- Email: [email protected]
Studies on the formation of methylglyoxal from dihydroxyacetone in Manuka (Leptospermum scoparium) honey
Carbohydrate Research 361 (2012) 7–11
Contents lists available at SciVerse ScienceDirect
Carbohydrate Research journal homepage: www.elsevier.com/locate/carres
Studies on the formation of methylglyoxal from dihydroxyacetone in Manuka (Leptospermum scoparium) honey Julia Atrott, Steffi Haberlau, Thomas Henle ⇑ Institute of Food Chemistry, Technische Universität Dresden, D-01062 Dresden, Germany
a r t i c l e
i n f o
Article history: Received 22 May 2012 Received in revised form 17 July 2012 Accepted 31 July 2012 Available online 8 August 2012 Keywords: Manuka honey Carbohydrate metabolism Dihydroxyacetone Methylglyoxal Quinoxaline Glycation
a b s t r a c t Dihydroxyacetone (DHA) and methylglyoxal (MGO) are unique carbohydrate metabolites of manuka honey. A method for the reliable quantification of DHA in honey samples was established, based on derivatization with o-phenylenediamine (OPD) and subsequent RP-HPLC with UV detection. The previously unknown reaction product of DHA and OPD was identified as 2-hydroxymethylquinoxaline by spectroscopic means. DHA was exclusively determined in 6 fresh manuka honeys originating directly from the beehive as well as 18 commercial manuka honey samples, ranging from 600 to 2700 mg/kg and 130 to 1600 mg/kg, respectively. The corresponding MGO contents varied from 50 to 250 mg/kg in fresh and 70 to 700 mg/kg in commercial manuka honey samples. A good linear correlation between DHA and MGO values in commercial manuka honeys was observed, resulting in a mean ratio of DHA to MGO of 2:1. In contrast to this, the DHA-to-MGO relation was much higher in fresh manuka honeys but approximated to a ratio of 2:1 while honey ripening. Heating experiments revealed that MGO formation based on thermal treatment as a consequence, for example, of caramelization in honey does not occur. DHA and MGO can serve as suitable unique quality parameter for manuka honey. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The presence of 1,2-dicarbonyl compounds such as 3-deoxyglucosone (3-DG), glyoxal (GO), and methylglyoxal (MGO) in honey was first described in 2004 by Weigel et al.1 as a consequence of sugar degradation or caramelization. Over a wide range of different honey types, 3-DG, which is the precursor for 5-hydroxymethylfurfural (HMF), could be determined in high amounts, varying between 80 and 1270 mg/kg and correlating with heating or storage conditions.1–3 Contents of GO and MGO generally were very low (up to 5 mg/kg). However, surprisingly high amounts of MGO (up to 760 mg/kg) were found exclusively in manuka honey.2 This honey is derived from the flowers of the manuka tree (Leptospermum scoparium) in New Zealand. It was clearly demonstrated that the pronounced antibacterial activity of New Zealand manuka honey directly originates from MGO.2 The amount of MGO correlates with the antibacterial activity in manuka honey,4,5 and, therefore, can be used as a tool for labelling the bioactivity of manuka honey. Manuka honey was shown to inhibit a wide range of microorganisms, including multiresistant strains.6–8 Furthermore, wound dressings containing manuka honey seem to be useful supports in clinical applications for wound healing.9,10 The origin and in particular the high concentration of MGO in manuka honey cannot be explained by sugar degradation in the ⇑ Corresponding author. Fax: +49 351 463 34138. E-mail address: [email protected] (T. Henle). 0008-6215/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.carres.2012.07.025
course of caramelization. In a recent report, Adams et al.11 identified dihydroxyacetone (DHA) as direct precursor for MGO formation in manuka honeys. In this study, the authors had determined DHA in nectar and fresh honey derived from manuka plants and proved that DHA is non-enzymatically transferred to MGO during honey storage. However, the origin of DHA stills remains enigmatic. DHA is a well-known degradation product of carbohydrates and was detected in caramelized mixtures next to other compounds as a result of the Maillard reaction.12 DHA was also found in naturally aged wine samples.13 Moreover, its presence was described in relation to fermentation processes of selected microorganisms.14 Furthermore, DHA is discussed as food additive, for example to enhance browning in baked foods,15 or for food preservation.16 In certain cosmetic products, namely self-tanning creams, DHA is the functional ingredient.17 Knowledge about the amount of DHA and MGO in honey is of prime importance for manufacturers as well as for labelling purpose. In the present study, a new method for the reliable quantification of DHA via RP-HPLC using pre-column derivatization with ophenylenediamine (OPD) is reported. The previously unknown derivative resulting from DHA and OPD is identified by spectroscopic means. DHA and MGO was measured for a number of commercially available and freshly produced manuka honeys in order to obtain information about the ratio between DHA and MGO and the transformation of DHA to MGO during storage. Furthermore, the influence of a thermal treatment on MGO content in
8
J. Atrott et al. / Carbohydrate Research 361 (2012) 7–11
manuka honey was studied in order to clarify whether the bioactive compound can be produced artificially in the final product. 2. Results and discussion 2.1. RP-HPLC analysis of DHA and characterization of DHA quinoxaline The analysis of 1,2-dicarbonyl compounds such as 3-desoxyglucosone (3-DG) or methylglyoxal (MGO) in honey is generally achieved by RP-HPLC of the corresponding quinoxalines resulting from derivatization with o-phenylenediamine (OPD). Under conditions applied for optimal derivatization of these compounds (phosphate buffer pH 6.5, incubation at room temperature), dihydroxyacetone (DHA) also formed an OPD-derivative eluting in the chromatogram at about 13 min (Fig. 1A). This reaction product was isolated using semi preparative RP-HPLC. The molar mass of the isolate was determined with 160 g/mol by LC-ESI-TOF-MS. The UV maximum at 319 nm was observed, varying only slightly when compared to the absorption maximum of 312 nm generally reported for quinoxalines. Based on the NMR data, the structure was identified as 2-hydroxymethylquinoxaline (2-quinoxalinemethanol, compound VII in Fig. 2). 2-Hydroxymethylquinoxaline has already been described as reaction product formed during incubation of sugars like glucose or ribose, respectively, with OPD.18,19 The identified quinoxalines derived from this reaction mixture were explained by the bufferinduced degradation of hexoses to 1,2-dicarbonyl fragments.19 However, the authors did not propose DHA being involved in quinoxaline formation. Up to now, it was generally accepted that DHA does not react with OPD to a stable derivative, and, therefore, cannot be quantified by means of RP-HPLC following derivatization with OPD.20 A possible reaction mechanism for the formation of 2-hydroxymethylquinoxaline from DHA and OPD is shown in Figure 2. First, a nucleophilic addition of one amino group of OPD (I) to the carbonyl carbon of DHA (II) occurs. The iminol (III) is formed by dehydration and is able to convert via the enaminol (IV) into the
ketoamine (V) according to the conversion of Amadori products. A second dehydration results in formation of an intermediate (VI), which is converted into the quinoxaline (VII), putatively by oxidation. Compared to dicarbonyl compounds like MGO or 3DG, complete derivatization of DHA with OPD takes more time for development of the quinoxaline. Therefore, it can be supposed that the formation of the keto group in (V) via enaminol (IV) is the rate determining step of the reaction. The oxidation of intermediate (VI) to the quinoxaline cannot be proven yet, but it may due to the oxidative properties of the reaction mixture, for example the influence of oxygen. Kinetic studies with varying incubation time, temperature, pH value, and molar ratio of DHA to OPD were performed in order to optimize derivatization of DHA to compound (VII) (data not shown). Finally, it could be shown that best results are obtained after incubation for 16 h at 37 °C using an acetate buffer with a pH value of 4.0 for sample preparation. Due to these differences in methodical conditions, reliable simultaneous quantification of DHA together with MGO was not possible. At pH value of 4.0, acid-induced formation of the MGO–quinoxaline was observed in carbohydrate solutions, and hence, MGO values would be overestimated. Otherwise, reaction of DHA and OPD is too slow at neutral pH, and therefore, DHA does not react completely with OPD under these conditions. Consequently, to reach optimal quinoxaline formation and best sensitivity for both analytes, two independent sample preparations and two chromatographic runs per sample are necessary. Performing DHA determination under these conditions, it became clear that notable amounts of DHA are also formed from the honey matrix itself, namely from glucose and fructose breakdown, during derivatization (Fig. 1B and C). This reproducible basis level of DHA has to be taken into account by using an artificial honey matrix as blank value.21 Calibration of (VII) using this sugar matrix as 10% solution in acetate buffer showed a good linearity (regression equation y = 0.15x + 5.66, R2 = 0.9944) in the range from 0.18 to 2.20 mmol/L DHA in sample solution, equivalent to DHA concentrations between 165 and 1980 mg/kg in honey. Quantification of DHA in manuka honey (Fig. 1D) was finally performed
Figure 1. RP-HPLC after derivatization with OPD of (A) DHA standard solution in water, (B) solution of fructose and glucose simulating honey matrix, (C) rape honey, (D) manuka honey.
J. Atrott et al. / Carbohydrate Research 361 (2012) 7–11
9
Figure 2. Proposed reaction mechanism for the reaction of DHA (II) with OPD (I) to 2-hydroxymethylquinoxaline (VII).
by matrix calibration using the above mentioned artificial honey matrix. LOD (limit of detection) and LOQ (limit of quantification) of DHA in test solutions were estimated with 0.01 and 0.12 mmol/L, respectively, representing a LOD of 10 mg/kg and LOQ of 110 mg/kg in honey while using 10% sample solution. Recovery of DHA added to the artificial honey matrix was determined with a mean of 95.3%. The relative repeatability for DHA quantification was assessed by repetitive analysis of one manuka honey sample and was at 7.8%. 2.2. Quantification of DHA and MGO in manuka honey Within this work, 18 commercial as well as 6 fresh manuka honey samples obtained directly from the beehive without further processing were analysed for their contents of DHA and MGO. The results are shown in Figure 3. In the fresh manuka honeys, DHA varied from 611 to 2724 mg/kg (median 1622 mg/kg). These fresh honey samples contained MGO in concentrations from 50 to 260 mg/kg (median 129 mg/kg), whereas in commercial manuka honeys, MGO contents were ranging from 66 to 684 mg/kg with a median value of 356 mg/kg, confirming the data in the literature.2,5 DHA in commercial samples varied from 127 to 1563 mg/ kg (median 735 mg/kg). A good linear correlation (regression equation y = 0.47x + 11.57, R2 = 0.8977) between the corresponding concentration of DHA and MGO in commercial manuka honeys was observed (Fig. 4, black squares), resulting in a mean ratio of DHA to MGO of 2:1. Compared to this, concentrations of DHA exceeded the values of MGO in fresh manuka honeys seven to sixteenfold (Fig. 4, open squares).
Figure 3. Concentrations of DHA and MGO in 6 fresh and 18 commercial manuka honeys (box plot, Whiskers indicate min/max values, = median value).
Figure 4. Relation of DHA to MGO in 18 commercial manuka honeys (dark squares), compared with 6 fresh manuka honeys before (opened squares) and after storage at 37 °C for 6 weeks (grey triangles).
During storage of these fresh samples at 37 °C for 6 weeks, the amount of DHA decreased considerably due to formation of MGO (Fig. 4, grey triangles). For stored fresh manuka honeys, a ratio of DHA to MGO ranging from 5 to 9 was calculated. A longer honey storage would be followed by a further increase of MGO content till there is reached a plateau which can be also observed for data reported by Adams et al.11 This conversion can be additionally characterized by the ratio of DHA to MGO which is approximating factor 2 in the course of honey ripening. Consequently, the discrimination between fresh and ‘ripened’ manuka honeys on the basis of defined DHA-to-MGO ratios seems to be a suitable tool for classification of manuka honeys and for monitoring reactions occurring during ripening. The simultaneous presence of DHA and MGO in certain relation to each other is a unique quality parameter for manuka honey. The findings of this study are in agreement with the results of Adams et al.11 These authors identified DHA as precursor for MGO formation by illustrating curves of DHA decrease and MGO increase while honey storage. Authors observed a similar relation of DHA to MGO after honey incubation. Adams et al.,11 however, had determined higher values of DHA and MGO in fresh manuka honeys, which can be due to the quantification method used by these authors. Other floral or honeydew honeys which were analysed in this study did not contain any DHA above LOD (Fig. 1C). The peak
10
J. Atrott et al. / Carbohydrate Research 361 (2012) 7–11
detectable for DHA in Figure 1C is due to the small amount of DHA formed from the honey matrix during derivatization (see above). Confirming the data of Weigel et al.,1 only small amounts of MGO ranging from 2 to 28 mg/kg were detected in these honeys. In contrast, other honeys from Leptospermum species sources like jellybush honey from Australia (Leptospermum polygalifolium) can also contain MGO and DHA in high amounts.20 Two samples of this honey type analysed in this study revealed an analogous DHA-toMGO ratio of 2:1. By evaluation of the data of Windsor et al.20 in the same way, it can be suggested that a similar conversion occurred in these honeys which can be illustrated by calculated DHA-to-MGO relation ranging from 18.5 till 0.7. However, these authors analysed partially very high values of MGO up to 1723 mg/kg by doing a simultaneous determination of DHA and MGO, raising doubts concerning a possible overestimation of MGO as a result of ‘neoformation’ during sample preparation, which cannot be dispelled, as several methodical details are missing in this paper.20 The phenomenon that DHA is not completely transferred to MGO during honey storage and that a constant ratio between DHA and MGO is obtained, might be due to the water content of honey being the limiting factor. MGO formation occurs by dehydration of DHA. A back reaction presumably does not occur since this transformation of DHA to MGO was described to be irreversible.22 The water content of 10 commercial manuka honeys was determined varied between 16.0% and 19.3% with a median value of 18.1%. This is in agreement with legislative regulations, which give very narrow limits for an acceptable water content in honey (maximum 20%).23 2.3. Influence of heat treatment on MGO levels in manuka honey Honey is not allowed to be heated.23 However, speculations concerning a fraudulent heat-treatment to increase the MGO content in order to obtain manuka honey with ‘optimized’ bioactivity, are conceivable. MGO may also increase while heating due to a potential release of ‘sugar-bound’ MGO.24 To study the effect of thermal treatment on MGO content, two commercial samples of manuka honey with different MGO contents and one sample of a rape honey were incubated at 70 °C for 10 min (resembling conditions of pasteurization) and up to 24 h for long-term heating. The results of these analyses, given in Table 1, show unambiguously that heat treatment of commercial rape honey did not lead to MGO formation, whereas the concentration of MGO in manuka honey remained constant while short heating and decreased markedly during long-term incubation at 70 °C. Moreover, the long-term storage of honey for 12 weeks at 37 °C resulted in a decrease of MGO in manuka honey, which was more pronounced in the sample which had been pre-heated for 10 min at 70 °C. These results clearly show that a thermal generation of MGO as a consequence of caramelization of sugars in honey is
Table 1 Concentration of MGO after thermal treatment of manuka and rape honeys at 70 °C, partially followed by storage at 37 °C for 12 weeks Incubation conditions
not possible. In general, treatment of honey with high temperatures is followed by massive changes in sensory properties like colour, smell, and taste, respectively. Moreover, a remarkable increase of 3-DG, resulting from degradation of glucose and fructose, and, in a consequence, of 5-hydroxymethylfurfural (HMF) will occur due to the heating processes, making such honeys inacceptable. 2.4. Conclusion Reliable analysis of DHA and MGO in honey was achieved by pre-column derivatization with OPD, followed by RP-HPLC with UV-detection. DHA is transformed to MGO during ripening of manuka honey, resulting in a constant DHA-to-MGO ratio of 2:1, which can be used as a quality index to monitor changes during storage and to classify commercial products. DHA and MGO are unique and naturally occurring constituents of manuka honey. 3. Materials and methods 3.1. Chemicals Methylglyoxal (MGO, 40% in water) and dihydroxyacetone (DHA) were obtained from Sigma–Aldrich (Steinheim, Germany) and the derivatizing agent o-phenylenediamine (OPD) from Fluka (Munich, Germany). Sodium acetate, sodium dihydrogenphosphate and disodium hydrogenphosphate were purchased from Grüssing (Filsum, Germany). Methanol (HPLC grade) and acetic acid were ordered from VWR Prolabo (Leuven, Belgium). Water used for buffer preparations and HPLC solvents was obtained using a Purelab plus purification system (USFilter, Ransbach-Baumbach, Germany). 3.2. Honey samples Manuka honeys were kindly provided by Manuka Health Ltd, Te Awamutu, New Zealand. In total, 18 commercial samples with varying methylglyoxal content and 6 samples of fresh manuka honey from different regions originating directly from the beehives without any further treatment were obtained. Other honey samples (rape, acacia, sunflower, bush flower, heather, chestnut, leatherwood, thyme, pine, and forest, in total 17 samples) were obtained from local supermarkets, or from Neuseelandhaus, Bergkamen, Germany (2 samples of jellybush honey), respectively. Water content of honey samples was measured refractrometrically according to Weigel et al.1 Additionally, an artificial honey matrix consisting of 46.5% fructose, 34.0% glucose, 1.5% sucrose, and 18.0% water was used.21 3.3. Honey storage and heat treatment To simulate a long-term storage, the 6 fresh manuka honeys were incubated at 37 °C for 6 weeks. To investigate the effect of thermal treatment, one rape honey and two manuka honeys were heated at 70 °C for 10 min as well as for 8 h and 24 h. Furthermore, samples heated for 10 min at 70 °C were incubated afterwards at 37 °C for 12 weeks to assess the combination of heating and storage of honey compared to a non-heated sample. 3.4. Sample preparation
MGO concentration (mg/kg)
70 °C
37 °C
Manuka 1
Manuka 2
Rape
0 min 0 min 10 min 10 min 8h 24 h
— +12 weeks — +12 weeks — —
100.7 90.6 107.9 67.4 68.7 35.9
307.2 268.1 329.1 241.8 237.8 132.8
2.3 5.3 3.1 5.5 3.8 3.9
The determination of MGO was performed according to Mavric et al.2 with slight modifications. For this, 1 mL aliquots of 10% (w/v) honey solutions in 0.5 M phosphate buffer (pH 6.5) were mixed with 300 lL phosphate buffer (pH 6.5) and 300 lL OPD solution (1% in phosphate buffer pH 6.5). Samples were incubated at room temperature overnight and membrane filtered (0.45 lm). For analysis of DHA, 1 mL aliquots of 10% honey solutions in 0.5 M acetate
J. Atrott et al. / Carbohydrate Research 361 (2012) 7–11
buffer (pH 4.0) were mixed with 300 lL acetate buffer (pH 4.0) and 300 lL OPD solution (1% in acetate buffer pH 4.0), followed by incubation for 16 h at 37 °C and membrane filtration (0.45 lm). 3.5. Analytical RP-HPLC HPLC analyses were performed using an Äkta basic system from Amersham Pharmacia Biotech (Uppsala, Sweden) with a pump P900 and an online degasser K-5004 (Knauer, Berlin, Germany) as well as an UV detector UV-900 and an auto sampler A-900. Peak evaluation was managed using the software UNICORN 4.11. The separation of quinoxalines was realized on a stainless steel column filled with Eurospher 100 RP18 material (250 mm 4.6 mm, 5 lm particle size, integrated pre column; Knauer, Berlin, Germany). The mobile phase were 0.075% acetic acid in water (solvent A) and a mixture of 80% methanol and 20% solvent A (solvent B). The gradient started with 40% solvent B for 1 min and then was elevated linearly to 100% B over a period of 20 min, was changed back to 40% B in 4 min and was held there for 7 min. The flow rate was 0.8 mL/ min, the separation was done at 30 °C, 20 lL sample solution was injected and peaks were detected by measurement of UV absorbance at 312 nm. Quantification was achieved by external calibration with standard solution for MGO or by matrix calibration using an artificial honey matrix according to Wahdan21 for DHA, respectively. The limits of detection (LOD) and quantification (LOQ) of DHA were calculated from blank values using artificial honey matrix (measurement of 10 independent blank values on one day).25 For determination of repeatability, 10 samples of a manuka honey were derivatized on different days. The recovery of DHA was estimated by spiking 10% solutions of artificial honey with varying concentrations of DHA and analysing as described above. 3.6. Isolation and identification of quinoxaline resulting from DHA and OPD Semipreparative HPLC was performed with a system consisting of two pumps K-1001 with mixing chamber, automatically driven valves for injection and fractionation, online degasser and UV detector K-2501 (all from Knauer, Berlin, Germany). Data handling was managed using the software Eurochrom 2000. The separation was realized on a stainless steel column filled with Eurospher 1008 C18 (33 mm 8 mm with integrated pre column; Knauer, Berlin, Germany). A DHA standard solution in water (2500 mg/L) after derivatization with OPD (1% in 0.5 M acetate buffer pH 4) for 16 h at 37 °C was used for fractionation. The derivative of DHA was separated with same solvents as mentioned above for analytical separations. The gradient started with 40% B for 1.2 min, elevated on 86% B in 14.2 min, on 95% in further 3 min, and on 100% in 1.8 min, then gradient changed back on 40% B in 3 min and was held there for further 7.8 min. The flow rate was set to 1.6 mL/min. Injection volume was 2 mL. The eluate collected from 16.4 to 19.4 min was evaporated to dryness, taken up in 2 mL water and lyophilized. Absorption spectrum of the isolate in water was recorded by a Specord S100 diode array spectrophotometer (Carl Zeiss, Jena, Germany). LC–MS measurement was realized by using a LC system 1100 Series (Agilent Technologies, PaloAlto, USA) with Mariner
11
ESI-TOF mass spectrometer (PerSeptive Biosystem, Framingham, USA). The chromatographic conditions were the same as described for analytical HPLC, 50 lL sample solution was injected. Electrospray ionization was used in the positive ionization mode. Full scan mass spectra were measured in mass range 100–1000 m/z in the tic-mode. For data acquisition the software Data Explorer Version 4.0.0.1 (Applied Biosystems, Foster City, USA) was used. 1 H NMR spectrum was recorded on a Bruker DRX 500 instrument (Rheinstetten, Germany) at 500 MHz. For this, 5 mg of the purified and lyophilized isolate was dissolved in 750 lL deuterium oxide. All chemical shifts are given in parts per million (ppm) relative to the internal HOD signal (4.70 ppm). Data of the isolated quinoxaline of DHA were as follows. 1H NMR (500 MHz, D2O), d [ppm]: 8.74 (s, H-3), 7.81–7.87 (m, H-5, H-8), 7.68–7.73 (m, H-6, H-7), 4.82 (s, H-10 A, H-10 B). Mass spectroscopy gave a m/z of 161 for [MH]+. UV spectroscopy gave a kmax of 319 nm. Acknowledgements The authors thank Dr. Uwe Schwarzenbolz, Institute of Food Chemistry, for his help during the LC-ESI-TOF-MS measurements, and furthermore, the members of the Institute of Organic Chemistry, namely Dr. Margit Gruner and Anett Rudolph, for recording the NMR data. References 1. Weigel, K. U.; Opitz, T.; Henle, T. Eur. Food Res. Technol. 2004, 218, 147–151. 2. Mavric, E.; Wittmann, S.; Barth, G.; Henle, T. Mol. Nutr. Food Res. 2008, 52, 483– 489. 3. Marceau, E.; Yaylayan, V. A. J. Agric. Food Chem. 2009, 57, 10837–10844. 4. Adams, C. J.; Boult, C. H.; Deadman, B. J.; Farr, J. M.; Grainger, M. N.; ManleyHarris, M.; Snow, M. J. Carbohydr. Res. 2008, 343, 651–659. 5. Atrott, J.; Henle, T. Czech. J. Food Sci. 2009, 27, 163–165. 6. Cooper, R. A.; Molan, P. C.; Harding, K. G. J. Appl. Microbial. 2002, 93(5), 857– 863. 7. Molan, P. Honey: Antimicrobial Actions and Role in Disease Management. In New Strategies combating bacterial infection, Ahmad, I., Aqil, F., Eds.; Wiley-VCH: Weinheim, Germany, 2008; pp 229–253. 8. George, N. M.; Cutting, K. F. Wounds 2010, 19, 231–236. 9. Molan, P. C. Int. J. Low. Extrem. Wounds 2006, 5, 40–54. 10. Simon, A.; Traynor, K.; Santos, K.; Blaser, G.; Bode, U.; Molan, P. eCAM 2009, 6, 165–173. 11. Adams, C. J.; Manley-Harris, M.; Molan, P. C. Carbohydr. Res. 2009, 344, 1050– 1053. 12. Severin, T.; Hiebl, J.; Popp-Ginsbach, H. Z. Lebensm. Unters. Forsch. 1984, 178, 284–287. 13. Laurie, V. F.; Waterhouse, A. L. J. Agric. Food Chem. 2006, 54, 4668–4673. 14. Green, S. R. U.S. 2948658, 1960 Aug 09. 15. Kirk, D. L. U.S. 3220850, 1965 Nov 30. 16. Oborsh, E. V.; Barkate, J. A.; Wesu, C.; Owen, T. M. CA 1054434 A1, 1979 May 15. 17. Monfrecola, G.; Prizio, E. Compr. Ser. Photosci. 2001, 3, 489–493. 18. Takagi, M.; Mizutani, M.; Tsuchiva, K. Bull. Univ. Osaka Prefecture, Ser. B: Agric. Life Sci. 1972, 24, 43–48. 19. Rizzi, G. P. J. Agric. Food Chem. 2004, 52, 953–957. 20. Windsor, S.; Pappalardo, M.; Brooks, P.; Williams, S.; Manley-Harris, M. J. Pharmacog. Phytother. 2012, 4, 6–11. 21. Wahdan, H. A. L. Infection 1998, 26, 30–35. 22. Fedoronko, M.; Koenigstein, J. Collect. Czech. Chem. Commun. 1969, 34, 3881– 3894. 23. Codex Alimentarius Commission: Revised Codex Standard for honey, Codex STAN 12-1981, Rev.1 (1987), Rev.2 (2001). 24. Stephens, J. C.; Schlothauer, R. C. WO 2010082845 A1, 2010 Jul 22. 25. Mocak, J.; Bond, A. M.; Mitchell, S.; Scollary, G. Pure Appl. Chem. 1997, 69, 297– 328.
Contents lists available at SciVerse ScienceDirect
Carbohydrate Research journal homepage: www.elsevier.com/locate/carres
Studies on the formation of methylglyoxal from dihydroxyacetone in Manuka (Leptospermum scoparium) honey Julia Atrott, Steffi Haberlau, Thomas Henle ⇑ Institute of Food Chemistry, Technische Universität Dresden, D-01062 Dresden, Germany
a r t i c l e
i n f o
Article history: Received 22 May 2012 Received in revised form 17 July 2012 Accepted 31 July 2012 Available online 8 August 2012 Keywords: Manuka honey Carbohydrate metabolism Dihydroxyacetone Methylglyoxal Quinoxaline Glycation
a b s t r a c t Dihydroxyacetone (DHA) and methylglyoxal (MGO) are unique carbohydrate metabolites of manuka honey. A method for the reliable quantification of DHA in honey samples was established, based on derivatization with o-phenylenediamine (OPD) and subsequent RP-HPLC with UV detection. The previously unknown reaction product of DHA and OPD was identified as 2-hydroxymethylquinoxaline by spectroscopic means. DHA was exclusively determined in 6 fresh manuka honeys originating directly from the beehive as well as 18 commercial manuka honey samples, ranging from 600 to 2700 mg/kg and 130 to 1600 mg/kg, respectively. The corresponding MGO contents varied from 50 to 250 mg/kg in fresh and 70 to 700 mg/kg in commercial manuka honey samples. A good linear correlation between DHA and MGO values in commercial manuka honeys was observed, resulting in a mean ratio of DHA to MGO of 2:1. In contrast to this, the DHA-to-MGO relation was much higher in fresh manuka honeys but approximated to a ratio of 2:1 while honey ripening. Heating experiments revealed that MGO formation based on thermal treatment as a consequence, for example, of caramelization in honey does not occur. DHA and MGO can serve as suitable unique quality parameter for manuka honey. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The presence of 1,2-dicarbonyl compounds such as 3-deoxyglucosone (3-DG), glyoxal (GO), and methylglyoxal (MGO) in honey was first described in 2004 by Weigel et al.1 as a consequence of sugar degradation or caramelization. Over a wide range of different honey types, 3-DG, which is the precursor for 5-hydroxymethylfurfural (HMF), could be determined in high amounts, varying between 80 and 1270 mg/kg and correlating with heating or storage conditions.1–3 Contents of GO and MGO generally were very low (up to 5 mg/kg). However, surprisingly high amounts of MGO (up to 760 mg/kg) were found exclusively in manuka honey.2 This honey is derived from the flowers of the manuka tree (Leptospermum scoparium) in New Zealand. It was clearly demonstrated that the pronounced antibacterial activity of New Zealand manuka honey directly originates from MGO.2 The amount of MGO correlates with the antibacterial activity in manuka honey,4,5 and, therefore, can be used as a tool for labelling the bioactivity of manuka honey. Manuka honey was shown to inhibit a wide range of microorganisms, including multiresistant strains.6–8 Furthermore, wound dressings containing manuka honey seem to be useful supports in clinical applications for wound healing.9,10 The origin and in particular the high concentration of MGO in manuka honey cannot be explained by sugar degradation in the ⇑ Corresponding author. Fax: +49 351 463 34138. E-mail address: [email protected] (T. Henle). 0008-6215/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.carres.2012.07.025
course of caramelization. In a recent report, Adams et al.11 identified dihydroxyacetone (DHA) as direct precursor for MGO formation in manuka honeys. In this study, the authors had determined DHA in nectar and fresh honey derived from manuka plants and proved that DHA is non-enzymatically transferred to MGO during honey storage. However, the origin of DHA stills remains enigmatic. DHA is a well-known degradation product of carbohydrates and was detected in caramelized mixtures next to other compounds as a result of the Maillard reaction.12 DHA was also found in naturally aged wine samples.13 Moreover, its presence was described in relation to fermentation processes of selected microorganisms.14 Furthermore, DHA is discussed as food additive, for example to enhance browning in baked foods,15 or for food preservation.16 In certain cosmetic products, namely self-tanning creams, DHA is the functional ingredient.17 Knowledge about the amount of DHA and MGO in honey is of prime importance for manufacturers as well as for labelling purpose. In the present study, a new method for the reliable quantification of DHA via RP-HPLC using pre-column derivatization with ophenylenediamine (OPD) is reported. The previously unknown derivative resulting from DHA and OPD is identified by spectroscopic means. DHA and MGO was measured for a number of commercially available and freshly produced manuka honeys in order to obtain information about the ratio between DHA and MGO and the transformation of DHA to MGO during storage. Furthermore, the influence of a thermal treatment on MGO content in
8
J. Atrott et al. / Carbohydrate Research 361 (2012) 7–11
manuka honey was studied in order to clarify whether the bioactive compound can be produced artificially in the final product. 2. Results and discussion 2.1. RP-HPLC analysis of DHA and characterization of DHA quinoxaline The analysis of 1,2-dicarbonyl compounds such as 3-desoxyglucosone (3-DG) or methylglyoxal (MGO) in honey is generally achieved by RP-HPLC of the corresponding quinoxalines resulting from derivatization with o-phenylenediamine (OPD). Under conditions applied for optimal derivatization of these compounds (phosphate buffer pH 6.5, incubation at room temperature), dihydroxyacetone (DHA) also formed an OPD-derivative eluting in the chromatogram at about 13 min (Fig. 1A). This reaction product was isolated using semi preparative RP-HPLC. The molar mass of the isolate was determined with 160 g/mol by LC-ESI-TOF-MS. The UV maximum at 319 nm was observed, varying only slightly when compared to the absorption maximum of 312 nm generally reported for quinoxalines. Based on the NMR data, the structure was identified as 2-hydroxymethylquinoxaline (2-quinoxalinemethanol, compound VII in Fig. 2). 2-Hydroxymethylquinoxaline has already been described as reaction product formed during incubation of sugars like glucose or ribose, respectively, with OPD.18,19 The identified quinoxalines derived from this reaction mixture were explained by the bufferinduced degradation of hexoses to 1,2-dicarbonyl fragments.19 However, the authors did not propose DHA being involved in quinoxaline formation. Up to now, it was generally accepted that DHA does not react with OPD to a stable derivative, and, therefore, cannot be quantified by means of RP-HPLC following derivatization with OPD.20 A possible reaction mechanism for the formation of 2-hydroxymethylquinoxaline from DHA and OPD is shown in Figure 2. First, a nucleophilic addition of one amino group of OPD (I) to the carbonyl carbon of DHA (II) occurs. The iminol (III) is formed by dehydration and is able to convert via the enaminol (IV) into the
ketoamine (V) according to the conversion of Amadori products. A second dehydration results in formation of an intermediate (VI), which is converted into the quinoxaline (VII), putatively by oxidation. Compared to dicarbonyl compounds like MGO or 3DG, complete derivatization of DHA with OPD takes more time for development of the quinoxaline. Therefore, it can be supposed that the formation of the keto group in (V) via enaminol (IV) is the rate determining step of the reaction. The oxidation of intermediate (VI) to the quinoxaline cannot be proven yet, but it may due to the oxidative properties of the reaction mixture, for example the influence of oxygen. Kinetic studies with varying incubation time, temperature, pH value, and molar ratio of DHA to OPD were performed in order to optimize derivatization of DHA to compound (VII) (data not shown). Finally, it could be shown that best results are obtained after incubation for 16 h at 37 °C using an acetate buffer with a pH value of 4.0 for sample preparation. Due to these differences in methodical conditions, reliable simultaneous quantification of DHA together with MGO was not possible. At pH value of 4.0, acid-induced formation of the MGO–quinoxaline was observed in carbohydrate solutions, and hence, MGO values would be overestimated. Otherwise, reaction of DHA and OPD is too slow at neutral pH, and therefore, DHA does not react completely with OPD under these conditions. Consequently, to reach optimal quinoxaline formation and best sensitivity for both analytes, two independent sample preparations and two chromatographic runs per sample are necessary. Performing DHA determination under these conditions, it became clear that notable amounts of DHA are also formed from the honey matrix itself, namely from glucose and fructose breakdown, during derivatization (Fig. 1B and C). This reproducible basis level of DHA has to be taken into account by using an artificial honey matrix as blank value.21 Calibration of (VII) using this sugar matrix as 10% solution in acetate buffer showed a good linearity (regression equation y = 0.15x + 5.66, R2 = 0.9944) in the range from 0.18 to 2.20 mmol/L DHA in sample solution, equivalent to DHA concentrations between 165 and 1980 mg/kg in honey. Quantification of DHA in manuka honey (Fig. 1D) was finally performed
Figure 1. RP-HPLC after derivatization with OPD of (A) DHA standard solution in water, (B) solution of fructose and glucose simulating honey matrix, (C) rape honey, (D) manuka honey.
J. Atrott et al. / Carbohydrate Research 361 (2012) 7–11
9
Figure 2. Proposed reaction mechanism for the reaction of DHA (II) with OPD (I) to 2-hydroxymethylquinoxaline (VII).
by matrix calibration using the above mentioned artificial honey matrix. LOD (limit of detection) and LOQ (limit of quantification) of DHA in test solutions were estimated with 0.01 and 0.12 mmol/L, respectively, representing a LOD of 10 mg/kg and LOQ of 110 mg/kg in honey while using 10% sample solution. Recovery of DHA added to the artificial honey matrix was determined with a mean of 95.3%. The relative repeatability for DHA quantification was assessed by repetitive analysis of one manuka honey sample and was at 7.8%. 2.2. Quantification of DHA and MGO in manuka honey Within this work, 18 commercial as well as 6 fresh manuka honey samples obtained directly from the beehive without further processing were analysed for their contents of DHA and MGO. The results are shown in Figure 3. In the fresh manuka honeys, DHA varied from 611 to 2724 mg/kg (median 1622 mg/kg). These fresh honey samples contained MGO in concentrations from 50 to 260 mg/kg (median 129 mg/kg), whereas in commercial manuka honeys, MGO contents were ranging from 66 to 684 mg/kg with a median value of 356 mg/kg, confirming the data in the literature.2,5 DHA in commercial samples varied from 127 to 1563 mg/ kg (median 735 mg/kg). A good linear correlation (regression equation y = 0.47x + 11.57, R2 = 0.8977) between the corresponding concentration of DHA and MGO in commercial manuka honeys was observed (Fig. 4, black squares), resulting in a mean ratio of DHA to MGO of 2:1. Compared to this, concentrations of DHA exceeded the values of MGO in fresh manuka honeys seven to sixteenfold (Fig. 4, open squares).
Figure 3. Concentrations of DHA and MGO in 6 fresh and 18 commercial manuka honeys (box plot, Whiskers indicate min/max values, = median value).
Figure 4. Relation of DHA to MGO in 18 commercial manuka honeys (dark squares), compared with 6 fresh manuka honeys before (opened squares) and after storage at 37 °C for 6 weeks (grey triangles).
During storage of these fresh samples at 37 °C for 6 weeks, the amount of DHA decreased considerably due to formation of MGO (Fig. 4, grey triangles). For stored fresh manuka honeys, a ratio of DHA to MGO ranging from 5 to 9 was calculated. A longer honey storage would be followed by a further increase of MGO content till there is reached a plateau which can be also observed for data reported by Adams et al.11 This conversion can be additionally characterized by the ratio of DHA to MGO which is approximating factor 2 in the course of honey ripening. Consequently, the discrimination between fresh and ‘ripened’ manuka honeys on the basis of defined DHA-to-MGO ratios seems to be a suitable tool for classification of manuka honeys and for monitoring reactions occurring during ripening. The simultaneous presence of DHA and MGO in certain relation to each other is a unique quality parameter for manuka honey. The findings of this study are in agreement with the results of Adams et al.11 These authors identified DHA as precursor for MGO formation by illustrating curves of DHA decrease and MGO increase while honey storage. Authors observed a similar relation of DHA to MGO after honey incubation. Adams et al.,11 however, had determined higher values of DHA and MGO in fresh manuka honeys, which can be due to the quantification method used by these authors. Other floral or honeydew honeys which were analysed in this study did not contain any DHA above LOD (Fig. 1C). The peak
10
J. Atrott et al. / Carbohydrate Research 361 (2012) 7–11
detectable for DHA in Figure 1C is due to the small amount of DHA formed from the honey matrix during derivatization (see above). Confirming the data of Weigel et al.,1 only small amounts of MGO ranging from 2 to 28 mg/kg were detected in these honeys. In contrast, other honeys from Leptospermum species sources like jellybush honey from Australia (Leptospermum polygalifolium) can also contain MGO and DHA in high amounts.20 Two samples of this honey type analysed in this study revealed an analogous DHA-toMGO ratio of 2:1. By evaluation of the data of Windsor et al.20 in the same way, it can be suggested that a similar conversion occurred in these honeys which can be illustrated by calculated DHA-to-MGO relation ranging from 18.5 till 0.7. However, these authors analysed partially very high values of MGO up to 1723 mg/kg by doing a simultaneous determination of DHA and MGO, raising doubts concerning a possible overestimation of MGO as a result of ‘neoformation’ during sample preparation, which cannot be dispelled, as several methodical details are missing in this paper.20 The phenomenon that DHA is not completely transferred to MGO during honey storage and that a constant ratio between DHA and MGO is obtained, might be due to the water content of honey being the limiting factor. MGO formation occurs by dehydration of DHA. A back reaction presumably does not occur since this transformation of DHA to MGO was described to be irreversible.22 The water content of 10 commercial manuka honeys was determined varied between 16.0% and 19.3% with a median value of 18.1%. This is in agreement with legislative regulations, which give very narrow limits for an acceptable water content in honey (maximum 20%).23 2.3. Influence of heat treatment on MGO levels in manuka honey Honey is not allowed to be heated.23 However, speculations concerning a fraudulent heat-treatment to increase the MGO content in order to obtain manuka honey with ‘optimized’ bioactivity, are conceivable. MGO may also increase while heating due to a potential release of ‘sugar-bound’ MGO.24 To study the effect of thermal treatment on MGO content, two commercial samples of manuka honey with different MGO contents and one sample of a rape honey were incubated at 70 °C for 10 min (resembling conditions of pasteurization) and up to 24 h for long-term heating. The results of these analyses, given in Table 1, show unambiguously that heat treatment of commercial rape honey did not lead to MGO formation, whereas the concentration of MGO in manuka honey remained constant while short heating and decreased markedly during long-term incubation at 70 °C. Moreover, the long-term storage of honey for 12 weeks at 37 °C resulted in a decrease of MGO in manuka honey, which was more pronounced in the sample which had been pre-heated for 10 min at 70 °C. These results clearly show that a thermal generation of MGO as a consequence of caramelization of sugars in honey is
Table 1 Concentration of MGO after thermal treatment of manuka and rape honeys at 70 °C, partially followed by storage at 37 °C for 12 weeks Incubation conditions
not possible. In general, treatment of honey with high temperatures is followed by massive changes in sensory properties like colour, smell, and taste, respectively. Moreover, a remarkable increase of 3-DG, resulting from degradation of glucose and fructose, and, in a consequence, of 5-hydroxymethylfurfural (HMF) will occur due to the heating processes, making such honeys inacceptable. 2.4. Conclusion Reliable analysis of DHA and MGO in honey was achieved by pre-column derivatization with OPD, followed by RP-HPLC with UV-detection. DHA is transformed to MGO during ripening of manuka honey, resulting in a constant DHA-to-MGO ratio of 2:1, which can be used as a quality index to monitor changes during storage and to classify commercial products. DHA and MGO are unique and naturally occurring constituents of manuka honey. 3. Materials and methods 3.1. Chemicals Methylglyoxal (MGO, 40% in water) and dihydroxyacetone (DHA) were obtained from Sigma–Aldrich (Steinheim, Germany) and the derivatizing agent o-phenylenediamine (OPD) from Fluka (Munich, Germany). Sodium acetate, sodium dihydrogenphosphate and disodium hydrogenphosphate were purchased from Grüssing (Filsum, Germany). Methanol (HPLC grade) and acetic acid were ordered from VWR Prolabo (Leuven, Belgium). Water used for buffer preparations and HPLC solvents was obtained using a Purelab plus purification system (USFilter, Ransbach-Baumbach, Germany). 3.2. Honey samples Manuka honeys were kindly provided by Manuka Health Ltd, Te Awamutu, New Zealand. In total, 18 commercial samples with varying methylglyoxal content and 6 samples of fresh manuka honey from different regions originating directly from the beehives without any further treatment were obtained. Other honey samples (rape, acacia, sunflower, bush flower, heather, chestnut, leatherwood, thyme, pine, and forest, in total 17 samples) were obtained from local supermarkets, or from Neuseelandhaus, Bergkamen, Germany (2 samples of jellybush honey), respectively. Water content of honey samples was measured refractrometrically according to Weigel et al.1 Additionally, an artificial honey matrix consisting of 46.5% fructose, 34.0% glucose, 1.5% sucrose, and 18.0% water was used.21 3.3. Honey storage and heat treatment To simulate a long-term storage, the 6 fresh manuka honeys were incubated at 37 °C for 6 weeks. To investigate the effect of thermal treatment, one rape honey and two manuka honeys were heated at 70 °C for 10 min as well as for 8 h and 24 h. Furthermore, samples heated for 10 min at 70 °C were incubated afterwards at 37 °C for 12 weeks to assess the combination of heating and storage of honey compared to a non-heated sample. 3.4. Sample preparation
MGO concentration (mg/kg)
70 °C
37 °C
Manuka 1
Manuka 2
Rape
0 min 0 min 10 min 10 min 8h 24 h
— +12 weeks — +12 weeks — —
100.7 90.6 107.9 67.4 68.7 35.9
307.2 268.1 329.1 241.8 237.8 132.8
2.3 5.3 3.1 5.5 3.8 3.9
The determination of MGO was performed according to Mavric et al.2 with slight modifications. For this, 1 mL aliquots of 10% (w/v) honey solutions in 0.5 M phosphate buffer (pH 6.5) were mixed with 300 lL phosphate buffer (pH 6.5) and 300 lL OPD solution (1% in phosphate buffer pH 6.5). Samples were incubated at room temperature overnight and membrane filtered (0.45 lm). For analysis of DHA, 1 mL aliquots of 10% honey solutions in 0.5 M acetate
J. Atrott et al. / Carbohydrate Research 361 (2012) 7–11
buffer (pH 4.0) were mixed with 300 lL acetate buffer (pH 4.0) and 300 lL OPD solution (1% in acetate buffer pH 4.0), followed by incubation for 16 h at 37 °C and membrane filtration (0.45 lm). 3.5. Analytical RP-HPLC HPLC analyses were performed using an Äkta basic system from Amersham Pharmacia Biotech (Uppsala, Sweden) with a pump P900 and an online degasser K-5004 (Knauer, Berlin, Germany) as well as an UV detector UV-900 and an auto sampler A-900. Peak evaluation was managed using the software UNICORN 4.11. The separation of quinoxalines was realized on a stainless steel column filled with Eurospher 100 RP18 material (250 mm 4.6 mm, 5 lm particle size, integrated pre column; Knauer, Berlin, Germany). The mobile phase were 0.075% acetic acid in water (solvent A) and a mixture of 80% methanol and 20% solvent A (solvent B). The gradient started with 40% solvent B for 1 min and then was elevated linearly to 100% B over a period of 20 min, was changed back to 40% B in 4 min and was held there for 7 min. The flow rate was 0.8 mL/ min, the separation was done at 30 °C, 20 lL sample solution was injected and peaks were detected by measurement of UV absorbance at 312 nm. Quantification was achieved by external calibration with standard solution for MGO or by matrix calibration using an artificial honey matrix according to Wahdan21 for DHA, respectively. The limits of detection (LOD) and quantification (LOQ) of DHA were calculated from blank values using artificial honey matrix (measurement of 10 independent blank values on one day).25 For determination of repeatability, 10 samples of a manuka honey were derivatized on different days. The recovery of DHA was estimated by spiking 10% solutions of artificial honey with varying concentrations of DHA and analysing as described above. 3.6. Isolation and identification of quinoxaline resulting from DHA and OPD Semipreparative HPLC was performed with a system consisting of two pumps K-1001 with mixing chamber, automatically driven valves for injection and fractionation, online degasser and UV detector K-2501 (all from Knauer, Berlin, Germany). Data handling was managed using the software Eurochrom 2000. The separation was realized on a stainless steel column filled with Eurospher 1008 C18 (33 mm 8 mm with integrated pre column; Knauer, Berlin, Germany). A DHA standard solution in water (2500 mg/L) after derivatization with OPD (1% in 0.5 M acetate buffer pH 4) for 16 h at 37 °C was used for fractionation. The derivative of DHA was separated with same solvents as mentioned above for analytical separations. The gradient started with 40% B for 1.2 min, elevated on 86% B in 14.2 min, on 95% in further 3 min, and on 100% in 1.8 min, then gradient changed back on 40% B in 3 min and was held there for further 7.8 min. The flow rate was set to 1.6 mL/min. Injection volume was 2 mL. The eluate collected from 16.4 to 19.4 min was evaporated to dryness, taken up in 2 mL water and lyophilized. Absorption spectrum of the isolate in water was recorded by a Specord S100 diode array spectrophotometer (Carl Zeiss, Jena, Germany). LC–MS measurement was realized by using a LC system 1100 Series (Agilent Technologies, PaloAlto, USA) with Mariner
11
ESI-TOF mass spectrometer (PerSeptive Biosystem, Framingham, USA). The chromatographic conditions were the same as described for analytical HPLC, 50 lL sample solution was injected. Electrospray ionization was used in the positive ionization mode. Full scan mass spectra were measured in mass range 100–1000 m/z in the tic-mode. For data acquisition the software Data Explorer Version 4.0.0.1 (Applied Biosystems, Foster City, USA) was used. 1 H NMR spectrum was recorded on a Bruker DRX 500 instrument (Rheinstetten, Germany) at 500 MHz. For this, 5 mg of the purified and lyophilized isolate was dissolved in 750 lL deuterium oxide. All chemical shifts are given in parts per million (ppm) relative to the internal HOD signal (4.70 ppm). Data of the isolated quinoxaline of DHA were as follows. 1H NMR (500 MHz, D2O), d [ppm]: 8.74 (s, H-3), 7.81–7.87 (m, H-5, H-8), 7.68–7.73 (m, H-6, H-7), 4.82 (s, H-10 A, H-10 B). Mass spectroscopy gave a m/z of 161 for [MH]+. UV spectroscopy gave a kmax of 319 nm. Acknowledgements The authors thank Dr. Uwe Schwarzenbolz, Institute of Food Chemistry, for his help during the LC-ESI-TOF-MS measurements, and furthermore, the members of the Institute of Organic Chemistry, namely Dr. Margit Gruner and Anett Rudolph, for recording the NMR data. References 1. Weigel, K. U.; Opitz, T.; Henle, T. Eur. Food Res. Technol. 2004, 218, 147–151. 2. Mavric, E.; Wittmann, S.; Barth, G.; Henle, T. Mol. Nutr. Food Res. 2008, 52, 483– 489. 3. Marceau, E.; Yaylayan, V. A. J. Agric. Food Chem. 2009, 57, 10837–10844. 4. Adams, C. J.; Boult, C. H.; Deadman, B. J.; Farr, J. M.; Grainger, M. N.; ManleyHarris, M.; Snow, M. J. Carbohydr. Res. 2008, 343, 651–659. 5. Atrott, J.; Henle, T. Czech. J. Food Sci. 2009, 27, 163–165. 6. Cooper, R. A.; Molan, P. C.; Harding, K. G. J. Appl. Microbial. 2002, 93(5), 857– 863. 7. Molan, P. Honey: Antimicrobial Actions and Role in Disease Management. In New Strategies combating bacterial infection, Ahmad, I., Aqil, F., Eds.; Wiley-VCH: Weinheim, Germany, 2008; pp 229–253. 8. George, N. M.; Cutting, K. F. Wounds 2010, 19, 231–236. 9. Molan, P. C. Int. J. Low. Extrem. Wounds 2006, 5, 40–54. 10. Simon, A.; Traynor, K.; Santos, K.; Blaser, G.; Bode, U.; Molan, P. eCAM 2009, 6, 165–173. 11. Adams, C. J.; Manley-Harris, M.; Molan, P. C. Carbohydr. Res. 2009, 344, 1050– 1053. 12. Severin, T.; Hiebl, J.; Popp-Ginsbach, H. Z. Lebensm. Unters. Forsch. 1984, 178, 284–287. 13. Laurie, V. F.; Waterhouse, A. L. J. Agric. Food Chem. 2006, 54, 4668–4673. 14. Green, S. R. U.S. 2948658, 1960 Aug 09. 15. Kirk, D. L. U.S. 3220850, 1965 Nov 30. 16. Oborsh, E. V.; Barkate, J. A.; Wesu, C.; Owen, T. M. CA 1054434 A1, 1979 May 15. 17. Monfrecola, G.; Prizio, E. Compr. Ser. Photosci. 2001, 3, 489–493. 18. Takagi, M.; Mizutani, M.; Tsuchiva, K. Bull. Univ. Osaka Prefecture, Ser. B: Agric. Life Sci. 1972, 24, 43–48. 19. Rizzi, G. P. J. Agric. Food Chem. 2004, 52, 953–957. 20. Windsor, S.; Pappalardo, M.; Brooks, P.; Williams, S.; Manley-Harris, M. J. Pharmacog. Phytother. 2012, 4, 6–11. 21. Wahdan, H. A. L. Infection 1998, 26, 30–35. 22. Fedoronko, M.; Koenigstein, J. Collect. Czech. Chem. Commun. 1969, 34, 3881– 3894. 23. Codex Alimentarius Commission: Revised Codex Standard for honey, Codex STAN 12-1981, Rev.1 (1987), Rev.2 (2001). 24. Stephens, J. C.; Schlothauer, R. C. WO 2010082845 A1, 2010 Jul 22. 25. Mocak, J.; Bond, A. M.; Mitchell, S.; Scollary, G. Pure Appl. Chem. 1997, 69, 297– 328.