Subscriber access provided by University of Newcastle, Australia

Review
A Review of the Analytical Methods used in the Quality Control of Honey
Consuelo Pita-Calvo, María Esther Guerra-Rodriguez, and Manuel Vazquez
J. Agric. Food Chem., Just Accepted Manuscript • Publication Date (Web): 04 Jan 2017
Downloaded from http://pubs.acs.org on January 4, 2017

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 57 Journal of Agricultural and Food Chemistry

1 A Review of the Analytical Methods used in the Quality Control of Honey

2 Consuelo Pita-Calvo, María Esther Guerra-Rodríguez, Manuel Vázquez*

4 Department of Analytical Chemistry, Faculty of Veterinary Science, University of

5 Santiago de Compostela, 27002-Lugo, Spain

7 * Corresponding author. Email: [email protected]

9 ABSTRACT

10 Honey is a natural sweet substance produced by bees (Apis mellifera). In this

11 work, the main parameters used in routine quality control of honey and the most

12 commonly used analytical methods for their determination were reviewed. Honey can

13 be adulterated with cheaper sweeteners or in an indirect way feeding the bees with

14 sugars. Therefore, methods for detecting and quantifying adulteration are necessary.

15 Chromatographic techniques are widely used in honey analysis. More recently,

16 techniques such as Raman, near-infrared, mid-infrared, and nuclear magnetic resonance

17 spectroscopy in combination with chemometric data processing have been proposed.

18 However, spectroscopy does not allow the determination of enzyme activities, one

19 criteria of great importance for the honey trade. Methylglyoxal is an interesting

20 compound for its antibacterial properties. Methods for its determination were also

21 reviewed.

22

23 Keywords: honey; quality control; adulteration; analytical methods;

24 hydroxymethylfurfural; methylglyoxal.

25

26

ACS Paragon Plus Environment 1

 Journal of Agricultural and Food Chemistry Page 2 of 57

27 INTRODUCTION

28 Honey is probably one of the most complex natural foods. It is the only sweetener that

29 does not require processing for its human consumption 1. Bees (Apis mellifera) make

30 honey using nectar of plants (blossom honey), secretions of plants, or excretions of

31 plant-sucking insects on plants (honeydew honey). Bees collect and transform the nectar

32 or honeydew and deposit, dehydrate, and store in the hives to allow the honey to ripen

33 and mature (EU Directive 110/2001) 2. Honey can be adulterated with cheaper

34 sweeteners or in an indirect way feeding the bees with sugars. Therefore, methods for

35 detecting and quantifying adulteration are necessary.

36 Blossom honey can be classified as 2 types: a) unifloral or monofloral honey and

37 b) multifloral honey. Unifloral honey has a higher market value due to limited

38 production and availability. Due to the existence of protected designation of origin

39 (PDO) and protected geographical indication (PGI), honey classification can also has

40 economic importance 3.

41 The carbohydrates of honey are formed by the action of several enzymes on

42 nectar sugars. Therefore, honey is a complex mixture mainly composed of

43 carbohydrates (70-80 % w/w), water (10-20 % w/w), and a large number of minor

44 components 4. Carbohydrates comprise about 95 % w/w by dry weight of honey. The

45 major carbohydrates are glucose and fructose, 65-80 % (w/w) of the total soluble solids,

46 and disaccharides of glucose and fructose having the glycosidic bond in different

47 positions and configurations 5.

48 Using gas chromatography (GC), 16 oligosaccharides in honey have been found,

49 including 11 disaccharides (turanose, sucrose, maltose, isomaltose, kojibiose, cellobiose

50 palatinose, gentiobiose, laminaribiose, neotrehalose, and nigerose) and 5 trisaccharides

51 (erlose, panose, isopanose, maltotriose, and theanderose) 6. Other trisaccharides include

ACS Paragon Plus Environment 2

Page 3 of 57 Journal of Agricultural and Food Chemistry

52 melezitose, isomaltotriose, raffinose, kestoses, and isomelezitose, which have been

53 identified by chromatographic methods. Tetrasaccharides in honey include

54 isomaltotetraose, maltotetraose, stachyose, nystose, fructosyl-isomelezitose, α-4´-

55 glucosyl-erlose, and α-6´-glucosyl-erlose 7.

56 The minor components in honey include, among others, organic acids, proteins,

57 enzymes, and hormones. Organic acids are responsible for the taste of honey. The main

58 proteins are globulin and albumin which come from pharyngeal glands of bees. The

59 enzymes come from the glands of bees being the most important diastase (amylase),

60 invertase, and glucose oxidase. The origins of the hormones (such as acetylcholine and

61 its precursor, choline) are plants or bees themselves. Other minor compounds are

62 flavonoids, phenolic acids, anthocyanins, vitamins, essential oils, pigments, sterols,

63 phospholipids, and minerals. The mineral substances include those occurring in the

64 largest quantities such as Ca, Mg, K, and P, and those in minor amounts such as Fe, Cu,

65 Zn, Co, Si, S, Mn, F, Mo, Cr, and I.

66 Honey has a high-energy value and many manufactured foods are elaborated

67 with honey as an ingredient. Furthermore, its oligosaccharides seem to present potential

68 prebiotic activity (prebiotic index: 3.38-4.24) in that they increase the populations of

69 Bifidobacteria and Lactobacilli 8. Sugars are also responsible for such properties as

70 viscosity, hygroscopy, and granulation 9. Honey has antibacterial and anti-inflammatory

71 properties that help healing of wounds and burns and treatment of gastric ulcer.

72 Additionally, honey has significant antioxidant activity 10.

73 Therefore, quality control is necessary to verify the presence of health-beneficial

74 properties of the honey. In this work, the main parameters used in the routine quality

75 control of honey and the analytical methods for its determination are reviewed.

76 Analytical methods to determine carbohydrate composition and detect and quantify

ACS Paragon Plus Environment 3

 Journal of Agricultural and Food Chemistry Page 4 of 57

77 adulteration in honey were reviewed in depth as well as those for methylglyoxal

78 (MGO), an interesting compound known for its antibacterial properties.

79

80 ANALYTICAL METHODS FOR QUALITY PARAMETERS

81 Honey quality is evaluated using its more significant physical, chemical, and

82 biochemical parameters. Such parameters have been classified according to their

83 contribution to honey quality (Table 1). Composition criteria of honey in accordance

84 with the European Legislation cited above are shown in Table 2. The International

85 Honey Commission (IHC, http://www.ihc-platform.net) has recommended methods to

86 use in the routine quality control of honey. In the following sections, the contribution of

87 each parameter to honey quality as well as recent advances in spectroscopic methods for

88 its determination is presented.

89

90 Chemical parameters. Moisture is related to the honey capability to remain

91 stable and to resist spoilage by yeast fermentation. The moisture content of honey

92 generally depends on its botanical origin, climatic conditions, and degree of maturity of

93 the honey as well as processing techniques and storage conditions 11,12. Moisture has an

94 influence largely on sensory characteristics of the product, especially its flavor and

95 granulation 13.

96 Moisture, together with the main sugars and other physicochemical parameters

97 of honey, has successfully been determined by near-infrared (NIR) and mid-infrared

98 (MIR) spectroscopy. The advantage of using infrared (IR) spectroscopy is that the

99 determination of several analytes in samples can be realized simultaneously. For light-

100 scattering samples such as honey, NIR transflectance spectroscopy is considered a better

101 analytical technique than NIR transmittance spectroscopy. This is because in the

ACS Paragon Plus Environment 4

Page 5 of 57 Journal of Agricultural and Food Chemistry

102 transflectance mode backscattered radiation is also collected and measured. Therefore, a

103 more reliable value of absorbance is obtained 14.

104 The moisture measurements in the NIR region were performed in the

105 transmittance 15, transflectance 16, or reflectance mode 17. Good prediction accuracy was

106 also obtained by Fourier transform near-infrared spectroscopy (FT-NIR) in the

107 transflection mode 18. Coefficients of determination (r2) of the models obtained by NIR

108 (transmittance or transflectance) and FT-NIR (transflection) spectroscopy ranged

109 between 0.96 and 1. Prediction standard errors (SEP) and standard errors of cross-

110 validation (SECV) were 0.16-0.3 % (w/w) and 0.08-0.3 % (w/w), respectively 15,16,18. A

111 study using NIR in reflectance mode reported a SECV of 3.1 % (w/w) and a value of r2

112 in cross-validation of 0.94 17.

113 In the MIR region, Fourier transform infrared spectroscopy with attenuated total

114 reflectance (FTIR-ATR) was also evaluated for determining moisture in honey. Good

115 prediction accuracy was obtained, similar to that obtained by NIR, with values of r2,

116 SEP, and SECV of 0.989, 0.24 % (w/w), and 0.46 % (w/w), respectively 19.

117 The pH is related to the stability and the shelf-life of honey and it can be used to

118 indicate microbial contamination 11. IR spectroscopy has been evaluated to determine

119 pH and acid parameters in honey. The prediction accuracy obtained by NIR
15,18
120 spectroscopy was poor and unreliable . However, pH and free acidity were

121 determined with acceptable accuracy by FTIR-ATR spectroscopy 19. For pH, the r2 and

122 SEP values were 0.868 and 0.16, respectively. For free acidity, r2 and SEP were 0.958

123 and 2 meq/kg, respectively.

124 The glucose/moisture ratio is an indicator used for predicting honey

125 crystallization. When glucose/moisture is 1.7 or lower, honey will not crystallize. The

126 fructose/glucose ratio has also been recommended and, when this ratio is higher than

ACS Paragon Plus Environment 5

 Journal of Agricultural and Food Chemistry Page 6 of 57

18,20
127 1.3, honey remains liquid or crystallize slowly . Fructose/glucose and
21
128 glucose/moisture ratios are useful for the classification of unifloral honeys . The

129 fructose/glucose and glucose/moisture ratios can be determined by FT-NIR and FTIR-

130 ATR spectroscopy 18,19. Better prediction accuracy was obtained using the latter. For the

131 fructose/glucose ratio, the r2 and SEP values obtained were 0.975 and 0.03, respectively,

132 and for the glucose/moisture ratio, the r2 and SEP values were 0.942 and 0.06,

133 respectively. Using FT-NIR spectroscopy, for the fructose/glucose ratio, r2 and SEP

134 were 0.820 and 0.09, respectively. For the glucose/moisture ratio, r2 and SEP were

135 0.849 and 0.12, respectively.

136 Ash content is an indicator of the mineral content. It is considered as a quality

137 criterion for honey origin. The blossom honeys have lower ash contents than the

138 honeydew ones (IHC). Generally, the blossom honeys have an ash content ≤ 0.6 %

139 (w/w) while the honeydew honeys, blends of honeydew and blossom honeys, or chesnut

140 honey ≤ 1.2 % (w/w) 4.

141 Proline is the main amino acid of honey and it is associated with its antioxidant

142 properties 22. The amount of proline is used as a criterion of honey ripeness and it can

143 also be used to detect adulteration with sugar. The level of proline decreases

144 significantly by adulterating the honey 23. European legislation does not set a minimum

145 value for the proline concentration in honey. However, honey with proline at less than

146 180 mg/kg is considered to be either non-ripe or adulterated 24.

147 The analytical methods proposed by the IHC and the Association of Official

148 Analytical Chemists (AOAC, http://www.aoac.org) are derived from the original Ough
25
149 photometric method . The IHC method introduces some significant changes that

150 lengthen the time of analysis. Concerning the AOAC method, the absorbance of a honey

151 sample without reagents is subtracted of that of the sample of honey. A comparison

ACS Paragon Plus Environment 6

Page 7 of 57 Journal of Agricultural and Food Chemistry

152 between the 3 methods was recently performed by analyzing honey samples of several

153 botanical origins. No statistically significant differences were found between the Ough

154 and IHC methods for all honey types investigated. Therefore, the extra treatment stages

155 proposed by the IHC method are pointless. On the other hand, the AOAC method

156 provided greater accuracy and time-saving. Quality parameters of the AOAC method

157 were studied. No statistically significant differences between the slopes of addition and

158 calibration lines were found and linearity range was confirmed up to 1800 mg/L. By

159 contrast, the matrix effect was important for the limits of detection (LOD) and

160 quantification (LOQ). The LOQs values were 20 and 61 mg/L using extern calibration

161 and standard additions method, respectively. Analytical recoveries ranged from 92 to

162 111 %, and precision was better than that reported by the IHC 26. The spectroscopic

163 techniques, FT-NIR and FTIR-ATR, were evaluated for determining proline. However,

164 prediction accuracy was poor and unreliable due to the low concentrations present in

165 honey 18,19.

166 Sucrose is an important sugar from a regulatory point of view. The analysis of

167 sucrose content has been mainly used to control adulteration with commercial sucrose.
27
168 Genuine honeys contain only about 5 % sucrose . A high amount of this sugar is

169 usually due to an early harvest of honey. Therefore, sucrose is not completely

170 transformed into glucose and fructose by the enzyme invertase 28. High levels of sucrose
29
171 may also be due to inadequate maturation of the honey or to the fact that the
30
172 beekeeper over-fed the bees with sugar during spring . On the other hand, during

173 honey storage the sucrose level can decrease by action of the invertase 30.

174 The measurement of water-insoluble content is used to detect an amount of

175 impurities higher than the permitted maximum value in honeys. Hydroxymethylfurfural

176 (HMF) is widely used as an indicator of honey quality deterioration produced by

ACS Paragon Plus Environment 7

 Journal of Agricultural and Food Chemistry Page 8 of 57

177 excessive heating or inadequate storage conditions. The HMF content in fresh honeys is

178 low. However a high concentration of this compound is present in old honeys, honey

179 that has been heated, stored in non-adequate conditions, or adulterated with invert sugar

180 or syrup 31. Extremely high amounts of HMF (>500 mg/kg) due to adulteration with

181 invert syrup have been reported 32.

182 The interest in the HMF determination in food is also related to its toxicity.

183 Ingestion, inhalation, or skin absorption are the three ways of exposure to HMF 33. HMF

184 and its derivatives (5-chloromethyl and 5-sulfidemethylfurfural) have been reported to
34-41
185 have genotoxic, cytotoxic, mutagenic, and carcinogenic effects . However, other

186 studies have suggested that HMF probably does not pose a serious risk to human health
40
187 .

188 Besides HMF, other furanic aldehydes such as 2-furaldehyde (furfural or 2-F)

189 and furan-3-carboxaldehyde (3-furaldehyde or 3-F) can be present in honey, although in

190 smaller quantities. Furanic acids such as furan-2-carboxylic acid (2-furoic acid or 2-FA)

191 and furan-3-carboxylic acid (3-furoic acid or 3-FA) have also been found in honey.

192 Analytical techniques used to determine HMF and other furanic aldehydes and

193 acids in honey are shown in Table 3. The analytical technique regularly used to

194 determine HMF is high-performance liquid chromatography (HPLC). This technique

195 avoids interferences and allows the simultaneous analysis of compounds. The

196 disadvantage of the GC is that it requires a previous derivatization of HMF. The IHC

197 proposed a HPLC method 42 for determining HMF. Honey is simply dissolved in water

198 and HMF is determined by reverse phase HPLC (RP-HPLC) with ultraviolet (UV)

199 detector (285 nm). Elution in isocratic mode with 90 % (v/v) methanol as mobile phase

200 is used. The IHC also proposed 2 spectrophotometric methods known as White method
43
201 and Winkler method 44. Differences between the 3 methods were only found at very

ACS Paragon Plus Environment 8

Page 9 of 57 Journal of Agricultural and Food Chemistry

202 low levels of concentration (about 5 mg/kg) without interest for assessing honey quality

203 (IHC).

204 The White and HPLC methods have recently been compared by several

205 researchers for its application to samples of honey with a very low HMF content (<4

206 mg/kg) 45. Quality parameters of both methods were determined. For the HPLC method,

207 the LOD and LOQ were affected by matrix effect. Using the standard additions method,

208 the LOD and LOQ were 0.27 and 0.83 mg/L, respectively. For the White method, the

209 values of LOD and LOQ were similar, 0.22 and 0.67 mg/L, respectively. 43 honey

210 samples were analyzed with both methods. For samples with HMF content between 1

211 and 4 mg/kg, no statistically significant differences were observed. However, the

212 precision was higher with the HPLC method (3.5 %) than with the White method (6.4

213 %). Therefore, the HPLC method would be more adequate for this range of

214 concentrations. For samples with a HMF content much lower (<1 mg/kg), the results

215 were inaccurate with either method.

216 Homogentistic acid (HA) is a phenolic acid (2,5-dihydroxyphenylacetic acid)

217 that has been identified in strawberry-tree (Arbutus unedo) honey by mass spectrometry

218 (MS) and nuclear magnetic resonance (NMR). Homogentistic acid has been suggested

219 as a marker for this variety of honey 46. In strawberry-tree honey, interference of HA in

220 the determination of HMF by the HPLC method proposed by the IHC has been

221 reported. HA completely prevented the quantification of HMF due to overlapping of

222 their corresponding chromatographic peaks. In order to prevent this interference, a new

223 RP-HPLC method using gradient elution (solvent A: 0.01 mol/L H2SO4, solvent B:

224 methanol) and detection at 291 nm was proposed. The separation was completed in 15

225 min, a time 20 % less than that of the IHC method. The LOD and LOQ of the method

226 proposed were 0.2 mg/L (1.9 mg/kg) and 0.4 mg/L (4.0 mg/kg), respectively. For HMF

ACS Paragon Plus Environment 9

 Journal of Agricultural and Food Chemistry Page 10 of 57

227 levels between 5 and 140 mg/kg honey, the repeatability and reproducibility were better

228 than about 2 and 3 %, respectively. Therefore, this method is much more precise than

229 the HPLC method proposed by the IHC for this range of concentrations. The method

230 presented also good accuracy and a high linearity interval (2-800 mg/kg) 47.

231 Some methods based on IR spectroscopy have been studied to determine HMF.

232 Poor prediction accuracies were obtained because the low concentrations of HMF in

233 honeys 15,17-19. A method by RP-HPLC for the simultaneous determination of HMF, 2-

234 furaldehyde, 3-furaldehyde, 2-furoic acid, 3-furoic acid, and methyl anthranilate (MA)
31
235 in honey has been reported . MA is the 2-aminobenzoic acid methyl ester. This

236 compound has been proposed as a marker for citrus honey. In order to prevent

237 overlapping interferences and pre-concentrate the analytes, a previous solid-phase

238 extraction on polymeric cartridges was used. The compounds were separated by

239 gradient elution (solvent A: 1 % (v/v) acetic acid–acetonitrile (97:3), solvent B: 50 %

240 (v/v) acetonitrile). Detection at 250 nm was used. Precision, accuracy, LOD, LOQ, and

241 linearity range for each component were reported. For HMF, the LOD and LOQ were

242 0.04 and 0.13 mg/L, respectively, and the linearity range was confirmed up to 50 mg/L.

243 Precision (RSD %) was 2.47 %. This method has significant advantages over the
47
244 method developed for HMF by Spano . It is more sensitive, allowing also the

245 simultaneous determination of other furanic aldehydes and acids as well as MA with

246 great sensitivity (LOD: 0.01-0.08 mg/L). On the other hand, the precision of both

247 methods is similar.

248 Another method to determine all these compounds, except MA, in honey has

249 been developed 48 as a modified method of that for HMF described by Spano 47. Elution

250 in gradient mode with 0.1 mol/L H2SO4 (solvent A) and methanol (solvent B) was used.

251 For HMF, the LOD and LOQ values reported were 0.003 and 0.010 mg/L. These were

ACS Paragon Plus Environment 10

Page 11 of 57 Journal of Agricultural and Food Chemistry

252 lower than those obtained by Nozal 31. A wider linear range of concentrations was also

253 obtained (0.01-100 mg/L). The precision and accuracy of the method were very good.

254 The repeatability and reproducibility values were 0.99 and 1.47 %, respectively.

255 Average analytical recovery was 97 %. For the other compounds, low LODs and LOQs,

256 wide linear intervals of concentration, and good accuracy and precision were also

257 obtained.

258

259 Biochemical parameters: Enzyme activity. Diastase and invertase activities

260 are also used as a marker for the freshness of honeys because they decrease in old or

261 heated honeys.

262 The IHC recommends 2 methods to determine honey diastase activity: Schade

263 method and Phadebas method. The Schade method (traditional) is based on the original
49 50 51
264 work of Schade , modified by White Jr and Pairent and Zurcher and Hadorn .

265 Diastase activity in Schade units is directly determined using starch as a substrate. The

266 Phadebas method measures the absorbance of a honey solution after the enzymatic

267 hydrolysis of an artificial substrate (cross-linked starch), which is directly proportional

268 to the diastase activity 52. A method similar to the Phadebas method is the Amylazyme

269 method which employs cross-linked amylose as a substrate (the linear fraction of starch)
53
270 . The Phadebas and Amylazyme methods use substrates commercially available in

271 tablet format (Phadebas test and Amylazyme tablets) and they are easier to apply than

272 the Schade method. The IHC has also recommended a photometric method for the
54
273 determination of invertase activity .

274 Recently, a direct potentiometric method to determine diastase activity was

275 developed. The method involves the formation of a starch-triiodide complex with

276 characteristic colors. The starch is hydrolyzed by the enzyme producing the release of

ACS Paragon Plus Environment 11

 Journal of Agricultural and Food Chemistry Page 12 of 57

277 triiodide iones which are measured using a platinum redox electrode. This method

278 presented a good correlation with both standard methods. It is much simpler and faster

279 and requires low-cost instrumentation. Furthermore, it did not require additional

280 dilutions and volume adjustments. Therefore, the errors associated with this procedure

281 are lower than those of the Schade method 55.

282 Physical parameters. Electric conductivity (EC) is a good criterion for

283 differentiate the botanical origin of honey (nectar or honeydew). The EC values of

284 blossom honeys are lower than those of honeydew honeys (Table 2). Blossom honey

285 and blends of these honeys have EC values ≤ 0.8 mS/cm. However, there are exceptions

286 to this rule which include eucalyptus, strawberry tree (Arbutus unedo), bell heather

287 (Erica), ling heather (Calluna vulgaris), lime (Tilia spp.), tea tree (Melaleuca spp.), and

288 manuka or jelly bush (leptospermum) honeys. Honeydew honey and chestnut honey,

289 and blends with blossom honey (except with those indicated above), have values ≥0.8

290 mS/cm. The EC values have also been used for the classification of unifloral honeys 21.

291 IR spectroscopy has been evaluated to determine the EC of honey. The

292 prediction accuracy obtained by FTIR-ATR spectroscopy was good (r2: 0.979, SEP:

293 0.05 mS/cm) 19. However, it was lower using NIR reflectance spectroscopy (r2: 0.88,

294 SECV: 0.01 mS/cm) 17 and FT-NIR spectroscopy in transflection mode (r2: 0.870, SEP:

295 0.14 mS/cm) 18.

296 Color is one of the most variable features of honey and depends mainly on the

297 botanical origin. Other factors that affect the color of the honey are its ash content,

298 temperature, and length of storage, as well as the presence of antioxidant pigments such
56
299 as flavonoids and carotenoids . Generally, honeydew honeys are darker than the

300 blossom honeys. Some blossom honeys such as chestnut and heather have also dark

301 color.

ACS Paragon Plus Environment 12

Page 13 of 57 Journal of Agricultural and Food Chemistry

302 The most used methods to determine the color of honey are based on optical
57 21
303 comparison using simple color grading after Pfund or Lovibond . Color

304 characteristics can be assessed by the CIE L* a* b* tristimulus method. The L* refers to

305 lightness and a* and b* refer to redness-greenness and yellowness-blueness,

306 respectively. The lighter honeys holds an L* value > 50 whereas dark honeys showed a
58
307 value < 50 . Color grading (mm Pfund scale) has also been determined by NIR

308 reflectance spectroscopy (r2: 0.97, SECV: 4.7 mm Pfund) 17.

309 Honey holds the property of rotating the plane of polarized light. Specific

310 rotation depends mainly on types and relative amounts of their sugars. The method
21,59
311 recommended by the IHC for its measure is based on published methods .

312 Polarimetric properties of honey (direct polarization, polarization after inversion,

313 polarization due to non-monosaccharides, and specific rotation in dry matter) have been

314 successfully calibrated using NIR spectroscopy 60.

315 The determination of the specific rotation can be used for the differentiation

316 between honeydew honeys and blossom honeys. Most of the honeydew honeys are

317 dextrorotatory, whereas nectar honeys are levorotatory. Furthermore, the specific

318 rotation can also be useful for the classification of unifloral blossom honeys 21.

319

320 ANALYTICAL METHODS FOR CARBOHYDRATE COMPOSITION OF

321 HONEY

322

323 Honey is composed primarily of carbohydrates. Therefore, this section is

324 dedicated to the review of analytical methods for determining carbohydrates in honey.

325 Furthermore, some of the quality parameters of honey are related to glucose, fructose,

ACS Paragon Plus Environment 13

 Journal of Agricultural and Food Chemistry Page 14 of 57

326 and sucrose contents. Analytical methods to detect and quantify adulteration of honey

327 with commercial sweeteners are also reviewed.

328

329 Determination of carbohydrates in honey. Instrumental analytical methods for

330 determining sugars in honey are usually based on chromatographic and spectroscopic

331 techniques as shown in Table 4.

332 Several chromatographic techniques have been used: HPLC with refractive

333 index detector (HPLC-RID), evaporative light scattering detection (HPLC-ELSD) or

334 diode array detector (HPLC-DAD), anion-exchange chromatography with pulsed

335 amperometric detection (HPAEC-PAD), gas chromatography with flame ionization

336 detector (GC-FID), and gas chromatography–mass spectrometry (GC-MS).

337 The main problems for oligosaccharides determination arises from their low

338 concentrations in honeys as well the lack of commercial standards. Furthermore, the

339 presence of structurally similar carbohydrates makes their chromatographic resolution

340 very difficult. The large contents of monosaccharides and the relatively high amounts of

341 disaccharides in honey also make it difficult to separate oligosaccharides with an

342 acceptable resolution. Therefore, they should be separated by a chromatographic run or

343 a previous extraction step.

344 HPAEC-PAD is a valuable technique for oligosaccharides determination. Sugars

345 are very weak acids (pKa: 12-13) being partially or totally ionized at high pH (12-14).

346 Therefore, they can be separated by anion-exchange chromatography. Oligosaccharides

347 profiles obtained with this technique have been used to detect adulteration of honey

348 with sugar syrups. Before the chromatographic separation of oligosaccharides, honey

349 samples were pre-treated with activated charcoal which removes mono- and

350 disaccharides and retains the fraction of oligosaccharides 61. HPAEC-PAD has good

ACS Paragon Plus Environment 14

Page 15 of 57 Journal of Agricultural and Food Chemistry

62
351 sensitivity, comparable to that of GC-FID . Furthermore, PAD proved to be more

352 selective and sensitive than the RI and ELS detections offering the possibility of using
63
353 gradient elution . However, this detection system is not as usual in analytical

354 laboratories as RI and UV detectors. The UV detector is more sensitive that the RI

355 detector but less than PAD.

356 HPLC-DAD has the disadvantage of requiring derivatization of carbohydrates

357 time-consuming. A method using this technique has been applied to characterize the di-

358 and tri- saccharides profiles of 5 samples of honey from different botanical origins

359 collected in different regions of Province of Buenos Aires (Argentina). The

360 carbohydrates were isolated from monosaccharides-rich matrix using a solid-phase

361 extraction procedure with porous graphitic carbon cartridges, the extraction being

362 quantitative. Then, they were derivatized with the chromophoric reagent 4-(3-methyl-5-

363 oxo-2-pyrazolin-1-yl) benzoic acid at 70 ºC for 60 min. Subsequently, the sugars were

364 separated and eluted at 40 min by RP-HPLC and detected at 271 nm. LODs (16-24

365 pmol) were calculated for 4 sugars and correspond to a 0.005 % (w/w) honey sample.

366 This apparent increase in sensitivity was due to the pre-concentration during the solid

367 phase extraction step 62.

368 Better resolution was provided using GC-MS that HPAEC-PAD for honey
27
369 oligosaccharides analysis . The main limitation of the GC if that a previous

370 derivatization is needed in order to obtain volatile carbohydrate compounds, a step that

371 may be long, difficult, and time consuming. Furthermore, very complex chromatograms

372 with a high degree of overlap can be obtained due to the large number of carbohydrate

373 isomers produced during the derivatization reaction. The trimethylsilyl oximes obtained

374 from reducing sugars produced only two chromatographic peaks and those from

375 nonreducing sugars one 64-66.

ACS Paragon Plus Environment 15

 Journal of Agricultural and Food Chemistry Page 16 of 57

376 The IHC has recommended several chromatographic methods to determine

377 sugars in honey. One of the methods involves HPLC-RID 67. It uses a chromatographic

378 column of amine-modified silica gel and 80 % (v/v) acetonitrile as the mobile phase.

379 Column oven and detector must be at 30 ºC. Another method uses HPAEC-PAD.

380 Honey samples are simply diluted with ultra-pure water, being sugars separated with an

381 anionic exchange column and eluted with a sodium hydroxide solution.

382 Two methods using GC-FID have been recommended: Pierce-Portallier method
68
383 and I.N.A. method 69. In the former, any sugars containing free aldehyde or ketone

384 groups, such as glucose and fructose, are derivatized to their corresponding oximes and

385 then silylated. In the second method, sugars are directly silylated. The silylated

386 derivatives obtained are separated and quantified by GC. Mannitol is used as the

387 internal standard.

388 Sugar composition in honey has been widely studied using chromatography. For

389 instance, amounts of oligosaccharides in 70 samples of Brazilian floral honeys from

70
390 different geographical regions have been determined using HPLC-RID . The

391 oligosaccharides identified included disaccharides (sucrose, maltose, nigerose, turanose,

392 isomaltose, and melibiose) and trisaccharides (panose, maltotriose, raffinose, and

393 melezitose). The major disaccharide found in all samples was maltose. Its mean content

394 ranged from 1.58 to 3.77 % (w/w) with an average value of 3.05 % (w/w). The lowest

395 amounts of oligosaccharides corresponded to melibiose (0.05-0.15 % (w/w) and panose

396 (0.03-0.08 % (w/w). The presence of melezitose in the floral honeys was probably due

397 to contamination with honeydew while the origin of raffinose was not clear. Raffinose
71
398 could arise from nectar or honeydew contamination .

399 Sugar composition in Algerian floral honeys from different botanical origins has

400 been also studied 72. Glucose, fructose, and 9 oligosaccharides were quantified in 50

ACS Paragon Plus Environment 16

Page 17 of 57 Journal of Agricultural and Food Chemistry

401 samples using HPAEC-PAD. The separation was completed in 40 min. using gradient

402 elution (solvent A: ultra-pure water; solvent B: 0.2 M NaOH). Maltose and sucrose

403 were the major disaccharides found. The mean values ranged from 0.47-3.42 % (w/w)

404 and from 0.48-5.26 % (w/w) for maltose and sucrose, respectively. Maltose was also the

405 main disaccharide in Brazilian honeys. Melibiose was not detected in all samples.

406 HPAEC-PAD has also been used to determine sugars in 63 samples of

407 Portuguese floral honeys 73. A good separation was obtained using isocratic elution with

408 a 50 mM NaOH solution for 25 min. LOQs of the analytical method were determined:

409 0.7, 1.2, 1.7, 0.8, 1.3, 1.2, and 0.9 % (w/w) for trehalose, glucose, fructose, sucrose,

410 melezitose, turanose, and maltose, respectively. Repeatability of the method was also

411 evaluated: 0.25, 0.40, 0.16, 0.13, 0.25, and 0.28 % (w/w) for glucose, fructose, sucrose,

412 melezitose, turanose, and maltose, respectively. The separation took place in less than

413 20 min. Maltose and turanose were the most important disaccharides present in the

414 samples in term of quantity. Values up to 2 % (w/w) (mean value 1.47 %) and up to 3 %

415 (w/w) (mean value 2.65 %) were found for maltose and turanose, respectively.

416 Trehalose was the minor disaccharide with amounts below LOQ for most of the

417 samples. Low amounts of sucrose were found; below LOQ for approximately 50 % of

418 the samples. Values of the melezitose up to 6.26 % (w/w) and a mean value of 2.19 %

419 (w/w) were found. These values were higher than those found by others 70,72.

420 Amounts of sugars in Spanish floral honeys by GC-MS have also been reported
74
421 . 109 samples of honey were analyzed. Sugars were derivatized to their corresponding

422 oximes with 2.5 % hydroxylamine chloride in pyridine at 75 ºC for 30 min. Then the

423 samples were persilylated (45 ºC for 30 min) using hexamethyldisilazane and

424 trifluoroacetic acid and subsequent centrifuged. The chromatographic method used was

425 that described by Sanz 75. Two capillary columns containing different stationary phases

ACS Paragon Plus Environment 17

 Journal of Agricultural and Food Chemistry Page 18 of 57

426 allowed the analysis of fructose and glucose as well as a high number of disaccharides

427 and trisaccharides. The separation was completed in less than 1 h. The mean amounts

428 obtained for fructose and glucose were 37.1 and 30.0 % (w/w), respectively. The main

429 disaccharides were turanose (2.16% w/w), maltulose (1.92 % w/w), isomaltose (1.35 %

430 w/w), maltose (1.54 % w/w), kojibiose (1.42 % w/w), and trehalulose (1.14 % w/w)

431 which were present in all samples and with a very high variability. The sucrose content

432 was very low in most of the samples. α,β-Trehalose (0.23-0.79 % w/w) was present in

433 all examined samples whereas α,α-trehalose only in 5. The total concentration of

434 trisaccharides was also highly variable in the samples (0.37-5.31 % w/w). Raffinose co-

435 eluted with 1-kestose. Erlose (0.01-3.0 % w/w), raffinose+1-kestose (0.02–0.81 %

436 w/w), and melezitose (0-0.81 % w/w) were the most abundant trisaccharides.

437 The chromatographic methods present some disadvantages. They are expensive

438 and time-consuming. The industry needs fast, accurate, easy-to-use, and low-cost

439 analytical methods. Therefore, methods based on IR and Raman spectroscopy have been

440 developed to analyze honey. They are rapid and inexpensive, not needing sample

441 preparation or expensive reagents. In addition, in comparison to some chromatographic

442 methods, the spectroscopy has the advantage of being non-destructive.

443 IR spectroscopy is a technique of great importance in the analysis of foods,

444 especially when it is used in combination with multivariate analysis. IR spectroscopy

445 has been applied to honey analysis: quality control, detection and quantification of

446 adulteration, and authentication of botanical or geographical origin of honey samples.

447 MIR spectroscopy presents advantages over NIR spectroscopy. MIR

448 spectroscopy is 10-100 times more sensitive than NIR spectroscopy and spectra in the

449 MIR region (2500–25000 nm) contain more information than those obtained in the NIR

450 region (750-2500 nm). The absorption of the major sugar in honey takes place in the

ACS Paragon Plus Environment 18

Page 19 of 57 Journal of Agricultural and Food Chemistry

451 region between 1500 and 750 cm-1. Another advantage of MIR spectroscopy is its better

452 resolution. Traditionally, sample handling was a problem in MIR spectroscopy.

453 Nowadays attenuated total reflectance (ATR) crystals have simplified the handling 76.

454 NIR spectroscopy allowed the determination of major compounds in honey.

455 Some minor constituents could also be determined. Fructose, glucose, sucrose, maltose,

456 and moisture have been determined in commercial honey samples by NIR spectroscopy
15
457 in transmittance mode . The r2 values between the predicted and reference

458 concentrations were 1.0, 0.97, 0.91, 0.86, and 0.93 for moisture, fructose, glucose,

459 sucrose, and maltose, respectively. SEP values were 0.16, 0.42, 0.34, 0.34, and 0.27 %

460 (w/w) for moisture, fructose, glucose, sucrose, and maltose, respectively. Sharper peaks

461 and better resolution were obtained using transmittance spectra rather than reflectance

462 spectra. Furthermore, the performance of the calibration was 30–70 % better. The

463 optical path length of the cuvette affected significantly the transmittance spectra of

464 samples. An optical path of 1 mm produced the least saturated spectra in the range

465 1300-2500 nm. Therefore, lower SECV values were obtained for all compounds

466 analyzed. NIR transflectance spectroscopy has also been used for the determination of
16
467 fructose and glucose . The r2 values obtained were 0.98 and 0.95 for fructose and

468 glucose, respectively. The SEP values were 0.35 and 0.70 % (w/w) for fructose and

469 glucose, respectively. The results obtained were similar to those obtained by Qiu 15.

470 FT-NIR spectroscopy in transflection mode has also been assessed for its

471 application to the simultaneous determination of sugars and main quality

472 physicochemical parameters in honeys. 421 samples of honey from different botanical

473 origins and countries were used in order to develop the calibration model. The most of

474 the samples were collected in Switzerland and Germany 18. The prediction accuracies

475 obtained were quite acceptable for fructose, glucose, sucrose, and total monosaccharide

ACS Paragon Plus Environment 19

 Journal of Agricultural and Food Chemistry Page 20 of 57

476 content (sum of fructose and glucose), as well as for electric conductivity and the

477 fructose/glucose and glucose/moisture ratios. For moisture, the precision accuracy was

478 very good. The r2 values of fructose and glucose were 0.810 and 0.884, respectively.

479 The values of SEP obtained were 1.6 and 1.3 % (w/w) for fructose and glucose,

480 respectively. The prediction accuracies of fructose and glucose were somewhat lower

481 than those obtained by other authors 15,16. The prediction accuracy of sucrose (SEP: 0.36

482 % (w/w); r2: 0.725) was lower than that obtained by Qiu 15 allowing a rough estimation

483 of the sucrose amount. However, the prediction accuracy was poor and unreliable for

484 minor sugars such as maltose, turanose, nigerose, erlose, trehalose, isomaltose,

485 kojibiose, melezitose, raffinose, gentiobiose, melibiose, and maltotriose. The authors

486 considered that the low prediction accuracies of minor sugars obtained are due to the

487 low contents of these compounds in honey, their insufficient separation by the HPLC

488 method used (IHC), and the non-specific absorption bands in NIR. Concerning maltose,
15
489 it has been determined with good accuracy of prediction by Qiu using NIR

490 spectroscopy although a calibration with fewer samples was used. The determinations

491 of HMF, proline, pH, and free acidity gave also poor prediction accuracy. NIR

492 spectroscopy has also been applied to the determination of low concentrations of

493 perseitol, a sugar that is specific of avocado honey 77.

494 FTIR spectroscopy with micro-attenuated total reflectance (FTIR-mATR) has

495 been applied to the determination of fructose, glucose, sucrose, and maltose in honeys

496 belonging to various floral and geographical origins around the world. Two calibration

497 models were developed; one used 42 standard mixtures (r2: 0.97-0.99) and another 45

498 samples of honey from different floral and geographical origins around the world (r2:

499 0.94-0.97). Both models showed a good correlation. However, a calibration model using

500 spectra of honey samples is preferred because it uses samples containing the same

ACS Paragon Plus Environment 20

Page 21 of 57 Journal of Agricultural and Food Chemistry

501 matrix as the real samples improving predictability. On the other hand, this calibration

502 model was developed using a wide variety of samples of different floral and regional

503 sources so that it proves to be more useful to predict amounts of the 4 sugars in a wide

504 variety of honeys 78.

19
505 FTIR-ATR spectroscopy was evaluated by Ruoff , for the simultaneous

506 determination of a high number of sugars and the main quality physicochemical

507 parameters in honey. Good precision accuracies were obtained for moisture, fructose,

508 glucose, sucrose, melezitose, total monosaccharide content (sum of fructose and

509 glucose), fructose/glucose and glucose/moisture ratios, electric conductivity, pH, and

510 free acidity. The prediction accuracy of the fructose (r2: 0.84) was similar to that

511 obtained by FT-NIR 18. Glucose (r2: 0.94) was predicted with accuracy similar to those

512 obtained by NIR 15,16 and FT-mATR 78. On the other hand, the prediction accuracy of

513 sucrose (r2: 0.91) was similar to that obtained by Tewari 78 and Qiu 15, but it was better

514 than that obtained by Ruoff 18. The prediction accuracy for HMF, proline, and minor

515 sugars (maltose, turanose, erlose, trehalose, isomaltose, and kojibiose) was rather poor
19
516 .

517 A partial least squares calibration model using second-derivative transformed

518 spectra (PLS-2nd derivative) was developed for the prediction of fructose, glucose,

519 sucrose, and maltose concentrations in honey from different geographical regions

520 around the world by FTIR-ATR. The reference method used to quantify sugars was

521 HPLC-ELSD 79. Calibration model was developed using a set of 45 standard mixtures.

522 The predicted concentrations had good agreement with those obtained by the reference

523 method (r2: 0.948-0.988). The correlations found were similar to those obtained in a

524 previous study by Tewari 78 (r2: 0.971-0.993) where a calibration model with standard

525 mixtures was also developed using FTIR-mATR in combination with PLS. The r2

ACS Paragon Plus Environment 21

 Journal of Agricultural and Food Chemistry Page 22 of 57

526 values for fructose, glucose and maltose were also similar to those found by Qiu 15 using

527 NIR and PLS (r2: 0.91-0.97). However, r2 for sucrose was better than that obtained by

528 Qiu 15 (r2: 0.86).

529 Glucose, fructose, sucrose, trehalose, maltose, melezitose, and turanose were

530 determined in samples of Portuguese honey by FTIR-ATR. HPAEC-PAD was used as

531 the reference method. Good PLS-1st derivative calibration models were obtained for

532 glucose (r2: 0.879), fructose (r2: 0.841), melezitose (r2: 0.890), and turanose (r2: 0.825)
73
533 . No calibration model was established for sucrose and trehalose due to the lower

534 amounts of these sugars in samples.

535 Raman spectroscopy is also a valuable technique to evaluate the quality of

536 honey. The main advantages of this over other spectroscopy techniques are that the

537 water present in the samples does not interfere with the Raman measurement. There is

538 only a minimal fluorescence interference of the honey matrix that varies from sample to

539 sample.

540 A method based on Fourier transform Raman (FT-Raman) spectroscopy in

541 combination with PLS was developed to determine fructose and glucose in honey. This

542 method was assessed using a standard HPLC-RID method. Both methods gave

543 statistically similar results being also equivalent with respect to their reproducibility 80.

544 Raman spectroscopy has also used for the determination of maltose and sucrose,

545 as well as fructose and glucose 81. A method using HPLC-RID was used as the reference

546 method. In this work, commercial honeys were used for the calibration data set. 2

547 multivariate techniques, artificial neural network (ANN) and PLS, were studied. Both

548 techniques provided very good results. The r2 values between actual and predicted

549 values of sugars ranged between 0.949 and 0.968 for PLS and 0.956 and 0.978 for
80
550 ANN. As compared to the work of Basoulis , a greater number of training and

ACS Paragon Plus Environment 22

Page 23 of 57 Journal of Agricultural and Food Chemistry

551 prediction samples have been used. Therefore, the calibration model obtained seems

552 more reliable in improving prediction efficiency.

553

554 Honey adulteration with sweeteners. Due to the their nutritional and healthy

555 properties and unique flavour, the commercial value of honey is much higher than that

556 of other sweeteners such as beet sugar, cane sugar, corn syrup (CS), high fructose corn

557 syrup (HFCS), and invert syrup (IS). Therefore, honey can be adulterated with these

558 sweeteners.

559 The addition of cane and beet invert syrup is usually difficult to detect. These

560 sweeteners can be added to honey imitating its natural sucrose–glucose–fructose profile.

561 Furthermore, adding only a moderate amount of invert syrup, the levels of glucose and

562 fructose remain within their normal range. However, adulteration with invert syrup may

563 be detected indirectly studying the oligosaccharides profile of honey samples. Genuine

564 and adulterated honeys show different profiles. Oligosaccharides in invert sugar are

565 produced as impurities during the hydrolysis of sugar into glucose and fructose 82.

566 The presence of high levels of HMF may be due to the adulteration of the honey

567 with inverted syrup since the acid hydrolysis of sucrose produces HMF 83. However,

568 high levels of HMF are also produced by excessive heating during the honey

569 manufacturing process or by inadequate storage conditions. Old honeys contain the

570 highest concentrations of HMF. Therefore, HMF alone cannot be used for detecting

571 adulteration with invert syrup.

572 HFCS is produced by hydrolysis and isomerization of corn starch. According to

573 their fructose content, this type of syrup is classified as: HFCS-42 (42 % fructose),

574 HFCS-55 (55 % fructose), and HFCS-90 (90 % fructose). HFCS presence significantly

575 affects the pH and HMF content. HFCS is also used to feed bees, especially in winter

ACS Paragon Plus Environment 23

 Journal of Agricultural and Food Chemistry Page 24 of 57

576 months and early spring 84. Different analytical techniques, such as isotopic analysis,

577 chromatography, and spectroscopy, have been used in order to detect honey adulteration

578 with different sweeteners (Table 5).

579 The stable carbon isotope ratio analysis (SCIRA) is a standard analytical
85-88
580 technique for detecting adulteration of honey . This technique is based on the
13
581 determination of the C/12C isotope ratio. According to their carbon metabolism, the

582 plants that produce substances used to adulterate honeys are classified as C3 or C4

583 plants. The C3 plants fix CO2 by the Calvin cycle while the C4 plants do using the
13
584 Hatch-Slack cycle. C3 plants have a lower C/12C ratio than C4 plants. Most of the

585 honey-contributing plants (wheat, rice, and sugar beets, among others) are C3 plants.

586 The C4 sugars most used to adulterate honeys are corn and cane sugar 84. SCIRA is an

587 expensive and time-consuming technique that also requires a considerable technical

588 skill.

589 Profiles of malto-oligosaccharides obtained by HPAEC-PAD were used to detect

590 adulteration with CS or HFCS with different degrees of isomerization: 20 % (20HFCS),

591 40 % (40HFCS), and 80 % (80HFCS). Honey samples were treated with activated

592 charcoal before analysis in order to remove mono and disaccharides 61. CS and 20HFCS

593 profiles showed oligosaccharides with a degree of polymerization (DP) up to 16 and at

594 relatively high concentrations. Oligosaccharides with DP up to 7 were detected in the

595 40HFCS and 80HFCS syrups and only oligosaccharides with DP up to 6 were

596 quantified presenting much lower concentrations than those found in the other syrups.

597 In honey samples, oligosaccharides up 7 DP were detected and their amounts were very

598 low and variable among samples. A comparison of oligosaccharide profiles of honey

599 samples and the same samples adulterated with one of the different syrups at different

600 amounts (5, 10, and 20 % (w/w)) enabled to detect adulteration with CS (down to 5 %

ACS Paragon Plus Environment 24

Page 25 of 57 Journal of Agricultural and Food Chemistry

601 (w/w)), 20HFCS, and 40HFCS. Adulteration with 80HFCS was not detected due to the

602 low amounts of oligosaccharides present in this syrup, comparable to those present in

603 honey samples.

604 A simple chromatographic separation of sugars by HPLC-RID was proposed to

605 detect honey adulteration with starch syrup. A characteristic chromatographic peak of

606 syrup can be found. No such peak at the same retention time is seen in chromatograms

607 of pure honeys. This characteristic peak is an overlapping peak of oligosaccharides with

608 a DP greater than 5. Therefore, this oligosaccharide peak was proposed as a syrup

609 indicator in adulterated honey, with a content of detectable syrup near 2.5 % (w/w). On

610 the other hand, the syrup content in adulterated honeys could be approximately

611 calculated using the height of this peak in the chromatograms. Using this method,

612 preliminary treatment of the samples is not required and no organic solvent is needed.

613 Therefore, the method proposed is simple, of low cost, environmentally

614 unobjectionable, and easy to perform in quality control 27.

615 Spectroscophic techniques (NIR, MIR, Raman) coupled with chemometric

616 methods have proven to be effective to recognize adulterated honey and to determine

617 the amount of adulterant added. For quality-grading purposes, discriminating between

618 different ranges of adulterant concentration might be enough.

619 NIR transflectance spectroscopy was evaluated to discriminate between

620 unadulterated honey and the same honey adulterated by different fructose:glucose

621 mixtures. Honey samples of 6 floral origins were used. Five classification methods were

622 evaluated. The best classification model obtained is known as least square support

623 vector machine (LS-SVM). Wavelet transformation (WT) was more effective than
89
624 principal component analysis (PCA) for selection of variables . More samples of

ACS Paragon Plus Environment 25

 Journal of Agricultural and Food Chemistry Page 26 of 57

625 different places and botanical origins could be included to expand the calibration model

626 and improve the robustness and accuracy of the method.

627 The use of fibre optic diffuse reflectance FT-NIR spectroscopy has been

628 assessed in order to differentiate between adulterated with HFCS and authentic blossom

629 honey samples. Classification of honey samples was realized using discriminant partial

630 least squares (D-PLS) analysis. The main bands responsible for the discrimination of

631 samples were in the range of 6000-10000 cm-1. Honey samples containing HFCS were

632 purchased in the market. The 100 % of genuine honeys and 95 % of adulterated honeys
90
633 from the test set were correctly classified . This method presented the advantage

634 compared to most methods of using adulterated honey samples not prepared in

635 laboratory. Therefore, adulterated honey samples were truly representative improving

636 the prediction efficiency.

637 FTIR-ATR was evaluated for determine the amount of beet medium invert

638 sugar added to 3 varieties of honey (clover, bluckwheat, and orange blossom honey) 91.

639 Two multivariate techniques, PLS and principal component regression (PCR), were

640 assessed. PLS-1st derivative models for the different varieties of honeys gave slightly

641 better results that those obtained using other calibration models. The r2 values for

642 validation set ranged between 0.946 and 0.956. SEP values were between 2.1 and 4.5 %

643 (w/w). Adulterated samples were classified into 3 groups according to the levels of beet

644 invert sugar added to honey. For this purpose, linear discriminant analysis (LDA) and

645 canonical variate analysis (CVA) were evaluated. PCA was used as the data

646 compression technique. The best predictive model was achieved using CVA, which

647 successfully classified 88.2-94.1 % of the validation set. On the other hand, a single-

648 calibration model including all varieties of honey gave lower correlations and higher

ACS Paragon Plus Environment 26

Page 27 of 57 Journal of Agricultural and Food Chemistry

649 SEC and SEP values than the individual models, and the percentage of correct

650 classification was reduced to 78.4 % for the validation set using CVA.

651 This methodology was used to study adulteration with cane medium invert sugar
92
652 by the same authors . The same varieties of honey were analyzed. The PLS-1st

653 derivative model gave SEP values between 2.8 to 3.5 % (w/w); the r2 values ranged

654 between 0.93 and 0.96. The PLS data compression technique was used in discriminant

655 analysis. Optimum classifications of 88.2-96.4 % and 89.3-94.1 % were achieved in the

656 validation set, using LDA and CVA, respectively. However, LDA was slightly better

657 for orange blossom honey and CVA for clover and bluckwheat. A single-calibration

658 model including all varieties of honey was also studied and gave worse results than

659 individual models.

660 An analytical method using FTIR-ATR spectroscopy was developed for

661 determining the amount added of CS to 3 different varieties of honeys (bluckwheat,

662 clover, and orange). Using PLS-1st derivative models, SEP values were in the range

663 1.99-2.22 % (w/w). The r2 values ranged between 0.899 and 0.940. A combined model

664 for the 3 different varieties of honey was studied and gave worse results than individual
93
665 models . This result was similar to that obtained by the same authors for other

666 adulterants, such as beet and cane medium invert sugar 91,92.

667 FTIR-ATR spectroscopy was applied to the discrimination of Irish artisanal

668 honeys from those adulterated with 1 of 5 sugar syrups: fully inverted beet syrup,

669 HFCS, partial invert cane syrup, dextrose syrup, and beet sucrose 94. A soft independent

670 modeling class analogy (SIMCA) model allowed correct identification of 96.2 % of

671 authentic honeys. Furthermore, 97.5 % of beet sucrose-adulterated honeys and 95.8 %

672 of honeys adulterated with dextrose syrup were correctly classified as adulterated

673 honeys. Most of the samples adulterated with the others syrups were incorrectly

ACS Paragon Plus Environment 27

 Journal of Agricultural and Food Chemistry Page 28 of 57

674 classified as genuine honeys. Therefore, the model may be used only to detect

675 adulteration with dextrose syrup or beet sucrose in Irish artisanal honeys. Minimum

676 levels of 10 % and 7 % (w/w) for beet sucrose or dextrose syrup, respectively, were

677 detected. On the other hand, a PLS model that allowed differentiate between authentic

678 honeys and honeys adulterated with partial invert cane syrup was obtained. Applying

679 this model to the samples adulterated with HFCS or fully inverted beet syrup, most of

680 them were identified as not cane partial inverted syrup-adulterated honeys, i.e., they

681 were classified as genuine honeys. This model allowed confirmation of the identity of

682 partial invert cane syrup-adulterated honeys with a correct classification rate of 91.7 %.

683 Using the 2 models, SIMCA and PLS, adulteration with beet sucrose, dextrose syrup, or

684 partial invert cane syrup can be detected.

685 FTIR-ATR spectroscopy was used to quantify CS, HFCS, and inverted sugar in

686 honeys of 4 states of México (Estado de México, Oaxaca, Chiapas, and Morelos). PLS

687 models for the 3 adulterants were developed. Good prediction accuracies were obtained.

688 The r2 values ranged between 0.976 and 0.999. SEP values were between 1.5 and 2.1 %

689 (w/w) for CS, 2.1-3.0 % (w/w) for HFCS, and 1.4-2.5 % (w/w) for inverted sugar.

690 Furthermore, a classification model based on SIMCA was developed in order to

691 discriminate the geographic origin of the Mexican honeys. Using this model a correct

692 classification rate of 100 % was achieved 95.

693 Raman spectroscopy was used to detect honey adulteration with cane and beet

694 medium invert sugar in 3 varieties of honey (bluckwheat, clover, and orange blossom

695 honey) 82. PLS analysis was used to predict amounts added of the adulterants. Models

696 were made for each adulterant and type of honey. The prediction models had an

697 acceptable accuracy. Values of r2 between 0.909 and 0.952 were obtained. SEP values

698 were between 1.56-2.20 % (w/w). As compared with results obtained by Sivakesava 91,92

ACS Paragon Plus Environment 28

Page 29 of 57 Journal of Agricultural and Food Chemistry

699 using FTIR-ATR, the r2 values were somewhat lower. However, SEP values were

700 better. Discriminant analysis by CVA with PCA compressed data was used in order to

701 classify adulterated honey based on the different concentration ranges for the

702 adulterants. About 96 % was the correct minimum classification rate obtained. Results
91,92
703 were better than those obtained by Sivakesava , using FTIR-ATR. CVA was also

704 used to differentiate between the 2 different types of adulterants regardless of the

705 honeys floral origin. The percentages of correct classification obtained were 90 %

706 (approximately) and 91 % for the calibration and validation sets, respectively. Better

707 predictions were obtained when discriminant models were developed for each type of

708 honey. In this case, the classification accuracy from all floral varieties was

709 approximately 96 % for both calibration and validation data sets.

710

711 ANALYTICAL METHODS FOR METHYLGLYOXAL AND OTHER 1, 2-

712 DICARBONYL COMPOUNDS

713 The 1,2-dicarbonyl compounds in honey, such as glyoxal (GO), 2-oxopropanal

714 or MGO, and 3-deoxyglucosone (3-DG), are formed from reducing carbohydrates by

715 the Maillard reaction or caramelization reactions. GO and MGO are considered markers

716 for an advanced state of glucose degradation. GO is the product of autoxidation for

717 glucose and MGO is formed by retroaldolization of the intermediate 3-DG 96. Analytical

718 methods used for determining MGO and other 1,2-dicarbonyl compounds in honey are

719 shown in Table 6.

720 1,2-dicarbonyl compounds were determined in honey by RP-HPLC with UV-

721 detection at 312 nm, preceded by derivatization to their corresponding quinoxalines

722 with orthophenylenediamine (OPD) 96. The derivatization procedure was done at room

723 temperature in the absence of light for 12 h. The chromatographic separation was

ACS Paragon Plus Environment 29

 Journal of Agricultural and Food Chemistry Page 30 of 57

724 carried out using gradient elution: solvent A: 0.075 % (v/v) acetic acid and solvent B:

725 80 % (v/v) aqueous methanol containing 0.075 % (v/v) acetic acid. The LODs obtained

726 were: 0.16, 0.03, and 0.04 mg/L for 3-DG, GO, and MGO, respectively. The ranges of

727 concentration used for the calibration showed good linearity: 0.32-81 mg/L for 3-DG,

728 0.06-1.7 mg/L for GO, and 0.07-2.2 mg/L for MGO. HMF, formed preferably under

729 acidic conditions from 3-DG, was separately quantified using RP-HPLC. Elution was

730 performed in isocratic mode with a buffer containing 15 % (v/v) methanol and 85 %

731 (v/v) 0.05 M phosphate-buffer (pH 5.5). The concentration of HMF was much lower

732 than that of 3-DG and no correlation between the 2 compounds was found.

733 Another 1,2-dicarbonyl compound, glucosone, unknown to occur in foods before

734 2004, was isolated and identified by HPLC-MS and NMR spectroscopy. The estimated

735 amount of glucosone in all samples analyzed was between 18 and 262 mg/kg, with an

736 average value of 90 mg/kg. On the other hand, an increase of 3-DG and HMF was

737 observed when the samples of honey were stored at 35 ºC and 45 ºC. However, the

738 concentrations of GO and MGO did not show changes and no significant increase in the

739 content of glucosone was observed. 3-DG proved to be a more sensitive indicator for

740 heat treatment than HMF 96.

741 GO, MGO and 3-DG were also determined in honey using the Mavric method,

742 which is a slight modification of the Weigel method 97. The LODs obtained were 0.3

743 mg/kg for 3-DG and 0.2 mg/kg for GO and MGO. Higher concentration ranges for

744 calibration curves were used showing good linearity: 10-500 mg/L for 3-DG, 0.1-20

745 mg/L for GO, and 0.1-300 mg/L for MGO. HMF was determined separately. The

746 amounts of 3-DG were also much higher than those of HMF, and no correlation

747 between the 2 compounds was found. Furthermore, GO and MGO were not affected by

748 storage conditions.

ACS Paragon Plus Environment 30

Page 31 of 57 Journal of Agricultural and Food Chemistry

749 HPLC-DAD coupled to mass spectrometry with an electrospray ionization

750 interface (HPLC-DAD-ESI-MS) has also been used for the determination of GO, MGO,

751 and 3-DG in honeys 98. 1,2-dicarbonyl compounds were derivatized with OPD before

752 analysis. Electrospray ionization in positive mode was performed. Gradient elution was

753 used: solvent A: methanol/acetic acid (99.97:0.03 v/v) and solvent B methanol/acetic

754 acid (99.85:0.15 v/v). The detection wavelength used was 312 nm. The estimated LODs

755 for 3-DG, GO, and MGO were 0.130, 0.083, and 0.063 µg/mL, respectively. LODs for

756 GO and MGO were higher than those reported by Weigel 96. However, the LOD of 3-

757 DG was similar. The concentration ranges used for the calibration curves were 3−520,

758 0.1-60, and 0.1-250 mg/L, for 3-DG, GO, and MGO, respectively.

759 Honey is valued for its healing properties and can be used as a topical

760 antibacterial agent for surface wound infections. Acidity, high osmolality, and hydrogen

761 peroxide contribute to the antimicrobial activity of honey. Non-peroxide factors include

762 lysozyme, phenolic acids, and flavonoids. MGO was identified as the component

763 mainly responsible for the non-peroxide antimicrobial activity in New Zealand manuka

764 honey (Leptsopermum scoparium) 97,99. Leptospermum honeys are well known for their

765 healing properties and include New Zealand manuka honey (Leptsopermum scoparium)

766 and Australian jelly bush honey (Leptospermum polygalifolium).

767 Amounts of MGO in New Zealand manuka honey (Leptsopermum scoparium)

768 between 38.4 and 761 mg/kg have been found97. These values were much higher than

769 those found in conventional honey 96-98 when MGO values below 13 mg/kg were found.

770 Concentrations of MGO in honeys from Australian native plants have also been
100
771 determined . The values obtained ranged from 8 to 1755 mg/kg, with an average

772 concentration of 357 mg/kg. The samples from the L. polygalifolium (jelly bush) floral

ACS Paragon Plus Environment 31

 Journal of Agricultural and Food Chemistry Page 32 of 57

773 source had the highest MGO concentrations (279-1755 mg/kg) with an average value of

774 604 mg/kg.

775 It has been reported that the high levels of MGO arise from the nonenzymatic

776 conversion of dihydroxyacetone (DHA) to MGO in maturating honey99. High levels of

777 DHA and low levels of MGO have been found in nectar and honey freshly produced by

778 bees.

779 A strong positive correlation between the amount of MGO in manuka honey and

780 antibacterial activity on Staphylococcus aureus was reported 101. The MGO content has

781 been used commercially as an indicator of the bioactivity of Leptospermum honeys.

782 MGO amount in honeys is expressed as a MGO number. For instance, ‘500 + MGO’

783 indicates that the honey contains a minimum amount of methylglyoxal of 500 mg/kg 100.

784 In order to classify manuka honey by antimicrobial strength, unique manuka

785 factor (UMF) is used. For its determination, the size of the zone of inhibition of

786 Staphylococcus aureus induced by the honey is compared with that induced by phenol

787 solutions. Honey having an antibacterial activity equivalent to that of a 5 % phenol

788 solution has a UMF 5+ and so on up to a 30 % phenol solution rated as UMF 30+ 102.

789 Manuka honey with a high UMF value has a good antibacterial activity. Therefore,

790 UMF 12+ manuka honey is commercialized for therapeutic purposes and that with

791 UMF 16+ would be more potent for specific medical applications.

792 MGO content is usually measured according to the above methods 96,97. These

793 methods are time-consuming and involve a number of steps and numerous reagents and

794 solvents. A direct method has been developed involving a mixed mode size-

795 exclusion/ligand exchange separation with 2 columns and with refractive index

796 detection 103. This method was compared to the Weigel method 96. The direct method

797 presented disadvantages. It proved to be much less sensitive with a detection limit ca.

ACS Paragon Plus Environment 32

Page 33 of 57 Journal of Agricultural and Food Chemistry

798 50 mg/kg. Therefore, it is not suitable to determine MGO in conventional honeys (non-

799 manuka honeys) with relatively low amounts of MGO. Another disadvantage is that the

800 standard additions method was necessary for quantification, known amounts of MGO

801 being added to a honey solution (50-900 mg/kg). Instead, the indirect method using

802 derivatization with OPD required an external calibration. On the other hand, better

803 resolution of the MGO peak was obtained using the indirect method, less prone to

804 interference with co-eluting peaks.

805 A method for the simultaneous determination of MGO and DHA, its precursor,

806 in Australian honeys has been proposed 104. The method involves a derivatization with

807 O-(2, 3, 4, 5, 6-pentafluorobenzyl) hydroxylamine·HCl (PFBHA). Subsequent analysis

808 of PFBHA derivatives by RP-HPLC with UV-detection at 263 nm was carried out.

809 Hydroxyacetone (HA) was used as an internal standard. Gradient elution was used for

810 separation: solvent A: 30 % (v/v) acetonitrile and solvent B: acetonitrile. The usable

811 linear range of the method for MGO was from 20 to 1800 mg/kg. This range was wider
103
812 than that used (50-900 mg/kg) by Adams . The method using PFBHA was less

813 sensitive than the methods developed by Weigel, Mavric, and Marshall 96-98. However, it

814 was sensitive enough to detect MGO and DHA in honeys from 6 Leptospermum species

815 (L. polygalifolium, L. trinervia, L. leavigatum, L. liversidgei, L. semibaccatum, and L.

816 speciosum) and blends with other species. High levels of MGO and DHA were found

817 ranging from 43 to 1723 mg/kg and from 412 to 2403 mg/kg for MGO and DHA,

818 respectively. High levels of DHA and MGO were present at the genus level, not being
97,103
819 restricted to manuka honey . This method presents some advantage over other
96,97
820 methods using HPLC-UV and previous derivatization with OPD which does not

821 detect monocarbonyl compounds such as DHA. Furthermore, the time for sample

822 preparation was reduced from 12-16 h to 2 h.

ACS Paragon Plus Environment 33

 Journal of Agricultural and Food Chemistry Page 34 of 57

823 The Windsor method was applied to the simultaneous determination of MGO,

824 DHA, and HMF in Austalian Leptospermun honeys 105. The PFBHA derivate of HMF

825 resulted in 2 chromatographic peaks corresponding to their cis and trans-isomers.

826 GC has also been employed to determine MGO in some New Zealand manuka

827 and kanuka honeys. A method using headspace solid-phase microextraction coupled to

828 gas chromatography with nitrogen/phosphorus detector (HS-SPME-GC–NPD) was

829 described 106. Before SPME, derivatization of MGO with OPD was carried out in vials

830 that were capped and left for 1 h to allow complete derivatization before analysis. 2,3-

831 Butanedione was used as an internal standard. The LOD of the method was 5 mg/kg.

832 This method was less sensitive than those developed by Weigel 96 and Mavric 97 using

833 HPLC-UV and derivatization with OPD. However, it was more sensitive than the direct

834 method developed by Adams 103 and the Windsor method 104 which uses HPLC-UV and

835 derivatization with PFBHA.

836 Besides chromatography, other analytical techniques have been used. A NMR
102
837 method was proposed . MGO was present in sample solutions as the mono- and

838 dihydrate which were independently measured. The values obtained were added to

839 determine the methylglyoxal concentration. The method was applied to samples of

840 manuka honey. The amounts of methylglyoxal ranged from 338 to 817 mg/kg. This

841 method was compared with the Mavric method using HPLC-UV detection and another

842 using HPLC coupled with time of flight mass spectrometry detection (HPLC/TOF-MS),

843 being MGO previously derivatized with OPD. The amounts of MGO obtained with

844 these 2 methods were in most of the cases somewhat higher than those obtained with the

845 NMR method. This could be explained by the Maillard reaction between sugars and

846 OPD which would generate MGO or by the conversion of DHA to MGO due to the

847 elevated temperature used during the process of derivatization 102. The NMR method

ACS Paragon Plus Environment 34

Page 35 of 57 Journal of Agricultural and Food Chemistry

848 presented advantages because it is applied directly on the diluted honey with neither

849 chromatographic separation nor sample derivatization.

850 Recently, MIR spectroscopy with ATR in combination with a PLS1 regression

851 model has been proposed to predict MGO content in Australian honeys from several
100
852 native plants . The reference method used derivatization of MGO with OPD and

853 posterior analysis by HPLC-UV detection. A r2 value of 0.75 between the predicted and

854 reference values was obtained. Robustness of the method could be improved for

855 commercial applications.

856 Sugars and several physicochemical parameters used in quality control can be

857 determined simultaneously with FTIR-ATR spectroscopy. Moreover, this technique

858 allows discrimination of adulterated honey samples containing several added

859 sweeteners, and determination of the amount added of each one. It is nondestructive,

860 rapid, requires only limited sample preparation, and is easy to perform. Therefore,

861 FTIR-ATR has been proposed for routine quality control by several researchers.

862 However, IR spectroscopy did not allow a good quantitative determination of HMF and

863 enzyme activities, two criteria of great importance for the honey trade to assess storage

864 and heat damage.

865 As conclusion, several analytical techniques are needed due to the complex

866 chemical composition of honey. Chromatographic techniques are widely used. More

867 recently, methods based on other techniques, such as Raman, IR, and NMR

868 spectroscopy, in combination with chemometric methods, have been proposed for

869 honey analysis. These methods rely on the inherent advantages of spectroscopy. They

870 are rapid, inexpensive, nondestructive, and not needing sample preparation or expensive

871 reagents. Official methods based on spectroscopic techniques should be developed and

872 validated for routine quality control of honeys.

ACS Paragon Plus Environment 35

 Journal of Agricultural and Food Chemistry Page 36 of 57

873

874

875 References

876 1. Iglesias, M.; De Lorenzo, C.; Polo, M.; Martin-Alverez, P.; Pueyo, E. Usefulness of

877 amino acid composition to discriminate between honeydew and floral honeys.

878 Application to honeys from a small geographic area. J. Agric. Food Chem. 2004, 52,

879 84-89.

880 2. Anonymous EU Directive 110/2001. http://eur-lex. europa.

881 eu/LexUriServ/LexUriServ. do?uri=OJ:L:2002:010:0047:0052:EN:PDF 2001,

882 3. Cajka, T.; Hajslova, J.; Pudil, F.; Riddellova, K. Traceability of honey origin based

883 on volatiles pattern processing by artificial neural networks. J. Chromatogr. a 2009,

884 1216, 1458-1462.

885 4. Ouchemoukh, S.; Louaileche, H.; Schweitzer, P. Physicochemical characteristics and

886 pollen spectrum of some Algerian honeys. Food Control 2007, 18, 52-58.

887 5. de la Fuente, E.; Sanz, M.L.; Martinez-Castro, I.; Sanz, J. Development of a robust

888 method for the quantitative determination of disaccharides in honey by gas

889 chromatography. J. Chromatogr. a 2006, 1135, 212-218.

890 6. Low, N.H.; Sporns, P. Analysis and Quantitation of Minor Di- and Trisaccharides in

891 Honey, Using Capillary Gas Chromatography. J. Food Sci. 1988, 53, 558-561.

ACS Paragon Plus Environment 36

Page 37 of 57 Journal of Agricultural and Food Chemistry

892 7. Ruiz-Matute, A.L.; Brokl, M.; Soria, A.C.; Sanz, M.L.; Martinez-Castro, I. Gas

893 chromatographic-mass spectrometric characterisation of tri- and tetrasaccharides in

894 honey. Food Chem. 2010, 120, 637-642.

895 8. Sanz, M.; Polemis, N.; Morales, V.; Corzo, N.; Drakoularakou, A.; Gibson, G.;

896 Rastall, R. In vitro investigation into the potential prebiotic activity of honey

897 oligosaccharides. J. Agric. Food Chem. 2005, 53, 2914-2921.

898 9. Cavia, M.; Fernandez-Muino, M.; Gomez-Alonso, E.; Montes-Perez, M.; Huidobro,

899 J.; Sancho, M. Evolution of fructose and glucose in honey over one year: influence of

900 induced granulation. Food Chem. 2002, 78, 157-161.

901 10. Gheldof, N.; Engeseth, N. Antioxidant capacity of honeys from various floral

902 sources based on the determination of oxygen radical absorbance capacity and

903 inhibition of in vitro lipoprotein oxidation in human serum samples. J. Agric. Food

904 Chem. 2002, 50, 3050-3055.

905 11. Conti, M.E. Lazio region (central Italy) honeys: a survey of mineral content and

906 typical quality parameters. Food Control 2000, 11, 459-463.

907 12. Khalil, M.I.; Sulaiman, S.A.; Gan, S.H. High 5-hydroxymethylfurfural

908 concentrations are found in Malaysian honey samples stored for more than one year.

909 Food Chem. Toxicol. 2010, 48, 2388-2392.

910 13. Martelo, M.S. Chemical characterization in honey harvested in Madrid's northern

911 mountains. Ars Pharm. 2000, 41, 387-396.

912 14. Osborne, B.G.; Fearn, T.; Hindle, P.H. Practical NIR spectroscopy with

913 applications in food and beverage analysis. Longman scientific and technical: 1993;

ACS Paragon Plus Environment 37

 Journal of Agricultural and Food Chemistry Page 38 of 57

914 15. Qiu, P.Y.; Ding, H.B.; Tang, Y.K.; Xu, R.J. Determination of chemical composition

915 of commercial honey by near-infrared spectroscopy. J. Agric. Food Chem. 1999, 47,

916 2760-2765.

917 16. García-Alvarez, M.; Huidobro, J.F.; Hermida, M.; Rodríguez-Otero, J.L. Major

918 components of honey analysis by near-infrared transflectance spectroscopy. J. Agric.

919 Food Chem. 2000, 48, 5154-5158.

920 17. Cozzolino, D.; Corbella, E. Determination of honey quality components by near

921 infrared reflectance spectroscopy. J. Apic. Res. 2003, 42, 16-20.

922 18. Ruoff, K.; Luginbühl, W.; Bogdanov, S.; Bosset, J.-.; Estermann, B.; Ziolko, T.;

923 Kheradmandan, S.; Amadò, R. Quantitative determination of physical and chemical

924 measurands in honey by near-infrared spectrometry. Eur. Food Res. Technol. 2007, 225,

925 415-423.

926 19. Ruoff, K.; Iglesias, M.T.; Luginbühl, W.; Bosset, J.-.; Bogdanov, S.; Amadò, R.

927 Quantitative analysis of physical and chemical measurands in honey by mid-infrared

928 spectrometry. Eur. Food Res. Technol. 2006, 223, 22-29.

929 20. de Rodriguez, G.; de Ferrer, B.; Ferrer, A.; Rodriguez, B. Characterization of honey

930 produced in Venezuela. Food Chem. 2004, 84, 499-502.

931 21. Bogdanov, S.; Ruoff, K.; Persano Oddo, L. Physico-chemical methods for the

932 characterisation of unifloral honeys: a review. Apidologie 2004, 35, 4-17.

933 22. Meda, A.; Lamien, C.; Romito, M.; Millogo, J.; Nacoulma, O. Determination of the

934 total phenolic, flavonoid and proline contents in Burkina Fasan honey, as well as their

935 radical scavenging activity. Food Chem. 2005, 91, 571-577.

ACS Paragon Plus Environment 38

Page 39 of 57 Journal of Agricultural and Food Chemistry

936 23. Can, Z.; Yildiz, O.; Sahin, H.; Turumtay, E.A.; Silici, S.; Kolayli, S. An

937 investigation of Turkish honeys: Their physico-chemical properties, antioxidant

938 capacities and phenolic profiles. Food Chem. 2015, 180, 133-141.

939 24. Bogdanov, S.; Lullmann, C.; Martin, P.; von der Ohe, W.; Russmann, H.; Vorwohl,

940 G.; Odo, L.; Sabatini, A.; Marcazzan, G.; Piro, R.; Flamini, C.; Morlot, M.; Lheritier, J.;

941 Borneck, R.; Marioleas, P.; Tsigouri, A.; Kerkvliet, J.; Ortiz, A.; Ivanov, T.; D'Arcy, B.;

942 Mossel, B.; Vit, P.; Int Honey Commission Honey quality and international regulatory

943 standards: review by the International Honey Commission. Bee World 1999, 80, 61-69.

944 25. Ough, C. Rapid Determination of Proline in Grapes and Wines. J. Food Sci. 1969,

945 34, 228-&.

946 26. Truzzi, C.; Annibaldi, A.; Illuminati, S.; Finale, C.; Scarponi, G. Determination of

947 proline in honey: Comparison between official methods, optimization and validation of

948 the analytical methodology. Food Chem. 2014, 150, 477-481.

949 27. Wang, S.; Guo, Q.; Wang, L.; Lin, L.; Shi, H.; Cao, H.; Cao, B. Detection of honey

950 adulteration with starch syrup by high performance liquid chromatography. Food Chem.

951 2015, 172, 669-674.

952 28. Azeredo, L.; Azeredo, M.; de Souza, S.; Dutra, V. Protein contents and

953 physicochemical properties in honey samples of Apis mellifera of different floral

954 origins. Food Chem. 2003, 80, 249-254.

955 29. Escuredo, O.; Míguez, M.; Fernández-González, M.; Carmen Seijo, M. Nutritional

956 value and antioxidant activity of honeys produced in a European Atlantic area. Food

957 Chem. 2013, 138, 851-856.

ACS Paragon Plus Environment 39

 Journal of Agricultural and Food Chemistry Page 40 of 57

958 30. Anklam, E. A review of the analytical methods to determine the geographical and

959 botanical origin of honey. Food Chem. 1998, 63, 549-562.

960 31. Nozal, M.; Bernal, J.; Toribio, L.; Jimenez, J.; Martin, M. High-performance liquid

961 chromatographic determination of methyl anthranilate, hydroxymethylfurfural and

962 related compounds in honey. J. Chromatogr. a 2001, 917, 95-103.

963 32. LoCoco, F.; Valentini, C.; Novelli, V.; Ceccon, L. High-performance liquid

964 chromatographic determination of 2-furaldehyde and 5-hydroxymethyl-2-furaldehyde in

965 honey. J. Chromatogr. a 1996, 749, 95-102.

966 33. Makawi, S.Z.A.; Taha, M.I.; Zakaria, B.A.; Siddig, B.; Mahmod, H.; Elhussein,

967 A.R.M.; Kariem, E.A.G. Identification and quantification of 5-hydroxymethyl furfural

968 HMF in some sugar-containing food products by HPLC. Pak. J. Nutr. 2009, 8, 1391-

969 1396.

970 34. Nassberger, L. Influence of 5-Hydroxymethylfurfural (5-Hmf) on the overall

971 Metabolism of Human Blood-Cells. Hum. Exp. Toxicol. 1990, 9, 211-214.

972 35. Bruce, W.R.; Archer, M.C.; Corpet, D.E.; Medline, A.; Minkin, S.; Stamp, D.; Yin,

973 Y.; Zhang, X. Diet, aberrant crypt foci and colorectal cancer. Mutat. Res. -Fund. Mol.

974 M. 1993, 290, 111-118.

975 36. Severin, I.; Dumont, C.; Jondeau-Cabaton, A.; Graillot, V.; Chagnon, M. Genotoxic

976 activities of the food contaminant 5-hydroxymethylfurfural using different in vitro

977 bioassays. Toxicol. Lett. 2010, 192, 189-194.

ACS Paragon Plus Environment 40

Page 41 of 57 Journal of Agricultural and Food Chemistry

978 37. Zhang, X.-.; Chan, C.C.; Stamp, D.; Minkin, S.; Archer, M.C.; Bruce, W.R.

979 Initiation and promotion of colonic aberrant crypt foci in rats by 5-hydroxymethyl-2-

980 furaldehyde in thermolyzed sucrose. Carcinogenesis 1993, 14, 773-775.

981 38. Surh, Y.J.; Liem, A.; Miller, J.A.; Tannenbaum, S.R. 5-Sulfooxymethylfurfural as a

982 possible ultimate mutagenic and carcinogenic metabolite of the Maillard reaction

983 product, 5-hydroxymethylfurfural. Carcinogenesis 1994, 15, 2375-2377.

984 39. Lee, Y.; Shlyankevich, M.; Jeong, H.; Douglas, J.; Surh, Y. Bioactivation of 5-

985 Hydroxymethyl-2-Furaldehyde to an Electrophilic and Mutagenic Allylic Sulfuric-Acid

986 Ester. Biochem. Biophys. Res. Commun. 1995, 209, 996-1002.

987 40. Janzowski, C.; Glaab, V.; Samimi, E.; Schlatter, J.; Eisenbrand, G. 5-

988 hydroxymethylfurfural: assessment of mutagenicity, DNA-damaging potential and

989 reactivity towards cellular glutathione. Food Chem. Toxicol. 2000, 38, 801-809.

990 41. Durling, L.J.; Busk, L.; Hellman, B.E. Evaluation of the DNA damaging effect of

991 the heat-induced food toxicant 5-hydroxymethylfurfural (HMF) in various cell lines

992 with different activities of sulfotransferases. Food Chem. Toxicol. 2009, 47, 880-884.

993 42. Jeuring, H.; Kuppers, F. High-Performance Liquid-Chromatography of Furfural and

994 Hydroxymethylfurfural in Spirits and Honey. J. Assoc. Off. Anal. Chem. 1980, 63,

995 1215-1218.

996 43. White, J. Spectrophotometric Method for Hydroxymethylfurfural in Honey. J Assoc

997 Off Agric Chem. 1979, 62, 509-514.

998 44. Zappala, A.; Fallico, B.; Arena, E.; Verzera, A. Methods for the determination of

999 HMF in honey: a comparison. Food Control 2005, 16, 273-277.

ACS Paragon Plus Environment 41

 Journal of Agricultural and Food Chemistry Page 42 of 57

1000 45. Truzzi, C.; Annibaldi, A.; Illuminati, S.; Finale, C.; Rossetti, M.; Scarponi, G.

1001 Determination of Very Low Levels of 5-(Hydroxymethyl)-2-furaldehyde (HMF) in

1002 Natural Honey: Comparison Between the HPLC Technique and the Spectrophotometric

1003 White Method. J. Food Sci. 2012, 77, C784-C790.

1004 46. Cabras, P.; Angioni, A.; Tuberoso, C.; Floris, I.; Reniero, F.; Guillou, C.; Ghelli, S.

1005 Homogentisic acid: A phenolic acid as a marker of strawberry-tree (Arbutus unedo)

1006 honey. J. Agric. Food Chem. 1999, 47, 4064-4067.

1007 47. Spano, N.; Casula, L.; Panzanelli, A.; Pilo, M.; Piu, P.; Scanu, R.; Tapparo, A.;

1008 Sanna, G. An RP-HPLC determination of 5-hydroxymethylfurfural in honey - The case

1009 of strawberry tree honey. Talanta 2006, 68, 1390-1395.

1010 48. Spano, N.; Ciulu, M.; Floris, I.; Panzanelli, A.; Pilo, M.I.; Piu, P.C.; Salis, S.; Sanna,

1011 G. A direct RP-HPLC method for the determination of furanic aldehydes and acids in

1012 honey. Talanta 2009, 78, 310-314.

1013 49. Schade, J.; Marsh, G.; Eckert, J. Diastase activity and hydroxy‐methyl‐furfural in

1014 honey and their usefulness in detecting heat alteration. J. Food Sci. 1958, 23, 446-463.

1015 50. White Jr, J.W.; Pairent, F. Report on the analysis of honey. J Assoc Off Agric Chem

1016 1959, 42, 341-348.

1017 51. Zurcher, K.; Hadorn, H. Eine einfache kinetische Methode zur Bestimmung der

1018 Diastasezahl in Honig. Deut Lebensmittel rundsch 1972, 68, 209-216.

1019 52. Oddo, L.P.; Pulcini, P. A scientific note on the Phadebas method for honeys with

1020 low enzyme content. Apidologie 1999, 30, 347-348.

ACS Paragon Plus Environment 42

Page 43 of 57 Journal of Agricultural and Food Chemistry

1021 53. McCleary, B.V.; Sturgeon, R. Measurement of alpha-amylase in cereal, food, and

1022 fermentation products. Cereal Foods World 2002, 47, 299-+.

1023 54. Bentabol Manzanares, A.; Hernández García, Z.; Rodríguez Galdón, B.; Rodríguez

1024 Rodríguez, E.; Díaz Romero, C. Physicochemical characteristics of minor monofloral

1025 honeys from Tenerife, Spain. LWT - Food Sci. Technol. 2014, 55, 572-578.

1026 55. Sakac, N.; Sak-Bosnar, M. A rapid method for the determination of honey diastase

1027 activity. Talanta 2012, 93, 135-138.

1028 56. Baltrusaityte, V.; Venskutonis, P.R.; Ceksteryte, V. Radical scavenging activity of

1029 different floral origin honey and beebread phenolic extracts. Food Chem. 2007, 101,

1030 502-514.

1031 57. Fell, R. Color Grading of Honey. Am. Bee J. 1978, 118, 782-&.

1032 58. González-Miret, M.L.; Terrab, A.; Hernanz, D.; Fernández-Recamales, M.A.;

1033 Heredia, F.J. Multivariate correlation between color and mineral composition of honeys

1034 and by their botanical origin. J. Agric. Food Chem. 2005, 53, 2574-2580.

1035 59. Junk, W.; Pancoast, H. Boiling point elevations of sugar solutions. Handbook of

1036 Sugars: For Processors, Chemists & Technologists, 1st Ed. (HM Pancoast, ed. ) 1973,

1037 42-66.

1038 60. García-Alvarez, M.; Ceresuela, S.; Huidobro, J.F.; Hermida, M.; Rodríguez-Otero,

1039 J.L. Determination of polarimetric parameters of honey by near-infrared transflectance

1040 spectroscopy. J. Agric. Food Chem. 2002, 50, 419-425.

ACS Paragon Plus Environment 43

 Journal of Agricultural and Food Chemistry Page 44 of 57

1041 61. Morales, V.; Corzo, N.; Sanz, M.L. HPAEC-PAD oligosaccharide analysis to detect

1042 adulterations of honey with sugar syrups. Food Chem. 2008, 107, 922-928.

1043 62. Arias, V.C.; Castells, R.C.; Malacalza, N.; Lupano, C.E.; Castells, C.B.

1044 Determination of oligosaccharide patterns a in honey by solid-phase extraction and

1045 high-performance liquid chromatography. Chromatographia 2003, 58, 797-801.

1046 63. Ni, C.; Zhu, B.; Wang, N.; Wang, M.; Chen, S.; Zhang, J.; Zhu, Y. Simple column-

1047 switching ion chromatography method for determining eight monosaccharides and

1048 oligosaccharides in honeydew and nectar. Food Chem. 2016, 194, 555-560.

1049 64. Horvath, K.; MolnarPerl, I. Simultaneous quantitation of mono-, di- and

1050 trisaccharides by GC-MS of their TMS ether oxime derivatives .2. In honey.

1051 Chromatographia 1997, 45, 328-335.

1052 65. Barez, J.; Garcia-Villanova, R.; Garcia, S.; Pala, T.; Paramas, A.; Sanchez, J.

1053 Geographical discrimination of honeys through the employment of sugar patterns and

1054 common chemical quality parameters. Eur Food Res Technol 2000, 210, 437-444.

1055 66. Sanz, M.; Sanz, J.; Martinez-Castro, I. Characterization of O-trimethylsilyl oximes

1056 of disaccharides by gas chromatography-mass spectrometry. Chromatographia 2002,

1057 56, 617-622.

1058 67. Bogdanov, S.; Vit, P.; Kilchenmann, V. Sugar profiles and conductivity of stingless

1059 bee honeys from Venezuela. Apidologie 1996, 27, 445-450.

1060 68. Pourtallier, J. Ueber die Benutzung der GC für die Bestimmung der Zucker in

1061 Honig. Z. für Bienenforschung 1967, 9, 217-221.

ACS Paragon Plus Environment 44

Page 45 of 57 Journal of Agricultural and Food Chemistry

1062 69. Sabatini, G.; Nanetti, A.; Maurizi, M.; Lercker, G. Studio del'origine dei mieli

1063 attraverso il profilo gaschromatografico dei componenti neutri. Rivista di merceologia

1064 1984, 23, 71-81.

1065 70. Leite, J.; Trugo, L.; Costa, L.; Quinteiro, L.; Barth, O.; Dutra, V.; De Maria, C.

1066 Determination of oligosaccharides in Brazilian honeys of different botanical origin.

1067 Food Chem. 2000, 70, 93-98.

1068 71. White, J.; Meloy, R.; Probst, J.; Huser, W. Sugars Containing Galactose Occur in

1069 Honey. J. Apic. Res. 1986, 25, 182-185.

1070 72. Ouchemoukh, S.; Schweitzer, P.; Bey, M.B.; Djoudad-Kadji, H.; Louaileche, H.

1071 HPLC sugar profiles of Algerian honeys. Food Chem. 2010, 121, 561-568.

1072 73. Anjos, O.; Campos, M.G.; Ruiz, P.C.; Antunes, P. Application of FTIR-ATR

1073 spectroscopy to the quantification of sugar in honey. Food Chem. 2015, 169, 218-223.

1074 74. de la Fuente, E.; Ruiz-Matute, A.I.; Valencia-Barrera, R.M.; Sanz, J.; Martínez

1075 Castro, I. Carbohydrate composition of Spanish unifloral honeys. Food Chem. 2011,

1076 129, 1483-1489.

1077 75. Sanz, M.; Sanz, J.; Martinez-Castro, I. Gas chromatographic-mass spectrometric

1078 method for the qualitative and quantitative determination of disaccharides and

1079 trisaccharides in honey. J. Chromatogr. a 2004, 1059, 143-148.

1080 76. Kelly, J.D.; Downey, G.; Fouratier, V. Initial study of honey adulteration by sugar

1081 solutions using midinfrared (MIR) spectroscopy and chemometrics. J. Agric. Food

1082 Chem. 2004, 52, 33-39.

ACS Paragon Plus Environment 45

 Journal of Agricultural and Food Chemistry Page 46 of 57

1083 77. Dvash, L.; Afik, O.; Shafir, S.; Schaffer, A.; Yeselson, Y.; Dag, A.; Landau, S.

1084 Determination by near-infrared spectroscopy of perseitol used as a marker for the

1085 botanical origin of avocado (Persea americana Mill.) honey. J. Agric. Food Chem. 2002,

1086 50, 5283-5287.

1087 78. Tewari, J.; Irudayaraj, J. Quantification of saccharides in multiple floral honeys

1088 using fourier transform infrared microattenuated total reflectance spectroscopy. J. Agric.

1089 Food Chem. 2004, 52, 3237-3243.

1090 79. Wang, J.; Kliks, M.M.; Jun, S.; Jackson, M.; Li, Q.X. Rapid analysis of glucose,

1091 fructose, sucrose, and maltose in honeys from different geographic regions using

1092 Fourier transform infrared spectroscopy and multivariate analysis. J. Food Sci. 2010,

1093 75, C208-C214.

1094 80. Batsoulis, A.; Siatis, N.; Kimbaris, A.; Alissandrakis, E.; Pappas, C.; Tarantilis, P.;

1095 Harizanis, P.; Polissiou, M. FT-Raman spectroscopic simultaneous determination of

1096 fructose and glucose in honey. J. Agric. Food Chem. 2005, 53, 207-210.

1097 81. Ozbalci, B.; Boyaci, I.H.; Topcu, A.; Kadilar, C.; Tamer, U. Rapid analysis of

1098 sugars in honey by processing Raman spectrum using chemometric methods and

1099 artificial neural networks. Food Chem. 2013, 136, 1444-1452.

1100 82. Paradkar, M.; Irudayaraj, J. Discrimination and classification of beet and cane

1101 inverts in honey by FT-Raman spectroscopy. Food Chem. 2002, 76, 231-239.

1102 83. Swallow, K.W.; Low, N.H. Determination of honey authenticity by anion-exchange

1103 liquid chromatography. J. AOAC Int. 1994, 77, 695-702.

ACS Paragon Plus Environment 46

Page 47 of 57 Journal of Agricultural and Food Chemistry

1104 84. Zábrodská, B.; Vorlová, L. Adulteration of honey and available methods for

1105 detection–a review. Acta Veterinaria Brno 2015, 83, 85-102.

1106 85. White, J.; Winters, K.; Martin, P.; Rossmann, A. Stable carbon isotope ratio analysis

1107 of honey: Validation of internal standard procedure for worldwide application. J. AOAC

1108 Int. 1998, 81, 610-619.

1109 86. Martı́n, I.G.; Macıá s, E.M.; Sánchez, J.S.; Rivera, B.G. Detection of honey

1110 adulteration with beet sugar using stable isotope methodology. Food Chem. 1998, 61,

1111 281-286.

1112 87. Padovan, G.; De Jong, D.; Rodrigues, L.; Marchini, J. Detection of adulteration of

1113 commercial honey samples by the 13 C/12 C isotopic ratio. Food Chem. 2003, 82, 633-

1114 636.

1115 88. Cabañero, A.I.; Recio, J.L.; Ruperez, M. Liquid chromatography coupled to isotope

1116 ratio mass spectrometry: a new perspective on honey adulteration detection. J. Agric.

1117 Food Chem. 2006, 54, 9719-9727.

1118 89. Zhu, X.; Li, S.; Shan, Y.; Zhang, Z.; Li, G.; Su, D.; Liu, F. Detection of adulterants

1119 such as sweeteners materials in honey using near-infrared spectroscopy and

1120 chemometrics. J. Food Eng. 2010, 101, 92-97.

1121 90. Chen, L.; Xue, X.; Ye, Z.; Zhou, J.; Chen, F.; Zhao, J. Determination of Chinese

1122 honey adulterated with high fructose corn syrup by near infrared spectroscopy. Food

1123 Chem. 2011, 128, 1110-1114.

1124 91. Sivakesava, S.; Irudayaraj, J. Detection of inverted beet sugar adulteration of honey

1125 by FTIR spectroscopy. J. Sci. Food Agric. 2001, 81, 683-690.

ACS Paragon Plus Environment 47

 Journal of Agricultural and Food Chemistry Page 48 of 57

1126 92. Sivakesava, S.; Irudayaraj, J. Prediction of inverted cane sugar adulteration of honey

1127 by Fourier transform infrared spectroscopy. J. Food Sci. -Chicago 2001, 66, 972-978.

1128 93. Sivakesava, S.; Irudayaraj, J. A rapid spectroscopic technique for determining honey

1129 adulteration with corn syrup. J. Food Sci. 2001, 66, 787-792.

1130 94. Kelly, J.D.; Petisco, C.; Downey, G. Application of Fourier transform midinfrared

1131 spectroscopy to the discrimination between Irish artisanal honey and such honey

1132 adulterated with various sugar syrups. J. Agric. Food Chem. 2006, 54, 6166-6171.

1133 95. Gallardo-Velazquez, T.; Osorio-Revilla, G.; Zuniga-de Loa, M.; Rivera-Espinoza,

1134 Y. Application of FTIR-HATR spectroscopy and multivariate analysis to the

1135 quantification of adulterants in Mexican honeys. Food Res. Int. 2009, 42, 313-318.

1136 96. Weigel, K.; Opitz, T.; Henle, T. Studies on the occurrence and formation of 1,2-

1137 dicarbonyls in honey. Eur. Food Res. Technol. 2004, 218, 147-151.

1138 97. Mavric, E.; Wittmann, S.; Barth, G.; Henle, T. Identification and quantification of

1139 methylglyoxal as the dominant antibacterial constituent of Manuka (Leptospermum

1140 scoparium) honeys from New Zealand. Mol. Nutr. Food Res. 2008, 52, 483-489.

1141 98. Marshall, S.M.; Schneider, K.R.; Cisneros, K.V.; Gu, L. Determination of

1142 Antioxidant Capacities, alpha-Dicarbonyls, and Phenolic Phytochemicals in Florida

1143 Varietal Honeys Using HPLC-DAD-ESI-MSn. J. Agric. Food Chem. 2014, 62, 8623-

1144 8631.

1145 99. Adams, C.J.; Manley-Harris, M.; Molan, P.C. The origin of methylglyoxal in New

1146 Zealand manuka (Leptospermum scoparium) honey. Carbohydr. Res. 2009, 344, 1050-

1147 1053.

ACS Paragon Plus Environment 48

Page 49 of 57 Journal of Agricultural and Food Chemistry

1148 100. Sultanbawa, Y.; Cozzolino, D.; Fuller, S.; Cusack, A.; Currie, M.; Smyth, H.

1149 Infrared spectroscopy as a rapid tool to detect methylglyoxal and antibacterial activity in

1150 Australian honeys. Food Chem. 2015, 172, 207-212.

1151 101. Atrott, J.; Henle, T. Methylglyoxal in Manuka Honey - Correlation with

1152 Antibacterial Properties. Czech J. Food Sci. 2009, 27, S163-S165.

1153 102. Donarski, J.A.; Roberts, D.P.T.; Charlton, A.J. Quantitative NMR spectroscopy for

1154 the rapid measurement of methylglyoxal in manuka honey. Anal. Method. 2010, 2,

1155 1479-1483.

1156 103. Adams, C.J.; Boult, C.H.; Deadman, B.J.; Farr, J.M.; Grainger, M.N.C.; Manley-

1157 Harris, M.; Snow, M.J. Isolation by HPLC and characterisation of the bioactive fraction

1158 of New Zealand manuka (Leptospermum scoparium) honey. Carbohydr. Res. 2008,

1159 343, 651-659.

1160 104. Windsor, S.; Pappalardo, M.; Brooks, P.; Williams, S.; Manley-Harris, M. A

1161 convenient new analysis of dihydroxyacetone and Methylglyoxal applied to Australian

1162 Leptospermum honeys. J. Pharmacogn. Phytother. 2012, 4, 6-11.

1163 105. Windsor, S.; Kavazos, K.; Brooks, P. The quantitation of hydroxymethylfurfural in

1164 Australian Leptospermum honeys. J. Pharmacogn. Phytother. 2013, 5, 21-25.

1165 106. Stephens, J.M.; Schlothauer, R.C.; Morris, B.D.; Yang, D.; Fearnley, L.;

1166 Greenwood, D.R.; Loomes, K.M. Phenolic compounds and methylglyoxal in some New

1167 Zealand manuka and kanuka honeys. Food Chem. 2010, 120, 78-86.

1168 107. Ajlouni, S.; Sujirapinyokul, P. Hydroxymethylfurfuraldehyde and amylase

1169 contents in Australian honey. Food Chem. 2010, 119, 1000-1005.

ACS Paragon Plus Environment 49

 Journal of Agricultural and Food Chemistry Page 50 of 57

1170 108. Bentabol-Manzanares, A.; Hernandez Garcia, Z.; Rodriguez Galdon, B.; Rodriguez

1171 Rodriguez, E.; Diaz Romero, C. Differentiation of blossom and honeydew honeys using

1172 multivariate analysis on the physicochemical parameters and sugar composition. Food

1173 Chem. 2011, 126, 664-672.

1174 109. León-Ruiz, V.; Vera, S.; González-Porto, A.V.; Andrés, M.P.S. Vitamin C and

1175 sugar levels as simple markers for discriminating spanish honey sources. J. Food Sci.

1176 2011, 76, C356-C361.

1177 110. Santos, A.; Moreira, R.F.A.; De Maria, C.A.B. Study of the principal constituents

1178 of tropical angico (Anadenanthera sp.) honey from the atlantic forest. Food Chem. 2015,

1179 171, 421-425.

1180 111. Nayik, G.A.; Dar, B.N.; Nanda, V. Physico-chemical, rheological and sugar profile

1181 of different unifloral honeys from Kashmir valley of India. Arab. J. Chem. 2015,

1182 112. Goodall, I.; Dennis, M.; Parker, I.; Sharman, M. Contribution of High-Performance

1183 Liquid-Chromatographic Analysis of Carbohydrates to Authenticity Testing of Honey.

1184 J. Chromatogr. a 1995, 706, 353-359.

1185 113. Cotte, J.; Casabianca, H.; Chardon, S.; Lheritier, J.; Grenier-Loustalot, M.

1186 Application of carbohydrate analysis to verify honey authenticity. J. Chromatogr. a

1187 2003, 1021, 145-155.

1188 114. Cordella, C.B.Y.; Militao, J.S.L.T.; Clement, M.C.; Cabrol-Bass, D. Honey

1189 characterization and adulteration detection by pattern recognition applied on HPAEC-

1190 PAD profiles. 1. Honey floral species characterization. J. Agric. Food Chem. 2003, 51,

1191 3234-3242.

ACS Paragon Plus Environment 50

Page 51 of 57 Journal of Agricultural and Food Chemistry

1192 115. Sanz, M.; Gonzalez, M.; de Lorenzo, C.; Sanz, J.; Martinez-Castro, I. A

1193 contribution to the differentiation between nectar honey and honeydew honey. Food

1194 Chem. 2005, 91, 313-317.

1195 116. Kaskoniene, V.; Venskutonis, P.R.; Ceksteryte, V. Carbohydrate composition and

1196 electrical conductivity of different origin honeys from Lithuania. Lwt-Food Science and

1197 Technology 2010, 43, 801-807.

1198 117. Cho, H.; Hong, S. Acacia honey quality measurement by near-infrared

1199 spectroscopy. J. Near Infrared Spectrosc. 1998, 6, A329-A331.

1200 118. Ha, J.; Koo, M.; Ok, H. Determination of the constituents of honey by near

1201 infrared spectroscopy. J. Near Infrared Spectrosc. 1998, 6, A367-A369.

1202

1203

ACS Paragon Plus Environment 51

 Journal of Agricultural and Food Chemistry Page 52 of 57

1204

1205 Table 1
1206 Classification of quality parameters of honey.
Parameters related to preservation Moisture
pH
Free acidity and total acidity
Parameters related to freshness Diastase activity
Invertase activity
Hydroxymethylfurfural
Parameters related to ripeness Proline
Moisture
Sucrose
Parameters related to the botanical Ash content
origin Moisture
Color
Electric Conductivity
Specific rotation
Fructose/glucose
Glucose/moisture
pH
Free acidity and total acidity
Proline
Sugar composition
Parameters related to adulteration Sucrose
Glucose-fructose-sucrose profile
Proline
Hydroxymethylfurfural
Parameters related to crystallization Fructose/glucose
Glucose/moisture
1207
1208
1209

ACS Paragon Plus Environment 52

Page 53 of 57 Journal of Agricultural and Food Chemistry

1210 Table 2

1211 Composition criteria of honey (EU Directive 110/2001).

Parameter Concentration
Fructose and glucose content (sum of both):
- Blossom honey ≥ 60 % (w/w)
- Honeydew honey and blends of honeydew honey ≥ 45 % (w/w)
with blossom honeys
Sucrose content, in general (with exceptions) ≤ 5 % (w/w)
Moisture content, in general (with exceptions) ≤ 20 % (w/w)
Water-insoluble content:
- in general ≤ 0.1 % (w/w)
- pressed honey ≤ 0.5 % (w/w)
Electric conductivity:
- Blossom honey (with exceptions) ≤ 0.8 mS/cm
- Honeydew and chestnut honey ≥ 0.8 mS/cm
Free acid, in general ≤ 50 milli-equivalents/kg
Diastase activity:
- In general ≥ 8 (Schale units)
- Honeys with low natural enzyme content ≥ 3 (Schale units)
- (e.g. citrus honey) and an HMF content ≤ 15 mg/kg
Hydroxymethylfurfural (HMF):
- In general except baker´s honey ≤ 40 mg/kg
- Honeys of declared origin from regions with ≤ 80 mg/kg
tropical climate and blends of these honeys
Honeys with low natural enzymatic level (e.g. ≤ 15 mg/kg
citrus honey)
1212
1213

ACS Paragon Plus Environment 53

 Journal of Agricultural and Food Chemistry Page 54 of 57

1214 Table 3

1215 Analytical techniques for the determination of hydroxymethylfurfural and other furanic

1216 alhehydes and acids in honey.

Compound Analytical techniques References

12,31,33,42,44,47,48,96,105,107
Hydroxymethylfurfural RP-HPLC (detection UV)

43,44
Hydroxymethylfurfural UV/VIS spectrophotometry

105
Hydroxymethylfurfural RP-HPLC (UV-detection),
previous derivatization with
PFBHA
31,48
Others furanic aldehydes RP- HPLC (detection UV)
and acids
(2-F, 3-F, 2- FA, 3- FA)
1217 2-F: 2-furfuraldehyde or furfural, 3-F: 3-furfuraldehyde, 2-FA: 2-furfuroic acid,

1218 3-FA: 3-furfuroic acid

1219 PFBHA: O-(2, 3, 4, 5, 6-pentafluorobenzyl) hydroxylamine·HCl

1220

1221

1222

1223

1224

1225

1226

1227

ACS Paragon Plus Environment 54

Page 55 of 57 Journal of Agricultural and Food Chemistry

1228 Table 4

1229 Analytical techniques for the determination of carbohydrates.

Analytical techniques References

27,54,67,70,80,108-111
HPLC-RID
79
HPLC-ELSD
62
HPLC-DAD
29,61,72,73,112-114
HPAEC-PAD
68,69,113,115,116
GC-FID
74,75
GC-MS
15,16,18,77,117,118
NIR spectroscopy
19,73,78,79
FTIR-ATR spectroscopy
80,81
Raman spectroscopy
1230

1231

ACS Paragon Plus Environment 55

 Journal of Agricultural and Food Chemistry Page 56 of 57

1232 Table 5

1233 Analytical techniques for the detection of honey adulteration with sweeteners.

Technique References
85-88
Stable carbon isotope ratio analysis (SCIRA)
27,61
Chromatography
89,90
NIR spectroscopy
91,92,94
FTIR-ATR spectroscopy
82
Raman spectroscopy
1234

1235

1236

1237

1238

1239

1240

ACS Paragon Plus Environment 56

Page 57 of 57 Journal of Agricultural and Food Chemistry

1241 Table 6.

1242 Analytical methods for the determination of methylglyoxal

Methods References
96,97
RP-HPLC with UV-detection, previous *

derivatization with OPD

104,105
RP-HPLC with UV-detection, previous **

derivatization with PFBHA

103
RP-HPLC with RI detection (direct

method)
102
HPLC-TOF-MS, previous derivatization

with OPD
98
HPLC-DAD-ESI-MS, previous *

derivatization with OPD

106
Headspace SPME/GC-NPD, previous

derivatization with OPD

102
NMR spectroscopy

100
MIR spectroscopy-ATR

1243
1244 OPD: ortho-phenylenediamine, PFBHA: O-(2, 3, 4, 5, 6-pentafluorobenzyl)
1245 hydroxylamine·HCl,
1246 * 3-deoxyglucosone (3-DG), methylglyoxal (MGO), and glyoxal (GO) were
1247 determined.
1248 ** hydroxymethylfurfural (HMF), methylglyoxal (MGO), and dihydroxyacetone

1249 (DHA) were determined.

ACS Paragon Plus Environment 57