The Application of SDS-PAGE Protein and HPTLC Amino Acid Profiling for Verification of Declared Variety and Geographical Origin of Honey

Proteins and amino acids are minor components of honey that are rarely used for its quality evaluation, although these components create its biological activity as well as can serve for overheating detection. The aim of the study was to use these indicators to confirm a declared on the label variety. Fifty-eight honey samples of 8 different varieties meeting the commercial quality requirements were used, including 28 local Polish and 30 commercial (mainly imported from EU and non-EU) honeys. For honey protein profiling previously used, polyacrylamide electrophoresis in denaturing conditions (SDS-PAGE) was applied whereas the free amino acid profile was analyzed by high-performance thin-layer chromatography (HPTLC) for the first time. As auxiliary indicators, the colorimetric determination of the protein content by the Bradford method, the activity of 5 glycolytic enzymes, including diastase, β-galactosidase, N-acetyl-β-glucosaminidase, α-mannosidase, and α-glucosidase, and the content of proline were used. It has been shown that based on the determined model protein SDS-PAGE profiles for selected monofloral honeys, it is possible to detect honeys of questionable variety based on the lack of specific protein bands or their diverging intensity. The HPTLC amino acid analysis can serve as a supporting control tool, capturing differences in the amino acid profile. Due to a great variation of multifloral honey, such assays are effective for monofloral honey only. The colorimetric assays, especially for total protein and β-galactosidase, can be also useful. The applied tools can be proposed for the initial verification of honey variety for cost reduction of officially recognized melissopalynological analysis.


Introduction
Honey is a popular food product, readily consumed for its taste, nutritional, and health benefits (Kumar and Bhowmik 2010;Rana et al. 2018). Good-quality honey can reach a high price on the market, but nevertheless, it is accepted by conscious consumers. Among the factors influencing the quality and properties of honey, three main groups can be distinguished: related to bees (species, health status, nutrition), geographical and climatic factors (including environmental pollution and floral origin), and factors related to the honey itself and its production (beehive technology, harvesting, maturation, decrystallization, handling, packaging, and storage) (Bett 2017;Mohammed 2020).
Depending on the main source used by bees, honey is categorized as nectar (mono-and multifloral) or honeydew, made from the secretions of trees and insects, mainly from the Aphidae family (Pita-Calvo and Vasquez 2017). Based on the botanical origin of the nectar, individual varieties of monofloral honeys are classified. While distinguishing nectar honey from honeydew honey is not difficult with the use of simple physicochemical analysis, the determination of the correct variety of nectar honey is more complicated and the only officially recognized method for determining the variety is the laborious and time-consuming melissopalynological analysis (von der Ohe et al. 2004;Thrasyvoulou et al. 2018). Alternative cost-and time-saving tools to verify the declared honey variety are still in demand. Multiple attempts are made to use chemical markers, phytochemicals mainly from the group of polyphenolic compounds (Gašić et al. 2017;Stanek and Jasicka-Misiak 2018;Beceril-Sanchez et al. 2021;Puścion-Jakubik et al. 2022), characteristic for a given variety or crop plants, and volatile compounds (Yang et al. 2018;Machado et al. 2020;Tananaki et al. 2022). Due to their dual origin (bee and plant), nitrogen compounds, mainly proteins, could be interesting but rarely studied group of honey components that can be used in quality assessment and variety verification (Baroni et al. 2002;Miłek et al. 2021;Muresan et al. 2022). Although there is little protein in honey (on average, from 0.3 g/100 g for nectar honeys to 0.6 mg/100 g for honeydew honey) (Bogdanov 2011), they are important ingredients, largely responsible for honey bioactivity. Glucose oxidase and active peptides, e.g., defensin-1, are recognized factors shaping antimicrobial activity (Kwakman and Zaat 2012;Bucekova et al. 2014). The main group of proteins is enzymes, such as diastase, invertase, glucose oxidase, acid phosphatase, catalase, and β-glucosidase, which due to their thermolability may be a good and sensitive indicator of honey quality (Machado de Melo et al. 2018;Miłek et al. 2021). Additionally, honey also contains free amino acids, among which proline usually predominates (Hermosin et al. 2003;Kowalski et al. 2017). Honey amino acid profile could be a potential indicator of honey variety (Balkanska et al. 2020;Pacholczyk-Sienicka et al. 2022).
Another aspect important for the classification of honey is its geographical origin. It has been shown that honeys purchased from local producers usually meet the requirements of the standards and show better health-promoting properties (Dżugan et al. 2018;Miłek et al. 2021). Often, commercially available honeys are blends from many countries, and in addition, inappropriate conditions are used in their production, i.e., too high temperature for liquefaction, which affects their parameters, including the content and activity of proteins.
One of the tools helpful in analyzing the protein profile of honey is polyacrylamide gel electrophoresis (PAGE). This technique was successfully used earlier to analyze the composition of honey protein fractions, to detect active enzymes (native-PAGE zymography), and the first attempts to use it to evaluate typical varietal patterns of protein bands were also made (Bilikova et al. 2015;Bocian et al. 2019;Borutinskaite et al. 2018;Miłek et al. 2021, Muresan et al. 2022. A similar assumption can be made for using the high-performance thin-layer chromatography (HPTLC) technique to create typical profiles of free amino acids in honey; however, to our knowledge, this technique has not yet been applied.
The aim of the study was to check the usefulness of honey protein and amino acid profiling to verify the variety declared on the label of both local and imported honeys. The study covered honey varieties typical for beekeeping in Central and Eastern Europe. Two not very common methods, SDS-PAGE and HPTLC, were tested as potential analytical tools in comparison to colorimetric standard measures of protein, diastase, and proline contents as well as selected glycolytic enzyme activity.

Reagents and Chemicals
All used reagents were of analytical grade. The amino acid standards, p-nitrophenyl substrates for enzymatic activity assays, Nonidet P-40 Substitute, dithiotreitol, Coomassie Brilliant Blue G-250, and 2-methoxyethanol were purchased form Sigma-Aldrich (Saint Louis, MO, USA). Ninhydrin, acetic acid, isopropanol, n-propanol, and hydrochloric acid were purchased from Chempur (Gliwice, Poland). Deionized water was used to prepare the solutions.

Honey Samples
At the beginning, 80 honey samples (35 local, L; and 45 commercial, C) were initially assessed based on obligatory regulations (Journal of Laws 2015) in terms of honey quality and only 58 samples (28 L and 30 C) that fulfilled the law limits regarding water content, free acidity, conductivity, and HMF content (data not shown) were used in further study. Local honeys (L) were purchased directly from beekeepers in South-Eastern Poland (Podkarpackie region) whereas commercial honeys (C), mainly imported from EU and non-EU countries (according to labeling), were bought in local supermarkets (Table S1). The following varietal honeys were analyzed: rape-Brassica napus (4 L and 4 C); acacia-Robinia pseudoacacia (4 L and 4 C); linden-Tilia sp. (4 L and 4 C); buckwheat-Fagopyrum esculentum (4 L and 4 C); deciduous honeydew (3 L and 4 C); coniferous honeydew-mainly Abies sp. (3 L and 4 C); heather-Calluna vulgaris (2 L and 2 C); and multifloral (4 L and 4 C). The variety was declared by beekeeper for local and by producer/distributor for commercial honey. Until the analysis, honeys were stored in a dark place at a temperature of 20-22 °C.

Protein Content
In honey extracts (20% w/v), soluble protein fraction was determined by the Bradford method according to Latimer 1 3 (2016). Accordingly, 1000 µL of Bradford reagent (G-250) was added to 100 μL of honey extract. Samples were incubated for 5 min at 20-22 °C and the absorbance (y) was read against blank sample at 595 nm using a spectrophotometer (Biomate 3, Thermo Scientific, Waltham, MA, USA). The results were calculated on the basis of a calibration curve 6.25-200 µg/per 20 µl of sample (x); (y = 31.6752x, r 2 = 0.9919). Bovine albumin was used as a standard protein.

Proline Content
The proline content was determined in accordance with the instructions contained in the Polish regulation, relating to the methods of honey analysis (Journal of Laws 2015). Honey solutions at a concentration of 2.5 g in 50 mL of distilled water were prepared. Five hundred microliters of each honey solution, 0.25 mL of concentrated formic acid, and 1 mL of a 3% solution of ninhydrin in 2-methoxyethanol were added to the test tube. The closed samples were heated at 100 °C for 15 min. After this time, the samples were cooled until ambient temperature is reached, and 5 mL of 50% isopropanol was added. The absorbance at 520 nm was measured (y) against the blank prepared for each honey (0.5 mL honey solution, 1.25 mL distilled water, 5 mL isopropanol). The results were calculated on the basis of calibration curve prepared for proline in the concentration range 0-20 µg/per 500 µl of sample (x); (y = 0.0259x, r 2 = 0.9655).

Enzyme Activity
Diastase activity in honey was determined using Phadebas Honey Diastase Test tablets (Magle AB, Sweden) by the spectrophotometric method by following the instructions provided by the manufacturer. The absorbance at 620 nm was measured against the control. The results were converted according to the formula included in the instruction and expressed as diastase number (DN).
The activities of four glycosidases, N-acetyl-βglucosaminidase (NAG), α-mannosidase (α-MAN), β-galactosidase (β-GAL), and α-glucosidase (α-GLU), using the method described by Tomczyk et al. (2019) were tested for honey samples. The reaction mixture contained 25 μL of 2 mM appropriate p-nitrophenyl substrate in 0.2 M citrate buffer (pH 5.5 for α-GLU and 4.0 for other enzymes) and 25 μL of honey aqueous solution (20% w/w) sample. In blank, 25 μL of suitable buffer was used instead of substrate solution. Samples were incubated at 37 °C for 60 min, and the reaction was then stopped by the addition of 250 μL of 0.5 M carbonate buffer at pH 10.5. Absorbance was measured on a microplate reader (EPOCH2, Biotek) at a wavelength of 400 nm. Results were expressed as enzymatic units mU (nmol/min/g) by multiplying the measured absorbance by a factor of 79.4.

Protein Profiling by SDS-PAGE
Samples of each honey (0.5 g) were prepared for denaturing electrophoresis as described previously (Tomczyk et al. 2022) by adding 0.5 mL of deionized water containing 2% Nonidet P-40 Substitute, and 2% dithiothreitol and mixed thoroughly on a rotator. Liquid honey samples were combined with 4 × concentrated Laemmli buffer (20 µL sample + 10 µL buffer). Samples were incubated for 5 min at 95 °C and then cooled before being applied to 15% denaturing gels (with 3% stacking gels). The samples (15 µL) were loaded onto 8 different gels, each of which contained samples of one species of honey. Electrophoresis was carried out on a Mini-Protean Tetra Cell apparatus (Bio-Rad Laboratories, Hercules, CA, USA) at 50 V for 15 min and then at 200 V for 1 h according to the standard method of Laemmli, with the BlueEasy Prestained Protein Marker (NIPPON Genetics EUROPE, Düren, Germany) as a molecular weight marker and Tris-Glycine-SDS buffer. After electrophoresis, gels were stained overnight with Coomassie Brilliant Blue G-250 and destained for 24 h with deionized water. Gels were scanned with Image Scanner III (GE Healthcare, Little Chalfont, UK) and processed by EPSON Scan software. Gels analysis was performed in ImageJ (1.52a) software to generate a graphical representation of each lane on the gel to assist in sample comparison. For better visualization, individual bands from the SDS-PAGE gel were contrasted with the profile obtained with the software.

HPTLC Free Amino Acid Profiling
Free amino acid profiles of all honey samples were obtained using CAMAG HPTLC system (automatic applicator Linomat 5, an automatic developing chamber ADC2, a derivatizer, and a visualization apparatus TLC Visualizer 2, CAMAG, Switzerland). Chromatographic separation was performed on silica gel 60 F254 HPTLC glass plates (20 × 10 cm), purchased from Merck (Darmstadt, Germany). The mobile phase was a mixture of n-propanol and water (7:3, v/v), according to Brunauer et al. (2014). Honey solutions (20% in water) were applied to the plate in a volume of 4 µL as bands 6 mm wide, 8 mm from the bottom edge of the plate. Standards of 21 protein amino acids were used and applied in a volume of 2 µL (1 mg/mL in distilled water). The plates were developed with a ninhydrin solution (0.2 g in 100 ml of 96% ethanol, acidified with 3 mL of acetic acid) and after spraying were heated at 105 °C for 3 min. The plates were photographed under visible light and analyzed with the Vision CATS software (CAMAG). Based on a comparison of the colors of the bands and their retention factor (Rf) values, specific amino acids for each honey varieties were identified.

Statistical Analysis
All analyses of quantitative parameters were performed in triplicates; data are shown as mean ± standard deviation. The correlations among obtained parameters were analyzed by Pearson coefficient (r). The significant differences were calculated by a one-way analysis of variance, followed by Tukey's honest significant difference test (p < 0.05) or t-test in the case of comparison of local and imported honey samples. Principal components analysis (PCA) was applied to find the relation between tested parameters. All calculations were made using the Statistica 13.1 software (StatSoft, Tulsa, OK, USA).

Protein and Proline Content
The analysis of protein content (Table 1) using the Bradford method showed that dark honeys, especially heather honeys, contain the highest amount of protein, which is in line with previous studies (Burzyńska et al. 2020;Miłek et al. 2021). However, the protein content of heather honeys varied depending on the geographical origin of the honey. Significantly lower values (by 65.48%) were found for imported heather honeys which were also more diversified. Lower values were obtained for buckwheat and honeydew honeys (from 90.19 to 177 mg/100 g depending on the variety). The lowest values among the analyzed samples were found for light honeys (rape, acacia, and linden). In this case, the average protein content ranged between 34.36 and 66.69 mg/100 g depending on the variety. When comparing local with imported honeys, it should be noted that local honeys are richer in protein, especially dark honeys (buckwheat and deciduous honeydew). For linden and coniferous honeydew honey, slightly higher protein content in commercial honeys was observed (from 2.29 to 7.78%); p > 0.05. Only commercial rape honeys had a much higher protein content than local honeys (30.02%); p < 0.05. A similar relationship was found when analyzing the proline content in varietal honeys (Fig. 1). It was shown that light honeys, mainly rape and acacia honeys, showed the lowest values of proline content (on average 33.79 mg/100 g). Much higher values were found for dark honeys (57.35-122.36 mg/100 g), and the highest values were obtained for heather honeys (110.51-140.12 mg/100 g). Truzzi et al. (2014) analyzed the proline content in Italian varietal honeys. The authors showed that the honeydew honeys contained 66.6 mg of proline in 100 g of honey, which confirms the results presented in this study. In turn, the proline contents of 25.3 mg/100 g and 39.7 mg/100 g were found for acacia and linden honey, respectively. The extended analysis was performed by Czipa et al. (2012) who showed that Hungarian acacia honeys contain from 20.7 to 44.7 mg of proline per 100 g of honey, which is consistent with the results shown in Fig. 1. Moreover, for acacia honeys originating from outside the EU, they showed on average a higher proline content. The linden honeys from the European Union countries analyzed by the authors contained on average 14% less proline compared to Polish honeys. The authors also studied less known honeys such as coriander honey, in which they found even 228.83 mg of proline in 100 g of the product. Because the proline content is considered an indicator of honey maturity and its possible adulteration, it is assumed that honey should contain at least 18 mg/100 g of this amino acid (Bogdanov et al. 1999). Thus, the tested samples meet this assumption well.

Enzymatic Activity
Diastase activity is the only enzymatic parameter which is limited by law regulations. The comparison of enzymatic Fig. 1 Proline content (on average) in varietal honeys: R, rape; L, linden; A, acacia; B, buckwheat; HR, heather; C, coniferous honeydew; D, deciduous honeydew. a,b,c Means sharing the same superscript capital letters do not differ significantly  Table 1.
The analyzed honeys showed diversified diastase activity depending on the variety, which was expressed as the variation coefficient (10-30%). Again, the highest values of up to 44.21 DN were found for dark honeys (honeydew from coniferous and deciduous honeydew and heather), while the lowest values (10.73 DN) were found for light honeys (acacia, rape, linden). Moreover, it has been shown that honeys from local apiaries show higher values (up to 57.66% for honeydew honeys) compared to honeys from trade. The results presented in Table 1 are similar to those presented by Wesołowska and Dżugan (2017), who tested Polish varietal honeys. The authors showed that among the tested samples, honeydew honeys showed the highest diastase activity (33.08 DN), and rape honeys were the lowest (15.35 DN). However, Nayik and Nanda (2015) showed a higher diastase activity in acacia honeys (15.51 DN) than in multifloral honeys (14.93 DN).
Changes in the activity of other glycolytic enzymes in honey depending on the variety and origin were analyzed (Table 1). It was found that heather and honeydew honeys had a much higher N-acetyl-β-glucosidase (NAG) activity (up to 43.70 mU) compared to other types of honey (from 4.29 to 10.46 mU). Heather and honeydew honeys also showed an increased content of α-mannosidase and β-galactosidase. The higher β-galactosidase activity was also observed in rape honey (32.15 mU), which is much greater compared to other glycosidases. Thus, this enzyme could be a potential marker in confirming the authenticity of rape, heather, and honeydew honeys.
Unlike diastase which is included in the requirements for the commercial quality of honey, other enzymes belonging to glycosidases group are not widely studied (Bogdanov 2016;Dżugan et al. 2021;. In other study, this enzyme was proposed as an indicator of honey thermal processing Sidor et al. 2021). The studies showed that in all of the tested honeys, the glycosidase activity decreases in sequence: β-GAL > α-GLU > α-MAN and the highest activities for linden than multifloral and acacia honeys were found . The results cited are in line with the present study. Moreover, reduced activity of the tested enzymes was found in fresh liquid honeys subjected to fixation and liquefied solid honey regardless of the technique used. The significant decrease found for glucosidase activity, especially β-galactosidase, as a result of honey processing by heat or ultrasound treatment indicated that the tested glycosidases are more sensitive to overheating than diastase ). However, it should be remembered that enzymatic activity may change under the influence of various factors, depending on the degree of honey maturity and the method of processing (Karabournioti and Zervalaki 2001;Zhang et al. 2021).
As many studies have proven that chemometric approach based on multivariate statistical analysis (PCA or LDA) can be used for honey classification in terms of botanical and geographical origin as well as to detect adulteration in honey (Oroian and Ropciuc 2017;Kaczmarek et al. 2019;Miłek et al. 2021;Tarapatskyy et al. 2021). Following the study of Tarapatskyy et al. (2021) where a clear clustering for 4 varieties of Polish honey using principal component analysis (PCA) has been observed, we use the same approach to check the relation between determined parameters (Fig. 2). The results indicated the strong positive correlation between protein content and tested enzymatic activities, excluding α-glucosidase, demonstrated by close proximity on the biplot (Fig. 2a). Similarly, weak correlation was observed for protein and free proline content (Fig. 2a). It can be observed that almost all tested parameters located close to the projection circle have a significant impact on the quality of honey. However, among examined parameters, the most significant impact on overall classification had protein and proline content as well as α-GLU and NAG activity, based on their location close to the projection circle.
Among the parameters studied, the activity of N-acetylglucosaminidase strongly correlates with the total protein content (r = 0.839), as well as the activity of β-galactosidase relatively well (r = 0.614). Moreover, the values for the protein correlated well with the proline content (r = 0.524), which correlated to a similar degree with NAG (r = 0.533). The PCA score plot (Fig. 2b) does not distinguish separate honey groups. However, samples with high values of tested parameters are located on the right part of the graph (positive PC1 and PC2 values for heather and positive PC1 and negative PC2 values for honeydew honey). Samples of pale honeys are grouped in the left part (negative PC1 values) of the graph but there is no clear varietal clustering. Similarly, Tarapatskyy et al. (2021) observed a close resemblance between multifloral and linden honey samples whereas buckwheat and honeydew honey formed clearly separated clusters. However, they considered a different set of variables, including detailed polyphenol, sugar, and elemental composition as well as physicochemical parameters. Fig. 3 Images of gels obtained for honey varieties. a Rape (R). b Acacia (A). c Linden (L). d Buckwheat (B). e Heather (HR). f Coniferous honeydew (CHD). M, protein mass marker (kDa). Standard profile, representative protein profiles considered as standard; questionable profile, example of a sample profile suspected of being adulterated or of lower quality ◂ 1 3

SDS-PAGE Protein Profiling
For the analysis of protein profiles using the SDS-PAGE technique, each variety was analyzed on a separate gel comparing local and commercial honeys. The gels resulted from electrophoretic separation were stained with Coomassie Brilliant Blue dye, and band patterns corresponding to individual proteins were obtained for each sample. As the honey solutions were applied to the gel at the same concentration, a direct comparison of the protein concentration in each honey due to the different intensity of staining bands after separation can be made. By processing in ImageJ software, protein profiles were obtained in the form of graphs, in which individual proteins are visualized in the form of peaks (Fig. 3).
Among the analyzed varieties, rape, linden, and buckwheat honeys were characterized by the greatest homogeneity of profiles between individual samples. Only 12.5% of rape (1 sample of 8), 12.5% linden, 12.5% acacia, 14% coniferous honeydew, 25% buckwheat (2 samples of 8), and 25% heather honey samples were found to be inconsistent with the varietal standard. Rape honey tracks are dominated by two protein bands (masses of about 59 and 70 kDa); the signal from a protein of about 95 kDa is also clearly marked. Only for one sample (R5, local honey) a different profile from the others was obtained; only two peaks from proteins with masses of about Fig. 4 The images of the gel obtained for deciduous honeydew honey samples. M, protein mass marker (kDa) 1 3 55 and 65 kDa were visible, which can contest the variety declared by the beekeeper. Moreover, in sample R10 (commercial blend), a low-weight protein fraction is visible at the lower edge of the gel. A typical profile, although different in terms of color intensity, was observed for acacia honey samples (Fig. 3b). It is dominated by two proteins, similar to rape honey (Fig. 3a). Two local samples (A1 and A3) had higher protein content (more intensive bands). The linden honeys have a typical stripe pattern, similar to rape, with slight differences in the intensity of the stripe corresponding to the protein with a mass of approx. 95 kDa visible for samples L2, L9, and L10 (Fig. 3c). In turn, buckwheat honey profiles are very characteristic for this variety, with an additional protein fraction in the range of 25-30 kDa, manifested by a wide peak with two maxima (Fig. 3d). In the upper part of the gels, a larger number of bands can be also clearly distinguished (in the range of 60-90 kDa). Two of the tested commercial honeys (B7 and B8) show much less intensely colored stripes, although still with the arrangement characteristic for buckwheat honey. Such profile differences may indicate worse quality of these honeys, either due to the wrong variety specification or overheating of the honey. This can only be verified by pollen analysis. Among the studied varieties of nectar honeys, heather honeys have the highest protein content (Table 1) and have a characteristic protein profile (Fig. 3e). It is dominated by three proteins with high masses (approx. 70 kDa, approx. 60 kDa, and approx. 48 kDa) as well as a very intense protein band with a mass of approx. 28 kDa. Among the analyzed samples, one (HR3-imported honey) does not show a typical profile, so it can be considered as a sample with an incorrectly defined variety. In the case of coniferous honeydew honeys, the profiles are more even; sample CHD4 (commercial) clearly stands out from the others, having additional peaks from proteins with masses of approx. 33 and 39 kDa and visible 2 peaks less intense than proteins and lower masses (Fig. 3f).
Similar protein profiles by SDS-PAGE have been obtained previously in other studies. Muresan et al. (2022) studied honeys of various geographic and botanical origins. The protein profiles obtained by them showed great similarity to those shown in this work; most of the main proteins visible on the gels were those with masses in the range of 45 to 85 kDa. They are considered to be proteins of bee origin (major royal jelly proteins, enzymes). The most intense bands were obtained in each case for a protein with a mass of about 57 kDa which they classified as MRJPs (50-60 kDa) (Lewkowski et al. 2019, Muresan et al. 2022). Alpha-glucosidase and MRJP-1 have been also identified as the main proteins of bee origin found in honey by Bocian et al. (2019). The low mass bands on the gels (up to several kDa) may be derived from peptides present in honey, e.g., royalisin and apisimin (Bilikova et al. 2015). Other proteins visible in the profiles may come from plants and constitute specific markers of botanical origin. Additional proteins with lower masses (in the range of 20-30 kDa) were obtained for honeys defined based on the melissopalynological analysis as buckwheat, sunflower, chestnut, or coriander (Muresan et al. 2022). Royal jelly proteins with a mass of 48 and 55 kDa (apalbumins) have been detected in honeys of various origins and are considered potential markers of honey quality and authenticity (Bilikova et al. 2015). Protein profiling by SDS-PAGE was also used as a tool to determine the quality of coniferous honeydew honeys. It has been shown that a high-mass protein (245 kDa) is characteristic of such honeys (Tomczyk et al. 2022). The comparison of honeydew honeys with other varieties made it possible to show nectar admixtures to some samples. As markers, proteins with lower masses (25-55 kDa), most probably derived from flower nectar (Tomczyk et al. 2022), were indicated. In the present study, the analysis was carried out on smaller gels; hence, it is impossible to demonstrate the presence or absence of the band at 245 kDa in the case of honeydew honeys. This indicates that more subtle differences between the varieties could be made visible by the use of high-resolution gels.
Proteomic analysis of varietal honeys in conjunction with the analysis of proteins derived from the pollen of various plants may also be a promising tool for the verification of the variety of honey (Baroni et al. 2002). Plant-based proteins and peptides may be promising markers of a botanical variety of honey, as demonstrated by the example of Manuka honey, in which 17 proteins derived from Leptospermum scoparium nectar were detected (Bong et al. 2021). The characteristic additional protein band was also found to be a varietal marker for Ziziphus jujuba honeys useful in the authentication of samples of this valuable honey (Zhang et al. 2019).
A high degree of variability was observed in the case of the analyzed deciduous honeydew honey (Fig. 4). This is probably related to the diversity of the source of the honey harvest, which is the honeydew, and to possible nectar admixtures in such honeys. In addition to typical proteins found in other varieties, additional peaks can be seen in the profiles of local honeydew honeys, derived from proteins with a mass of approx. 25-35 kDa (samples DHD1 and DHD3; a wide band with 3 visible maxima), or approx. 40 kDa (additional peak in sample DHD2). Commercial samples showed less variation. In the case of coniferous honeydew honeys, the profiles are more even; sample CHD4 clearly stands out from the others, having additional peaks from proteins with masses of approx. 33 and 39 kDa and visible 2 peaks less intense than proteins and lower masses.
As can be expected, the greatest diversity in the profiles of multifloral honeys was observed (Fig. 5). The sample M3 has additional proteins, similar to buckwheat honey. Samples of commercial multifloral honeys are low in proteins, which confirm the lower quality of these types of blends.
Previously, apart from the SDS-PAGE, also the native-PAGE technique was used to evaluate the quality of local honeys compared to imported blends. It has been shown that commercially available honeys are often characterized by lower protein content (Miłek et al. 2021). Using the native electrophoresis technique, different band patterns were obtained than in SDS-PAGE, which makes it impossible to directly compare the results with those presented in this study. It has been shown, however, that the zymography technique may be useful, e.g., by showing differences in the occurrence of amylase isoforms in honeys of various origins (Miłek et al. 2021). Zymography has also been used to characterize buckwheat honey. An in-depth proteomic analysis allowed for the identification of plant proteins (mainly enzymes from the hydrolases, lyases, and oxidoreductase classes) in addition to the main bee proteins of origin (from the MRJP family) (Borutinskaite et al. 2018).

Amino Acid Profile
The amino acid profiles were obtained for each variety on a separate thin-layer plate. Usually, profiling is performed by the HPLC method; for the first time, we used the HPTLC technique, which is cheaper and allows for simultaneous analysis of up to 20 samples on one chromatographic plate. Exemplary images of ninhydrin-derivatized HPTLC plates for the analyzed honey varieties are presented in Fig. 6.
All tested samples showed the presence of proline, visible as a yellow band at Rf = 0.23-0.26 (Fig. 6). Its concentration varied, but no clear relationship was found between the amount of proline and the botanical origin of honey. The most intense violet band present in all samples at Rf = 0.41 belongs to an unidentified compound, most probably not belonging to protein amino acids. For buckwheat honey, the characteristic amino acid is isoleucine (red band at Rf = 0.50), which can be considered a marker of this variety (Fig. 6b). It is also present in other honey varieties, but in a much lower concentration. Literature data confirm the exceptionally high content of isoleucine and leucine in Polish buckwheat honeys, compared to other varieties (Janiszewska et al. 2012). In samples of honeydew honeys, especially from deciduous honeydew (Fig. 5d), there is a distinct band assigned to aspartic acid (red, Rf = 0.28).
Generated profiles for tested honey varieties are presented in Fig. 7 which can be used as the standard for comparison with analyzed samples. The observed differences mainly concerned the lower activity of proline which is known as a marker of good honey quality (R3, A2, L10 samples). However, in some cases, different image was observed, suggesting the unclear variety declaration in questionable samples (B8: lack of isoleucine, CHD6; DHD8: lack of aspartic acids; or HR3: unexpected isoleucine and aspartic acid).
As in the case of protein profiles, also the amino acid profiles of multifloral honeys were very diversified. Additional amino acids appeared in them, such as leucine and tyrosine as well as aspartic acid (Fig. 8).
We propose the use of a relatively simple tool, which is thin-layer chromatography for the initial selection of honeys of adulterated variety or as a support for verifying honey variety by melissopalynology. Among tested honeys, 12.5% of acacia, 25% of buckwheat and heather, and 37.5% of linden honey were pointed out as questionable based on the analysis of amino acid profiles using the HPTLC method. Samples classified as not inconsistent with the reference profiles require variety verification by pollen analysis. The exemplary profiles of the lower quality samples shown in Fig. 6 were characterized either by negligible intensity of the amino acid bands or additional bands from the unusual amino acids of the given variety. In the case of a sample of heather honey (HR3), HPTLC analysis confirms the earlier hypothesis that it is honey with an incorrectly defined variety. It is distinguished by the presence of an intense band probably derived from isoleucine, which may suggest a high proportion of buckwheat nectar.
Amino acid profile analysis has previously been used as a tool to support the identification of the botanical origin of honey as well as the diversification of geographic origin (Cometto et al. 2003;Hermosin et al. 2003;Sun et al. 2017;Kowalski et al. 2017). Using chromatographic analysis supported by chemometric analysis, it was also shown that the content of individual free amino acids in honey and the ratio of exo-to endogenous amino acids vary depending on the origin of the honey. Such data may be useful in confirming the authenticity of honey as an extension of standard methods of quality evaluation and melissopalynological analysis limitation (Łozowicka et al. 2021).

Conclusions
The research showed the possibility of using the SDS-PAGE protein profile and the amino acid HPTLC to verify the beekeeper's or producer's declaration regarding the variety of Polish local and imported honey samples. The proposed indicators make it possible to confirm or question the declared variety and refer questionable samples to melissopalynological analysis. As an aid, a simple Bradford protein determination and the β-galactosidase activity can be used as a valuable marker, especially for heather, rape, and honeydew honeys. The proposed analytical tools can allow for the initial screening of honeys available on the market in terms of the Fig. 8 Exemplary HPTLC profiles obtained for multifloral honey samples adulteration of variety declaration which can limit the cost of pollen analysis in beekeeping. However, the implementation of methods in beekeeping practice requires verification for a larger number of honey samples. Moreover, the effectiveness of the proposed methods for honey varieties other than those produced in Central and Eastern Europe requires separate studies. The validation of the usefulness of proteins and amino acids as markers of honey authenticity would provide a simple and affordable tools that beekeeping practice is waiting for.