Evaluation of a Microwave-Assisted Extraction Methodfor Lignan Quantification in Flaxseed Cultivars and SelectedOil Seeds

Simona Mihaela Nemes & Valérie Orsat

Received: 1 June 2011 /Accepted: 2 August 2011 /Published online: 17 August 2011# Springer Science+Business Media, LLC 2011

Abstract An optimized microwave-assisted extraction(MAE) method was evaluated through repeatability, recov-ery and efficiency testing. The repeatability tests, performedby three users over time, were not significantly different.The recovery of lignan throughout the extraction, prepara-tion and analysis steps is 97.5% with a coefficient ofvariation <1%. The MAE method is efficient for extractinglignans from the plant matrix, and it achieves significantlyhigher extraction yields than the two established referencemethods. The applicability of the MAE method wasdemonstrated by extracting lignans from a variety of plantsamples. The secoisolariciresinol diglucoside (SDG) con-tent of seven flaxseed cultivars grown in Saint-Mathieu-de-Beloeil, QC, in 2009 ranged from 14.6 to 18.9 mg SDG pergram of seed. Flax hulls produced in Winchester, ON, in2010 were very rich in lignan; their SDG content was40.0 mg/g of flax hull. Sesame seeds contained 0.18–1.89 mg SECO (aglycone of SDG) per gram of seed, withsignificant differences among black, white and brownsesame seed. Chia seeds contained 0.99–1.29 mg SECOper gram of seed, with significant differences amongbatches of seeds. Sunflower seeds had 0.046 mg SECOper gram of sample and almonds had 0.029 mg SDG pergram of sample. The optimized and evaluated MAE methodis recommended for the general analytical quantification oflignans in plant samples.

Keywords MAE . Secoisolariciresinol . SDG . Sesame .

Chia . Sunflower . Almonds

Introduction

Lignans are a major class of phytoestrogens that are foundin decreasing order of concentration in oilseeds, nuts, soyproducts and cereals (Thompson et al. 2006). They areconcentrated in the bran layer of cereals and the seed coatof oil seeds. Although plant foods in general may containup to 2 mg lignan/100 g of food product, flaxseed andsesame seeds have more than 300 mg lignan/100 g of seeds(Peterson et al. 2010). Flaxseed contains the lignansecoisolariciresinol diglucoside (SDG) which is naturallyoccurring as part of a complex macromolecule. Animportant body of research has contributed to the elucida-tion of the composition and the structure of the flaxseedlignan macromolecule in the past 10 years (Ford et al.2001; Kamal-Eldin et al. 2001; Struijs et al. 2007, 2008,2009). The flaxseed lignan macromolecule includes,besides SDG, 3-hydroxy-3-methyl-glutaric acid (HMGA),p-coumaric acid glucoside, ferulic acid glucoside and theflavonoid herbacetin diglucoside. A sketch of the lignanmacromolecule (Fig. 1) was proposed by Peterson et al.(2010) upon reviewing the latest studies on the composi-tion and the structure of the flaxseed lignan. The backboneof the macromolecule is made of units of SDG ester linkedto HMGA; the number of backbone units (represented bythe letter n in Fig. 1) can vary between 1 and 7, with anaverage of 3.

In order to correlate the health effects conferred by SDGconsumption in humans, the quantity of SDG present inflaxseed products used in epidemiological and lignansupplementation studies has to be known, whether theseproducts are SDG extracts, whole or defatted flaxseed meal(Adolphe et al. 2010; Peterson et al. 2010). Adolphe et al.(2010) summarized the health effects of SDG consumptionupon reviewing only those animal and human studies that

S. M. Nemes (*) :V. OrsatBioresource Engineering Department, Macdonald Campusof McGill University,Macdonald–Stewart Building, 21111 Lakeshore Road,Sainte-Anne-de-Bellevue, QC, Canada H9X 3V9e-mail: simona.nemes@mail.mcgill.ca

Food Anal. Methods (2012) 5:551–563DOI 10.1007/s12161-011-9281-6

reported the dosage of SDG in the flaxseed products used.It appears that SDG benefits human health upon biocon-version in the digestive tract to the more potent mammalianlignans: enterodiol and enterolactone. The mammalianlignans have antioxidant and weak oestrogenic or anti-oestrogenic effects, thus providing protection againstcardiovascular diseases, the metabolic syndrome andcertain types of cancer.

The demand for known lignan contents in foods resultedin the publication of a variety of analytical methods forlignan analysis and a number of databases reporting lignancontents in foods. A recent inventory of publishedphytoestrogen databases (Schwartz et al. 2009) has identi-fied the need for a generally applicable sample preparation

methodology. The inconsistencies in the lignan values invarious databases were attributed to the use of differentmethods of extraction. Moreover, the limitations of acidicand enzymatic hydrolyses, such as artefact formation and/orincomplete extraction from the matrix, led to the underes-timation of the results. Two main trends in samplepreparation methodologies that were used for the construc-tion of some of the databases were identified; theseinvolved lignan extraction by methanolysis or alcoholicextraction followed by alkaline hydrolysis, both followedby enzymatic hydrolysis (Table 1).

Smeds et al. (2007) found that the yield of lignanextraction can be increased if an additional mild hydrolysisstep is introduced between the alkaline and enzymatic

Fig. 1 Structure of the flaxseed lignan macromolecule according to Peterson et al. (2010)

Table 1 Trends in extraction methodologies used for lignan quantification in foods and construction of databases and the values of lignan contentin flaxseed achieved with these methodologies

Method description Specifications Analysed samples Results (mg lignan/gfresh weight flaxseed)a

References

Extraction with 0.3 M NaOH in 70%Me–OH at 60 °C for 1 h. Extractacidification to pH 5–6 with aceticacid. Centrifugation and drying ofpooled supernatants under N2stream

Enzymatic hydrolysis of driedextracts

Dutch food 3.01 aglycone (5.71 SDGequivalent)

Milder et al.(2005)

Food 3.35 aglycone (6.35 SDGequivalent)

Penalvo et al.(2005)

Japanese food – Penalvo et al.(2008)

Mild acidic hydrolysis of driedextracts with 0.01 M HCl in70% Me–OH at 60 °C for1 h, followed by enzymatichydrolysis

Cereals, oil seeds andnuts

1.69 aglycone (3.20 SDGequivalent)

Smeds et al.(2007)

Drying step eliminated.Enzymatic hydrolysis ofacidified liquid extract

Cereals – Smeds et al.(2009)

Extraction with 70% Me–OH at 60–70 °C for 2 h twice. Evaporationof extracts under vacuum at 60 °C,followed by hydrolysis with 1 MNaOH for 3 h, and neutralizationwith acetic acid

Purification of neutralizedextract by SPE followed byenzymatic hydrolysis andfinal purification by SPE

Canadian food 3.79 aglycone (7.18 SDGequivalent)

Thompson et al.(2006)

Dietary supplements – Thompson et al.(2007)

a The aglycone values represent the sum of SECO, pinoresinol, lariciresinol and matairesinol

552 Food Anal. Methods (2012) 5:551–563

hydrolysis steps. Schwartz et al. (2009) identified thissequential three-step hydrolysis methodology as trend-setting in lignan analysis and an indicator that furthermethod development is required for analytical lignanextraction. Although the additional mild hydrolysis stepwas not used later by Smeds et al. (2009) for lignananalysis in cereals, it is worth further investigation.

Although significant advancements have been done inthe fields of extraction and analysis of lignans in the past10 years, the efficient extraction of lignans from plantmatrices remains a challenge. It is known that lignancontents in flaxseed vary not only with the cultivatedvariety (Eliasson et al. 2003; Johnsson et al. 2000) but alsowith the location and the year of growth (Westcott and Muir2003). These are natural challenges that complicate the taskof lignan quantification in foods as it would be impracticaland tedious to analyse various cultivated varieties of oilseeds and cereal grains harvested in different years atvarious locations. However, the methods of extraction andanalysis of lignans are challenges that the researcher has achoice over. The composition of flaxseed extracts vary withthe extraction method used. If alcoholic extraction followedby alkaline hydrolysis or direct alkaline hydrolysis is used,the reported lignan is SDG (Eliasson et al. 2003; Johnssonet al. 2000; Nemes and Orsat 2009, 2010). If the flaxseedalcoholic or alkaline extracts are further hydrolysed byacidic or enzymatic hydrolysis, a variety of lignanaglycones are reported besides secoisolariciresinol (SECO).For example, when enzymatic hydrolysis has been used,matairesinol, lariciresinol and pinoresinol have beenreported by many, as can be seen from Table 1; in additionto these, isolariciresinol and demethoxy-secoisolariciresinolhave been reported by Sicilia et al. (2003). Enzymatichydrolysis appears to be a better choice than acidichydrolysis as the latter causes the transformation oflariciresinol into isolariciresinol and the dehydration (lossof a water molecule) of SECO and demethoxy-SECO(Sicilia et al. 2003). The formation of lignan artefactsduring the extraction increases the cost and the difficulty ofthe chromatographic analysis as more standard lignancompounds are required. As opposed to enzymatic andacidic hydrolysis, alkaline hydrolysis is more efficient androbust (no artefact formation), and the chromatographicanalysis of flaxseed lignans is straightforward as only SDGis present (Oomah and Hosseinian 2008).

The fact that SDG was the only lignan found in theflaxseed lignan macromolecule (Peterson et al. 2010)corroborated with earlier findings according to whichSDG was as well the only lignan found by high-performance liquid chromatography (HPLC) with massspectrometry and nuclear magnetic resonance spectroscopyanalysis of flaxseed extracts obtained by alcoholic extrac-tion followed by alkaline hydrolysis (Ford et al. 2001).

Further proof was provided by Eliasson et al. (2003) whoused reversed-phase HPLC with SDG, syringaresinol,pinoresinol and matairesinol as lignan standards foridentification, but found only SDG in flaxseed extractsobtained by direct alkaline hydrolysis.

The most efficient methods for lignan extraction fromflaxseed use direct alkaline hydrolysis with dilute NaOH(Eliasson et al. 2003; Nemes and Orsat 2009). Alkalinehydrolysis has the role of breaking the ester bonds betweenSDG and HMGA, thus releasing the SDG from thecomplex lignan macromolecule. The lignan macromoleculehas been extracted often with aqueous alcohol or withalcoholic mixtures (Eliasson et al. 2003; Johnsson et al.2000, 2002; Kamal-Eldin et al. 2001; Thompson et al.2006). However, the SDG content of the macromoleculecan only be analysed upon hydrolysis. The alcoholicextraction step is time-consuming; some authors have used4 h (Thompson et al. 2006) and others as much as 16 h(Johnsson et al. 2000) or even up to 48 h (Eliasson et al.2003). The duration of alcoholic extraction is crucial for aquantitative recovery of the lignan macromolecule; thelonger the extraction time, the higher the yield of the lignanmacromolecule, with a high yield being obtained after 48 hof alcoholic extraction. Moreover, a direct alkaline hydro-lysis method was developed (1-h hydrolysis at roomtemperature with 1 M NaOH) that gave results higher than48-h alcoholic extraction followed by alkaline hydrolysis(Eliasson et al. 2003). A new microwave-assisted extraction(MAE) method that used a direct hydrolysis approach wasdeveloped by Nemes and Orsat (2009, 2010). The MAEmethod was used for SDG extraction from defatted flaxseedmeal and achieved a 6% increase in the extraction yield asopposed to the conventional direct hydrolysis methoddeveloped by Eliasson et al. (2003), with additional benefitssuch as: reduction in extraction time by 95%, reduction inthe NaOH concentration by half and an internal standard(o-coumaric acid) recovery of 97%. MAE methods publishedby others showed improvements in the extraction yields ofSDG from flax hull (Zhang and Xu 2007) and flaxseed cake(Beejmohun et al. 2007) as opposed to conventional methodssuch as stirring extraction and Soxhlet extraction.

This article proposes to bring further proof in terms ofefficiency and repeatability for our previously developedoptimized MAE method, as well as to demonstrate itsapplicability for fast and efficient analytical quantificationof lignans in flaxseed cultivars and other oil seed samples.The first objective of this article was to evaluate ourpreviously developed MAE method by: (1) assessing itsrepeatability when performed by different users over time,(2) assessing its recovery of standard SDG compound and(3) assessing its efficiency by comparison with knowntrend-setting methods and investigating the possibility offurther increasing the extraction yield using a sequential

Food Anal. Methods (2012) 5:551–563 553

hydrolysis approach. The second objective was to use theevaluated MAE method for the analysis of SDG in differentflaxseed cultivars, flaxseed hulls and other oil seeds.

Materials and Methods

Chemicals

The reference lignan standards of HPLC grade—SDG (molecular weight, 686.7; purity, 97.6%) andanhydro-secoisolariciresinol (ANSECO; molecularweight, 344.4; purity, >99%)—were purchased fromChromadex (Santa Ana, CA, USA); SECO (molecularweight, 362.4; purity, ≥95%) was purchased from Sigma-Aldrich (Oakville, ON, Canada). The solvents acetonitrile,methanol and hexane of HPLC grade were obtained fromFisher Scientific (Ottawa, ON, Canada). The reagents sodiumhydroxide ≥98%, sulphuric acid 95–98% ACS, phosphoricacid ≥85% and dipotassium hydrogen phosphate 98% werepurchased from Sigma-Aldrich.

Samples

The flaxseed (Linum usitatissiumum L.) used for the MAEmethod efficiency assessment and validation was purchasedin 2009 from a local grocery store (Montreal, QC, Canada);the country of provenance was Guatemala, as specified onthe package. The flaxseed cultivars (CDC Bethune,McBeth, Prairie Blue, Flanders, 09LS01, CRGL 8.1 andCRGL 8.2) analysed for SDG content were provided by Mr.Yves Dion from Centre de Recherche sur les Grains Inc.(CEROM). All the cultivars or advanced breeding lineswere grown in Saint-Mathieu-de-Beloeil, (QC, Canada) in2009. The Omega-3 Flax Hull, produced in May 2010, wasprovided by Dr. Nam Fong Han from Natunola Health Inc.(Winchester, ON, Canada). Other samples used for lignananalysis were purchased from local grocery stores in 2008and 2010 as follows: white and black sesame (Sesamumindicum L.) seeds and black chia (Salvia hispanica L.)seeds were purchased in 2008; white and black chia seeds,brown sesame seeds, sunflower seeds and almonds werepurchased in 2010. All the samples were divided intobatches of 100–300 g, packed in plastic bags (the flaxhullwas vacuum-packed) and stored at −18 °C until use.

Microwave Extraction Apparatus

The MAE experiments were carried out in a mono-mode(focused) microwave apparatus (Star System 2, CEM,Mathews, USA) with a nominal power of 800 W andmicrowave frequency of 2.45 GHz. The temperature wasmonitored by a built-in IR temperature sensor placed at the

bottom of the extraction vessel. The extraction vessels(250 ml) were made of borosilicate glass and were fittedwith a Graham-type reflux condenser.

HPLC Analysis of Lignans

All extracts were analysed in triplicate (coefficient ofvariation, <5%) with an Agilent 1100 series HPLC. TheChemstation software for LC systems (Rev. B.01.03 [204],Agilent Technologies) was used for instrument control andchromatographic data analysis. The chromatograms wererecorded at 280 nm using a variable wavelength detector.The separation of lignans (SDG, SECO, ANSECO) wascarried out at 25 °C on a reversed-phase C18 column(Discovery; 5 μm, 25 cm×4.6 mm; Sigma-Aldrich) fittedwith a guard cartridge (Supelguard; 5 μm, 2 cm×4 mm;Sigma-Aldrich) using a slightly modified version of thegradient elution method developed by Johnsson et al.(2000). The two solvents were: (A) 0.01 M phosphatebuffer, pH 2.8, containing 5% acetonitrile and (B) acetoni-trile. Originally, solvent A decreased from 100% to 70% ata rate of 1% per minute over a period of 30 min. In thiswork, the rate of decrease of solvent A was maintained at1% per minute, but it was extended over a period of50 min. This extended gradient method allowed thequantification of the three lignans SDG, SECO andANSECO together; their respective elution times were19.5, 26.5 and 38.5 min. Standard curves were built usingsix levels of lignan concentrations ranging from 5 to200 μg lignan per millilitre methanol; in all cases, thecorrelation coefficients were r > 0.999. The followingequations were used for quantification: SDG = peak height/0.4494, SECO = peak height/1.054, ANSECO = peakheight/0.8065.

Sample Preparation

Microwave-Assisted Extraction

Experiments were carried out in order to assess therepeatability and the efficiency of the optimized MAEmethod and for the quantification of lignans in variousflaxseed cultivars, flaxseed hulls, sesame seeds, chia seeds,almonds and sunflower seeds. Some experiments werecarried out using defatted flaxseed meal (DFM) that wasobtained by extracting flaxseed meal twice with hexane(1:6 sample-to-liquid ratio, grams per millilitre) for 1 hunder magnetic stirring at room temperature (Nemes andOrsat 2010). The MAE experiments were carried outfollowing our previously published methodology (Nemesand Orsat 2009, 2010). Briefly, samples of 1 g of defattedor non-defatted flaxseed meal were hydrolysed with 50 mlof 0.5 M NaOH at 135 W; the microwave power was

554 Food Anal. Methods (2012) 5:551–563

applied intermittently (30 s on/off) for 3 min. Thetemperature of the extracts rose from room temperature(22–23 °C) to about 67 °C over the 3-min span. Theextracts contain water-soluble carbohydrates and proteinswhich must be removed before HPLC analysis. In order todo so, the extracts were acidified to pH 3 by the addition of5.55 ml of 5 N H2SO4 whilst measuring the pH with anAccumet 25 instrument (Fisher Scientific), then methanolwas added to the acidified extracts in a proportion of 2:1 (v/v, millilitres per millilitre), and the extracts were kept for15 min under magnetic stirring (3,000 rpm). The solid andliquid phases were separated by 10 min of centrifugation at3,000 rpm using a Spinette centrifuge (InternationalEquipment Company, Needham Heights, MA, USA). IfDFM was used for MAE, an aliquot of the clear liquidphase was passed through 0.22-μm Whatman Puradisc(13 mm) nylon syringe filters into 2-ml vials and thenanalysed by HPLC. If non-defatted flaxseed meal sampleswere used for MAE, the liquid phase was transferred into around bottom flask and the extract was stripped ofmethanol and concentrated to 3–5 ml by vacuum rotaryevaporation (30 min, 65 °C, 195 rpm) using a BuchiRotavapor 205 equipped with a B490 heating bath. Theaqueous concentrated extracts were transferred into 15-mlcentrifuge tubes and hexane was added in a proportion of2:1 (v/v, millilitres per millilitre). The tubes were shaken tofacilitate the defatting and then centrifuged for 10 min at3,000 rpm to speed up the separation of the organic phase,which was then removed with a pipette. The concentrationof the extracts was adjusted with methanol to bring thetheoretical sample-to-liquid ratio to a maximum of 10:1(milligrams per millilitre) and then the extracts were filteredthrough 0.22-μm membrane filters and analysed by HPLC.

Cumulative Effects of MAE, Sample Preparation and HPLCAnalysis on the Recovery of SDG Standard

The methodologies described above for MAE, samplepreparation and HPLC analysis, were applied in order toassess the recovery of SDG standard; 120 μg SDGdissolved into 300 μl methanol was used per MAE test.The percentage of SDG recovered was calculated from theamount of SDG used per test upon HPLC analysis.

In order to assess the efficiency of the optimized MAEmethod, two reference methods were carried out forcomparison purposes.

Reference Method A

A sequential extraction involving alkaline hydrolysisfollowed by mild acidic hydrolysis was performed asdescribed by Smeds et al. (2007). Briefly, samples of0.5 g DFM were hydrolysed with 24 ml of 0.3 M NaOH in

70% methanol for 1 h at 60 °C and then the pH wasadjusted to 5–6 with 750 μl glacial acetic acid; the extractswere centrifuged and the liquid phase was dried by vacuumrotary evaporation (65 °C, 195 rpm). Note that this type ofalkaline hydrolysis was also used by Milder et al. (2004,2005) and Penalvo et al. (2005, 2008). The dry extractswere dissolved in 24 ml of 0.01 M HCl in 70% methanoland then hydrolysed for 1 h at 60 °C. Methanol was addedto the extracts to bring the theoretical concentration tomaximum 10 mg sample per millilitre and then the extractswere filtered through 0.22-μm membrane filters andanalysed by HPLC.

Reference Method B

Alcoholic extraction followed by alkaline hydrolysis wascarried out following the procedures described by Thompsonet al. (2006). Briefly, 0.5 g DFM was extracted twice with25 ml of 70% methanol at 60 °C for 2 h and then the extractswere centrifuged and the liquid phase dried by vacuumrotary evaporation (65 °C, 195 rpm). The dry extracts weredissolved in 30 ml of 1 M NaOH and hydrolysed at roomtemperature for 3 h, and then the pH was brought to 7 by theaddition of 1.7 ml glacial acetic acid. The concentration ofextracts was adjusted with methanol to correspond to amaximum of 10 mg DFM/ml, and then the extracts werefiltered through 0.22-μm membrane filters and analysed byHPLC.

Sequential MAE

A sequential MAE experiment was carried out in order totest the hypothesis that alkaline hydrolysis (MAE) followedby mild acidic hydrolysis (MAE-AC) might increase thetotal yield of lignan. This hypothesis was considered basedon the findings of Smeds et al. (2007). The MAE-ACexperiment was designed keeping in mind that the acidityof the solvent and the extraction temperature might bedirectly related to lignan artefact formation. Sicilia et al.(2003) found that both enzymatic hydrolysis (in sodiumacetate buffer, pH 5, 37 °C) and acidic hydrolysis (in 1 MHCl, 95 °C) led to artefact formation, but the harsh acidichydrolysis produced more lignan artefacts. Furthermore, itis known that for stability, lignan extracts must be kept atpH 2–4 (Nollet 2000), preferably 3 (Eliasson et al. 2003;Johnsson et al. 2000), in order to prevent the ionization ofthe hydroxylic groups which would decrease lignans’retention times, thus resulting in HPLC quantificationlosses. Therefore, MAE-AC was carried out at pH 3 inorder to preserve the integrity of extracted lignans and toavoid quantification losses. The sequential MAE wascarried out as follows: samples of 1 g DFM were extractedby MAE and then the pH was adjusted to 3 with 5.55 ml of

Food Anal. Methods (2012) 5:551–563 555

5 N H2SO4; furthermore, the extracts were subjected toMAE-AC which investigated the time (3 and 9 min) andtemperature (60 and 90 °C) of hydrolysis according to theexperimental design presented below (“Experimental Designand Statistical Analyses”). The temperature of extracts afterMAE was about 67 °C; therefore, before running MAE-ACat 60 °C, the extracts needed to be cooled down to therequired temperature in a cold water bath. For practicalreasons, the microwave power was kept at 135 W for theMAE-AC as well. Given that the MAE-AC extracts alreadyhave a pH of 3, the ensuing preparation steps in view ofHPLC analysis were done as described above for the MAEexperiments.

Experimental Design and Statistical Analyses

One-way ANOVA and Tukey–Kramer HSD tests forpairwise comparison of means were carried out for therepeatability and efficiency assessments of MAE, thecomparison of lignan contents in various flaxseed cultivarsand the comparison of lignan contents among differentbatches of sesame and chia seeds. The number ofreplications for the pairwise comparison tests ranged fromtwo to five per sample. The sequential MAEwas designed asa two-factor, two-level full factorial screening design with tworeplications. The coded and actual levels for the factorsstudied were as follows: time (−1)=3 min and (+1)=9 min;temperature (−1)=60 °C and (+1)=90 °C. The design wasanalysed using response surface procedures. The significanceof tests was established at p values <0.05 in all cases. Thestatistical analyses were carried out with SAS 9.2 TS2M2 orJMP 8 (SAS Institute Inc., Cary, NC, USA).

Results and Discussion

MAE Repeatability

Three users tested the repeatability of the optimized MAEmethod by performing three extractions each over 3 weeks(one MAE per user per week) using the setup in our

laboratory. The non-defatted flaxseed meal (flaxseed grownin Guatemala) used was from the same batch. The dataobtained are presented in Table 2. The MAE tests wereperformed equally well by the three users judging from theresults of the pairwise comparison test, by which the threemeans did not differ from each other significantly (p=0.551).The highest coefficient of variation was 4.03%. In theliterature, coefficients of variation ranging from 1% to 7%were reported for two well-known lignan extraction methods.The direct hydrolysis method reported by Eliasson et al.(2003) was tested by two analysts over 3 days; thecoefficients of variation for SDG were smaller than7%. The complex lignan extraction method reported byThompson et al. (2007) was tested for variation within aday and between different days; the obtained coefficientsof variation ranged from 1% to 5.2% and from 1.4% to6.3%, respectively. Our results fall within the publishedrange of acceptable coefficients of variation. Our opti-mized MAE method has excellent repeatability demon-strated not only over time (low coefficients of variation peruser over 3 weeks) but also between users (non-significantone-way ANOVA and pairwise comparison test).

Recovery of SDG Standard

Four MAE tests were carried out in order to assess (1) thepotential of MAE to decompose SDG to SECO orANSECO and (2) the magnitude of the cumulative effectsof MAE, additional preparations for non-defatted samplesand HPLC analysis on the recovery of SDG standard. Therewere no SECO or ANSECO detected in the MAE extracts,meaning that the method did not produce lignan aglyconeand/or artefacts. The mean value for the recovery of SDGwas 97.5% with a standard deviation of 0.61 and acoefficient of variation <1%. We have previously reported97% recovery of o-coumaric acid used as the internalstandard for MAE in the presence of flaxseed samples(Nemes and Orsat 2009). Lignan recovery results account-ing for the cumulative effects of the extraction, preparationand analysis procedures were also reported by Thompson etal. (2007); the percentage of recovered lignans ranged from

Table 2 Repeatability of the optimized MAE method tested by three users over 3 weeks

Individual performingthe MAE test

Mean value of SDG(mg/g fresh weight flaxseed)

Standarddeviation

Coefficient ofvariationa (%)

Pairwise comparisonb,Tukey–Kramer HSD test

1 10.4 0.37 3.56 A

2 10.7 0.43 4.03 A

3 10.3 0.24 2.34 A

a The coefficient of variation was calculated as percentage from the mean (standard deviation × 100/mean)b The same letter (A) under the pairwise comparison test heading indicates that there is no significant difference (p=0.551) between the three meanvalues of SDG

556 Food Anal. Methods (2012) 5:551–563

73.8 to 92.3. Our optimized MAE method has a better SDGstandard recovery, which could be explained in part notonly by the much simpler procedure (less manipulationsteps) but also by its robustness (no lignan degradation).Although the cumulative losses were very small (2.55%), amultiplication correction factor was computed (100/97.5=1.026) in order to be able to account for the lossesoccurring during analytical MAE, sample preparation andanalysis of SDG in oil seeds. Thus, original lignan valuescan be presented side by side with the corrected values.

MAE Efficiency

The efficiency of the optimized MAE method was assessedby a comparison with two well-known published methods(referred to in this paper as reference methods A and B).The samples extracted with the three methods were fromthe same batch of DFM (flaxseed grown in Guatemala).The results are presented in Table 3. The SDG extractionyield obtained with MAE was the highest and wassignificantly different (p<0.0001) from the extraction yieldsobtained with the two reference methods. The SDGrecoveries of reference methods A and B relative to theMAE method were 78.6% and 73.4%, respectively. Asopposed to reference method B, method A has an improvedinitial extraction step. Although both methods use 70%methanol, for method A this is supplemented with 0.3 MNaOH. However, this concentration of NaOH is not highenough to ensure a complete SDG extraction from the matrix.Eliasson et al. (2003) have demonstrated that alcoholicextraction followed by alkaline hydrolysis is superseded bya simple direct alkaline hydrolysis with 1 M NaOH for 1 h atroom temperature. In addition, we have previously proventhat the optimized MAE using 0.5 M NaOH achieved anincrease of 6% in the SDG extraction yield as opposed to thedirect hydrolysis method (Nemes and Orsat 2009).

Sequential MAE (MAE-AC)

This experiment was carried out in order to verify whetherprolonged MAE extraction after acidification of extracts to

pH 3 with H2SO4 improves the lignan extraction yield fromflaxseed samples. This hypothesis was inspired by thefindings of Smeds et al. (2007) which indicated thatintroducing an additional acidic hydrolysis step betweenthe alkaline and enzymatic hydrolysis steps increased thelignan extraction yield. The assumption is that the increasein the lignan extraction yield was due to the additionalacidic hydrolysis step rather than to the subsequentenzymatic hydrolysis. It is known that the enzymatichydrolysis does not break the lignan macromoleculeefficiently, its main benefit being the production of lignanaglycones from the already released lignan glucosides(Oomah and Hosseinian 2008). The DFM used for thisexperiment was from the same batch that was used for theMAE efficiency tests (flaxseed grown in Guatemala). Thesequential MAE did not produce SECO or ANSECO. Theeffects of the MAE-AC factors time (3 and 9 min) andtemperature (60 and 90 °C) and their interaction did nothave significant effects on the SDG extraction yield; theirrespective p values were 0.162, 0.333 and 0.130. Thepredicted maximum value for SDG was 22.9 mg/DFM(standard error = 0.22) for 3 min MAE-AC at 60 °C. Thisresult is well depicted in the response surface plot (Fig. 2).The maximum SDG value of 22.9 mg/g DFM obtainedwith MAE-AC is equal to the MAE reference SDG valueshown in Table 3. This indicates that the additional MAE-AC step is not necessary as it does not increase the SDGextraction yield. Therefore, it can be concluded that theMAE of SDG from flaxseed is complete as a furtherincrease of SDG is not possible and robust as SDG is nottransformed into SECO or ANSECO.

The complete proof of repeatability and efficiency,provided above for the optimized MAE method, alongwith its simplicity and rapidity, makes MAE the method ofchoice for analytical quantification of SDG in flaxseed andother plant samples. In order to demonstrate its utility andthe range of samples that can be analysed with it, theoptimized MAE method was further applied for theanalytical quantification of lignans in various flaxseedcultivars, flax hull, sesame seeds, chia seeds, sunflowerseeds and almonds.

Table 3 Assessment of MAE efficiency by comparison with two reference methods

Extraction method Mean value of SDG(mg/g DFM)

Standard deviation No. of replicates Pairwise comparisona,Tukey–Kramer HSD test

MAE 22.9 0.26 4 A

Reference method A 18.0 0.43 3 B

Reference method B 16.8 0.03 3 C

aDifferent letters under the pairwise comparison test heading indicate that there are significant differences (p<0.0001) between the mean values ofSDG obtained with the three methods of extraction

Food Anal. Methods (2012) 5:551–563 557

Quantification of SDG in Flaxseed Cultivars

The optimized and evaluated MAE method was used forthe quantification of SDG in flaxseed cultivars grown in theQuébec province in 2009. The results of the pairwisecomparison tests, presented in Table 4, are the mean ofthree replicates and are expressed as milligrams SDG pergram of fresh weight flaxseed meal. In order to account forthe cumulative losses of SDG that occurred during MAE,sample preparation and HPLC analysis, the corrected SDGvalues are also presented in Table 4. The corrected valuewas calculated by multiplying the mean SDG values by1.026 (correction factor computed based on the recovery of

SDG standard, presented in “Recovery of SDG Standard”).The letters assigned based on the pairwise comparisontest categorize the Québec flaxseed cultivars into threegroups—A, B and C—which vary from each othersignificantly (p<0.0001). Group A has the highest SDGvalue (18.9 mg/g flaxseed meal) and includes only thecultivar 09LS01. A model of HPLC chromatogram forMAE extracts of the 09LS01 flax cultivar is shown inFig. 3. Group B is for the medium SDG values (16.5–17.3) and includes four cultivars: CDC Bethune, Flanders,CRGL 8.1 and 8.2. Group C is for the low SDG values(14.6–14.9) and includes the cultivars McBeth and PrairieBlue. For comparison purposes, the flaxseed sample(unknown cultivar grown in Guatemala) bought from alocal grocery store in 2009 was also included in the tablein order to highlight that all the flaxseed cultivars grownin Québec analysed by MAE contained significantly moreSDG per gram of fresh weight flaxseed meal.

Previous reports for SDG content in flaxseed cultivars,presented on an oil-free dry meal (OFDM) basis, include11.9–25.9 mg SDG per gram OFDM in Swedish flaxseedcultivars (Eliasson et al. 2003), 11.7–24.1 mg SDG pergram OFDM in Swedish and Danish cultivars (Johnsson etal. 2000), and 12.9–14.3 mg SDG per gram OFDM inAmerican cultivars (Madhusudhan et al. 2000). Westcott etal. (2002) reported SDG concentrations ranging from 11.1to 17.6 mg/g defatted meal for flax cultivars grown atvarious locations in Canada between 1995 and 1998. Otherpublished SDG concentration ranges include 1–10 μM(0.686–6.867 mg) SDG per gram flaxseed and 1.2–1.7%SDG (12–17 mg SDG per gram meal) in commercial flaxmeal obtained from Canadian cultivars (Daun et al. 2003).The latter SDG concentration range is similar to that

Fig. 2 Response surface plot for SDG during MAE-AC

Table 4 Quantification of SDG in flaxseed cultivars grown in Québec by MAE

Flaxseed cultivars Mean value of SDG(mg/g fresh weightflaxseed meal)

Standarddeviation

Coefficient ofvariationa (%)

Pairwise comparisonb,Tukey–Kramer HSD test

Corrected SDG valuec

(mg/g fresh weightflaxseed meal)

09LS01 18.4 0.49 2.66 A 18.9

CRGL 8.1 16.8 0.28 1.67 B 17.3

CDC Bethune 16.3 0.25 1.53 B 16.7

Flanders 16.2 0.07 0.43 B 16.6

CRGL 8.2 16.1 0.16 0.99 B 16.5

McBeth 14.5 0.22 1.52 C 14.9

Prairie Blue 14.2 0.23 1.62 C 14.6

Unknown cultivard 10.4 0.33 3.17 D 10.7

a The coefficient of variation was calculated as percentage from the mean (standard deviation × 100/mean)b Different letters under the pairwise comparison test heading indicate that there are significant differences (p<0.0001) between the mean values ofSDG obtained for the various flaxseed cultivarsc The mean value of SDG was multiplied with the correction factor of 1.026 in order to obtain the corrected SDG valued The flaxseed bought from a local grocery store was an unknown cultivar grown in Guatemala

558 Food Anal. Methods (2012) 5:551–563

published by Westcott et al. (2002) for defatted meal forCanadian cultivars. One publication reported SDG contentsof three Canadian flaxseed cultivars expressed on a freshweight flaxseed basis as follows: cultivar AC Linora had9 mg SDG per gram of seed, cultivar Flanders had 7.4 mgSDG per gram of seed, and cultivar Linola 989 had 15 mgSDG per gram of seed (Spence et al. 2003). Thompson et al.(1997) analysed ten flaxseed cultivars grown in Manitoba in1989 by in vitro fermentation with human faecal inoculumand found 0.96–3.15 μM lignan per gram seed and greatvariations between the cultivars. However, the coefficients ofvariation calculated from the published data were 11.7–45.8%, which are much greater than the current acceptablecoefficients of variation for lignan analysis methods (max,7%), as specified in “MAE Repeatability.” The coefficientsof variation recorded for the MAE of flaxseed cultivarsshown in Table 4 range from 0.43% to 3.17%.

Lignan values expressed on a fresh weight basis arerequired for database construction (Milder et al. 2005;Thompson et al. 2006) as they give valuable information interms of lignan concentration per edible serving and can beused in epidemiological and supplementation studies inorder to find relations between SDG content and healthbenefits (Adolphe et al. 2010; Peterson et al. 2010). TheSDG content of flaxseed cultivars obtained with MAEcannot be compared with those expressed on an oil-free drybasis for the Danish and Swedish cultivars or with thoseexpressed on a defatted meal basis from Canadian cultivars.Based on the results obtained with MAE for the extractionof defatted and non-defatted meal from the same flaxseedmaterial (Table 4, 10.43 mg SDG per gram non-defattedmeal, and Table 3, 22.9 mg SDG per gram defatted meal), itcan be concluded that the SDG concentrations in defattedflaxseed materials should be at least 200% higher. Even if

the results were expressed on a fresh weight flaxseed basis,the comparison would be unjust towards flaxseed cultivarsanalysed with methods less efficient than the optimizedMAE method. Moreover, the ranges for SDG concentrationin flaxseed or flaxseed meal depend on the year of growth,the cultivated variety, the location of growth (Eliasson et al.2003; Johnsson et al. 2000; Westcott et al. 2002; Westcottand Muir 2003), and the methodology of extraction andanalysis (Daun et al. 2003; Oomah and Hosseinian 2008).Here, we have demonstrated that the flaxseed cultivarsgrown in the Québec province can be classified into threelignan cultivar groups (high, medium and low) based ontheir SDG content. Such a classification could be veryuseful for narrowing down varieties of flaxseed that arelikely to produce similar amounts of lignan when grown insimilar conditions. High lignan flaxseed varieties aredesired for flax dehulling processes in order to obtain highlignan flax hull. In addition, high lignan flaxseed cultivarsare attractive for incorporation into health food productsand dietary supplements. The optimized MAE method isrecommended for future applications such as the analysis ofSDG in flaxseed cultivars in view of database constructionand also for adding lignan values to the Canadian qualitydata tables for flaxseed. The quality data for the Canadianflaxseed published annually by the Canadian Grain Com-mission include the oil, protein and free fatty acid contents,the fatty acid composition and the iodine values (Puvirajah2011). The flaxseed cultivars registered for production inCanada must have quality attribute values that are notstatistically significantly lower than those of the checkflaxseed cultivar. The check cultivar is determined everyyear by the Flax Evaluation Committee. The current checkcultivar, which has been used for the past few years, isFlanders (Anonymous 2008, 2011). Although the lignancontent is not one of the quality attributes being monitoredyet by the Canadian Grain Commission, if this would be thecase, the statistical analysis presented in Table 4 indicatesthat the cultivars McBeth and Prairie Blue do not have therequired minimum lignan content. Setting a minimumstandard for lignan content for flaxseed cultivar registrationwould contribute to an improved nutraceuticals quality ofnewer flaxseed cultivars.

Quantification of Lignans in Other Samples

The flax hull contained on average 39.0 mg SDG per gramof fresh weight sample (SD = 0.1; SDG value corrected forlosses, 40.0 mg/g), which was 2.12 to 3.73 times higherthan the SDG contents of the flaxseed cultivars presented inTable 4. Oomah and Sitter (2009) reported 31.2 mg SDGper gram defatted flax hull, as analysed with a directhydrolysis method. The authors investigated six methods ofoil extraction from the hulls prior to SDG extraction and

Fig. 3 Model of HPLC chromatogram for MAE extract of 09LS01flaxseed cultivar. This chromatogram is also representative for MAEextracts of flax hull, defatted and non-defatted flaxseed meal, and it issimilar with a previously published HPLC chromatogram for MAE ofDFM (Nemes and Orsat 2010)

Food Anal. Methods (2012) 5:551–563 559

found that oil extraction with acetone and ethanol resultedin defatted hulls with significantly lower SDG content asopposed to oil extraction by cold press, hexane, petroleumether and supercritical CO2. Although Oomah and Sitter(2009) analysed defatted flax hull and we analysed fresh(non-defatted) hull, some sources of variation, whichaccount for the marked difference between the two results,can be identified. Both studies analysed flax hull from thesame producer. However, the material was produced indifferent years (the present study used flax hull produced in2010), and apparently, different cultivars were used as rawmaterial for the dehulling process.

Flax hull has a naturally high concentration of SDG thathas potential for future epidemiological or supplementationstudies. For example, only 12.5 g of flax hull (calculatedbased on the MAE analysis of flax hull) would be requiredin order to provide the 500 mg SDG, which, according toAdolphe et al. (2010), taken on a daily basis would reducethe risk of cardiovascular diseases.

In order to demonstrate the applicability of the MAEmethod for lignan analysis in plant samples other thanflaxseed, MAE was carried out for sesame seeds, chiaseeds, sunflower seeds and almonds. The SECO content ofsesame and chia seeds seemed to fall in the same range ofvalues; therefore, they are presented together in Table 5.SECO values corrected for losses are presented as well as itis assumed that SECO has a recovery similar to SDG in theabsence of artefact formation during MAE and additionalsample preparation. The differences in SECO concentrationin sesame seeds were significant and large enough in orderto be attributed to different seed colours (varieties) ratherthan different samples. Black sesame had the highest SECOconcentration, 1.898 mg/g of fresh weight sample, which isabout 2.2 and 10.2 times higher than that found in whiteand brown sesame seeds, respectively. However, the SECOconcentration does not correlate with the intensity of theseed colour since brown sesame had the lowest concentra-tion, rather than medium, as opposed to the white and black

Table 5 Quantification of SECO in sesame and chia seeds by MAE

Sample information Mean value SECO(mg/g fresh weight meal)

Standard deviation Pairwise comparisona,Tukey–Kramer HSD test

Corrected SECO valueb

(mg/g fresh weight meal)

White sesame 0.849 0.002 C 0.871

Brown sesame 0.181 0.008 D 0.186

Black sesame 1.850 0.003 A 1.898

White chia 1.257 0.135 B 1.289

Black chia sample 1 0.974 0.058 C 0.999

Black chia sample 2 1.026 0.070 BC 1.053

a Different letters under the pairwise comparison test heading indicate that there are significant differences (p < 0.0001) between the mean valuesof SECO obtained for the tested samplesb The mean value of SECO was multiplied with the correction factor of 1.026 in order to obtain the corrected SECO value

Fig. 4 Model of HPLC chromatogram for MAE extract of blacksesame seeds. This chromatogram is also representative for MAEextracts of white and brown sesame seeds

Fig. 5 Model of HPLC chromatogram for MAE extracts of black chiaseeds. This chromatogram is also representative for MAE extracts ofwhite chia seeds

560 Food Anal. Methods (2012) 5:551–563

seeds. A model of HPLC chromatogram for MAE extract ofblack sesame is shown in Fig. 4.

Lignan aglycone values (sum of SECO, pinoresinol,matairesinol and lariciresinol) for sesame seeds (seed colouris assumed to be white as it is the common one sold ingrocery stores) reported in the literature by Thompson et al.(2006) and Milder et al. (2005) were 0.079 and 0.393 mglignan aglycone per gram of fresh seeds, respectively. It isnot known whether SECO is naturally conjugated tocarbohydrates or other compounds in sesame seeds. Ifsesame contains oil-soluble conjugated SECO, this wouldbe lost during sample defatting with hexane. Therefore,only non-defatted sesame meal samples were extracted withMAE, which could explain in part the much higher resultsobtained with MAE. In the literature, lignans other thanSECO were reported both in sesame oil and in defattedsesame samples. The oil-soluble lignans sesamolin andsesamin were found in concentrations up to 2.97 and7.12 mg/g of fresh weight sesame samples, respectively, byMoazzami and Kamal-Eldin (2006). The polar sesaminoldiglucoside and triglucoside were found in alcoholicextracts (non-hydrolysed) of defatted sesame samples inconcentrations up to 4.93 and 15.6 mg/g defatted sample,respectively, by Moazzami et al. (2006). Unlike for SECO,there was no significant difference found in the content ofoil-soluble lignans and lignan glucosides between whiteand black sesame seeds. The fact that the SECO yields for

white sesame seeds obtained with MAE were much higherthan those reported in the literature (sum of lignanaglycones) can also be attributed to the high efficiency ofMAE for releasing lignans from plant matrices as opposedto other lignan extraction methods.

The only significant difference in SECO content for chiaseeds was recorded between white chia and black chiasample 1. Here, the variations in SECO content seem to beattributable to the different samples rather than differentcolours (varieties) as both white and black chia sample 1are not significantly different from black chia sample 2. Thetwo black chia samples and white sesame have similarSECO contents. A model of HPLC chromatogram for MAEextract of black chia is shown in Fig. 5.

To the best of our knowledge, this is the first report of SECOin chia seeds. However, SECO was previously reported byPowell and Plattner (1976) and Plattner and Powell (1978) inSalvia plebeia R. brown seeds. Both chia—S. hispanica andS. plebeia—belong to the genus Salvia L. As a common trait,they are both mucilaginous seeds (Dweck 2000), which isalso shared by flaxseeds. In S. plebeia, SECO is found in theform of a SECO diester and a SECO branched fatty diester.The first diester was extracted with a mixture of pentane andhexane and yielded SECO, 12-methyltetradecanoic andferulic acids upon alkaline hydrolysis (Powell and Plattner1976). The second diester was extracted with hexane andyielded SECO and 12-methyltetradecanoic acid upon alkaline

Fig. 6 Model of HPLCchromatogram for MAE extractof sunflower seeds

Fig. 7 Model of HPLC chro-matogram for MAE extract fromalmonds

Food Anal. Methods (2012) 5:551–563 561

hydrolysis (Plattner and Powell 1978). Only non-defattedchia samples were extracted with MAE. If SECO is ester-linked to fatty and/or cinnamic acids in chia as it is in S.plebeia, then all SECO found in chia would be released byMAE.

Samples of non-defatted sunflower seeds and almondswere also extracted with MAE. Sunflower seeds contained0.045 mg SECO per gram of fresh weight sample with astandard deviation of 0.005; after multiplication with thecorrection factor for losses (1.026), the SECO value is0.046 mg/g of fresh weight sample. Almonds contained0.028 mg SDG per gram of fresh weight sample with astandard deviation of 0.001; the SDG value corrected forlosses and its equivalent SECO value are 0.029 and0.015 mg/g of fresh weight sample, respectively. Othershave reported lignan contents for sunflower seeds andalmonds (sum of SECO, matairesinol, pinoresinol andlariciresinol) as follows: 0.009 mg lignan per gram of freshweight sunflower (Milder et al. 2005), 0.002 mg lignan pergram of fresh weight sunflower and 0.001 mg lignan pergram of fresh weight almonds (Thompson et al. 2006). Thelignan contents found in both sunflower seeds and almondswith MAE are much higher than the values published in theliterature. Models of HPLC chromatograms for MAEextracts of sunflower seeds and almonds are shown inFigs. 6 and 7, respectively.

The lignan yields obtained with MAE for samples otherthan flaxseed are much higher than the range of valuespublished in the literature, thus further demonstrating theefficiency of the MAE method and its potential ofbecoming a generally applicable method for analyticallignan quantification in plants. In most plant sources,lignans occur conjugated to a variety of carbohydrates;thus, in order for the optimized MAE to be generallyapplicable, it has to be followed by an optimized enzymatichydrolysis method.

Conclusions

The novelty of this work resides in the complete proof ofefficiency, repeatability and reliability of an optimizedMAE method for the analytical quantification of lignansin plant materials. In addition, the applicability of the MAEmethod was demonstrated for a variety of plant samplessuch as flaxseed cultivars, flaxseed hulls, sesame seeds,chia seeds, sunflower seeds and almonds. Therefore, it canbe concluded that the MAE method was evaluated for thegeneral analytical quantification of lignans in these sam-ples. For extending its applicability to the general analyticalquantification of lignans in plants, the MAE method has tobe followed by an optimized enzymatic hydrolysis. Theoptimized MAE method is recommended for applications

such as the classification of flaxseed cultivars based on theirlignan content, for construction of lignan databases forplant foods and for adding lignan values to the quality datatables for flaxseed.

Acknowledgements The authors thank the Natural Sciences andEngineering Research Council of Canada (NSERC), the FondsQuébécois de la Recherche sur la Nature et les Technologies(FQRNT), the Canada Foundation for Innovation (CFI) and the ABIPprogramme of Agriculture and Agri-Food Canada for the financialsupport. Thanks are due to Dr. Nam Fong Han from Natunola HealthInc. for providing the flax hulls and to Mr. Yves Dion from Centre deRecherche sur les Grains (CEROM) for providing the flaxseedcultivars. The help of Miss Yanti Yusoff, Mr. Maxime Ouelette-Payeur and Mr. Scott Hong Liang for carrying out the MAErepeatability tests is much appreciated.

References

Adolphe JL, Whiting SJ, Juurlink BHJ, Thorpe LU, Alcorn J (2010)Health effects with consumption of the flax lignan secoisolaricir-esinol diglucoside. Br J Nutr 103:929–938. doi:10.1017/S0007114509992753

Anonymous (2008) Flaxseed quality in Canada—Canadian flaxquality shows long term consistency and improvement. CanadianGrain Commission. http://www.grainscanada.gc.ca/flax-lin/trend-tendance/qfc-qlc-eng.htm

Anonymous (2011) Appendix C—Operating procedures for the registra-tion of a flax cultivar for production in Western Canada. PrairieRecommending Committee for Oil Seeds (PRCO). http://www.pgdc.ca/pdfs/oc/2011%20PRCO%20Operating%20Procedures%20Appendix%20C%20%28Flax%29.pdf. Accessed March 2011,pp 1–18

Beejmohun V, Fliniaux O, Grand E, Lamblin F, Bensaddek L, ChristenP, Kovensky J, Fliniaux MA, Mesnard F (2007) Microwave-assisted extraction of the main phenolic compounds in flaxseed.Phytochem Anal 18:275–282

Daun JK, Barthet VJ, Chornick TL, Duguid SD (2003) Structure,composition, and variety development of flaxseed. In: ThompsonLU, Cunnane SC (eds) Flaxseed in human nutrition. AOCS,Champaign, pp 1–40

Dweck AC (2000) The folklore and cosmetic use of various Salviaspecies. In: Kintzisos SE (ed) Sage: the genus Salvia. HarwoodAcademic, Amsterdam, pp 1–26

Eliasson C, Kamal-Eldin A, Andersson R, Aman P (2003) High-performance liquid chromatographic analysis of secoisolaricire-sinol diglucoside and hydroxycinnamic acid glucosides inflaxseed by alkaline extraction. J Chromatogr A 1012:151–159

Ford JD, Huang KS, Wang HB, Davin LB, Lewis NG (2001)Biosynthetic pathway to the cancer chemopreventive secoisolar-iciresinol diglucoside-hydroxymethyl glutaryl ester-linked lignanoligomers in flax (Linum usitatissimum) seed. J Nat Prod64:1388–1397

Johnsson P, Kamal-Eldin A, Lundgren LN, Aman P (2000) HPLCmethod for analysis of secoisolariciresinol diglucoside in flax-seeds. J Agric Food Chem 48:5216–5219

Johnsson P, Peerlkamp N, Kamal-Eldin A, Andersson RE,Andersson R, Lundgren LN, Aman P (2002) Polymericfractions containing phenol glucosides in flaxseed. FoodChemistry 76:207–212

Kamal-Eldin A, Peerlkamp N, Johnsson P, Andersson R, AnderssonRE, Lundgren LN, Amar P (2001) An oligomer from flaxseed

562 Food Anal. Methods (2012) 5:551–563

composed of secoisolariciresinoldiglucoside and 3-hydroxy-3-methyl glutaric acid residues. Phytochemistry 58:587–590

Madhusudhan B, Wiesenborn D, Schwarz J, Tostenson K, Gillespie J(2000) A dry mechanical method for concentrating the lignansecoisolariciresinol diglucoside in flaxseed. Lebensmittel-Wissenschaft and Technologie 33:268–275

Milder IEJ, Arts ICW, Venema DP, Lasaroms JJP, Wahala K, HollmanPCH (2004) Optimization of a liquid chromatography–tandemmass spectrometry method for quantification of the plant lignanssecoisolariciresinol, matairesinol, lariciresinol, and pinoresinol infoods. J Agric Food Chem 52:4643–4651

Milder IEJ, Arts ICW, van de Putte B, Venema DP, Hollman PCH(2005) Lignan contents of Dutch plant foods: a databaseincluding lariciresinol, pinoresinol, secoisolariciresinol and mat-airesinol. Br J Nutr 93:393–402

Moazzami A, Kamal-Eldin A (2006) Sesame seed is a rich source ofdietary lignans. Journal of the American Oil Chemists’ Society83:719–723. doi:10.1007/s11746-006-5029-7

Moazzami AA, Andersson RE, Kamal-Eldin A (2006) HPLC analysisof sesaminol glucosides in sesame seeds. J Agric Food Chem54:633–638. doi:10.1021/jf051541g

Nemes SM, Orsat V (2009) Microwave-assisted extraction ofsecoisolariciresinol diglucoside—method development. Foodand Bioprocess Technology 1–9. doi:10.1007/s11947-009-0213-z

Nemes SM, Orsat V (2010) Screening the experimental domain for themicrowave-assisted extraction of secoisolariciresinol diglucosidefrom flaxseed prior to optimization procedures. Food and Bio-process Technology 3:300–307. doi:10.1007/s11947-009-0212-0

Nollet LML (2000) Food analysis by HPLC, 2nd edn. Marcel Dekker,New York

Oomah BD, Hosseinian FS (2008) Phytoestrogens. In: Hurst WJ (ed)Methods of analysis for functional foods and nutraceuticals.CRC, Boca Raton, pp 1–84

Oomah BD, Sitter L (2009) Characteristics of flaxseed hull oil. FoodChemistry 114:623–628

Penalvo JL, Haajanen KM, Botting N, Adlercreutz H (2005) Quantifi-cation of lignans in food using isotope dilution gas chromatography/mass spectrometry. J Agric Food Chem 53:9342–9347

Penalvo JL, Adlercreutz H, Uehara M, Ristimaki A, Watanabe S(2008) Lignan content of selected foods from Japan. J AgricFood Chem 56:401–409

Peterson J, Dwyer J, Adlercreutz H, Scalbert A, Jacques P,McCullough ML (2010) Dietary lignans: physiology andpotential for cardiovascular disease risk reduction. Nutr Rev68:571–603. doi:10.1111/j.1753-4887.2010.00319.x

Plattner RD, Powell RG (1978) A secoisolariciresinol branched fattydiester from Salvia plebeia seed. Phytochemistry 17:149–150.doi:10.1016/s0031-9422(00)89701-0

Powell RG, Plattner RD (1976) Structure of a secoisolariciresinoldiester from Salvia plebeia seed. Phytochemistry 15:1963–1965

Puvirajah AS (2011) Quality of Western Canadian Flaxseed 2010.Canadian Grain Commission. www.grainscanada.gc.ca, pp 1–14

Schwartz H, Sontag G, Plumb J (2009) Inventory of phytoestrogendatabases. Food Chemistry 113:736–747

Sicilia T, Niemeyer HB, Honig DM, Metzler M (2003) Identificationand stereochemical characterization of lignans in flaxseed andpumpkin seeds. J Agric Food Chem 51:1181

Smeds AI, Eklund PC, Sjoholm RE, Willfor SM, Nishibe S, DeyamaT, Holmbom BR (2007) Quantification of a broad spectrum oflignans in cereals, oilseeds, and nuts. J Agric Food Chem55:1337–1346

Smeds AI, Jauhiainen L, Tuomola E, Peltonen-Sainio P (2009)Characterization of variation in the lignan content and composi-tion of winter rye, spring wheat, and spring oat. J Agric FoodChem 57:5837–5842. doi:10.1021/jf9004274

Spence JD, Thornton T, Muir AD, Westcott ND (2003) The effect offlax seed cultivars with differing content of α-linolenic acid andlignans on responses to mental stress. J Am Coll Nutr 22:494–501

Struijs K, Vincken J-P, Verhoef R, van Oostveen-van Casteren WHM,Voragen AGJ, Gruppen H (2007) The flavonoid herbacetindiglucoside as a constituent of the lignan macromolecule fromflaxseed hulls. Phytochemistry 68:1227–1235

Struijs K, Vincken JP, Verhoef R, Voragen AGJ, Gruppen H (2008)Hydroxycinnamic acids are ester-linked directly to glucosylmoieties within the lignan macromolecule from flaxseed hulls.Phytochemistry 69:1250–1260

Struijs K, Vincken J-P, Doeswijk TG, Voragen AGJ, Gruppen H (2009)The chain length of lignan macromolecule from flaxseed hulls isdetermined by the incorporation of coumaric acid glucosides andferulic acid glucosides. Phytochemistry 70:262–269.doi:10.1016/j.phytochem.2008.12.015

Thompson LU, Rickard SE, Cheung F, Kenaschuk EO, Obermeyer WR(1997) Variability in anticancer lignan levels in flaxseed. Nutritionand Cancer 27:26–30. doi:10.1080/01635589709514497

Thompson LU, Boucher BA, Zhen L, Cotterchio M, Kreiger N (2006)Phytoestrogen content of foods consumed in Canada, includingisoflavones, lignans, and coumestan. Nutrition & Cancer 54:184–201

Thompson LU, Boucher BA, Cotterchio M, Kreiger N, Liu Z (2007)Dietary phytoestrogens, including isoflavones, lignans, andcoumestrol, in nonvitamin, nonmineral supplements commonlyconsumed by women in Canada. Nutrition and Cancer 59:176–184

Westcott ND, Muir AD (2003) Chemical studies on the constituens ofLinum spp. In: Muir AD, Westcott ND (eds) Flax: the genusLinum. Routledge, London, pp 55–73

Westcott ND, Muir AD, Lafond G, McAndrew DW,MayW, Shirtliffe S,Bruulsema TW (2002) Factors affecting the concentration of anutraceutical lignan in flaxseed. In: Bruulsema TW (ed) Proceed-ings of the Symposium on Fertilizing Crops for Functional Foods,Potash & Phosphate Institute/Potash & Phosphate Institute ofCanada (PPI/PPIC), Indianapolis, Indiana, USA, pp 1–4

Zhang W, Xu S (2007) Microwave-assisted extraction of secoisolar-iciresinol diglucoside from flaxseed hull. Journal of the Scienceof Food and Agriculture 87:1455–1462

Food Anal. Methods (2012) 5:551–563 563