Macro‐ and microelements in pumpkin seed oils: Effect of processing, crop season, and country of origin

Abstract Macro‐ and microelements in the samples of virgin and cold pressed pumpkin seed oils produced in Croatia through two consecutive crop seasons were determined by inductively coupled plasma–optical emission spectroscopy (ICP‐OES). Croatian oils were also compared to oils from Slovenia and Austria in order to assess differences in the element content. Magnesium, potassium, calcium, sodium, selenium, and iron were the dominant elements in all pumpkin seed oils. Their amounts together with barium, strontium, manganese, copper were up to ninefold higher (p ≤ 0.05) in virgin compared to cold pressed pumpkin seed oils. These differences occur due to the different processing conditions which include salt addition, heat treatment, and higher degree of equipment ware out during virgin pumpkin seed oil production. As the sodium level increases with the addition of salt, virgin pumpkin seed oil could be considered its hidden source and producers should pay attention to the amount added. Contents of cobalt, copper, selenium, and thallium significantly differed (p ≤ 0.05) between the two crop seasons. Principal component analysis revealed clear differences between samples with different origin that can be explained by the specifics in the production processes of each country. In comparison with Austrian and Slovenian, Croatian pumpkin seed oils had significantly lower contents of sodium, potassium, calcium, magnesium, and tin while bismuth and selenium were higher.

On the other hand, Joebstl, Bandoniene, Meisel, and Chatzistathis (2010) compared the impact of geographical origin of pumpkin seeds on mineral composition of rare earth elements in pumpkin seed oil and concluded that, according to their concentrations, it is possible to distinguish pumpkin seed oils from Styria, Lower Austria, and China using discriminant analysis. However, possible influence of the production processes specifics related to the country of origin was not investigated.
Since it is known that the concentration of mineral substances in oils is mostly affected by production conditions and climatic conditions of a particular cultivation year (Schuster et al., 1983), this paper was aimed to determine and compare the quantities of macro-and microelements in the samples of virgin and cold pressed pumpkin oils produced in 2011 and 2012 in northern Croatia. Also to get insight into the processing parameters specific to the country of origin, Croatian oils were compared with the ones for oils from Slovenia and Austria.

| Pumpkin seed oils
Pumpkin seed oil samples from Croatia, Slovenia, and Austria were used in this study. Croatian oils had been collected from various pro- were conducted on 31 virgin oils and 8 cold pressed oils from Croatia as well as 9 virgin oils from Slovenia and 6 from Austria.

| Reagents
All the standards used in the analyses fulfilled the Atomic Spectroscopy Standard purity level. For ICP analysis, a multielement standard solution VIII: Al, B, Ba, Be, Bi Ca, Cd, Co, Cr, Cu, Fe, Ga, K, Li, Mg, Mn, Na, Ni, Pb, Se, Sr, Te, Tl, Zn obtained by Merck (Darmstadt, Germany) was used. Trace metals III: Ba, Ca, Mo, Na, K, Mg were purchased by Perkin Elmer Life and Analytical Sciences (Shelton, USA).
Tin, antimony, silver, vanadium, silicon, and lanthanum standards for atomic absorption were obtained from Carlo Erba Reagenti SpA (Rodano, Italy). The acids used for the digestion of samples in the microwave oven and for diluting the standard solutions fulfilled the proanalysis purity level. Nitric acid 65% and hydrochloric acid ≥37% were obtained from Carlo Erba Reagenti SpA (Rodano, Italy) and Sigma-Aldrich Chemie GmbH (Steinheim, Germany), respectively.

| Microwave digestion
Mars Xpress (CEM Corporation, Matthews, USA) microwave oven was used for microwave digestion of samples. Digestion was made according to the application note for fats and oils. Visual inspection of samples after digestion showed no signs of precipitation, turbidity, or residual carbon, and no precipitate was present on the filter after sample filtration. In the process of digestion, 1 g of the sample is put into a Teflon tube, to which 5 ml of 65% nitric acid and 1 ml of 36.5% hydrochloric acid are added. The microwave digestion was performed in four stages (first stage: p = 75 W, t = 100°C, t retention = 10 min; second stage: p = 150 W, t = 150°C, t retention = 10 min; third stage: p = 225 W, t = 190°C, t retention = 20 min; fourth stage: p = 300 W, t = 230°C, t retention = 30 min). The rate of temperature increase in all stages was 15 min. After cooling, the samples were quantitatively transferred into volumetric vessels of 50 ml and diluted with deionized water of <0.05 μS/cm conductivity.

| ICP-OES method
The metals in sample solutions were quantified by inductively coupled plasma-optical emission spectrometry (ICP-OES Optima 7000; Perkin Elmer, USA), in accordance with the official method (HRN EN ISO 11885, 2010

| Method validation
The validation of the method included the following parameters: linearity, limit of detection (LOD), limit of quantification (LOQ), and recovery. The limits of detection and quantification were determined by a tenfold measurement of a blind-trial solution at a particular wavelength characteristic for each metal. The standard deviation was calculated and the triple value of the obtained standard deviation (3 × SD) was taken as the limit of detection, while the ninefold value of the obtained standard deviation (9 × SD) was taken as the limit of quantification. For calcium, magnesium, sodium, and potassium, the values of the axial (ax) as well as the radial (rad) view are provided owing to detector saturation that occurs during axial determination with higher concentrations of those elements, which are common in food samples, so there is often a switch to the radial view. For all the other elements, the values are expressed only for the axial view (Table 1). The recovery of the method was determined by spiking (prior to digestion) a pumpkin seed oil sample of known base values with the standard solutions in three different concentrations representing the lowest, middle, and highest concentrations of the calibration range. Recovery is expressed as the middle ratio of the calculated (measured value reduced by base value) and the actual added amount of elements (Table 1).

| Statistical analysis
The concentrations of micro-and macroelements in the examined pumpkin seed oils were determined and expressed as the mean values, relative standard deviations, and the ranges of results ob-

| RE SULTS AND D ISCUSS I ON
Subgroups of Croatian pumpkin seed oil samples were formed based on the production process (virgin and cold pressed) ( Table 2) and production year (2011 and 2012) ( Notes. <LOQ: below the limit of quantification. *Significant influence of production process (p ≤ 0.05).
TA B L E 2 Concentrations of elements (mg/kg) in Croatian virgin and cold pressed pumpkin seed oils subgroups and Slovenia) ( Table 4). Among determined elements, the following groups were defined as follows: (a) alkali and alkaline earth metals; (b) transition metals; and (c) post-transition metals, metalloids, and lanthanum. Out of 32 measured elements, 14 elements (Ag, As, Be, Cd, La, Li, Mo, Ni, Pb, Sb, Si, Te, V, and Zn) were below quantification limits shown in Table 1 in all analyzed samples.

| Alkali and alkaline earth metals
Elements from this group, especially magnesium, potassium, calcium, and sodium, play important roles in many physiological functions of the human body. Since their intake levels have been connected to many different conditions and diseases such as hypertension, heart attack, and various gastrointestinal cancers, it is considered that homeostasis of these elements greatly contributes to the overall well-being (Blaine, Chonchol, & Levi, 2015;Glasdam, Glasdam, & Peters, 2016 TA B L E 3 Concentrations of elements (mg/kg) in Croatian pumpkin seed oils from 2011 and 2012 subgroups and processed cereals which are generally considered "healthy" (Magriplis et al., 2011).
According to the mean values obtained though this study ( Statistical analysis applied to the obtained results had shown that production process causes significant differences in the content of dominant alkali and alkaline earth metals. In cold pressed oils, the mean concentration of sodium was about 9 times lower (p ≤ 0.001), in their work showed no statistically significant differences between the seeds pertaining to different production year.

| Transition elements
In Croatian virgin pumpkin seed oils investigated in this study ( were below quantification in all samples. The mean values of manganese, copper, and iron were lower (p ≤ 0.001) in the cold pressed oils, while chrome was not detected. As it was mentioned, the differences in the concentrations of these elements can be partly explained by the residues from the mentioned salt addition (De la Guardia & Garrigues, 2015). However, heavy metals originating from the machines ware out are more likely the cause especially considering that iron, chromium, and manganese are the main constituents of stainless steel which is the main food processing equipment construction material (Kamerud, Hobbie, & Anderson, 2013). Namely, roasting of pumpkin seed dough during virgin oil production process takes place in the steel frying pans which are equipped with metal scrapers that prevent the dough sticking to and burning on the pan bottom. Due to the needed close contact of the pan and the scraper, abrasion, which can lead to higher iron, chromium, and manganese contents in the virgin seed oil, occurs.
In their study, Juranovic et al. (2003) obtained considerably higher values of elements from this group. The concentration of chrome (6.8 mg/kg) was hundred times higher than the mean for virgin pumpkin seed oils in this research, while the concentrations of copper and iron were about 10 times higher (12.1 mg/kg for copper; 16.1 mg/kg for iron; and 3.2 mg/kg for zinc). The average concentrations of cadmium and molybdenum were 1.7 mg/kg and 0.80 mg/ kg, respectively. Nickel, manganese, and cobalt were not detected. In that year, 31 occurrences of acid rain were detected by the Zagreb measurement station. In 2011, the situation was fairly different. In that year in the area of the Varaždin measurement station, 481.2 mm of precipitation was recorded, with the average annual temperature of 11.2°C and only eight occurrences of acid rain detected by the Zagreb measurement station (Ostroški, 2012). Possible impact of the aforementioned climatic conditions, that is, a smaller amount of precipitation in 2011 is evident in higher concentrations of copper, cobalt, and vanadium in comparison with their concentrations in 2012. Gfrerer and Zischka (2003) analyzed pumpkin seed oil samples from the Austrian province of Styria using the ICP mass spectrometer after microwave digestion and detected varying concentrations of cadmium, arsenic, lead, and mercury from this element group. As these elements are considered toxic heavy metal contaminants, their levels in food have to be closely monitored (Zhu, Fan, Wang, Qu, & Yao, 2011).

| Post-transition metals, metalloids, and lanthanum
The elements from this group found in the highest average amount in the virgin oils (Table 2) were selenium (2.783 mg/kg), aluminum (0.835 mg/kg), tin (0.768 mg/kg), and thallium (0.288 mg/ kg). While bismuth (0.027 mg/kg) was considerably low, cadmium, lead, zinc, boron, gallium, lanthanum, tellurium, antimony, silicon, and arsenic were not detected. Compared to cold pressed oils, concentration of selenium (p ≤ 0.05) was higher in virgin pumpkin seed oils, while tin (p ≤ 0.001) and bismuth (p ≤ 0.01) were lower. Also, thallium was present in virgin oils but it was not detected in any of the cold pressed samples. Juranovic et al. (2003) found that concentrations of both lead and aluminum were below the detection limits of their method (0.166 and 0.920 mg/kg, respectively). Similar findings regarding the concentration of lead being below the detection limit of 0.22 mg/kg were found by Cindric et al. (2007), who also reported the concentration of aluminum was 1.50 mg/kg which is somewhat higher than the levels detected in our study. Maximum amount of lead allowed in oils is set to 0.1 mg/kg by the Commission Regulation (EC) No 1881. As lead was not detected in any of the samples, the results of this study show satisfactory safety related to this parameter. Kreft, Stibilj, and Trkov (2002) analyzed the concentration of selenium in pumpkin seed oils and seeds and concluded that pumpkin seed oil could not be considered a source of selenium. The levels of selenium for oils in their research were below the method limit

| Comparison between macro-and microelements in virgin pumpkin seed oils according to country of origin
Results for the contents of analyzed elements in virgin pumpkin seed oils from Croatia, Slovenia, and Austria are shown in Table 4 as it was mentioned that the origin of potassium in virgin pumpkin seed oil is the salt added during its production, it can be assumed that Austrian and some Slovenian producers add higher amounts of salt during the preparation of pumpkin seed dough for roasting. Also, as the intensity of heat treatment causes better transfer of elements from seeds to oil, Austrian as well as Slovenian producers of virgin pumpkin seed oil could be using higher temperatures and/or longer roasting times in comparison with Croatian producers. Significant differences were also noted in the contents of barium and strontium from the alkali and alkaline earth metals group. In transition element group, only copper was significantly higher in Croatian oils, while in the post-transition metals, metalloids, and lanthanum group, significant differences were noted for the selenium, tin, and bismuth which were again higher in Slovenian and Austrian oils. Since the differences between countries considering the four elements present in the highest amounts in the samples are most notable for sodium and considering that kitchen salt is added in the process of oil production, it can be assumed that there are differences in the amount of salt used in the production process in each of the three countries.
Distribution of various trace elements in seed oils is dependent on their content in the soil. Some of these elements, such as rare earth elements, derive from the underlying rock and are therefore related to the geographical origin (Bandoniene, Zettl, Meisel, & Maneiko, 2013). On the other hand, environmental conditions, that is, air and soil pollution can significantly influence the content of elements, especially heavy metals in the edible oils (Angelova, Ivanova, & Ivanov, 2005 To further investigate the noted differences between the countries of origin, principal components analysis (PCA), which enables efficient reduction of data dimensionality, was used. PCA is often applied to examine the interrelationships between macro-and mi- In addition, Figure 2 shows the projection of samples of virgin pumpkin seeds oils in the plane of factors 1 and 2. Considering the aim of this research, it is clearly evident that all Austrian and most of the Slovenian oils are grouped in the first and second quadrant, that is, in the area in which the first factor was assigned positive values. In addition, if we compare Slovenian and Austrian oils, it is notable that all the Austrian oils are located further to the right in relation to Slovenian oils, which is a consequence of a higher concentration of elements that are strongly positively correlated with F1-magnesium, potassium, sodium, calcium, and tin.
Only three Croatian oils are on the right side of coordinate system while the rest have negative values for F1 and are dispersed across second and third quadrant showing. Such a distribution can be explained by differences in the contents of elements for which positive correlations F2 were established, that is, bismuth and selenium. Consequently, owing to the aforementioned reasons, in the two-factor plane grouping of samples with the same country of origin is evident and confirms specifics in the applied production processes.

| CON CLUS ION
According to the results of this investigation, it can be concluded that the content of macro-and microelements in pumpkin seed oils varies significantly under the influence of processing conditions, crop seasons, and countries of origin. Magnesium, potassium, calcium, and sodium were dominant elements in both oil types; however, their content was significantly higher in virgin pumpkin seed oils compared to cold pressed. These differences occur due to the addition of kitchen salt for pumpkin seed dough preparation and also because of the heat-induced damage to cells and release of cellular material during roasting. Even though the quantities of sodium found in pumpkin seed oil are relatively low and do not pose a direct health threat, with the growing evidence on the importance of detection of hidden salt sources, reported findings of this research surely call for attention. On the other hand, crop season characteristics, that is, the amount of precipitation, were accounted for the minor but significant differences in the content of copper and cobalt. In addition, comparison of samples in regard to the country of origin revealed that sodium, potassium, calcium, and magnesium contents were significantly higher in Slovenian and Austrian oils than in Croatian oils. These differences confirm the existence of applied processing specifics, mainly related to the amounts of salt addition.

This work was supported by the Faculty of Food Technology and
Biotechnology University of Zagreb and Bioinstitut Ltd. Čakovec.
We thank the organizers of the exhibition of pumpkin seed oils of north-western Croatia for providing the pumpkin seed oil samples.

E TH I C A L S TATEM ENT
The study did not involve any human or animal experimentation.

CO N FLI C T O F I NTE R E S T
The authors have declared no conflict of interest.