine oil supplements, these properties are also expected to be dependent on the precise molecular nature of these agents [e.g., as ethyl esters detectable in product I, as partial or triacylglycerols, and their substitutional status (positional distribution) with respect to the glycerol backbone], their relative concentrations, and, indeed, the chemical nature and levels of all other components therein (e.g., vitamin A and E derivatives). The detection of relatively low concentrations of components that are known, or likely, to exert adverse effects on human health, such as LOPs (53), trans-fatty acids (50, 51), and, to a much lesser extent, cholesterol, is also an important factor for consideration. Therefore, we conclude that multicomponent monitoring of these products by 1D and/or 2D NMR spectroscopic techniques provides a wealth of molecular information that is likely to be of much value to those conducting clinical studies on the therapeutic actions putatively offered by these materials.

	ACKNOWLEDGMENTS

 
The authors are very grateful to the Department of Chemistry, Queen Mary University of London, for the provision of NMR facilities; to the Special Trustees of St. Bartholomew's Hospital for financial support; and to Max McCann for helpful discussions.

Manuscript received June 20, 2003

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

1. Kromhout, D., E. B. Bosschieter, and C. de Lezenne Coulander. 1985. The inverse relation between fish consumption and 20-year mortality from coronary heart disease. N. Engl. J. Med. 312: 1205–1209.[Abstract]

2. Ascherio, A., E. B. Rimm, M. J. Stampfer, E. L. Giovannucci, and W. C. Willett. 1995. Dietary intake of marine n-3 fatty acids, fish intake, and the risk of coronary disease among men. N. Engl. J. Med. 332: 977–982.[Abstract/Free Full Text]

3. Kremer, J. M., D. A. Lawrence, G. F. Petrillo, F. Gayle, L. L. Litts, P. M. Mullaly, R. I. Rynes, R. P. Stocker, N. Parhami, N. S. Greenstein, B. R. Fuchs, A. Mathur, D. R. Robinson, R. I. Sperling, and O. Bigaouette. 1995. Effects of high dose fish oil on rheumatoid arthritis after stopping non-steroidal anti-inflammatory drugs. Clinical and immune correlates. Arthritis Rheum. 38: 1107–1114.[Medline]

4. Dyerberg, J., H. O. Bang, and N. Hjoorne. 1975. Fatty acid composition of the plasma lipids in Greenland Eskimos. Am. J. Clin. Nutr. 28: 958–966.[Abstract/Free Full Text]

5. Goodnight, S. H., Jr., W. S. Harris, and W. E. Connor. 1981. Effect of n-3 fatty acids on platelet function and composition. Blood. 58: 880–885.[Abstract/Free Full Text]

6. Harker, L. A., A. B. Kelly, S. R. Hanson, W. Krupski, A. Bass, B. Osterud, G. A. FitzGerald, S. H. Goodnight, and W. E. Connor. 1993. Interruption of vascular thrombus formation and vascular lesion formation in primates. Circulation. 87: 1017–1029.[Abstract/Free Full Text]

7. von Schacky, C., P. Angerer, W. Kothny, K. Theisen, and H. Mudra. 1999. The effect of dietary omega-3 fatty acids on coronary atherosclerosis. A randomized, double-blind placebo-controlled trial. Ann. Intern. Med. 130: 554–562.[Abstract/Free Full Text]

8. Hazra, A., S. Kumar Tripathi, and A. Ghosh. 1999. Pharmacology and therapeutic potential of n-3 polyunsaturated fatty acids, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DMA) in fish oils. Ind. J. Pharmacol. 31: 247–264.

9. Goodnight, S. H., Jr., W. S. Harris, W. E. Connor, and D. R. Illingworth. 1982. Polyunsaturated fatty acids, hyperlipidaemia and thrombosis. Arteriosclerosis. 2: 87–113.[Free Full Text]

10. McLennan, P. L., M. Y. Abeywardena, and J. S. Charnock. 1988. Prevention of ventricular fibrillation after occlusion-reperfusion in rats. Am. Heart J. 116: 709–717.[CrossRef][Medline]

11. Billman, G. E., H. Hallaq, and A. Leaf. 1994. Abolition of ventricular fibrillation in dogs. Proc. Natl. Acad. Sci. USA. 91: 4427–4430.[Abstract/Free Full Text]

12. Vreugdenhil, M., C. Bruehl, R. A. Voskuyl, J. X. Kang, A. Leaf, and W. J. Wadman. 1996. Polyunsaturated fatty acids modulate sodium and calcium channels in CA1 neurons. Proc. Natl. Acad. Sci. USA. 98: 12559–12563.

13. Malasanos, T. H., and P. W. Stackpoole. 1991. Biological effects of omega-3 fatty acids in diabetes mellitus. Diabetes Care. 14: 1160–1179.[Abstract]

14. Chen, L. H., and S. Xi. 1996. Fish oil retards the progression of murine AIDS. The International Conference on Highly Unsaturated Fatty Acids in Nutrition and Disease Prevention in Barcelona, Spain. 1996.

15. Takahata, H. 1996. Prevention of neuronal cell death with docosahexaenoic acid. The International Conference on Highly Unsaturated Fatty Acids in Nutrition and Disease Prevention in Barcelona, Spain. 1996.

16. American Oil Chemists' Society Press. 1997. Fatty Acid Composition by GLC. Marine oils. Official Methods and Recommended Practices of the American Oil Chemists' Society. 4th edition. American Oil Chemists' Society Press, Champaign, IL. 16–89.

17. Kajishima, T., S. Aoki, E. Nishimoto, E. Hasegawa, E. Moriyoshi, and H. Nagayama. 1996. Quantitative determination of EPA and DHA in refined fish oil by capillary gas chromatography. J. Jpn. Oil Chem. Soc. 45: 335–339.

18. Igarashi, T., Y. Hirata, M. Nonaka, and S. Wada. 1997. Nondestructive evaluation with NMR on DHA triglycerides in refined fish oils. In Magnetic Resonance Spectroscopy in Food Science. H. Saito and G. Webb, editors. Himeji Institute of Technology, Hyogo, Japan. 103–106.

19. Esterbauer, H. 1985. Lipid peroxidation products: formation, chemical properties and biological activities. In Free Radicals in Liver Injury. G. Poli, K. H. Cheeseman, M. U. Diazani, and T. F. Slater, editors. IRL Press, London. 29–47.

20. Bell, J. C., J. C. Brown, and P. J. Sadler. 1989. NMR studies of biofluids. NMR Biomed. 2: 246–256.[Medline]

21. Nicholson, J. K., and I. D. Wilson. 1989. High-resolution proton NMR spectroscopy of biological fluids. Prog. NMR Spectrosc. 21: 449–501.

22. Grootveld, M., A. Sheerin, M. Atherton, A. D. Millar, E. J. Lynch, D. R. Blake, and D. P. Naughton. 1996. Applications of high resolution NMR analysis to the study of inflammatory diseases at the molecular level. In Biomedical Applications of Spectroscopy, Advances in Spectroscopy. Vol. 25. R. J. H. Clark and R. E. Hester, editors. John Wiley and Sons, Ltd., Chichester. 295–327.

23. Aursand, M., and H. Grasdalen. 1992. Interpretation of the ¹³C NMR spectra of omega-3 fatty acids and lipid extracted from the white muscle of Atlantic salmon (Salmo salar). Chem. Phys. Lipids. 62: 239–251.[CrossRef]

24. Aursand, M., J. R. Rainuzzo, and H. Grasdalen. 1993. Quantitative high-resolution ¹³C and ¹H nuclear magnetic resonance of -3 fatty acids from white muscle of Atlantic salmon (Salmo salar). J. Am. Oil Chem. Soc. 70: 971–981.

25. Sacchi, R., I. Medina, S. P. Aubourg, F. Addeo, and L. Paolillo. 1993. Proton nuclear magnetic resonance rapid and structure-specific determination of -3 polyunsaturated fatty acids in fish lipids. J. Am. Oil Chem. Soc. 70: 225–228.

26. Medina, I., and R. Sacchi. 1994. Acyl stereospecific analysis of tuna phospholipids via high resolution ¹³C nuclear magnetic resonance spectroscopy. Chem. Phys. Lipids. 70: 53–61.[CrossRef]

27. Aursand, M., L. Jorgensen, and H. Grasdalen. 1995. Positional distribution of -3 fatty acids in marine lipid triacylglycerols by high-resolution ¹³C nuclear magnetic resonance spectroscopy. J. Am. Oil Chem. Soc. 72: 293–297.

28. Igarashi, T., M. Aursand, Y. Hirata, I. Gribbestad, S. Wada, and M. Nonaka. 2000. Nondestructive quantitative determination of docosahexaenoic acid and n-3 fatty acids in fish oils by high-resolution ¹H nuclear magnetic resonance spectroscopy. J. Am. Oil Chem. Soc. 77: 737–748.[CrossRef]

29. Gunstone, F. D., and S. Seth. 1994. A study of the distribution of eicosapentaenoic acid and docosahexaenoic acid between the and ß glycerol chains in fish oils by ¹³C NMR spectroscopy. Chem. Phys. Lipids. 72: 119–126.[CrossRef]

30. Aue, W. P., E. Bartholdi, and R. R. Ernst. 1976. Two-dimensional spectroscopy: application to nuclear magnetic resonance. J. Chem. Phys. 64: 2229–2246.[CrossRef]

31. Braunschweiler, L., and R. R. Ernst. 1983. Coherence transfer by isotropic mixing—application to proton correlation spectroscopy. J. Magn. Reson. 53: 521–528.

32. Bax, A., and D. G. Davies. 1985. MLEV-17-based two-dimensional homonuclear magnetization transfer spectroscopy. J. Magn. Reson. 65: 355–360.

33. Marion, M., and K. Wuthrich. 1983. Application of phase sensitive two-dimensional correlated spectroscopy (COSY) for measurements of ¹H-¹H spin-spin coupling constants in proteins. Biochem. Biophys. Res. Commun. 113: 967–974.[CrossRef][Medline]

34. Grubbs, F. E. 1969. Procedures for Detecting Outlying Observations in Samples. Technometrics. 11: 1–21.

35. IUPAC-IUB Commission on Biochemical Nomenclature (CBN). 1967. The nomenclature of lipids. Eur. J. Biochem. 2: 127–131.[Medline]

36. Haywood, R. M., A. W. D. Claxson, G. E. Hawkes, D. P. Richardson, D. P. Naughton, G. Coumbarides, J. Hawkes, E. J. Lynch, and M. C. Grootveld. 1995. Detection of aldehydes and their conjugated hydroperoxydiene precursors in thermally-stressed culinary oils and fats: investigations using high resolution proton NMR spectroscopy. Free Radic. Res. 22: 441–482.[Medline]

37. Willker, W., and D. Leibfritz. 1998. Assignment of mono- and polyunsaturated fatty acids in lipids of tissues and body fluids. Magn. Reson. Chem. 36 (Suppl.): 79–84.[CrossRef]

38. Pouchert, C. J., and J. Behnke. 1993. The Aldrich Library of ¹³C and ¹H FT NMR Spectra, 1st edition. Vol. 1. Aldrich Chemical Co., Inc., Milwaukee, WI.

39. Silwood, C. J. L., and M. Grootveld. 1999. Application of high-resolution, two-dimensional ¹H and ¹³C nuclear magnetic resonance techniques to the characterization of lipid oxidation products in autoxidized linoleoyl/linolenoylglycerols. Lipids. 34: 741–756.[CrossRef][Medline]

40. Saeed, S., and N. K. Howell. 1999. High-performance liquid chromatography and spectroscopic studies extracted from frozen Atlantic mackerel. J. Am. Oil Chem. Soc. 76: 391–397.

41. Nicholson, J. K., M. J. Buckingham, and P. J. Sadler. 1983. High resolution ¹H N.M.R. studies of vertebrate blood and plasma. Biochem. J. 211: 605–615.[Medline]

42. Gunstone, F. D. 1991. High resolution NMR studies of fish oils. Chem. Phys. Lipids. 59: 83–89.[CrossRef]

43. Gunstone, F. D. Polyunsaturated fatty acids. (Accessed March 3, 2003, at http://www.lipid.co.uk/infores/nmr/nmrpufa.)

44. Gunstone, F. D. Regiospecific analysis. (Accessed March 3, 2003, at http://www.lipid. co.uk/infores/nmr/nmrregio.)

45. Sacchi, R., I. Medina, L. Paolillo, and E. I. Addeo. 1994. High-resolution ¹³C NMR olefinic spectra of DHA and EPA acids, methyl esters and triacylglycerols. Chem. Phys. Lipids. 69: 65–73.[CrossRef]

46. Sacchi, R., F. Addeo, and L. Paolillo. 1997. ¹H and ¹³C NMR of virgin olive oil. An overview. Magn. Reson. Chem. 35 (Suppl.): 133–145.[CrossRef]

47. Japan Health Food and Nutrition Food Association. 1994. Specification Standards for Processed Refined Fish Oil-Containing Eicosapentaenoic Acid (EPA) and Docosahexaenoic Acid (DHA) in Health Food for your Guidance. Japan Health Food and Nutrition Food Association, Tokyo, Japan.

48. British Nutrition Foundation Task Force. 1992. British Nutrition Foundation, The Report of the British Nutrition Foundation's Task Force. Unsaturated Fatty Acids: Nutritional and Physiological Significance. Chapman and Hall, London. 157.

49. Crook, J. E., and J. S. McLaughlin. 1966. Methyl alcohol poisoning. J. Occup. Med. 8: 467–470.[Medline]

50. Oomen, C. M., M. C. Ocke, E. J. M. Feskens, M. A. J. van Erp-Baart, F. J. Kok, and D. Kromhout. 2001. Association between trans-fatty acid intake and 10-year risk of coronary heart disease in the Zutphen Elderly Study: a prospective population-based study. Lancet. 357: 746–751.[CrossRef][Medline]

51. Enig, M. G. Trans Fatty Acid Fact Sheet. (Accessed April 7, 2003, at http://www.enig.com/0001tlb.html.)

52. Harris, W. S., M. L. Zucker, and C. A. Dujovne. 1987. Omega-3 fatty acids in type IV hyperlipidemia: fish oils versus methyl esters (Abstract). Am. J. Clin. Nutr. 45: 858.

53. Grootveld, M., C. J. L. Silwood, P. Addis, A. W. D. Claxson, B. Bonet Serra, and V. Viana. 2001. Health effects of oxidised heated oils. Foodserv. Res. Internat. 13: 39–53.

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This Article Abstract Full Text (PDF) All Versions of this Article: D300017-JLR200v1 44/12/2406    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Siddiqui, N. Articles by Grootveld, M. Search for Related Content PubMed PubMed Citation Articles by Siddiqui, N. Articles by Grootveld, M. Social Bookmarking                      What's this?