Levels of trace elements and rare earth elements in honey from Jordan

Honey is a common sweetener in the Jordanian diet with an annual consumption of approximately one thousand tons, two-thirds of which are imported. It is believed that the elemental profile of honey is an indicator of safety and botanical and geographic origin. In the literature, there are a lack of studies concerning the levels of major and trace elements in honey in Jordan. A total of 46 elements, including 15 rare earth elements (REEs), were analyzed by inductively coupled plasma-mass spectrometry (ICP-MS) in 18 monofloral and multifloral imported honey samples and 12 multifloral local samples. Regarding monofloral samples, Black Forest samples had the highest total metal content, while acacia samples had the lowest total metal content. Local multifloral honey had the largest Sr and total REE levels, while it had the lowest Mn levels. Very low levels of toxic elements were found in all samples, indicating the safety of honey in Jordan for human consumption. The results of this study showed that a large number of samples (> 100) and the application of advanced statistical models are required to discriminate between multifloral imported and local honey.


Introduction
Honey is a valuable natural product that is marketed with a wide range of nutritional and medicinal properties. Honey is produced by bees (Apis mellifera) from the nectar of flowers Jordan, especially plant biodiversity. Jordan's plant diversity is remarkably high considering its size and aridity. The total number of recorded vascular plant species in Jordan is now over 2,622 species belonging to 113 families and approximately 810 genera (Al-Eisawi 1982). Flowering plants are the dominant group of species, which makes Jordan an excellent place for beekeeping activities. Generally, most Jordanian beekeepers migrate their colonies to the Jordan valley between October and April to avoid cold effects during the winter season and to benefit from early flowering of cultivated and wild plants in the Jordan valley, harvesting their multifloral honey by the end of April. Afterwards, the beekeepers transport their hives to the uplands for further honey collection from wild flowering plants growing there (Zaitoun et al. 2000).
Honey contains a variety of macro-and microelements in the range of 0.02-1.03%. Minerals and trace elements in honey seem to be good indicators of environmental pollution or the botanical and geographic origin of the honey (Solayman et al. 2016). Potassium (K) is the most abundant major metal in honey. Other major metals include sodium (Na), calcium (Ca), magnesium (Mg), and phosphorous (P) (Hernández et al. 2005). The rest of the elements are considered trace elements. Some of these elements are essential due to their involvement in metabolic processes and have several physiological and biochemical functions for correct cellular metabolism. These elements include cobalt (Co), copper (Cu), chromium (Cr), iron (Fe), manganese (Mn), molybdenum (Mo), nickel (Ni), selenium (Se), and zinc (Zn) (Squadrone et al. 2020a). The rest of the trace elements are considered nonessential. However, some trace elements exert toxicity even at low concentrations and are considered potentially toxic, such as mercury (Hg), cadmium (Cd), arsenic (As), and lead (Pb). Rare earth elements (REEs), an emerging group of nonessential elements, were detected in honey and have been used as geological tracers (Squadrone et al. 2020a).
Multielemental analysis has been recommended as a suitable tool for evaluating the mineral composition of honey. Inductively coupled plasma-atomic emission spectroscopy (ICP-AES) and inductively coupled plasma-mass spectrometry (ICP-MS) are the most popular techniques due to their compatibility with most elements in the periodic table, large sensitivity, and large linear dynamic range (Devillers et al. 2002;Kiliç Altun et al. 2017;Squadrone et al. 2020a;Wetwitayaklung et al. 2018).
Multivariate data analytical methods usually accompany the elemental profile for predicting mainly the botanical origin and, to a lesser extent, the geographic origin of honey. Principal component analysis (PCA), discriminant analysis (DA), and cluster analysis are the preferred methods for performing predictive analyses. PCA and DA are the most popular due to their ease of application and interpretation (Lazarević et al. 2012;Conti et al. 2018;Berriel et al. 2019). However, a successful prediction necessitates that honey samples be obtained from producers or purchased from reliable stores. Additionally, the origin of samples must be ascertained prior to the application of data analytical methods. This process is accomplished through poollen analysis, which is known as a slow process. Also, a compromise between the number of samples and number of variables must be made (Maione et al. 2019).
The objective of this study was to determine the levels of 46 elements in imported honey (monofloral and multifloral) and local honey (multifloral) from Jordan. The elements were categorized into major elements (Na, Mg, P, K, and Ca), essential trace elements (Fe, Zn, Mn, Cu, Ni, Cr, Se, Mo, and Co), nonessential trace elements (Li, Be, Al, V, Ga, Sr, Ag, Sn, Cs, Ba, Bi, and U), potentially toxic trace elements (As, Cd, In, Tl, and Pb), and rare earth elements (Y, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu). Levels of elements were evaluated for nutritional and safety assessment and the possibility of characterizing the botanical and geographic origin of the honey.

Reagents
Chemicals and standards used were analytical grade. HNO 3 (69% w/w, extra pure) and H 2 O 2 (35% w/w, extra pure) were purchased from Sigma (St. Louis, MO, USA). Ultrapure deionized water (Milli-Q water) was employed to prepare standard solutions and sample solutions. A stock solution of combined elements (10.0 mg/L each) was purchased from Merck (Darmstadt, Germany). All plastic containers, pipette tips, polypropylene flasks, Pyrex glass digestion tubes, and reagents that came into contact with samples or standards were checked for contamination.

Sampling and sample preparation
Local honey samples (n=12) were obtained directly from commercial honey producers, and imported honey samples (n=18) were purchased from local markets. The thirty sample brands are available to Jordanian customers from either producers or public stores. However, other local honey brands are produced but not available to public consumers. These brands were not included in this study. All samples were labeled with their floral type. Samples were stored in brown bottles at room temperature. All 12 local samples were multifloral from harvest year 2019. Details of imported samples, including imported country and floral type, are presented in Table 1S.
A 1.00 g honey sample in triplicate was mixed with 6.0 mL HNO 3 and 2.0 mL H 2 O 2 as a catalyst and digested using ETHOS 1-advanced Microwave Digestion Lab Station (Milestone S.r.l., Italy). The thermal program included a gradual rise in temperature for 12 min to reach 200°C, a constant temperature at 200°C for 20 min, and finally a gradual decline for 12 min to reach room temperature. After digestion was completed, as indicated by the appearance of a colorless, completely clear, and homogeneous solution, the clear mixture was left to cool down, and the contents of the tubes were transferred to digestion tubes, evaporated to near dryness at 110°C, and then diluted to 20 mL with 1% HNO 3 solution.
The obtained solutions were measured for the targeted elements using ICP-MS (Agilent 7500a Series ICP-MS). The ICP-MS configuration and operating parameters, including selected isotopes, are presented in Table 2S.

Calibration and quality control
External calibration was employed by sequential dilution of the multielemental calibration standard to prepare six calibration standards ranging from 0.10 to 20 μg/L for REEs, 0.10 to 100 μg/L for trace and potentially toxic elements, and from 10 to 1000 μg/L for major elements. Absolute intensities, counts per second (Cps), versus concentration were employed in the statistical analysis. The linear least squares method was applied for the calculation of statistical parameters for calibration curves. Elemental concentrations were calculated from the respective calibration curves.
Calibration curves were linear over the range of the three calibration ranges with correlation coefficients better than 0.999.
The limit of detection (LOD), defined as 3S b /a, and the limit of quantitation (LOQ), defined as 10S b /a, where S b is the standard deviation of the intercept and "a" is the slope of the calibration curve, were calculated for each element. The method limit of detection (MLOD) and method limit of quantitation (MLOQ) were calculated by multiplying the LOD and LOQ by the dilution factor. The precision (% RSD), defined as (S a /a) x 100, where S a is the standard deviation of the slope of the calibration curve, was also calculated.
Quality control studies were performed on a composite sample prepared from the digest of the 30 honey samples. The composite sample was certified for targeted elements by the standard addition method (Tahboub et al. 2021) and was employed for further quality control measurements.
Accuracy and recovery results were above 94%, and precision results (% RSD) were less than 6% for all determined elements. The accuracy, precision, and recovery results on certified samples and postspiked certified samples are presented in Table 4S.

Statistical analysis
JMP PRO 14 software (jmp-statistical discovery from SAS, NC, USA) was employed for the calculation and evaluation of statistical parameters. Q-Q plots were employed to assess the normality of the data; if the variable could be assumed to be normally distributed, a t-test for unequal variances, which was insensitive to heterogeneous variables, was employed to discriminate between local and imported honey samples. Then, the nonparametric Wilcoxon test was used. Q-Q plots indicated that the individual levels of the elements Mn, Ni, Ba, La, Ce, Pr, Nd, Sm, and Gd, and ∑REEs were not normally distributed. Detailed Q-Q plots are presented in Table 5S.

Results and discussion
The analytical results for the basic 31 elements and 15 rare earth elements (REEs) in targeted honey samples (n=30) are displayed in Tables 1 and 2. Levels for each targeted element (mg/kg or μg/kg) include the mean ± SD and range (minimum-maximum). Additionally, the elements were categorized as local or imported, and the level of each element in its category is reported in Tables 1 and 2 as the mean ± SD. The basic elements in Table 1 were classified into major elements (n=5) (Na, Mg, P, K, and Ca), essential trace elements (n=9) (Cr, Mn, Fe, Co, Ni, Cu, Zn, Se, and Mo), nonessential elements (n=12) (Li, Be, Al, V, Ga, Sr, Ag, Sn, Cs, Ba, Bi, and U), and potentially toxic elements (n=5) (As, Cd, In, Tl, and Pb). The rare earth elements (REEs), n=15 (Y, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu), are presented in Table 2. The classifications of selected basic elements and selected REESs based on floral type are presented in Table 3. Selected elements were those having significant concentrations.

Major elements
The major elements Na, Mg, P, K, and Ca represented more than 90% of the total elemental content in most foods, including honey, and their levels are related to their presence in soil, fertilizer, and irrigation water. A wide range, max/min > 10, was observed for all major elements. Their decreasing order in honey was K>> Ca> P> Na> Mg (Table 1). The same order held for imported honey samples; however, in local samples, Na was more abundant than P. Additionally, total major elements (∑major) were significantly more abundant in imported honey samples (Table 1). The dependence of major elemental content on botanical origin was demonstrated between the monofloral Black Forest (∑major = 2131 mg/kg) and monofloral acacia (∑major= 420 mg/kg) ( Table 3). The levels of major elements in local and imported multifloral honey samples were similar (Table 3).

Essential trace elements
Levels of essential trace elements in honey samples varied between 19.4 mg/kg for Fe and 0.011 mg/kg for Co. A wide range, max/min > 10, was observed for all essential trace elements. Their decreasing order in honey was Fe>> Mn> Zn> Cu> Cr> Ni> Se> Mo> Co (Table 1).
Iron, Fe, is an element required for essential functions in cells related to oxygen transport, oxidase activities, and energy metabolism (Squadrone et al. 2020c). Levels of Fe varied between 7.03 mg/kg in an imported multifloral sample and 49.3 mg/kg in a local multifloral sample.
Manganese, Mn, is a metalloenzyme involved in various metabolic processes (Squadrone et al. 2020c). Mn levels in honey varied between 0.13 mg/kg in an imported monofloral sample (acacia) and 13.3 mg/kg in an imported Black Forest monofloral sample.
Zinc, Zn, is utilized in various physiological functions and is ubiquitous in the body; three general functional classes (catalytic, structural, and regulatory) define the role of Zn in biological processes. Zn is important for regulation of gene expression, stabilizing DNA, and has several specific functions in zinc enzymes (Jurowski et al. 2014). Zn levels varied between 0.48 mg/kg in an imported acacia monofloral sample and 3.72 mg/kg in an imported multifloral sample.
Copper, Cu, is classified as an essential element due to the wide range of enzymes that use this element as a cofactor. However, the symptoms of Cu deficiency are diverse. On the other hand, an excess of Cu may cause genetic disorders such as Wilson disease (Jurowski et al. 2019a). Cu levels varied between 0.145 mg/kg in a local multifloral sample and 2.34 mg/kg in another local multifloral sample. Chromium, Cr, is a very problematic element due to the number of its oxidation states, especially Cr(III) and Cr(VI). Cr(III) is the most abundant form in the environment and plays a role in glucose metabolism. Cr(III) deficiency may be associated with cardiovascular diseases and numerous system disorders (Jurowski et al. 2019b). Levels of Cr varied between 0.06 mg/kg in a local multifloral sample and 0.31 mg/kg in another local multifloral sample.
Other essential elements, Ni, Se, Mo, and Co, were detected at lower levels and had less impact on nutrition or safety assessments. No significant differences were observed between local and imported samples for all targeted essential trace elements except for Mn, where its mean in imported samples was 5 times larger than local samples.

Nonessential trace elements
Among the 12 targeted nonessential elements, Al, Sr, Ba, and Sn had significant levels and were found in decreasing order: Al>> Sr> Ba> Sn (Table 1).
Aluminum, Al, a common metal in the earth's crust, was detectable in all samples, ranging from 2.67 mg/kg in an imported acacia monofloral sample to 21.9 mg/kg in an imported multifloral sample. Mean levels of Al in local and imported samples were similar.
Strontium, Sr, was measured at significant levels in all samples, ranging from 0.075 mg/kg in an imported acacia monofloral sample to 1.62 mg/kg in an imported multifloral sample. Mean Sr levels between local and imported samples were significantly different. Sr could be employed as an elemental marker to distinguish between local and imported honey.
For barium, Ba, levels varied between 0.059 mg/kg in an imported multifloral sample and 1.06 mg/kg in a Black Forest monofloral sample. There was no significant difference in mean Ba levels between local and imported samples.
Tin and Sn were found at low levels in honey, with the lowest concentration of 0.033 mg/kg in an imported Black Forest monofloral honey sample and the highest concentration of 0.186 mg/kg in a local multifloral sample. Sn levels in local samples were slightly higher than those in imported samples. Potentially toxic trace elements Thallium, Tl, and indium, In, had lower levels and were not considered for toxicity and assessment. Arsenic, As, is a highly toxic element. Its levels were not detected (ND) in 50% of the samples and were up to 6.9 μg/kg in an imported multifloral sample. Cadmium, Cd, and lead, Pb, are common pollutants that can enter the food chain after contamination of flowering plants. Cd levels were low in most samples and reached 6.7 μg/kg in an imported Black Forest monofloral sample. Pb levels were low in most samples and reached 31.5 μg/kg in the same imported Black Forest monofloral sample. Levels of Cd and Pb as well as As were much lower than regulated levels in food (Paz 2017).

Rare earth elements
REEs are concentrated in different geological environments, and their presence in honey is related to geographic origin and geochemical soil composition. Significant levels were detected for La, Ce, Pr, Nd, and Sm in decreasing order: Ce> Nd> La> Pr> Sm (Table 2). Total REE levels varied between 41 μg/kg in an imported acacia monofloral sample and 583 μg/kg in a local multifloral sample with a mean of 145 μg/kg and standard deviation (SD) of 157 μg/kg. Local samples had significantly larger means than imported samples. Significantly larger deviations, mainly in imported samples, were due to the difference in geographic origins between imported countries.

Safety assessment
Honey is a popular sweetener in Jordanian diets. An adult usually consumes 10-50 g daily. Thus, honey may contribute significantly to the total daily intake of elements. Table 4 presents the levels of estimated daily intake (EDI), recommended daily intake (RDI) (Paz 2017), and percentage contribution of each metal. The percentage contributions of Na, Mg, P, K, Ca, Mn, and Zn were negligible relative to their respective EDI values (<1.0%). Little percentage contributions were observed for the rest of the essential elements. Levels of potentially toxic elements (As, Cd, and Pb) were negligible to their RDI values, indicating the safety of honey in Jordanian markets, either local or imported, for human consumption.

Assessment of botanical and geographic origins
Botanical and geographic origins are usually considered in the assessment and authentication of honey (Gallmann 2007;Salvador et al. 2019;Squadrone et al. 2020b;Squadrone et al. 2020b;Zhou et al. 2018). A single factor ANOVA test is usually employed for discrimination between means of variables (elements) in two populations (local and imported). A p> 0.05 indicates no significant difference between variables. However, ANOVA is sensitive to the normality and heterogeneity of the data. In our study, Q-Q plots were employed to assess the normality of the levels of each element in the thirty samples. Q-Q plots revealed that the levels of essential trace elements Ni and Mn; nonessential elements Ba; REEs La, Ce, Pr, Nd, Sm, and Gd; and ∑REEs were not normally distributed (Table 5S). Additionally, the variances of elements in local   (Tables 1 and 2). Thus, ANOVA could not be applied and was replaced by alternative t-tests (unequal variances) for normally distributed elements and the nonparametric Wilcoxon test for nonnormally distributed elements. Both tests are insensitive to heterogeneous variables. As in ANOVA, p> 0.05 indicates no significant difference (NS). The pvalues in Table 1 indicate no significant differences between local and imported honey samples for all basic elements except the nonessential elements Al and Sr and the potentially toxic element Pb. Environmental factors may have contributed to the differences in the levels of Pb. The p-values in Table 2 indicate significant differences between local and imported honey samples for all targeted REEs except Nd.
A closer look at imported samples consisted of Black Forest monofloral samples (n=4), Acacia monofloral samples (n=4), and multifloral samples (n=10). All local samples were multifloral (n=12). Thus, botanical origin was considered in the assessment and authentication. The levels of selected elements (mean ± SD) are presented in Table 3.
The precision of results of monofloral samples (Table 3) was much smaller than those in Tables 1 and 2. ∑major elements in Black Forest samples were much larger than those in acacia samples. While it was easier to discriminate between monofloral and multifloral honey, more factors are needed to discriminate between multifloral local and multifloral imported samples. Aluminum, Al, strontium, Sr, and ∑REEs were significantly different between multifloral local and imported honey samples. These elements may be considered chemical markers for tentative discrimination between local and imported multifloral samples.
Pearson's correlation coefficients between targeted elements in honey samples were computed and are presented in Table 5.
In this study, we combined monofloral Black Forest and acacia samples in one category under the title "monofloral." The other two categories were multifloral imported and multifloral local. Rare earth elements in the three categories were positively correlated with each other (1.00 > r > 0.80) and negatively correlated with other elements; thus, they were not considered in this study.
Selected correlations were based on categories with positive correlation coefficients (r> 0.50). Among the 60 selected correlations, the largest positive correlation coefficients were distributed as 22 for multifloral local samples, 11 for multifloral imported samples, and 27 for monofloral samples. The correlations Na/Mg, Na/Se, Mg/Se, Cr/Fe, Cr/Co, Cr/Ni, Fe/Ni, and Al/Sr were distinctive for multifloral local samples. The correlations K/Fe, K/Cu, ∑major/Cu, Fe/Cu, Co/Mo, and Se/Sn were distinctive for multifloral imported samples.

Comparison with previous studies
A large number of studies (above 200) have been reported for multielemental measurements of basic elements (major and trace) in honey from popular honey-producing countries. A summary of these results was reported in a review by Solayman et al. (2016).   Data availability All data generated or analyzed during this study are included in the published article and its supplementary information files.

Declarations
Ethics approval Not applicable Consent to participate Not applicable