Development of an LC-MS/MS Method for Quantitation of Western Honeybee (Apis mellifera) α-Glucosidase III as a Potential Honey Authenticity Marker

Abstract

Western honeybee (Apis mellifera) α-glucosidase III (HBG-III), which is secreted from the hypopharyngeal glands of honeybees, plays a role in converting nectar into honey. Consequently, hypothesizing that HBG-III is a suitable marker of honey authenticity, we developed an analytical method to determine the HBG-III content and investigated its applicability to various commercial products. Following extraction from honey using phosphate-buffered saline, HBG-III was concentrated using an ultrafiltration membrane and subsequently fragmented with trypsin and lysyl endopeptidase mixture. The specific peptide fragments were used for quantitation by liquid chromatography-tandem mass spectrometry. The established method was validated for linearity, accuracy, precision, and the limit of quantitation (LOQ). As a result, the calibration curve was linear in the range of 0.01-0.3 μM, the mean recovery ranged from 73.8 to 89.2 %, the within-laboratory reproducibility (RSD_wr) ranged from 3.9 to 6.5 %, and the LOQ was 1.9 mg/kg. An investigation of HBG-III concentrations in 65 honey products available on the Japanese market revealed that the HBG-III content of 15 low-priced honey products was below the LOQ. This suggested that these products may be adulterated with non-honey syrups. Therefore, this method can serve as an effective tool to verify the authenticity of honey products.

Abbreviations

HBG-III, honeybee α-glucosidase III; HorRat, Horwitz ratio; LC, liquid chromatography; LC-MS/MS, liquid chromatography-tandem mass spectrometry; LOD, limit of detection; LOQ, limit of quantitation; MRJP, major royal jelly proteins; MS, mass spectrometry; NMR, nuclear magnetic resonance; PBS, phosphate-buffered saline; QTOF, quadrupole time-of-flight; RSD_r, repeatability; RSD_wr, within-laboratory reproducibility; Tris-HCl, tris (hydroxymethyl) aminomethane hydrochloride.

INTRODUCTION

Honey, a food with a long history, is one of the oldest nutritious and valuable foods known to humankind, as depicted in the beekeeping reliefs of the Pabasa tomb in ancient Egypt [1]. Research on honey, which continues to this day, has revealed that its main components are fructose and glucose and that it contains several types of enzymes in addition to various nutritional components [2]. Honey has several health benefits. It effectively aids in fatigue recovery and energy replenishment after exercise and is known to have an intestinal regulating effect as it is rich in oligosaccharides, phenols, and carbohydrates [3]. Moreover, it offers a healthy balance of amino acids, minerals, and protein and is rich in polyphenols including quercetin, which has strong antioxidant and antibacterial properties [4].

Owing to these useful properties, the history of honey is intertwined with that of fighting food fraud. High-quality honey, such as Manuka honey from New Zealand, has long commanded high prices [5]. Therefore, methods of counterfeiting honey, such as diluting it with syrup, are rampant [6]. Typically, cheap industrial syrups, such as high-fructose corn, rice, and sugar beet syrups, are used for adulteration. Therefore, a detection method using thin-layer chromatography with starch-derived polysaccharides as markers was developed to detect syrup in honey [7]. Stable carbon isotope ratio analysis was subsequently developed to identify syrups derived from C4 plants, such as corn and sugarcane [8]. However, this method cannot be used to detect syrups derived from C3 plants such as beets, which have the same carbon isotope signature as some common nectar sources of honey. Other analytical methods have been developed that target markers for various adulterated syrups [9, 10, 11]. Recently, nuclear magnetic resonance (NMR) has also been used to comprehensively investigate honey components [12, 13, 14, 15, 16]. However, as syrup production methods have diversified, it has become increasingly difficult to distinguish between adulterated syrup and pure honey using these methods.

In our previous study, we developed a highly sensitive and simple method for detecting foreign amylase, which is often to mimic diastase activity, an indicator of honey quality, using native polyacrylamide gel electrophoresis activity staining [17]. Using this method, we investigated honey products sold in Japan and found that foreign amylase was present in most low-cost honeys (less than 100 yen per 100 g). However, in countries such as Japan that do not use diastase activity as a quality indicator, adulteration may not be detected if the product is diluted with syrup and no foreign amylase is added. The reason why diastase activity is not used as a quality indicator is that honey is known to be heat-treated at about 60 °C to facilitate bottling and to prevent microbial fermentation [18], and there is concern that this process may overheat the honey, reducing its diastase activity.

In recent years, it has been suggested that the quality of honey and honeybee products should be defined in terms of the physiological functions of their authentic ingredients [19, 20, 21, 22, 23, 24, 25, 26]. The most specific components of honey are proteins derived from honeybees [27]. Albeit in minute contents, honey contains enzymes related to carbohydrate metabolism, such as α-glucosidase [28], glucose-6-oxidase [29], β-glucosidase [30], and the above-mentioned diastase [31], as well as major royal jelly proteins (MRJPs) [25, 32, 33, 34, 35, 36]. Among the latter, apalbumin1 (MRJP1) is the most abundant. Consequently, a method has been developed to use this component as a marker of honey authenticity [37]; however, the validity of the method has not been evaluated. Moreover, it is important to consider that although royal jelly is expensive, MRJP can be easily extracted in large quantities from royal jelly and added to honey. On the other hand, among the three types of western honeybee (Apis mellifera) α-glucosidase isozymes, honeybee α-glucosidase III (HBG-III) (amino acid sequence length: 567, molecular mass: 65,565 Da, Database (UniProt) accession: Q17058) is known to be involved in honey production because it decomposes sucrose, the main component of nectar, into glucose and fructose [38]. In addition, like α-amylase, HBG-III is secreted from the pharyngeal glands of honeybees, is present in honey, and listed in the reviewed Swiss-Prot database in UniProtKB. Furthermore, the most important honey worldwide, in terms of economic importance and other criteria, is produced by western honeybee, which is the most widely used beekeeping species worldwide [39, 40, 41]. Therefore, it was considered to be a suitable marker for evaluating the authenticity of honey produced by this species of honeybee.

With these facts in mind, in this study, we selected western honeybee HBG-III as a marker, fragmented it using proteases, and developed a reliable method to quantify the specific peptides obtained using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The technique of decomposing proteins with proteases and measuring the resulting peptide fragments using LC-MS/MS has recently been used as a method to analyze allergens in food [42, 43, 44, 45, 46, 47]. Subsequently, we applied the developed method to quantify HBG-III in honey products sold in Japan, and the α-glucosidase activity was also measured according to the method described in a previous study [19]. Our results revealed that inexpensive honey contains almost no HBG-III and exhibits low α-glucosidase activity, indicating that HBG-III is a suitable marker for assessing honey authenticity.

MATERIALS AND METHODS

Samples. Authenticated honey samples were obtained with the cooperation of the NATIONAL HONEY FAIR TRADE CONFERENCE (Tokyo, Japan). Commercially available honey samples and sugar beet syrup were purchased from retail outlets or online stores in Japan. Samples were stored in a refrigerator at 5 °C until use.

Sample preparation. After dissolving 4 g of the honey in Dulbecco's phosphate-buffered saline (PBS: Nacalai Tesque Inc., Kyoto, Japan), the volume was adjusted to 20 mL and the solution was centrifuged for 10 min (15,000 × G, 20 °C). The supernatant (10 mL) was transferred to 10 kDa Amicon^® Ultra 15 centrifugal filters (Merck KGaA, Darmstadt, Germany) and centrifuged for 30 min (5,000 × G, 20 °C). PBS (10 mL) was added to the obtained concentrate, which was then centrifuged for 20 min (5,000 × G, 20 °C). This operation was repeated six times to ensure sugar removal. The volume of the obtained concentrate was made up to 2 mL with water, and the solution was centrifuged for 10 min (15,000 × G, 20 °C). The resultant supernatant was used as honey extract. Subsequently, the honey extract (250 µL) was dried using a centrifugal evaporator (CC-105, Tomy Seiko Co., Ltd., Tokyo, Japan) for 80 min at 40 °C and dissolved in 100 μL of 50 mM Tris (hydroxymethyl) aminomethane hydrochloride (Tris-HCl) buffer (pH 8.0). After denaturing the honey protein using a heat block (DTU-N, Taitec Corporation, Saitama, Japan) for 90 min at 95 °C, 10 µL of 100 mM dithiothreitol (DTT) solution was added. The mixture was then incubated for 30 min at room temperature in the dark. Next, after adding 5 μL of 1 M iodoacetamide (IAA) solution and incubating at room temperature in the dark (30 min), 500 μL of 50 mM Tris-HCl buffer (pH 8.0) was added. To digest the honey proteins into peptide fragments, 2 μL of 0.2 μg/μL Trypsin/Lys-C Mix (Promega Corporation, Madison, WI, USA) solution was added, and the sample was subsequently incubated for 16 h at 37 °C in a water bath (Thermo Minder SD mini, Taitec Corporation). Following digestion, 30 μL of 20 % trifluoroacetic acid, and 50 μL of 0.5 μM internal standard solution (described in the Quantitation method section) were added and mixed. The entire volume of the solution was then loaded onto a solid-phase extraction column (InertSep HLB, 10 mg/1 mL, GL Sciences Inc., Tokyo, Japan) that was pre-conditioned with 250 μL methanol containing 0.5 % formic acid followed by 250 μL of 0.1 % formic acid solution. After loading the sample solution, the column was washed with 250 μL of 0.1 % formic acid solution, and the target peptides were eluted with 250 μL methanol containing 0.5 % formic acid. Next, 250 µL of 0.1 % formic acid was added to the eluate and the solution was filtered through a membrane filter (0.2 µm, polypropylene, Membrane Solutions LLC, Auburn, WA, USA), and analyzed using LC-MS/MS.

α-Glucosidase activity. The α-glucosidase activity was measured according to the method described in our previous study [19].

Detection of foreign amylase. Foreign amylase was detected using the method described in our previous study [19].

LC-quadrupole time-of-flight (QTOF)-MS conditions for peptide fragment analysis. Liquid chromatography (LC) was performed with a mobile phase of 0.1 % formic acid (solution A) and 0.1 % formic acid/acetonitrile (solution B), at a flow rate of 0.15 mL/min, under the gradient conditions: 2 % B (initial conditions) to 50 % B over 82 min. Following elution, a washing step was conducted, wherein B was increased to 90 % across 6 min and allowed to flow for a further 5 min. Separation was conducted using a Nexera X2 series instrument (Shimadzu Corporation, Kyoto, Japan) with a Develosil RPAQUEOUS-AR column (2 × 150 mm, particle size 3 μm) (Nomura Chemical Co., Ltd., Aichi, Japan). The column temperature was set at 40 °C and the injection volume of each solution was 3 μL. Mass spectrometry (MS) was performed using a TripleTOF 5600+ instrument (Sciex LLC, Framingham, MA, USA) under the following conditions: Electrospray ionization (+), data acquisition, information-dependent acquisition (IDA): desolvation temperature, 500 °C; ion spray voltage, 4,500 V; curtain gas, N₂ at 25 psi; nebulizer gas, zero air at 50 psi; turbo gas, zero air at 50 psi; mass range, m/z 50-1000. Analyst TF software Ver.1.7.1 (Sciex LLC) was used to detect peaks in the recorded chromatograms. The data files were processed with ProteinPilot version 5.0 software (Sciex LLC) using the Paragon algorithm [48].

LC-MS/MS conditions for quantitation. LC was run with a mobile phase of solution A and solution B, at a flow rate of 0.5 mL/min, under the gradient conditions: 2 % solution B at 0 min, 40 % solution B at 12 min, 80 % solution B at 14 min, 80 % solution B at 16 min, 2 % solution B at 16.1 min, and 2 % solution B at 25 min. Separation was performed using a 1290 Infinity II Series instrument (Agilent Technologies, Inc., Santa Clara, CA, USA) with an Acquity UPLC Peptide BEH C18 column (2.1 × 50 mm, particle size 1.7 μm) (Waters Corporation, Milford, MA, USA). The column temperature was set at 40 °C and the injection volume of each solution was 3 μL. MS was performed using an Ultivo instrument (Agilent Technologies, Inc.) under the following conditions: Dynamic multiple reaction monitoring mode; electrospray ionization (+); drying gas, 12 L/min of N₂ at 300 °C; sheath gas, 11 L/min of N₂ at 250 °C; nebulizer gas, N₂ at 55 psi; capillary voltage, 3,000 V; nozzle voltage, 1,000 V; and fragmentor voltage, 102 V. Subsequently, peaks in the obtained chromatogram were detected using the ‘Data Acquisition’ function in MassHunter software Ver.1.2 (Agilent Technologies, Inc.).

Quantitation method. Synthetic peptides (shown in Table 1, purity ≥ 95 %) were purchased from Cosmo Bio Co., Ltd. (Tokyo, Japan), and dissolved in methanol containing 0.5 % formic acid to prepare a 100 μM/L stock solution. Standard solutions were prepared by diluting stock solutions of peptide A and peptide B with a mixture of 0.5 % formic acid in methanol and 0.1 % formic acid solution (1:1). Thus, solutions with concentrations ranging from 0.01 to 0.3 μM were achieved, including an internal standard (peptide-A-SI labeled with stable isotope of arginine and peptide-B-SI labeled with stable isotope of lysine) solution, at a concentration of 0.05 μM. The calibration curve for each peptide was constructed from the peak area ratio of the internal standard solution and standard solution using a linear regression curve. The obtained calibration curve was then used to calculate the concentration of each peptide in the test solution, taking into account the sample amount. HBG-III is a glycosylated protein with a signal sequence of 17 amino acid residues [49]. Therefore, the HBG-III content (mg/kg) in the sample was converted from the molar concentration using the molecular mass (63,758 Da) subtracted the signal sequence and ignoring glycosylation.

Table 1. LC-MS/MS operating conditions for HBG-III quantitation.

Abbreviation of peptide	Sequence	Precursor ion (m/z)	Charge state	Product ion (m/z)	Fragment typea	RT (min)	Collision energy (eV)
A	IYTHDIPETYNVVR	574.0	+3	743.4c	b6	6.4	17
				751.4	y6		21
				650.4	y5		17
A-SIb	IYTHDIPETYNVV-R(13C6,15N4)	577.3	+3	743.4c	b6	6.4	17
B	NIEPYNNYYIWHPGK	636.6	+3	776.4c	y12	7.5	17
				840.9	y13		17
				228.1	b2		17
B-SIb	NIEPYNNYYIWHPG-K(13C6,15N2)	639.3	+3	780.4c	y12	7.5	17

^a Type b contains the N-terminus, type y contains the C-terminus, and the numbers indicate the total number of residues in the peptide minus 1.

^b Stable isotopic-labeled peptide.

^c Quantifier ion (no mark is qualifier ion).

Method validation. The method was validated according to AOAC guidelines for single laboratory validation of chemical methods for dietary supplements and botanicals [50]. To validate the developed method, the linearity of the calibration curve, selectivity, accuracy, precision, limit of quantitation (LOQ), and limit of detection (LOD) were confirmed. The linearity was evaluated at approximately eight equally spaced concentration levels. Selectivity and precision were evaluated by running Chinese blend honey twice on the same day and repeated on five different days. Accuracy was evaluated through fortified recovery tests using sugar beet syrup without HBG-III, which were conducted twice on the same day and repeated on five different days. The fortified concentrations were equivalent to 4, 10, and 25 mg/kg HBG-III. In addition, precision was confirmed in the fortified recovery tests. For accuracy, since high-purity HBG-III is not commercially available, crude HBG-III extracted from Japanese milk vetch honey was concentrated using ultrafiltration, and its concentration was confirmed under the measurement conditions of this method. The precision results were calculated using one-way analysis of variance to determine repeatability (RSD_r) and within-laboratory reproducibility (RSD_wr), and Horwitz ratio (HorRat)_r values were calculated by dividing the RSD_wr by the predicted value calculated using the Horwitz equation. The LOQ and LOD were calculated from the standard deviation (SD) of within-laboratory reproducibility for 4 mg/kg fortified recovery tests, and were determined as 3 × SD and 10 × SD, respectively.

RESULTS AND DISCUSSION

Sample preparation. HBG-III was extracted from honey by dissolving the samples in PBS, removing insoluble substances such as pollen by centrifugation, and then concentrating the supernatant by ultrafiltration. While dialysis [19, 25, 51, 52, 53] and various precipitation protocols [41, 54, 55, 56] are commonly used methods for separating proteins from honey, we opted for a 10 kDa ultrafiltration membrane. This decision was based on the marker protein being solely HBG-III, with a molecular mass of 65 kDa. Honey contains significant levels of reducing sugars such as glucose and fructose. If these sugars are not removed, the Maillard reaction during trypsin and lysyl endopeptidase mixture digestion may lead to the modification of basic amino acids. Consequently, we confirmed that washing the extract post-ultrafiltration with PBS six times effectively removes the reducing sugars. In addition, we used a polymer-based cartridge, InertSep HLB, to desalt the peptides after digestion. Our results revealed a high recovery of the target peptides upon elution with methanol containing 0.5 % formic acid (data not shown).

Selection of target peptides and optimization for LC-MS/MS operating conditions. MS-based protein analysis generally involves digesting the protein into peptide fragments using trypsin, and then selecting and quantifying specific peptides as targets. Therefore, we first decomposed HBG-III derived from Japanese milk vetch honey extract with trypsin and lysyl endopeptidase, and performed peptide mapping. The digestion products were analyzed using the IDA mode of LC-QTOF-MS. The resulting MS/MS spectra were searched using a ProteinPilot Paragon algorithm against protein databases [downloaded from UniProtKB] to generate a list of proteins, peptide sequences, intensities, and modification information. The results for the peptide fragments of HBG-III obtained are shown in Table S1 (see J. Appl. Glycosci. Web site). Next, the MS/MS spectral intensities and peak shapes of the obtained peptide fragments were matched using Skyline software Ver. 20.1 [57] (Fig. S1; see J. Appl. Glycosci. Web site). The peptide fragments used for quantitation should be 7-20 amino acid residues (because shorter fragments have poor selectivity and longer fragments have poor MS ionization efficiency), should not contain methionine or cysteine residues that are easily oxidized, and should be less affected by modification [43]. Therefore, two suitable peptides derived from HBG-III were selected: peptide-A (IYTHDIPETYNVVR) and peptide-B (NIEPYNNYYIWHPGK). The uniqueness of the two selected peptides was evaluated by searching against the NCBI non-redundant database using the protein-protein BLAST query, BLASTp, to verify that these peptides are specific to HBG-III and are unique to honeybee (Apis spp.). The amino acid sequence identity with HBG-III from other honeybee listed in UniProtKB was 95.4 % in A. cerana japonica, 96.3 % in A. cerana indica, 96.5 % in A. dorsata, and 94.4 % in A. florea. The amino acid sequence of peptide-A differed at two positions in A. cerana japonica and one position in A. cerana indica, and peptide-B differed at one position in A. florea. Thus, these peptides cannot be used as targets for honey of species such as A. cerana, which are currently restricted to Asia [39]. Finally, to optimize the LC-MS/MS operating conditions for the target peptides, synthetic peptides, including peptides labeled with stable isotopes, were obtained. These were used to determine the operating conditions for the precursor ions and product ions based on the peaks and intensities observed (Table 1).

Optimization of the denaturation method. The denaturation method for HBG-III was optimized using Chinese milk vetch honey and Hungarian acacia honey. We initially used a 50 mM Tris-HCl buffer (pH 8.0) containing 6 M urea; however, this led to a C-terminally carbamoylated target peptide fragment. Consequently, we proceeded to denature solely by heating at 95 °C in 50 mM Tris-HCl buffer (pH 8.0) to determine the optimal denaturation time, and found that stable HBG-III values were generated at 90 min (Fig. 1). In addition, we found that the conversion value to HBG-III was lower for peptide-B than for peptide-A. It has been reported that trypsin has higher substrate specificity for arginine residues than for lysine residues [58]. In this study, lysyl endopeptidase was added simultaneously with trypsin to promote degradation at lysine residues, but it was presumed that the substrate specificity of trypsin was more affected. Based on these results, we selected a denaturation time at 95 °C for 90 min, with peptide-A as the target for quantitation.

Fig. 1. Optimization of denaturation time at 95 °C using Chinese milk vetch and Hungarian acacia honeys.

　Plots of HBG-III content (mg/kg) versus time (min) for peptide-A and peptide-B from milk vetch and acacia. Values are a mean of duplicates.

Method validation. The linearity of the eight-point calibration curve was satisfactory (r² > 0.99) for peptide-A across the concentration range of 0.01-0.3 μM, corresponding to a HBG-III concentration range of 2-60 mg/kg in the sample.

Selectivity was confirmed by comparing the quantifier and qualifier ion values using Chinese blend honey. The average values obtained were 30.8 mg/kg for the quantifier ion (m/z 574.0 > 743.4), 29.8 mg/kg for the qualifier ion (m/z 574.0 > 650.4), and 29.6 mg/kg for the qualifier ion (m/z 574.0 > 751.4). These results indicated equivalent quantitative values, confirming the ability of the method to measure HBG-III selectively.

RSD_r and RSD_wr of the developed method were evaluated through fortified recovery tests using sugar beet syrup and performing measurements on Chinese blend honey. As shown in Table 2, both the RSD_r and the RSD_wr were below 6.5 %, while the HorRat_r values ranged from 0.8 to 1.1, indicating satisfactory results across all parameters. The accuracy of the developed method was also evaluated through fortified recovery tests using sugar beet syrup. As shown in Table 2, the recovery rates at each fortified level (4, 10, and 25 mg/kg) were 73.8, 86.2, and 89.2 %, respectively. Notably, the recovery rate at the 4 mg/kg fortified level was slightly lower than the range recommended by the AOAC guidelines. However, this discrepancy was considered reasonable given that HBG-III is a polymeric substance, suggesting the influence of irreversible adsorption onto the ultrafiltration membrane.

Table 2. Accuracy and precision based on recovery test, and precision based on actual sample.

Sample	Fortified (mg/kg)	Meana (mg/kg)	Recoverya (%)	RSDrb (%)	RSDwrc (%)	HorRatrd
Sugar beet syrup	4	2.95	73.8	4.7	6.4	0.9
	10	8.62	86.2	6.5	6.5	1.1
	25	22.3	89.2	2.9	3.9	0.8
Chinese blend	-	30.7	-	4.6	5.2	1.1

^a Mean of 10 replicates

^b Relative standard deviation (Repeatability)

^c Relative standard deviation (Within-laboratory reproducibility)

^d HorRat_r = RSD_wr (found, %) / RSD_wr (predicted, %), Predicted value was calculated as C^−0.15

The LOQ and LOD, calculated from the SD of the RSD_wr for the 4 mg/kg fortified recovery tests, were 1.9 and 0.6 mg/kg, respectively. These values were considered acceptable for assessing the authenticity of honey.

Application to authenticated samples. The authenticated samples were collected on-site and their authenticity was confirmed. The 104 samples were collected from 23 nectar sources in 16 countries, including Japan, Argentina, Hungary, and China. The results are shown in Table S2 (see J. Appl. Glycosci. Web site), and the concentration distribution is shown in Fig. S2 (see J. Appl. Glycosci. Web site). The mean HBG-III value was 30.5 mg/kg, with the highest content found in Japanese buckwheat honey (69.7 mg/kg) and the lowest found in New Zealand multifloral honey (11.5 mg/kg). The results in Table S3 (see J. Appl. Glycosci. Web site), show that the HBG-III contents of acacia honey tended to be lower than those for milk vetch honey and multifloral honey; however, the differences in the number of samples and the wide range of contents did not suggest a significant difference. Country-specific results also did not indicate a significant difference. Therefore, in this investigation, no bias in the distribution of HBG-III was identified according to the country or type of nectar source. The amount of enzymes in honey depends on various factors, such as the physiological state of the bees, the strength of the bee family, the temperature, and the nectar flow [59]. The enzymes are also added during nectar collection and transfer in the hive. The results of this investigation suggest that these factors all influence the distribution of HBG-III in honey.

Application to commercial samples. The results for the commercially available samples are listed in Table 3; typical selected reaction-monitoring chromatograms are presented in Fig. S3 (see J. Appl. Glycosci. Web site). Among the 15 lower-priced honey samples priced (less than 100 yen per 100 g), all except one (No. 14, from Myanmar), presented HBG-III values below the LOQ (1.9 mg/kg). Additionally, the α-glucosidase activity measured at the same time was below or slightly above the LOQ (0.005 U/g). These results suggest that α-glucosidase activity, which can be easily measured, may be used as a marker for adulterated syrup. However, in the domestically produced honey samples (Nos. 45 and 51), the α-glucosidase activity was low, but the quantitative value of HBG-III was significantly higher than the LOQ. Furthermore, when the correlation between α-glucosidase activity and HBG-III was confirmed for the 50 honeys that contained HBG-III, the coefficient of determination (r²) was low (0.52). This result was attributed to be due to the influence of heat treatment during the manufacturing process, since α-glucosidase activity is more susceptible to the effects of heating than the HBG-III content (Fig. S4; see J. Appl. Glycosci. Web site). Therefore, the HBG-III content, which is less affected by heat treatment, was considered more reliable for determining syrup adulteration. Furthermore, foreign amylase was detected in 13 samples in which the HBG-III content was below the LOQ; however, two samples (Nos. 15 and 17) were identified in which the HBG-III content was below the LOQ and no foreign amylase was detected. These samples may have been adulterated with syrup, without the addition of foreign amylase. Based on these results, HBG-III is considered a suitable authenticity marker even for samples that have been adulterated with syrup and do not contain foreign amylase.

Table 3. α-Glucosidase activities, detected foreign amylase, and HBG-III contents in honey samples.

No.	Honey source	Country	Price (yen/100 g)	α-Glucosidase activity (U/g)a	Foreign amylasea, b	HBG-III (mg/kg)a
1	Blend	China	36.8	< 0.005	Aspergullus	< 1.9
2	Blend	China	39.7	< 0.005	Geobacillus	< 1.9
3	Blend	China	45.2	< 0.005	Geobacillus	< 1.9
4	Blend	China	49.7	< 0.005	Aspergullus	< 1.9
5	Blend	China	49.8	0.008	Aspergullus	< 1.9
6	Blend	China	49.8	0.008	Geobacillus	< 1.9
7	Blend	China	49.8	0.025	Aspergullus	< 1.9
8	Blend	China	55.4	< 0.005	Geobacillus	< 1.9
9	Blend	China	59.2	< 0.005	Geobacillus	< 1.9
10	Blend	China	59.2	< 0.005	Geobacillus	< 1.9
11	Blend	Thailand	59.7	< 0.005	Aspergullus	< 1.9
12	Blend	China	59.7	0.006	Aspergullus	< 1.9
13	Blend	China	59.8	0.017	Aspergullus	< 1.9
14	Blend	Myanmar	69.1	0.201	N.D.	14.9
15	Blend	China	78.9	0.005	N.D.	< 1.9
16	Blend	China	107.8	0.161	N.D.	29.9
17	Blend	China	111.6	< 0.005	N.D.	< 1.9
18	Blend	Myanmar	119.8	0.026	N.D.	10.1
19	Blend	China	143.2	0.125	N.D.	17.2
20	Blend	Vietnam	147.2	0.030	N.D.	10.2
21	Milk vetch	China	150.0	0.047	N.D.	18.7
22	Blend	Mexico	154.0	0.038	N.D.	17.1
23	Acacia	China	162.0	0.041	N.D.	22.1
24	Blend	Argentina, New Zealand	164.5	0.078	N.D.	19.8
25	Acacia	China	182.7	0.099	N.D.	16.1
26	Sunflower	Ukraine	185.3	0.052	N.D.	11.8
27	Blend	China, Argentina	185.3	0.145	N.D.	8.8
28	Blend	Romania, Ukraine	199.2	0.068	N.D.	16.2
29	Blend	Canada	199.2	0.027	N.D.	18.7
30	Blend	Brazil	217.1	0.076	N.D.	7.7
31	Blend	Hungary, Canada	224.5	0.224	N.D.	12.6
32	Milk vetch	China	238.7	0.142	N.D.	16.1
33	Blend	Canada	263.3	0.310	N.D.	17.5
34	Blend	Argentina, Canada, Hungary, Japan	275.2	0.072	N.D.	15.9
35	Multifloral	Japan	285.5	0.212	N.D.	24.8
36	Milk vetch	Japan	370.3	1.243	N.D.	46.2
37	Lemon	Italy	383.1	0.034	N.D.	6.7
38	Acacia	Italy	383.1	0.072	N.D.	10.5
39	Acacia	Hungary	383.1	0.023	N.D.	18.9
40	Acacia	Switzerland	392.0	0.464	N.D.	22.9
41	Multifloral (honeycomb)	Canada	450.0	0.250	N.D.	28.4
42	Acacia (honeycomb)	Hungary	490.0	0.275	N.D.	8.3
43	Cherry blossoms	Japan	518.5	0.435	N.D.	24.9
44	Acacia	Japan	549.7	0.582	N.D.	41.0
45	Multifloral	Japan	552.0	0.007	N.D.	14.6
46	Multifloral	Japan	555.6	0.435	N.D.	25.9
47	Mandarin orange	Japan	592.6	0.504	N.D.	30.4
48	Multifloral	Japan	638.6	0.068	N.D.	9.2
49	Multifloral	Japan	670.0	0.415	N.D.	33.7
50	Acacia (honeycomb)	Hungary	671.0	0.284	N.D.	9.2
51	Multifloral	Japan	756.0	0.008	N.D.	16.9
52	Acacia	Japan	816.7	0.480	N.D.	23.2
53	Acacia	Japan	835.0	0.420	N.D.	24.3
54	Milk vetch	Japan	883.3	0.573	N.D.	32.6
55	Milk vetch	Japan	892.6	0.101	N.D.	6.7
56	Acacia	Japan	900.0	0.277	N.D.	16.0
57	Milk vetch	Japan	954.5	0.413	N.D.	20.5
58	Time	Greece	1000.0	0.206	N.D.	15.9
59	Acacia	Japan	1035.7	0.396	N.D.	11.3
60	Milk vetch	Japan	1035.7	0.531	N.D.	18.3
61	Lavender	France	1600.0	0.032	N.D.	12.4
62	Rosemary	France	1600.0	0.032	N.D.	13.9
63	Multifloral (honeycomb)	Japan	1600.8	0.581	N.D.	30.6
64	Acacia (honeycomb)	Japan	1922.0	0.621	N.D.	60.5
65	Apple (honeycomb)	Japan	3055.5	0.552	N.D.	27.1

^a Mean of duplicates.

^b Foreign amylase was detected by native PAGE activity staining method and LC-QTOF-MS.

CONCLUSIONS

As methods for the manufacture of syrup-adulterated honey have diversified and become increasingly sophisticated, it has become difficult to distinguish them from pure honey using marker substances derived from syrups. Therefore, we selected HBG-III as a marker and developed a quantitative method for its analysis using LC-MS/MS. The established method was validated for linearity, selectivity, precision, accuracy (recovery test), LOQ, and LOD. The HBG-III content in 104 authenticated honey samples ranged from 11.5 to 69.7 mg/kg; in contrast, the HBG-III contents of 65 honey products available on the Japanese market revealed that the HBG-III content of 15 low-priced honey products was below the LOQ. These low-priced honey samples also displayed low α-glucosidase activity and a high detection rate of foreign amylase. In two samples, the HBG-III levels were below the LOQ, and no foreign amylase was detected. These results indicate that HBG-III is suitable as an authenticity marker even for honey adulterated with syrup that does not contain foreign amylase. However, the target peptide used in this method has a different sequence from the HBG-III of honeybees such as A. cerana, and therefore cannot be applied to determine the authenticity of those honeys. In addition, although the amount of HBG-III in authenticated honey is estimated to be 10 mg/kg or more, some honey samples contained high concentrations of HBG-III. Therefore, if such honey is adulterated with syrup, HBG-III alone cannot determine whether it has been adulterated. Hence, to further increase confidence in the accurate identification of adulterated samples, supplementation with multiple methods such as NMR and foreign amylase detection is required.

CONFLICTS OF INTEREST

The authors of this study are staff members of the contracted analytical laboratory, Japan Food Research Laboratories.

ACKNOWLEDGMENTS

We sincerely thank the NATIONAL HONEY FAIR TRADE CONFERENCE for their cooperation in obtaining authenticated honey samples. We would also like to thank Editage (www.editage.jp) for English language editing.

REFERENCES

• [1]  Al Naggar Y, Codling G, Giesy JP, Safer A. Beekeeping and the need for pollination from an agricultural perspective in Egypt. Bee World. 2018; 95: 107-12.
• [2]  Bogdanov S, Jurendic T, Sieber R, Gallmann P. Honey for nutrition and health: A review. J Am Coll Nutr. 2008; 27: 677-89.
• [3]  Jiang L, Xie M, Chen G, Qiao J, Zhang H, Zeng X. Phenolics and carbohydrates in buckwheat honey regulate the human intestinal microbiota. Evid Based Complement Alternat Med. 2020; 2020: 6432942.
• [4]  Hossen MS, Ali MY, Jahurul MHA, Abdel-Daim MM, Gan SH, Khalil MI. Beneficial roles of honey polyphenols against some human degenerative diseases: A review. Pharmacol Rep. 2017; 69: 1194-205.
• [5]  Rogers KM, Somerton K, Rogers P, Cox J. Eliminating false positive C4 sugar tests on New Zealand Manuka honey. Rapid Commun Mass Spectrom. 2010; 24: 2370-4.
• [6]  Fakhlaei R, Selamat J, Khatib A, Razis AFA, Sukor R, Ahmad S, et al. The toxic impact of honey adulteration: a review. Foods. 2020; 9: 1538.
• [7]  White JW Jr. Detection of honey adulteration by carbohydrate analysis. J Assoc Off Anal Chem. 1980; 63: 11-8.
• [8]  Soares S, Amaral JS, Oliveira MBPP, Mafra I. A comprehensive review on the main honey authentication issues: Production and origin. Compr Rev Food Sci Food Saf. 2017; 16: 1072-100.
• [9]  Du B, Wu L, Xue X, Chen L, Li Y, Zhao J, et al. Rapid screening of multiclass syrup adulterants in honey by ultrahigh-performance liquid chromatography/quadrupole time of flight mass spectrometry. J Agric Food Chem. 2015; 63: 6614-23.
• [10]  Ruiz-Matute AI, Rodríguez-Sánchez S, Sanz ML, Martínez-Castro I. Detection of adulterations of honey with high fructose syrups from inulin by GC analysis. J Food Compost Anal. 2010; 23: 273-6.
• [11]  Xue X, Wang Q, Li Y, Wu L, Chen L, Zhao J, et al. 2‑Acetylfuran-3-glucopyranoside as a novel marker for the detection of honey adulterated with rice syrup. J Agric Food Chem. 2013; 61: 7488-93.
• [12]  Bertelli D, Lolli M, Papotti G, Bortolotti L, Serra G, Plessi M. Detection of honey adulteration by sugar syrups using one-dimensional and two-dimensional high-resolution nuclear magnetic resonance. J Agric Food Chem. 2010; 58: 8495-501.
• [13]  Rhee Y, Shilliday ER, Matviychuk Y, Nguyen T, Robinson N, Holland DJ, et al. Detection of honey adulteration using benchtop ¹H NMR spectroscopy. Anal Methods. 2023; 15: 1690-9.
• [14]  Ribeiro ROR, Mársico ET, Carneiro CS, Monteiro MLG, Júnior CC, Jesus EFO. Detection of honey adulteration of high fructose corn syrup by Low Field Nuclear Magnetic Resonance (LF ¹H NMR). J Food Eng. 2014; 135: 39-43.
• [15]  Schievano E, Sbrizza M, Zuccato V, Piana L, Tessari M. NMR carbohydrate profile in tracing acacia honey authenticity. Food Chem. 2020; 309: 125788.
• [16]  Siddiqui AJ, Musharraf SG, Choudhary MI, Rahman AU. Application of analytical methods in authentication and adulteration of honey. Food Chem. 2017; 217: 687-98.
• [17]  Takahashi Y, Yoshida I, Yokozeki T, Igarashi T, Fujita K. Investigation of foreign amylase adulteration in honey distributed in Japan by rapid and improved native PAGE activity staining method. J Appl Glycosci. 2023; 70: 67-73.
• [18]  Singh I, Singh S. Honey moisture reduction and its quality. J Food Sci Technol. 2018; 55: 3861-71.
• [19]  Azeredo LC, Azeredo MAA, de Souza SR, Dutra VML. Protein contents and physicochemical properties in honey samples of Apis mellifera of different floral origins. Food Chem. 2003; 80: 249-54.
• [20]  Bíliková K, Wu G, Šimúth J. Isolation of a peptide fraction from honeybee royal jelly as a potential antifoulbrood factor. Apidologie. 2001; 32: 275-83.
• [21]  Bíliková K, Mirgorodskaya E, Bukovská G, Gobom J, Lehrach H, Šimúth J. Towards functional proteomics of minority component of honeybee royal jelly: The effect of post-translational modifications on the antimicrobial activity of apalbumin2. Proteomics. 2009; 9: 2131-8.
• [22]  Brudzynski K, Abubaker K, St-Martin L, Castle A. Re-examining the role of hydrogen peroxide in bacteriostatic and bactericidal activities of honey. Front Microbiol. 2011; 2: 213.
• [23]  Cooper RA, Molan PC, Harding KG. The sensitivity to honey of Gram-positive cocci of clinical significance isolated from wounds. J Appl Microbiol. 2002; 93: 857-863.
• [24]  Molan PC. Why honey is effective as a medicine: 2. The scientific explanation of its effects. Bee World. 2001; 82: 22-40.
• [25]  Šimúth J, Bíliková K, Kováčová E, Kuzmová Z, Schroder W. Immunochemical approach to detection of adulteration in honey: Physiologically active royal jelly protein stimulating TNF-α release is a regular component of honey. J Agric Food Chem. 2004; 52: 2154-8.
• [26]  White JW Jr, Rudyj ON. The protein content of honey. J Apic Res. 1978; 17: 234-8.
• [27]  Erban T, Shcherbachenko E, Talacko P, Harant K. The unique protein composition of honey revealed by comprehensive proteomic analysis: Allergens, venom-like proteins, antibacterial properties, royal jelly proteins, serine proteases, and their inhibitors. J Nat Prod. 2019; 82: 1217-26.
• [28]  Ohashi K, Sawata M, Takeuchi H, Natori S, Kubo T. Molecular cloning of cDNA and analysis of expression of the gene for α-glucosidase from the hypopharyngeal gland of the honeybee Apis mellifera L. Biochem Biophys Res Commun. 1996; 221: 380-5.
• [29]  Schepartz AI. The glucose oxidase of honey II. Stereochemical substrate specificity. Biochim Biophys Acta. 1965; 96: 334-6.
• [30]  Pontoh J, Low NH. Purification and characterization of β-glucosidase from honeybees (Apis mellifera). Insect Biochem Mol Biol. 2002; 32: 679-90.
• [31]  Babacan S, Pivarnik LF, Rand AG. Honey amylase activity and food starch degradation. J Food Sci. 2002; 67: 1625-30.
• [32]  Di Girolamo F, D'Amato A, Righetti PG. Assessment of the floral origin of honey via proteomic tools. J Proteomics. 2012; 75: 3688-93.
• [33]  Hanes J, Šimúth J. Identification and partial characterization of the major royal jelly protein of the honey bee (Apis mellifera L.). J Apic Res. 1992; 31: 22-6.
• [34]  Malecová B, Ramser J, O'Brien JK, Janitz M, Júdová J, Lehrach H, et al. Honeybee (Apis mellifera L.) mrjp gene family: Computational analysis of putative promoters and genomic structure of mrjp1, the gene coding for the most abundant protein of larval food. Gene. 2003; 303: 165-75.
• [35]  Schmitzová J, Klaudiny J, Albert S, Schröder W, Schreckengost W, Hanes J, et al. A family of major royal jelly proteins of the honeybee Apis mellifera L. Cell Mol Life Sci. 1998; 54: 1020-30.
• [36]  Šimúth J. Some properties of the main protein of the honeybee (Apis mellifera) royal jelly. Apidologie. 2001; 32: 69-80.
• [37]  Bilikova K, Krakova TK, Yamaguchi K, Yamaguchi Y. Major royal jelly proteins as markers of authenticity and quality of honey. Arh Hig Rada Toksikol. 2015; 66: 259-67.
• [38]  Nishimoto M, Kubota M, Tsuji M, Mori H, Kimura A, Matsui H, et al. Purification and substrate specificity of honeybee, Apis mellifera L., α-glucosidase III. Biosci Biotechnol Biochem. 2001; 65: 1610-6.
• [39]  Han F, Wallberg A, Webster MT. From where did the Western honeybee (Apis mellifera) originate? Ecol Evol. 2012; 2: 1949-57.
• [40]  Kubota M, Tsuji M, Nishimoto M, Wongchawalit J, Okuyama M, Mori H, et al. Localization of α-glucosidases I, II, and III in organs of European honeybees. Biosci Biotechnol Biochem. 2004; 68: 2346-52.
• [41]  Lichtenberg-Kraag B. Evidence for correlation between invertase activity and sucrose content during the ripening process of honey. J Apic Res. 2014; 53: 364-73.
• [42]  Ippoushi K, Wakagi M, Hashimoto N, Takano-Ishikawa Y. Absolute quantification of the α, α′, and β subunits of β-conglycinin from soybeans by liquid chromatography/tandem mass spectrometry using stable isotope-labelled peptides. Food Res Int. 2019; 116: 1223-8.
• [43]  New LS, Schreiber A, Stahl-Zeng J, Liu HF. Simultaneous analysis of multiple allergens in food products by LC-MS/MS. J AOAC Int. 2018; 101: 132-45.
• [44]  Sato K, Jinno S, Nakamura Y, Eto S, Inagaki M. Absolute quantification of bovine lactadherin to screen the anti-rotavirus activity of dairy ingredients. J Dairy Sci. 2023; 106: 2261-70.
• [45]  Sealey-Voyksner J, Zweigenbaum J, Voyksner R. Discovery of highly conserved unique peanut and tree nut peptides by LC-MS/MS for multi-allergen detection. Food Chem. 2016; 194: 201-11.
• [46]  Torii A, Seki Y, Arimoto C, Hojo N, Iijima K, Nakamura K, et al. Development of a simple and reliable LC-MS/MS method to simultaneously detect walnut and almond as specified in food allergen labelling regulations in processed foods. Curr Res Food Sci. 2023; 6: 100444.
• [47]  Xiong W, Parker CH, Boo CC, Fiedler KL. Comparison of allergen quantification strategies for egg, milk, and peanut in food using targeted LC-MS/MS. Anal Bioanal Chem. 2021; 413: 5755-66.
• [48]  Shilov IV, Seymour SL, Patel AA, Loboda A, Tang WH, Keating SP, et al. The Paragon Algorithm, a next generation search engine that uses sequence temperature values and feature probabilities to identify peptides from tandem mass spectra. Mol Cell Proteomics. 2007; 6: 1638-55.
• [49]  Nishimoto M, Mori H, Moteki T, Takamura Y, Iwai G, Miyabuchi Y, et al. Molecular cloning of cDNAs and genes for three α-glucosidases from European honeybees, Apis mellifera L., and heterologous production of recombinant enzymes in Pichia pastoris. Biosci Biotechnol Biochem. 2007; 71: 1703-16.
• [50]  AOAC International. Official methods of analysis of AOAC International. Appendix K: Guidelines for dietary supplements and botanicals. 2012.
• [51]  Bauer L, Kohlich A, Hirschwehr R, Siemann U, Ebner H, Scheiner O, et al. Food allergy to honey: Pollen or bee products? Characterization of allergenic proteins in honey by means of immunoblotting. J Allergy Clin Immunol. 1996; 97: 65-73.
• [52]  Wang J, Kliks MM, Qu W, Jun S, Shi G, Li QX. Rapid determination of the geographical origin of honey based on protein fingerprinting and barcoding using MALDI TOF MS. J Agric Food Chem. 2009; 57: 10081-8.
• [53]  Won SR, Lee DC, Ko SH, Kim JW, Rhee HI. Honey major protein characterization and its application to adulteration detection. Food Res Int. 2008; 41: 952-6.
• [54]  Bocian A, Buczkowicz J, Jaromin M, Hus KK, Legáth J. An effective method of isolating honey proteins. Molecules. 2019; 24: 2399.
• [55]  Mohammed SEA, Azim MK. Characterisation of natural honey proteins: Implications for the floral and geographical origin of honey. Int J Food Sci Technol. 2012; 47: 362-8.
• [56]  Steinhorn G, Sims IM, Carnachan SM, Carr AJ, Schlothauer R. Isolation and characterisation of arabinogalactan-proteins from New Zealand kanuka honey. Food Chem. 2011; 128: 949-56.
• [57]  MacLean B, Tomazela DM, Shulman N, Chambers M, Finney GL, Frewen B, et al. Skyline: An open source document editor for creating and analyzing targeted proteomics experiments. Bioinformatics. 2010; 26: 966-8.
• [58]  Perry W, Cole PW, Murakami K, Inagami T. Specificity and mechanism of clostripain catalysis. Biochemistry. 1971; 10: 4246-52.
• [59]  Persano Oddo L, Piazza MG, Pulcini P. Invertase activity in honey. Apidologie. 1999; 30: 57-65.