Sei sulla pagina 1di 38

Subscriber access provided by Kaohsiung Medical University

Food and Beverage Chemistry/Biochemistry


Effect of Roasting on Oligosaccharide Abundance in Arabica Coffee Beans
TIAN TIAN, Samara Freeman, Mark Corey, J. Bruce German, and Daniela Barile
J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.8b02641 • Publication Date (Web): 03 Sep 2018
Downloaded from http://pubs.acs.org on September 4, 2018

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 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 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.

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 37 Journal of Agricultural and Food Chemistry

Effect of Roasting on Oligosaccharide Abundance in Arabica Coffee Beans

Tian Tian,ǂ Samara Freeman,† Mark Corey,§ J. Bruce German, ǂ, † and Daniela Barile*,ǂ,†

ǂDepartment of Food Science and Technology, University of California, Davis, California 95616,

United States

† Foods for Health Institute, University of California, Davis, One Shields Avenue, Davis,

California 95616, United States

§ Keurig Green Mountain, Inc., Waterbury, Vermont 05676, United States

* To whom correspondence should be addressed. Tel: +1-530-752-0976; Fax: +1-530-752-4759;

E-mail: dbarile@ucdavis.edu

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 2 of 37

1 ABSTRACT: Emerging research into the bioactivities of indigestible carbohydrates is

2 illuminating the potential of various foods and food streams to serve as novel sources of

3 health-promoting compounds. Oligosaccharides (OS) are widely present in milks and some

4 plants. Our previous research demonstrated the presence of OS in brewed coffee and spent coffee

5 grounds. Armed with this new knowledge, the next step toward improving the utilization of these

6 valuable components involved investigating the effect of roasting on the formation and

7 abundance of coffee OS. In the present study, we used advanced mass spectrometry to analyze a

8 variety of coffee samples and demonstrated that a great structural diversity and increased

9 abundance of OS is associated with higher roasting intensity. The present investigation also

10 evaluated methods for OS extraction and fractionation. A preparative-scale chromatographic

11 method, based on activated carbon, was developed to isolate enough amounts of OS from coffee

12 to enable future confirmation of prebiotic and other in vitro activities.

13 KEYWORDS: Coffea arabica, mass spectrometry, oligosaccharides, prebiotics

14

ACS Paragon Plus Environment


Page 3 of 37 Journal of Agricultural and Food Chemistry

15 INTRODUCTION

16 Roasting of coffee beans is a very important step in coffee processing. It is associated with the

17 development of organoleptic parameters that are important indicators of the quality of the roasted

18 beans. The two major types of polysaccharides that involve structural modifications in green

19 coffee beans during roasting are type II arabinogalactans and galactomannans.1 A series of

20 reactions involving polysaccharides have been studied, including depolymerization, dehydration,

21 cleavage of carbon-carbon linkages, their incorporation in melanoidins through transglycolation

22 reactions, and some structural modifications at the sugar reducing end.2, 3 4, 5. Additionally

23 previous research has indicated that the roasting process opens the cell wall matrix, enables the

24 hydrolysis reactions of polysaccharides, and promotes the formation and release of

25 oligosaccharides (OS).6

26 OS are carbohydrates generally consisting of two to twenty monomers linked by a variety of

27 O-glycosidic bonds.7 Other than sucrose, tiny amount of raffinose (0.77%) and stachyose

28 (0.51%), there is no evidence for the presence of naturally occurring OS in green coffee beans. 8

29 Only a few OS hypothesized to derive from the structural modifications of coffee

30 polysaccharides during roasting,7, 9, 10 have been found in roasted coffee beans. Our previous

31 research established an OS extraction and purification method, and elucidated the composition

32 and concentration of OS in dark roasted coffee beans and spent coffee grounds.11

33 OS are targets of new investigations because they exhibit highly selective prebiotic activity.

34 Prebiotics are dietary ingredients that cannot be digested by human-produced digestive enzymes,

35 yet they provide a health benefit to the host mediated by selectively stimulating the growth

36 and/or activity of one or a limited number of host gut microbiota.13 Several food OS, including

37 galactooligosaccharides and inulin, have been reported to act as prebiotics.14 Additionally, a

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 4 of 37

38 recent in vitro study demonstrated that bovine milk OS can block attachment of pathogens to the

39 epithelial cells and interact directly with intestinal cells. 15

40 Only a few studies have explored the biological activities of coffee OS. In one study, purified

41 coffee mannooligosaccharides resisted degradation when exposed to a series of digestive enzyme;

42 moreover, the mannooligosaccharides were fermented by human fecal bacteria, and the products

43 of fermentation included beneficial short chain fatty acids.16 In another study, consumption of

44 coffee was positively associated with an increased population and metabolic activity of

45 Bifidobacterium spp. in feces of healthy adult volunteers who consumed three cups of coffee

46 daily for three weeks.17 In a third study, Escherichia coli and Clostridium spp. were decreased,

47 whereas Bifidobacterium spp. were increased in the colons of mice that consumed coffee,

48 compared with their numbers in a control group.18

49 With an increasing interest in identifying sources of selective prebiotics to serve as novel

50 food ingredients to improve human health, there is a considerable demand for natural,

51 concentrated sources of novel oligosaccharides with improved bioactivity. However, it is

52 extremely difficult to obtain functional oligosaccharides on a large scale since oligosaccharides

53 have diverse sugar sequences and are found in only trace amounts in the obvious sources (e.g.

54 bovine milk). Armed with this new knowledge, the next step toward improving the utilization of

55 these valuable components involved investigating the effect of conventional coffee processing on

56 the formation of OS. The degree of roasting is one of the key factors that could determine the

57 type and amount of OS formed and hence transferred into brewed coffee. Several investigators

58 discussed the effect of roasting on degradation and structural modification of carbohydrates,

59 especially polysaccharides, in coffee beans.2, 12 However, the structural modifications of coffee

60 carbohydrates induced by roasting remain incompletely characterized.

ACS Paragon Plus Environment


Page 5 of 37 Journal of Agricultural and Food Chemistry

61 Investigating the effect of roasting on the composition and abundance of coffee OS is important

62 as it could lead to an improved understanding of the structures of coffee OS. In the present study,

63 we used advanced mass spectrometry to analyze a variety of coffee samples and demonstrated

64 that a great structural diversity and increased abundance of OS is associated with higher roasting

65 intensity.

66 Controlling roasting parameters, and optimizing both the extraction in a preparative purification

67 methods will lead to a more effective and targeted extraction of OS. Here, we studied the effect

68 of roasting intensity on the composition and abundance of coffee OS and selected the optimum

69 roasting conditions for the production of OS. The present investigation also evaluated methods

70 for OS extraction and fractionation. A preparative-scale chromatographic method, based on

71 activated carbon, was developed to isolate enough amounts of OS from coffee to enable future

72 confirmation of prebiotic and other in vitro activities.

73

74 MATERIALS AND METHODS

75 Chemicals and Reagents. D-galactose, D-glucose, D-mannose, L-arabinose, D-xylose,

76 L-rhamnose, and L-allose were from Sigma-Aldrich (St. Louis, MO, USA). Analytical grade

77 standard 2’-fucosyllactose (2’-FL) was purchased from V-Laboratories (Covington, LA, USA).

78 A commercial sucrose/D-glucose/D-fructose kit was from R-biopharm AG (Darmstadt,

79 Germany). The total carbohydrate colorimetric assay kit was purchased from BioVision,

80 Incorporated (Milpitas, CA, USA). The Tri-Sil HTP (HDMS:TMCS:Pyridine) reagent was

81 purchased from Thermo Fisher Scientific (Waltham, MA, USA). The Discovery® DSC-C8

82 1-mL solid-phase extraction columns and the Discovery DSC-C8 column packing material

83 were purchased from SUPELCO (Bellefonte, PA, USA). The Extract-Clean TM Carbo porous

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 6 of 37

84 graphitized carbon cartridges (150 mg) were purchased from GRACE Davison Discovery

85 Science (Deerfield, IL, USA). The activated carbon charcoal (50–200 mesh) was purchased from

86 FisherbrandTM (Columbia, MD, USA).

87 The 60-mL filtration tubes with a Teflon Frits were purchased from SUPELCO (Bellefonte,

88 PA, USA). The 0.22-µm MILLEX GP filter units were purchased from Merck Millipore Ltd.

89 (Tullagreen, Carrigtwohill, Co. Cork, IRL). All solvents used were HPLC-MS grade (Fisher

90 Scientific, Fiar Lawn, NJ).

91 Processing and Characterization of Coffee Beans. Raw or roasted ground Coffea arabica

92 were provided by Keurig® Green Mountain, Inc. (Waterbury, VT, USA). A blend of coffee beans

93 sourced from Southeast Asia, East Africa, and South America was used. Beans were ground

94 using a No. 5 Ditting® grinder. The coffee bean samples were stored at –20 oC until ready to use.

95 To study the effect of roasting intensity on the composition and abundance of OS, seven

96 samples of the blended beans were roasted and coded from R1 to R7, where R1 was the least

97 roasted and R7 was the most heavily roasted. Green beans from the same blend were also

98 analyzed. The roasting levels of beans were indicated by the color characteristics of the ground

99 beans, including the Agtron/SCAA color measurement and CIE L*a*b*. The Agtron numbers

100 were obtained using a roast classification system (the Specialty Coffee Association of America

101 and Agtron Inc. (Reno, NV, USA)). Twenty grams of ground beans were transferred to a plastic

102 Petri dish and compared to the color disk. Both samples and the color disk were viewed and

103 compared against a black neutral density background provided in the kit. The colorimetric

104 measurement was performed using a portable Chromameter CR-410 (Konica Minolta, Ramsey,

105 NJ, USA) on the ground coffee beans directly. The chromameter was calibrated with a standard

106 white plate (D65 illuminant, Y=93.4, x = 0.3162, y = 0.3342). Values for L* (lightness), a*

ACS Paragon Plus Environment


Page 7 of 37 Journal of Agricultural and Food Chemistry

107 (red-green), and b* (yellow-blue) were obtained. The chroma [c = (a2+b2)1/2] and the hue [arctan

108 (b/a)] were calculated accordingly.19-21

109 Small Scale OS Isolation and Purification for Characterization and Quantification. OS

110 from ground C. arabica beans were isolated and purified as previously described.11 Briefly, a 5-g

111 sample of ground coffee beans was boiled with 100 mL of nanopure water (Direct-Q® 5 UV

112 Remote Water Purification System, Millipore Sigma, Burlington, MA, USA) for 20 min under

113 constant stirring, resulting in a liquid coffee extract. The liquid coffee extract was prepared in

114 triplicate. Total dissolved solids (Table 1) were measured (Ultramete II™ 4P, Myron L®

115 Company, Carlsbad, CA, USA) to check efficiency of extraction. The pH values (Table 1) of the

116 resulting liquid coffee extract were also measured with SevenCompact S220-Basic, pH/Ion

117 benchtop meter (Mettler-Toledo, LLC, Columbus, OH, USA). The majority of lipids and some

118 proteins in the coffee extract were removed using the Folch method by mixing one volume of

119 coffee extract with four volumes of a mixture of chloroform and methanol (2:1, vol:vol). The

120 mixture was mixed vigorously with a vortex mixer, and then was centrifuged at 4,000 ×g for 30

121 min at 4 oC. The upper aqueous layer containing OS was vacuum dried, re-suspended in

122 nanopure water, and subjected to stepwise solid-phase extraction. The OS-containing solution

123 was first loaded to a Discovery® DSC C8 column to remove residual lipids, small peptides, and

124 polysaccharides. The entire eluate was then loaded onto the porous graphitized carbon column.

125 The hydrophilic OS were bonded by porous graphitized carbon cartridge. The bonded OS

126 fraction was eluted stepwise with 20% acetonitrile, then 40% acetonitrile/0.05% trifluoacetic

127 acid. Those OS-rich eluate fractions were combined, was vacuum dried and suspended in

128 nanopure water for LC-MS and GC-FID analysis.

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 8 of 37

129 Characterization of OS by Chip-Quadrupole-Time-of-Flight Mass Spectrometry. Prior

130 to LC-MS analysis, dried OS samples were reconstituted in 200 µL of nanopure water, with

131 2’-FL being added as internal standard at the final concentration of 0.005 g/L. Mass spectrometry

132 analysis was performed with an Agilent 6520 NanoChip-LC-Quadrupole-Time-of-Flight (QToF)

133 with a microfluidic nanoelectrospray PGC-Chip (II) (Agilent Technologies, Santa Clara, CA,

134 USA) according to the previously published method, with minor modification.22 The

135 micro-fluidic PGC-Chip (II) (Agilent Technologies, Santa Clara, CA, USA) contained an

136 enrichment column (4 mm; 40 nL) and an analytical column (43 mm × 75 µm), both packed with

137 porous graphitized carbon. Chromatographic separation was performed by binary solvent

138 gradients of 3% acetonitrile/0.1% formic acid (solvent A) in water and 90% acetonitrile/0.1%

139 formic acid in water (solvent B). The columns were initially equilibrated with 100% solvent A

140 with a flow rate of 0.3 µL/min for the nano pump and 4 µL/min for the capillary pump. The

141 65-min gradient was programmed as follows: 0 to 2.5 min, 0%B; 2.5 to 20 min, 0 to 16% B; 20

142 to 30 min, 16 to 40% B; 30 to 40 min, 40% to 100% B, followed by 100% B for 10 min; 50 to

143 55.01 min, 100 to 0% B, and re-equilibrium at 0% B for10 min.

144 Mass spectra were acquired in positive mode within a m/z range of 450–2,500. OS were

145 detected in their protonated form ([M+H]+). Internal mass calibration was conducted using two

146 reference masses (m/z 922.009798, 1221.990637, ESI-ToF tuning mix G1969-85000, Agilent

147 Technologies). Automated precursor selection was employed based on ion abundance for tandem

148 MS analysis, performing up to 6 MS/MS spectra per individual MS when precursor ions were

149 above ion abundance threshold. The threshold for peak selection was set at 200 ion counts for

150 MS and 5 ion counts for MS/MS. The acquisition rate was 0.63 spectra/s. The isolation width for

ACS Paragon Plus Environment


Page 9 of 37 Journal of Agricultural and Food Chemistry

151 tandem MS was set on medium (~ 4 m/z). The collision energy was set at 1.8V/100 Da with an

152 offset of –3.6V.

153 MS data were analyzed using the molecular feature extraction function of Mass Hunter

154 Qualitative Analysis software version B.06.00 and Mass Hunter Profinder B.06.00 (Agilent

155 Technologies). The putative OS structures were extracted through the “find compound by

156 formula” algorithm. The compounds were filtered with mass range from 450 to 2500 m/z and

157 retention time between 5 and 30 min. Target compounds had an ion count higher than 1,000,

158 either singly or doubly charged, and a typical isotopic distribution of small biological molecules.

159 OS molecular formulas were determined from the deconvoluted mass list with in-house software,

160 with an error as low as 10 ppm. All the OS compositions were confirmed by tandem MS analysis.

161 The abundance of each OS was calculated as (Area os/ Area i.s. ×average Area i.s.). Average area

162 of the internal standard was calculated over all roasting levels and replicates.

163 Characterization and Quantification of Constituent Monosaccharides by GC-FID. A

164 gas chromatograph with a flame-ionization detector (GC-FID) was used to characterize the

165 constituent monosaccharides and their absolute amounts by employing methanolysis and

166 trimethylsilylation derivatization following published procedure.23 Purified coffee OS (0.5 mL)

167 were dried, dissolved in 0.5 mL MeOH/0.5 M HCl, and incubated at 80 oC for 16 h. The reaction

168 mixture was cooled to room temperature and dried under a stream of nitrogen. The dried sample

169 was mixed with 250 µL methanol and dried under a stream of nitrogen. The mixture was

170 trimethylsilylated by incubating with TriSil® reagent (300 µL) for 1 h at 80 oC. The reaction

171 mixture was cooled to room temperature and excess solvent was removed under a stream of

172 nitrogen. A 1-mL aliquot of hexane was added to the dried sample and centrifuged at 10,000 rpm

173 for 10 min to separate salts. The solution was dried under a stream of nitrogen and suspended in

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 10 of 37

174 150 µL of hexane prior to injection (1 µL) into a GC coupled to an FID controlled by a

175 Hewlett-Packard ChemStation. Separation was carried out using a DB-1 fused-silica capillary

176 column (30 m × 250 µm i.d., 0.25 µm film thickness, J&W Scientific, Folsom, CA, USA).

177 Samples were injected in the pulsed split mode with a split ratio of 5:1. The injector and the FID

178 temperature was 280 oC. The GC oven temperature was programmed from 120 to 200 oC at 1.5
o
179 C/min, 200 oC for 5 min isothermal, and a post run of 2 min at 250 oC. Hydrogen was used as

180 carrier gas at a flow rate of 2.5 mL/min in constant flow mode. Calibration curves were built for

181 the following monosaccharide standards: glucose, galactose, mannose, arabinose, xylose, and

182 rhamnose. Allose was used as internal standard.

183 Statistical Analysis. The differences in constituent monosaccharides and the extrapolated

184 amount of OS among different samples were analyzed using ANOVA followed by Tukey’s

185 post-hoc test. All statistical analyses were conducted in R software, version 3.1.2. The threshold

186 of significant difference was set at p ≤ 0.05.

187 Preparative-scale Purification of Coffee OS. The entire workflow for the preparative scale

188 purifications is depicted in Figure 1

189 Five grams of the dark roasted coffee beans were subjected to hot water extraction as

190 previously described. The resulting extract was mixed with cold ethanol (1:2 vol:vol) and stored

191 at –30 oC for 1 h to precipitate the majority of proteins, polysaccharides, and melanoidins. The

192 mixture was centrifuged at 4,225 ×g for 30 min. The supernatant was defatted using the Folch

193 method as previously described for the small scale. The aqueous layer was vacuum dried and

194 re-suspended in water prior to solid-phase extraction by an in-house packed column.

195 The C8 column was packed by dissolving 15 g of the Discovery DSC-C8 column packing

196 material in 60 mL acetonitrile. The slurry was transferred to a 60 mL filtration tube with a

10

ACS Paragon Plus Environment


Page 11 of 37 Journal of Agricultural and Food Chemistry

197 Teflon Frit on the bottom. The column elution time was set for 30 min and the remaining

198 acetonitrile was eluted slowly at ambient atmosphere pressure. The column was sealed with a

199 Teflon Frit on the top. The packed column was washed three times with 100 mL of water to

200 remove the residual acetonitrile and condition the column at ambient atmosphere pressure.

201 The coffee OS extract was loaded on the manually packed C8 column and washed with 150

202 mL of water. The eluate was collected in three 50-mL fractions (C8 eluate fraction 1, 2 and 3).

203 To further purify the OS in this eluate, each fraction was incubated with 2 g of activated carbon

204 for 3 h. The activated carbon was packed into a12 mL filter tubes with a Teflon Frit on the

205 bottom. The remaining solution was eluted slowly and collected (AC eluate fraction 1). The

206 column was sealed with a Teflon Frit on the top and washed three times with 10 mL water to

207 remove residual monosaccharides and sucrose (AC eluate fraction 2). The OS were eluted with

208 30 mL of 20% acetonitrile (AC eluate fraction 3, target fraction containing OS). All eluate

209 fractions were dried under vacuum, combined, resuspended in water, and passed through a 0.22

210 µm MILLEX GP filter unit (Merck Millipore Ltd., Tullagreen, Carrigtwohill, Co. Cork, IRL) to

211 remove residual activated carbon powder before further analysis.

212 A microplate colorimetric carbohydrate assay (Biovision) was used to quantify the purified

213 coffee OS. A commercial 2 mg/mL glucose standard (0, 2, 4, 6, 8, and 10 µL) was used to

214 establish a calibration curve. The final volume of each calibration point was adjusted to 30 µL

215 with nanopure water. A 30 µL aliquot of purified coffee OS was used at adjusted dilution.

216 Glucose standards and samples were incubated with 150 µL of sulfuric acid (98%) at 85 oC for

217 15 min. After incubation, 30 µL of developer (provided by the manufacturer) was added to each

218 well. Samples were mixed for 5 min and absorbance was measured at 490 nm by a SpectraMax

11

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 12 of 37

219 190 Microplate Reader (Molecular Devices, San Jose, CA, USA). The quantity of total

220 carbohydrates was calculated based on the calibration curve.

221 RESULTS AND DISCUSSION

222 Characteristic Parameters of the Coffee Beans and the Brewed Extract. The

223 Agtron/SCAA roast measurement and CIE L*a*b* color characteristics of the ground coffee

224 beans from the different roasting levels are indicated in Table 1. The Agtron number of samples

225 varied between 90 and 25, which corresponds, respectively, to the degrees of roasting intensity

226 varying from the lightest roast with distinguishable coffee flavor characteristics (90) to the

227 espresso roast (25). L* and b* values, and calculated chroma and hue values of roasted beans

228 were significantly lower (p < 0.001) than that of the green beans, and also decreased as the

229 roasting intensity increased. The a* value increased from green beans to light roasted beans, and

230 then decreased along the increased roasting intensity.

231 The significant inverse correlation between L*, b* values and intensity of roasting were well

232 discussed in previous publications.19, 20 However, the discussions about the change in a* value

233 during roasting were controversial. Moss and Otten,24 and Somporn et al. 20reported an elevated

234 a* value with increased levels of roasting. Bicho et al.25 reported a sharp increase of a* values in

235 roasted beans as compared to green beans, followed by a decreased trend with a continued

236 roasting process, which was also observed in this study.

237 The change in pH and total dissolved solids in brewed extract as a result of roasting level are

238 also shown in Table 1. Prior to roasting, the pH of green coffee brew was 5.73, close to the value

239 of 5.8 reported by Ramalakshmi et al.26 The pH first decreased to 4.82 in the brew of the lightest

240 roasted beans, and increased up to 5.32 with elevated roasting intensity. A previous study also

241 observed the same trend.27 The decrease in pH was attributed to the formation of formic, acetic,

12

ACS Paragon Plus Environment


Page 13 of 37 Journal of Agricultural and Food Chemistry

242 glycolic, and lactic acids as the coffee beans were roasted to medium level. The increase in pH in

243 the dark roasted beans are possibly caused by the destruction of organic acids formed during

244 roasting and those present initially (citric acid, malic acid, and chlorogenic acids).28

245 An increased total dissolved solids as a result of heavier roasting intensity was also noticed.

246 This result can be partially due to the effect of roasting on opening the cell wall structures and

247 facilitating the release of polysaccharides and OS. 6

248 Effects of Roasting on the Formation and Abundance of OS. The diversity in OS

249 structures, including monosaccharide compositions, linkages, and isomers and the range of sizes

250 provides the basis for a selective support of probiotic growth. Only bacteria equipped with a

251 specific set of glycoside hydrolases, transporters and other molecules contributed to the OS

252 degradation would be able to cleave the various monosaccharides and utilize them as a carbon

253 source.29 Therefore, even different isomers with the same composition are not equivalent in

254 biological function and separation and differentiation achieved at structural isomer level is crucial

255 for the structure-function analysis of OS.

256 The effect of roasting intensity on the formation and abundance of OS was studied by

257 analyzing the green beans and seven samples roasted to different intensities (Table 2). Using

258 tandem MS/MS, we identified and confirmed a total of 26 OS structures, including isomers

259 corresponding to 16 unique OS compositions, via analysis of characteristic fragments. The OS

260 compositions, retention time, and accurate mass for all samples are displayed in Table 2. The

261 identified OS ranged from 3 to 15 hexose (hex) or hexose-pentose (pent) combinations. The

262 relative abundances, monitored as peak areas, of 26 OS were averaged for each set of replicates.

263 The peak areas of selected OS are shown in Figure 2 and Figure 3, and the data for all the OS in

264 coffee beans are shown in Figures S1–S3. In general, hexose OS exhibited a greater level of

13

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 14 of 37

265 structural diversity with a wider degree of polymerization (DP) range and more isomeric

266 structures with higher abundance, compared with hexose-pentose OS. A recent model study also

267 reported the presence of hybrid hexose-pentose OS and suggested that the production of hybrid

268 structures was due to non-enzymatic transglycosylation reactions, which may involve the

269 arabinose side chains of coffee arabinogalactan.5

270 The OS profiles of the coffee samples varied among different roasting levels. Green bean

271 samples only contained simple OS of (Hex)3 and (Hex)4, whereas OS with up to 15

272 monosaccharide units were found in roasted coffee samples (Table 2). Figure 2 presents the

273 relative abundance of selected small hexose OS (DP 3 to 4). These small OS were more often

274 present in green beans, and light to light medium roasted beans (Table 2). Three OS, i.e. (Hex)3

275 isomer1&3, and (Hex)4 isomer1, were found both in green and roasted beans, and their relative

276 abundances were higher in green beans (Fig. 2). Some unique structures, i.e. (Hex)3 isomer 2,

277 and (Hex)4 isomer 3 &4, were only found in green beans and were in moderate abundance.

278 (Hex)4 isomer 2 and (Hex)3–Pent were only present in light to light medium samples (R1, R2 and

279 R3). Arabinogalactans are known to be more susceptible to thermal degradation during roasting,

280 indeed, a 10 to 20-fold decrease in the molecular weight of arabinogalactans was noticed even

281 after a light roast.2 The small OS found in roasted coffee beans, particularly those

282 pentose-containing structures, are possibly derived from the removal of arabinose side chains,

283 the fission of the galactose backbone, followed up by the non-enzymatic transglycosylation

284 reactions.5 These small OS can be rapidly degraded during the roasting process, resulting in their

285 low abundance in green or light roasted beans.1, 10

286 Roasting levels appeared to have a direct effect on the OS of DP 5-15. Except for the (Hex)5

287 isomer 1, only present in very dark roasted beans (R7), and the (Hex)4-Pent only present in

14

ACS Paragon Plus Environment


Page 15 of 37 Journal of Agricultural and Food Chemistry

288 R5-R7, the OS made of 5 to 13 monomers were present in all roasted coffee beans. However, a

289 higher number of larger OS structures was associated with more intense roasting. (Hex)14 was

290 found in light medium to very dark roasted beans (R3–R7) and (Hex)15 was only found in

291 medium-very dark beans (R4–R7) (Table 2). A greater number of isomeric structures was found

292 for (Hex)5–(Hex)7, whereas only a single isomer was found for larger compounds with DP 8-15.

293 The changes in relative abundance of the selected OS made of 5–15 monomers in roasted beans

294 are shown in Figure 3. Overall, the relative abundance of individual OS increased with elevated

295 intensity of roasting, with (Hex)6 isomer 2 being the only exception. Moreover, an interesting

296 trend was noticed between the relative abundance and the degree of polymerization. The relative

297 abundance decreased from (Hex)5 (relatively high) to (Hex)9, and increased again from (Hex)10 to

298 (Hex)15..

299 Structurally, coffee arabinogalactans have a backbone of β-(1→3)-linked galactopyranosyl

300 residues frequently substituted at the O-6 position by side chains formed by 1→6-linked

301 β-galactosyl, or at O-3 position by α-arabinosyl, rhamnoarabinosyl and rhamnoarabinoarabinosyl,

302 and glucuronic acid residues.3, 30, 31 Coffee galactomannans have been described as linear

303 polysaccharides with a main backbone composed of β-(1→4)-linked mannopyranose residues,

304 which are sometimes interspersed with β -(1→4)-linked glucopyranose. Some acetylated

305 mannopyrannose residues in the backbones are also present in O-2 and or O-3 position. The side

306 chains usually occur at O-6, with side chains of a single α-(1→6)-linked D-galactose residue or a

307 single arabinose residue. 32, 33 A series of publications reported changes in coffee

308 polysaccharide structures during roasting.1, 2, 9, 34, 35 Nunes and Coimbra35 observed the

309 de-polymerization and de-branching of galactomannans and arabinogalactans, as well as

310 decreased size of the arabinosyl side chains in arabinogalactans with an increase in degree of

15

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 16 of 37

311 roasting. They subsequently identified structural modifications on galactomannan, including

312 those that involved in Maillard reactions, caramelization, isomerization, oxidation, and

313 decarboxylation, that were due to roasting process.9 Redgewell et al.2 investigated the degree and

314 nature of polysaccharide degradation at different roasting levels; they reported that

315 arabinogalactans and galactomannans were degraded up to 60% and 36%, respectively, after a

316 dark roasting. Although those studies only demonstrated the changes in polysaccharides, both the

317 fission of the backbone and cleavage of the side chains can lead to the generation of OS.

318 In the absence of commercial standards for the novel oligosaccharides identified in coffee, it

319 is difficult to achieve the quantification purpose by mass spectrometry. Gas chromatography

320 therefore was used to obtain the total oligosaccharides amount extrapolated from the total

321 monosaccharides amount, also reveal the presence of the individual constituent monosaccharides.

322 Table 3 illustrate the absolute amount of constituent monosaccharides and the extrapolated total

323 amount of OS (calculated as the sum of constituent monosaccharides amounts measured by GC)

324 in raw and roasted beans.

325 All six constituent monosaccharides were present in roasted beans (R1–R7) in various

326 amounts, as shown in Table 3. Although the amount of constituent monosaccharides varied,

327 galactose and mannose consistently dominated the compositions, with lower amounts of

328 arabinose and glucose, and trace amounts of rhamnose and xylose. Galactose and mannose were

329 possibly units of OS derived from the backbone of arabinogalactans and galactomannans.

330 Arabinose was possibly a monomer of OS derived from the arabinogalactans side chain. Glucose

331 may have derived either from the galactomannans or arabinogalactans backbone. Xylose was

332 also identified in another study and proposed to derive from mannose through isomerization and

333 oxidative decarboxylation during the roasting process.36 Nunes et al.31 reported the presence of

16

ACS Paragon Plus Environment


Page 17 of 37 Journal of Agricultural and Food Chemistry

334 rhamnoarabinosyl and rhamnoarabinoarabinosyl side chains in coffee arabinogalactan-proteins,

335 which could explain the low level of rhamnose found in the present study. There was a positive

336 correlation between the total amounts of OS and degree of roasting. ANOVA indicated a

337 significant difference among the total amount (mg/g coffee beans) of OS in beans roasted at

338 different levels (p = 3.37 × 10-6). This trend corresponded well with the individual OS identified

339 through mass spectrometry as their relative abundance increased with elevated degree of

340 roasting.

341 Galactose and mannose were consistently the dominant sugars throughout roasting.

342 Galactose concentrations remained relative high throughout all the roasting levels, whereas the

343 increase of mannose was more significant (Table 3) with the elevated roasting intensities. Recent

344 studies showed that coffee arabinogalactans and galactomannans undergo remarkably different

345 structural changes during roasting.2, 37 Arabinogalactans are particularly labile to thermal

346 degradation, and in our study they began to depolymerize after a light roasting. The OS derived

347 through the fission and debranching of arabinogalactans were generated after light roasting and

348 remained in high concentration with elevated roasting intensities. Considering that galactose

349 residues are the building blocks of coffee arabinogalactans, the high concentration of galactose at

350 all roasting levels can be explained. On the other hand, galactomannans were moderately

351 degraded even in dark roasted beans, and their extractability increased probably due to changes

352 in cell wall structures.2 The mannose-containing OS derived from galactomannan may slowly

353 start to appear under elevated level of roasting, resulting in a more significant increase in

354 absolute concentration.

355 Interestingly, the amount of arabinose in OS increased until light-medium roasting (R3) and

356 decreased with elevated intensity of roasting. Arabinose residues in arabinogalactans are

17

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 18 of 37

357 sensitive to thermal degradation,12 and these residues are involved in the formation of

358 melanoidins38-40 and acids.28 Moreira et al.41 further demonstrated structural modifications of

359 α-(1→5)-l-arabinotriose induced by roasting, including dehydration, oxidation, and cleavage of a

360 carbon-carbon bond at the reducing end of OS, and reported the products of these modifications,

361 e.g. aldehydes, dialdehydes, and acids. Accordingly, the decreased amount of arabinose may

362 have resulted from the further structural modification caused by higher intensity of roasting.

363 The concentration of glucose in coffee OS was stable in beans roasted at all degrees, which is

364 in agreement with previous published results.2 OS of green beans did not contain arabinose,

365 rhamnose, or xylose; however, they contained high amounts of glucose.

366 Previous work reported high amounts of sucrose, up to 90 mg/g, in green coffee beans

367 (especially in Coffea arabica species).3, 42 Although porous graphitized carbon cartridges were

368 used to purify OS and remove residual sucrose and glucose, a small amount may have remained.

369 A commercial sucrose/D-glucose/D-fructose kit was used to quantify free glucose, fructose, and

370 sucrose in the purified green coffee bean extract, and only 0.47 mg/g sucrose and 0.15 mg/g free

371 D-glucose were present. This amount could not explain the high amount (2.039±0.042 mg/g

372 coffee bean) of glucose in green coffee bean samples. In the present work, we found green beans

373 contained two unique hexose OS isomers, (Hex)4 isomer 3 and 4, with moderate relative

374 abundance. The beans also contained a high abundance of (Hex)3 isomer 3. The higher

375 concentration of glucose may be due to the presence of those short chain hexose OS.

376 Preparative-scale Purification of Coffee OS. The discovery of potential prebiotic activity

377 requires generation of an adequate amount of OS of interest in high purity in order to test the

378 activity. Several methods, such as liquid-liquid extraction, ion-exchange resin and size-exclusion

379 chromatography, activated carbon chromatography, and nano filtration have been developed to

18

ACS Paragon Plus Environment


Page 19 of 37 Journal of Agricultural and Food Chemistry

380 purify and enrich OS from food mixtures. 43-45 Among those, activated carbon chromatography

381 was the most efficient purification method and has been used in a wide spectrum of applications

382 to remove impurities and concentrate OS. For example, it has been tested for separating

383 maltopentaose from mixtures of simple sugars and other maltooligosaccharides,46 isolating

384 hexose OS from honey,47 purifying fructooligosaccharides from a fermentative broth,45 and etc.

385 The purification consists of three steps: absorption of OS onto the activated carbon, washing the

386 column with water, and desorption of OS by solvents, e.g. ethanol or acetonitrile in water

387 solution.

388 Following characterization of the OS present in roasted coffee, we performed a preparative

389 scale purification to obtain adequate pure OS to perform in vitro tests of probiotic activity.

390 Because coffee roasted to the highest intensity (R7, Agtron # 25) was the most abundant source

391 of OS, the R7 sample was selected for extraction and purification of OS in preparative scale for

392 testing. In-house columns packed with C8-silica gel base octyl bonding material successfully

393 removed proteins, peptides, and some browning pigments and activated carbon treatment of

394 samples successfully concentrated the coffee OS. C8 column fractionation yielded three 50-mL

395 fractions. The first 50-mL fraction was completely transparent, whereas the second and third

396 fractions were slightly brown. The LC-MS results revealed that OS were primarily in the first

397 two fractions, whereas trace levels were found in the third fraction. However, all fractions after

398 purification on the C8 column still contained some impurities and therefore were subject to

399 additional activated carbon absorption of OS for further purification.

400 After incubating the purified OS samples with activated carbon, there was no detectible OS

401 remaining in AC eluate fraction 1 and 2. After desorption of the OS with acetonitrile/water

402 solution (resulting in AC eluate fraction 3), the final solution contained 7.807±0.376 mg OS as

19

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 20 of 37

403 measured by a total carbohydrate assay kit conducted according to the instructions. No glucose

404 or sucrose was detected in the final OS-containing solution as confirmed by a commercial

405 sucrose/D-glucose/D-fructose assay kit.

406 Activated carbon successfully removed impurities while absorbing the OS. The brown color

407 of the final OS-containing solution possibly derived from a small amount of either polyphenols

408 or melanoidins that remained even after two stages of solid-phase extraction. Melanoidins are the

409 final products of the Maillard browning reactions that occur during roasting.39 Melanoidins

410 account for up to 29% (w/w) of dry matter in brewed coffee.39 The chemical structures of

411 melanoidins are largely unknown, yet several researchers proposed that melanoidins mainly

412 consist of the degradation products of carbohydrates formed in the early stages of the Maillard

413 reaction; these degradation products are polymerized through aldol-type condensation and may

414 also be linked by amino compounds.39, 48-50 The incorporation of OS into melanoidin without

415 degradation of the glycosidic bonds during the Maillard reaction has been reported.48 A few

416 attempts have been performed to evaluate the implication of melanoidins on gut microbiota.

417 Ames et al. demonstrated that model melanoidins increased the counts of bacteroides and

418 clostridia, and bifidobacteria from 6 h to 24 h incubation through the in vitro study.51 Jaquet et al.

419 reported the increase in the population and the metabolic activity of Bifidobacterium spp. in a

420 3-week coffee consumption clinical trial on healthy adult volunteers, which may result from the

421 implication of melanoidins metabolism by the gut microbiota. 17 However, melanoidins in coffee

422 were possibly present with OS or polysaccharides and therefore the clinical study result cannot

423 be directly related to the effect of melanoidins.

424 The characterization of OS in roasted coffee beans in this study highlighted coffee as a

425 putative source of functional OS. This study explored the difference in OS profiles among beans

20

ACS Paragon Plus Environment


Page 21 of 37 Journal of Agricultural and Food Chemistry

426 roasted to different levels. The size, numbers of structural isomers, and abundance of OS were all

427 elevated with increased intensity of roasting. Coffee OS were characterized with a wide range of

428 DP with various constituent monosaccharides and possibilities of branching, all of which

429 provided the basis for selective simulation of a specific gut microbiota species. Furthermore, our

430 results suggest that the formation and extraction of diverse OS structures can be optimized by

431 manipulating the degree of roasting. OS purification using an in-house packing column allowed

432 us to isolate sufficient amounts for future testing of bioactivity. The application of active carbon

433 OS adsorption especially aided removal of impurities and concentration of target OS. Other

434 techniques, such as membrane filtration, could also be applied or combined with activated

435 carbon in future studies to further investigate coffee OS and their potential prebiotic activities.

436 ACKNOWLEDGEMENTS

437 The authors acknowledge financial support from the Keurig Green Mountain, Inc., Waterbury,

438 Vermont USA, and thank C. J. Dillard for editing this manuscript.

439 ABBREVIATIONS USED

440 OS, oligosaccharides; DP, degree of polymerization; LC, liquid chromatography; 2’-FL,

441 2’-fucosyllactose; QToF, quadrupole time-of-fight; GC, gas chromatography; FID, flame

442 ionization detector; MRS, De Man-Rogose-Sharp.

443 CONFLICT OF INTEREST

444 MC was an employee of Keurig Green Mountain, Inc. at the time the work was performed. The

445 remaining authors declare that the research was conducted in the absence of any commercial

446 relationships that could be construed as a potential conflict of interest.

447

21

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 22 of 37

448 REFERENCES

449 1. Redgwell, R.; Fischer, M. Coffee carbohydrates. Braz. J. Plant Physiol. 2006, 18, 165-174.

450 2. Redgwell, R. J.; Trovato, V.; Curti, D.; Fischer, M. Effect of roasting on degradation and

451 structural features of polysaccharides in Arabica coffee beans. Carbohydr. Res. 2002, 337,

452 421-431.

453 3. Bradbury, A. G. W., Chemistry I: Non-volatile Compounds. In Coffee: recent developments,

454 Clarke, R.; Vitzthum, O., Eds. Blackwell Science: Oxford, 2008; pp 1-17.

455 4. Moreira, A. S.; Nunes, F. M.; Simões, C.; Maciel, E.; Domingues, P.; Domingues, M. R. M.;

456 Coimbra, M. A. Transglycosylation reactions, a main mechanism of phenolics incorporation in

457 coffee melanoidins: Inhibition by Maillard reaction. Food Chem. 2017, 227, 422-431.

458 5. Moreira, A. S.; Simões, J.; Nunes, F. M.; Evtuguin, D. V.; Domingues, P.; Coimbra, M. A.;

459 Domingues, M. R. M. Nonenzymatic transglycosylation reactions induced by roasting: new

460 insights from models mimicking coffee bean regions with distinct polysaccharide composition. J.

461 Agric. Food Chem. 2016, 64, 1831-1840.

462 6. Oosterveld, A.; Harmsen, J.; Voragen, A.; Schols, H. Extraction and characterization of

463 polysaccharides from green and roasted Coffea arabica beans. Carbohydr. Polym. 2003, 52,

464 285-296.

465 7. Eggleston, G.; Cote, G. L., Oligosaccharides in food and agriculture. In Oligosaccharides in

466 food and agriculture, Eggleston, G.; Cote, G. L., Eds. American Chemical Society: Washington,

467 DC, 2003; pp 1-14.

468 8. Knopp, S.; Bytof, G.; Selmar, D. Influence of processing on the content of sugars in green

469 Arabica coffee beans. Eur. Food Res. Technol. 2006, 223, 195.

22

ACS Paragon Plus Environment


Page 23 of 37 Journal of Agricultural and Food Chemistry

470 9. Nunes, F. M.; Reis, A.; Domingues, M. R. M.; Coimbra, M. A. Characterization of

471 galactomannan derivatives in roasted coffee beverages. J. Agric. Food Chem. 2006, 54,

472 3428-3439.

473 10. Arya, M.; Rao, L. J. M. An impression of coffee carbohydrates. Crit. Rev. Food Sci. Nutr.

474 2007, 47, 51-67.

475 11. Tian, T.; Freeman, S.; Corey, M.; German, J. B.; Barile, D. Chemical characterization of

476 potentially prebiotic oligosaccharides in brewed coffee and spent coffee grounds. J. Agric. Food

477 Chem. 2017, 65, 2784-2792.

478 12. Oosterveld, A.; Voragen, A.; Schols, H. Effect of roasting on the carbohydrate composition

479 of Coffea arabica beans. Carbohydr. Polym. 2003, 54, 183-192.

480 13. Gibson, G. R.; Probert, H. M.; Van Loo, J.; Rastall, R. A.; Roberfroid, M. B. Dietary

481 modulation of the human colonic microbiota: updating the concept of prebiotics. Nutr. Res. Rev.

482 2004, 17, 259-275.

483 14. Barreteau, H.; Delattre, C.; Michaud, P. Production of oligosaccharides as promising new

484 food additive generation. Food Technol. Biotechnol. 2006, 44, 323.

485 15. Ryan, J. T.; Slattery, H.; Hickey, R. M.; Marotta, M. Bovine milk oligosaccharides as

486 anti-adhesives against the respiratory tract pathogen Streptococcus pneumoniae. Int. Dairy J.

487 2018.

488 16. Asano, I.; Hamaguchi, K.; Fujii, S.; IINO, H. In vitro digestibility and fermentation of

489 mannooligosaccharides from coffee mannan. Food Sci. Technol. Res. 2003, 9, 62-66.

490 17. Jaquet, M.; Rochat, I.; Moulin, J.; Cavin, C.; Bibiloni, R. Impact of coffee consumption on

491 the gut microbiota: a human volunteer study. Int. J. Food Microbiol. 2009, 130, 117-121.

23

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 24 of 37

492 18. Nakayama, T.; Oishi, K. Influence of coffee (Coffea arabica) and galacto-oligosaccharide

493 consumption on intestinal microbiota and the host responses. FEMS Microbiol. Lett. 2013, 343,

494 161-168.

495 19. Almeida, A. A. P.; Farah, A.; Silva, D. A.; Nunan, E. A.; Glória, M. B. A. Antibacterial

496 activity of coffee extracts and selected coffee chemical compounds against enterobacteria. J.

497 Agric. Food Chem. 2006, 54, 8738-8743.

498 20. Somporn, C.; Kamtuo, A.; Theerakulpisut, P.; Siriamornpun, S. Effects of roasting degree on

499 radical scavenging activity, phenolics and volatile compounds of Arabica coffee beans (Coffea

500 arabica L. cv. Catimor). Int. J. Food Sci. Technol. 2011, 46, 2287-2296.

501 21. Corso, M. P.; Vignoli, J. A.; de Toledo Benassi, M. Development of an instant coffee

502 enriched with chlorogenic acids. J. Food Sci. Technol. 2016, 53, 1380-1388.

503 22. Barile, D.; Marotta, M.; Chu, C.; Mehra, R.; Grimm, R.; Lebrilla, C.; German, J. Neutral and

504 acidic oligosaccharides in Holstein-Friesian colostrum during the first 3 days of lactation

505 measured by high performance liquid chromatography on a microfluidic chip and time-of-flight

506 mass spectrometry. J. Dairy Sci. 2010, 93, 3940-3949.

507 23. Ayestarán, B.; Guadalupe, Z.; León, D. Quantification of major grape polysaccharides

508 (Tempranillo v.) released by maceration enzymes during the fermentation process. Anal. Chim.

509 Acta 2004, 513, 29-39.

510 24. Moss, J.; Otten, L. A relationship between colour development and moisture content during

511 roasting of peanuts. Can. Inst. Food Sci. Technol. J. 1989, 22, 34-39.

512 25. Bicho, N. C.; Leitão, A. E.; Ramalho, J. C.; Lidon, F. C. Use of colour parameters for roasted

513 coffee assessment. Food Sci. Technol. (Campinas) 2012, 32, 436-442.

24

ACS Paragon Plus Environment


Page 25 of 37 Journal of Agricultural and Food Chemistry

514 26.Ramalakshmi, K.; Kubra, I. R.; Rao, L. J. M. Physicochemical characteristics of green coffee:

515 comparison of graded and defective beans. J. Food Sci. 2007, 72.

516 27. Mwithiga, G.; Jindal, V. Changes in properties of coffee brew due to roasting. World Appl.

517 Sci. J. 2007, 2, 527-535.

518 28. Ginz, M.; Balzer, H. H.; Bradbury, A. G.; Maier, H. G. Formation of aliphatic acids by

519 carbohydrate degradation during roasting of coffee. Eur. Food Res. Technol. 2000, 211, 404-410.

520 29. Goh, Y. J.; Klaenhammer, T. R. Genetic mechanisms of prebiotic oligosaccharide

521 metabolism in probiotic microbes. Annual review of food science and technology 2015, 6,

522 137-156.

523 30. Redgwell, R. J.; Curti, D.; Fischer, M.; Nicolas, P.; Fay, L. B. Coffee bean arabinogalactans:

524 acidic polymers covalently linked to protein. Carbohydr. Res. 2002, 337, 239-253.

525 31. Nunes, F. M.; Reis, A.; Silva, A. M.; Domingues, M. R. M.; Coimbra, M. A.

526 Rhamnoarabinosyl and rhamnoarabinoarabinosyl side chains as structural features of coffee

527 arabinogalactans. Phytochemistry 2008, 69, 1573-1585.

528 32. Nunes, F. M.; Domingues, M. R.; Coimbra, M. A. Arabinosyl and glucosyl residues as

529 structural features of acetylated galactomannans from green and roasted coffee infusions.

530 Carbohydr. Res. 2005, 340, 1689-1698.

531 33. Navarini, L.; Gilli, R.; Gombac, V.; Abatangelo, A.; Bosco, M.; Toffanin, R. Polysaccharides

532 from hot water extracts of roasted Coffea arabica beans: isolation and characterization.

533 Carbohydr. Polym. 1999, 40, 71-81.

534 34. Nunes, F. M.; Coimbra, M. A. Chemical characterization of the high molecular weight

535 material extracted with hot water from green and roasted arabica coffee. J. Agric. Food Chem.

536 2001, 49, 1773-1782.

25

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 26 of 37

537 35. Nunes, F. M.; Coimbra, M. A. Chemical characterization of galactomannans and

538 arabinogalactans from two arabica coffee infusions as affected by the degree of roast. J. Agric.

539 Food Chem. 2002, 50, 1429-1434.

540 36. Moreira, A. S.; Simões, J.; Pereira, A. T.; Passos, C. P.; Nunes, F. M.; Domingues, M. R. M.;

541 Coimbra, M. A. Transglycosylation reactions between galactomannans and arabinogalactans

542 during dry thermal treatment. Carbohydr. Polym. 2014, 112, 48-55.

543 37. Leloup, V.; Liardon, R. In Analytical characterization of coffee carbohydrates, COLLOQUE

544 SCIENTIFIQUE INTERNATIONAL SUR LE CAFE, 1993; ASIC ASSOCIATION

545 SCIENTIFIQUE INTERNATIONALE: 1993; pp 863-863.

546 38. Bekedam, E. K.; Loots, M. J.; Schols, H. A.; Van Boekel, M. A.; Smit, G. Roasting effects

547 on formation mechanisms of coffee brew melanoidins. J. Agric. Food Chem. 2008, 56,

548 7138-7145.

549 39. Moreira, A. S.; Nunes, F. M.; Domingues, M. R.; Coimbra, M. A. Coffee melanoidins:

550 structures, mechanisms of formation and potential health impacts. Food Funct. 2012, 3, 903-915.

551 40. Nunes, F. M.; Cruz, A. C.; Coimbra, M. A. Insight into the mechanism of coffee melanoidin

552 formation using modified “in bean” models. J. Agric. Food Chem. 2012, 60, 8710-8719.

553 41. Moreira, A. S.; Coimbra, M. A.; Nunes, F. M.; Domingues, M. R. M. Roasting-induced

554 changes in arabinotriose, a model of coffee arabinogalactan side chains. Food Chem. 2013, 138,

555 2291-2299.

556 42. Murkovic, M.; Derler, K. Analysis of amino acids and carbohydrates in green coffee. J.

557 Biochem. Biophys. Methods 2006, 69, 25-32.

26

ACS Paragon Plus Environment


Page 27 of 37 Journal of Agricultural and Food Chemistry

558 43. Hernández, O.; Ruiz-Matute, A. I.; Olano, A.; Moreno, F. J.; Sanz, M. L. Comparison of

559 fractionation techniques to obtain prebiotic galactooligosaccharides. Int. Dairy J. 2009, 19,

560 531-536.

561 44. Botelho-Cunha, V. A.; Mateus, M.; Petrus, J. C.; de Pinho, M. N. Tailoring the enzymatic

562 synthesis and nanofiltration fractionation of galacto-oligosaccharides. Biochem. Eng. J. 2010, 50,

563 29-36.

564 45. Nobre, C.; Teixeira, J.; Rodrigues, L. Fructo-oligosaccharides purification from a

565 fermentative broth using an activated charcoal column. New Biotechnol. 2012, 29, 395-401.

566 46. Lee, J.-W.; Kwon, T.-O.; Moon, I.-S. Adsorption of monosaccharides, disaccharides, and

567 maltooligosaccharides on activated carbon for separation of maltopentaose. Carbon 2004, 42,

568 371-380.

569 47. Morales, V.; Sanz, M.; Olano, A.; Corzo, N. Rapid separation on activated charcoal of high

570 oligosaccharides in honey. Chromatographia 2006, 64, 1-6.

571 48. Cämmerer, B.; Jalyschko, W.; Kroh, L. W. Intact carbohydrate structures as part of the

572 melanoidin skeleton. J. Agric. Food Chem. 2002, 50, 2083-2087.

573 49. Kato, H.; Tsuchida, H. Estimation of melanoidin structure by pyrolysis and oxidation. Prog.

574 Food Nutr. Sci. 1981, 5, 147-156.

575 50. Yaylayan, V.; Kaminsky, E. Isolation and structural analysis of Maillard polymers: caramel

576 and melanoidin formation in glycine/glucose model system. Food Chem. 1998, 63, 25-31.

577 51. Ames, J. M.; Wynne, A.; Hofmann, A.; Plos, S.; Gibson, G. R. The effect of a model

578 melanoidin mixture on faecal bacterial populations in vitro. British Journal of Nutrition 1999, 82,

579 489-495.

580

27

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 28 of 37

581

Figure Legends

Figure 1. The workflow for preparative scale production of coffee OS.

Figure 2. Relative abundances, expressed as peak areas, of small hexose OS (DP3~4) present in

green beans and/or roasted beans, as analyzed by LC-NanoChip-QToF.

Figure 3. Relative abundances, expressed as peak areas, of selected medium-large hexose OS

(DP5~15) present in roasted beans, as analyzed by LC-NanoChip -QToF.

28

ACS Paragon Plus Environment


Page 29 of 37 Journal of Agricultural and Food Chemistry

Table 1. CIE L*a*b*, Chroma, and Hue Color Characteristics of the Green Beans and Ground Roasted Coffee Samples (R1–R7), and

pH and total dissolved of Their Liquid Coffee Extracts

brewed extract
dry ground bean characteristics
characteristics

Sample1 total

roast Agtron dissolved

classification2 number L* a* b* Chroma3 Hue4 pH solids5

green
raw NA 56.41±0.90 h 1.69±0.06 h 17.30±0.21 h 17.38±0.21 h 1.47±0.00 h 5.73±0.00 h 1978±101 h
beans

R1 light 90 25.36±0.11 i 9.79±0.11 i 13.21±0.17 i 16.44±0.20 i 0.93±0.00 i 4.82±0.04 i 2187±31 hi

R2 moderately light 75 25.20±0.16 i 9.27±0.04 j 12.49±0.22 j 15.55±0.20 j 0.93±0.01 i 4.86±0.01 i 2158±41 hi

R3 light medium 65 23.01±0.15 j 8.63±0.06 k 9.87±0.11 k 13.11±0.12 k 0.85±0.00 j 4.99±0.01 j 2380±10 i

R4 medium 50 19.27±0.18 k 5.95±0.09 l 5.02±0.12 l 7.78±0.14 l 0.70±0.01 k 5.08±0.00 k 2337±46 i

R5 moderately dark 40 18.76±0.03 l 5.34±0.25 m 4.26±0.06 m 6.84±0.16 m 0.67±0.03 l 5.16±0.03 l 2201±3 i

R6 dark 35 18.19±0.07 kl 5.01±0.11 n 3.70±0.18 n 6.23±0.19 n 0.64±0.01 k 5.20±0.01 l 2338±96 hi

R7 very dark 25 17.05±0.04 m 3.68±0.04 o 2.04±0.02 o 4.21±0.04 o 0.51±0.00 m 5.32±0.00 m 2352±201 i

1
Samples were ground beans; green beans were from the same batch as roasted beans. Data represents the mean of n = 3.
2
Bulk roast classification is based on the Agtron Scores.

29

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 30 of 37

3
Chroma = (a*2+b*2)1/2
4
Hue = arctan (b*/a*).
5
Total dissolved solids (ppm) were measured at 18 °C.

Numbers followed by the same letter, within a column, are not significantly different (p > 0.05).

30

ACS Paragon Plus Environment


Page 31 of 37 Journal of Agricultural and Food Chemistry

Table 2. OS found in green beans and seven roasted beans.

retention accurate exact mass Presence in individual samples

OS time mass mass error2 green

compositions1 (minutes) (g/mol) (g/mol) (ppm) beans R1 R2 R3 R4 R5 R6 R7

(Hex)2Pent 4.39 474.158 474.158 0.84    

(Hex)3 isomer1 6.81 504.169 504.169 0.00        

(Hex)3 isomer2 11.24 504.171 504.169 3.97 

(Hex)3 isomer3 14.42 504.180 504.169 21.82    

(Hex)3Pent 15.28 636.217 636.211 9.43   

(Hex)4 isomer1 11.23 666.223 666.222 1.50        

(Hex)4 isomer2 12.49 666.223 666.222 1.50   

(Hex)4 isomer3 12.34 666.222 666.222 0.00 

(Hex)4 isomer4 14.83 666.222 666.222 0.00 

(Hex)4Pent 14.08 798.266 798.264 2.51   

(Hex)5 isomer1 13.42 828.275 828.275 0.00 

(Hex)5 isomer2 15.17 828.275 828.275 0.00       

31

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 32 of 37

(Hex)5 isomer3 16.74 828.275 828.275 0.00       

(Hex)5Pent 17.42 960.317 960.316 1.04   

(Hex)6 isomer1 16.36 990.327 990.328 1.01       

(Hex)6 isomer2 17.97 990.327 990.328 1.01       

(Hex)7 isomer1 18.14 1152.379 1152.380 0.87       

(Hex)7 isomer2 21.43 1152.378 1152.380 1.74       

(Hex)8 19.52 1314.431 1314.433 1.52       

(Hex)9 20.66 1476.483 1476.485 1.35       

(Hex)10 21.73 1638.538 1638.539 0.61       

(Hex)11 22.61 1800.593 1800.592 0.56       

(Hex)12 23.5 1962.645 1962.644 0.51       

(Hex)13 24.11 2124.700 2124.697 1.41       

(Hex)14 24.85 2286.749 2286.750 0.44     

(Hex)15 25.28 2448.795 2448.803 3.27    

1
Hex: hexose; Pent: Pentose; (Hex)n: OS composed of n hexose units.
2
mass error = (accurate mass-exact mass)/exact mass ×106.

32

ACS Paragon Plus Environment


Page 33 of 37 Journal of Agricultural and Food Chemistry

Table 3. Constituent Monosaccharides1 and Extrapolated Total OS in Beans Roasted at Different Intensities as Analyzed by Gas

Chromatography2

Extrapolated
sample arabinose rhamnose xylose mannose galactose glucose
total OS

green ND ND ND 0.249±0.013 f 0.443±0.041 e 2.039±0.042 a 2.730±0.013 d

R1 0.496±0.016 b 0.059±0.004 b 0.085±0.004 b 0.514±0.019 ef 1.380±0.086 cd 0.557±0.051 b 3.090±0.101 d

R2 0.450±0.082 b 0.058±0.008 b 0.115±0.036 ab 0.538±0.073 e 1.355±0.228 d 0.477±0.045 b 2.994±0.472 d

R3 0.658±0.012 a 0.107±0.018 a 0.172±0.010 a 0.998±0.020 d 2.059±0.055 ab 0.658±0.137 b 4.653±0.058 bc

R4 0.460±0.032 b 0.060±0.003 b 0.126±0.006 ab 1.214±0.015 cd 1.891±0.119 b 0.680±0.198 b 4.432±0.375 bc

R5 0.458±0.023 b 0.053±0.002 b 0.121±0.003 ab 1.335±0.073 bc 1.812±0.091 bc 0.526±0.074 b 4.306±0.118 c

R6 0.547±0.007 ab 0.066±0.002 b 0.152±0.012 a 1.572±0.129 b 2.400±0.130 a 0.569±0.028 b 5.305±0.248 b

R7 0.482±0.026 b 0.063±0.001 b 0.124±0.007 ab 2.597±0.093 a 2.492±0.016 a 0.671±0.126 b 6.429±0.014 a

1
Monosaccharides are expressed as mean mg/g coffee bean ± SD (n = 3); ND, not detected.
2
Numbers followed by the same letter, within a column, are not significantly different (p > 0.05).

33

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 34 of 37

Figure 1.

34

ACS Paragon Plus Environment


Page 35 of 37 Journal of Agricultural and Food Chemistry

Figure 2.

green
beans

35

ACS Paragon Plus Environment


Journal of Agricultural and Food Chemistry Page 36 of 37

Figure 3

36

ACS Paragon Plus Environment


Page 37 of 37 Journal of Agricultural and Food Chemistry

Table of content graph

37

ACS Paragon Plus Environment

Potrebbero piacerti anche