Extraction of Glucuronoarabinoxylan from Quinoa Stalks

Subscriber access provided by Columbia University Libraries

Article

Extraction of glucuronoarabinoxylan from quinoa stalks (Chenopodium quinoa Willd.) and evaluation of xylooligosaccharides produced by GH10 and GH11 xylanases Daniel Martin Salas-Veizaga, Rodrigo Villagomez, Javier A Linares-Pastén, Cristhian Álvaro Carrasco, Maria Teresa Álvarez, Patrick Adlercreutz, and Eva Nordberg-Karlsson J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.7b01737 • Publication Date (Web): 11 Aug 2017 Downloaded from http://pubs.acs.org on August 14, 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 43

Journal of Agricultural and Food Chemistry

1 1

Extraction of glucuronoarabinoxylan from quinoa stalks (Chenopodium quinoa Willd.) and

2

evaluation of xylooligosaccharides produced by GH10 and GH11 xylanases

3

4

Daniel Martin Salas-Veizaga1,2, Rodrigo Villagomez3, Javier A. Linares-Pastén1, Cristhian

5

Carrasco4, María Teresa Álvarez2, Patrick Adlercreutz1, Eva Nordberg Karlsson1*

6

7

*Corresponding Author: Eva Nordberg-Karlsson, Division of Biotechnology, Lund

8

University, Lund, Sweden. E-mail: [email protected]

9

10

1

11

Sweden

12

2

13

Box 3239 – La Paz, Bolivia

14

3

15

124, 22100 – Lund, Sweden

16

4

17

Andrés, P.O. Box 12958 – La Paz, Bolivia

Biotechnology, Department of Chemistry, Lund University, P.O. Box 124, 22100 – Lund,

Instituto de Investigaciones Fármaco Bioquímicas, Universidad Mayor de San Andrés, P.O.

Center of Analysis and Synthesis, Department of Chemistry, Lund University, P.O. Box

Instituto de Investigación y Desarrollo de Procesos Químicos, Universidad Mayor de San

18

19

Key words: Xylan, XOs, , Bacillus halodurans, Rhodothermus marinus, endo-xylanase.

20

ACS Paragon Plus Environment


Page 2 of 43

2 21

ABSTRACT

22

Byproducts from quinoa are not yet well explored sources of hemicellulose or products

23

thereof. In this work, xylan from milled quinoa stalks was retrieved to 66% recovery by

24

akaline extraction using 0.5 M NaOH at 80°C, followed by ethanol precipitation. The

25

isolated polymer eluted as a single peak in SEC with a molecular weight > 700 kDa.

26

Analysis by FT-IR and NMR, combined with acid hydrolysis to monomers, showed that the

27

polymer was built of a backbone of β-(1→4) linked xylose residues that were substituted

28

by 4-O-methylglucuronic acids, arabinose and galactose in an approximate molar ratio of

29

114:23:5:1. NMR analysis also indicated presence of α-(1→5) linked arabinose

30

substituents in dimeric or oligomeric forms.

31

The main xylooligosaccharides (XOs) produced after hydrolysis of the extracted

32

glucuronoarabinoxylan-polymer by thermostable glycoside hydrolases (GH) from families

33

10 and 11 were xylobiose and xylotriose, followed by peaks of putative substituted XOs.

34

Quantification of the unsubstituted XOs using standards, showed that the highest yield

35

from the soluble glucuronoarabinoxylan fraction was 1.26 g/100 g xylan fraction, only

36

slightly higher than the yield (1.00 g/100g xylan fraction) from the insoluble fraction

37

(p0.05).

38

This study shows that quinoa stalks represent a novel source of glucuronoarabinoxylan,

39

with a substituent structure that allow limited production of XOs by GH10 or GH11

40

enzymes.

41


Page 3 of 43


3 42

INTRODUCTION

43

Agricultural residues from quinoa (Chenopodium quinoa Willd.) have increased alongside

44

an increasing consumption of its highly nutritional grains. 1- 3 Bolivia, one of the major

45

quinoa exporters in the world, produces around 28 000 tons of quinoa annually.2

46

Consequently, similar or even higher quantities of agricultural byproducts are produced

47

per year, mainly seed coats (locally called ‘mojuelo’) and stalks.1 Quinoa stalks are

48

cylindrical and porous, and possess different phenotypes typical of the quinoa variety.4

49

The quinoa stalks are currently neither used as feed nor for any other purposes, and as a

50

consequence the stalks accumulate at the cultivation sites posing an agricultural waste

51

problem.2 However, these agricultural residues could emerge as a renewable resource for

52

lignin, cellulose and/or hemicellulose that potentially could be utilized if methodologies

53

for their extraction were developed, allowing valorization of the material.5

54

Hemicelluloses make up 20-50 % of the lignocellulosic biomass present on earth and are

55

chemically heterogeneous polymers formed by mainly pentoses (xylose, arabinose),

56

hexoses (mannose, glucose, galactose) and acetylated sugars.6,7 Based on structure,

57

hemicelluloses can be divided in D-xylans and D-xyloglucans, D-mannans and D-

58

(galacto)glucomannans, mixed-linkage β-D-glucans and L-arabino-D-galactans.8 Xylans, the

59

major hemicellulose component, have a backbone of β-(1→4)-linked xylose polymers, in

60

which the xylose residues are variably substituted with e.g. acetyl, arabinose and 4-O-

61

methylglucuronic acid groups.9

62

Xylans can be obtained from various agricultural raw materials, using different types of

63

extraction methods, including biological, chemical or physico/chemical treatments.10 Thus



Page 4 of 43

4 64

far, alkaline methods have resulted in the highest yields of the xylan fraction. Despite this

65

fact, other methods such as treatment by hot water/steam or enzymatic treatment

66

(laccase/cellulase-based) have lately gained interest, providing milder, environmentally

67

friendly methodologies.11-13

68

Xylans have several applications in industrial, chemical, and alimentary fields.8 In food and

69

feed related applications, xylooligosaccharides (XOs) have for example attracted interest

70

as components in feed formulations and as prebiotics in functional foods, being non-

71

digestible ingredients that beneficially affect the host by selective stimulation of one or a

72

limited number of beneficial bacteria in the gastrointestinal tract.14 XOs can be produced

73

from xylans by different methods, including chemical treatments, auto hydrolysis, or

74

enzymatic hydrolysis methods.15 XOs production by enzymatic hydrolysis possess certain

75

advantages including: a) suitable degree of polymerization (DP) of obtained products (not

76

too many monosaccharides),16 b) lack of undesirable and/or inhibitory by products (e.g.

77

furfural), and c) specific activity towards the target substrates.17 The enzymes used,

78

however, need to be stable under the processing conditions.18

79

In this study, the possibility of extracting xylan from quinoa stalks is evaluated, the

80

extracted polymer is analyzed and thermostable xylanases from glycoside hydrolase

81

families (GH) 10 and GH 11 are used to enzymatically degrade the polymer to XOs.

82


Page 5 of 43


5 83

MATERIALS AND METHODS

84

Biological Materials

85

Quinoa stalks

86

Fresh stalks from quinoa (C. quinoa Willd.), variety “Real Blanca”4 were collected from the

87

Bolivian Altiplano, close to Challapata (18°54’48.11”S; 66°46’43.86”W) and Santuario de

88

Quillacas (19°13’48.33”S; 66°56’13”W), Oruro - Bolivia. After collection, the stalks were

89

dried and chopped by a knife-milling machine. Milled stalks were further sieved (1 - 1.7

90

mm) using a portable sieve shaker and stored in a cold room (4°C) until its usage.

91

Enzymes for XOs production

92

Six different xylanase variants were evaluated for XOs production: five in house-produced

93

enzymes variants (BhXyn10AK80R, BhHXyn10A, and BhXyn10AH from Bacillus halodurans

94

and RmXyn10A, RmXyn10ACM from Rhodothermus marinus, Figure 1) and, one

95

commercial preparation (PENTOPAN®, Novozymes, Bagsvaerd, Denmark). All in house-

96

produced enzymes were produced in recombinant form in Escherichia coli BL21(DE3),

97

using expression plasmids from the pET-system (Novagen). The genes encoding

98

BhXyn10AK80R and BhXyn10AH from B. halodurans and,19,20 RmXyn10A, RmXyn10ACM

99

from R. marinus, 21-23 were cloned under the control of the T7-lac promoter as described

100

previously.19-23 BhHXyn10A (encoding an N-terminal His-tag) was cloned and produced for

101

the first time in this study. A synthetic gene with codons optimized for E. coli was

102

constructed using Optimizer Server (http://genomes.urv.es/OPTIMIZER/) and synthesized

103

by GenScript (NJ. USA). The synthetic gene was transferred from the propagation plasmid

104

pUC57::HXynAS to the expression vector pET28b between the restriction sites NdeI and



Page 6 of 43

6 105

NotI giving the plasmid pET28b::HXynAS, resulting in a recombinant BhHXyn10A fused

106

with an N-terminal hexa-histidine tag.

107

The enzyme variants were produced in shake flask cultivations using E. coli BL21(DE3)

108

induced by IPTG in the mid log phase. Each construct was purified from the respective cell

109

pellet by immobilized metal ion affinity chromatography, using the histidine tag included

110

in the cloning design, following the protocol described by Faryar et al. (2005).19

111

Extraction of xylan from quinoa stalks

112

Prior to any treatment, quinoa stalks were washed with several volumes of boiling water

113

to eliminate any saponins. This process was done until the water did not show any foam

114

(Figure 2).2 Saponins are known to be present in most parts of the quinoa plant, but have

115

not yet been quantified in stalks, although their presence is likely based on the observed

116

foaming.

117

Three methods were then tested for extraction of the major hemicellulose (xylan): hot

118

water extraction, enzyme aided extraction in aqueous solution, and alkaline extraction.

119

Hot water extraction was conducted according to the procedure by Immerzeel et al.24; 10

120

g of milled quinoa stalks were mixed with 150 mL of ultrapure water and were autoclaved

121

for 15 h at 121 °C. After autoclaving, the samples were twice centrifuged at 2700 g for 20

122

min. Supernatants were collected and pellets were resuspended in 50 mL of ultrapure

123

water and again centrifuged. Both supernatants were pooled and mixed 50:50 with 95 %

124

v/v ethanol to precipitate hemicellulose at 4 °C overnight The 50% ethanol concentration

125

was chosen (to reduce ethanol requirements) after comparison with samples precipitated

126

using 80% ethanol (v/v), and resulted in corresponding yields (data not shown). The


Page 7 of 43


7 127

precipitate was subsequently washed three times with 50 % v/v ethanol in all xylan

128

extraction methodologies. Finally, samples were centrifuged 29 600 g for 20 min to

129

separate the pellet from the supernatant; the pellet was cooled down to - 80 °C for 90 min

130

and finally freeze dried.

131

Enzyme aided extraction was carried out as described by Rico et al.13 A 3.75 g portion of

132

milled quinoa stalks was mixed with 112 mL of phosphate buffer (0.1 M, pH 6) and ABTS

133

(0.1 mM) in a 300 mL Erlenmeyer flask (three replicates were done). Laccase from

134

Aspergillus niger (Sigma-Aldrich) was added to a final concentration of 2.5 U/mg, and

135

incubated for 72 h, 37 °C under continuous shaking at 150 rpm. Supernatants were

136

separated, collected and kept at 4 °C. The pelleted fraction was again suspended in 112

137

mL phosphate buffer to which β - glucosidase (1 U/mg) (Sigma-Aldrich) and cellulase (0.25

138

U/mg) (Sigma-Aldrich) were added, followed by incubation 72 h, 37 °C at 150 rpm. The

139

supernatant was filtered and kept at 4 °C. The supernatants from laccase and β -

140

glucosidase/cellulase treatments were pooled and mixed, 50:50 with 95 % v/v ethanol and

141

left overnight 4 °C. The precipitate was washed three times with 50 % v/v ethanol,

142

centrifuged and freeze dried.

143

Alkaline extraction was conducted according to Sun et al.25; 10 g of milled quinoa stalks

144

were treated with 300 mL NaOH 2 % w/v (0.5 M) at 80 °C in a water bath for 90 min. The

145

samples were then cooled down and filtered through a cotton fabric filter, pH was

146

balanced to 5 with glacial acetic acid (99 %), and then the hemicellulose fraction was

147

precipitated overnight at 4 °C after an addition of ethanol 95 % to the supernatant to

148

reach 50 % v/v. The precipitate was washed with ethanol (50 % v/v) three times,



Page 8 of 43

8 149

centrifuged and freeze dried. Three consecutive extractions were done to evaluate the

150

efficiency of the extraction step. All extraction procedures were performed in triplicates.

151

Analyses of the xylan fraction extracted from quinoa stalks

152

Size-Exclusion Chromatography

153

Samples (2 mg/mL) were injected (10 µl) on a Shodex OH-Pak SB-806HQ column (8.0 mm x

154

300 mm, exclusion limit 20 000 kDa, target molecular weight range: 100 – 20 000 kDa)

155

eluted with ultra-pure water at 0.5 mL/min, at room temperature for 60 min. A refractive

156

index detector (Shodex RI-101, Showa Denko K.K., Japan) was used for detection. Pullulan

157

standards (Shodex® P-82, molecular weight range: 6 100 – 642 000 Da) prepared at 1

158

mg/mL concentration were used to calibrate the column and then estimate a “pullulan

159

equivalent” molecular weight (MW).

160

Monosaccharide content

161

From each freeze dried extract, 10 mg portions were weighed into tubes, and 175 µl

162

sulfuric acid 72 % w/w was added, followed by incubation for 60 min in a water bath at 30

163

°C. Afterwards, H2SO4 was diluted to 4 % w/w and the samples were heated to 100 °C for 3

164

h. Finally, the tubes were cooled down, vortexed and centrifuged for 1 min, 9 400 g at

165

room temperature.

166

The supernatant (1 mL) was transferred to small cup and pH was set to 5 with 0.1 M

167

Ba(OH)2. Then, the samples were centrifuged for 5 min at 2 700 g and filtered with a 0.2

168

µm membrane filter. Monosaccharides were separated and then quantified using High-

169

Performance Anion Exchange Chromatography with Pulse Amperometric Detection

170

(HPAEC-PAD) (Dionex, Sunnyvale, CA, USA) using a PA20 column and a mobile phase (0.5


Page 9 of 43


9 171

mL/min) of 0.75 mM NaOH, post column 200 mM NaOH for 23 min. Standards were used

172

to identify the peaks in the chromatograms, and contained arabinose, glucose, galactose,

173

xylose and mannose (Sigma-Aldrich) in a concentration range of 0.5 - 20 µg/mL.

174

Fourier Transform Infrared Spectroscopy analysis (FT-IR)

175

The freeze-dried hemicellulosic fraction alkaline-extracted from quinoa stalks was

176

analyzed by FT-IR. Triplicate samples were recorded in transmittance mode from 4000 to

177

500 cm-1 and spectra were obtained at a resolution of 4 cm-1. using Nicolet iS5

178

spectrometer- iD1 transmission accessory (Thermo Scientific, Madison, USA).

179

Nuclear Magnetic Resonance (NMR)

180

Nuclear Magnetic Resonance spectra in D2O were recorded with a Bruker DRX 400 MHz at

181

400 Mhz (1H) and at 100 MHz (13C). A total of 50 mg of extracted quinoa stalks xylan were

182

dissolved in 1 mL of D2O and 0.6 mL of the solution was transferred to an NMR sample

183

tube. Chemical shifts are given in ppm relative to TMS and using 10 µl of acetone as

184

internal standard for 0.6 mL of sample (2.225 ppm for 1H and 215.9 ppm 13C=O).

185

Bruker standard parameter sets with the pulse sequences according to the Bruker

186

nomenclature were used for recording 1D and 2D spectra: Proton Nuclear Magnetic

187

Resonance (1H-NMR/Pulse sequence: zg30), Carbon Nuclear Magnetic Resonance (13C-

188

NMR/zgdc30), H-H Correlation Spectroscopy (COSY/cosygpqf), H-H Total Correlation

189

Spectroscopy (TOCSY/mlevph), H-C Heteronuclear Multiple Quantum Coherence

190

(HMQC/hmqcgpqf), and H-C Heteronuclear Multiple Bond Correlation

191

(HMBC/hmbcgpqf).”

192



Page 10 of 43

10 193

Enzymatic XOs production and analysis

194

Buffer (100 µl, of either Glycine/NaOH pH 9 for subsequent B. halodurans xylanase

195

incubation or Phosphate/Na pH 7 for the R. marinus and PENTOPAN® xylanase

196

incubations) or ultrapure water (100 µl) was dispensed in microtitre plates, followed by

197

addition of 100 µl of extracted hemicellulose, suspended to 2 % w/v in ultrapure water to

198

reach 1 % final concentration. (The soluble and insoluble fractions of the extracted

199

hemicellulose were added to separate wells). Then, 1 µl of the respective enzyme (initial

200

concentration 1.2 mg/mL) was added, and incubated for 12 h at 60 °C. After the

201

incubation, the content of each well was transferred to 1.5 mL microtubes and heated for

202

5 min at 110 °C to stop the enzymatic reaction, followed by cooling on ice for 20 min and

203

centrifugation (5 min, 9500 g at room temperature) to collect the supernatant. All

204

reactions were performed in triplicate.

205

Production of XOs was analyzed by HPAEC-PAD (Dionex, Sunnyvale, CA, USA) using a 250

206

mm x 4 mm i.d., 8.5 µm, CarboPac PA200 column and guard column, 50 mm x 4 mm, of

207

the same material (Dionex), a mobile phase (0.5 mL/min) of constant 100 mM NaOH, and

208

a gradient of sodium acetate 0 - 20 min of 10 - 160 mM and 20 - 25 min of 160 - 400 mM.

209

Standards of xylose (X1) (Sigma-Aldrich), xylobiose (X2), xylotriose (X3), xylotetraose (X4),

210

xylopentaose (X5) and xylohexaose (X6) (Megazyme, Ireland) were made in a

211

concentration range of 0.5 - 20 µM as described by Falck et al.23

212


Page 11 of 43


11 213

Statistical analysis

214

Statistical analysis was conducted to detect possible differences between the yields of XOs

215

in buffer or water solution, and the yields of XOs from soluble or insoluble xylan, the

216

analysis were performed by the Wilcoxon non-parametric test using SYSTAT 11 Statistical

217

software.

218

219

RESULTS AND DISCUSSION

220

Extraction of xylan, as the major hemicellulose from quinoa stalks

221

Quinoa stalks has, to our knowledge, not previously been used as a source of polymeric

222

xylan. Hence, three methods were screened to choose the most feasible extraction

223

conditions for the raw material. The methods included: one enzymatic method using

224

laccase, β - glucosidase/cellulase (37 °C, 72 h), one hot water method using steam (121 °C,

225

15 h) and one alkaline method using 0.5 M NaOH (80 °C, 90 min). These methodologies

226

have previously been used to either improve xylan saccharification (i.e. in enzymatic

227

treatment of Eucalyptus residues11) or extract xylan polymers from other raw materials

228

(e.g. hot water/steam extraction of xylan from wheat bran24 and alkaline extraction of

229

xylan from wheat straw19).

230

Successful extraction of xylan from the milled stalks was only achieved by the alkaline

231

method, as verified by acid hydrolysis and monosaccharide analysis by HPAEC-PAD (Figure

232

3B). Hot water treatment at 121 °C did not result in release of xylan (as judged by lack of



Page 12 of 43

12 233

detectable monosaccharides, data not shown). A higher severity factor (increased

234

temperature and treatment time) or at least a higher temperature is probably necessary

235

to release xylan from the stalk material. An increase in extraction temperature (from 121

236

to 185 °C) did for example result in improved yield of xylan from wheat bran, despite

237

shorter exposure time and unchanged severity factor24. The laccase, β -glucosidase/

238

cellulase treatment also gave negative results (data not shown). This method had the

239

objective to break interactions between hemicellulose and lignin and loosen up cellulose

240

structures, thereby promoting release of hemicellulose. The negative result shows that it

241

is probably necessary to combine this method with other pretreatments to release xylan

242

from remaining lignin-carbohydrate complexes. In trials on Eucalyptus feedstock, enzyme

243

treatment was for example combined with peroxide and alkaline steps, which resulted in

244

increased overall yields.11

245

Hemicelluloses are proposed to be linked to lignin by ester or ether bonds, of which the

246

ester linkages are easily cleaved by alkali.26-28 Indeed, by alkaline extraction, xylan was

247

successfully obtained from the milled quinoa stalks. The dried xylan fraction formed a

248

white, soft, cotton-like structure (Figure 3A) and was, based on the composition of

249

monosaccharides (Figure 3B), apparently free of monosaccharides from other polymers

250

(only monosaccharides relevant for xylans were detected).

251

The combined yield after three consecutive extractions was on average 10.9 ± 0.4 g xylan

252

per 100 g of the dry raw material (Table 1A). Each extraction repetition resulted in

253

approximately 50 % reduction in the amount of extracted xylan, with average values (from

254

the 10 g portion) of 0.61 ± 0.03 g, 0.35 ± 0.02 g, and 0.12 ± 0.03 g of xylan. Quinoa stalks


Page 13 of 43


13 255

(from the same species variety) have a reported xylan content of 16.7 g per 100 g dry

256

matter (averaged from the xylan content at two collection sites),2 showing that the

257

current extraction procedure resulted in 66 % recovery. This is compatible to yields

258

reported for other types of raw materials such as rye straw, maize bran, and corn stover

259

(Table 2). To further increase xylan recovery, a higher concentration of NaOH has been

260

suggested,16 but was here avoided, as it can lead to formation of undesired byproducts

261

that reduce the purity of the xylan.

262

Characterization of the xylan from quinoa stalks

263

The extracted fraction eluted as a single peak in SEC, with a molecular weight significantly

264

higher than the pullulan standards (maximum 642 kDa). This is high compared to the Mw

265

reported from xylans from some other resources, but the range of molecular weights

266

observed also varies broadly. Arabinoxylan from wheat bran is for example reported to

267

have a Mw of 470 – 600 kDa,35 while alkaline extracted xylans from cornstover have a

268

reported Mw of 49 kDa.36

269

The polymer was further analyzed by a combination of monosaccharide analysis (by

270

HPAEC-PAD) after sulfuric acid hydrolysis, FT-IR and NMR. The monosaccharide content

271

was (mg/g xylan fraction): 610 ± 120 (xylose), 28 ± 31 (arabinose) and 7 ± 9 (galactose)

272

(Figure 3B), leading to an apparent molar ratio estimation of 114:5:1. The relatively large

273

variation in arabinose and galactose, was due to between-batch variation (triplicates of

274

three batches of alkaline extracted material was analysed, Figure 3B).

275

When also considering the glucuronic acid content (the latter estimated by NMR below, as

276

these residues cannot be detected using the sulfuric acid hydrolysis method) a good



Page 14 of 43

14 277

overall purity, of at least 77 %, was obtained directly from alkaline extraction, which

278

represents a bioprocessing advantage.

279

FT-IR analysis of the polymer (Figure 4) corroborated a glucuronoxylan structure. A major

280

peak, characteristic of 1→4 linked xylan,37, 86 was present at 1040 cm-1 and is reflecting C-

281

O and C-C stretching and/or C-O bending.39 In addition, a peak at 896 cm-1, is in the region

282

characteristic for the β-type of glycosidic linkage.39,40 Two additional peaks, at 1413 and

283

1603 cm-1 could, according to Capek & Matulova,40 be assigned to antisymmetric or

284

symmetric vibrations of deprotonated COO- of uronic acids.

285

NMR analysis, using 1D and 2D (COSY, HMQC, HMBC and TOCSY), revealed presence of

286

three major monosaccharides (xylose, 4-methyl-glucuronic acid and arabinose).

287

The anomeric proton identified at 4.48 ppm belonged to the same spin system as the

288

signals at 4.11, 3.77, 3.56, 3.37 and 3.28 (TOCSY correlations) and the corresponding

289

assignation were determined by COSY correlations (Table 3). The HMQC spectrum was

290

used to assign the 13C chemical shifts as shown in Table 3 and their assignments were

291

confirmed by HMBC correlations. These 1H and 13C chemical shifts are consistent with

292

those reported for the internal β-xylanopyranoses.41 Moreover, a strong HMBC correlation

293

between the H-1 (δ 4.48 ppm) and C-4 (δ 76.9 ppm) also indicate a (1→4) glycosidic bond

294

between the xylose units (Figure 5). Unfortunately, it was not possible to identify the

295

terminal and reducing ends, due to the high MW of the polymer.

296

In the same way, COSY and TOCSY experiments starting from the anomeric proton at 5.28

297

ppm, revealed the presence of five protons belonging to glucuronic acid. This was also

298

corroborated by the 13C chemical shifts in the HMQC and HMBC spectra (Table 3). The


Page 15 of 43


15 299

HMBC spectra also provided a strong correlation between the methyl group (δ 3.46 ppm)

300

and the C- 4 (δ 82.4 ppm), suggesting a methoxylation on position 4 (Figure 5). Therefore,

301

the results obtained in this study were compared with previous data reported for 4-O-

302

methylglucuronic acid and they were in agreement with the α epimer.42,43 It is important

303

to note that this monosaccharide was not detected by HPAEC-PAD, as these residues are

304

not liberated by the sulfuric acid hydrolysis conditions. The integration of the anomeric

305

signals at 4.48 ppm (internal xylose) and 5.28 ppm (glucuronic acid), suggests an

306

approximate ratio of 5:1, in a first estimation of the 4-O-methylglucuronic acid content in

307

the polymer, changing the overall molar ratio to 114:23:5:1 (xylose: 4-methylglucuronic

308

acid: arabinose: galactose).

309

The anomeric proton at 5.09 ppm corresponded to the third monosaccharide, for which

310

the identification and assignation of the 1H and 13C chemical shifts (Table 3) was done as

311

previously described. These chemical shifts are in good agreement with those reported for

312

α (1→5) glycosidic linkages between arabinofuranose residues,44 which was confirmed by

313

a strong HMBC correlation between the H-1 (δ 5.09 ppm) and C-5 (δ 67.4 ppm) (Figure 5),

314

suggesting that the arabinose is mainly branching as a dimer or larger oligosaccharide. An

315

alternative interpretation of the H-1 - C-5 correlation is to consider the pyranic isomer,

316

however, the chemical shifts45 and biosynthetic background46,47 are not consistent with

317

the pyranic form. Reports on oligomeric or α (1→5) - linked arabinofuranose substituents

318

to xylan polymers are not commonly found in literature. A few cases from dicotyledonous

319

plants however exist and include xylan from Sage (Salvia officinalis) for which

320

arabinofuranose residues are suggested to be located mostly at O-3 of xylose in the form



Page 16 of 43

16 321

of short branched or unbranched oligomeric chains that include 1→5 linkages (found as

322

the terminal, 1,3-, 1,5-and 1,3,5-linked) 40. Oligomeric arabinofuranoside substituents (1,2

323

linked) have also been reported from green leaves of Litsea glutinosa (Lauraceae)48,

324

showing that different oligomeric arabinofuranoside substitutions may be present in

325

different dicotyledonous species.

326

Consistently with the apparent low amounts of galactose reported by HPAEC-PAD, this

327

monosaccharide could not be detected by any NMR assay. Due to the large MW of the

328

extracted xylan, it was also not possible to determine the 4-O-Methylglucuronic acid

329

bonds or the arabinose bonds that connect the substituents to the xylose backbone.

330

Hence, to achieve ultimate proof of the bonding of the substituents to the xylan

331

backbone, further studies on oligomerized material would be necessary.

332

Despite this fact and based on the information found in this study, a dicotyledonous

333

glucuronoarabinoxylan structure, including oligomeric arabinosylation, can be suggested

334

as likely for the xylan extracted from the quinoa stalks. As alkaline extraction (here with

335

0.5 M NaOH) is known to not preserve acetyl groups in xylan,49 this type of substituent

336

cannot be determined and it is clear that other types of groups are aiding the water

337

solubility of the extracted polymer. The 4-O-methylglucuronic acids might be candidates

338

for this, explaining the high ratio of water soluble polymer (63 %), which was kept in the

339

first, second and third extraction repetition (Table 1B). The xylose content in the polymer,

340

is based on the composition analysis 80% (xylose per total monosaccharides and uronic

341

acids), which is in range with what is reported for cotton stalks and wheat straw (Table 2).

342


Page 17 of 43


17 343

Enzymatic production of XOs

344

Endo-xylanases from GH10 and GH11 were used to hydrolyse the soluble and insoluble

345

fractions of quinoa xylan into XOs. From GH10, five variants of two xylanases were used

346

(BhXyn10AK80R,19 BhHXyn10A, BhXyn10AH20 from B. halodurans and RmXyn10A22 and

347

RmXyn10ACM21,22 from R. marinus), of which BhHXyn10A, with an N-terminal tail of 16

348

amino-acids (SSGLVPRGSH) and a six His-residues, was produced and analyzed for the first

349

time (Figure 1). From GH11, the commercial xylanase PENTOPAN® was used.

350

The major products were in all cases xylobiose and xylotriose, although the

351

chromatograms from the HPAEC-analysis also showed distinct peaks that likely

352

correspond to substituted oligosaccharides (see example in Figure 6). Due to the lack of

353

standards, yields ( g XOs / g xylan polymer) could only be calculated for the non-

354

substituted XOs, and the relatively low overall yields, are judged to be a consequence of

355

the substituents (e.g. glucuronic acid, or oligomeric arabinosylation) that either hinder

356

efficient hydrolysis by individual endo-xylanases or result in substituted XOs that could not

357

be quantified. Low yields of XOs from xylan using single enzymes, is however not unusual.

358

It has for example been shown by Immerzeel et al that use of RmXyn10ACM on the

359

arabinoxylan extracted from wheat bran resulted in conversion of 3-5% of the material to

360

XOS.

361

The yields of nonsubstituted XOs were in the same range as previously reported for

362

enzymatic XOs-production using xylan extracted from stalks of other dicotyledons50 (e.g.

363

tobacco stalks, cotton stalks or sunflower stalks). The DP-ratio clearly differed between



Page 18 of 43

18 364

the enzymes, while variants of a single enzyme generally displayed small or no difference

365

in product profile (with the exception of BhHXyn10A, as explained below), Table 4 and 5.

366

Highest yields were shown using RmXyn10ACM on both soluble (1.26 ± 0.17 g total

367

XOs/100 g xylan) and insoluble (1.00 ± 0.05 g total XOs/100 g xylan) xylan fractions,

368

respectively. Resulting XOs were mainly X2 (major product) and X3 (Table 4 and 5). These

369

are the most common XOs produced by endo-xylanase hydrolysis (previously reported for

370

other enzyme from e.g. Aspergillus niger,50 A. foetidus,17 and Thermoascus auranticus51).

371

The degree of polymerization obtained reflects the sub-site interactions in the respective

372

enzyme, showing that the B. halodurans enzyme variants (BhXyna10AK80R and

373

BhXyn10AH) should have pronounced interactions in the -2 or +2 sub-sites (-2 is normally

374

well conserved in GH1052), while more distal substrate interactions may be weak or

375

missing as very low amounts of X3 were produced.

376

The different DP-range observed for XOs produced by BhHXyn10A (and lower total yield)

377

indicate that the N-terminally positioned His-tag change and may obstruct substrate

378

binding. RmXyn10A, RmXyn10ACM, and PENTOPAN® produced more X3, indicating

379

stronger interactions with -3 or +3 sub sites. The low amount of X1 produced by

380

PENTOPAN® show that this enzyme has a lower substrate affinity to -1 or +1 subsites than

381

to -2 or +2 sub sites, and it is also known that GH11 enzymes do not allow substitution in

382

the -1 or +1 binding sites.52


Page 19 of 43


19 383

The product profile (with predominant X2 and X3) is beneficial for uptake in

384

Bifidobacterium as well as in probiotic bacteria from the genus Lactobacillus, which have

385

previously been shown to take up and utilize XOs with DP 2-3.19,23,24,53

386

A small but significant difference (p

Extraction of Glucuronoarabinoxylan from Quinoa Stalks

Recommend Documents