Quick determination of erucic acid in mustard oils and seeds

Erucic acid is known to have negative effects on the human health and is therefore regulated in nutritional products such as vegetable oils. In order to determine the content of erucic acid in mustard oil an oil must be pressed from seeds, the oil derivatized into fatty acid methyl esters and then analyzed by gas chromatography. By using thermally assisted hydrolysis and methylation we were able to develop a method that can determine the content of erucic acid in mustard oils directly from the mustard seeds, thus avoiding time consuming pressing and off-line derivatization steps. Eleven samples have been tested and the results are in good agreement with conventional oil analysis. It could further be shown that even mustard varieties which are listed as erucic acid free can produce certain amounts of this fatty acid under certain environmental conditions, which supports the need for a fast and reliable screening method which enables analysis directly from the seeds.


Introduction
Mustard grains have been used for centuries in many different applications.They all belong to the Brassicaceae family either under the genus Brassica with Brassica juncea (brown mustard) and Brassica nigra (black mustard) or the genus Sinapis with Sinapis alba (white or yellow mustard) [1,2].
The composition of oil content in Brassicaceae is of interest due to the possible formation of erucic acid, a naturally occurring 22 carbon acid with double bond on C-13.Erucic acid (cis-13-docosenoic acid) belongs to the family of n-9 ("omega-9 ′′ ) fatty acids and is built by elongation of oleic acid, a n-9 18 carboxylic acid [3].The formation of erucic acid is influenced by varietal as well as environmental influences such as temperature during germination, irrigation, day length, and soil fertility [4,5].Jones, Coutts, and Hawkes showed in 2007 that glucosinolate levels in Brassica napus (rapeseed) increased due to beet western yellow virus infection, leading to an erucic acid content of up to 44% [6].This shows that even erucic acid-free varieties can form erucic acid again due to external influences.
To date, mustard oil is scarcely used in the European diet but is very common in Asia and especially the North of India [3].As a diet rich in erucic acid may have a negative impact on health by accumulation of triacylglycerol in the heart, so-called myocardial lipidosis [7], the maximum levels of erucic acid in Europe are regulated since 1976.The maximum content of erucic acid in oils, fats, and food containing oils or fats was initially limited to 5% of the total level of fatty acids in the fat component.[8].In 2019 the Commission Regulation 2019/1870, amending regulation 1881/2006, entered into force where the maximum content of erucic acid in oils and fats was set to 20 g/kg, in mustard oil to 50 g/kg and in table mustard to 35 g/kg.These maximum levels again refer to the total level of fatty acids in the fat component of the food.It was further stated that, with the approval of the competent authority, the maximum level does not apply to mustard oil produced and consumed locally [9].However, Rastogi et al. found, that the overall effect of mustard oil is beneficial due to the high content of α-linolenic acid [10].
The main application for mustard seeds in Europe is table mustard (a.k.a.mustard).It was shown that already two servings of table mustard could surpass the Australian TDI (tolerable daily intake) of erucic acid.This holds true especially for younger people due to their lower body weight [11].This study led to a re-evaluation request of erucic acid by the EFSA.In 2016 the EFSA Panel on Contaminants in the Food Chain (CONTAM Panel) published a scientific opinion on erucic acid in feed and food and recommended generating more analytical data especially for foods consumed by infants and small children [12].
Thermally assisted hydrolysis and methylation or short THM is an analytical pyrolysis technique that was first applied by Challinor in 1989 [13].In contrast to conventional pyrolysis where typically high temperatures are used to enable breaking of C-C bonds, THM is carried out in the presence of an organic base, mostly tetramethylammonium hydroxide, at lower temperatures.Due to the high alkalinity polar bonds such as esters, ethers, amides, etc. are cleaved and in a second step a methyl group is transferred to the thus formed products.The technique has proven to be very valuable in the analysis of natural materials such as carbohydrates, lignocellulosics, soil, or proteins, but also for synthetic materials such as resins, polyesters, polyamides and more [14].
The analysis of fatty acids has also been a topic of research, for example their distribution in mosquitos and different nutrition [15], zooplankters [16], fungi [17], identification of oils in paints and varnishes [18][19][20][21], plant and animal oils [22,23], or the use of alternative methylation reagents [24,25].One major issue known in the THM analysis of fatty acids is the tendency of polyunsaturated fatty acids to isomerize or degrade [26].However, by proper selection of conditions, in particular lower concentration of the TMAH solution and lower pyrolysis temperatures, these negative effects can be minimized [15].
In this study we have investigated the possibility to determine the fatty acid composition, in particular the amount of erucic acid, of mustard seeds with thermally assisted hydrolysis and methylation.In contrast to the conventional approach, where an oil has to be pressed from the seeds and fatty acids are determined by transesterification and GC analysis, the pyrolytic technique is fast and can be carried out with extremely low sample amounts directly from the seeds.

Samples
In this study 11 samples of mustard seeds (Table 1) were investigated.All samples were sourced from commercial suppliers and harvested either in 2019 or 2020.

Oil pressing
A household oil press (Rommelsbacher Ölpresse OP 700 Emilio) was used to press few milliliters of single-variety oil from each seed.The intermediate batches were discarded.All oils were yellow in color and not cloudy.
A comparison was made to ensure that the fatty acid distribution in these oils was comparable to oils produced commercially.One and the same batch of mustard seeds was pressed in the laboratory with a household mill and pressed at a company with a conventional industrial scale oil mill (screw press).

FAME analysis
Derivatization of fatty acids to fatty acid methyl esters (FAME) was done in a two-step transesterification process followed by GC-MS analysis.Aliquots of the oils were mixed with 200 µL 0.2 M Sodium methoxide (Alfa Aesar, Thermo Fisher Scientific, Heysham, UK) and incubated at 60 • C for 45 min under shaking.After addition of 70 µL of M methanolic sulfuric acid, samples were incubated a further 30 min at 60 • C. FAME were extracted by addition of 600 µL saturated sodium chloride and 500 µL n-hexane (Suprasolv, Merck KGaA, Darmstadt, Germany).Pentadecanoic acid (C15:0, Sigma-Aldrich, St. Louis, MO, USA) was used as internal standard.GC-MS analysis was performed as described in [28] using a Trace 1300 gas chromatograph coupled to an ISQ QD single-quadrupole mass spectrometer equipped with a PTV injector (both Thermo Scientific, Waltham, MA, USA).Injector temperature was kept constant at 240 • C and samples were injected with a 1/100 split ratio onto a DB-23 column (60 m × 0.25 mm, 0.25 µm film thickness; Agilent, Santa Clara, CA, USA).Helium was used as carrier gas at a constant flow rate of 1.5 mL min − 1 .Oven temperature was kept constant at 55 • C for 2 min, then raised to 190 • C at 20 • C min − 1 and further to 240 • C at 3 • C min − 1 , followed by a constant period of 4 min at 240 • C before returning to start conditions.The injector and ion source temperatures were set to 240 • C and 250 • C, respectively.Full scans from m/z 40-400 were recorded and SIM scans at m/z 55, 67, 74 and were used for quantification with the Supelco 37 Component FAME Mix (Sigma-Aldrich, St. Louis, MO, USA) as external standard.Six-point calibration using a linear regression model was performed with concentrations of the individual fatty acids ranging from 2 to 100, 4-200 or 6-300 µg/mL.Peak areas of the SIM scans were used and the abundance ratio of each individual fatty acid methyl ester to the internal standard C15:0 was used for correction.Results are expressed as percentage of the total fatty acids in each sample (wt%).

Thermally assisted hydrolysis and methylation-GC-MS
Thermally assisted hydrolysis and methylation (THM) experiments were carried out with a Gerstel pyrolyzer system attached to a Trace GC Ultra (Thermo Electron Corp.) equipped with a capillary column Supelco SP-2330 (30 m x 0.32 mm × 0.2 µm), and a Polaris ion trap mass spectrometer (Thermo Electron Corp.).Pyrolysis was performed at 450 • C for 30 s unless mentioned otherwise.The cold injection system (CIS) was heated from 150 • to 280 • C with 10 • C s − 1 , the thermal transfer unit (TDU) programmed from 50 • to 280 • C with 720 • C min − 1 and the TDU transfer set to 300 • C. The GC column temperature conditions were as follows: initial temperature 100 • C, hold for 1 min, increase at 10 • C min − 1 to 230 • C, and hold this temperature for 5 min.Helium gas flow was set at 2 mL min − 1 , the total split flow was 100 mL min − 1 .Mass spectra were recorded under electron impact ionization at 70 eV electron energy in the range from m/z 35-500.For thermally assisted hydrolysis and methylation 3 µL of 10% aqueous tetramethylammonium hydroxide (TMAH, Fluka) solution were added to the 30-100 µg of sample in a closed bottom vial.In order to homogenize the samples, the several seeds were ground cryogenically with a Spex Sample Prep Freezer Mill 6770.

Pressing and analysis of mustard oils
All oils were subjected to conventional transesterification to FAMEs with subsequent GC-MS analysis.The resulting reference data is depicted in Fig. 1, showing the fatty acid composition of all analyzed mustard seeds processed in triplicates.Up to 15 different fatty acids were quantified per variety, with the most abundant ones being either oleic or erucic acid.Quantification was performed by external standard calibration, with the linear regression model revealing R 2 values > = 0.995 for each of the 37 individual standard FAMEs.
Operating conditions such as temperature or rotation speed might have an impact on oil yield and composition.In order to rule out any possible differences in the oil milling process between the household and the industrial scale mill, aliquots of the Martigena variety were processed in both types of mills.The resulting oils where derivatized to FAMEs and analyzed by GC-MS and did not show any significant difference in their fatty acid composition.It was therefore concluded that the household mill based pressing could be considered representative also for large-scale oil milling.

Thermally assisted hydrolysis and methylation
In a first step the reaction conditions, which were already optimized in a previous study for a different pyrolyzer [16], were investigated.One major difference is the use of different quartz vials with a closed bottom in contrast to using vials open on both sides and quartz wool to hold the sample in place.The vials have an additional slit in the side where the sample can enter into the GC interface, but this is limiting the maximal sample volume to about 5 µL.By using these vials, the reproducibility and linearity of the system was drastically increased.Calibration was carried out using a polyunsaturated fatty acid methyl ester standard containing 37 different FAMEs by injecting 0.5, 1 and 2 µL of the reference solution into the quartz vials and subjecting them to pyrolysis.For quantification the TIC (total ion current) was used.Each calibration point was measured in triplicate and via the slope of the calibration curve a relative response factor for each FAME was determined.In contrast to the conventional FAME analysis, palmitoleic acid could not be measured as it was coeluting with methoxybenzoic acid methyl ester, for vaccenic acid no standard was available, but additionally an 18:2 isomer was found which was not detected in the conventional analysis of oils.This latter isomer is most likely formed via isomerization during the pyrolysis process and was quantified using the calibration data of its structural isomer linolenic acid methyl ester.
In a first approach the method was tested with six seed samples obtained from Polarka Po2 and the results compared to the pyrolysis of the oil obtained from the same seed (3 replicates).As can be seen in Fig. 2 no relevant difference between the composition derived from the seeds and the oil can be seen.The highest deviation can be found in the values obtained for erucic acid with 23.3% for the seeds and 24.9% for the oils (standard deviations 3.6% and 2.4% respectively).This is insofar a necessary prerequisite for the desired analytical approach as it clearly shows that there is no compositional difference between the oil composition in the seeds and the oil composition of the pressed mustard

oil.
The analysis of all 11 seed samples was carried out in triplicate and the standard deviations for all fatty acids were found to be in an acceptable range (see Fig. 3 and Table 2).The trends are clearly the same when comparing Figs. 1 and 3 with, as already mentioned above, the major difference being the detection of an isomer of octadecadienoic acid (18:2 FAME) in the THM experiments.When looking at the absolute values listed in Table 2, the largest deviation is found for linolenic acid.This fatty acid is rather evenly distributed in all samples in concentrations of 12.1 + /-1.8%, however, the results derived from the THM experiments with 2.7 + /-1% are substantially lower.The content of linoleic acid on the other hand is rather similar in both methods, only the sample Raketa (Rk) has a value of 35.5% versus 26.5% in conventional and pyrolytic analysis, respectively.When we have a look at the newly formed 18:2 fatty acid isomer which is usually found in ranges of 3-4%, the Rk sample is again an exemption with 8.4%.This suggests that part of the linolenic acid is isomerized into the 18:2 isomer during the pyrolysis or to be more precise the thermally assisted hydrolysis and methylation process.Additionally, the loss of a double bond must occur, otherwise the low values for linolenic acid cannot be explained properly.This assumption is also supported by the slightly higher values for the unsaturated palmitic and stearic acid and the slightly higher values for oleic acid, which would be formed from polyunsaturated fatty acids such as linoleic or linolenic acid.
Most important for this study, however, is the determination of erucic acid and especially, if its content is above or below the 2 wt% allowance mark [9].Fig. 4 shows the comparison of the results obtained with both methods.The three Martigena seeds as well as the Raketa seed have the lowest amounts of erucic acid, all of them around 1%.Only for Martigena Ma1 the results of conventional analysis (1.0%) and thermally assisted hydrolysis and methylation (2.1%) differ, therefore not allowing a safe judgement from the latter result alone.Next lowest concentration is found in Raduga seed with 5.7%, followed by Carnella with 25%.The result for Raduga is in so far surprising as this variety should be free of erucic acid [28].The other seeds show erucic acid concentrations of more than 35%, all values as derived from the THM experiments.

Conclusions
We have successfully proven that thermally assisted hydrolysis and methylation aka THM is a powerful analytical tool for the determination of fatty acids in oil seeds.Without the need to produce an oil in the first step and do a time-consuming sample derivatization, the erucic acid content of mustard oils could be directly determined from the seeds with very good accuracy.The overall determination of the fatty acid composition was also possible with the exception of linolenic acid which seems to undergo isomerization and hydrogenation reactions under the chosen reaction conditions.Out of eleven seeds tested, seven could be identified as being not suited for the production of mustard oil, even though one of these seeds, Raduga, is listed as 'free of erucic acid' in the Austrian Descriptive Variety List (BSL) [28].This underlines the impact of growth conditions and environmental effects on the production of erucic acid by mustard plants, even in cases with favorable genetic foundation.The finding further highlights the demand for fast and reliable analytical methods for routine analysis of mustard seeds, such as THM.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. Comparison of THM-GC-MS results for six seed samples and three oils pressed from Polarka seed Po2.

Fig. 3 .
Fig.3.Fatty acid analysis of the eleven seeds as obtained from THM-GC-MS.Abundance of fatty acids is expressed as percentage of the total fatty acids in each sample.

Table 1
List of seed samples.
Fig.1.Conventional FAME analysis of the eleven oil samples after pressing from the seeds.B.Schwarzinger et al.

Table 2
Comparison of fatty acid composition obtained with conventional FAME analysis and THM-GC-MS.Results are given in %wt of total fatty acid content.