Xylanases to Improve the Nutritional Value of High Fibre

Diets based on Corn and Wheat DDGS

PhD thesis

Mads Brøgger Pedersen

Department of Animal Science,

Aarhus University

&

Enzyme Development,

DuPont Industrial Biosciences

2015

iii

PREFACE

To meet the requirements for obtaining the PhD degree in Animal Science, this thesis

has been submitted to the Graduate School of Science and Technology, Aarhus University. The

original research presented in this thesis was carried out in the period of January 2012 to

December 2014. The PhD project was conducted as an Industrial PhD project in collaboration

between Molecular Nutrition and Cell Biology, Department of Animal Science, Aarhus

University and DuPont Industrial Biosciences, Brabrand, Denmark, and partially supported by

The Danish Ministry of Science, Innovation, and Higher Education. The work presented in this

thesis was carried out at the facilities of both Aarhus University and DuPont Nutrition

Biosciences. Additional experimental work was carried out during a 1½ month visit at the

Department of Food Chemistry and Phytochemistry, Karlsruhe Institute of Technology (KIT),

Germany.

I would particularly like to thank my university supervisors Dr. Helle Nygaard Lærke,

and Prof. Knud Erik Bach Knudsen, and my company supervisors Dr. Shukun Yu and Group

Manager Søren Dalsgaard for their professional guidance, discussion, and advice throughout

the project.

I would also like to thank Dr. Susan Arent, DuPont Industrial Biosciences, for repeated

scientific discussions and contribution, Dr. Luis Romero and Dr. Maria Walsh, Danisco Animal

Nutrition, for their contribution to the animal trials, and Head of Department Jens Frisbæk

Sørensen, DuPont Industrial Biosciences, for allowing me this exceptional opportunity of

working in the cross-field between industry and academia.

Prof. Mirko Bunzel is thanked for allowing me the opportunity of working in his

department at Karlsruhe Institute of Technology (KIT), Germany, and PhD student Judith

Schäfer is thanked for supporting me in the laboratory during my visit.

Stina Greis Handberg and Lisbeth Märcher, and the staff at the Intensive Research

Facilities, Department of Animal Science, are thanked for their outstanding contribution to the

animal trials and chemical analyses.

My colleagues and fellow PhD students at the Department of Animal Science and my

colleagues at DuPont Industrial Biosciences are also thanked for their assistance and positive

attitude throughout the past three years.

A special thank goes to my family and friends - and particularly to Karen for her never-

ending support.

Mads Brøgger Pedersen

Aarhus, December 2014

iv

Supervisors:

Senior Researcher Helle Nygaard Lærke, PhD, Department of Animal Science, Aarhus

University, Tjele, Denmark

Professor Knud Erik Bach Knudsen, PhD, Department of Animal Science, Aarhus University,

Tjele, Denmark

Senior Staff Scientist Shukun Yu, PhD, Enzyme Development, DuPont Industrial Biosciences,

Brabrand, Denmark

Group Manager, Søren Dalsgaard, Enzyme Development, DuPont Industrial Biosciences,

Brabrand, Denmark

Assessment committee:

Professor Anne S. Meyer, PhD, Department of Chemical and Biochemical Engineering,

Technical University of Denmark, Lyngby, Denmark

Professor Ruurd Zijlstra, PhD, Department of Agricultural, Food and Nutritional Science,

University of Alberta, Edmonton, Alberta, Canada

Associate Professor Jan Værum Nørgaard, PhD, Department of Animal Science, Aarhus

University, Tjele, Denmark (chair)

v

TABLE OF CONTENTS

PREFACE ....................................................................................................................................................................... iii

SUMMARY ................................................................................................................................................................... vii

SAMMENDRAG ......................................................................................................................................................... viii

ABBREVIATIONS ....................................................................................................................................................... ix

LIST OF PAPERS ......................................................................................................................................................... xi

INTRODUCTION .......................................................................................................................................................... 1

BACKGROUND ............................................................................................................................................................. 5

Distillers dried grains with solubles .............................................................................................................. 7

Inclusion of distillers dried grains with solubles in diets for pigs ................................................. 8

Non-starch polysaccharides in cereals ......................................................................................................... 9

Arabinoxylan ...................................................................................................................................................... 9

Arabinoxylan crosslinking .......................................................................................................................... 10

Cellulose ............................................................................................................................................................. 11

β-glucan .............................................................................................................................................................. 11

Difference in fibre complexity between corn- and wheat DDGS .................................................. 11

Non-starch polysaccharides in relation to feed for swine ................................................................... 12

Improving fibre utilization in swine by exogenous enzymes ............................................................ 13

Xylanases ........................................................................................................................................................... 13

Use of xylanase in animal feed ................................................................................................................... 15

METHODS OVERVIEW ........................................................................................................................................... 17

Near infrared reflectance spectroscopy ..................................................................................................... 19

Analysis of non-digestible carbohydrates ................................................................................................. 19

Reversed phase high performance liquid chromatography with ultraviolet detection .......... 21

Enzymatic degradation of DDGS in vitro .................................................................................................... 21

In vivo digestion model ..................................................................................................................................... 22

Principal component analysis ........................................................................................................................ 23

AIM AND HYPOTHESES ......................................................................................................................................... 25

PAPER I ........................................................................................................................................................................ 29

PAPER II ....................................................................................................................................................................... 43

PAPER III ..................................................................................................................................................................... 53

PAPER IV...................................................................................................................................................................... 63

vi

GENERAL DISCUSSION .......................................................................................................................................... 91

Variation in DDGS constituents across origins ........................................................................................ 93

Compositional variations between corn and wheat DDGS ................................................................. 94

Modification of corn DDGS during processing ......................................................................................... 95

Enzymatic degradability of DDGS ................................................................................................................. 96

In vitro degradation of DDGS...................................................................................................................... 96

In vivo digestibility of DDGS ....................................................................................................................... 97

Comparison between in vitro and in vivo degradation of DDGS ................................................... 99

Applied methodologies .................................................................................................................................. 100

CONCLUSIONS AND PERSPECTIVES ............................................................................................................. 103

Conclusions ......................................................................................................................................................... 105

Perspectives ....................................................................................................................................................... 106

REFERENCES .......................................................................................................................................................... 107

vii

SUMMARY

The increased feed prices for cereal grains and protein crops have forced the feed

industry to look for alternatives to the conventional feedstuffs. One such alternative is

distillers dried grain with solubles (DDGS), the major co-product from dry-grind based

ethanol production. The main challenges limiting the use of DDGS in feeds for non-ruminant

animals is the inferior protein quality, potential high mycotoxin content, compositional

variations, and high levels of non-starch polysaccharides (NSP).

The overall aim of this PhD project was to map the variation and differences in

especially NSP composition of these feedstuffs, and to improve the digestive utilization of

corn- and wheat DDGS (c- and wDDGS) for pigs by enzymatic degradation of the fibre matrix.

A total of 138 corn, wheat and mixed cereal DDGS samples were investigated for

compositional variability of common nutrients, whereas a detailed analysis of the NSP and

ester-linked ferulic acid dehydrodimer (DFA) and –trimer (TriFA) profile was performed on

63 DDGS samples. Analysis of samples from five different ethanol plants showed that the

individual plants were capable of producing cDDGS with an individual conserved

compositional characteristic. Furthermore, investigation of corn and corresponding DDGS

showed that the most readily degradable arabinoxylan (AX) from endosperm is likely

modified during DDGS processing, and that the DFAs and TriFAs are not modified, thus,

leaving the AX in DDGS less degradable than that of the parent grain. The higher cell wall

complexity in c- than wDDGS was illustrated by a higher proportion of insoluble NSP, higher

arabinose/xylose - and uronic acid/xylose-ratio, and more than five times the content of DFAs

and TrifAs in c- than wDDGS. The higher complexity of c- than wDDGS had large implications

of the enzymatic degradability and digestibility, in vitro and in vivo, as wDDGS was markedly

more degradable than cDDGS. Four different xylanases from glycoside hydrolase (GH) family

10 and -11 were investigated, in vitro, alone and in combination with protease and phytase.

The GH10 xylanase showed higher degradability than GH11 on cDDGS substrate, due to the

differences in substrate affinity between the two GH families. In addition, it was found both in

vitro and in vivo, that the depolymerizing of DDGS fibre depended on the substrate affinity of

the individual xylanases. Both GH10 and -11 treatments increased the apparent ileal

digestibility of NSPs and generated low molecular weight components with potential prebiotic

effects in pigs fed a wDDGS diet. The GH11 xylanase had numerically greater effect than the

GH10 xylanase, likely due to a higher stability in the upper gastro-intestinal tract, illustrated

by a four times higher recovery of enzyme activity in ileum.

In summary, the compositional variations and differences in NSP complexity between c-

and wDDGS had large implications on the enzymatic degradation of these two substrates. In

addition, in vitro and in vivo experiments illustrated that enzymes may comprise different

degradation efficiency in relation to their substrate affinity and stability towards the physical

environment occurring in the gastro intestinal tract of pigs. To further increase the enzymatic

degradation of DDGS, more research is needed to evidently identify the structural components

most accountable for the enzymatic recalcitrance of especially cDDGS.

viii

SAMMENDRAG

De øgede foderpriser på korn og protein-foder har tvunget foderindustrien til at lede

efter alternativer til de konventionelle foderstoffer. Et sådant alternativ er tørret bærme

(DDGS), det primære biprodukt fra kornbaseret produktion af ethanol. De største

begrænsninger ved brugen af DDGS i foder til enmavede dyr er den ringere proteinkvalitet,

potentielt højt indhold af mykotoksiner, variation i sammensætning og et højt niveau af ikke-

stivelses polysakkarider (NSP).

De overordnede formål med dette ph.d.-projekt var at kortlægge variationen af

nærringstoffer i DDGS foder med særligt fokus på NSP sammensætningen, og at forbedre

foderudnyttelsen af majs- og hvede baseret DDGS til svin ved enzymatisk nedbrydning af

fibermatricen.

I alt blev 138 DDGS prøver fra majs, hvede og blandede korn undersøgt for variation af

basale næringsstoffer, mens en mere detaljeret analyse af NSP og esterbundet ferulinsyre

dehydrodimer (DFA) og -trimer (TriFA) profiler blev udført på 63 DDGS prøver. Analyse af

prøver fra fem forskellige ethanol-producenter viste, at de enkelte producenter var i stand til

at producere majs DDGS med en individuel og karakteristisk sammensætningen. Analyse af

majs og DDGS viste, at den lettest nedbrydelige arabinoxylan (AX) fra endosperm er blevet

ændret igennem produktionen af DDGS og, at DFA’er og TriFA’er forbliver uændrede. Dette

betyder, at den tilbageværende AX i majs DDGS er mindre nedbrydelige end AX i majs.

Sammenlignet med hvede DDGS havde majs DDGS en mere kompleks cellevægstruktur,

illustreret ved en højere andel af uopløselig NSP, højere arabinose/xylose - og

uronsyre/xylose-ratio, og mere end fem gange større indhold af DFA’er og TriFA’er. Den

højere kompleksitet i majs end hvede DDGS havde store konsekvenser for den enzymatiske

nedbrydning og fordøjelighed, in vitro og in vivo, eftersom hvede DDGS var markant mere

nedbrydelige end majs DDGS. Fire forskellige xylanaser fra glycosidhydrolase (GH) familie 10

og -11 blev undersøgt in vitro, alene og i kombination med protease og phytase. GH10

xylanasen viste højere nedbrydning af majs DDGS end GH11 på grund af forskellene i

substrataffinitet mellem de to GH familier. Desuden blev det konstateret, både in vitro og in

vivo, at depolymeriseringen af DDGS fiber afhang af de individuelle xylanasers

substrataffinitet. Både GH10 og -11 øgede den ileale fordøjelighed af NSP'er og genererede

lavmolekylære komponenter med potentielle prebiotiske effekter i grise fodret med en hvede

DDGS diæt. GH11 xylanasen havde numerisk større effekt end GH10 xylanasen, sandsynligvis

på grund af en højere stabilitet i den øvre mave-tarm-kanal.

Samlet set viste forskellen i NSP-kompleksitet mellem majs- og hvede DDGS at have

store konsekvenser for den enzymatiske nedbrydning af disse to substrater. Desuden viste

både in vitro og in vivo eksperimenter, at enzymer har forskellig nedbrydningsevne i forhold

til deres substrataffinitet og stabilitet over for det fysiske miljø i mave-tarmkanalen af grise.

For yderligere at øge den enzymatiske nedbrydning af DDGS, er der behov for mere forskning

for at identificere de strukturelle komponenter der hindrer den enzymatiske nedbrydning af

især majs DDGS mest.

ix

ABBREVIATIONS

AX arabinoxylan

AXOS arabinoxylan-oligosaccharides

CAID coefficient of apparent ileal digestibility

CATTD coefficient of apparent total tract digestibility

cDDGS corn DDGS

CHO carbohydrates

CV coefficient of variation

DDGS distillers dried grains with solubles

DE digestible energy

DFA ferulic acid dehydrodimer

GC gas chromatography

GH glycoside hydrolase

LMW low molecular weight

ME metabolizable energy

NCP non-cellulosic polysaccharides

NDC non-digestible carbohydrates

NIRS near infrared reflectance spectroscopy

NSP non-starch polysaccharides

PCA principal component analysis

RP-HPLC reversed phase high performance liquid chromatography

TAXI triticum aestivum xylanase inhibitor

TLXI thaumatin-like xylanase inhibitor

TriFA ferulic acid dehydrotrimer

UV ultraviolet

wDDGS wheat DDGS

XIP xylanase inhibitor protein

x

xi

LIST OF PAPERS

xii

LISTOFPAPERS

xiii

Thisthesisisbasedontheworkpresentedinthefollowingpapers,whicharereferredtobyromannumbersthroughoutthethesis.

I: MadsB.Pedersen, SørenDalsgaard,KnudErikB.Knudsen, ShukunYu, andHelleN.Lærke.CompositionalProfileandVariationofDistillersDriedGrainswithSolublesfromVariousOriginswithFocusonNon‐StarchPolysaccharides.Anim.FeedSci.Technol.2014,197:130‐141

II: MadsB.Pedersen,MirkoBunzel,JudithSchäfer,KnudErikB.Knudsen,JensF.Sørensen, Shukun Yu, andHelleN. Lærke. Ferulic AcidDehydrodimer and –DehydrotrimerProfilesofDistillersDriedGrainswithSolublesfromDifferentCerealSpecies.J.Agric.FoodChem.2015,63:2006‐2012

III: MadsB.Pedersen,SørenDalsgaard,SusanArent,RikkeLorentzen,KnudErikB.Knudsen, ShukunYu, andHelleN. Lærke. Xylanase andProtease IncreaseSolubilization of Non‐Starch Polysaccharides and Nutrient Release of Corn‐andWheatDistillersDriedGrainswithSolubles.Biochem.Eng.J.2015,98:99‐106

IV: MadsB.Pedersen, Shukun Yu, Susan Arent, Søren Dalsgaard, Knud Erik B.Knudsen,andHelleN.Lærke.XylanaseIncreasedtheIlealDigestibilityofNon‐Starch Polysaccharides and Concentration of Low Molecular Weight Non‐DigestibleCarbohydratesinPigsFedHighLevelsofwheatDDGS

RevisedversionacceptedforpublicationinJ.Anim.Sci.,March2015

Parts of thework carried out in this PhDprojectwere incorporated into two internationalpatentapplications:

‐ S.L.Arent,B.S.Laursen,E.G.Kiarie,L.F.R.Millán,Z.Zhang,R.Lau,Z.Yu,M.B.Pedersen,andS.Dalsgaard.2014.Method.WO2014020143A2

‐ S.L.Arent,B.S.Laursen,E.G.Kiarie,L.F.R.Millán,Z.Zhang,R.Lau,Z.Yu,F.K.Kooy, T. J. H. Van And C. Koops, M.B. Pedersen, and S. Dalsgaard. 2014.Application.WO2014020142A1

xiv

1

INTRODUCTION

2

INTRODUCTION

3

The increased feed prices for cereal grains and protein crops have forced the feed

industry to look for alternative feedstuffs to replace part of the conventionally feedstuff for

livestock. The rapid increase in production of grain ethanol through the past decade has made

a high amount of the primarily produced co-product, distillers dried grains with solubles

(DDGS) with an estimated US-production of 43 million tonne in 2014, available for animal

feed.

Due to the fermentation of starch to ethanol, the non-fermentable nutrients in the grain

are concentrated approximately 3-fold in the produced DDGS. Thus, DDGS is a rich source of

significant amounts of crude protein, amino acids, phosphorus and fat for animal feed.

However, the main challenges limiting the use of DDGS in the feeds for non-ruminant animals

is the inferior protein quality, potential high mycotoxin content, and high levels of non-starch

polysaccharides (NSP).

The digestive system of non-ruminants does not secrete enzymes capable of digesting

NSP. Instead, NSP is subjected to microbial modification, degradation and fermentation

throughout the gastro intestinal tract. Of the different end-products of microbial fermentation,

only short chain fatty acids are available as energy for the animal. In addition, high inclusion

levels of DDGS in the animal diet may increase the endogenous losses, increase the energy

requirement for maintenance, and increase the amount of nutrients not available for the

animal due to encapsulation within the fibre matrix.

Thus, there is a need to increase the ability of the animal to utilize the energy of DDGS,

and to limit the loss of energy associated with high NSP content.

The work presented in this PhD thesis describes the variation among DDGS sources of

various origins with particular emphasis on detailed characterizing of the NSP fraction,

modification of the NSP-fraction occurring during production, and the differences in

complexity of the NSP fraction between different DDGS sources. In addition, the enzymatic

degradation of DDGS by different exogenous enzymes is investigated both in vitro, and in vivo,

and the differences in enzymatic degradability and digestibility of difference DDGS sources

are discussed in relation to their compositional differences.

4

5

BACKGROUND

6

BACKGROUND

7

Distillers dried grains with solubles

Distillers dried grains with solubles (DDGS) is the main co-product of ethanol

production from cereal grains. In the United States corn is the major crop used for production

of bioethanol, whereas in cooler climates like Europe and Canada wheat is primarily used

instead of corn [1, 2]. In addition to corn and wheat, some ethanol plants may use a mixture of

different crops e.g. sorghum (US), and triticale, barley, and rye (EU), depending on availability

and cost. In Brazil, sugar cane is the major crop used for ethanol production [3].

Bio ethanol is produced by two major industrial technologies; wet milling and dry-grind

processing [4]. Wet milling processing produces gluten meal and gluten feed as major co-

products, whereas DDGS is the major co-product of dry-grind processing [3].

Figure 1. Schematic diagram of a dry-grind ethanol production [3, 4]

The dry-grind processing is designed to maximize the production of ethanol and to

minimize waste. The dry-grind process consists of several major steps including grinding,

liquefaction, saccharification, fermentation, distillation, and recovery of co-products [3],

Figure 1. Grinding is the first step, which breaks the whole grain into smaller particles,

consequently increasing the surface area of the starchy endosperm. Liquefaction reduces the

size of the starch polymers by addition of thermostable α-amylase under high temperature

(100°C) and mechanical shear provided by jet-cooking. After the slurry is cooled down,

Cereal grain Grinding Slurry mixing

Liquefaction (Jet)Cooker Fermentation Distillation

Ethanol Whole stillage

Centrifuge

Thin stillage

Evaporation

Condensed distillers solubles

Coarse solids Wet distillers

grains

Rotary dryer Dry distillers

grains

Distillers dried grains with solubles

CO2

Oil

Rotary dryer

8

glycoamylase is added to break the dextrinized mash into glucose monomers. During the 48-

72 h fermentation, yeast converts glucose into ethanol and CO2. Distillation separates ethanol

from the mash by evaporation to a final ethanol purity of 99.9%. The remaining mash is now

referred to as whole stillage, and is centrifuged to produce thin stillage and coarse solids.

Some ethanol plants may extract some of the oil from the thin stillage before it is condensed

and mixed with the wet grains and dried to produce DDGS [3, 4].

During ethanol production approximate 2/3 of the starting material (starch) is

converted into ethanol and CO2, consequently concentrating the unfermentable nutrients,

such as protein, oil, minerals, and fibre approximately three-fold in the DDGS [5, 6]. The

chemical composition of DDGS is generally acknowledged as being of great variability due to a

combination of several factors; variability in the grain, differences in processing technologies

between ethanol plants, and varying degree of accuracy in the analysis of DDGS samples [7-9].

A typical chemical composition of corn DDGS (cDDGS) and wheat DDGS (wDDGS) is listed in

Table 1.

Thus, DDGS is a rich source of significant amounts of crude protein, amino acids,

phosphorus and fat for animal feed. However, the main challenges limiting the use of DDGS in

feeds for non-ruminant animals is the inferior protein quality caused by heat damaging of

protein from excessive drying, potential high mycotoxin content, and high levels of non-starch

polysaccharides (NSP) [5].

Table 1. Typical chemical composition of corn- and wheat distillers dried grains with solubles (% of

dry matter)

Corn DDGS Wheat DDGS N, samples Mean Range S.D. N, samples Mean Range S.D. Crude protein 44 27.9 (23-35) 2.4 18 38.1 (32-46) 4.1 Crude fibre 33 7.4 (6-11) 1.1 12 7.7 (6-9) 0.9 Neutral detergent fibre 17 36.6 (28-51) 5.8 11 32.6 (22-47) 7.5 Acid detergent fibre 19 13.6 (9-19) 3.3 10 14.0 (7-22) 5.3 Ether extract 37 10.8 (3-18) 2.4 15 5.4 (3-7) 1.1 Ash 36 4.5 (3-6) 0.6 15 5.3 (5-7) 0.6 S.D., standard deviation. Adapted from Olukosi and Adebiyi [6].

Inclusion of distillers dried grains with solubles in diets for pigs

DDGS may be included in diets for pigs in all phases of production. An inclusion level up

to 30% of DDGS has been applied without negative impact on growth performance in nursery

pigs (2-3 weeks post-weaning) and grower-finisher pigs [5]. Due to the relative high content

of oil in cDDGS, high inclusion levels have been reported to negatively affect pork fat quality

[10], thus, it is recommended to withdraw DDGS from the diets 3-4 weeks before slaughter.

DDGS may replace soybean meal in diets for gestating sows without negative effects on sow or

litter performance, and be fed up to 30% DDGS in lactating sows [5].

The concentration of digestible energy (DE) and metabolizable energy (ME) in cDDGS is

similar to that of corn, provided that the oil is not removed during processing [5]. However,

BACKGROUND

9

since DE and ME systems may overestimate energy availability in protein- and fibre-rich

feedstuffs, net energy systems that account for the variable utilization of nutrients are

considered better to manage the inclusion of alternative feedstuffs such as DDGS [11].

Non-starch polysaccharides in cereals

In cereals, such as corn and wheat, the NSP originates from the primary and secondary

cell walls in the grain tissues e.g. endoperm, aleurone, testa, pericarp and tip cap (corn). The

major NSP fractions in corn and wheat, and consequently also in c- and wDDGS, are

arabinoxylan (AX), cellulose, and β-glucan that together with lignin and other cell wall

components build up the skeletal framework and provides structure and protection from the

outside environment [12, 13], Figure 2.

Figure 2. Component tissue structure of: A, wheat [14]; B, corn [15]. C, fluorescent antibody-labeling of

arabinoxylan and β-glucan in aleurone cell wall from wheat grain: orange/red, arabinoxylan; green, β-

glucan; yellow, presence of both arabinoxylan and β-glucan [16].

Arabinoxylan

Arabinoxylan consists of a linear backbone of D-xylose residues linked by β-1,4- linkages

mainly substituted with α-L-arabinose residues at the O-2 position, O-3 position, both O-2 and

O-3 position, or unsubstituted [13, 16], Figure 3A. The arabinose to xylose ratio (A:X) and

uronic acid to xylose ratio (UA:X) indicates the average degree of AX substitution, and is often

used as indication of the structural complexity of the AX [17]. The AX structure is different

10

among the botanical grain fractions, with the outer layers encompassing higher complexity

than endosperm AX, due to a higher content of glucuronic acids (or its 4-O methyl ester) and

galactose residues. This AX is referred to as glucuronoarabinoxylan or heteroxylan [12, 16,

17], Figure 3B. In addition, this heteroxylan may be further substituted by short side chains

of 2-4 sugars residues, acetic acids, and ferulic acids ester-linked to arabinose [18].

Figure 3. Main structural features of arabinoxylan from: A, endosperm; B, outer tissues of cereal grains. A,

arabinose; X, xylose; G, galactose; Ga, glucuronic acid; F, Ferulic acid; uX, unsubstituted xylose; dX, di-

substituted xylose; mX3, mono-substituted xylose; mX2, O-2 monosubstituted xylose. From Saulnier et al.

[16].

Arabinoxylan crosslinking

In cereals, esterified hydroxycinnamic acids such as ferulic acid, p-coumaric acid, and

sinapic acid can participate in photochemically or oxidatively induced coupling of cell wall

polysaccharides such as arabinoxylans and pectins along with cell wall trapped proteins and

lignin like polymers [19-21]. While p-coumaric acid is mainly bound to lignin, ferulic acid and

its derivatives, which are the quantitatively most important cross-links in the plant cell wall,

are primarily bound to arabinoxylans and pectins [22]. Ferulate in particular has been found

to form dimers (DFA), trimers (TriFA), and higher ferulate oligomers potentially cross-linking

two or more polysaccharide chains, Figure 4. Evidence that DFAs are indeed attached to

arabinoxylans has been reported by use of carbohydrases and mild acidic hydrolysis followed

by structural characterization in bamboo and corn bran [23, 24], and from corn bran insoluble

fibre [25, 26]. The cross-linking of AX by DFA and TriFA strengthens the cell wall with

implications for plant physiology, the plant’s defense mechanisms against pathogens, food

science, and nutrition [22, 27, 28].

BACKGROUND

11

Figure 4. Schematic illustration of the hypothetical coupling of two arabinoxylan chains via a ferulic acid

dehyrdrotrimer. Ara, arabinose; FA, ester-linked ferulic acid; Xyl, xylose backbone. From Bunzel [22].

Cellulose

Cellulose consist of linear homopolymers of D-glucose residues linked by β-1,4-linkages,

Figure 5. These glucan chains are stabilized by numerous intermolecular hydrogen-bonds

between the hydroxyl groups of glucose residues from adjacent chains, thus, forming

microfibrils. Cellulose is organized in crystalline regions within the microfibrils interspersed

by non-crystalline amorphous regions. Cellulose is present in all plant tissues, however, to a

much higher degree in the outer tissues [13, 29-31].

Figure 5. Structure of the hydrogen-bonding pattern in cellulose. From Festucci-Buselli et al. [30].

β-glucan

Like cellulose, β-glucan is formed from a linear homopolymer of D-glucose residues.

However, these residues are linked by two to three consecutive β-1,4-linkages that are

separated by a single β-1,3-linkage, thus, making the β-glucan less ordered compared to

cellulose [17, 31, 32]. In addition, β-glucan is more easily degradable than AX in the small

intestine, thus, only associated with minor effects on physico-chemical properties such as

luminal viscosity [33].

Difference in fibre complexity between corn- and wheat DDGS

The nutrient composition and characteristic of DDGS reflect in part the composition in

the parent grain, as only starch is removed during the fermentation process [34, 35]. As

illustrated in Table 1; cDDGS contain more fat than wDDGS, whereas wDDGS has a higher

content of crude protein than cDDGS, in line with previous characterizations of parent grains

[36].

12

The NSP fraction in both corn and wheat [37] and c- and wDDGS [34], describes a

greater fraction of soluble NSP in wheat than in corn. Corn and wheat have approximately

similar content of cellulose, whereas the content of AX is reported slightly higher in wheat

than corn [37]. The higher A:X and UA:X ratio in corn than in wheat [34, 37], indicate a more

complex structure of the AX present in corn compared to wheat, due to a higher degree of AX

substitution. Corn AX is described to comprise greater branch density and complexity than

wheat AX [28, 38-40]. In addition, the higher complexity of the fibre matrix in corn than wheat

is further substantiated by the markedly higher content of DFAs and TriFAs reported in corn

and wheat insoluble fibre [27, 41, 42], and whole grain flours [28].

Collectively, the fibre matrix in cDDGS is expected to be more complex than wDDGS in

terms of a more substituted AX and by having a markedly increase in potential polysaccharide

cross-linkages, which combined is likely to have implications on the enzymatic degradation

and overall digestibility.

Non-starch polysaccharides in relation to feed for swine

The digestive system of non-ruminants does not secrete enzymes capable of degrading

NSP. Instead, NSP is subjected to microbial fermentation throughout the gastro intestinal tract

[43, 44]. The fermentation of NSP occurs primarily in the large intestine, however, depending

on the type of NSP a considerable depolymerisation and fermentation may also occur in the

small intestine of pigs [45-48]. The main end-products of microbial fermentation are short

chain fatty acids (SCFA), hydrogen, methane, and heat [49-51]. The produced SCFA, such as

acetate, propionate, and butyrate are rapidly absorbed and have been shown to supply up to

30% of the energy requirement of pigs [44]. On the other hand the remaining end-products

(hydrogen, methane, and heat) are considered as energy loss, as these end-products are not

available to the pig as energy source, thus, decreasing the efficiency of energy utilization [52,

53]. In addition, the undigested fibre itself and excretion of microbial biomass is also regarded

as energy loss. The degree of microbial fermentation of NSP is related, among other things, to

the solubility of the NSP, illustrated by the high total tract digestibility of easily solubilizable

AX from wheat [54, 55], whereas insoluble and lignified branched AX from wheat secondary

cell walls remain practically undegraded in pigs [55, 56]. Further research has indicated that

the microbial degradation of the soluble NSP fraction is selective and that the molecular

weight distribution may explain differences in SCFA production [57]. Increased maturity of

the pigs will positively affect the NSP digestibility due to an increased volume of the gastro

intestinal tract, longer retention time, and higher activity of cellulytic bacteria in combination

with a lower feed intake per unit of live weight [58, 59].

A high content of NSP in animal diets may directly or indirectly affect the digestive

utilization and absorption of other nutrients. As mentioned above, chemical properties of the

NSP (e.g. solubility, lignification and branching) affect the digestion of the NSP itself. In

addition, NSP in the cell walls may encapsulate potential available nutrients such as proteins,

thus, suppressing the availability of these nutrients for the endogenous enzymes in the small

intestine [33, 54, 60]. Depending on the type, NSP may contribute to the physical properties of

BACKGROUND

13

the digesta such as viscosity (soluble NSP) and water-binding capacity, potentially affecting

digesta transit time, bulking properties, microbial activity, and endogenous loses [43, 61-64].

In addition, especially insoluble NSP may increase mucus secretion in the small intestine due

to the physical effects on the gut wall, consequently causing damage to the mucus layer [63].

Furthermore, feeding high levels of NSP will increase the total empty weight of the gastro

intestinal tract [65], stimulate intestinal epithelial cell proliferation [66], increase endogenous

secretions [67], increase maintenance energy requirement [67], and increase satiety and

consequently potentially decrease voluntary feed intake [68].

Considering the type of NSP present in DDGS, which is primarily insoluble AX and

cellulose, the effects caused by high inclusion levels of DDGS in diets for pigs will potentially

be; dilution of energy in feed, increased endogenous loses primarily due to the mechanical

effects on the gut wall, encapsulation of nutrients, and increased energy requirement for

maintenance.

Improving fibre utilization in swine by exogenous enzymes

Due to the increased usage of high-fibre co-products, such as DDGS, and the negative

effects associated with the high NSP content as described above, ways to increase the

utilization of energy and/or limit the loss of energy from feeding these feedstuffs are essential.

One approach is the addition of exogenous fibre degrading enzymes to improve nutrient

utilization in diets for animal feed [69, 70].

As in the parent grain, corn and wheat, AX comprises the largest fraction of NSP in both

c- and wDDGS. Thus, the main focus of this PhD project was aimed towards degradation of AX.

Due to the structural complexity, complete degradation of AX requires a broad range of

enzyme activities, like β-1,4-endoxylanase, α-D-glucuronidase, acetyl-xylan esterase, α-L-

furanosidase, and ferulic acid esterase [71, 72], Figure 6. Of these NSP-degrading enzymes

only the effect of xylanase on DDGS degradation was investigated in this PhD project.

Figure 6. The structural components of arabinoxylan/heteroxylan and the enzymes responsible for its

degradation. Adapted from Hövel et al. [71].

Xylanases

In nature, xylanases are an important part of the enzyme-system involved in the

metabolism of xylan as carbon source found in e.g. bacteria and fungi populations [73, 74].

14

Endo-β-1,4-xylanases (EC 3.2.1.8, hereon referred to as xylanase) are hydrolytic enzymes

which cleave the β-1,4-linkages within the xylan backbone [75, 76]. Xylanases belong to two

main families based on their primary sequence and structure; glycoside hydrolase family (GH)

-10 and -11 [72, 77]. In addition to GH10 and -11, xylanases are also described to belong to

GH5, -8, -30, and -43 families [78]. The three-dimensional structure of enzymes is essential for

the substrate affinity, molecular mechanism, and overall function of the enzyme. Xylanases

belonging to the GH-10 family is generally characterized by a (β/α)8 barrel structure, Figure

7A, whereas GH-11 xylanases are generally characterized by a β-jelly-roll structure, Figure

7B [78, 79].

Figure 7. Representative structures of xylanases from glycoside hydrolase family -10 (A) and -11 (B). Red,

α-helix; blue, β-strand/sheets . Adapted from Collins et al. [79].

Figure 8. Schematic overview of endoxylanase degradation of arabinoxylan (AX) into arabinoxylan-

oligosaccharides (AXOS). Red arrows indicate site of hydrolysis. Adapted from Lagaert et al. [80].

BACKGROUND

15

In addition to the different structural characteristics among xylanases, different

substrate affinities are also reported. GH10 xylanases have been shown to exhibit greater

catalytic versatility or lower substrate specificity than xylanases from the GH11 family. GH11

xylanases primarily cleave unsubstituted regions of the xylan backbone, whereas GH10 is

capable of cleaving in more substituted regions, hence being less hindered by the presence of

substitutions [78, 81]. The GH-11 xylanases generally have a smaller molecular size (<30 kDa)

than GH-10 xylanases (>30 kDa), and it has been proposed that the smaller size may enable

the GH-11 xylanase in penetrating the cell wall network [82].

Endo-xylanases solubilize arabinoxylan by cleaving the β-1,4-glycosyl linkages within

the β-1,4-xylose backbone of insoluble as well as soluble arabinoxylan, thus, partially

solubilizing insoluble arabinoxylan and fragmenting soluble arabinoxylan into arabinoxylan-

oligosaccharides (AXOS), Figure 8 [81]. The generated AXOS are generally acknowledged as

being associated with potential prebiotic effects [83-85].

A part of the plant’s defense mechanism against pathogens is the production of proteins

which can inhibit xylanase activity, so-called xylanase inhibitors [86]. The Triticum aestivum

xylanase inhibitor (TAXI) has been isolated and characterized from wheat, rye, and barley and

is described to specifically inhibit GH11 xylanases [87]. The xylanase inhibitor protein (XIP)

has been isolated from wheat, rye, barley, corn, and rice described to inhibit both GH10 and -

11 xylanases [88]. Both TAXI and XIP is characterized as a competitive inhibitor, thus,

interacting with the enzyme active site by mimicking the substrate [87, 88]. The thaumatin-

like xylanase inhibitor (TLXI) has been identified in wheat, and is described as an non-

competitive inhibitor towards GH11- but not GH10 xylanases [86].

Use of xylanase in animal feed

Next to phytase, NSP degrading enzymes (mainly xylanase) comprise the second largest

fraction of commercial enzymes sold to the feed industry particularly for poultry feed, but are

commonly used in pig diets too [89]. The effects of NSP degrading enzyme addition in animal

diets have been reviewed by several authors [11, 68-70, 90, 91]. Collectively, the results

indicate that NSP degrading enzymes may improve NSP and nutrient digestibility, provided

that the NSP constituents in the diets match the substrate affinity of the enzymes or that the

substrate is limiting for the nutrient digestion. In addition, it is generally acknowledged that

further research is needed to elucidate, among other things; NSP characterization of the

feedstuff in detail (especially “new” co-products such as DDGS), substrate affinity of applied

enzymes in relation to feed composition, and detailed characterization of the degradation

products.

The effects of NSP-degrading enzymes application in DDGS containing diets are

currently inconclusive; Widyaratne et al. [92] reported an increase in ileal digestibility of

energy by xylanase addition in wheat but not in diets containing 40% wDDGS, Yanez et al. [93]

reported no effect of xylanase addition in diets containing 44% corn and wheat co-fermented

DDGS, Jones et al. [94] reported no effect of enzyme addition on growth performance in diets

containing 30 % cDDGS, whereas Emiola et al. [95] reported an improved ileal nutrient- and

16

total tract energy digestibility and growth performance by adding a combination of xylanase,

β-glucanase and cellulase enzymes in pigs fed 30% wDDGS.

Several factors such as the degree of lignification, cross-linkages, structural organization,

and the time the digesta remains in the fore gut, will affect and/or limit the effects of

exogenous enzyme addition. Based on studies with pigs showing that the outer layers of

wheat and rye are almost resistant to degradation after passage through the large intestine,

Back Knudsen [17] suggested that it is primarily the endosperm and aleurone layers that

potentially can be degraded by exogenous enzymes under in vivo conditions, Figure 9.

Figure 9. Potential effects of exogenous enzymes in cereal cell wall degradation, illustrated by cross-section

of the outer layers and the endosperm of corn, wheat, and barley. From Bach Knudsen [17].

17

METHODS OVERVIEW

18

METHODS OVERVIEW

19

Near infrared reflectance spectroscopy

To characterize the composition of macronutrients (e.g. crude protein, fat, ash, and

starch) in 138 DDGS samples [Paper I], near infrared reflectance spectroscopy (NIRS) was

applied instead of conventional wet chemistry methods [37, 96-98], which were used in Paper

IV. NIRS is a non-destructive analytical technique which relates the light absorbed by a sample

to its chemical and physical composition, with instant results. Due to its rapid and precise

measures, NIRS is often used in agriculture to determine the composition of feedstuff [99-

102]. However, the successful application and prediction obtained by NIRS relies on several

compositional analyses and calibration of spectral data, which requires a broad number of

samples with varying compositional profiles.

In principle, the milled samples were scanned from 1100 to 2498 nm on a FOSS

NIRSystems. The spectral data were predicted by Aunir (AB Agri, UK) for the composition of

moisture, fat (both ether extract and acid hydrolysis), crude protein, crude fibre, ash, starch,

total sugars, neutral- and acid detergent fibre, using the calibration available for DDGS.

Analysis of non-digestible carbohydrates

Quantification of non-digestible carbohydrates (NDC), non-starch polysaccharides, and

low molecular weight NDC was performed using a modification of the Englyst [103] and

Uppsala [104, 105] procedures, essentially as described by Bach Knudsen [37]. The NSP-

procedure was applied on the DDGS samples analyzed in Paper I and –III, whereas the more

detailed NDC-procedure was applied in analysis of the wDDGS used in Paper IV.

In principle, the NSP procedure consists of three parallel runs (A, B, C), whereas an

additional run is include for the determination of NDC (D), Figure 10. Procedure A,

determines the total NSP by removal of starch, ethanol precipitation of soluble fibres, swelling

of cellulose, hydrolysis of NSP into monomers, reduction of neutral sugars to alcohols

followed by derivatization into alditol acetates, and finally determination of these using gas

chromatography (GC). Procedure B, determines the non-cellulosic polysaccharides (NCP) by

omitting swelling of cellulose. Procedure C, determines the insoluble NSP by extraction of the

soluble NSP in phosphate buffer at 100°C for 1 h followed by centrifugation and removal of

supernatant. Procedure D, determines the total carbohydrates (CHO) by direct swelling and

hydrolysis, thus, omitting removal of starch. The content of uronic acids was determined in a

separate step by colorimetry. Klason lignin was determined gravimetrically as the residue

resistant to hydrolysis by 2 M H2SO4 for 2 h at 100 °C.

20

Figure 10. Analytical procedure for determination of different fractions of non-starch polysaccharides. CHO,

carbohydrates; NCP, non-cellulosic polysaccharides, NSP, non-starch polysaccharides.

Based on the four parallel runs in the NSP procedure, the following calculations were

performed;

Cellulose was calculated as:

Cellulose = NSPglucose, 12 M – NCPglucose, 2 M

Total NSP as:

Total NSP = rhamnose + fucose + arabinose + xylose + mannose + galactose +

NCP-glucose + uronic acids + cellulose

Soluble NSP as:

Soluble NSP = Total NSP – insoluble NSP

Sample

Removal of starch as A as A

Precipitation of soluble NSP

Extraction of soluble NSP

Swelling, 12 M H

2SO

4

Add 2 M H2SO

4 Swelling, 12 M

H2SO

4

Hydrolysis, 2M H2SO

4

Portion to uronic acid

Reduction and derivatization

Determination GC

as A as A as A

Procedure A, Total NSP

Procedure B, Total NCP

Procedure C, Insoluble NSP

Procedure D, Total CHO

METHODS OVERVIEW

21

Total NDC as:

Total NDC = Total CHO – starch

Low molecular weight (LMW) NDC as:

LMW-NDC = Total CHO – Total NSP - starch

Reversed phase high performance liquid chromatography with ultraviolet

detection

To investigate and quantify ester-linked DFA and TriFA in DDGS samples [Paper II], a

validated method using reversed phase high performance liquid chromatography (RP-HPLC)

with ultraviolet detection (UV) was applied [27]. Compared to previously applied

methodologies e.g. GC and HPLC methods [41, 106], this two-gradient RP-HPLC/UV method

does not require derivatization and allows for a complete separation of monomers, DFAs, and

TriFAs [27]. Analysis of lignin-rich matrices (e.g. corn stover) might cause problems with the

UV detection due to co-elution of lignin compounds with DFAs and TriFAs. For analysis of

such matrices HPLC-mass spectrometry methods are proposed [28]. The use of the RP-

HPLC/UV method requires synthesis of internal standard and isolation of standard

compounds for accurate validation, which can be very time consuming.

In principle, the DDGS samples were washed twice with both ethanol (80 % v/v) and

acetone to lower the spectral noise. After alkaline hydrolysis of the ester-linkages, the samples

were acidified and internal standard was added. Extraction was carried out using diethyl

ether and samples dried under a stream of nitrogen. Finally, the samples were redissolved in

tetrahydrofuran/water (1:1) and analyzed.

Enzymatic degradation of DDGS in vitro

To investigate the enzymatic degradation of c- and wDDGS - and to compare the

efficiency of different enzymes, experiments were setup using purified enzymes [Paper III].

Four different xylanases were evaluated alone and in combination with protease and phytase

according to four full factorial experiments. In addition, the enzymatic degradation of

insoluble DDGS was also investigated based on substrate prepared as described in the NSP

procedure-C.

The major difference between the applied methodology and a commonly used in vitro

digestion method [107], is that no endogenous digestive enzymes were added and pH was

kept constant throughout the incubation. However, the applied experimental conditions (i.e.

pH 6.0, reaction time 4 h, temperature 39°C) was chosen to mimic the physical conditions

occurring in the small intestine of pigs [108], whereas stirring was at 1100 rpm. The

experimental setup and following analysis was chosen since it allowed a direct comparison of

the effectiveness on DDGS degradation by different enzymes in a time effective approach.

The effect of enzyme addition was evaluated based on the ability to simultaneously

solubilize pentosan and protein by use of analytical methods that both were compliable with

the use of micro titer plates. Quantification of soluble pentosan was measured principally as

described by Rouau and Surget [109] using a continuous flow injection apparatus with

22

autosampler unit, Figure 11. In principle, the supernatant was hydrolyzed to monomers by

glacial acetic acid/HCl, 37%, (50:1) at 96 °C, and mixed with phloroglucinol (1,3,5-

trihydroxybenzene), consequently developing a colored complex. The absorbance was

measured at 510 and 550 nm against a xylose standard.

Quantification of soluble protein was determined by use of a commercial available

protein quantification kit (BCA Protein Assay Kit, Pierce) against a bovine serum albumin

standard. The first step involves the chelation of copper with protein under alkaline

conditions to produce a light blue complex. Secondly, bicinchoninic acid reacts with the

reduced copper-ion resulting in a purple colored complex that is measured at 562 nm [110].

Figure 11. A, continuous flow injection apparatus for analysis of soluble pentosan; B, corresponding

autosampler unit (Skalar Analytical B.V., Breda, Netherlands)

In vivo digestion model

To investigate the effects of enzyme addition on nutrient digestibility in vivo, two animal

trials were performed using ileum cannulated pigs fed either cDDGS or wDDGS. For wDDGS

the pigs were fed a control diet and three different enzyme treatments according to a double

4×4 Latin square design [Paper IV], whereas for cDDGS the pigs were fed five diets according

to a double 5×5 Latin square design [Supplementary data]. Due to protection of potentially

Intellectual Properties, only the results from the control diet without enzymes are presented

as Supplementary data in the General discussion section.

Implementation of a simple T-cannula in the distal ileum allows for collection of ileal

digesta and faeces and enable repeated samples on the same animal, which reduces the

variability and number of experimental animals, Figure 12 [111]. Ileal digestibility reflects

the endogenous enzymatic digestion, exogenous enzymatic digestion, microbial fermentation

in the small intestine, endogenous secretions into the gut, and absorption [112]. By addition

of an indigestible marker (e.g. chromic oxide as used in this thesis) the total extent of

digestion in the small- and large intestine can be determined.

The apparent ileal digestibility of organic matter, starch, crude protein, HCl-fat, and

NSPs were calculated by the indicator method relative to dietary and ileal concentrations of

chromic oxide [113], according to Equation 1 below. A prerequisite of the indicator method is

A B

METHODS OVERVIEW

23

that the marker is indigestible and not separated from the remaining feed matrix throughout

the gastrointestinal tract.

𝐷𝑖𝑔𝑒𝑠𝑡𝑖𝑏𝑖𝑙𝑖𝑡𝑦 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑜𝑓 𝑋 = 1 −𝐶𝑟2𝑂3,𝑑𝑖𝑒𝑡 × 𝑋𝑖𝑙𝑒𝑢𝑚

𝐶𝑟2𝑂3,𝑖𝑙𝑒𝑢𝑚 × 𝑋𝑑𝑖𝑒𝑡 (𝟏)

, where X is the concentration of a specific nutrient in the diet and ileum digesta

Figure 12. Implementation of simple T-cannula in the pig distal ileum

Principal component analysis

Principal component analysis (PCA) was applied to the compositional data to detect

distributions and separations among DDGS sources of various origins [Paper I], and to

investigate to which cell wall constituents the DFAs and TriFAs shared the highest correlation

[Paper II].

PCA is a common statistical method that transforms a number of correlated variables

into a smaller number of uncorrelated variables called principal components. The generated

score plot allow the DDGS samples to be classified or grouped according to their overall

composition, whereas the corresponding loading plot highlight which components are

responsible for the separation. In addition, the loading plot illustrate which components are

correlated with each other, as these will be located close to one another in the plot [114].

24

25

AIM AND HYPOTHESES

26

AIM AND HYPOTHESES

27

The overall aim of the work described in this thesis was to investigate the variation and

differences in NSP composition of DDGS, and to improve the digestibility of c- and wDDGS as

feedstuff by enzymatic modification of the fibre structure.

The work described in this thesis was divided into three work packages, each with

specific aims;

1. To perform a systematic investigation of the variation in the chemical and

physicochemical properties of different DDGS sources with emphasis on the fibre

matrix. If possible, also relate specific process technologies to DDGS composition.

2. To study, in vitro, the effect of different xylanases alone and in combination with other

enzymes for the degradation of c- and wDDGS. The enzyme performance is evaluated

based on the ability to simultaneously solubilize pentosan and protein from DDGS.

3. To study, in vivo, the effects of xylanase and protease addition in ileal cannulated pigs

fed c- or wDDGS as sole provider of macronutrients, according to a double 5×5 Latin

square design or a double 4×4 Latin square design, respectively. The effect of enzyme

addition is evaluated by apparent ileal digestibility of macronutrients and NSP, and

apparent total tract digestibility of NSP.

It is hypothesized that:

Different raw materials and ethanol production technologies give rise to variation in

the composition and structural complexity of NSP in c- and wDDGS

The susceptibility of c- and wDDGS sources to enzymatic degradation using exogenous

enzymes is dependent on the chemical composition

Xylanases (along with protease and phytase) applied in feed will increase fibre

degradation and nutrient release both in vitro and in vivo, thus, affecting ileal

digestibility of NSP and macro nutrients

28

29

PAPER I

Compositional Profile and Variation of

Distillers Dried Grains with Solubles from

Various Origins with Focus on Non-Starch

Polysaccharides

Mads B. Pedersen, Søren Dalsgaard, Knud Erik B.

Knudsen, Shukun Yu, and Helle N. Lærke

Published in:

Animal Feed Science and Technology, 2014,

197:130-141.

30

Animal Feed Science and Technology 197 (2014) 130–141

Contents lists available at ScienceDirect

Animal Feed Science and Technology

journal homepage: www.elsevier.com/locate/anifeedsci

Compositional profile and variation of Distillers Dried Grainswith Solubles from various origins with focus on non-starchpolysaccharides

M.B. Pedersen a,b,∗, S. Dalsgaard a, K.E. Bach Knudsen b, S. Yu a, H.N. Lærke b

a Enzyme Development, DuPont Nutrition Biosciences ApS, Brabrand DK-8220, Denmarkb Department of Animal Science, Aarhus University, Foulum, Tjele DK-8830, Denmark

a r t i c l e i n f o

Article history:Received 27 February 2014Received in revised form 24 July 2014Accepted 26 July 2014

Keywords:Chemical compositionCornDDGSNon-starch polysaccharidesVariationWheat

a b s t r a c t

Corn-, wheat- and mixed cereal Distillers’ Dried Grains with Solubles (DDGS) wereinvestigated for compositional variability among DDGS origins, ethanol plants, and therelationship between corn and corresponding DDGS. A total of 138 DDGS samples wereanalyzed by use of Near Infrared Reflectance Spectroscopy for common constituents, while63 DDGS samples along with 11 corn samples were characterized for their non-starchpolysaccharide (NSP) content. The results indicated that the compositional profile of DDGSreflected the nutrient content of the parent grain but with a greater content of remainingnutrients (e.g. protein, fat, fibre and minerals) after fermentation of starch to ethanol. CornDDGS differentiated from wheat DDGS by a greater content of fat (P≤0.006), insoluble-NSP (P<0.001), uronic acids (P<0.001), cellulose (P=0.032), and arabinose/xylose (P<0.001)– and uronic acid/xylose-ratio (P<0.001). Wheat DDGS differentiated from corn DDGS bya greater content of ash (P=0.001), soluble-NSP (P<0.001), and Klason lignin (P<0.001).Among the three sources of DDGS, the greatest variation was observed for the contentof soluble-NSPs, especially soluble arabinoxylan. Based on the compositional profiles ofthe DDGS, principal component analysis allowed for a visual differentiation of corn DDGSfrom five different ethanol plants, indicating the potential of each ethanol plant to produceDDGS with consistent compositional characteristics. Furthermore, investigation of cornand corresponding DDGS indicated that the NSP fraction is modified during the fermenta-tion process, especially arabinoxylan, by an increase in soluble arabinoxylan proportion inDDGS. In addition, the arabinose/xylose (P<0.001) and uronic acid/xylose-ratio (P<0.001)were greater for corn, compared with corresponding DDGS, indicating modifications of theendosperm arabinoxylan during the fermentation and drying process.

© 2014 Elsevier B.V. All rights reserved.

Abbreviations: ADF, acid detergent fibre; AH, acid hydrolysis; A/X, arabinose-to-xylose; Cel./NSP, cellulose-to-NSP; CV, coefficient of variation CF,crude fibre; CP, crude protein; DF, dietary fibre; DM, dry matter; DDGS, distillers dried grain with solubles; EE, ether extract; I-NSP, insoluble NSP; NCP,non-cellulosic polysaccharide; aNDFom, neutral detergent fibre assayed with heat stable amylase exclusive residual ash; NIRS, near infrared reflectancespectroscopy; NSP, non-starch polysaccharides; PCA, principal component analysis; S-Ara, soluble arabinose; S-NSP, soluble NSP; S-Xyl, soluble xylose;T-NSP, total NSP; UA/X, uronic acid-to-xylose.

∗ Corresponding author at: DuPont Nutrition Biosciences ApS, Edwin Rahrs Vej 38, Brabrand DK-8220, Denmark. Tel.: +45 8943 5327.E-mail addresses: Mads.brogger.pedersen@dupont.com, madsb.pedersen@agrsci.dk (M.B. Pedersen).

http://dx.doi.org/10.1016/j.anifeedsci.2014.07.0110377-8401/© 2014 Elsevier B.V. All rights reserved.

M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141 131

1. Introduction

Distillers’ Dried Grains with Solubles (DDGS) is the major co-product from the dry-grind production of bioethanol fromcereal grains. After the conversion of grain starch to ethanol during the fermentation process, the non-fermentable partremains, with an increase of all nutrients except starch compared with those in the parent grain. For both corn- andwheat DDGS an increase of approximately 3-fold of nutrients such as protein, fat, vitamins, minerals and fibre is observed(Widyaratne and Zijlstra, 2007; Stein and Shurson, 2009).

Compositional variability of DDGS from various bioethanol plants has previously been reported for both corn (Spiehset al., 2002; Belyea et al., 2004; Batal and Dale, 2006; Liu, 2011), and wheat origins (Nyachoti et al., 2005; Bandegan et al.,2009; Cozannet et al., 2010). The variability in the chemical composition of DDGS is due to a number of factors includingdifferences in processing technologies among the bioethanol plants, and variability in chemical composition of the grains(Liu, 2011; Olukosi and Adebiyi, 2013). The majority of the reported compositional profiles of DDGS have focused mainlyon common constituents such as crude protein (CP), crude fibre (CF), neutral detergent fibre, acid detergent fibre (ADF), fat,minerals, and amino acids (Liu, 2011; Olukosi and Adebiyi, 2013). Non-starch polysaccharides (NSP) make up 25–30% of theDDGS, with the two major components of the NSP being arabinoxylan and cellulose. Arabinoxylan consists of d-xylose unitsjoined by �-linkages substituted with arabinose residues along the chain (Kim et al., 2008). The substitution of arabinoseoccurs randomly allowing other substitutes such as d-glucuronic acid and acetyl groups to attach to the xylan backbone(Bedford, 1995). These random substitutes together with feruloylated arabinose residues induce the arabinoxylan cross-linking to form strong heterogeneous intermolecular complexes, affecting the potential for enzymatic degradation alongwith the potential encapsulation of nutrients within the cell wall (Hartley, 1973; Lequart et al., 1999; Lapierre et al., 2001;Piber and Koehler, 2005). Despite the high content of NSPs in DDGS detailed characterization of the NSPs in DDGS is limited asonly a few reports describe the NSP composition (Widyaratne and Zijlstra, 2007; de Vries et al., 2013). Characterization of theNSP composition in DDGS of various origins may contribute to the overall understanding and interpretation of opportunitiesand limitations in DDGS degradation in vitro and in vivo, which is particularly useful for feed manufactures and enzymeproducers targeting DDGS degradation.

The current study describes the compositional profiles and extent of variation in DDGS from three different grain origins;corn, wheat, and mixed cereal. A total of 138 DDGS samples from three sources of grain and 24 different bioethanol plantswere analyzed based on the common constituents, with detailed NSP profiles of 63 DDGS samples. Multivariate data analysiswas applied to the compositional data to determine grouping and separation of DDGS samples according to grain originsand bioethanol plants.

2. Materials and methods

2.1. Materials

A total of 138 DDGS samples were collected from 24 ethanol plants in the U.S. and E.U., covering the following threeparent grain origins; corn, wheat, and a mix of cereal (mixed).

A total of 72 corn DDGS samples were collected from 21 different ethanol plants in the U.S.; 11 DDGS samples along with11 corn samples were sampled over 11 months with one sample every month from a plant in Nevada (P1), 17 samples weresampled over 1½ months from a plant in Minnesota (P2), 10 samples were sampled over 20 days from another plant inMinnesota (P3), 10 samples were sampled over 20 days from a plant in Wisconsin (P4), 8 samples were sampled over 8 daysfrom a plant in Iowa (P5), and 16 samples were supplied by a DDGS supplier in Iowa, representing samples from 7 plants inMinnesota, 4 plants in Iowa, 2 plants in Nevada, and 1 plant in Illinois, Indiana and Kentucky.

A total of 56 wheat DDGS samples were collected from 2 different ethanol plants in the E.U., with 46 samples from a plantin the U.K. sampled over a period of app. 3 months, and 10 samples from a plant in France sampled over 20 days.

Finally, 10 DDGS samples of a mixed grain origin were sampled over 20 days from a plant in Sweden. The parent grainsused in the production of the mixed DDGS were wheat, triticale, barley, and rye. However, their individual proportion isunknown.

2.2. Chemical analysis

All 138 DDGS samples (72 corn, 56 wheat, and 10 mixed) and 11 corn samples were milled (0.5 mm) and scanned from1100 to 2498 nm by near infrared reflectance spectroscopy (NIRS) on a FOSS NIRSystems 5000 (Foss). The spectral data werepredicted by Aunir, AB Agri Ltd., UK, for the composition of: moisture, fat ether extract (EE), fat acid hydrolysis (AH), CP(N × 6.25), CF, ash, starch, total sugars, aNDFom and ADF, using the calibration available for DDGS and corn, respectively.

A total of 63 DDGS samples (47 corn, 11 wheat, and 5 mixed) and 11 corn samples were quantified for total and solubleNSP content along with their constituent sugars by gas–liquid chromatography for neutral sugars and by a colorimetricmethod for uronic acids, with procedure and calculations according to Bach Knudsen (1997), with the modification that2 mol/L sulfuric acid for 1 h was used to hydrolyze the non-cellulosic polysaccharides (NCP) rather than 1 mol/L sulfuric

132 M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141

Table 1Compositional profile of corn-, wheat-, and mixed cereal DDGS (g/kg DM).

Corn DDGS Wheat DDGS Mixed DDGS

Meana Rangeb S.D. (CV) Meanc Ranged S.D. (CV) Meane Rangef S.D. (CV)

Moisture 87g (65–124) 8 (0.10) 76g (68–87) 2 (0.02) 110h (74–127) 15 (0.14)Fat (EE) 91g (65–118) 15 (0.17) 52h (44–65) 8 (0.16) 47h (44–50) 2 (0.05)Fat (AH) 111g (84–135) 14 (0.13) 73h (65–88) 8 (0.11) 69h (66–74) 3 (0.04)Crude protein 314g (271–364) 21 (0.07) 334g (303–379) 28 (0.09) 366g (338–383) 13 (0.03)Crude fibre 77g (64–95) 6 (0.07) 67g (55–88) 9 (0.14) 62g (56–76) 6 (0.10)Ash 71g (54–90) 7 (0.09) 91h (81–100) 4 (0.05) 87g,h (80–102) 7 (0.08)Starch 60g (29–139) 27 (0.45) 40g (<10–88) 42 (1.03) 24g (<10–37) 9 (0.39)Total sugars 90g (54–126) 17 (0.19) 98g (46–124) 22 (0.23) 126g (99–142) 13 (0.10)aNDFom 351g (302–397) 24 (0.07) 306g (273–342) 26 (0.08) 302g (289–312) 9 (0.03)ADF 101g (89–119) 6 (0.06) 105g,h (95–122) 8 (0.07) 120h (115–123) 3 (0.02)

Based on near infrared reflectance spectroscopy. ADF, acid detergent fibre; AH, acid hydrolysis; EE, ether extract; aNDFom, neutral detergent fibre; S.D.,standard deviation.

a Avg. of means from 21 plants.b N = 72.c Avg. of means from 2 plants.d N = 56.e Mean from 1 plant.f N = 10.

g,hMeans that do not share a letter are significant different at P<0.05.

acid for 2 h. Klason lignin was measured gravimetrically as the residue resistant to hydrolysis by 2 mol/L sulfuric acid (BachKnudsen, 1997).

2.3. Calculations and statistical analysis

Content of NCP-glucose was calculated as

NCP-glucose = NSPglucose (2 mol/L H2SO4)

Content of cellulose was calculated as

cellulose = NSPglucose (12 mol/L H2SO4)–NCP-glucose

Total non-starch polysaccharides (T-NSP) as

T-NSP = rhamnose + fucose + arabinose + xylose + mannose + galactose + NCP-glucose + uronic acids + cellulose

Soluble NSP (S-NSP) as

S-NSP = T-NSP − insoluble NSP(I-NSP)

Dietary fibre (DF) as

DF = T-NSP + Klason lignin

A one-way ANOVA was applied for comparison among means of the DDGS compositions from various origins followedby a Tukey pair wise comparison with overall significance level at P=0.05, using Minitab 16 (Minitab Inc.). Data from thecompositional analysis were categorized and analyzed by principal component analysis (PCA) with evenly spread 7-foldpartial cross-validation after mean centering and unit variance scaling, using Evince 2.5.5 software (UmBio AB) to detectdistributions and separations among the groups, as previously discussed (Shewry et al., 2013).

3. Results

3.1. Compositional variation in DDGS of various origins

The results of the compositional analyses of the DDGS are presented in Table 1. Common for all DDGS, aNDFom and CPrepresented the major fractions, with a large extent of the aNDFom fraction present as ADF. Corn DDGS contained a greatercontent of fat (AH) compared with wheat- and mixed DDGS with an increase of 38 g/kg DM (P=0.005) and 42 g/kg DM(P=0.019), respectively. Furthermore, a greater aNDFom content was observed in corn DDGS with an increase of 45 g/kg DM(P=0.057) and 49 g/kg DM (P=0.118) compared with wheat- and mixed DDGS, respectively. Wheat DDGS contained 20 g/kgDM (P=0.001) greater ash content than corn DDGS. Among the three sources of DDGS, the greatest coefficient of variation(CV) was observed for the content of starch, with CVs of 0.39–1.03. For both corn- and wheat DDGS, the second and thirdgreatest CVs was observed for total sugars (0.19 and 0.23) and fat (EE) (0.17 and 0.16), respectively. Corn- and mixed DDGS

M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141 133

Table 2Non-starch polysaccharide profile of corn-, wheat-, and mixed DDGS (g/kg DM).

Corn DDGS Wheat DDGS Mixed DDGS

Meana Rangeb S.D. (CV) Meanc Ranged S.D. (CV) Meane Rangef S.D. (CV)

Total NSPTotal 283g (250–337) 20 (0.09) 262g (242–291) 9 (0.04) 247g (238–257) 8 (0.03)Soluble 31g (16–65) 8 (0.47) 67h (53–80) 1 (0.02) 65h (54–76) 8 (0.13)Cellulose 67g (52–91) 8 (0.16) 50h (35–67) 16 (0.32) 54g,h (51–59) 3 (0.06)

NCPXylose

Total 77g (67–100) 7 (0.10) 86g (70–93) 7 (0.08) 78g (74–85) 5 (0.07)Soluble 6g (1–16) 3 (0.62) 23h (15–32) 5 (0.22) 21h (17–25) 3 (0.12)

ArabinoseTotal 62g (56–72) 4 (0.07) 57g,h (51–62) 0 (0.00) 49h (46–52) 3 (0.05)Soluble 7g (2–15) 3 (0.45) 17h (12–22) 3 (0.15) 14g,h (12–17) 2 (0.13)

GlucoseTotal 28g (21–44) 4 (0.13) 33g (27–37) 1 (0.05) 28g (27–31) 2 (0.05)Soluble 3g (0–16) 4 (1.90) 11g (1–21) 10 (0.89) 8g (3–13) 3 (0.46)

MannoseTotal 17g (12–20) 2 (0.12) 16g (13–18) 2 (0.13) 20g (18–22) 2 (0.09)Soluble 6g (4–9) 1 (0.19) 7g (4–8) 1 (0.18) 13h (11–15) 2 (0.13)

GalactoseTotal 15g (13–21) 2 (0.11) 11h (10–12) 1 (0.11) 10h (10–10) 0 (0.02)Soluble 3g (2–5) 1 (0.29) 6h (4–7) 1 (0.18) 6h (6–6) 0 (0.05)

Uronic acidsTotal 16g (14–20) 1 (0.08) 8h (7–9) 1 (0.12) 7h (4–8) 2 (0.26)Soluble 5g (3–6) 1 (0.11) 3h (2–4) 0 (0.15) 3h (2–4) 1 (0.19)

Klason lignin 25g (15–47) 7 (0.26) 66h (44–93) 21 (0.32) 82h (74–90) 6 (0.07)A/X ratio 0.80g (0.71–0.85) 0.0 (0.05) 0.66h (0.62–0.70) 0.1 (0.09) 0.63h (0.61–0.67) 0.0 (0.04)UA/X ratio 0.20g (0.16–0.23) 0.0 (0.08) 0.09h (0.08–0.11) 0.0 (0.21) 0.08h (0.06–0.11) 0.0 (0.30)

A/X ratio, arabinose-to-xylose ratio; CV, coefficient of variation; NCP, non-cellulosic polysaccharides; NSP, non-starch polysaccharides; S.D., standarddeviation; UA/X ratio, uronic acid-to-xylose ratio.

a Average of means from 21 plants.b N = 47.c Average of means from 2 plants.d N = 11.e Mean from 1 plant.f N = 5.

g,hMeans that do not share a letter are significant different at P<0.05.

had CVs on moisture content of 0.10 and 0.14, respectively, and wheat- and mixed DDGS had CVs on CF content of 0.14 and0.10, respectively.

Detailed analysis of the NSP profile of 63 samples (47 corn DDGS, 11 wheat DDGS, and 5 mixed DDGS) is presented inTable 2. Corn DDGS had the greatest content of NSP (286 g/kg DM), followed by wheat- (262 g/kg DM), and mixed DDGS(247 g/kg DM). Wheat- and mixed DDGS did not differ (P>0.05) in S-NSP content (ranging from 53 to 80 g/kg DM), while amarkedly lesser content (P<0.001) was observed in corn DDGS (16–65 g/kg DM). The distribution of constituent sugars for allthree DDGS origins was in the order of xylose > arabinose > NCP-glucose > mannose > galactose. No difference was observedin arabinoxylan content (P=0.555) across the three DDGS origins with corn- (123–172 g/kg DM), wheat- (121–155 g/kg DM),and mixed DDGS (120–136 g/kg DM). Despite the comparable content of arabinoxylan, corn DDGS contained 5 g/kg DM(P=0.226) and 13 g/kg DM (P=0.022) more arabinose compared with wheat- and mixed DDGS, respectively, accompanied bya 8 g/kg DM lesser content of xylose compared with wheat DDGS, hence yielding a greater arabinose/xylose-ratio (A/X-ratio)in corn DDGS of 0.80 compared with that of wheat- (0.66, P<0.001), and mixed DDGS (0.63, P=0.002), respectively. Corn DDGScontained approximately twice as much uronic acids (16 g/kg DM, P<0.001), along with a greater cellulose content (67 g/kgDM) compared with both wheat- (50 g/kg DM, P=0.32), and mixed DDGS (54 g/kg DM, P=0.248). Furthermore, corn DDGScontained a markedly greater uronic acid/xylose-ratio (UA/X-ratio) of 0.20, compared with both wheat- (0.09, P<0.001), andmixed DDGS (0.08, P<0.001). The greatest Klason lignin content was observed in mixed DDGS (82 g/kg DM), compared withwheat- (66 g/kg DM, P=0.105), and corn DDGS (25 g/kg DM, P<0.001). All three DDGS origins had comparable contents oftotal mannose and galactose. Except the soluble mannose content (P<0.001), no differences in compositional profile wereobserved between wheat- and mixed DDGS (P>0.05). Regardless of DDGS origin, greater CVs were observed for the solubleconstituent sugars compared with the total constituents. Soluble-NCP-glucose had greatest CVs with values of 1.90, 0.89,and 0.46, for corn-, wheat-, and mixed DDGS, respectively. Overall, corn DDGS had greater CVs for both total- and solubleconstituents, followed by wheat- and mixed DDGS, respectively. For corn DDGS especially soluble xylose (S-Xyl) and solublearabinose (S-Ara) had high CVs with values of 0.62 and 0.45, respectively.

The PCA illustrated a clear clustering and separation between corn DDGS to one side, and wheat- and mixed cereal DDGS onthe other (Fig. 1A). The corresponding loading plot (Fig. 1B), indicated the components most responsible for the separation,

134 M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141

Fig. 1. Principal component analysis of DDGS samples of corn-, wheat, and mixed cereal origin: (A, B) scores and loadings plots of PCA model constructedfrom common constituent profiles as model variables (N = 138), (C, D) scores and loadings plots of PCA model constructed from NSP profile as modelsvariables (N = 63), (E, F) scores and loadings of PCA model constructed from combined common constituents and NSP profiles as models variables (N = 63).ADF, acid detergent fibre; AH, acid hydrolysis; Ara, arabinose; Cel, cellulose; DF, dietary fibre; EE, ether extract; Gal, galactose; Glu, NCP-glucose; I, insoluble;Lig, Klason lignin; Man, mannose; NDC, non digestible carbohydrate; NDF, neutral detergent fibre; S, soluble; T, total; Xyl, xylose.

M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141 135

Fig. 2. Principal component analysis of corn DDGS samples originating from five different ethanol plants in the US: (A, B) scores and loadings of PCAmodel constructed from combined common constituents and NSP profiles as models variables (N = 31). ADF, acid detergent fibre; AH, acid hydrolysis; Ara,arabinose; Cel, cellulose; DF, dietary fibre; EE, ether extract; Gal, galactose; Glu, NCP-glucose; I, insoluble; Lig, Klason lignin; Man, mannose; NDC, nondigestible carbohydrate; NDF, neutral detergent fibre; S, soluble; T, total; Xyl, xylose.

with corn DDGS differentiating due to a general greater content of fat, aNDFom, CF, and starch. Wheat and mixed DDGSdifferentiated from corn DDGS by a general greater content of ash, CP, and total sugars. The corn- and wheat DDGS samplesindicated a comparable variation between individual samples in the score plot. The PCA model illustrated clear separationand clustering between the three different DDGS origins based on the NSP characteristics, with only a minor overlap betweenwheat- and mixed DDGS (Fig. 1C). The corresponding loading plot revealed the constituents with the greatest effect on theseparations, with corn DDGS differentiating due to large content of insoluble- and total constituents along with greatercontent of cellulose, and wheat- and mixed DDGS differentiating due to greater content of soluble constituents and Klasonlignin (Fig. 1D). A combined PCA model on both NIRS- and NSP compositional data on 63 samples, revealed the clearestseparation among the three DDGS origins (Fig. 1E), compared with the two PCA models based individually on NIRS- and NSPcompositional data (Fig. 1A and C). The corresponding loading plot (Fig. 1F), strongly underlined the constituents differingthe most and being most responsible for the grouping between the three DDGS origins. Corn DDGS was greater in insoluble-and total NSP, cellulose, and starch, and wheat- and mixed DDGS greater in soluble NSPs, total sugars, and ash. Furthermore,the loading plot revealed that mixed DDGS differed from wheat DDGS by having a greater content of mannose, total sugars,ADF, and CP.

3.2. Compositional variation in corn DDGS from various ethanol plants

Table 3 presents the results of corn DDGS samples collected from 5 different ethanol plants; P1 (N = 11), P2 (N = 17), P3(N = 10), P4 (N = 10), and P5 (N = 8), of which 56 samples were analyzed for common constituents, and 31 samples for theNSP profile. For starch content P1, P4, and P5 had the greatest CVs of 0.10–0.11. Furthermore, P1 and P5 had greater CVs inmoisture content of 0.11 and 0.16, respectively, than P2–P4 and greater CVs were observed for total sugar content for P1(0.09) and P2 (0.11). Across all of the common constituents, the five plants had overall average CVs of 0.07, 0.05, 0.03, 0.05,and 0.05, for P1, P2, P3, P4, and P5, respectively.

The five plants had varying CVs with regard to S-NSPs with values of 0.07–0.49. However, only small variation in CVs wasobserved for T-NSPs (288–325 g/kg DM) with values of 0.01–0.04 across all five plants. The CVs were relatively consistentfor cellulose content among the different plants. On the other hand Klason lignin content differed with CVs of 0.13–0.22.The CVs with regard to soluble constituent sugars varied markedly between the five plants with overall CVs of 0.03–1.49.Each of the five plants had conserved A/X-ratios (0.71–0.79) with corresponding low CVs with values of 0.01–0.02, togetherwith conserved UA/X-ratios (0.17–0.23). There was an individual grouping between the five plants, indicating a conservedcompositional profile of DDGS from each ethanol plant (Fig. 2A), and a greater compositional variation of P1, followed by P5,P4, P2, and P3, respectively. The corresponding loading plot (Fig. 2B), illustrated that P1 differed from the remaining plantsby a greater content of total- and insoluble NSP (P<0.05), cellulose (P<0.05, except when compared with P2), and uronic acid(P<0.05, except when compared with P4), P4 differed mainly by a greater content of fat(EE) (P<0.05), P5 differed based on agreater content of ash (P<0.05), total sugars (P<0.05), and total NCP-glucose (P<0.05), while P3 differed from P2 mainly by agreater content of fat (AH) (P<0.05).

136 M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141

Tab

le

3C

omp

osit

ion

al

pro

file

s

of

corn

DD

GS

from

five

eth

anol

pla

nts

(g/k

g

DM

).

Plan

t

P1

(N

=

11;

11)a

Plan

t

P2

(N

=

17;

5)a

Plan

t

P3

(N

=

10;

5)a

Plan

t

P4

(N

=

10;

5)a

Plan

t

P5

(N

=

8;

5)a

Sam

pli

ng

per

iod

11

mon

ths

1½

mon

ths

20

day

s

20

day

s

8

day

s

Mea

n

Ran

ge

CV

Mea

n

Ran

ge

CV

Mea

n

Ran

ge

CV

Mea

n

Ran

ge

CV

Mea

n

Ran

ge

CV

Moi

stu

re

80e

(65–

91)

0.11

98b,

c(9

0–10

8)

0.05

103b

(95–

109)

0.04

87d

,e(7

8–93

)

0.05

90c,

d(7

8–12

4)

0.16

Fat

(EE)

86c

(79–

93)

0.06

87c

(81–

101)

0.05

84c

(79–

92)

0.04

92b

(83–

103)

0.06

83c

(80–

87)

0.02

Fat

(AH

)

106c

(101

–115

)

0.05

112b

(105

–126

)

0.04

106c

(102

–114

)

0.03

112b

(104

–122

)

0.05

105c

(100

–108

)

0.02

Cru

de

pro

tein

317c

(297

–337

)

0.05

351b

(337

–364

)

0.02

349b

(343

–356

)

0.01

320c

(309

–333

)

0.02

319c

(313

–332

)

0.02

Cru

de

fibr

e

85b

(74–

95)

0.08

76d

(73–

82)

0.03

81b,

c(7

8–86

)

0.03

83b,

c(8

0–86

)

0.03

79c,

d(7

7–81

)

0.02

Ash

68d

,e(6

2–73

)

0.06

65e

(61–

70)

0.04

70c,

d(6

8–73

)

0.02

72c

(66–

78)

0.05

83b

(74–

90)

0.06

Star

ch

51d

(43–

59)

0.10

69b

(65–

78)

0.05

39e

(36–

45)

0.07

58c

(49–

69)

0.10

64b

(57–

79)

0.11

Tota

l su

gars

72d

(62–

87)

0.09

64e

(54–

78)

0.11

89c

(86–

95)

0.03

94c

(85–

103)

0.06

102b

(92–

114)

0.07

aND

Fom

373b

(346

–397

)

0.04

345d

(330

–363

)

0.02

367b,

c(3

59–3

80)

0.02

359c

(350

–368

) 0.

02

362b,

c(3

57–3

65)

0.01

AD

F

113b

(108

–119

)

0.03

109b

(104

–116

)

0.03

111b

(108

–114

)

0.02

102c

(98–

107)

0.03

105c

(99–

114)

0.05

Tota

l NSP

Tota

l

325b

(313

–337

)

0.02

305c

(329

–349

)

0.04

293c

(290

–296

)

0.01

288c

(280

–299

)

0.03

294c

(283

–309

)

0.04

Solu

ble

29b

(18–

37)

0.19

37b

(18–

65)

0.49

29b

(23–

33)

0.15

25b

(23–

27)

0.07

31b

(26–

41)

0.18

Cel

lulo

se

79b

(74–

91)

0.06

74b,

c(6

5–79

)

0.08

67c

(64–

70)

0.03

68c

(64–

71)

0.04

68c

(64–

76)

0.07

NC

P Xyl

ose

Tota

l

94b

(88–

100)

0.04

88c

(84–

91)

0.03

80d

(80–

81)

0.01

79d

(74–

85)

0.06

78d

(75–

83)

0.04

Solu

ble

5b(1

–8)

0.48

8b(1

–16)

0.78

5b(2

–6)

0.33

4b(2

–6)

0.39

4b(2

–6)

0.34

Ara

bin

ose

Tota

l

69b

(65–

72)

0.04

62c

(60–

64)

0.03

63c

(62–

63)

0.01

61c

(59–

65)

0.05

59c

(57–

62)

0.04

Solu

ble

7b(3

–9)

0.25

8b(3

–14)

0.54

7b(5

–7)

0.13

6b(4

–7)

0.19

7b(6

–8)

0.15

Glu

cose

Tota

l

27c

(22–

29)

0.06

25c

(23–

25)

0.04

28c

(27–

29)

0.03

25c

(24–

28)

0.07

34b

(29–

44)

0.17

Solu

ble

1b(0

–5)

1.35

4b(0

–14)

1.41

2b(0

–5)

1.17

0b(0

–1)

1.49

3b(0

–8)

1.04

Man

nos

eTo

tal

17d

(14–

18)

0.06

19b,

c(1

8–20

)

0.04

18b,

c(1

7–19

)

0.05

18c

(17–

19)

0.04

20b

(19–

20)

0.03

Solu

ble

7c,d

(5–8

)

0.13

9b(8

–9)

0.09

6d(6

–7)

0.08

7c,d

(6–7

)

0.06

8b,c

(8–8

)

0.03

Gal

acto

seTo

tal

20b

(18–

21)

0.05

17c

(17–

18)

0.03

17c,

d(1

7–17

)

0.01

17c,

d(1

6–18

)

0.04

16d

(15–

17)

0.05

Solu

ble

3b(2

–4)

0.17

3b(2

–5)

0.31

3b(3

–3)

0.09

3b(3

–3)

0.03

3b(3

–4)

0.09

Uro

nic

acid

sTo

tal

19b

(18–

20)

0.04

17c

(16–

18)

0.06

18c

(17–

18)

0.02

18b,

c(1

7–18

)

0.03

17c

(16–

17)

0.02

Solu

ble

5b(5

–6)

0.08

5b(3

–6)

0.26

5b(5

–5)

0.04

5b(5

–6)

0.07

5b(4

–6)

0.08

Kla

son

lign

in

38b

(28–

47)

0.13

33b

(29–

42)

0.18

30b

(23–

41)

0.22

31b

(24–

38)

0.18

34b

(27–

42)

0.17

A/X

-rat

io

0.73

d(0

.71–

0.74

)

0.01

0.71

e(0

.70–

0.72

) 0.

01

0.78

b(0

.77–

0.80

)

0.01

0.77

b,c

(0.7

5–0.

79)

0.02

0.76

c(0

.75–

0.78

)

0.01

UA

/X-r

atio

0.20

c(0

.19–

0.21

)

0.03

0.20

c(0

.17–

0.22

) 0.

08

0.22

b(0

.21–

0.22

)

0.02

0.23

b(0

.22–

0.23

)

0.04

0.22

b(0

.20–

0.23

)

0.05

AD

F,

acid

det

erge

nt

fibr

e;

AH

, aci

d

hyd

roly

sis;

CV

, coe

ffici

ent

of

vari

atio

n;

EE, e

ther

extr

act;

NC

P,

non

-cel

lulo

sic

pol

ysac

char

ides

;

aND

Fom

, neu

tral

det

erge

nt

fibr

e.a

N

com

mon

con

stit

uen

ts

anal

ysis

;

N

NSP

anal

ysis

.b–

eM

ean

s

that

do

not

shar

e

a

lett

er

are

sign

ifica

nt

dif

fere

nt

at

P<0.

05.

M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141 137

Table 4Compositional profile of corn and corresponding DDGS (g/kg DM).

Corn (N = 11) Corn DDGS (N = 11) Ratioa

Mean Range S.D. (CV) Mean Range S.D. (CV)

Moisture 139b (130–146) 6 (0.04) 80c (65–91) 9 (0.11) –Fat (EE) 35b (34–37) 1 (0.03) 86c (79–93) 5 (0.06) 2.4Fat (AH) 39b (37–41) 1 (0.04) 106c (101–115) 5 (0.05) 2.7Crude protein 83b (75–91) 6 (0.08) 317c (297–337) 15 (0.05) 3.8Crude fibre 25b (23–27) 1 (0.04) 85c (74–95) 7 (0.08) 3.4Ash 10b (9–11) 1 (0.05) 68c (62–7-3) 4 (0.06) 6.5Starch 723b (705–736) 11 (0.01) 51c (43–59) 5 (0.10) –Total sugars 51b (46–59) 4 (0.08) 72c (62–87) 7 (0.09) 1.4aNDFom 93b (79–104) 8 (0.08) 373c (346–397) 15 (0.04) 4.1ADF 38b (33–44) 3 (0.09) 113c (108–119) 4 (0.03) 3.0Total NSPTotal 79b (67–91) 7 (0.08) 325c (313–337) 8 (0.02) 4.1Soluble 6b (2–10) 3 (0.39) 29c (18–37) 6 (0.19) 4.5Cellulose 17b (14–20) 2 (0.12) 79c (74–91) 5 (0.06) 4.6NCP

XyloseTotal 23b (20–27) 2 (0.09) 94c (88–100) 3 (0.04) 4.0Soluble 1b (0–1) 1 (0.97) 5c (1–8) 2 (0.48) 7.6

ArabinoseTotal 18b (15–20) 1 (0.07) 69c (65–72) 2 (0.04) 3.8Soluble 1b (0–2) 1 (0.63) 7c (3–9) 2 (0.25) 5.8

GlucoseTotal 7b (6–8) 1 (0.09) 27c (22–29) 2 (0.06) 3.7Soluble 1b (0–1) 1 (0.97) 1b (0–5) 2 (1.35) 2.4

MannoseTotal 2b (2–3) 0 (0.10) 17b (14–18) 1 (0.06) 7.5Soluble 1b (1–1) 0 (0.14) 7c (5–8) 1 (0.13) 6.8

GalactoseTotal 6b (4–7) 1 (0.13) 20c (18–21) 1 (0.05) 3.6Soluble 1b (1–2) 0 (0.38) 3c (2–4) 1 (0.17) 2.5

Uronic acidsTotal 5b (4–6) 1 (0.10) 19c (18–20) 1 (0.04) 3.5Soluble 1b (1–2) 0 (0.19) 5c (5–6) 0 (0.08) 3.5

Klason lignin 10b (7–15) 2 (0.23) 38c (28–47) 5 (0.13) 3.9A/X-ratio 0.77b (0.74.0.79) 0.0 (0.02) 0.73c (0.71–0.74) 0.0 (0.01)UA/X-ratio 0.23b (0.22–0.24) 0.0 (0.03) 0.20c (0.19–0.21) 0.0 (0.03)Cel./T-NSP ratio 0.21b (0.20–0.23) 0.0 (0.05) 0.24c (0.20–0.23) 0.0 (0.05)

ADF, acid detergent fibre; AH, acid hydrolysis; A/X ratio, arabinose-to-xylose ratio; CV, coefficient of variation; EE, ether extract; NCP, non-cellulosicpolysaccharides; aNDFom, neutral detergent fibre; NSP, non-starch polysaccharides; UA/X ratio, uronic acid-to-xylose ratio.

a Average corn DDGS-to-corn ratio.b,cMeans that do not share a letter are significant different at P<0.05.

3.3. Compositional variation between corn and corresponding corn DDGS

The results of the compositional profile of corn and corresponding DDGS, sampled simultaneously from the ethanol plantover a period of 11 month, are presented in Table 4. For common constituents, DDGS had an increase in all componentsbut starch, with average increase of 3.4-fold. The greatest increase was observed for the content of ash, aNDFom, and CP,whereas the least increase was observed for the content of total sugars, and fat, respectively. Furthermore, DDGS had greaterCVs of fat and CF content compared with the parent grain, whereas corn on the other hand had greater CVs of CP, ADF, andaNDFom content, across the 11 months. Comparable CVs of ash and total sugar content was observed between corn andcorresponding DDGS.

The NSP compositional profiles revealed that DDGS had a 4.1- and 4.5 times as high content of T-NSP and S-NSP as in corn,respectively. The S-Xyl and S-Ara content in DDGS increased to levels 7.6 and 5.8 times the content in corn, respectively.Furthermore, the content of total- and soluble mannose was markedly increased in DDGS compared with corn with valuesof 7.5- and 6.8-fold greater than in corn, respectively. The A/X- and UA/X-ratio were 0.04- (P<0.001) and 0.05 (P<0.001) unitsless in DDGS than in corn, respectively. On the other hand cellulose/NSP-ratio was 0.03 units greater (P<0.001) for DDGSthan in corn. Corn DDGS had reduced CVs of T-NSP (0.06 units) and S-NSP (0.20 units) than the corresponding corn, andexcept for soluble NCP-glucose, this was the case for all constituent sugars, both total- and soluble content.

4. Discussion

In line with our expectations, the nutrient composition of DDGS varied between DDGS from both different ethanol plantsand DDGS originating from different parent grains (Spiehs et al., 2002; Olukosi and Adebiyi, 2013). Furthermore, as expected,

138 M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141

we also observed that the nutrient composition of the DDGS in part reflected the composition in the parent grain, as onlystarch was removed during the fermentation process (Widyaratne and Zijlstra, 2007; Gibreel et al., 2011). The greater contentof fat in corn DDGS compared with wheat DDGS, along with the greater content of protein in wheat DDGS than in corn DDGScorresponds to previous characterizations of content in parent grains (Belitz et al., 2009). Overall, the major constituentanalysis of the DDGS corresponds to previous published results (Cromwell et al., 1993; Belyea et al., 2004; Widyaratne andZijlstra, 2007; Kim et al., 2008; Liu, 2008). The cereals used in the production of the mixed DDGS are unknown, however themajor cereal fraction is presumably wheat, considering the similar composition of wheat- and mixed DDGS, based on thePCA-model and compositional analysis.

Similar to previous studies describing the NSP profiles of corn and wheat (Bach Knudsen, 1997), and corn- and wheatDDGS (Widyaratne and Zijlstra, 2007), we also observed greater content of soluble NSPs in wheat DDGS compared withcorn DDGS, and a greater content of insoluble NSPs in corn DDGS compared with wheat DDGS. It can be speculated that theslightly greater NCP-glucose content of wheat DDGS compared with corn DDGS is caused by the greater content of �-glucanpresent in wheat compared with corn, as described in literature (Bach Knudsen, 1997). However, it has been speculated thatalso �-glucan and mannans from yeast are present in the DDGS (de Vries et al., 2013). Both corn- and wheat DDGS containedsimilar amounts of total arabinoxylan (approximately 140 g/kg DM). However, the greater content of both arabinose anduronic acids present in corn DDGS compared with wheat DDGS, yielding greater A/X- and UA/X-ratios, indicate a morecomplex structure of the heteroxylan in corn DDGS than in wheat DDGS. This is in line with previous studies reporting agreater branch density and complexity of corn arabinoxylan compared with that of wheat (Bedford, 1995; Saulnier et al.,1995a; Jilek and Bunzel, 2013; Yang et al., 2013).

The markedly greater content of Klason lignin present in wheat DDGS compared with corn DDGS is also in line withprevious reported results present in parent grains (Bach Knudsen, 1997; Bunzel et al., 2011). However, the Klason lignincontent might overestimate the “true” lignin (complex phenolic polymers) content of the DDGS. Klason lignin is not a welldefined chemical matter, but an empirical residue consisting of materials not solubilized by 12 mol/L sulfuric acid (Davinet al., 2008). Klason lignin determined in corn and wheat insoluble fibre have been shown to include, besides true lignin;N (proteins), residual fat and waxes, as well as cutin, with true lignin content equal to approximately 1/3 and 1/4 of themeasured Klason lignin content for wheat and corn, respectively (Bunzel et al., 2011). However, other nitrogen sources thanproteins, such as Maillard products, can potentially contribute to nitrogen content of the Klason lignin fraction. The Maillardreaction can occur between reducing sugars and lysine residues during the processing of DDGS, as a result of the additionof condensed solubles to the wet distillers cake during drying, consequently damaging the protein fraction (Fastinger andMahan, 2006; Pahm et al., 2009; Kim et al., 2012). Furthermore, changes in the corn protein structure during the ethanolproduction have been described to induce a greater fraction of DDGS protein associated with the cell wall material, comparedwith that of corn (Yu and Nuez-Ortin, 2010). The herein presented high CV values of Klason lignin content in both corn- (0.26)and wheat DDGS (0.32) from the different ethanol plants along with their corresponding different processing technologies,further indicates the presence of non-lignin sources, such as Maillard products, may be present in the Klason lignin fractionof DDGS. Regardless of DDGS origin, the greatest CV values were observed for the S-NSP constituent sugar monomers. It canbe speculated that this observation is related to the amount of condensed solubles added to the wet distillers cake duringdrying. However, it should also be noted that the relative small S-NSP content present could directly give rise to greater CVs.Furthermore, degradation of insoluble NSP occurring during the fermentation process could increase the content of solubleNSP, due to the presence of endogenous fibre degrading enzymes from yeast, exogenous fibre degrading enzymes added toincrease ethanol yield, and mechanical- and heat (pre)treatment.

When sampled over a period of 8 days to 11 months the five ethanol plants showed, in spite of some variation, capability ofproducing DDGS with conserved compositional profiles, indicated by PCA models. The variation in starch content, especiallyobserved for three of the five plants, is directly related to the effectiveness of the fermentation process at each individualethanol plant, which varies according to process technology and fermentation time, yeast, and enzymes used in the produc-tion. Furthermore, the relatively high CV in moisture content is related to different intensities of the DDGS drying process,which is a crucial step in the DDGS production, due to the negative effects on nutritional value associated with excessive heatdamage of especially lysine, and the high energy expenses associated with drying. Similarly large variation was observedfor the content of Klason lignin, as discussed above. Despite the consistent content of total NSP and total constituent sugarsfrom each individual ethanol plant, large variation was observed for the S-NSPs fractions, as discussed above. P1 showedgreater overall variation in compositional profile, which could be explained by P1 having the far greatest sampling periodof 11 months, hence probably encountering greater variation in raw material. Despite the narrowest sampling period of 8days, P5 on the other hand also showed relatively large overall variation.

The simultaneous sampling of both corn and corresponding DDGS over 11 months allowed for investigating the changesin compositional profile after the fermentation and drying process. Although the corn and the DDGS were sampled simul-taneously, it should be noted that the corn and DDGS did not originate from the same fermentation batch, which wouldbe practically impossible to obtain. The observed changes between corn and corresponding DDGS may not be consideredcommon for all ethanol plants, underlined by the presented compositional variations between the 24 corn DDGS plants. Asexpected, the DDGS reflected the nutrient content of the parent grain (Spiehs et al., 2002). All other constituents than starchwere concentrated in the DDGS compared with corn, with an average 3.4-fold increase for the common constituents, similarto previous studies (Urriola et al., 2010; de Vries et al., 2013). The relatively small increase in fat content of 2.4–2.7 fold mayindicate that the ethanol plant have extracted some of the oil from the thin stillage before condensing and blending with

M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141 139

the grain fraction and drying to make DDGS. The marked increase in mannose content in DDGS compared with corn, mostlikely originate from yeast cell wall mannans, as discussed previously (de Vries et al., 2013). Furthermore, a marked shiftin solubility of xylose and arabinose was observed for DDGS, indicating a modification of the arabinoxylan fraction duringprocessing. Corn showed a consistently greater content of substituted xylan with both greater A/X- and UA/X-ratios, com-pared with corresponding DDGS. Unpublished results from our research group indicate the A/X-ratio in corn differs amongthe different botanical grain fractions, with endosperm having a markedly greater A/X-ratio compared with that of cornbran, which is in line with previous published results in corn grain, flour and bran (Bach Knudsen, 1997; Rose et al., 2009).Similarly, UA/X-ratio has been reported greater in corn flour, compared with corn bran (Bach Knudsen, 1997). The shift inA/X- and UA/X ratio between corn and DDGS indicates modification of the greater substituted xylan (e.g. in endosperm)during the fermentation process. It must be noted, however, that the total content of arabinoxylan and uronic acids presentin corn endosperm is very low compared with corn bran.

Though highly substituted, corn endosperm arabinoxylan is not cross-linked to the same degree as corn bran heteroxylans(Saulnier and Thibault, 1999; Bunzel et al., 2006). Hence, corn endosperm arabinoxylan is potentially more susceptible toenvironmental changes, such as heat processing and presence of fibre degrading enzymes and other accessory enzymes,which might lead to an increased degradation and solubilization of arabinoxylan. It can be speculated that the solubilizedarabinoxylan is more susceptible to participate in Maillard reactions, by having a potentially larger fraction of reducing endsdue to hydrolysis. The arabinoxylan parting in Maillard reactions would be unable to determine as part of the constituentsugars in the NSP-procedure used in this study, since they would most likely end up in the Klason lignin fraction. On the otherhand, cellulose appear to remain unmodified during processing, due to its strong rigid structure and anchorage in the cellwall matrix (Van Eylen et al., 2011), which consequently leads to the greater cellulose/NSP-ratio for DDGS compared withcorn in the current study. Endosperm arabinoxylan comprise approximately 20% of the total content of arabinoxylan in corn,calculated with data from Bach Knudsen (1997) and Watson (1987). Hence, considering that the endosperm arabinoxylanonly comprise a minor fraction of the total arabinoxylan content of the grain, and that the cellulose fraction is concentratedin DDGS, the outcome will be that the majority of NSP remain unmodified or potentially more complex in DDGS than in theparent grain.

The use of multivariate data analysis, PCA, on compositional data of DDGS proved useful to visually distinguish betweenDDGS covering three different feedstock origins; corn, wheat, and mixed cereals. Furthermore, PCA was able to provideinformation concerning the most conserved components of each DDGS origin, hence the components most responsible forthe individual clusters and groupings of DDGS. By combining both compositional data of the common constituents and NSPprofiles of the DDGS, a more clear separation between the groups was observed.

DDGS has a reputation of having variable nutrient composition and protein quality, and a high content of mycotoxins,which has limited its use in swine feed (Stein et al., 2006; Pedersen et al., 2007; Anderson et al., 2012). High quantities ofDDGS in feed increase the content of dietary fibre, associated with negative effects on nutrient digestibility. To increasenutrient digestibility of feed formulations containing DDGS, studies have investigated the effects of adding fibre degradingenzymes to animal diets containing DDGS (Jones et al., 2010; Yoon et al., 2010; Yánez et al., 2011). However, the results areinconclusive. This may relate to the complexity of the cell wall matrix in corn DDGS, and that the most readily degradablearabinoxylan for the fibre degrading enzymes already have been modified during DDGS production, as discussed above. Theseobservations indicate that the fibre degrading enzymes applied for degradation of corn DDGS need to be targeted towardshighly complex substrates. With respect to enzymatic degradation, corn bran has been acknowledged as a recalcitrantsubstrate, as a consequence of the highly branched structure of the arabinoxylan (Saulnier et al., 2001; Agger et al., 2010).The cross-linking and complexity of the NSP matrix is influenced by many factors including; content of substituted xylan,ferulic acid cross-linking, and lignin and structural proteins interacting with arabinoxylan (Saulnier et al., 1995b). Despite theethanol plants’ capability of individually producing DDGS with a conserved NSP profile, variation in the degree of substitutedxylan from plant to plant exists. For corn DDGS, the differences in A/X- (0.71–0.85) and UA/X-ratio (0.16–0.23) underline theheterogeneous structure of DDGS NSPs. It can be speculated that the enzymatic degradation potential of the DDGS NSPs issterically hindered by a greater proportion of substituted xylan (Rose et al., 2009). Hence, the DDGS samples from 24 differentethanol plants analyzed in this study might yield different results in relation to digestibility and enzymatic degradationpotential of the NSPs. Difference in animal performance may be observed when feeding either corn- or wheat DDGS, dueto the observed large variation in NSP cell wall complexity between the two DDGS sources likely affecting digestibility.However, it should be noted that the overall degradation potential of the DDGS NSPs, is affected by a combination of severalfactors (as described above), a number of which not determined in this study.

5. Conclusion

The current study showed a large variability in the chemical composition of common constituents, and in non-starchpolysaccharide profiles of corn-, wheat, and mixed DDGS. Despite variations, the major fractions across all origins were; CPand aNDFom with a large degree of the aNDFom fraction being ADF. All samples also had a large content of sugars, whilecorn DDGS contained more fat than the other two. Wheat- and mixed DDGS showed a markedly greater content of S-NSPthan in corn, which on the other hand had greater contents of total- and insoluble NSPs. Of the NSP fraction, all samples hadthe greatest content of arabinoxylan. However, the A/X- and UA/X-ratio differed with greater ratios for corn DDGS than inthe other two. DDGS samples from five different ethanol plants showed that each plant were capable, despite variations,

140 M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141

of producing DDGS with an individual conserved compositional profile. Finally, analysis of corn and corresponding DDGSsamples showed that the content of substituted xylan and S-NSP is modified during the production of DDGS.

Conflict of interest

The authors declare that there are no conflicts of interest.

Acknowledgements

The following people are acknowledged for their assistance in collecting DDGS samples; M. Chow, D. KO, A. Riley, D. Ward,G. Jones, C. Miller-Fosmore, and E. Seward. The authors would also like to thank the technical staff, especially L. Märcher andS.G. Handberg for assistance. This project was partially supported by The Danish Ministry of Science, Innovation and HigherEducation.

References

Agger, J., Vikso-Nielsen, A., Meyer, A.S., 2010. Enzymatic xylose release from pretreated corn bran arabinoxylan: differential effects of deacetylation anddeferuloylation on insoluble and soluble substrate fractions. J. Agric. Food Chem. 58, 6141–6148.

Anderson, P.V., Kerr, B.J., Weber, T.E., Ziemer, C.J., Shurson, G.C., 2012. Determination and prediction of digestible and metabolizable energy from chemicalanalysis of corn coproducts fed to finishing pigs. J. Anim. Sci. 90, 1242–1254.

Bach Knudsen, K.E., 1997. Carbohydrate and lignin contents of plant materials used in animal feeding. Anim. Feed Sci. Technol. 67, 319–338.Bandegan, A., Guenter, W., Hoehler, D., Crow, G.H., Nyachoti, C.M., 2009. Standardized ileal amino acid digestibility in wheat distillers dried grains with

solubles for broilers. Poult. Sci. 88, 2592–2599.Batal, A.B., Dale, N.M., 2006. True metabolizable energy and amino acid digestibility of distillers dried grains with solubles. J. Appl. Poult. Res. 15, 89–93.Bedford, M.R., 1995. Mechanism of action and potential environmental benefits from the use of feed enzymes. Anim. Feed Sci. Technol. 53, 145–155.Belitz, H.D., Grosch, W., Schieberle, P., 2009. Food Chem., fourth ed. Berlin, Springer.Belyea, R.L., Rausch, K.D., Tumbleson, M.E., 2004. Composition of corn and distillers dried grains with solubles from dry grind ethanol processing. Bioresour.

Technol. 94, 293–298.Bunzel, M., Ralph, J., Brüning, P., Steinhart, H., 2006. Structural identification of dehydrotriferulic and dehydrotetraferulic acids isolated from insoluble

maize bran fiber. J. Agric. Food Chem. 54, 6409–6418.Bunzel, M., Schüßler, A., Tchetseubu Saha, G., 2011. Chemical characterization of klason lignin preparations from plant-based foods. J. Agric. Food Chem.

59, 12506–12513.Cozannet, P., Lessire, M., Gady, C., Métayer, J.P., Primot, Y., Skiba, F., Noblet, J., 2010. Energy value of wheat dried distillers grains with solubles in roosters,

broilers, layers, and turkeys. Poult. Sci. 89, 2230–2241.Cromwell, G.L., Herkelman, K.L., Stahly, T.S., 1993. Physical, chemical, and nutritional characteristics of distillers dried grains with solubles for chicks and

pigs. J. Anim. Sci. 71, 679–686.Davin, L.B., Jourdes, M., Patten, A.M., Kim, K.W., Vassao, D.G., Lewis, N.G., 2008. Dissection of lignin macromolecular configuration and assembly: comparison

to related biochemical processes in allyl/propenyl phenol and lignan biosynthesis. Nat. Prod. Rep. 25, 1015–1090.de Vries, S., Pustjens, A.M., Kabel, M.A., Salazar-Villanea, S., Hendriks, W.H., Gerrits, W.J.J., 2013. Processing technologies and cell wall degrading enzymes

to improve nutritional value of dried distillers grain with solubles for animal feed: an in vitro digestion study. J. Agric. Food Chem. 61, 8821–8828.Fastinger, N.D., Mahan, D.C., 2006. Determination of the ileal amino acid and energy digestibilities of corn distillers dried grains with solubles using

grower-finisher pigs. J. Anim. Sci. 84, 1722–1728.Gibreel, A., Sandercock, J.R., Lan, J., Goonewardene, L.A., Scott, A.C., Zijlstra, R.T., Curtis, J.M., Bressler, D.C., 2011. Evaluation of value-added components of

dried distiller’s grain with solubles from triticale and wheat. Bioresour. Technol. 102, 6920–6927.Hartley, R.D., 1973. Carbohydrate esters of ferulic acid as components of cell-walls of Lolium multiflorum. Phytochem 12, 661–665.Jilek, M.L., Bunzel, M., 2013. Dehydrotriferulic and dehydrodiferulic acid profiles of cereal and pseudocereal flours. Cereal Chem. J. 90, 507–514.Jones, C.K., Bergstrom, J.R., Tokach, M.D., DeRouchey, J.M., Goodband, R.D., Nelssen, J.L., Dritz, S.S., 2010. Efficacy of commercial enzymes in diets containing

various concentrations and sources of dried distillers grains with solubles for nursery pigs. J. Anim. Sci. 88, 2084–2091.Kim, B.G., Kil, D.Y., Zhang, Y., Stein, H.H., 2012. Concentrations of analyzed or reactive lysine, but not crude protein, may predict the concentration of

digestible lysine in distillers dried grains with solubles fed to pigs. J. Anim. Sci. 90, 3798–3808.Kim, Y., Mosier, N.S., Hendrickson, R., Ezeji, T., Blaschek, H., Dien, B., Cotta, M., Dale, B., Ladisch, M.R., 2008. Composition of corn dry-grind ethanol by-products:

DDGS, wet cake, and thin stillage. Bioresour. Technol. 99, 5165–5176.Lapierre, C., Pollet, B., Ralet, M.C., Saulnier, L., 2001. The phenolic fraction of maize bran: evidence for lignin–heteroxylan association. Phytochemistry 57,

765–772.Lequart, C., Nuzillard, J.-M., Kurek, B., Debeire, P., 1999. Hydrolysis of wheat bran and straw by an endoxylanase: production and structural characterization

of cinnamoyl-oligosaccharides. Carbohydr. Res. 319, 102–111.Liu, K., 2008. Particle size distribution of distillers dried grains with solubles (DDGS) and relationships to compositional and color properties. Bioresour.

Technol. 99, 8421–8428.Liu, K., 2011. Chemical composition of distillers grains, a review. J. Agric. Food Chem. 59, 1508–1526.Nyachoti, C.M., House, J.D., Slominski, B.A., Seddon, I.R., 2005. Energy and nutrient digestibilities in wheat dried distillers’ grains with solubles fed to growing

pigs. J. Sci. Food Agric. 85, 2581–2586.Olukosi, O.A., Adebiyi, A.O., 2013. Chemical composition and prediction of amino acid content of maize- and wheat-distillers’ dried grains with soluble.

Anim. Feed Sci. Technol. 185, 182–189.Pahm, A.A., Pedersen, C., Stein, H.H., 2009. Standardized ileal digestibility of reactive lysine in distillers dried grains with solubles fed to growing pigs. J.

Agric. Food Chem. 57, 535–539.Pedersen, C., Boersma, M.G., Stein, H.H., 2007. Digestibility of energy and phosphorus in ten samples of distillers dried grains with solubles fed to growing

pigs. J. Anim. Sci. 85, 1168–1176.Piber, M., Koehler, P., 2005. Identification of dehydro-ferulic acid-tyrosine in rye and wheat: evidence for a covalent cross-link between arabinoxylans and

proteins. J. Agric. Food Chem. 53, 5276–5284.Rose, D.J., Patterson, J.A., Hamaker, B.R., 2009. Structural differences among alkali-soluble arabinoxylans from maize (Zea mays), rice (Oryza sativa), and

wheat (Triticum aestivum) brans influence human fecal fermentation profiles. J. Agric. Food Chem. 58, 493–499.Saulnier, L., Vigouroux, J., Thibault, J.-F., 1995a. Isolation and partial characterization of feruloylated oligosaccharides from maize bran. Carbohydr. Res. 272,

241–253.Saulnier, L., Marot, C., Chanliaud, E., Thibault, J.-F., 1995b. Cell wall polysaccharide interactions in maize bran. Carbohydr. Polym. 26, 279–287.

M.B. Pedersen et al. / Animal Feed Science and Technology 197 (2014) 130–141 141

Saulnier, L., Thibault, J.-F., 1999. Ferulic acid and diferulic acids as components of sugar-beet pectins and maize bran heteroxylans. J. Sci. Food Agric. 79,396–402.

Saulnier, L., Marot, C., Elgorriaga, M., Bonnin, E., Thibault, J.F., 2001. Thermal and enzymatic treatments for the release of free ferulic acid from maize bran.Carbohydr. Polym. 45, 269–275.

Shewry, P.R., Hawkesford, M.J., Piironen, V., Lampi, A.-M., Gebruers, K., Boros, D., Andersson, A.A.M., Åman, P., Rakszegi, M., Bedo, Z., Ward, J.L., 2013. Naturalvariation in grain composition of wheat and related cereals. J. Agric. Food Chem. 61, 8295–8303.

Spiehs, M.J., Whitney, M.H., Shurson, G.C., 2002. Nutrient database for distiller’s dried grains with solubles produced from new ethanol plants in Minnesotaand South Dakota. J. Anim. Sci. 80, 2639–2645.

Stein, H.H., Gibson, M.L., Pedersen, C., Boersma, M.G., 2006. Amino acid and energy digestibility in ten samples of distillers dried grain with solubles fed togrowing pigs. J. Anim. Sci. 84, 853–860.

Stein, H.H., Shurson, G.C., 2009. Board-invited review: the use and application of distillers dried grains with solubles in swine diets. J. Anim. Sci. 87,1292–1303.

Urriola, P.E., Shurson, G.C., Stein, H.H., 2010. Digestibility of dietary fiber in distillers coproducts fed to growing pigs. J. Anim. Sci. 88, 2373–2381.Van Eylen, D., van Dongen, F., Kabel, M., de Bont, J., 2011. Corn fiber, cobs and stover: enzyme-aided saccharification and co-fermentation after dilute acid

pretreatment. Bioresour. Technol. 102, 5995–6004.Watson, S.A., 1987. Structure and composition. In: Watson, S.A., Ramstad, P.E. (Eds.), Corn: Chemistry and Technology. Am. Assoc. Cereal Chem., St. Paul,

MN.Widyaratne, G.P., Zijlstra, R.T., 2007. Nutritional value of wheat and corn distiller’s dried grain with solubles: digestibility and digestible contents of energy,

amino acids and phosphorus, nutrient excretion and growth performance of grower-finisher pigs. Can. J. Anim. Sci. 87, 103–114.Yánez, J.L., Beltranena, E., Cervantes, M., Zijlstra, R.T., 2011. Effect of phytase and xylanase supplementation or particle size on nutrient digestibility of diets

containing distillers dried grains with solubles cofermented from wheat and corn in ileal-cannulated grower pigs. J. Anim. Sci. 89, 113–123.Yang, J., Maldonado-Gómez, M.X., Hutkins, R.W., Rose, D.J., 2013. Production and in vitro fermentation of soluble, non-digestible, feruloylated oligo- and

polysaccharides from maize and wheat brans. J. Agric. Food Chem.Yoon, S.Y., Yang, Y.X., Shinde, P.L., Choi, J.Y., Kim, J.S., Kim, Y.W., Yun, K., Jo, J.K., Lee, J.H., Ohh, S.J., Kwon, I.K., Chae, B.J., 2010. Effects of mannanase and

distillers dried grain with solubles on growth performance, nutrient digestibility, and carcass characteristics of grower-finisher pigs. J. Anim. Sci. 88,181–191.

Yu, P., Nuez-Ortin, W.G., 2010. Relationship of protein molecular structure to metabolisable proteins in different types of dried distillers grains with solubles:a novel approach. Br. J. Nutr. 104, 1429–1437.

43

PAPER II

Ferulic Acid Dehydrodimer and –

Dehydrotrimer Profiles of Distillers Dried

Grains with Solubles from Different Cereal

Species

Mads B. Pedersen, Mirko Bunzel, Judith Schäfer,

Knud Erik B. Knudsen, Jens F. Sørensen, Shukun

Yu, and Helle N. Lærke

Published in:

Journal of Agricultural and Food Chemistry, 2015,

63:2006-2012

44

Ferulic Acid Dehydrodimer and Dehydrotrimer Profiles of Distiller’sDried Grains with Solubles from Different Cereal SpeciesMads B. Pedersen,*,†,‡ Mirko Bunzel,§ Judith Schafer,§ Knud Erik B. Knudsen,‡ Jens F. Sørensen,†

Shukun Yu,† and Helle N. Lærke‡

†Enzyme Development, DuPont Industrial Biosciences ApS, Edwin Rahrs Vej 38, DK-8220 Brabrand, Denmark‡Department of Animal Science, Aarhus University, Blichers Alle 20, DK-8830 Tjele, Denmark§Department of Food Chemistry and Phytochemistry, Karlsruhe Institute of Technology (KIT), Kaiserstrasse 12, 76131 Karlsruhe,Germany

ABSTRACT: Ferulic acid dehydrodimers (DFA) and dehydrotrimers (TriFA) ester-linked to plant cell wall polymers maycross-link not only cell wall polysaccharides but also other cell wall components including proteins and lignin, thus enhancing therigidity and potentially affecting the enzymatic degradation of the plant cell wall. Corn, wheat, and mixed-cereal distiller’s driedgrains with solubles (DDGS) were investigated for composition of DFAs and TriFAs by reversed phase high-performance liquidchromatography with ultraviolet detection. Corn DDGS contained 5.3 and 5.9 times higher contents of total DFAs than wheatand mixed-cereal DDGS, respectively. Furthermore, the contents of total TriFAs were 5.7 and 6.3 times higher in corn DDGSthan in wheat and mixed-cereal DDGS, respectively. In addition, both corn grains and corresponding DDGS had similar profilesof individual DFAs and TriFAs, indicating that ferulic acid cross-links in the corn cell wall are presumably not modified duringfermentation and DDGS processing.

KEYWORDS: corn, dehydrodimers, dehydrotrimers, distiller’s dried grains with solubles (DDGS), ferulic acid, wheat

■ INTRODUCTION

Distiller’s dried grains with solubles (DDGS) is a coproductfrom the grain-based ethanol industry mainly used as animalfeed.1 The production of DDGS includes several major steps,including grinding, liquefaction, saccharification, fermentation,distillation, and recovery of coproducts. Due to the removal ofmost of the starch fraction during the dry-grind ethanolfermentation process, DDGS has 3−3.5 times higher nutrient(e.g., protein, minerals, and vitamins) and nonstarch poly-saccharides (NSP) contents than the parent grains such as cornand wheat.1,2 The NSP content in DDGS makes up 250−300g/kg of dry matter (DM) with arabinoxylan and cellulose as thetwo major NSP fractions.2 When fed in high level tononruminants, the high content of NSP in DDGS may causenutritional challenges as the nonruminant digestive system isincapable of digesting NSP.In cereals, esterified hydroxycinnamic acids such as ferulic

acid, p-coumaric acid, and sinapic acid can participate inphotochemically or oxidatively induced coupling of arabinox-ylans along with cell wall trapped proteins and lignin-likepolymers.3−5 Whereas p-coumaric acid is mainly bound tolignin, ferulic acid and its derivatives, which are thequantitatively most important cross-links in the cereal cellwall, are primarily bound to arabinoxylans and pectins.6

Ferulate in particular has been found to form dimers (DFA),trimers (TriFA), and higher ferulate oligomers potentiallycross-linking two or more polysaccharide chains, thusstrengthening the cell wall with vital implications for plantphysiology, food science, nutrition, and the plant’s defensemechanisms against pathogens.6−8 DFA cross-links aredescribed as an early barrier against the fungus Fusarium

graminearum and the resistance of maize to Giberella ear rot9

and to Giberella stalk rot.10 Ferulate-based cross-links impairthe enzymatic degradation of plant cell walls11,12 and, thus,affect DDGS digestibility in relation to animal nutrition.To our knowledge there are no previous papers describing

the DFA and TriFA profiles of DDGS or the possibleimplications of the fermentation process on the DFA andTriFA profiles, despite their potentially large implications foranimal nutrition. In addition, because compositional variationamong DDGS samples may induce challenges when applied inanimal feed, the broad extent of DDGS samples analyzed in thisstudy will elucidate these potential implications. Character-ization of the DFA and TriFA composition in DDGS ofdifferent cereal species may contribute to the overall under-standing and interpretation of opportunities and limitations inDDGS degradation and digestibility, which is particularly usefulfor feed manufacturers and enzyme producers targeting DDGSdegradation.The current study describes the compositional profile and

extent of variation of DFA and TriFA contents in a wide varietyof DDGS samples from three different cereal origins: corn (Zeamays), wheat (Triticum aestivum), and a mixture of differentcereal species, including wheat (T. aestivum), barley (Hordeumvulgare L.), rye (Secale cereale), and triticale (a hybrid of wheatand rye). A total of 63 DDGS samples covering 24 differentethanol plants from North America and Europe along with 11

Received: October 24, 2014Revised: January 22, 2015Accepted: February 8, 2015Published: February 9, 2015

Article

pubs.acs.org/JAFC

© 2015 American Chemical Society 2006 DOI: 10.1021/jf505150gJ. Agric. Food Chem. 2015, 63, 2006−2012

unprocessed corn grain samples were analyzed by reversedphase high-performance chromatography (RP-HPLC) withultraviolet (UV) detection for the identification of 12 differentcell wall-bound DFAs and TriFAs.

■ MATERIALS AND METHODSMaterials and Instrumentation. Standard compounds of 8-

8(aryltetralin)-DFA, 8-8-DFA, 5-5-DFA, 8-O-4-DFA, 8-5-(benzofuran)-DFA, 8-5-DFA, 8-5(decarboxylated)-DFA, 5-5/8-O-4-TriFA, 8-8(aryltetralin)/8-O-4-TriFA, 8-O-4/8-O-4-TriFA were iso-lated as described by Jilek and Bunzel8 and Bunzel et al.13 The 5-5(methylated)-DFA (monomethylated 5-5-dehydrodiferulic acid) asinternal standard8 was kindly provided by Prof. Dr. John Ralph,University of Wisconsin, Madison, WI, USA.An analytical phenyl-hexyl HPLC column (250 × 4.6 mm, 5 μm

particle size) was purchased from Phenomenex (Germany). RP-HPLCwas carried out with the following instrumentation for DFA and TriFAanalysis: LC-20AT pumps, CTO-20AC column oven, SPD-M20Aphotodiode array detector (PDA), SIL-20AC autosampler, DGU-20Adegasser, and CBM-20A communication bus module from Shimadzu,Japan. All solvents used were of HPLC grade, and water was distilledand deionized.Sample Materials. A total of 63 DDGS samples were collected

from 24 ethanol plants in the United States and European Union(EU), covering the following three parent grain origins: corn (Z.mays), wheat (T. aestivum), and a mix of cereals grains [wheat (T.aestivum), barley (H. vulgare L.), rye (S. cereale)], and triticale (ahybrid of wheat and rye). In addition, 11 unprocessed corn grain (Z.mays) samples were also collected. All samples have previously beencharacterized for contents of common constituents and NSP profilesin our laboratory.2

A total of 47 corn DDGS samples were collected from 21 differentethanol plants in the United States; 11 DDGS samples along with 11unprocessed corn grain samples were sampled over 11 months with 1sample every month from a plant in Nevada (P1); 5 samples weresampled over 1.5 months from a plant in Minnesota (P2); 5 sampleswere sampled over 20 days from another plant in Minnesota (P3); 5samples were sampled over 20 days from a plant in Wisconsin (P4); 5samples were sampled over 8 days from a plant in Iowa (P5); and 16samples were supplied by a DDGS supplier in Iowa, representingsamples from 7 plants in Minnesota, 4 plants in Iowa, 2 plants inNevada, and 1 plant each in Illinois, Indiana, and Kentucky.A total of 11 wheat DDGS samples were collected from 2 different

ethanol plants in the EU, with 6 samples from a plant in the UnitedKingdom sampled over a period of approximately 3 months and 5samples from a plant in France sampled over 20 days.Finally, 5 DDGS samples of a mixed cereal grains were sampled

over 20 days from a plant in Sweden. The parent grains used in the

production of the mixed DDGS were wheat, triticale, barley, and rye;however, their individual proportions are unknown.

All 63 DDGS samples and 11 unprocessed corn grain samples weremilled (0.5 mm) prior to analysis using a centrifugal mill (RetschZM200).

Pretreatment, Alkaline Hydrolysis, and Extraction ofPhenolic Acids. Samples of 50, 175, or 200 mg were weighed outfor corn DDGS, corn, and wheat and mixed-cereal DDGS,respectively. All samples were washed twice with 10 mL of 80% (v/v) ethanol and twice with 10 mL of acetone followed by centrifugationand removal of supernatant between and after extractions. Theresidues were dried under a fume hood. Saponification (5 mL of 2mol/L NaOH) of the residues was carried out under nitrogen,protected from light, and under continuous stirring for 18 h, followinga procedure described by Dobberstein and Bunzel.7 Followingacidification with 0.95 mL of concentrated HCl, 25 μg of 5-5(methylated)-DFA was added as internal standard for thequantification of DFAs and TriFAs. Extraction was carried out threetimes with 4 mL of diethyl ether. The diethyl ether extracts were driedunder nitrogen and redissolved in 500 μL of THF/water (50:50 (v/v))using ultra sonication and used directly for the determination of DFAsand TriFAs.

Phenyl-Hexyl RP-HPLC-PDA Analysis of DFAs and TriFAs.The procedure basically followed a protocol described by Dobbersteinand Bunzel.7 DFAs and TriFAs were analyzed using a Luna phenyl-hexyl column (250 mm × 4.6 mm i.d., 5 μm particle size, plus 3 mm ×4.6 mm i.d. guard column) and a ternary gradient made up of 1 mMaqueous trifluoroacetic acid (TFA), acetonitrile (ACN), and methanol(MeOH) at the flow rate of 1.0 mL/min. The sample (20 μL) wasinjected, and the separation was performed at 45 °C, withchromatograms monitored at 280 nm. The following gradients wereused: eluent A, 1 mM aqueous TFA; eluent B, ACN/1 mM aqueousTFA [90:10 (v/v)]; eluent C, MeOH/1 mM aqueous TFA [90:10 (v/v)]. Separation of DFAs and TriFAs was carried out with initially 85%A, 15% B, and 0% C, linear over 15 min to 82% A, 18% B, and 0% C;linear over 5 min to 80% A, 20% B, and 0% C; linear over 5 min to72% A, 25% B, and 3% C; linear over 5 min to 70% A, 25% B, and 5%C; linear over 10 min to 65% A, 30% B, and 5% C; held for 5 min; andlinear over 10 min to 55% A, 45% B, and 5% C, following a rinsing andan equilibration step.

Identification of 8-8(Tetrahydrofuran)-DFA in DDGS Sam-ples. Because no purified standard was available for the 8-8-(tetrahydrofuran)-DFA, the presence of this particular dimer wasvalidated by its relative retention time and UV spectrum character-istics, as described by Dobberstein and Bunzel.7 Furthermore, theelution order for the 8-8(tetrahydrofuran)-DFA was validated byscreening for its mass (m/z 403 [M−1]) using liquid chromatography−mass spectrometry, basically as described by Jilek and Bunzel.8 Finally,gas chromatography−mass spectrometry of a silylated extract wasexemplarily applied to validate the presence of the 8-8-(tetrahydrofur-

Table 1. Validation of Applied Correction Factors from Reference in Relation to Increased Analyte Concentration Ranges

Dobberstein and Bunzel7 current study

range tested (μg/mL) correlation coefficient correction factor range tested (μg/mL) correlation coefficient

8-8(aryltetralin)-DFA 2.5−12.5 0.9975 4.350 6.0−60.0 0.99348-8-DFA 2.5−12.5 0.9893 1.936 a a8-8(tetrahydrofuran)-DFA 2.5−12.5 0.9931 6.040 a a8-5-DFA 6.0−30.0 0.9984 1.509 6.0−60.0 0.99645-5-DFA 6.0−30.0 0.9980 1.514 14.0−140.0 0.99748-8(aryltetralin),8-O-4-TriFA 6.0−30.0 0.9864 3.591 a a8-O-4-DFA 6.0−30.0 0.9997 0.845 10.0−100.0 0.99508-5(benzofuran)-DFA 6.0−30.0 0.9997 4.597 11.0−110.0 0.99655,5,8-O-4-TriFA 2.5−12.5 0.9839 2.089 5.0−50.0 0.99585,5(methylated)-DFA 6.0−30.0 0.9981 1.000 6.0−60.0 0.99718-O-4,8-O-4-TriFA 1.0−5.0 0.9979 b a a8,5(decarboxylated)-DFA 2.5−12.5 0.9992 1.341 6.0−60.0 0.9839

aNot determined as concentration range did not exceed that of Dobberstein and Bunzel.7 bNot determined due to lack of standard compound.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/jf505150gJ. Agric. Food Chem. 2015, 63, 2006−2012

2007

an)-DFA based on the mass spectral characteristics described by

Schatz et al.14

Calculations and Statistical Analysis. The contents of DFAs

and TriFAs were calculated relative to the area of the internal standard

by using the correction factors reported by Dobberstein and Bunzel7

(Table 1).A one-way ANOVA was applied for comparison among means

followed by Tukey’s pairwise comparison with overall significance level

at P = 0.05, using Minitab 16 (Minitab Inc.). The compositional data

were analyzed by principal component analysis (PCA) with evenly

spread seven-fold partial cross-validation after mean centering, and

unit variance scaling, using Evince 2.5.5 software (UmBio AB), was

performed as previously discussed.15 PCA was used to detect

correlations of DFAs and TriFAs with previously reported data on

composition of common nutrients and NSPs from Pedersen et al.2

■ RESULTS

Detection of Ferulic Acid Dehydrodimers andDehydrotrimers and Validation of Correction Factors.A chromatogram of the standard compounds of the DFAs andTriFAs monitored at 280 nm is shown in Figure 1, indicating anacceptable separation of eight dimers and four trimers. The 8-8(tetrahydrofuran)-DFA was not incorporated into the stand-ard mix (see also Identification of 8-8(Tetrahydrofuran)-DFAin DDGS Samples) because a purified standard compound wasnot available. Because the concentration range of DFAs andTriFAs in the samples exceeded the ranges validated byDobberstein and Bunzel,7 increased concentration ranges ofDFA and TriFA standards were analyzed for linear responsesand are presented in Table 1. The linearity was validated byinspecting plots of the area on the y-axis versus analyteconcentration on the x-axis, with observed correlation factors

Figure 1. Separation of ferulic acid dehydrodimers and dehydrotrimers by using RP-HPLC/UV with detection at 280 nm.

Figure 2. Separation of ferulic acid dehydrodimers and dehydrotrimer in corn DDGS with detection at 280 nm.

Table 2. Composition of Ferulic Acid Dehydrodimers and Dehydrotrimers in DDGS of Various Origins

corn DDGS wheat DDGS mixed DDGS

mean’ rangeb SD meanc ranged SD meane rangef SD

∑DFA (μg/g DM) 3310g (2811−4825) 308 621h (543−698) 3 556h (476−664) 69∑8-8-DFA (% DFA) 20g (16−23) 1 30h (26−34) 3 32h (30−34) 2∑8-5-DFA (% DFA) 35g (32−40) 2 31h (25−36) 4 32gh (31−33) 15-5 (% DFA) 27g (25−28) 1 21h (21−22) 1 19i (18−20) 18-O-4 (% DFA) 18g (17−20) 1 18g (16−20) 2 18g (17−18) 0

∑TriFA (μg/g DM) 440g (332−651) 52 77h (65−97) 6 70h (57−88) 115-5,8-O-4 (%TriFA) 69g (56−80) 5 78g (75−82) 2 72g (69−74) 28-8,8-O-4 (%TriFA) 31g (20−44) 5 22g (18−25) 2 28g (26−31) 2

aAverage of means from 21 plants. Means in the same row that do not share a letter (g, h, i)are significantly different at P < 0.05. bN = 47. cAverageof means from 2 plants. dN = 11. eMean from 1 plant. fN = 5.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/jf505150gJ. Agric. Food Chem. 2015, 63, 2006−2012

2008

ranging from 0.9839 to 0.9974 for the eight DFAs and TriFAstested. Visual inspection of the graphs indicated a minortendency toward curve flattening for the highest concentrationsof the 8-5(decarboxylated)-DFA, which may lead to a slightunderestimation of this particular dimer. Inspection of thegraphs for the remaining DFAs and TriFAs indicated aconsistent linear response between area and concentration ofanalyte measured in the samples.Variation in Ferulic Acid Dehydrodimer and Dehy-

drotrimer Profiles in DDGS of Various Origins. A typicalchromatogram of the analysis of alkaline extractable DFAs andTriFAs in DDGS is presented in Figure 2, demonstrating thesuitability of this chromatographic method for DDGS samples.The results of the DFA and TriFA compositional analysis of theDDGS are listed in Table 2. The eight DFAs and the twoTriFAs used to quantify the total content of DFA and TriFAwere identified in all three DDGS origins. The dentity of thesedimers and trimers was verified by comparison of relativeretention times and comparison of the UV spectra with those ofthe isolated standard compounds. In addition, the 8-O-4/8-O-4-TriFA and the 8-5/8-O-4-TriFA were also identified bycomparison to isolated standard compounds. However, becausethe amounts of these trimers were not sufficient for anappropriate validation, these TriFAs were not quantitated. CornDDGS contained 5.3 and 5.9 times the content of total alkalineextractable DFAs than in wheat DDGS (P < 0.001) and mixedDDGS (P < 0.001), respectively. Similarly, the observedcontents of total alkaline extractable TriFAs were 5.7 and 6.3times greater in corn DDGS than in wheat DDGS (P < 0.001)and mixed-cereal DDGS (P < 0.001), respectively.For corn DDGS the ∑8-5-DFA coupled dimers comprised

the greatest proportion of the total DFA content, whereas forwheat and mixed-cereal DDGS the ∑8-8-DFA and ∑8-5-DFAcoupled dimers contributed equally to the total DFA content.In corn DDGS the proportion of the 5-5-DFA was higher than

in wheat (P < 0.001) and mixed-cereal DDGS (P < 0.001).Furthermore, corn DDGS contained a greater proportion forthe ∑8-5-DFA coupled dimers than in wheat DDGS (P =0.024). On the other hand, wheat and mixed-cereal DDGScontained a greater proportion of total DFA for the ∑8-8-DFAcoupled dimers compared with corn DDGS (P < 0.001). Nodifference between the different DDGS origins was observed incontribution of the 8-O-4-DFA (P = 0.481) to the total DFAcontents.Independent of DDGS origin, higher amounts of 5-5,8-O-4-

TriFA than 8-8,8-O-4-TriFA were observed. Both the 5-5,8-O-4-TriFA and 8-8,8-O-4-TriFA proportions were overallsignificantly different among the three DDGS origins (P =0.029); however, the difference was not strong enough to besignificant for pairwise comparisons among origins (P ≥ 0.09).The PCA loading plot illustrated with which compositional

fractions of the DDGS (for DDGS composition see alsoPedersen et al.2) the DFA and TriFA shared the highestcorrelation, presented for corn DDGS (Figure 3A) and wheatand mixed-cereal DDGS (Figure 3B). Both corn DDGS andwheat and mixed-cereal DDGS showed correlation of the totalDFAs to the contents of insoluble xylose, insoluble uronicacids, and insoluble galactose. In addition, corn DDGS DFAscorrelated to insoluble arabinose, insoluble dietary fiber, andcrude fiber, whereas wheat and mixed-cereal DDGS furthercorrelated to the total arabinose content. The TriFAs in cornDDGS correlated to Klason lignin, cellulose, neutral detergentfiber, and crude fiber, whereas the TriFAs in wheat and mixed-cereal DDGS correlated to total and insoluble uronic acids,insoluble arabinose, total and insoluble nondigestible carbohy-drates, and cellulose (for principal component 1).Analyses of corn DDGS samples from five different ethanol

plants (P1, N = 11; P2, N = 5; P3, N = 5; P4, N = 5; P5, N = 5)revealed a greater content of total DFAs for P1 (4030 μg/gDM, P < 0.001) compared with P2 (3371 μg/g DM), P3 (3317

Figure 3. Loading plot of principal component analysis of corn DDGS (A) and wheat and mixed-cereal DDGS (B) constructed by combining DFAand TriFA profiles with additional compositional data from Pedersen et al. as model variables. ADF, acid detergent fiber; AH, acid hydrolysis; Ara,arabinose; Cel, cellulose; DF, dietary fiber; DFA, ferulic acid dehydrodimer; EE, ether extract; Gal, galactose; Glu, NCP-glucose; I, insoluble; Lig,Klason lignin; Man, mannose; NDC, nondigestible carbohydrates; NDF, neutral detergent fiber; S, soluble; T, total; TriFA, ferulic aciddehydrotrimer; Xyl, xylose.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/jf505150gJ. Agric. Food Chem. 2015, 63, 2006−2012

2009

μg/g DM), P4 (3292 μg/g DM), and P5 (3509 μg/g DM).However, the contents of total TriFAs were not differentamong the ethanol plants with an overall range of 426−504 μg/g DM (P = 0.12). The coefficient of variation (CV) rangedfrom 0.04 to 0.09 among the plants for the content of totalDFAs with highest CVs for P1, whereas CVs of 0.01−0.02 wasobserved for the total TriFA contents among the ethanolplants.Variation in Ferulic Acid Dehydrodimer and Dehy-

drotrimer Profile between Corn and CorrespondingDDGS. The DFA and TriFA profiles of corn and correspondingDDGS, sampled simultaneously from the ethanol plant (P1)over a period of 11 months, are presented in Table 3. DDGShad a 3.6-fold increase in total DFA contents along with a 3.7-fold increase in total TriFA contents compared with corn. Nodifference in the proportions of ∑8-8-DFA (P = 0.773) and∑8-5-DFA (P = 0.290) of total DFA and of 5-5,8-O-4-TriFA(P = 0.719) and 8-8,8-O-4-TriFA (P = 0.719) of total TriFAswas observed between corn and corresponding DDGS (Table3).

■ DISCUSSIONSaponification with 2 mol/L NaOH at room temperaturecleaves ester linkages. Thus, DFAs and TriFAs ester-linked topolymeric cell wall components were analyzed in this study.The applied chromatographic methodology allowed forseparation and quantification of eight DFAs and two TriFAs,along with the identification of two additional TriFAs.Validation of the extended analyte concentration ranges (ifcompared to Dobberstein and Bunzel7) indicated that the 8-5(decarboxylated)-DFA might be slightly underestimatedcompared to the other DFAs. Because the 8-5-DFA, the 8-5(benzofuran)-DFA, and the 8-5(decarboxylated)-DFA pre-sumably have the same precursor in vivo, the results of the 8-5-DFAs were reported as the sum of 8-5-coupled dimers.6,8 Also,the contents of the 8-8-DFAs were reported as the sum of the8-8-coupled dimers according to Dobberstein and Bunzel,7

although it is likely that these three 8-8-DFAs occur asindividual DFAs in vivo.6

Because only starch is removed during the fermentationprocess, the nutrient composition of DDGS is expected topartially reflect the composition of the parent grain.2,16 Thehigher content of total DFAs and TriFAs observed in cornDDGS compared to wheat DDGS is in line with previousreports of the DFA and TriFA contents in corn and wheatinsoluble fiber7,17,18 and whole grain flours.8 Also, the observeddominating proportion of the ∑8-5-DFAs (35% of total DFA)in corn DDGS is well in line with previous reports of 39%8 and42%7 in corn, just as the proportion of the 8-O-4-DFA (18% of

total) in corn DDGS is consistent with previous reports of18%8 and 17%7 in corn. For corn DDGS the proportions of 5-5-DFA (27% of total) and the ∑8-8-DFAs (20% of total) wereslightly different from previous reports on corn ranging from 18to 19% and from 21 to 26% of total DFAs, respectively. Forwheat DDGS the observed proportion of the∑8-5-DFAs (31%of total) was lower than the previous reports of 49%8 and 44%.7

On the other hand, the observed proportions of the ∑8-8-DFAs (30% of total) and the 8-O-4-DFA (18% of total) inwheat DDGS were roughly comparable to previous reportsranging between 24 and 25% and between 15 and 16%,respectively.7,8

The 5-5,8-O-4-TriFA was the dominating TriFA in cornDDGS, just as reported for the parent grain.7,8 Despite beingdominant, the proportion of the 5-5,8-O-4-TriFA reported hereis overestimated because a recent study quantitated five TriFAsin corn,8 whereas only two were used for quantification in thisstudy. As shown in Figure 2, at least two other trimers wereidentified in corn DDGS. The same consideration needs to bedone for wheat DDGS, as the same study of Jilek and Bunzel8

indicates an equal proportions of both the 5-5,8-O-4-TriFA andthe 8-O-4,8-O-4-TriFA in wheat.Arabinoxylan consists of a D-xylose backbone joined by β-1,4-

linkages substituted with L-arabinose residues along the xylosechain. The arabinose substitution occurs randomly, allowingother substitutes such as D-glucuronic acid, acetyl groups,galactose, and feruloylated arabinose residues to be attached tothe xylan backbone.19 Evidence that DFAs are indeed attachedto arabinoxylans has been reported for the 5-5-DFA by use ofcarbohydrases and mild acidic hydrolysis followed by structuralcharacterization in bamboo and corn bran20,21 and for the 8-O-4-DFA and 8-8-DFA from corn bran insoluble fiber.22,23 Theuse of PCA multivariate data analysis based on the combinationof DFA and TriFA profiles with compositional data previouslyreported2 indicated a clear correlation between DFAs andTriFAs in the DDGS and the content of insoluble arabinoxylan(sum of anhydrous arabinose and xylose), uronic acids,insoluble galactose, and cellulose. In addition to pointing outthat DFAs are attached to arabinoxylan as mentioned above,these results logically indicate a similar association for theTriFAs. The observed correlation of DFAs to insolublegalactose is in line with previous studies describing highlycomplex feruloylated heteroxylan side chains containing up totwo galactose residues from corn bran.24,25 The observedgreater content of DFAs in DDGS from P1 is likely related tothe higher content of insoluble arabinoxylan in DDGS from P1than from P2−P5, as reported by Pedersen et al.2

The simultaneous sampling of both corn grains andcorresponding DDGS over 11 months allowed for investigating

Table 3. Composition of Ferulic Acid Dehydrodimers (DFA) and Dehydrotrimers (TriFA) in Corn and Corresponding DDGS

corn (N = 11) DDGS (N = 11)

meana range SD mean range SD ratiob

∑DFA (μg/g DM) 1135b (976−1419) 140 4030c (3594−4828) 366 3.6∑8-8-DFA (% DFA) 18b (17−19) 1 18b (17−18) 1∑8-5-DFA (% DFA) 37b (36−38) 1 38b (36−40) 15-5 (% DFA) 27b (26−28) 1 26c (25−27) 08-O-4 (% DFA) 18b (17−18) 0 18c (18−19) 0

∑TriFA (μg/g DM) 137b (107−158) 17 504c (388−651) 83 3.75-5,8-O-4 (%TriFA) 70b (63−76) 4 71b (62−74) 48-8,8-O-4 (%TriFA) 30b (24−37) 4 29b (24−38) 4

aMeans in the same row that do not share a letter (b, c) are significantly different at P < 0.05. bAverage corn-to-corn DDGS ratio.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/jf505150gJ. Agric. Food Chem. 2015, 63, 2006−2012

2010

the potential changes in DFA and TriFA profiles after thefermentation and drying process. Although corn grains andDDGS were sampled simultaneously, it should be noted thatdue to practical challenges the samples did not originate fromthe same fermentation batch. Furthermore, it should be notedthat the results may not be the same for all ethanol plants,emphasized by the variation in DFAs and TriFAs content fromthe 21 corn DDGS ethanol plants. The 3.6- and 3.7-foldincreases in DFA and TriFA contents, respectively, in DDGSare in accordance with the 3.4-fold average increase in nutrientsfrom the same set of samples reported by Pedersen et al.2 andwith previous studies reporting a similar nutrient increase inDDGS compared with the parent grain.26,27

Overall, the analysis of corn and DDGS indicated that theindividual composition of DFAs and TriFAs is conserved,indicating that DFAs and TriFAs are basically not modifiedduring ethanol fermentation and subsequent drying of DDGS.Previous characterization of NSP profiles of the samples led tothe hypothesis that only arabinoxylan from corn endosperm,which comprises only a minor fraction of the total arabinoxylan,was modified during DDGS production.2 Considering the latterand the fact that corn endosperm arabinoxylan containsmarkedly less feruloylated cross-linkages compared to cornbran heteroxylan,28−30 the observed conservation of the DFAand TriFA patterns in corn DDGS is well in line. Because themajor fraction of feruloylated arabinoxylan remains unmodified,the remaining arabinoxylan in DDGS is probably more complexthan in the parent grain because the readily degradablearabinoxylan was modified during processing.Large differences in enzymatic degradability were reported

for corn- and wheat-based substrates,31 with a markedly greaterdegradation potential for wheat than for corn. Furthermore,with regard to enzymatic degradation, corn DDGS26 andespecially corn bran have been described as recalcitrantsubstrates due to the highly complex structure of thearabinoxylan.32−34 Especially the DFA and TriFA cross-linksare suspected to play a major role in hindering the enzymaticdegradation of cereal cell walls.9−12 Thus, the results from thisstudy on DFA and TriFA contents may indicate a generallyhigher potential for enzymatic degradation of wheat DDGScompared to corn DDGS. As a consequence of the difference incell wall complexity, differences in digestibility, microbialfermentation, and resistance to fiber degrading enzymes maybe observed when feeding either wheat DDGS or corn DDGS.The use of DDGS in swine feed has been limited due to a

reputation of having variable nutrient composition and proteinquality, potentially high content of mycotoxins, and high fibercontent. The high fiber content in DDGS is associated withnutritional challenges and negative effect on nutrientdigestibility when fed in high quantities. To overcome thenegative effects associated with high fiber content, studies havebeen performed to investigate the effects of adding fiberdegrading enzymes to animal diets containing DDGS.35−37

However, the results are inconclusive.The results presented in this study may indicate that fiber-

degrading enzymes applied for the degradation of DDGS needto be targeted toward highly complex substrates, especiallywhen applied to diets high in corn DDGS.In conclusion, the current study demonstrated a large

variation in the content of DFAs and TriFAs between DDGSof three different origins. Corn DDGS contained 5.3 and 5.7times greater content of total DFAs and TriFAs, respectively,compared with wheat DDGS. The contents of total DFA and

TriFA correlated to the content of insoluble arabinoxylan,uronic acids, insoluble galactose, and cellulose in DDGS. Asimilar proportional profile of the individual DFAs and TriFAswas observed in both corn and corresponding DDGS,indicating that ferulic acid cross-links in the corn cell wall arepresumably not modified during fermentation and DDGSprocessing. The results presented herein may add to the overallunderstanding and implication of compositional differencesbetween DDGS sources in relation to animal nutrition andenzyme digestibility.

■ AUTHOR INFORMATIONCorresponding Author*(M.B.P.) Phone: +45 8943 5327. E-mail: madsb.pedersen@anis.au.dk.FundingThis project was partially supported by The Danish Ministry ofScience, Innovation and Higher Education.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThe following people are acknowledged for their assistance incollecting DDGS samples from different bioethanol plants: M.Chow, D. Ko, A. Riley, D. Ward, G. Jones, C. Miller-Fosmore,and E. Seward.

■ ABBREVIATIONS USEDACN, acetonitrile; CV, coefficient of variation; DDGS,distillers’ dried grains with solubles; DFA, ferulic aciddehydrodimer; DM, dry matter; HPLC, high-performanceliquid chromatography; MeOH, methanol; NSP, nonstarchpolysaccharides; PCA, principal component analysis; PDA,photodiode array detector; RP-HPLC, reversed phase HPLC;TriFA, ferulic acid dehydrotrimer; TFA, trifluoroacetic acid;UV, ultraviolet

■ REFERENCES(1) Stein, H. H.; Shurson, G. C. Board-invited review: the use andapplication of distillers dried grains with solubles in swine diets. J.Anim. Sci. 2009, 87, 1292−303.(2) Pedersen, M. B.; Dalsgaard, S.; Knudsen, K. E. B.; Yu, S.; Lærke,H. N. Compositional profile and variation of distillers dried grains withsolubles from various origins with focus on non-starch polysaccharides.Anim. Feed Sci. Technol. 2014, 197, 130−141.(3) Ralph, J.; Grabber, J. H.; Hatfield, R. D. Lignin-ferulate cross-linksin grasses: active incorporation of ferulate polysaccharide esters intoryegrass lignins. Carbohydr. Res. 1995, 275, 167−178.(4) Bunzel, M.; Ralph, J.; Lu, F.; Hatfield, R. D.; Steinhart, H. Ligninsand ferulate-coniferyl alcohol cross-coupling products in cereal grains.J. Agric. Food Chem. 2004, 52, 6496−6502.(5) Piber, M.; Koehler, P. Identification of dehydro-ferulic acid-tyrosine in rye and wheat: evidence for a covalent cross-link betweenarabinoxylans and proteins. J. Agric. Food Chem. 2005, 53, 5276−5284.(6) Bunzel, M. Chemistry and occurrence of hydroxycinnamateoligomers. Phytochem. Rev. 2010, 9, 47−64.(7) Dobberstein, D.; Bunzel, M. Separation and detection of cell wall-bound ferulic acid dehydrodimers and dehydrotrimers in cereals andother plant materials by reversed phase high-performance liquidchromatography with ultraviolet detection. J. Agric. Food Chem. 2010,58, 8927−8935.(8) Jilek, M. L.; Bunzel, M. Dehydrotriferulic and dehydrodiferulicacid profiles of cereal and pseudocereal flours. Cereal Chem. J. 2013,90, 507−514.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/jf505150gJ. Agric. Food Chem. 2015, 63, 2006−2012

2011

(9) Bily, A. C.; Reid, L. M.; Taylor, J. H.; Johnston, D.; Malouin, C.;Burt, A. J.; Bakan, B.; Regnault-Roger, C.; Pauls, K. P.; Arnason, J. T.;Philogene, B. J. R. Dehydrodimers of ferulic acid in maize grainpericarp and aleurone: resistance factors to Fusarium graminearum.Phytopathology 2003, 93, 712−719.(10) Santiago, R.; Reid, L. M.; Arnason, J. T.; Zhu, X.; Martinez, N.;Malvar, R. A. Phenolics in maize genotypes differing in susceptibility togibberella stalk rot (Fusarium graminearum Schwabe). J. Agric. FoodChem. 2007, 55, 5186−5193.(11) Grabber, J. H.; Hatfield, R. D.; Ralph, J. Diferulate cross-linksimpede the enzymatic degradation of non-lignified maize walls. J. Sci.Food Agric. 1998, 77, 193−200.(12) Grabber, J. H.; Ralph, J.; Hatfield, R. D. Ferulate cross-links limitthe enzymatic degradation of synthetically lignified primary walls ofmaize. J. Agric. Food Chem. 1998, 46, 2609−2614.(13) Bunzel, M.; Funk, C.; Steinhart, H. Semipreparative isolation ofdehydrodiferulic and dehydrotriferulic acids as standard substancesfrom maize bran. J. Sep. Sci. 2004, 27, 1080−1086.(14) Schatz, P. F.; Ralph, J.; Lu, F.; Guzei, I. A.; Bunzel, M. Synthesisand identification of 2,5-bis-(4-hydroxy-3-methoxyphenyl)-tetrahydro-furan-3,4-dicarboxylic acid, an unanticipated ferulate 8-8-couplingproduct acylating cereal plant cell walls. Org. Biomol. Chem. 2006, 4,2801−2806.(15) Shewry, P. R.; Hawkesford, M. J.; Piironen, V.; Lampi, A.-M.;Gebruers, K.; Boros, D.; Andersson, A. A. M.; Åman, P.; Rakszegi, M.;Bedo, Z.; Ward, J. L. Natural variation in grain composition of wheatand related cereals. J. Agric. Food Chem. 2013, 61, 8295−8303.(16) Widyaratne, G. P.; Zijlstra, R. T. Nutritional value of wheat andcorn distiller’s dried grain with solubles: digestibility and digestiblecontents of energy, amino acids and phosphorus, nutrient excretionand growth performance of grower-finisher pigs. Can. J. Anim. Sci.2007, 87, 103−114.(17) Renger, A.; Steinhart, H. Ferulic acid dehydrodimers asstructural elements in cereal dietary fibre. Eur. Food Res. Technol.2000, 211, 422−428.(18) Bunzel, M.; Ralph, J.; Marita, J. M.; Hatfield, R. D.; Steinhart, H.Diferulates as structural components in soluble and insoluble cerealdietary fibre. J. Sci. Food Agric. 2001, 81, 653−660.(19) Izydorczyk, M. S.; Biliaderis, C. G. Cereal arabinoxylans:advances in structure and physicochemical properties. Carbohydr.Polym. 1995, 28, 33−48.(20) Ishii, T. Isolation and characterization of a diferuloylarabinoxylan hexasaccharide from bamboo shoot cell-walls. Carbohydr.Res. 1991, 219, 15−22.(21) Saulnier, L.; Crepeau, M.-J.; Lahaye, M.; Thibault, J.-F.; Garcia-Conesa, M. T.; Kroon, P. A.; Williamson, G. Isolation and structuraldetermination of two 5,5′-diferuloyl oligosaccharides indicate thatmaize heteroxylans are covalently cross-linked by oxidatively coupledferulates. Carbohydr. Res. 1999, 320, 82−92.(22) Allerdings, E.; Ralph, J.; Schatz, P. F.; Gniechwitz, D.; Steinhart,H.; Bunzel, M. Isolation and structural identification of diarabinosyl 8-O-4-dehydrodiferulate from maize bran insoluble fibre. Phytochemistry2005, 66, 113−124.(23) Bunzel, M.; Allerdings, E.; Ralph, J.; Steinhart, H. Cross-linkingof arabinoxylans via 8-8-coupled diferulates as demonstrated byisolation and identification of diarabinosyl 8-8(cyclic)-dehydrodifer-ulate from maize bran. J. Cereal Sci. 2008, 47, 29−40.(24) Allerdings, E.; Ralph, J.; Steinhart, H.; Bunzel, M. Isolation andstructural identification of complex feruloylated heteroxylan side-chains from maize bran. Phytochemistry 2006, 67, 1276−1286.(25) Saulnier, L.; Vigouroux, J.; Thibault, J.-F. Isolation and partialcharacterization of feruloylated oligosaccharides from maize bran.Carbohydr. Res. 1995, 272, 241−253.(26) de Vries, S.; Pustjens, A. M.; Kabel, M. A.; Salazar-Villanea, S.;Hendriks, W. H.; Gerrits, W. J. J. Processing technologies and cell walldegrading enzymes to improve nutritional value of dried distillers grainwith solubles for animal feed: an in vitro digestion study. J. Agric. FoodChem. 2013, 61, 8821−8828.

(27) Urriola, P. E.; Shurson, G. C.; Stein, H. H. Digestibility ofdietary fiber in distillers coproducts fed to growing pigs. J. Anim. Sci.2010, 88, 2373−2381.(28) Bunzel, M.; Ralph, J.; Bruning, P.; Steinhart, H. Structuralidentification of dehydrotriferulic and dehydrotetraferulic acidsisolated from insoluble maize bran fiber. J. Agric. Food Chem. 2006,54, 6409−6418.(29) Gallardo, C.; Jimenez, L.; García-Conesa, M. T. Hydroxycin-namic acid composition and in vitro antioxidant activity of selectedgrain fractions. Food Chem. 2006, 99, 455−463.(30) Saulnier, L.; Thibault, J.-F. Ferulic acid and diferulic acids ascomponents of sugar-beet pectins and maize bran heteroxylans. J. Sci.Food Agric. 1999, 79, 396−402.(31) Rose, D.; Inglett, G. A method for the determination of solublearabinoxylan released from insoluble substrates by xylanases. FoodAnal. Methods 2011, 4, 66−72.(32) Agger, J.; Vikso-Nielsen, A.; Meyer, A. S. Enzymatic xyloserelease from pretreated corn bran arabinoxylan: differential effects ofdeacetylation and deferuloylation on insoluble and soluble substratefractions. J. Agric. Food Chem. 2010, 58, 6141−6148.(33) Faulds, C. B.; Kroon, P. A.; Saulnier, L.; Thibault, J.-F.;Williamson, G. Release of ferulic acid from maize bran and derivedoligosaccharides by Aspergillus niger esterases. Carbohydr. Polym. 1995,27, 187−190.(34) Saulnier, L.; Marot, C.; Elgorriaga, M.; Bonnin, E.; Thibault, J. F.Thermal and enzymatic treatments for the release of free ferulic acidfrom maize bran. Carbohydr. Polym. 2001, 45, 269−275.(35) Anderson, P. V.; Kerr, B. J.; Weber, T. E.; Ziemer, C. J.;Shurson, G. C. Determination and prediction of digestible andmetabolizable energy from chemical analysis of corn coproducts fed tofinishing pigs. J. Anim. Sci. 2012, 90, 1242−1254.(36) Pedersen, C.; Boersma, M. G.; Stein, H. H. Digestibility ofenergy and phosphorus in ten samples of distillers dried grains withsolubles fed to growing pigs. J. Anim. Sci. 2007, 85, 1168−1176.(37) Stein, H. H.; Gibson, M. L.; Pedersen, C.; Boersma, M. G.Amino acid and energy digestibility in ten samples of distillers driedgrain with solubles fed to growing pigs. J. Anim. Sci. 2006, 84, 853−860.

Journal of Agricultural and Food Chemistry Article

DOI: 10.1021/jf505150gJ. Agric. Food Chem. 2015, 63, 2006−2012

2012

52

53

PAPER III

Xylanase and Protease Increase

Solubilization of Non-Starch

Polysaccharides and Nutrient Release of

Corn- and Wheat Distillers Dried Grains

with Solubles

Mads B. Pedersen, Søren Dalsgaard, Susan Arent,

Rikke Lorentzen, Knud Erik B. Knudsen, Shukun

Yu, and Helle N. Lærke

Published in:

Biochemical Engineering Journal, 2015, 98:99-106

54

Biochemical Engineering Journal 98 (2015) 99–106

Contents lists available at ScienceDirect

Biochemical Engineering Journal

jo ur nal home page: www.elsev ier .com/ locate /be j

Regular article

Xylanase and protease increase solubilization of non-starchpolysaccharides and nutrient release of corn- and wheatdistillers dried grains with solubles

Mads B. Pedersen a,b,∗, Søren Dalsgaard a, Susan Arent a, Rikke Lorentsen a,Knud Erik B. Knudsen b, Shukun Yu a, Helle N. Lærke b

a Enzyme Development, DuPont Industrial Biosciences ApS, Edwin Rahrs Vej 38, DK-8220 Brabrand, Denmarkb Department of Animal Science, Aarhus University, Blichers Allé 20, DK-8830 Tjele, Denmark

a r t i c l e i n f o

Article history:Received 6 January 2015Received in revised form 13 February 2015Accepted 20 February 2015Available online 27 February 2015

Keywords:Distillers dried grains with solublesEnzymesEnzymatic activityOptimizationXylanaseProtease

a b s t r a c t

The use of distiller dried grains with solubles (DDGS) as alternative to conventional animal feed for non-ruminants is challenged by the high content of non-starch polysaccharides and varying protein quality.In this study the enzymatic degradation of corn- and wheat DDGS was evaluated, in vitro, by use of fourxylanases from two different glycoside hydrolase families, GH10 and GH11, along with protease andphytase. Wheat DDGS showed the highest degree of enzymatic degradation due to a lower degree of cellwall complexity compared with that of corn DDGS. For corn DDGS,the combination of xylanase and pro-tease yielded the highest degree of enzymatic degradation, indicating close association of arabinoxylanand protein within the cell wall matrix. Collectively, the GH10xylanase degraded DDGS more efficientlythan the GH11 xylanases, due to the complexity of the DDGS substrate and the substrate affinity of theGH10xylanase. The current in vitro results indicate a high potential of xylanase in combination withprotease to efficiently degrade DDGS and promote nutrient release in diets for non-ruminant animals.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

During the past decade attention on implementing alternativesto conventional feed stuffs (e.g., corn, wheat, and soybean meal) inanimal production has become more and more prevalent. The mainreason is the increased demand for and consequently increasedcosts of conventional raw materials together with an increasedavailability of new low-cost raw materials, many of these beingco-products from the grain processing industries. One of these co-products is distillers dried grains with solubles (DDGS), a driedco-product from the production of fuel ethanol [1,2].

Abbreviations: ADF, acid detergent fiber; AH, acid hydrolysis; A:X, ara-binose:xylose ratio; BSA, bovine serum albumin; cDDGS, corn DDGS; DDGS,distillers dried grains with solubles; DF, dietary fiber; DM, dry matter; EE, etherextract; GH, glucoside hydrolase family; NCP, non-cellulosic polysaccharides; NDF,neutral detergent fiber; NIRS, near infrared reflectance spectroscopy; NSP, non-starch polysaccharides; PLS, partial least square; TAXI, Triticumaestivumxylanaseinhibitors; UA:X, uronicacid:xylose ratio; wDDGS, wheat DDGS; XIP, xylanasein-hibiting proteins.

∗ Corresponding author at: Department of Animal Science, Aarhus University,Blichers Allé 20, DK-8830 Tjele, Denmark. Tel.: +45 8715 7967; fax: +45 8943 5115.

E-mail address: madsb.pedersen@anis.au.dk (M.B. Pedersen).

As a feed ingredient, DDGS has high potential with a digestible-and metabolizable energy content similar to that of corn, along witha high content of digestible phosphorous [3,4]. In general, DDGS hasgreater concentrations of protein, fat, vitamins and minerals com-pared to the parent grain [5,6]. However, disadvantages of usingDDGS in the feeding of non-ruminant animals are the 3–3.5 foldhigher content of non-starch polysaccharides (NSP), compared tothe parent grain [3,7]. Since non-ruminant animals have a limitedcapacity to utilize NSP, high inclusion levels will inevitably limitfeed utilization [8,9,10]. Varying protein quality and the risk ofmycotoxin contamination further challenge a high inclusion rateof DDGS in animal diets.

The NSP in DDGS originate from cell walls in the botanical grainfractions; aleurone layer, pericarp, endosperm, germ and tip cap(corn), with arabinoxylan and cellulose as the major components ofthe NSP fraction [7]. Arabinoxylan consists of d-xylose units joinedby �-linkages and substituted with arabinose residues along thechain [11,12], including other substitutes like d-glucuronic acidand acetyl groups [13]. These substitutes together with feruloylatedarabinose residues contribute to arabinoxylan cross-linking to formstrong intermolecular complexes, affecting the enzymatic degra-dation and encapsulation of nutrients [14,15]. Xylanases belong totwo main families based on their primary sequence and structure

http://dx.doi.org/10.1016/j.bej.2015.02.0361369-703X/© 2015 Elsevier B.V. All rights reserved.

100 M.B. Pedersen et al. / Biochemical Engineering Journal 98 (2015) 99–106

[16], i.e., glycoside hydrolase family 10 and 11 (GH10, GH11). GH10xylanases have been shown to exhibit greater catalytic versatility orbroader substrate specificity than xylanases from the GH11 family.GH11 cleaves primarily unsubstituted regions of the arabinoxylanbackbone, whereas GH10 is capable of cleaving in more substitutedregions, hence being less hindered by the presence of substitutions[17,18].

Several studies have been conducted to investigate the effects ofNSP degrading enzymes on nutrient digestibility of DDGS contain-ing diets. However, the in vivo results are inconclusive [11,19,20],providing evidence for further investigation of the effect of NSPdegrading enzymes and other enzymes on DDGS hydrolysis andnutrient release.

The current study was undertaken to investigate the enzymaticdegradation of corn DDGS (cDDGS) and wheat DDGS (wDDGS), invitro, using four different xylanases, alone and in combination withprotease and phytase. It is hypothesized that the xylanases com-prise different affinities toward c- and wDDGS, thus, affecting DDGSdegradation differently. In addition, it is hypothesized that additionof protease and phytase may contribute to an increased degrada-tion by disrupting the feed matrix interactions through hydrolysisof protein and phytate. The effect of enzyme treatment was eval-uated by NSP degradation defined as pentosan solubilization, andnutrient release defined as protein solubilization.

2. Materials and methods

2.1. Chemicals

Phloroglucinol (1,3,5-trihydroxybenzene) and d-Xylose(C5H10O5) were from Merck, and BCA Protein Assay Kit wasfrom Thermo Fischer Scientific. All other chemicals used were ofanalytical grade.

2.2. Materials

The wDDGS was supplied from a European bioethanol plant andcDDGS from a North American bioethanol plant. These two sam-ples were previously analyzed and characterized together with 136DDGS samples by Pedersen et al. [7].

Purified endo-1,4-�-xylanases (EC 3.2.1.8); XylA (GH10), XylB(GH11), XylC (GH11) and XylD (GH11) from different microbialorigins, protease (EC 3.4.21.62) of Bacillus origin (syn. Multifect P-3000, Danisco Animal Nutrition, Marlborough, UK), and purifiedmicrobial phytase (EC 3.1.3.26), were supplied by DuPont Indus-trial Biosciences, Denmark. The purity of the enzyme solutionwas confirmed using gel electrophoresis (NU-PAGE Bis/Tris precastgel, Invitrogen, Life Technologies Corp.). Thermostable ˛-amylase(E-BLAAM 53.7 U/mg) and amyloglucosidase (E-AMGDF 36 U/mg)were obtained from Megazyme International.

2.3. Chemical analysis and calculations

2.3.1. Near infrared reflectance spectroscopyGround DDGS (Retsch ZM 200 centrifugal mill fitted with a

0.5 mm sieve) was scanned from 1100 to 2498 nm using nearinfrared reflectance spectroscopy (NIRS) on a FOSS NIRSystems5000 (Foss). The spectral data were predicted by Aunir (AB AgriLtd., UK) for the composition of moisture, fat (ether extract), fat(acid hydrolysis), protein, crude fiber, ash, starch, total sugars, neu-tral detergent fiber (NDF) and acid detergent fiber (ADF), using thecalibration available for DDGS as previously described [7].

2.3.2. Compositional analysis of NSPTotal- and soluble NSP along with their constituent sugars were

determined by gas–liquid chromatography for neutral sugars and

by a colorimetric method for uronic acids, basically as describedearlier [21], except that 2 M of sulfuric acid for 1 h was used for thehydrolysis of the non-cellulosic polysaccharides (NCP) rather than1 M of sulfuric acid for 2 h. Klason lignin was measured gravimetri-cally as the residue resistant to hydrolysis by 2 M sulfuric acid afterswelling with 12 M sulfuric acid [21].

Percentage solubilization of total pentosan as

0.88 × measured pentose(� g/ml)Total content of anhydrous xylose and arabinose(� g DM/ml)

× 100

Percentage solubilization of total protein as

measured protein(� g/ml)Total content of protein(� g DM/ml)

× 100

2.4. Enzyme purification and quantification

The protein content of the purified enzyme solutions were quan-tified using a NanoDrop 1000 spectrophotometer (Thermo FischerScientific Inc.) based on UV absorbance at 280 nm according toLambert Beer’s Law with extension coefficient calculated fromsequence. The preparation of protease was performed just prior toincubations, by diluting the stock solution in ice-cooled MQ-waterand mixed while kept on ice. One protease unit was defined asthe release of 1.0 �g of phenolic compound, expressed as tyrosineequivalents, from a casein substrate per minute.

2.5. Substrate preparation

DDGS was ground and sieved (<212 �m) before mixing with25 mM citrate buffer pH 6.0–10% w/v suspension followed by finalpH-adjustment. Under constant stirring, 175 �l/well of this suspen-sion was dispensed into 96-well plates using Biomek NX (BeckmanCoulter, Inc.). Prepared substrate plates were stored at −20 ◦C, andthawed just prior to incubations.

For the preparation of insoluble substrate (hereon referred to asinsoluble DDGS), removal of soluble NSP was performed basicallyas previously described [21]. Ground DDGS (<212 �m) in sodiumacetate/CaCl2-buffer (0.1 M/20 mM, pH 5.0) was mixed with ther-mostable �-amylase (E-BLAAM 53.7 U/mg) and incubated at 100 ◦Cfor 1 h with frequent mixing. Complete degradation of starch wasdone by incubation with amyloglucosidase (E-AMGDF 36 U/mg) for2 h at 60 ◦C. After removal of starch, the soluble NSP was extracted insodium phosphate buffer (0.2 M, pH 7.0) at 100 ◦C for 1 h, followedby centrifugation at 3000 rpm. The pellet was then thoroughlywashed sequentially with the phosphate buffer, ethanol (85% v/v),and acetone, with centrifugation and discard of supernatant inbetween these washes. The sample was placed at room temper-ature until completely dried.

2.6. Enzymatic hydrolysis of DDGS by xylanase, protease andphytase

For each of the four xylanase treatments (0.2 g xylanase/kg feed)a full factorial 32 experiment was setup in duplicates with 2 factorsin 3 levels; protease (0 U/kg feed, 4.3 × 105 U/kg feed, 8.6 × 105 U/kgfeed) and phytase (0 �g/well, 0.1 g/kg feed, 0.2 g/kg feed). A volumeof 25 �l of mixed enzyme solution was transferred to the thawed96-well substrate plates and mixed. Then the substrate plates weresealed and placed in iEMS incubators (Thermo Scientific) at 39 ◦C,1100 rpm for 4 h. After incubation, the solution was transferredto a 96-well filter plate (0.22 �m) and centrifuged at 3600 rpm,5 ◦C until the retentate was completely dried. Finally, the filtratewas mixed, proper diluted and aliquots taken out for protein- andpentosan quantification.

M.B. Pedersen et al. / Biochemical Engineering Journal 98 (2015) 99–106 101

2.7. Enzymatic hydrolysis of insoluble DDGS by combinations ofxylanases and protease

The effects of four xylanase treatments (0.2 g xylanase/kg feed)were investigated as mono components and in combination withprotease (8.6 × 105 U/kg feed) on the solubilization of pentosan andprotein from insouble c- and wDDGS. 88 mg of the prepared insol-uble DDGS substrate was weighed into 1.5 ml Eppendorf tubes andmixed with citrate buffer (25 mM, pH 6.0) and enzymes solutionto a final reaction volume of 1.0 ml. The incubations were carriedout at 39 ◦C and a stirring speed of 1300 rpm on an EppendorfThermoMixer incubator for 4 h. After incubation, the samples werefiltered and analyzed for soluble pentosan and protein content.Experiments were performed in duplicates.

2.8. Analysis of hydrolysis products

2.8.1. Protein quantificationSoluble protein was quantified using the BCA Protein Assay Kit

with the assay range of 50–2000 �g/ml. In 96-well plates 25 �l sam-ple was mixed with 200 �l premixed assay reagent and incubatedat 37 ◦C and stirring at 1100 rpm for 30 min. The absorbance wasmeasured spectrophotometrically at 562 nm and quantified againsta bovine serum albumin standard curve. The measured protein con-tent was corrected for the amounts of added enzymes. Additionalexperiments were performed to verify that protein quantificationwas not interfered by the pentosan content present in the samples.

2.8.2. Pentosan quantificationPerformed principally as previously described [22]. Soluble pen-

tosan was quantified using a San+ + Continous Flow Analyzer(Skalar Analytical B.V., Breda, Netherlands) equipped with 96-wellautosampler unit. The sample was mixed with acetic acid reagent(glacial acetic acid with 2% v/v HCl (37%)) for total hydrolysis andphloroglucinol (1,3,5-trihydroxybenzene). The mixture was heatedto 96 ◦C followed by cooling and absorbance measurement at 510and 550 nm. Pentosan was quantified against a xylose standardcurve (50–500 �g/ml). The pentosan content was determined asmono-sugars. To determine the proportion of solubilized pentosanof total pentosan, the mono-sugars were converted to polysaccha-rides equivalents (anhydro sugars) by a conversion factor of 0.88,and calculated as described above.

2.9. Experimental design and statistical analysis

The 32 full factorial experiments (2 factors in 3 levels) [23] wasdesigned and the response data analyzed using PLS regression anal-ysis with backward elimination of insignificant interactions andoutliers, with an overall significance level at P = 0.05 (Modde 9.1,Umetrics, Umeå, Sweden). A PLS regression model defines a linearrelationship y = f(X) between two blocks of variables: (1) the matrixX contains the 32 combinations of different enzymes, and (2) thevector y is the individual responses i.e., pentosan and protein [24].The PLS models were considered satisfactory when R2 and Q2 wereabove 0.5. Means from incubations with c- and wDDGS were com-pared based on the 95% confidence intervals generated from thePLS modelling.

For the comparison of means from the incubations with insolu-ble DDGS, a one-way ANOVA was applied followed by a Tukey’s pairwise comparison with overall significance level at P = 0.05, usingMinitab 16 (Minitab Inc.).

Table 1Compositional profile of corn- and wheat DDGS g/kg dry matter.

Corn DDGS Wheat DDGS

Fat (EE) 79 57Fat (AH) 102 76Crude protein 304 366Crude fiber 95 70Ash 70 93Starch 50 7Total sugar 62 79NDF 387 330ADF 118 105

Total NSP 328 (29) 266 (76)

Cellulose 91 62

NCPXylose 94 (5) 81 (22)Arabinose 70 (7) 55 (17)Mannose 14 (5) 14 (7)Galactose 18 (2) 10 (6)Glucose 22 (5) 33 (21)Uronic acid 18 (5) 8 (2)

Klason lignin 28 86

A:X-ratio 0.74 0.68UA:X-ratio 0.19 0.10

EE, ether extract; AH, acid hydrolysis; NDF, neutral detergent fiber; ADF, acid deter-gent fiber; NSP, non-starch polysaccharides; NSP, non-starch polysaccharides; NCP,non-cellulosic polysaccharides; A:X, arabinose:xylose; UA:X, uronic acid:xylose;Values in brackets are soluble NSP.

3. Results

3.1. Compositional profile of corn and wheat DDGS

The compositional profiles of c- and wDDGS are listed in Table 1.The content of soluble NSP was greater in w- than cDDGS, whereascDDGS had greater contents of both insoluble and total NSP. Thecontent of arabinoxylan (sum of anhydrous arabinose and xylose)was slightly greater in c- than wDDGS, and cDDGS had greater con-tent of uronic acids than wDDGS. In addition, cDDGS had higherarabinose:xylose ratio (A:X) and uronic acid:xylose ratio (UA:X)than wDDGS.

3.2. Effects of xylanases on DDGS hydrolysis

For cDDGS the four xylanase treatments had varying effects onthe solubilization of pentosan and protein compared with controlwithout enzyme. Compared with the control, the greatest increasein pentosan solubilization from cDDGS was observed with XylAtreatment (77.3%) followed by XylC (43.6%), XylB (12.2%), and XylD(7.2%), respectively (Fig. 1A). In addition, XylA increased the solu-bilization of protein from cDDGS by 11.1% followed by XylD (6.1%),XylC (4.1%), and XylB (2.6%), respectively, compared to the control(Fig. 1B).

For wDDGS, the variation in pentosan solubilization was lesspronounced between the individual xylanase treatments comparedwith cDDGS. For wDDGS, the greatest increase in pentosan solubi-lization compared to the control was observed for XylD treatment(26.7%) followed by XylA (23.3%), XylC (18.3%), and XylB (17.0%),respectively (Fig. 2A). The greatest increase in protein solubilizationcompared to control was observed for XylD treatment (30.9%) fol-lowed by XylC (15.2%), XylB (15.1%), and XylA (14.2%), respectively(Fig. 2B).

Based on modelling, the four individual xylanase treatments sol-ubilized 10.1–16.8% and 80.8–87.5% of the total pentosan content,and 13.0–14.1% and 23.5–27.0% of the total protein content for c-and wDDGS, respectively.

102 M.B. Pedersen et al. / Biochemical Engineering Journal 98 (2015) 99–106

Control XylA XylB XylC XylD

Solu

biliz

ed p

ento

san

(mg/

ml)

0,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5Xylanase Xylanase, protea se, (p hytase)

A

e

b

d

c

d,e

a

c

b

c

Cont rol XylA XylB XylC XylD

Solu

biliz

ed p

rote

in (m

g/m

l)

0

1

2

3

4

5XylanaseXylanase, protease, (phytase)

B

e

c,d

a

d,e

c

d,e

b,c

d,e

a,b

Fig. 1. Solubilization of pentosan (A) and protein (B) from cDDGS with xylanase alone and in combination with protease and phytase (for XylA and XylD) under optimizedconditions. Error bars indicate 95% confidence intervals. Bars with different letters are significantly different at P < 0.05.

Con trol XylA XylB XylC XylD

Solu

biliz

ed p

ento

san

(mg/

ml)

2

4

6

8

10

12

Xylanase Xylanase, protease

c

a,b a,bb b b a,b

aa,b

Control XylA XylB XylC XylD

Solu

biliz

ed p

rote

in (m

g/m

l)

0

2

4

6

8

10

12 XylanaseXylanase, protease

c

b,c

a

b,c

a

b,c

aa,b

a

A B

Fig. 2. Solubilization of pentosan (A) and protein (B) from wDDGS with xylanase alone and in combination with protease under optimized conditions. Error bars indicate95% confidence intervals. Bars with different letters are significantly different at P < 0.05.

3.3. Effects of protease and phytase in combination withxylanases on DDGS hydrolysis

Coefficients of the PLS regression analyses on the effect ofprotease and phytase addition to the four xylanase treatmentsand are presented in Table 2. Protease addition significantlyincreased the solubilization of protein in both cDDGS (P < 0.0001)and wDDGS (P ≤ 0.02) across all four xylanase treatments. Fur-thermore, protease significantly increased the solubilization ofpentosan in cDDGS across all four xylanase treatments (P < 0.0001),whereas no significant effect was observed for protease addi-tion on pentosan solubilization in wDDGS (P ≥ 0.094). Phytasehad no significant effect on the solubilization of pentosan orprotein in neither c- nor wDDGS. However, significantly posi-tive interactions between protease and phytase were observed onthe solubilization of protein with the XylB treatment in cDDGS(P = 0.009).

3.4. Optimizing hydrolysis of DDGS by xylanases in combinationwith protease and phytase

Modelling of the cDDGS response data predicted the greatestsimultaneous solubilization of pentosan and protein at 0.2 g/kgfeed phytase and 8.6 × 105 U/kg feed protease for XylA and XylDtreatment, and 8.6 × 105 U/kg feed protease for XylB and XylC treat-ment (Fig. A.1). The greatest increase in pentosan solubilization waspredicted for XylA treatment (113.5%) followed by XylC (71.8%),

XylD (46.1%), and XylB (38.6%) compared with the control. Fur-thermore, optimized conditions predicted the greatest increase inprotein solubilization for the XylA treatment (30.8%) followed byXylD (25.9%), XylC (20.6%), and XylB (16.9%) compared with thecontrol. When compared with the effects of xylanase treatment byitself, the optimized conditions with protease and phytase furtherincreased the solubilization of pentosan and protein by 19.7–36.3%and 15.1–20.2%, respectively.

Modelling of the wDDGS response data predicted the great-est simultaneous solubilization of pentosan and protein at8.6 × 105 U/kg feed protease for all xylanase treatments (Fig. A.2).Compared to control, the optimized conditions increased the sol-ubilization of protein by XylD treatment (51.5%) followed by XylA(50.4%), XylC (44.9%), and XylB (42.2%), respectively. The optimizedconditions with protease increased the protein solubilization by15.7–32.9% compared with the four individual xylanase treatments.No effect of the optimized conditions was observed on the solubi-lization of pentosan.

For cDDGS, the predictability of the PLS models were satisfactorywith R2 = 0.79-0.95 and Q2 = 0.85-0.98 for pentosan solubilization,and R2 = 0.65-0.79 and Q2 = 0.73-0.97 for protein solubilization. ForwDDGS, the predictability of the PLS models of XylA and XylCwere satisfactory for protein solubilization with R2 = 0.70–0.84 andQ2 = 0.65–0.78, whereas, the PLS models for XylB and Xyl D wereunsatisfactory. The PLS models for prediction of pentosan solu-bilization in wDDGS was unsatisfactory with R2 = 0.00–0.41 andQ2 = −0.20 to 0.21, indicating the maximum solubilization of pen-

M.B. Pedersen et al. / Biochemical Engineering Journal 98 (2015) 99–106 103

Table 2PLS regression models on the release of pentosan and protein by xylanase, protease, and phytase, with parameters and interactions.

Xylanase A Xylanase B Xylanase C Xylanase D

Corn DDGS

Pentosan (N = 17) Pentosan (N = 17) Pentosan (N = 16) Pentosan (N = 18)

Coefficient P-value Coefficient P-value Coefficient P-value Coefficient P-value

Protease 202.2 <0.0001 164.6 <0.0001 165.5 <0.0001 214.4 <0.0001Phytase 7.3 0.393 −27.2 0.162 0.5 0.974 21.4 0.197Pro × Pro −42.8 0.003 −55.0 0.035 −54.3 0.033Constant 2891.8 <0.0001 1775.6 <0.0001 2360.9 <0.0001 1901.8 <0.0001R2 0.95 0.79 0.86 0.85Q2 0.98 0.85 0.92 0.93

Protein (N = 18) Protein (N = 17) Protein (N = 17) Protein (N = 18)Protease 273.1 <0.0001 249.0 <0.0001 215.4 <0.0001 266.9 <0.0001Phytase −12.1 0.782 −29.5 0.058 −16.6 0.527 −3.9 0.906Pro × Pro −52.3 0.039Pro × Phy 46.4 0.009Constant 3818.2 <0.0001 3497.7 <0.0001 3542.5 <0.0001 3660.6 <0.0001R2 0.65 0.79 0.75 0.77Q2 0.73 0.97 0.84 0.82

Wheat DDGS

Pentosan (N = 18) Pentosan (N = 17) Pentosan (N = 16) Pentosan (N = 17)

Coefficient P-value Coefficient P-value Coefficient P-value Coefficient P-value

Protease −90.5 0.250 −10.2 0.883 99.6 0.091 −166.3 0.114Phytase −91.1 0.247 2.5 0.971 2.0 0.972 −168.5 0.110Pro × Pro −159.7 0.049Constant 10333.0 <0.0001 10003.7 <0.0001 10486.7 <0.0001 10457.3 <0.0001R2 0.16 0.00 0.41 0.29Q2 −0.10 −0.20 0.13 0.21

Protein (N = 16) Protein (N = 18) Protein (N = 18) Protein (N = 18)Protease 934.4 <0.0001 688.1 0.003 756.1 <0.0001 522.8 0.020Phytase 35.2 0.761 −87.0 0.660 −280.0 0.060 −260.8 0.215Constant 8113.7 <0.0001 7675.7 <0.0001 7532.2 <0.0001 8228.6 <0.0001R2 0.84 0.46 0.70 0.36Q2 0.78 0.29 0.65 0.22

Scaled and centered coefficients of the quadric models fitted with PLS, using Modde 9.1 software. Blanks indicate backward eliminated insignificant interactions at P > 0.05.Pro protease; Phy phytase.

tosan in wDDGS was achieved by xylanase treatment, thus, notaffected by additional protease or phytase treatment.

Based on modelling, the individual optimized conditions withfour xylanases, protease and phytase predicted 13.1–20.2% and80.6–84.3% solubilization of the total pentosan content, and14.8–16.6% and 29.3–31.2% solubilization of the total protein con-tent for c- and wDDGS, respectively.

3.5. Effects of xylanase and protease on solubilization of pentosanand protein from insoluble DDGS fraction

Similar to the previous observations on DDGS substrate, thefour xylanase treatments varied in the solubilization of pentosanand protein for both insoluble c- and wDDGS (Figs. 3 and 4). Thegreatest solubilization of insoluble cDDGS was observed for XylAwith an increased solubilization of 2.2 mg pentosan/ml and 1.8 mgprotein/ml. The remaining three xylanase treatments increasedthe solubilization of pentosan between 0.4 and 1.2 mg/ml andprotein between 0.6 and 1.2 mg/ml with XylC > XylB > XylD com-pared to control with a solubilization of 0.1 mg/ml and 0.2 mg/ml,respectively. Overall, the four xylanase treatments increasedthe solubilization of pentosan by 338–2052% and protein by311–1012% compared with the control (Fig. 3).

The greatest solubilization of pentosan from insoluble wDDGSwas observed for XylC treatment of 2.4 mg/ml followed by XylAof 2.3 mg/ml. XylD and XylB solubilized more pentosan in w- thancDDGS, with an increase of 1.5 mg/ml and 1.7 mg/ml, respectively

(Fig. 4A). The xylanase treatments increased protein solubilizationby 0.6–0.8 mg/ml compared to control of 0.2 mg/ml. Overall, thefour xylanase treatments increased the solubilization of pentosanby 926–1171% and the solubilization of protein by 421–561% com-pared with the control.

Protease addition increased the solubilization of pentosan andprotein from both insoluble c- and wDDGS, with the most pro-nounced effect on protein solubilization. Addition of proteaseincreased the solubilization of pentosan by 0.6 mg/ml (572%) and0.6 mg/ml (298%) compared with control for insoluble cDDGS(Fig. 3A) and wDDGS (Fig. 4A), respectively. Furthermore, pro-tease increased the protein solubilization by 9.4 mg/ml (5263%),and 7.1 mg/ml (4875%) for insoluble cDDGS (Fig. 3B) and wDDGS(Fig. 4B), respectively.

The effect of combining xylanase and protease was mostpronounced on the protein solubilization with an increase of8.5–9.9 mg/ml for insoluble cDDGS, and 6.7–7.1 mg/ml for insolu-ble wDDGS compared with xylanase treatment alone. Furthermore,the combination of xylanase and protease increased the solubi-lization of pentosan of 0.5–0.7 mg/ml for insoluble cDDGS, and0.2–0.4 mg/ml for insoluble wDDGS compared to xylanase treat-ment alone.

4. Discussion

In the current study supplementation with exogenous enzymeshad great effect on solubilizing pentosan and protein from both

104 M.B. Pedersen et al. / Biochemical Engineering Journal 98 (2015) 99–106

Control Pro XylA XylB XylC XylD

Solu

biliz

ed p

ento

san

(mg/

ml)

0,0

0,5

1,0

1,5

2,0

2,5

3,0

3,5 Xylanase Xylanase, protease

g

e

b

a

e,f

d d

c

f

d

Control Pro XylA XylB XylC XylD

Solu

biliz

ed p

rote

in (m

g/m

l)

0

2

4

6

8

10

12 XylanaseXylanase, protease

f

b

c

a

d,e

a

d

b

e,f

bA B

Fig. 3. Solubilization of pentosan (A) and protein (B) from insoluble cDDGS with xylanase alone and in combination with protease. Pro protease. Error bars indicate S.D. Barswith different letters are significantly different at P < 0.05.

Con trol Pro XylA XylB XylC XylD

Solu

biliz

ed p

ento

san

(mg/

ml)

0,0

0,5

1,0

1,5

2,0

2,5

3,0

Xylanase Xylanase, protea se

f

e

b,c

a

d

ca,b,c

a,b

d

b,c

Control Pro XylA XylB XylC XylD

Solu

biliz

ed p

rote

in (m

g/m

l)

0

2

4

6

8

10 XylanaseXylanase, pro teaseA B

f

c

d

a,b

e

a

d

a,b

d,e

b

Fig. 4. Solubilization of pentosan (A) and protein (B) from insoluble wDDGS with xylanase alone and in combination with protease. Pro protease. Error bars indicate S.D. Barswith different letters are significantly different at P < 0.05.

c- and wDDGS, indicating a high potential of these enzymes todegrade DDGS fiber and release nutrients in vitro. However, largedifferences was observed in the solubilization of pentosan betweenthe two substrates, illustrated by the highest enzymatic solubi-lization equals approximately 20% and 85% of the total in c- andwDDGS, respectively. Differences in the compositional complexitybetween c- and wDDGS are the most likely reason for the markedlydifferent responses to enzymatic treatments between the two sub-strates. The observed compositional characteristics of the c- andwDDGS used this study were in line with previous reports on DDGS[5,7] and parent grains [21]. Despite similar arabinoxylan content,cDDGS contained a greater fraction of insoluble arabinoxylan andcDDGS had higher A/X–and UA/X ratio, which is indicative for ahigher average degree of substitution in c- than wDDGS. Further-more, a higher ferulic acid dehydrodimers and–dehydrotrimerscrosslinking of arabinoxylan [25,26,27] in c- than wDDGS mayimpair the enzymatic degradation of the plant cell wall [28,29]. Ithas been found that corn contains approximately 5–7 times morediferulates in the insoluble dietary fiber fraction compared to wheat[30,31]. The combined greater A:X- and UA:X-ratio, along with anexpected greater content of ferulic acids in cDDGS than wDDGSimplies a more complex structure of the heteroxylans in cDDGS,which is in line with previous reports on the greater branching andcomplexity of corn arabinoxylan compared to wheat [13,27,32,33].The observed greater enzymatic effect on degradation of AX inwDDGS than cDDGS is well in line with a previous study describing

higher enzymatic degradation of wheat than corn [34]. In addi-tion to the solubilization of pentosan, xylanase addition furtherincreased the release of protein from both c- and wDDGS, indi-cating close association of pentosan and protein in the cell wall.Fiber degrading enzymes have previously been reported to increasein vitro digestibility of protein in cDDGS, however, with no reportedeffect on solubilization of NSP [35].

The use of purified xylanases, deprived of side activities, allowedfor a direct comparison of xylanases originating from two differentGH families. Considering the difference in catalytic versatility andsubstrate affinity of GH10 (XylA) and GH11 (XylB, XylC, and XylD)[17,18], the superior effect of XylA on the degradation of cDDGSis likely due to the greater substrate complexity than wDDGS. Thethree GH11 xylanases showed different effects on especially cDDGSdegradation, suggesting that XylC poses greater affinity towardinsoluble and more complex substrates than XylB and XylD. DespiteGH-relation, xylanases may encompass differences in substrateaffinity, and pH and temperature optimum, which may affect thedegree of enzymatic degradation [18]. The individual performanceof the four xylanases on the insoluble cDDGS was in line withthe results on cDDGS. As cDDGS contain lower amounts of solu-ble NSP than wDDGS, the change in substrate composition causedby removing soluble NSPs is logically less pronounced. However,removal of the soluble NSPs may affect the overall solubilizationof pentosan, as the presence of both soluble and insoluble sub-strate will compete for the xylanase activity, which potentially will

M.B. Pedersen et al. / Biochemical Engineering Journal 98 (2015) 99–106 105

limiting the solubilization from the insoluble NSP fraction. Thelarger difference in performance between the individual xylanaseson pentosan solubilization from insoluble wDDGS than wDDGSmay be explained by the different substrate affinities amongxylanases toward the insoluble NSP fraction (Fig. 4A).

As expected, addition of protease increased the solubilization ofprotein from both c- and wDDGS, and from insoluble c- and wDDGSfractions. Compared to the DDGS substrate, the effect of proteaseaddition was more pronounced on the protein solubilization frominsoluble DDGS substrate. It can be speculated that removal ofthe soluble protein increase the protease activity toward degrada-tion of the insoluble protein, consequently increasing the proteinsolubilization. Addition of protease increased the solubilization ofpentosan from cDDGS, indicating a close interaction of protein andarabinoxylan in the cell wall, as previously reported [16]. The latterwas further confirmed, as the greatest solubilization of both pen-tosan and protein was achieved by combining both xylanase andprotease, indicating synergistically interactions between the twodifferent hydrolases. Protein has previously been described to belocated within the aleurone layer of cereals [36,37], thus, it can bespeculated that the addition of xylanase would increase the acces-sibility of cell wall-encapsulated protein by opening up the cellwall structure through arabinoxylan degradation. Addition of pro-tease could potentially hydrolyze endogenous xylanase inhibitorspresent in cereals; i.e., Triticum aestivum xylanase inhibitors (TAXI)and Xylanase Inhibiting Proteins (XIP) [38,39,40]. Inactivation ofthese xylanase inhibitors would consequently improve xylanaseperformance, if inhibited. Addition of protease to wDDGS had noadditional effects on pentosan release, likely due to the fact thatthe solubilization of pentosan was apparently already maximizedby the addition of xylanase alone.

Endo-xylanases solubilize arabinoxylan (pentosan) by cleavingthe �-1,4-glycosyl linkages within the �-1,4-xylose backbone ofinsoluble as well as soluble arabinoxylan, thus, partially solubiliz-ing insoluble arabinoxylan and fragmenting soluble arabinoxylaninto arabinoxylan-oligosaccharides [17]. The large difference in theproportion of solubilized pentosan out of total pentosan betweenc- and wDDGS may indicate differences in which botanical grainfractions the pentosan is solubilized from in the two DDGS sources.In corn, endosperm arabinoxylan comprise approximately 20% ofthe total content of arabinoxylan, whereas endosperm arabinoxy-lan in wheat comprise approximately 25%, calculated based ondata from Bach Knudsen [21] and Watson [41]. As the maximumenzymatic solubilization of pentosan equals approximately 20% thetotal pentosan in cDDGS, it can be speculated that the majorityof the pentosan is solubilized from the endosperm fraction. Cornendosperm arabinoxylan is likely the most readily accessible ara-binoxylan for enzymatic degradation, as this arabinoxylan is notcross-linked to the same degree as arabinoxylan in the bran [42,43].Furthermore, corn bran arabinoxylan has been acknowledged as arecalcitrant substrate regarding enzymatic degradation due to thehighly branched structure [44,45]. For wDDGS, the maximum enzy-matic solubilization of pentosan equals approximately 85% of thetotal pentosan in wDDGS, indicating that the pentosan is solubi-lized from other botanical grain fractions besides the endosperm.To further increase the enzymatic degradation of arabinoxylan inDDGS (especially in cDDGS), other minor enzyme activities may benecessary, such as L-˛-arabinofuranosidase, feruloyl esterase, and˛-D-glucuronidase.

Overall, phytase had no significant effects on the solubilizationof neither pentosan nor protein in both c- and wDDGS. Phytate-bound P in corn is in the range from 61 to 77% of total P [46],whereas values of phytate-bound P in cDDGS have been reported tobe 30–35% of total P [20,47]. A decrease in phytate-bound P in fer-mentation products compared to parent grains has been reportedpreviously [48,49]. Partial degradation of phytate during the

fermentation process may be mediated by the presence of endoge-nous phytase in yeast or exogenous phytase added to increase thestarch degradation and ethanol yield [50].

5. Conclusion

The use of xylanases in combination with protease poses a largepotential regarding the degradation of arabinoxylan and release ofnutrients from both c- and wDDGS. Furthermore, positive effect ofprotease addition on top of xylanase on cDDGS degradation sug-gests close protein and fiber interactions in the DDGS matrix. Largedifferences in xylanase performance were observed between c- andwDDGS, which is related to the different degree of complexity ofthe arabinoxylan fraction and fiber matrix. The results presented inthis study indicate a large potential of using xylanase in combina-tion with protease for efficient degradation of DDGS and nutrientrelease.

Acknowledgements

This project was partially supported by The Danish Ministryof Science, Innovation and Higher Education. We acknowledge T.Christensen and M. K. Sandfær for assistance in acquiring phytase,M. Chow, D. Ko and A. Riley for assistance in collecting DDGS sam-ples, and H. Etzerodt, L. Märcher and S. G. Handberg for technicalassistance.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.bej.2015.02.036.

References

[1] K. Liu, Chemical composition of distillers grains, a review, J. Agric. Food Chem.59 (2011) 1508–1526.

[2] Y.V. Wu, Determination of neutral sugars in corn distillers’ dried grains corndistillers’ dried solubles, and corn distillers’ dried grains with solubles, J.Agric. Food Chem. 42 (1994) 723–726.

[3] H.H. Stein, G.C. Shurson, Board-invited review: the use and application ofdistillers dried grains with solubles in swine diets, J. Anim. Sci. 87 (2009)1292–1303.

[4] W.G. Nuez Ortín, P. Yu, Nutrient variation and availability of wheat DDGS,corn DDGS and blend DDGS from bioethanol plants, J. Sci. Food Agric. 89(2009) 1754–1761.

[5] G.P. Widyaratne, R.T. Zijlstra, Nutritional value of wheat and corn distiller’sdried grain with solubles: digestibility and digestible contents of energy,amino acids and phosphorus, nutrient excretion and growth performance ofgrower-finisher pigs, Can. J. Anim. Sci. 87 (2007) 103–114.

[6] P.H. Robinson, K. Karges, M.L. Gibson, Nutritional evaluation of fourco-product feedstuffs from the motor fuel ethanol distillation industry in theMidwestern USA, Anim. Feed Sci. Technol. 146 (2008) 345–352.

[7] M.B. Pedersen, S. Dalsgaard, K.E.B. Knudsen, S. Yu, H.N. Lærke, Compositionalprofile and variation of distillers dried grains with solubles from variousorigins with focus on non-starch polysaccharides, Anim. Feed Sci. Technol.197 (2014) 130–141.

[8] K.E. Bach Knudsen, B.B. Jensen, I. Hansen, Digestion of polysaccharides andother major components in the small and large intestine of pigs fed on dietsconsisting of oat fractions rich in beta-d-glucan, Br. J. Nutr. 70 (1993) 537–556.

[9] M. Le Gall, A. Serena, H. Jørgensen, P.K. Theil, K.E. Bach Knudsen, The role ofwhole-wheat grain and wheat and rye ingredients on the digestion andfermentation processes in the gut – a model experiment with pigs, Br. J. Nutr.102 (2009) 1590–1600.

[10] O. Adeola, A.J. Cowieson, BOARD-INVITED REVIEW: opportunities andchallenges in using exogenous enzymes to improve nonruminant animalproduction, J. Anim. Sci. 89 (2011) 3189–3218.

[11] S.Y. Yoon, Y.X. Yang, P.L. Shinde, J.Y. Choi, J.S. Kim, Y.W. Kim, K. Yun, J.K. Jo, J.H.Lee, S.J. Ohh, I.K. Kwon, B.J. Chae, Effects of mannanase and distillers driedgrain with solubles on growth performance nutrient digestibility, and carcasscharacteristics of grower-finisher pigs, J. Anim. Sci. 88 (2010) 181–191.

[12] M.S. Izydorczyk, C.G. Biliaderis, Cereal arabinoxylans: advances in structureand physicochemical properties, Carbohydr. Polym. 28 (1995) 33–48.

[13] M.R. Bedford, Mechanism of action and potential environmental benefits fromthe use of feed enzymes, Anim. Feed Sci. Technol. 53 (1995) 145–155.

106 M.B. Pedersen et al. / Biochemical Engineering Journal 98 (2015) 99–106

[14] R.D. Hartley, Carbohydrate esters of ferulic acid as components of cell-walls ofLolium multiflorum, Phytochemistry 12 (1973) 661–665.

[15] C. Lequart, J.-M. Nuzillard, B. Kurek, P. Debeire, Hydrolysis of wheat bran andstraw by an endoxylanase: production and structural characterization ofcinnamoyl-oligosaccharides, Carbohydr. Res. 319 (1999) 102–111.

[16] C.B. Faulds, G. Mandalari, R.B. Lo Curto, G. Bisignano, P. Christakopoulos, K.W.Waldron, Synergy between xylanases from glycoside hydrolase family 10 andfamily 11 and a feruloyl esterase in the release of phenolic acids from cerealarabinoxylan, Appl. Microbiol. Biotechnol. 71 (2006) 622–629.

[17] P. Biely, M. Vrsanská, M. Tenkanen, D. Kluepfel, Endo-�-1,4-xylanase families:differences in catalytic properties, J. Biotechnol. 57 (1997) 151–166.

[18] G. Paes, J.-G. Berrin, J. Beaugrand, GH11 xylanases:structure/function/properties relationships and applications, Biotechnol. Adv.30 (2012) 564–592.

[19] C.K. Jones, J.R. Bergstrom, M.D. Tokach, J.M. DeRouchey, R.D. Goodband, J.L.Nelssen, S.S. Dritz, Efficacy of commercial enzymes in diets containing variousconcentrations and sources of dried distillers grains with solubles for nurserypigs, J. Anim. Sci. 88 (2010) 2084–2091.

[20] J.L. Yanez, E. Beltranena, M. Cervantes, R.T. Zijlstra, Effect of phytase andxylanase supplementation or particle size on nutrient digestibility of dietscontaining distillers dried grains with solubles cofermented from wheat andcorn in ileal-cannulated grower pigs, J. Anim. Sci. 89 (2011) 113–123.

[21] K.E. Bach Knudsen, Carbohydrate and lignin contents of plant materials usedin animal feeding, Anim. Feed Sci. Technol. 67 (1997) 319–338.

[22] X. Rouau, A. Surget, A rapid semi-automated method for the determination oftotal and water-extractable pentosans in wheat flours, Carbohydr. Polym. 24(1994) 123–132.

[23] T. Lundstedt, E. Seifert, L. Abramo, B. Thelin, Å. Nyström, J. Pettersen, R.Bergman, Experimental design and optimization, Chemom. Intell. Lab. Syst.42 (1998) 3–40.

[24] B. Liebmann, A. Friedl, K. Varmuza, Applicability of near-infraredspectroscopy for process monitoring in bioethanol production, Biochem. Eng.J. 52 (2010) 187–193.

[25] M. Choct, Feed non-starch polysaccharides: chemical structures andnutritional significance, Feed Mill. Int. (1997) 13–26.

[26] F. Piston, C. Uauy, L. Fu, J. Langston, J. Labavitch, J. Dubcovsky,Down-regulation of four putative arabinoxylan feruloyl transferase genesfrom family PF02458 reduces ester-linked ferulate content in rice cell walls,Planta 231 (2010) 677–691.

[27] M.L. Jilek, M. Bunzel, Dehydrotriferulic and dehydrodiferulic acid profiles ofcereal and pseudocereal flours, Cereal Chem. 90 (2013) 507–514.

[28] J.H. Grabber, R.D. Hatfield, J. Ralph, Diferulate cross-links impede theenzymatic degradation of non-lignified maize walls, J. Sci. Food Agric. 77(1998) 193–200.

[29] J.H. Grabber, J. Ralph, R.D. Hatfield, Ferulate cross-links limit the enzymaticdegradation of synthetically lignified primary walls of maize, J. Agric. FoodChem. 46 (1998) 2609–2614.

[30] M. Bunzel, J. Ralph, J.M. Marita, R.D. Hatfield, H. Steinhart, Diferulates asstructural components in soluble and insoluble cereal dietary fibre, J. Sci. FoodAgric. 81 (2001) 653–660.

[31] D. Dobberstein, M. Bunzel, Separation and detection of cell wall-bound ferulicacid dehydrodimers and dehydrotrimers in cereals and other plant materialsby reversed phase high-performance liquid chromatography with ultravioletdetection, J. Agric. Food Chem. 58 (2010) 8927–8935.

[32] L. Saulnier, J. Vigouroux, J.-F. Thibault, Isolation and partial characterization offeruloylated oligosaccharides from maize bran, Carbohydr. Res. 272 (1995)241–253.

[33] J. Yang, M.X. Maldonado-Gómez, R.W. Hutkins, D.J. Rose, Production andin vitro fermentation of soluble non-digestible feruloylated oligo- andpolysaccharides from maize and wheat brans, J. Agric. Food Chem. 62 (2014)159–166.

[34] D. Rose, G. Inglett, A Method for the determination of soluble arabinoxylanreleased from insoluble substrates by xylanases, Food Anal. Methods 4 (2011)66–72.

[35] S. de Vries, A.M. Pustjens, M.A. Kabel, S. Salazar-Villanea, W.H. Hendriks, W.J.J.Gerrits, Processing technologies and cell wall degrading enzymes to improvenutritional value of dried distillers grain with solubles for animal feed: anin vitro digestion study, J. Agric. Food Chem. 61 (2013) 8821–8828.

[36] A.D. Evers, L. O’Brien, A.B. Blakeney, Cereal structure and composition, Aust. J.Agric. Res. 50 (1999) 629–650.

[37] P.R. Shewry, N.G. Halford, Cereal seed storage proteins: structures, propertiesand role in grain utilization, J. Exp. Bot. 53 (2002) 947–958.

[38] K. Fierens, K. Brijs, C.M. Courtin, K. Gebruers, H. Goesaert, G. Raedschelders, J.Robben, S. Van Campenhout, G. Volckaert, J.A. Delcour, Molecularidentification of wheat endoxylanase inhibitor TAXI-I, member of a new classof plant proteins, FEBS Lett. 540 (2003) 259–263.

[39] G.O. Elliott, W.R. McLauchlan, G. Williamson, P.A. Kroon, A wheat xylanaseinhibitor protein (XIP-I) accumulates in the grain and has homologues inother cereals, J. Cereal Sci. 37 (2003) 187–194.

[40] A.V. Gusakov, Proteinaceous inhibitors of microbial xylanases, Biochem.(Mosc.) 75 (2010) 1185–1199.

[41] S.A. Watson, Structure and composition, in: S.A. Watson, P.E. Ramstad (Eds.),Corn Chemistry and Technology, Am. Assoc. Cereal Chem., St. Paul, MN, 1987,pp. 53–82.

[42] L. Saulnier, J.-F. Thibault, Ferulic acid and diferulic acids as components ofsugar-beet pectins and maize bran heteroxylans, J. Sci. Food Agric. 79 (1999)396–402.

[43] M. Bunzel, J. Ralph, P. Brüning, H. Steinhart, Structural identification ofdehydrotriferulic and dehydrotetraferulic acids isolated from insoluble maizebran fiber, J. Agric. Food Chem. 54 (2006) 6409–6418.

[44] L. Saulnier, C. Marot, M. Elgorriaga, E. Bonnin, J.F. Thibault, Thermal andenzymatic treatments for the release of free ferulic acid from maize bran,Carbohydr. Polym. 45 (2001) 269–275.

[45] J. Agger, A. Vikso-Nielsen, A.S. Meyer, Enzymatic xylose release frompretreated corn bran arabinoxylan: differential effects of deacetylation anddeferuloylation on insoluble and soluble substrate fractions, J. Agric. FoodChem. 58 (2010) 6141–6148.

[46] W. Eeckhout, M. De Paepe, Total phosphorus, phytate-phosphorus andphytase activity in plant feedstuffs, Anim. Feed Sci. Technol. 47 (1994) 19–29.

[47] F.N. Almeida, H.H. Stein, Effects of graded levels of microbial phytase on thestandardized total tract digestibility of phosphorus in corn and corncoproducts fed to pigs, J. Anim. Sci. 90 (2012) 1262–1269.

[48] S.E.O. Mahgoub, S.A. Elhag, Effect of milling soaking, malting, heat-treatmentand fermentation on phytate level of four Sudanese sorghum cultivars, FoodChem. 61 (1998) 77–80.

[49] J. Liang, B.-Z. Han, M.J.R. Nout, R.J. Hamer, Effects of soaking germination andfermentation on phytic acid, total and in vitro soluble zinc in brown rice, FoodChem. 110 (2008) 821–828.

[50] E. Khullar, J.K. Shetty, K.D. Rausch, M.E. Tumbleson, V. Singh, Use of phytasesin ethanol production from e-mill corn processing, Cereal Chem. 88 (2011)223–227.

63

PAPERIV

XylanaseIncreasedtheIlealDigestibilityofNon‐StarchPolysaccharidesand

ConcentrationofLowMolecularWeightNon‐DigestibleCarbohydratesinPigsFed

HighLevelsofwheatDDGS

MadsB.Pedersen,ShukunYu,SusanArent,SørenDalsgaard,KnudErikB.Knudsen,andHelleN.

Lærke

Revisedversionacceptedforpublicationin:

JournalofAnimalScience,March2015

64

PAPER IV

65

RUNNING TITLE: Wheat DDGS and xylanase in swine diets

TITLE

Xylanase increased the ileal digestibility of non-starch polysaccharides and concentration of

low molecular weight non-digestible carbohydrates in pigs fed high levels of wheat DDGS1,2

AUTHORS

M. B. Pedersen*,†,3, S. Yu*, S. Arent*, S. Dalsgaard*, K. E. Bach Knudsen†, H. N. Lærke†

AUTHOR AFFILIATIONS

*Enzyme Development, DuPont Industrial Biosciences ApS, Edwin Rahrs Vej 38, DK-8220

Brabrand, Denmark

†Department of Animal Science, Aarhus University, Blichers Allé 20, DK-8830 Tjele, Denmark

Footnotes

1M. Walsh and A. Riley are thanked for their assistance in sourcing wheat DDGS. The authors

would also like to thank staff at Aarhus University, Department of Animal Science, and DuPont

Industrial Biosciences for technical assistance. This project was partially supported by The

Danish Ministry of Science, Innovation and Higher Education.

2M.B.P, S.Y., S.A., and S.D. are or have been employees of DuPont Industrial Biosciences

3Corresponding author: madsb.pedersen@anis.au.dk

66

ABSTRACT: The objective was to study the effect of a commercially available xylanase (DAN)

and an experimental xylanase (EX), and EX in combination with protease (EXP), on the

degradation of non-starch polysaccharides (NSP) and coefficient of apparent ileal digestibility

(CAID) of nutrients in wheat Distillers Dried Grains with Solubles (wDDGS). The control and

three enzyme diets contained 96% DDGS supplemented with vitamins, minerals, L-lysine,

chromic oxide in addition to enzyme premix. Eight ileal cannulated pigs were used in a 4× 4

Latin Square design. The experimental period lasted 7 d; 4 d for adaptation, and the ileal

digesta and spot-sampling of feces were collected for 8 h on d 5 and 7. Digesta samples were

analyzed for composition of total non-digestible carbohydrates (NDC) total- and soluble NSP,

low molecular weight (LMW)-NDC, OM, crude protein, fat, starch and marker. Compared with

the control diet the addition of DAN, EX, and EXP increased the CAID of NSP-AX by 32% (P <

0.001), 28% (P = 0.001), and 24% (P = 0.004), respectively. In addition, EXP increased the

CAID of NCP-glucose by 21% compared with control (P = 0.005). Compared to control the

addition of EX, EXP and DAN decreased the concentration of soluble AX in ileal digesta by 40%

(P < 0.0001), 40% (P < 0.0001), and 21% (P = 0.022), respectively. Furthermore, the addition

of DAN, EXP, and EX increased the concentration of LMW-AX in ileal digesta by 40% (P =

0.0001), 36% (P = 0.0006), and 24% (P = 0.023), respectively, compared with control.

Addition of EX and EXP decreased the concentration of soluble NSP in ileal digesta by 25% (P

= 0.001) and 26% (P < 0.001), respectively, compared with the control diet. Addition of DAN

(P < 0.0001) and EXP (P = 0.013) increased the arabinose-to-xylose ratio in the insoluble AX

fraction in ileal digesta compared with the control diet. In addition, a small but significant

increase in uronic acid-to-xylose ratio was observed in the ileal insoluble NSP-fraction for the

DAN treatment (P < 0.0001), compared with the control diet. Enzyme addition did not affect

CAID of OM, crude protein, starch, and fat (P > 0.3).

PAPER IV

67

KEYWORDS: animal feed, distillers dried grains with solubles, ileal digestibility, non-starch

polysaccharides, protease, xylanase

INTRODUCTION

Wheat Distillers Dried Grains with Solubles (wDDGS) is a by-product from the grain-

based ethanol industry. In cooler climates like Europe and Western Canada, wheat instead of

corn is used for ethanol production (Avelar et al., 2010; Lyberg et al., 2013). wDDGS is rich in

crude protein, fat, minerals and dietary fiber, and along with the rapid growth of the ethanol

industry, increasing quantities of wDDGS are available for use in animal feed (Widyaratne and

Zijlstra, 2007). The high content of dietary fiber in wDDGS may cause nutritional challenges

when implemented at high levels in feed formulations for monogastric animals (Theander et

al., 1989).

Non-starch polysaccharides (NSP) originate from the cell walls in the botanical grain

fractions and make up approximately 25-30 % of DM in wDDGS, with arabinoxylan (AX) and

cellulose as the two major fractions (Pedersen et al., 2014). Arabinoxylan consist of a D-xylose

backbone joined together by β-1,4-linkages substituted with L-arabinose residues along the

chain, and may be linked together by e.g. ferulic acid cross-linkages potentially impairing the

enzymatic degradation of the cell wall (Bunzel, 2010; Izydorczyk and Biliaderis, 1995; Ishii,

1997). Dietary supplementation of exogenous xylanase and protease may depolymerize cell

walls and structural proteins in vivo, thus, potentially increase the release of nutrients and

fiber degradation products of a lower molecular size.

The current study focused on the effects of two different xylanases and a combination of

one of the xylanases with protease on diets fed to growing pigs at high wDDGS inclusion rate.

It was hypothesized that the two xylanases have different substrate affinities, which affect

68

fiber degradation and depolymerization dynamics in vivo. The degradation of the fiber matrix,

composition of ileal digesta, coefficient of apparent ileal digestibility (CAID), coefficient of

apparent total tract digestibility (CATTD) of NSPs and macronutrients were examined.

MATERIALS AND METHODS

The animal experiment was conducted according to license obtained from the Danish

Animal Experiments Inspectorate, Danish Ministry of Food, Agriculture and Fisheries, Danish

Veterinary and Food Administration and in compliance with the Danish Ministry of Justice,

Law no. 253 of March 8 2013 concerning experiments with animals and care of experimental

animals.

Enzymes and Experimental Diets

The effect of two different β-1,4-endoxylanases and one protease were investigated;

DAN, xylanase belonging to the glycoside hydrolysis family (GH) 11 (Danisco Xylanase 8000G,

Danisco Animal Nutrition, Marlborough, UK); EX, an experimental xylanase belonging to the

GH10 family; EXP, EX in combination with protease (Multifect P-3000, Danisco Animal

Nutrition, Marlborough, UK). Enzymes were supplied as a dry premixed formulation by

DuPont Industrial Biosciences ApS, Denmark.

The four experimental diets were as follows: Control, containing no exogenous xylanase

or protease; DAN, containing 25,000 U/kg feed of DAN-xylanase; EX, containing 25,000 U/kg

feed of EX-xylanase; EXP, containing 25,000 U/kg feed of EX-xylanase in combination with

protease of 40,000 U/kg feed. Phytase (Phyzyme XP, Danisco Animal Nutrition, Marlborough,

UK) was applied in all four experimental diets at 500 FTU/kg feed.

PAPER IV

69

One U of protease activity is equal the amount of enzyme that releases 1 µg of tyrosine

equivalent per min from a casein substrate at pH 7.5 and 40°C. Xylanase activity was analyzed

prior to preparation of premixes (concentrated mix of phytase, xylanase and colored corn

grits on a heat treated wheat based carrier). One U of xylanase activity is equal to the amount

of enzyme that releases 0.5 µmol of reducing sugar equivalents per min from wheat

arabinoxylan at pH 4.2 and 50°C.

WDDGS was produced by Ensus Limited, Yarm, UK. The experimental diets were

supplied with feed supplements, enzyme premix, and dust binder (glycerol) as listed in Table

1.

Animals, Housing and Feeding

Eight male pigs (Duroc x Landrace-Yorkshire) with an average BW of 32.4 ± 2.4 kg were

fitted surgically with a permanent simple T-cannula 15 cm anterior to the ileal-cecal junction

following the procedure by Jørgensen et al. (1992) and allowed to recover for two weeks

before the study. During the recovery period, the pigs were gradually adapted to the

experimental diets by feeding increased levels of DDGS.

The eight pigs were housed individually in 3 x 2 m pens without bedding and with

elevated plastic grids covering half of the pen, which allowed the pigs to rest and stay dry in

an environmentally controlled room, and temperature maintained at 18°C±2°C. The pens

allowed freedom of movement during the entire experiment and visual- and nose-to-nose

contact between pigs in adjacent pens. A feeder and a nipple drinker were installed in each

pen.

The pigs were fed each experimental diet for seven days according to a double 4 × 4

Latin Square design, to give eight observations per diet. The pigs were weighed at the

70

beginning of each period, and the daily feed allowance adjusted on period basis according to

the estimated average body weight on day 4. To avoid feed remnants the pigs were fed

restrictively, equal to 70% of the daily requirement for finisher pigs according to the Danish

recommendations, corresponding on average to 2.1 x maintenance (i.e. 106 kcal ME kg-1

BW0.75) using an estimated ME of 2600 kcal kg-1. The pigs were fed three times a day in three

equal meals at 0730 h, 1530 h and 2330 h. Diets were provided as dry meal and water

supplied ad libitum throughout the experiment period.

Sampling and Sample Preparation

On day 5 and 7, ileal effluent was collected continuously for 8 h after the morning meal

using plastic bags attached to the open T-cannula barrel. Two to three drops of an aqueous

solution of 0.2% sodium azide (Sigma-Aldrich) were added to each collection bag to prevent

microbial activity. The bags were removed whenever they were filled with digesta or at least

once every 30 min. Furthermore, spot samples of feces were also collected on day 5 and 7.

Collected feces and digesta samples were pooled per pig over each collection period and

stored in a freezer at -20°C. When the trial was completed all pooled samples were freeze-

dried, milled (<0.5 mm), and stored at room temperature (22°C) until further analyses.

Chemical Analyses

Dry matter of diets and freeze-dried digesta and feces were determined by drying to

constant weight at 105 °C for 20 h, and ash analyzed according to the AOAC method (AOAC,

2006), and chromium(III) oxide as described by (Schürch et al., 1950). Diets and ileal digesta

were analyzed for crude protein (nitrogen × 6.25) determined by the Dumas method (Hansen,

PAPER IV

71

1989), HCl-fat according to the Stoldt procedure (Stoldt, 1952), and starch determined by the

enzymatic-colorimetric method (Bach Knudsen, 1997), with samples analyzed in duplicates.

The NSP in diets, ileal digesta and feces samples was determined essentially according to

Bach Knudsen (1997), with the modification that 2 M sulfuric acid for 1 h was used to

hydrolyze the non-cellulosic polysaccharides (NCP) rather than 1 M sulfuric acid for 2 h. The

total content of non-digestible carbohydrates (NDC) including low molecular weight (LMW)

residues, typically of degree of polymerization less than 10, was determined by direct acid

hydrolysis without starch removal and alcohol precipitation and corrected for starch content.

Klason lignin was measured gravimetrically as the residue resistant to hydrolysis by 2 M

sulfuric acid for 1 h (Bach Knudsen, 1997).

To determine residual ileal xylanase activity 5.0 g of diet and 0.1 g of freeze dried ileal

digesta were extracted by stirring for 10 min at room temperature (22 °C) in McIlvaine buffer,

pH 5.0. After filtration and dilution, the samples were equilibrated at 50 °C and a 60 mg

Xylazyme tablet (Megazyme International, Wicklow, Ireland) was added. After 60 min

incubation, the reaction was stopped with a 2% Tris(hydroxymethyl)aminomethane solution.

After centrifugation for10 min at 1,500 × g the optical density was measured at 590 nm. The

xylanase activity in samples was quantified relative to a standard curve with increasing

amounts of known DAN-xylanase activity. All samples and standards were analyzed in

duplicates.

Calculations and statistical Analyses

The CAID and CATTD of organic matter, starch, crude protein, HCl-fat, and NSPs were

calculated by the indicator method relative to dietary and ileal concentrations of chromic

72

oxide (Schürch et al., 1952). Calculations were based on the average nutrient composition

across the four diets.

The concentration of LMW-NDC sugars was calculated as the difference between total

NDC determined without ethanol precipitation (corrected for starch content) and total NSP

content determined by the procedure including alcohol precipitation. Arabinoxylan content

was calculated as the sum of anhydrous arabinose and xylose.

Residual xylanase activity in ileal digesta was calculated as:

𝑅𝑒𝑠𝑖𝑑𝑢𝑎𝑙 𝑖𝑙𝑒𝑎𝑙 𝑎𝑐𝑡𝑖𝑣𝑖𝑡𝑦, % = 100 × 𝐶𝑟𝑓𝑒𝑒𝑑 × (𝐴𝑖𝑙𝑒𝑎𝑙 − 𝐴𝑖𝑙𝑒𝑎𝑙,𝑐𝑜𝑛𝑡𝑟𝑜𝑙)

𝐴𝑑𝑖𝑒𝑡 × 𝐶𝑟𝑖𝑙𝑒𝑎𝑙

, where Cr is the chromic oxide content and A is the xylanase activity, with all values on DM

basis.

Data were analyzed using the GLM procedure of Minitab 16 (Minitab Inc., USA), with

enzyme treatment and week as fixed factors, and pig as random factor followed by Tukey’s

pair wise comparison with overall significance level at P = 0.05. Values in the tables were

reported as means and pooled SD.

RESULTS

Diet composition

Composition of wDDGS and the experimental diets is listed in Table 2. The four diets

were identical with an overall average crude protein content of 36.5% of DM, total NSP (26.9%

of DM), HCl-fat (7.9% of DM), and starch (1.6% of DM). The major NSP fraction was

determined as insoluble NSP (21.3% of DM) while the soluble NSP fraction only comprised 5.6%

of DM. The NSP-AX content comprised 15.1% of DM, cellulose (6.0% of DM), and LMW-NDC

(5.1% of DM). On average the Klason lignin content comprised 8.3% of DM across the four

PAPER IV

73

diets. Finally, an arabinose:xylose ratio (A:X) of 0.61 and uronic acid:xylose ratio (UA:X) of

0.11 were observed for the four diets. Content of the remaining NSP constituents is listed in

Table 2. Due to the high inclusion rate of wDDGS, the diet composition was approximately

equal to that of wDDGS.

Effect of enzymes on apparent ileal digestibility of non-starch polysaccharides and macro

nutrients

The CAID of NDC-AX, NSP components and macro nutrients is listed in Table 3.

Compared with the control diet the addition of DAN increased CAID of total NSP by 26% (P =

0.027), followed by EXP with a 23% increase (P = 0.049). Furthermore, the addition of DAN

increased the CAID of NSP-AX by 32% (P < 0.001), EXP by 28% (P = 0.001), and EX by 24% (P

= 0.004), compared with control. In addition, EXP further increased the CAID of NCP-glucose

by 21% compared with control (P = 0.005). No difference was observed in the CAID of NDC-

AX among the treatments ranging from 0.21 to 0.24. (P = 0.574). CAID of the remaining

constituent NSP sugars i.e. galactose, mannose, and uronic acids were not different for the

enzyme treatments compared with control and with overall means varying in the range of

0.16 to 0.19, 0.09 to 0.12, and 0.23 to 0.25, respectively (P > 0.5, data not shown). There was

no observed effect of enzyme addition on CAID of OM, starch, protein (N), or HCl-fat (P > 0.3)

(Table 3). No statistically significant differences were observed between the three enzyme

treatments on neither CAID of NSPs nor macronutrients (P > 0.05).

Effect of enzymes on non-starch polysaccharide content and composition in ileal digesta,

and residual xylanase activity

74

Addition of EX and EXP decreased the concentration of soluble NSP in ileal digesta by 25%

(P = 0.001) and 26% (P < 0.001), respectively, compared with control. All enzyme treatments

decreased the concentration of soluble NSP-AX in ileal digesta compared with control (Table

4). For EX and EXP it was reduced by 40% (P < 0.0001), and DAN reduced it by 21% (P =

0.022). The total concentration of LMW-NDC in ileal digesta increased by 34% and 29% with

addition of DAN (P = 0.005) and EXP (P = 0.015), respectively, compared with control.

Furthermore, the addition of DAN increased the concentration of LMW-AX in ileal digesta by

40% (P = 0.0001), EXP by 36% (P = 0.0006), and EX by 24% (P = 0.023), compared with

control.

Addition of DAN (P < 0.0001) and EXP (P = 0.013) increased the A:X ratio in the

insoluble NSP-AX fraction in ileal digesta compared with control. In addition, a small but

significant increase in UA:X ratio was observed in the ileal insoluble NSP-fraction for DAN

treatment (P < 0.0001), compared with control.

Across all enzyme treatments an increase in the proportion of LMW-AX out of total NDC-

AX in ileal digesta was observed compared with the control (P < 0.003) (Table 4).

Furthermore, EX and EXP (P = 0.0001) and DAN (P = 0.04) treatment decreased the

proportion of soluble NSP-AX significantly compared to control. In addition, the DAN

treatment significantly decreased the proportion of insoluble-AX in ileal digesta compared to

the other three treatments (P < 0.0001).

The residual xylanase activity measured in ileal digesta varied significantly among the

three xylanase treatments, with a 80% recovery of xylanase activity observed for the DAN

treatment (P < 0.001), followed by 22% recovery for the EXP treatment, and 20% recovery for

the EX treatment, indicating large differences in enzyme stability in the gastrointestinal tract

of pigs.

PAPER IV

75

Effect of enzymes on apparent total tract digestibility of organic matter and non-starch

polysaccharides

Enzyme addition had no observed effect on CATTD of OM (0.63 to 0.65), total NSP (0.47

to 0.50) or NSP constituents (P > 0.2). Collectively, mannose (0.84 to 0.86) had the greatest

CATTD of the NSP constituents, followed by NCP-glucose (0.75 to 0.76), xylose (0.53 to 0.55),

galactose (0.46 to 0.48), arabinose (0.43 to 0.45), uronic acids (0.41 to 0.44), and cellulose

(0.28 to 0.30), respectively. Across diets an A:X of 0.80 was observed in collected feces with no

significant differences between dietary treatments (data not shown).

DISCUSSION

In the current study supplementation with exogenous enzymes had great effect on

depolymerizing the AX fraction of the NSP in wDDGS, indicating a high potential of the

xylanases to degrade wDDGS fiber. Across enzyme treatments the increased CAID of NSP-AX

further led to an increase in the concentration of LMW-AX in ileal digesta, indicating that the

exogenously added enzymes had been active under in vivo conditions. The observed higher

CAID of NSP-AX than NDC-AX for all treatments is due the fact that the in vivo generated LMW-

AX is not measured in the NSP procedure, which leads to an overestimation of the ileal

digestibility of AX for the CAID of NSP-AX, as a part remains in the gut as LMW-residues.

Furthermore, because the AID of NDC-AX did not differ significantly between diets, enzyme

addition did not increase microbial fermentation or absorption of AX fragments through the

small intestine.

The use of both GH10 (EX and EXP) and GH11 (DAN) xylanases in this study allowed for

a comparison between enzymes from these two xylanase families and their corresponding

76

differences in wDDGS depolymerizing dynamics in vivo. Xylanases of the GH10 family have

shown to exhibit greater catalytic versatility and lower substrate specificity, whereas GH11

xylanases on the other hand exhibit greater specificity towards unsubstituted regions of AX

(Biely et al., 1997; Fujimoto et al., 2004; Paes et al., 2012; Pell et al., 2004). The ratio of

arabinose – and uronic acid to xylose, A:X and UA:X, is indicative for the average degree of

substitution of the arabinoxylan. The nature and frequency of arabinoxylan substitutions

differ among the different botanical grain fractions in wheat with the A:X ratio ranging from

0.4 to 0.5 in the aleurone layer, 0.8 to 0.9 in the endosperm, and 1.1 to 1.2 in the outer

pericarp (Barron et al., 2007). DAN, which is a GH11 xylanase, increased A:X and UA:X ratio in

ileal digesta, which corresponds well with the higher specificity for breakdown of

unsubstituted regions of the arabinoxylan chain for this particular xylanase, thus, leaving

behind the more substituted AX. When considering the lower substrate specificity of the GH10

xylanase and the relatively high cell wall complexity in wDDGS, the GH10 xylanase (EX and

EXP treatment) could be speculated to be more efficient towards degradation of AX in DDGS

than the GH11 xylanase (DAN treatment). In contrast, DAN induced the highest degradation,

which is likely related to the significantly higher recovery of xylanase activity in ileum for the

DAN treatment than the EX and EXP treatments. Despite the low recovery of xylanase activity

in ileal effluent, the observed positive effects of both the EX and EXP treatment may indicate a

high potential of the GH10 xylanase for DDGS degradation, especially if the in vivo stability can

be improved.

Endo-xylanases cleave the β-1,4-glycosyl linkages within the β-1,4-xylose backbone of

insoluble as well as soluble AX, thus, partially solubilizing insoluble AX and fragmenting

soluble AX and AX solubilized from insoluble AX into LMW-AX (Biely et al., 1997). All enzyme

treatments increased the concentration of LMW-AX in ileal digesta. The observed differences

PAPER IV

77

in the proportions of NDC-AX components in ileal digesta among treatments, however, may

indicate different depolymerization dynamics in vivo between the two different xylanases

used in this study. As the proportion of soluble-AX out of total NDC-AX decreases and the

proportion of insoluble-AX is not affected for the EX and EXP treatment compared to the

control diet, it can be speculated that the EX-xylanase may have a higher specificity for

depolymerizing soluble-AX rather than insoluble-AX into LMW-AX components in vivo. A

possible explanation for this observation is to be found in the higher degree of substitution of

the AX in the soluble- than in the insoluble fraction (Table 2) in relation to the low substrate

specificity of the EX-xylanase. The proportions of insoluble-AX and soluble-AX both decreased

for the DAN treatment compared to the control diet, indicating that the DAN-xylanase may

depolymerize both insoluble- and soluble-AX into LMW-AX components in vivo. However, it

cannot be concluded whether the depolymerization of insoluble-AX by DAN-xylanase directly

generates LMW-AX components or that the generation of LMW-AX components occurs

through a soluble-AX intermediate. The in vivo generated LMW-AX components (i.e.

arabinoxylan-oligosaccharides) are generally acknowledged as being associated with

potential prebiotic effects (Broekaert et al., 2011; Gibson et al., 2004; Van Craeyveld et al.,

2010).

Feeding wDDGS as the only carbohydrate source in the diet (96%, as fed) allowed for a

direct investigation of the enzymes’ capability of degrading DDGS fiber, deprived of NSP

substrates originating from other feed components. Collectively, the DAN xylanase had a

larger effect on CAID of NSPs and induced a higher concentration of LMW-NDCs in ileal

digesta than the two EX xylanase treatments. Conserving enzymatic activity throughout the

upper gastrointestinal tract is crucial to enable the full degradation potential of exogenous

enzymes. Gastrointestinal proteolysis by pepsin and pancreatic juice along with low stomach

78

pH are all factors that potentially may lead to total or partial inactivation of exogenous

enzyme activity (Morgavi et al., 2001). Thus, it can be speculated that the numerically higher

enzymatic degradation observed with the DAN treatment compared with the two EX xylanase

treatments is related to the approximately four times higher residual ileal xylanase activity for

the DAN treatment, consequently leading to an increased NSP hydrolysis.

We have previously described how the most readily degradable AX in corn is degraded

and modified during DDGS processing (Pedersen et al., 2014). Logically, the same should be

true for wDDGS. Generally it can be speculated that the effect of xylanase addition in diets

high in wDDGS (and DDGS in general) may be less effective, as the most readily degradable AX

already is degraded during DDGS production. Thus, leaving behind a more complex and

consequently not as easily accessible AX substrate for the exogenous xylanase; e.g. AX with

ester-linked ferulic acids which may cross-link two or more polysaccharide chains by forming

dimers, 78trimers and maybe even higher oligomers (Bunzel, 2010; Dobberstein and Bunzel,

2010; Jilek and Bunzel, 2013), which have previously been described to impair the enzymatic

degradation potential of the cell wall (Grabber et al., 1998a; Grabber et al., 1998b). As

different xylanases can comprise very different substrate specificities, a xylanase with a high

specificity towards complex AX substrates is required for degradation of the fiber matrix in

DDGS.

Addition of xylanase has been reported to improve both ileal and total tract nutrient (e.g.

amino acids, crude protein, DM, neutral detergent fiber, and energy) digestibility and

performance in swine when applied in wheat based diets (Barrera et al., 2004; Diebold et al.,

2004; Woyengo et al., 2008), and in diets containing wheat milling co-products (Nortey et al.,

2008; Nortey et al., 2007; Yin et al., 2000). However, the results obtained with xylanase

addition are not consistent and several studies showing no effects have also been described

PAPER IV

79

(Rosenfelder et al., 2013). As for xylanase addition in wDDGS diets, the effects are also

inconclusive; Widyaratne et al. (2009) reported an increase in ileal digestibility of energy by

xylanase addition in wheat but not in diets containing 40% wDDGS, Yanez et al. (2011)

reported no effect of xylanase addition in diets containing 44% of co-fermented corn and

wheat DDGS, whereas Emiola et al. (2009) reported an improved ileal nutrient- and total tract

energy digestibility and growth performance by adding a combination of xylanase, β-

glucanase and cellulose enzymes in pigs fed 30% wDDGS. The effects of xylanase addition may

be more or less pronounced depending on the response parameters measured to evaluate

xylanase performance. As illustrated in this study; xylanase addition did not significantly

affect apparent ileal digestibility of macronutrients, whereas on the other hand detailed fiber

analyses (NSP and NDC analyses) of the different xylanase substrates in the feed matrix and

xylanase degradation products (LMW NDCs) revealed a markedly effect of the xylanase

treatments through the small intestine.

Addition of protease in combination with xylanase did not have any additional effects on

nutrient digestibility or production of LMW components, except on the CAID of NCP-glucose

observed for the EXP treatment. As a results of the broad substrate affinity GH10 xylanases

are described to be able to hydrolyze certain β-1,4-glycopyranosyl linkages e.g. present in β-

glucan (Biely et al., 1997), which may explain the increased CAID of NCP-glucose for the EXP

treatment. However, despite the presence of the same xylanase the EX treatment did not

significantly increase CAID of NCP-glucose. Furthermore, the competition for substrate

between endogenous secreted proteolytic enzymes and exogenous protease may in part

explain why no overall additional effects were observed by protease addition. The Klason

lignin content in the wDDGS (8.3% of DM) is relatively high compared to what would be

expected based on the content in wheat (1.9% of DM) (Bach Knudsen, 1997), which may

80

indicate the presence of Maillard products (damaged protein) (Bunzel et al., 2011). Maillard

products is a common concern in relation to DDGS quality, due to the associated negative

effects on protein digestibility (Kim et al., 2012; Pahm et al., 2009), and especially low lysine

digestibility reported in wDDGS diets (Nyachoti et al., 2005; Lan et al., 2008). The presence of

Maillard products may also have limited the potential additional effects of protease addition

on e.g. CAID of CP, and potentially AID of NSPs closely associated with proteins in the cell wall

matrix. Furthermore, we have previously suggested that AX degradation products in DDGS

may be more reactive in the formation of Maillard products during DDGS drying, due to the

larger fraction of reducing ends after hydrolysis (Pedersen et al., 2014).

Despite the observed effect of exogenous enzymes on ileal digestibility of NSPs, no

significant effect was observed for the CATTD of OM or NSPs, presumably because the

depolymerized ileal NSP fragments was not limiting for the microbial fermentation in the

large intestine or quantitatively insufficient. Furthermore, it can be speculated that the

relative short adaption period of five and seven days to each treatment was not sufficient to

stabilize the microbiota to the changes occurring in the small intestine by the exogenous

enzymes, thus, potentially masking the effects of xylanase addition on CATTD.

In conclusion, addition of xylanases to pig diets containing wDDGS can increase the CAID

of NSPs and generate LMW-NDC components with potential prebiotic effects. The dynamics of

DDGS fiber depolymerizing is depended on the substrate affinity of the xylanases. To further

increase the enzymatic degradation of wDDGS, the applied enzymes must have affinity

towards highly complex substrates and preserve enzyme activity throughout the small

intestine of the pig.

PAPER IV

81

LITERATURE CITED

AOAC. 2006. Official methods of analysis. Assoc. Off. Anal. Chem., Arlington, VA.

Avelar, E. et al. 2010. The effect of feeding wheat distillers dried grain with solubles on growth

performance and nutrient digestibility in weaned pigs. Anim. Feed Sci. Technol.160: 73-77.

Bach Knudsen, K. E. 1997. Carbohydrate and lignin contents of plant materials used in animal

feeding. Anim. Feed Sci. Technol.67: 319-338.

Barrera, M., M. Cervantes, W. C. Sauer, A. B. Araiza, and N. Torrentera. 2004. Ileal amino acid

digestibility and performance of growing pigs fed wheat-based diets supplemented with

xylanase. J. Anim. Sci. 82: 1997-2003.

Barron, C., A. Surget, and X. Rouau. 2007. Relative amounts of tissues in mature wheat

(Triticum aestivum L.) grain and their carbohydrate and phenolic acid composition. J.

Cereal Sci. 45: 88-96.

Biely, P., M. Vršanská, M. Tenkanen, and D. Kluepfel. 1997. Endo-β-1,4-xylanase families:

differences in catalytic properties. J. Biotechnol. 57: 151-166.

Broekaert, W. F. et al. 2011. Prebiotic and other health-related effects of cereal-derived

arabinoxylans, arabinoxylan-oligosaccharides, and xylooligosaccharides. Crit. Rev. Food Sci.

Nutr. 51: 178-194.

Bunzel, M. 2010. Chemistry and occurrence of hydroxycinnamate oligomers. Phytochem. Rev.

9: 47-64.

Bunzel, M., A. Schüßler, and G. Tchetseubu Saha. 2011. Chemical characterization of klason

lignin preparations from plant-based foods. J. Agric. Food Chem. 59: 12506-12513.

Diebold, G., R. Mosenthin, H. P. Piepho, and W. C. Sauer. 2004. Effect of supplementation of

xylanase and phospholipase to a wheat-based diet for weanling pigs on nutrient

82

digestibility and concentrations of microbial metabolites in ileal digesta and feces. J. Anim.

Sci. 82: 2647-2656.

Dobberstein, D., and M. Bunzel. 2010. Separation and detection of cell wall-bound ferulic acid

dehydrodimers and dehydrotrimers in cereals and other plant materials by reversed phase

high-performance liquid chromatography with ultraviolet detection. J. Agric. Food Chem.

58: 8927-8935.

Emiola, I. A., F. O. Opapeju, B. A. Slominski, and C. M. Nyachoti. 2009. Growth performance and

nutrient digestibility in pigs fed wheat distillers dried grains with solubles-based diets

supplemented with a multicarbohydrase enzyme. J. Anim. Sci. 87: 2315-2322.

Fujimoto, Z., S. Kaneko, A. Kuno, H. Kobayashi, I. Kusakabe, and H. Mizuno. 2004. Crystal

structures of decorated xylooligosaccharides bound to a family 10 xylanase from

Streptomyces olivaceoviridis E-86. J. Biol. Chem. 279: 9606-9614.

Gibson, G. R., H. M. Probert, J. Van Loo, R. A. Rastall, and M. B. Roberfroid. 2004. Dietary

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

Rev. 17: 259-275.

Grabber, J. H., R. D. Hatfield, and J. Ralph. 1998a. Diferulate cross-links impede the enzymatic

degradation of non-lignified maize walls. J. Sci. Food Agric. 77: 193-200.

Grabber, J. H., J. Ralph, and R. D. Hatfield. 1998b. Ferulate cross-Links Limit the Enzymatic

Degradation of Synthetically Lignified Primary Walls of Maize. J. Agric. Food Chem. 46:

2609-2614.

Hansen, B. 1989. Determination of nitrogen as elementary N, an alternative to kjeldahl. Acta.

Agric. Scandinavica 39: 113-118.

Ishii, T. 1997. Structure and functions of feruloylated polysaccharides. Plant Sci. J. 127: 111-

127.

PAPER IV

83

Izydorczyk, M. S., and C. G. Biliaderis. 1995. Cereal arabinoxylans: advances in structure and

physicochemical properties. Carbohydr. Polym. 28: 33-48.

Jilek, M. L., and M. Bunzel. 2013. Dehydrotriferulic and dehydrodiferulic acid profiles of cereal

and pseudocereal flours. Cereal Chem. J. 90: 507-514.

Jørgensen, H., K. Jakobsen, and B. O. Eggum. 1992. The influence of different protein, fat and

mineral levels on the digestibility of fat and fatty acids measured at the terminal ileum and

in faeces of growing pigs. Acta. Agric. Scandinavica, Sec. A – Anim. Sci. 42: 177-184.

Kim, B. G., D. Y. Kil, Y. Zhang, and H. H. Stein. 2012. Concentrations of analyzed or reactive

lysine, but not crude protein, may predict the concentration of digestible lysine in distillers

dried grains with solubles fed to pigs. J. Anim. Sci. 90: 3798-3808.

Laerke, H. N., M. A. Mortensen, M. S. Hedemann, K. E. Bach Knudsen, J. L. Penalvo, and H.

Adlercreutz. 2009. Quantitative aspects of the metabolism of lignans in pigs fed fibre-

enriched rye and wheat bread. Br. J. Nutr. 102: 985-994.

Lan, Y., F. O. Opapeju, and C. M. Nyachoti. 2008. True ileal protein and amino acid

digestibilities in wheat dried distillers’ grains with solubles fed to finishing pigs. Anim. Feed

Sci. Technol.140: 155-163.

Lyberg, K., J. Borling, and J. E. Lindberg. 2013. Characterization and nutrient evaluation of wet

and dried wheat distillers’ grain for growing pigs. Anim. Feed Sci. Technol.186: 45-52.

Morgavi, D. P., K. A. Beauchemin, V. L. Nsereko, L. M. Rode, T. A. McAllister, A. D. Iwaasa, Y.

Wang, and W. Z. Yang. 2001. Resistance of feed enzymes to proteolytic inactivation by

rumen microorganisms and gastrointestinal proteases. J. Anim. Sci. 79: 1621-1630.

Nortey, T. N., J. F. Patience, J. S. Sands, N. L. Trottier, and R. T. Zijlstra. 2008. Effects of xylanase

supplementation on the apparent digestibility and digestible content of energy, amino

84

acids, phosphorus, and calcium in wheat and wheat by-products from dry milling fed to

grower pigs. J. Anim. Sci. 86: 3450-3464.

Nortey, T. N., J. F. Patience, P. H. Simmins, N. L. Trottier, and R. T. Zijlstra. 2007. Effects of

individual or combined xylanase and phytase supplementation on energy, amino acid, and

phosphorus digestibility and growth performance of grower pigs fed wheat-based diets

containing wheat millrun. J. Anim. Sci. 85: 1432-1443.

Nyachoti, C. M., J. D. House, B. A. Slominski, and I. R. Seddon. 2005. Energy and nutrient

digestibilities in wheat dried distillers’ grains with solubles fed to growing pigs. J. Sci. Food

Agric. 85: 2581-2586.

Paes, G., J.-G. Berrin, and J. Beaugrand. 2012. GH11 xylanases: Structure/function/properties

relationships and applications. Biotechnol. Adv. 30: 564-592.

Pahm, A. A., C. Pedersen, and H. H. Stein. 2009. Standardized ileal digestibility of reactive

lysine in distillers dried grains with solubles fed to growing pigs. J. Agric. Food Chem. 57:

535-539.

Pedersen, M. B., S. Dalsgaard, K. E. B. Knudsen, S. Yu, and H. N. Lærke. 2014. Compositional

profile and variation of Distillers Dried Grains with Solubles from various origins with

focus on non-starch polysaccharides. Anim. Feed Sci. Technol. 197: 130-141

Pell, G., E. J. Taylor, T. M. Gloster, J. P. Turkenburg, C. Fontes, L. M. A. Ferreira, T. Nagy, S. J.

Clark, G. J. Davies, and H. J. Gilbert. 2004. The mechanisms by which family 10 glycoside

hydrolases bind decorated substrates. J. Biol. Chem. 279: 9597-9605.

Rosenfelder, P., M. Eklund, and R. Mosenthin. 2013. Nutritive value of wheat and wheat by-

products in pig nutrition: A review. Anim. Feed Sci. Technol.185: 107-125.

Schürch, A. F., E. W. Crampton, S. R. Haskell, and L. E. Lloyd. 1952. The use of chromic oxide in

digestibility studies with pigs fed ad libitum in the barn. J. Anim. Sci. 11: 261-265.

PAPER IV

85

Schürch, A. F., L. E. Lloyd, and E. W. Crampton. 1950. The use of chromic oxide as an index for

determining the digestibility of a diet: two figures. J. Nutr. 41: 629-636.

Stoldt, W. 1952. Vorschlag zur vereinheitlichung der fettbestimmung in lebensmitteln fette

und seifen/(fette, seifen, anstrichmittel). 54: 206-207.

Theander, O., E. Westerlund, P. Åman, and H. Graham. 1989. Plant cell walls and monogastric

diets. Anim. Feed Sci. Technol. 23: 205-225.

Van Craeyveld, V., E. Dornez, U. Holopainen, E. Selinheimo, K. Poutanen, J. Delcour, and C. M.

Courtin. 2010. Wheat bran ax properties and choice of xylanase affect enzymic production

of wheat bran-derived arabinoxylan-oligosaccharides. Cereal Chem. J. 87: 283-291.

Widyaratne, G. P., J. F. Patience, and R. T. Zijlstra. 2009. Effect of xylanase supplementation of

diets containing wheat distiller’s dried grains with solubles on energy, amino acid and

phosphorus digestibility and growth performance of grower-finisher pigs. Can. J. Anim. Sci.

89: 91-95.

Widyaratne, G. P., and R. T. Zijlstra. 2007. Nutritional value of wheat and corn distiller’s dried

grain with solubles: Digestibility and digestible contents of energy, amino acids and

phosphorus, nutrient excretion and growth performance of grower-finisher pigs. Can. J.

Anim. Sci. 87: 103-114.

Woyengo, T. A., J. S. Sands, W. Guenter, and C. M. Nyachoti. 2008. Nutrient digestibility and

performance responses of growing pigs fed phytase- and xylanase-supplemented wheat-

based diets. J. Anim. Sci. 86: 848-857.

Yanez, J. L., E. Beltranena, M. Cervantes, and R. T. Zijlstra. 2011. Effect of phytase and xylanase

supplementation or particle size on nutrient digestibility of diets containing distillers dried

grains with solubles cofermented from wheat and corn in ileal-cannulated grower pigs. J.

Anim. Sci. 89: 113-123.

86

Yin, Y. L., J. D. G. McEvoy, H. Schulze, U. Hennig, W. B. Souffrant, and K. L. McCracken. 2000.

Apparent digestibility (ileal and overall) of nutrients and endogenous nitrogen losses in

growing pigs fed wheat (var. Soissons) or its by-products without or with xylanase

supplementation. Livest. Prod. Sci. 62: 119-132.

PA

PE

R I

V

87

T

ab

le 1

. In

gred

ien

t co

mp

osi

tio

n o

f b

asal

die

ts (

% a

s-fe

d b

asis

)

Ingr

edie

nt,

%

C

on

tro

l

DA

N

E

X

E

XP

W

hea

t D

DG

S

95

.9

9

5.9

95

.9

9

5.9

G

lyce

rol

1

.5

1

.5

1

.5

1

.5

Ca

lciu

m c

arb

on

ate

1.0

1.0

1.0

1.0

L

-lys

ine-

HC

l

0.4

0.4

0.4

0.4

M

ono

calc

ium

ph

osp

ha

te

0

.2

0

.2

0

.2

0

.2

Ch

rom

ium

(III

) ox

ide

0

.3

0

.3

0

.3

0

.3

Vit

am

in-m

icro

min

era

l pre

mix

1

0

.2

0

.2

0

.2

0

.2

Sod

ium

ch

lori

de

0

.02

0.0

2

0

.02

0.0

2

Co

ntr

ol p

rem

ix2

0.5

-

-

- D

AN

pre

mix

2

-

0

.5

-

-

EX

pre

mix

2

-

-

0

.5

-

EX

P p

rem

ix2

-

-

-

0.5

1P

rovi

ded

th

e fo

llo

win

g q

uan

titi

es o

f m

icro

min

eral

s p

er k

ilo

gram

of

com

ple

te d

iet;

Cu

, 15

mg

as c

op

per

su

lfat

e; F

e, 8

4 m

g as

iro

n s

ulf

ate;

I, 0

.21

mg

as

po

tass

ium

iod

ate;

Mn

, 42

mg

as m

anga

nes

e su

lfat

e; S

e, 0

.30

mg

as s

od

ium

sel

enit

e; a

nd

Zn

, 13

0 m

g as

zin

c o

xid

e.

Pro

vid

ed t

he

foll

ow

ing

qu

anti

ties

of

vita

min

s p

er k

ilo

gram

of

com

ple

te d

iet;

vit

amin

A, 4

,20

0 I

U a

s vi

tam

in a

ceta

te; v

itam

in B

12

, 0.0

21

mg;

vit

amin

D3

,

42

0 I

U;

vita

min

E,

69

IU

as

dl-

α-t

oco

ph

eryl

ace

tate

; vi

tam

in K

3,

2.1

mg

as m

enad

ion

e; b

ioti

n,

0.0

52

5 m

g; d

-pan

toth

enic

aci

d,

10

.5 m

g as

cal

ciu

m

pan

toth

enat

e; n

iaci

n, 2

1 m

g; p

yri

do

xin

e, 3

.15

mg

as p

yrid

oxi

ne

hyd

roch

lori

de;

rib

ofl

avin

, 9.9

mg;

an

d t

hia

min

, 2.1

mg

as t

hia

min

e m

on

on

itra

te.

2P

rem

ix i

ncl

ud

e a

hea

t tr

eate

d w

hea

t as

en

zym

e ca

rrie

r an

d c

on

cen

trat

ed a

mo

un

t o

f p

hyt

ase,

exp

erim

enta

l en

zym

es (

for

EX

, D

AN

, an

d E

XP

), a

nd

colo

red

co

rn g

rits

.

88

T

ab

le 2

. An

alyz

ed n

utr

ien

t co

mp

osi

tio

n o

f w

hea

t D

DG

S an

d e

xper

imen

tal d

iets

(%

of

DM

)

E

xper

imen

tal d

iets

Wh

eat

DD

GS

C

on

tro

l

DA

N

E

X

E

XP

H

Cl-

Fa

t 8

.4

7

.9

7

.9

7

.9

8

.0

Ash

5

.1

6

.4

6

.5

6

.5

6

.6

Sta

rch

1

.3

1

.6

1

.6

1

.6

1

.6

Cru

de

pro

tein

3

7.0

36

.6

3

6.6

36

.4

3

6.3

T

ota

l NSP

1

27

.2 (

6.2

)

27

.3 (

5.8

)

26

.6 (

5.4

)

26

.3 (

5.2

)

27

.5 (

6.1

)

C

ellu

lose

5

.7

5

.9

5

.9

5

.8

6

.3

NC

P

X

ylo

se

9.6

(2

.2)

9

.6 (

2.0

)

9.1

(1

.6)

9

.2 (

1.7

)

9.5

(2

.0)

Ara

bin

ose

5

.9 (

1.6

)

5.8

(1

.4)

5

.7 (

1.3

)

5.6

(1

.3)

5

.8 (

1.3

) G

luco

se

2.3

(1

.0)

2

.3 (

1.1

)

2.3

(1

.1)

2

.2 (

0.9

)

2.2

(1

.4)

Ma

nn

ose

1

.2 (

0.5

)

1.2

(0

.5)

1

.2 (

0.5

)

1.2

(0

.5)

1

.2 (

0.5

) G

ala

cto

se

1.3

(0

.5)

1

.3 (

0.5

)

1.3

(0

.5)

1

.3 (

0.5

)

1.3

(0

.5)

Uro

nic

aci

d

1.1

(0

.3)

1

.0 (

0.3

)

1.0

(0

.3)

1

.0 (

0.3

)

1.1

(0

.4)

LM

W r

esid

ues

2

X

ylo

se

0.7

0.5

0.9

0.9

0.5

A

rab

ino

se

0.8

0.7

0.8

0.9

0.7

G

luco

se

3.3

2.5

2.4

3.0

2.4

M

an

no

se

0.4

0.3

0.4

0.4

0.4

G

ala

cto

se

0.4

0.3

0.4

0.4

0.4

U

ron

ic a

cid

0

.3

0

.3

0

.3

0

.3

0

.3

Kla

son

lig

nin

8

.3

8

.3

8

.5

7

.8

8

.1

A:X

ra

tio

0

.62

(0

.72

)

0.6

1 (

0.7

2)

0

.63

(0

.81

)

0.6

1 (

0.7

5)

0

.61

(0

.65

) U

A:X

ra

tio

0

.11

(0

.14

)

0.1

1 (

0.1

6)

0

.11

(0

.19

)

0.1

1 (

0.1

8)

0

.11

(0

.19

) 1N

on

-sta

rch

po

lysa

cch

arid

es, p

reci

pit

ated

by

80

% e

tOH

in t

he

NSP

pro

ced

ure

2D

eter

min

ed a

s d

iffe

ren

ce b

etw

een

no

n-d

iges

tib

le c

arb

oh

ydra

tes

and

no

n-s

tarc

h p

oly

sacc

har

ides

A:X

, ara

bin

ose

:xyl

ose

; LM

W, l

ow

mo

lecu

lar

wei

ght;

NSP

, no

n-s

tarc

h p

oly

sacc

har

ides

; UA

:X, u

ron

ic a

cid

:xyl

ose

Val

ues

in b

rack

ets

are

solu

ble

NSP

PA

PE

R I

V

89

T

ab

le 3

. C

oef

fici

ents

of

app

aren

t il

eal

dig

esti

bil

ity

of

nu

trie

nts

an

d N

SPs

in g

row

er-f

inis

her

pig

s fe

d w

hea

t D

DG

S d

iets

co

nta

inin

g

exo

gen

ou

s en

zym

es

E

xper

imen

tal d

iets

Co

ntr

ol

D

AN

EX

EX

P

P

oo

led

SD

P-v

alu

e O

rga

nic

ma

tter

0

.49

0.5

0

0

.49

0.4

9

0

.02

9

0

.86

6

Cru

de

pro

tein

0

.63

0.6

4

0

.63

0.6

2

0

.02

3

0

.50

0

HC

l-F

at

0.7

3

0

.73

0.7

3

0

.73

0.0

23

0.8

77

St

arc

h

0.6

9

0

.72

0.7

5

0

.70

0.0

71

0.3

53

N

SP1

0

.25

a

0.3

2b

0.2

9a,

b

0

.31

b

0

.04

1

0

.02

4

Cel

lulo

se

0.1

2

0

.13

0.1

1

0

.12

0.0

61

0.9

10

N

CP

-glu

cose

0

.37

a

0.4

1a,

b

0

.39

a,b

0.4

5b

0.0

39

0.0

07

A

X

0.3

1a

0

.41

b

0

.39

b

0

.40

b

0

.03

7

<

0.0

01

N

DC

-AX

2

0.2

1

0

.24

0.2

4

0

.23

0.0

46

0.5

74

1N

on

-sta

rch

po

lysa

cch

arid

es, p

reci

pit

ated

by

80

% e

tOH

in t

he

NSP

pro

ced

ure

2D

eter

min

ed b

y d

irec

t h

ydro

lysi

s w

ith

ou

t et

OH

pre

cip

itat

ion

AX

, ara

bin

oxy

lan

; NC

P, n

on

-cel

lulo

sic

po

lysa

cch

arid

es; N

DC

, no

n-d

iges

tib

le c

arb

oh

ydra

tes;

NSP

, no

n-s

tarc

h p

oly

sacc

har

ides

; SD

, sta

nd

ard

dev

iati

on

a-

bW

ith

in a

ro

w, m

ean

s w

ith

ou

t a

com

mo

n s

up

ersc

rip

t d

iffe

r (P

< 0

.05

)

90

Table 4. Concentration and fractionation of NDC and residual enzyme activity in ileal digesta

of pigs fed wheat DDGS diets containing different enzyme treatments (% of DM)

Experimental diets Control DAN EX EXP Pooled SD P-value Insoluble NSP1

Total 29.6 27.4 29.4 29.0 1.333 0.099 AX 15.3a 13.7b 14.8a 14.5a,b 0.720 0.002

Soluble NSP1

Total 8.4a 7.3a,b 6.4b 6.2b 0.903 <0.001 AX 4.4a 3.5b 2.6c 2.6c 0.558 <0.001

LMW sugars2

Total 6.4a 8.6b 7.8a,b 8.3b 1.095 0.005 AX 5.0a 7.0b 6.2b 6.8b 0.718 <0.001

Proportion of total NDC-AX3

I-AX 0.62a 0.57b 0.63a 0.61a 0.025 <0.001 S-AX 0.18a 0.15b 0.11c 0.11c 0.022 <0.001 LMW-AX 0.21a 0.29b 0.26b 0.28b 0.026 <0.001

Insoluble A:X ratio1 0.59a 0.64c 0.60a,b 0.61b 0.013 <0.001 Insoluble UA:X ratio1 0.11a 0.12b 0.11a 0.11a 0.003 <0.001 Residual xylanase activity, % 80b 20a 22a 7.0 <0.001 1Non-starch polysaccharides, precipitated by 80% etOH in the NSP procedure 2Determined as difference between non-digestible carbohydrates and non-starch polysaccharides 3Determined by direct hydrolysis without etOH precipitation

A:X, arabinose:xylose; AX, arabinoxylan; I-AX, insoluble AX; LMW, low molecular weight; NDC, non-

digestible carbohydrates; S-AX, soluble AX; SD, standard deviation; UA:X, uronic acid:xylose a-cWithin a row, means without a common superscript differ (P < 0.05)

91

GENERAL DISCUSSION

92

GENERAL DISCUSSION

93

The overall aim of this PhD project was to improve the digestibility of macronutrients in

c- and wDDGS for pigs by enzymatic degradation of the fibre matrix, and to map the variation

and differences especially in NSP composition of these feedstuffs.

The compositional variation between different DDGS sources was identified based on

the content of common constituents and detailed NSP profiles [Paper I], and the differences in

complexity of the fibre matrix was additionally identified by quantification of ester-linked

DFAs and TriFAs [Paper II]. To investigate the enzymatic degradation of c- and wDDGS,

several enzymes were screened alone and in combination, in vitro [Paper III], thus, providing

basis for the preparation of enzyme treatments of wDDGS tested in vivo in growing pigs

[Paper IV].

In addition to the results presented in Paper I-IV, additional in vitro and in vivo

experiments were also performed. However, due to protection of potentially Intellectual

Properties only results obtained from the additional pig study unsupplemented cDDGS diet

will be included for discussion.

This section discusses in general the results presented in the four papers and the

additional pig study.

Variation in DDGS constituents across origins

The general compositional profile and variation between DDGS sources (mostly cDDGS)

have previously been reported and reviewed in the literature, mainly focusing on common

nutrients i.e. DM, crude protein, oil, ash, starch, total carbohydrate, crude fibre, acid- and

neutral detergent fibre [9, 115-117], amino acid composition [115, 116, 118, 119], and

mineral composition [116, 119-121]. Despite the high content of NSP in DDGS, the NSP profile

of DDGS has previously only been described on a limited number of samples [34, 122],

suggesting a need for further analysis on a more broad set of samples.

The variability between different DDGS sources was screened by analyzing 138 samples

from corn, wheat, and mixed cereal origin for common constituents by use of NIRS and

determination of the NSP profile of 63 samples [Paper I], along with the quantification of

ester-linked DFA and TriFA in the same 63 samples [Paper II]. cDDGS comprised the largest

fraction of the collected DDGS samples with a total of 72 samples originating from 21 different

ethanol plants in the US. Despite a comparable sized sample set of 56 samples, the collected

wDDGS samples only originated from two different ethanol plants. The number of mixed

cereal DDGS was the smallest, as the ten collected samples were all from the same ethanol

plant. As corn is by far the most used crop for ethanol production, the differences in amount

and extent of DDGS origins used in these studies may well illustrate the overall amount and

availability of DDGS produced worldwide.

The observed variation, expressed as coefficient of variation (CV), among DDGS samples

was primarily caused by grain origin, and secondly by different ethanol plants, as illustrated

by PCA of grain origins and ethanol plants based on compositional analyses [Paper I].

Collectively for all three DDGS origins, the greatest CVs were observed for the content of

starch, fat, soluble NCP (especially xylose, arabinose, glucose, and mannose), and Klason lignin.

94

The high CV of starch content is directly related to the effectiveness of the fermentation

process at the various ethanol plants. The high CV of fat content is likely caused by some

ethanol plants extracting oil from the thin stillage during DDGS production, whereas the high

CVs of soluble NCPs most likely are caused by a combination of several factors; differences in

the amount of condensed solubles added to the wet distillers cake during drying, and

solubilization of NCPs occurring during DDGS production (e.g. presence of fibre degrading

enzymes from yeast, addition of exogenous fibre degrading enzymes, and mechanical- and

heat treatment). However, the relative low soluble NCP content in the samples will directly

give rise to greater CVs. Furthermore, differences in the amount of yeast in the DDGS might be

the cause of the high CVs of NCP-glucose and mannose, due the content of β-glucan and

mannans present in yeast [122]. The high CVs of Klason lignin content among samples of the

same origin might indicate differences in the drying intensity, as Maillard products caused by

excessive heat during drying might contribute to the Klason lignin fraction [123].

Analyses of 56 cDDGS samples from five different ethanol plants illustrated that each

ethanol plant was capable of consistently producing DDGS with a conserved compositional

profile, as illustrated by PCA models [Paper I]. Variation in sampling period tended to affect

the CVs, as the ethanol plant with the greatest sampling period, 11 months, encompassed the

greatest variation [Paper I, II]. An aim was to relate certain process technologies to the

compositional profile of the DDGS. Of particular interest was to investigate the process

technologies, which we believed would affect the NSP composition in the DDGS the most;

addition of phytase and jet cooking (2×2 factorial). However, it became clear that it was

impossible to fulfill this goal, as not all combinations of technologies actually existed. In

addition, categorizing of the 21 different corn-based ethanol plants according to process

technology also turned out unsuccessful, as not all information on production process was

available or technologies were intertwined.

Compositional variations between corn and wheat DDGS

Large compositional variation was observed between c- and wDDGS; cDDGS contained

more fat than wDDGS, whereas the opposite was the case for protein [Paper I], in line with

previous characterization of parent grains [36]. More interestingly, large differences were

observed in the composition of the fibre matrix between the two DDGS sources. The content

of NSP was similar in both c- and wDDGS, however, the content of insoluble NSP was greater

in cDDGS than wDDGS, and the content of soluble NSP was accordingly higher in wDDGS. The

higher content of arabinose and uronic acid in cDDGS than wDDGS gave rise to higher A:X-

and UA:X ratio, indicating a higher degree of substitution of the arabinoxylan (AX) in cDDGS

[Paper I]. The higher degree of cell wall complexity of c- than wDDGS was further evidenced

by quantification of the ester-linked DFA and TriFA [Paper II], described to induce cross-

linking between polysaccharide chains, thus, strengthening the cell wall with vital

implications for its degradation [22, 27, 28]. cDDGS contained more than five times the

content of ester-linked DFA and TriFA than wDDGS, illustrating a higher degree of cell wall

rigidity in cDDGS. The combined greater A:X- and UA:X-ratio, along with a greater content of

GENERAL DISCUSSION

95

ferulic acids in cDDGS than wDDGS implies a more complex structure of the heteroxylans in

cDDGS, which is in line with previous reports on the greater branching and complexity of corn

AX compared to wheat [28, 38-40].

The use of PCA based on the combination of compositional data from Paper I and –II

indicated a clear correlation between DFAs and TriFAs in the DDGS and the content of

insoluble AX (sum of anhydrous arabinose and xylose), uronic acids, insoluble galactose, and

cellulose. The observed correlation of DFAs to insoluble galactose is in line with previous

studies describing highly complex feruloylated heteroxylan side chains containing up to two

galactose residues from corn bran [18, 39]. These results indicate that both DFA and TriFA in

c- and wDDGS are attached to AX and may be associated with side chains with conserved

profiles.

Modification of corn DDGS during processing

Simultaneous collection of corn and corresponding DDGS samples over 11 month

allowed for investigating the potential changes in composition occurring during DDGS

production. However, it must be noted that although the corn and DDGS samples were

sampled on the same day they did not originate from exactly same fermentation batch.

Initially we aimed to have raw materials and DDGS from the same batch, however, after

having visited an ethanol plant it became clear that this was practically impossible; DDGS is

produced continuously forming enormous heaps and the 48-72 h fermentation time would

make the sampling difficult for the personnel at the plant to manage. In our case the raw

material was collected on the same day as the DDGS was produced (determined as the

warmest DDGS in the heap). Unfortunately we were not able to arrange a similar sampling

strategy for wDDGS or from other corn based ethanol plants. Therefore, the obtained results

may not be considered common for all DDGS origins and ethanol plants.

As expected, the compositional profile of cDDGS in part reflected the composition of

corn with an average concentration of constituents of 3.4 times, due to the depletion of starch

during fermentation. Compared to corn, DDGS had markedly higher solubility of xylose and

arabinose, which is indicative of AX modification during DDGS production. In addition, the A:X

and UA:X ratio was significantly lower in DDGS than corn, most likely caused by modification

(e.g. hydrolysis, and contribution in Maillard reactions) of the more substituted AX from

endosperm during processing. Endosperm AX only comprises approximately 20% of the total

AX content in corn, and is likely more readily degradable as it contains markedly less

feruloylated cross-linkages compared to corn bran AX [24, 124, 125]. DDGS had a higher

cellulose:NSP ratio than corn, indicating that cellulose remains unmodified in DDGS [Paper I].

The relative distribution of the DFAs and TriFAs was conserved in DDGS, indicating that DFAs

and TriFAs are basically not modified during ethanol fermentation and the subsequent drying

of DDGS [Paper II]. The combined results of Paper I and –II suggest a more complex structure

of the remaining AX in DDGS than in the parent grain since the readily degradable endosperm

AX was modified and degraded during processing. These results may have implications of the

96

enzymatic degradability of DDGS and logically point towards the use of enzymes with a higher

specificity towards complex heteroxylan.

Enzymatic degradability of DDGS

Two essentially different approaches were applied to investigate the aim of improving

the digestibility of c- and wDDGS by enzymatic degradation; in vitro [Paper III] and in vivo

[Paper IV] experiments. Collectively, the results illustrated that NSP in wDDGS was more

easily degradable and had higher digestibility than cDDGS, caused by the differences in NSP

complexity between the two substrates.

In vitro degradation of DDGS

Four xylanases from glycosidic hydrolase families (GH) 10 and -11 were screened, in

vitro, for the solubilization of pentosan and protein from c- and wDDGS alone and in

combination with protease and phytase [Paper III]. The use of purified xylanases, deprived of

side activities, allowed for a direct comparison of xylanases originating from the two different

GH families.

The four xylanases showed varying capacity to degrade especially cDDGS; GH10 xylanase

solubilized markedly more pentosan and protein than the three GH11 xylanases, likely due to

the greater specificity of this xylanase towards more highly substituted AX [78, 81]. In

addition, the three GH11 xylanases solubilized different levels of pentosan and protein from

cDDGS, indicative of different substrate specificities between these three xylanases, despite of

the GH relation. The four xylanases showed similar individual ranking and performance on

insoluble cDDGS substrate, whereas no differences in degradation of wDDGS were observed

indicating that the solubilization of pentosan had reached a maximum plateau. On the other

hand, the GH10 and one GH11 xylanase solubilized more pentosan from insoluble wDDGS

than the two remaining GH11 xylanases. The highest solubilization by addition of xylanase

was approximately 15% and 85% of the total pentosan in c- and wDDGS, respectively,

illustrating the higher degradability of w- than cDDGS. The observed increased degradability

of wDDGS than cDDGS are well in line with a previous study describing higher enzymatic

degradability of wheat than corn [126].

As expected, addition of protease increased the solubilization of protein from both c-

and wDDGS, and from insoluble c- and wDDGS fractions. More interestingly, protease addition

alone increased the solubilization of pentosan from cDDGS, indicating a close interaction of

protein and AX in the cell wall, as previously reported [127]. The latter was further confirmed,

as the greatest solubilization of both pentosan and protein was achieved by combining both

xylanase and protease, which indicates synergistic interactions between the two different

hydrolases. It can be speculated that the addition of xylanase would increase the accessibility

of protein for the protease by opening up the cell wall structure through degradation.

Similarly, protease might increase the accessibility of AX for the xylanase by degradation of

cell wall associated proteins, or inactivate endogenous xylanase inhibitors leading to an

increased xylanase activity.

GENERAL DISCUSSION

97

The large difference in the proportion of solubilized pentosan between c- and wDDGS

indicate differences in which botanical grain fractions the pentosan is solubilized from in the

two DDGS sources. In corn, endosperm AX comprises approximately 20% of the total AX

content, whereas endosperm AX in wheat comprises approximately 25% [Paper III]. As the

maximum pentosan solubilization achieved by combining xylanase and protease equals to

approximately 20% of the pentosan in cDDGS, it can be speculated that the majority of

solubilized AX originates from the endosperm fraction. Endosperm AX is most likely the most

readily degradable AX in corn. For wDDGS, the maximum pentosan solubilization achieved by

combining xylanase and protease equals to approximately 85% of the pentosan in wDDGS,

indicating solubilization of AX from other botanical grain fractions besides the endosperm in

wheat.

In vivo digestibility of DDGS

In line with the observed differences in enzymatic degradation of NSP from c- and

wDDGS in vitro [Paper III], differences was also observed in ileal and faecal digestibility of

NSP in c- and wDDGS fed to growing pigs [Paper IV, Supplementary data], Table 1. The

differences in coefficient of apparent ileal digestibility (CAID) of NSP, especially of AX, in

control diets are caused by the higher complexity of cDDGS than wDDGS, as discussed above

[Paper I and –II]. Despite comparable coefficient of apparent total tract digestibility (CATTD)

of NSP between c- and wDDGS, the results indicate that the large intestinal microbiota

ferment cellulose prior to AX in cDDGS, whereas the opposite is the case for wDDGS, Table 1.

These results bring evidence to the large implications on NSP digestibility/degradability

caused by the compositional differences between c- and wDDGS.

To explore the aim of improving the nutrient utilization of wDDGS by use of enzymes, in

vivo, the effect of two different xylanases (GH10 and -11) and GH10 xylanase in combination

with protease were investigated in ileum-cannulated growing pigs fed wDDGS [Paper IV].

Supplementation with exogenous enzymes depolymerized the AX fraction in wDDGS.

Compared to the control diet, enzyme addition increased the AID of NSP-AX, and as a result

increased the concentration of LMW-AX in the ileal digesta. However, since the in vivo

generated LMW-AX is not measured in the NSP procedure, the CAID of NSP-AX overestimates

the ileal digestibility of AX, as a part remains in the gut as LMW-residues. This was evident as

the CAID of NDC-AX (analyses incorporating both NSP and LMW) was lower than CAID of

NSP-AX. Because the AID of NDC-AX did not differ significantly between diets, enzyme

addition did not increase microbial fermentation or absorption of AX fragments in the small

intestine.

98

Table 1. Coefficients of apparent ileal- and total tract digestibility of macronutrients and non-starch

polysaccharides in pigs fed corn- and wheat DDGS unsupplemented control diets

Wheat DDGS, N=8 [Paper IV]

Corn DDGS, N=10

[Supplementary data]1

Apparent ileal digestibility Organic matter 0.49 (±0.03) 0.44 (±0.03) Crude protein 0.63 (±0.03) 0.56 (±0.03) Starch 0.69 (±0.07) 0.90 (±0.01) HCl fat 0.73 (±0.05) 0.77 (±0.04) Non-starch polysaccharides 0.25 (±0.08) 0.05 (±0.08)

Cellulose 0.12 (±0.11) 0.15 (±0.08) Arabinoxylan 0.31 (±0.11) -0.05 (±0.09)

Apparent total tract digestibility Organic matter 0.63 (±0.02) 0.64 (±0.02) Non-starch polysaccharides 0.47 (±0.03) 0.41 (±0.08)

Cellulose 0.28 (±0.04) 0.48 (±0.06) Arabinoxylan 0.49 (±0.03) 0.24 (±0.12)

1Basically carried out as described in Paper IV, regarding feeding, sampling, and analyses. In brief, ten

ileum-cannulated growing pigs were fed five cDDGS diets (96% as fed) according to a double 5×5 Latin

square design. Values in brackets are standard deviation. Data presented are from the unsupplemented

control diet.

Collectively, the GH11 xylanase had a numerically larger effect on CAID of NSPs and

induced a higher concentration of LMW-NDCs in ileal digesta than GH10 xylanase with or

without protease. Conserving enzymatic activity throughout the upper gastrointestinal tract is

crucial to enable the full effectiveness of exogenous enzymes. Gastrointestinal proteolysis by

pepsin and pancreatic juice along with low stomach pH are all factors that potentially may

lead to total or partial inactivation of exogenous enzyme activity [128]. Thus, it can be

speculated that the higher enzymatic degradation observed with the GH11 treatment

compared with the two GH10 xylanase treatments is related to the approximately four times

higher residual ileal xylanase activity for the GH11 treatment, consequently leading to an

increased NSP hydrolysis. Overall, addition of protease in combination with xylanase did not

have any additional effects on nutrient digestibility or production of LMW components.

Differences were observed in the dynamics of AX depolymerization between the two

xylanases; the GH11 xylanase increased A:X and UA:X ratio in ileal digesta, which corresponds

well with the higher specificity for breakdown of unsubstituted regions of the AX chain for

this particular xylanase, thus, leaving behind the more substituted AX. Furthermore, as the

proportion of soluble-AX out of total NDC-AX decreased and the proportion of insoluble-AX

was not affected for the GH10 xylanase treatments compared to the control diet, it can be

speculated that the GH10-xylanase have a higher specificity for depolymerizing soluble-AX

rather than insoluble-AX in vivo. This observation can be partly explained by the higher

degree of AX substitution in the soluble- than in the insoluble fraction in relation to the

specificity of the GH10 xylanase. For the GH11 treatment both the proportion of insoluble-

and soluble-AX decreased compared to the control diet, indicating that the GH11-xylanase

depolymerize both insoluble- and soluble-AX into LMW-AX components in vivo. However, it

GENERAL DISCUSSION

99

cannot be concluded whether this depolymerization of insoluble-AX directly generates LMW-

AX components or that the generation of LMW-AX components occurs through a soluble-AX

intermediate.

Comparison between in vitro and in vivo degradation of DDGS

Similar to the in vitro experiments [Paper III], wDDGS was the only carbohydrate source

in the animal diets [Paper IV]. This allowed for a direct investigation of the enzymes’

capability of degrading DDGS fibre both in vitro and in vivo, deprived of NSP substrates

originating from other feed components. In addition, the wDDGS used both in vitro and in vivo

originated from the same ethanol plant. Consequently the compositional profile of the wDDGS

used in Paper III and -IV was similar.

The experimental conditions between the in vitro and in vivo experiments were

obviously very different, which is likely to have affected the experimental outcome. Compared

to the in vitro conditions (4 h reaction time at 39°C, 1100 rpm, pH 6.0), the in vivo conditions

are far more complex; e.g. low stomach pH and pepsin, pancreatic enzymes, microbial

fermentation in both small- and large intestine, dynamic endogenous secretion, and

removal/absorption of end products. All these factors may affect the efficacy of exogenous

enzymatic addition. Furthermore, the evaluation parameters were also different as the

enzyme performance was evaluated based on solubilization of pentosan and protein in vitro,

and by digestibility of macro nutrients and NSP, and composition of ileal digesta in vivo.

Finally, the enzyme dosage was approximate 20 times higher in vitro than in vivo.

The major difference between the in vitro and in vivo results was observed on the effect

of protease addition. Addition of protease in vitro increased the solubilization of protein and

pentosan, whereas no additional effect of protease addition was observed on ileal or faecal

digestibility in vivo. This observation is likely explained by the presence of endogenous

secreted proteolytic enzymes competing with exogenous protease over substrate in vivo,

whereas no other proteases than exogenously added protease was present in vitro.

The different substrate affinities between GH10 and -11 xylanases were confirmed both

in vitro by the more efficient degradation on the more substituted cDDGS by GH10 than GH11

xylanase, and in vivo by the higher affinity of GH11 xylanase for the less substituted insoluble-

AX. Collectively, the GH10 and -11 xylanases were capable of solubilizing approximately equal

amounts of pentosan in vitro, which corresponds to the approximately equal effect of the two

xylanases on ileal digestibility of AX in vivo. However, the in vivo stability of the two xylanases

was markedly different with four times the residual ileal activity observed for the GH11

xylanase compared to the GH10, which is the likely cause for the observed numerically higher

ileal digestibility of AX and content of LMW-AX in ileal digesta for GH11 treatment.

Solubilization of pentosan [Paper III] and ileal digestibility of AX [Paper IV] are not

directly comparable. Therefore, in order to compare these two measurements, the following

calculation was applied to estimate and align the in vivo- with the in vitro results, under the

assumption that digested AX was considered solubilized;

100

𝑃𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓𝑠𝑜𝑙𝑢𝑏𝑖𝑙𝑖𝑧𝑒𝑑 𝐴𝑋 𝑖𝑛 𝑖𝑙𝑒𝑢𝑚, %

= 100 × 𝑑𝑖𝑔𝑒𝑠𝑡𝑒𝑑 𝑁𝐷𝐶_𝐴𝑋𝑖𝑙𝑒𝑢𝑚,𝑔 + 𝑆_𝐴𝑋𝑖𝑙𝑒𝑢𝑚,𝑔 + 𝐿𝑀𝑊_𝐴𝑋𝑖𝑙𝑒𝑢𝑚,𝑔

𝑁𝐷𝐶_𝐴𝑋𝑑𝑖𝑒𝑡,𝑔 (𝟐)

, calculated per 1000 g of feed.

Similar for both in vitro [Paper III] and in vivo [Paper IV] experiments, addition of

exogenous enzymes was capable of solubilizing and degrading AX. In vitro, addition of

exogenous enzymes was capable of solubilizing ~80-85% of the AX fraction from wDDGS

(~70% for control with no enzymes), whereas the effect of enzyme addition was markedly

lower in vivo (53-57%), Table 2. These observed differences is explained by a combination of

several factors, including but not limited to; particle size, enzyme-to-feed ratio, enzyme

stability in vivo, pH, passage time in vivo, mixing in vivo, and analysis. Especially particle size

can potentially have large implications on solubility and enzymatic degradation; the surface

area increases when the particle size is lowered, thus, increasing the accessibility of the

substrate for the enzymes [91]. In addition, the approximately 20 times higher enzyme-to-

feed ratio in vitro may logically have led to an increase in AX solubility by increased hydrolysis.

Furthermore, an increased mixing (1100 rpm in vitro) will increase the interaction between

enzyme and substrate, which likely will increase the degree of hydrolysis. As soluble pentosan

comprises both LMW-AX and soluble-AX, the proportion of solubilized pentosan in vitro is

overestimated, since this proportion was calculated based on the content of NSP-AX instead of

NDC-AX (both NSP-AX and LMW-AX). Assuming the same level of LMW-AX in the wDDGS used

in vitro as in vivo, the overestimation equals ~8 %-units in vitro. Finally, the assay conditions

under which the AX solubility is determined may be different from actual AX solubility in vivo.

For instance, in the NSP-procedure, soluble AX is determined by boiling the sample in

phosphate buffer for 1 h; conditions that are very different from in vivo conditions. On the

other hand, the measurement of pentosan in the supernatant, in vitro, is a direct measure of

the AX solubility under the applied assay conditions.

Table 2. Calculated proportions of solubilized arabinoxylan in ileum from pigs fed wheat DDGS,

based on data from Paper IV

Dietary treatments Control GH11 GH10 GH10+Protease Proportion of arabinoxylan solubilized, %1 51.2(±9) 57.1(±5) 52.5(±5) 53.4(±8) 1Calculated according to Equation 2. Values in brackets are standard deviation.

Applied methodologies

The application of NIRS for prediction of the commonly measured nutrients provided

acceptable compositional data, illustrated as the compositional profiles of the DDGS reflected

the characteristics of the parent grains. Given that the main focus of this PhD project was on

the NSP fraction, the application of NIRS for compositional prediction of common nutrients

was considered adequate.

Compared to the commonly applied methodologies for analysis of fibre in feedstuff, such

as the neutral- and acid detergent fibre methods [129, 130], the application of the NSP-

GENERAL DISCUSSION

101

procedure [Paper I and -III] allowed for a detailed quantification of the NSP constituents, their

(in)solubility, and the cellulose content. In addition, based on the monomeric sugar ratio (A:X

and UA:X) the NSP-data was indicative of the average structural arrangement of the AX.

Application of the NDC-procedure [Paper IV] provided further information as this

methodology, in addition to the NSP-procedure, quantifies the amount of LMW-residues.

However, the NDC-procedure only quantifies the LMW-residues, thus, not providing

structural information of the individual AXOS. Instead, 1H-13C NMR spectroscopy may be

applied to gain further structural insight, as this methodology recently has provided detailed

structural information of AXOS generated by enzymatic treatment [131].

The applied RP-HPLC/UV methodology for quantification of ester-linked DFAs and

TriFAs provided detailed information of the potential AX cross-linkages. As this methodology

relies on the availability of purified DFAs and TriFAs and synthesized internal standard it is

considered relatively time consuming. A high number of samples may challenge this

methodology as the requirement for especially internal standard increases. In addition, the

range of purified standards (and corresponding correction factors) is limiting for the number

of DFAs and TriFAs available for validated quantification.

The degree of compositional detail obtained from the NSP- and NDC-procedure, and the

RP-HPLC/UV methodology facilitated the interpretation of the result obtained from Paper III

and –IV, both on enzymatic degradation and substrate affinity among different xylanases.

The application of multivariate data analysis, PCA, provided useful information to

visually distinguish between DDGS sources based on their compositional profile [Paper I], and

to correlate compositional data from Paper I and –II, indicative of the structural arrangement

of NSP and DFA/TriFA in the grain.

The methodologies applied for the in vitro screening of enzymatic degradation of DDGS

was chosen primarily due to the high throughput potential [Paper III]. Regarding this matter

the applied methodologies proved successful and elucidated the enzymatic degradation of

both c- and wDDGS, with the limitations inherently associated with a simple model.

Application of e.g. in vitro digestion methodologies [107] may have increased the correlation

between the observed in vitro and in vivo effects of enzyme addition. However, the applied

approach provided further insight into the challenges feed enzymes may face in vivo and the

importance of maintaining activity throughout the stomach and small intestine in pigs.

The in vivo digestion model was successful to investigate the effects of enzyme addition

on the digestibility of NSP and other nutrients [Paper IV]. As the applied enzymes were aimed

for degradation of the NSP matrix, the emphasis was on the analysis of the NSP in ileal digesta,

where the effect of enzyme addition it presumed most prevalent. The spot-sampling of faeces

may be perceived as an additional “bonus”, and it should be noted that the faeces collected on

day 7 may have been influenced by the 8 h ileal collection on day 5. However, to limit this

potential bias, no samples were taken on day 6. This digestion model builds on the use of the

indicator method, which rely on the indigestible marker is not separated from the feed matrix

for measurement of digestibility.

102

103

CONCLUSIONS AND PERSPECTIVES

104

CONCLUSIONS AND PERSPECTIVES

105

Conclusions

This PhD thesis focused on the compositional variability and NSP profile of DDGS from

various origins, and the differences in structural complexity of the NSP. In addition, different

enzymes were investigated both in vitro and in vivo for the degradation of both c- and wDDGS.

The following conclusions were drawn from the studies:

The chemical composition differed between DDGS from corn, wheat and mixed cereal

origin, with each origin comprising individual compositional characteristics.

The NSP fraction in cDDGS is more complex than wDDGS illustrated by the higher

degree of AX substitution, greater insoluble AX fraction, and the markedly higher

content of potential polysaccharide cross-linkages (DFA and TriFA) in cDDGS.

Ethanol plants are capable of producing cDDGS with a conserved individual

compositional profile, likely reflecting the individual production processes.

The most readily degradable AX from endosperm is likely modified during the

production of cDDGS.

The ester-linked DFA and TriFA are not modified during processing, indicating that

the AX in cDDGS may be more inaccessible for e.g. exogenous enzymes compared to

the AX in corn.

The enzymatic degradation, in vitro and in vivo, is markedly higher for wDDGS than

cDDGS, caused by the structural differences of the NSP fraction.

Xylanases of the GH10 family degrade cDDGS more efficiently than GH11, in vitro, due

to differences in substrate affinity.

The differences in substrate affinity between the GH10 and -11 xylanases were

confirmed both in vitro and in vivo.

Both GH10 and -11 xylanase increased the ileal digestibility of AX, and increased the

content of LMW-AX in ileal digesta.

The GH11 xylanase was numerically more efficient in degradation of wDDGS in vivo,

due to the superior stability through the stomach and small intestine and illustrated

by the four times higher xylanase activity recovered in ileum.

In summary, the compositional variations and differences in AX complexity between c-

and wDDGS had large implications on the efficiency of enzymatic degradation. In addition, in

vitro and in vivo experiments illustrated that enzymes have different efficiency of degradation

in relation to their substrate affinity and stability towards the physical environment occurring

in the gastro intestinal tract of pigs.

106

Perspectives

The results of the work conducted in this PhD thesis indicate that enzymes are capable

of degrading NSP in DDGS, and that their substrate affinity and in vivo stability have great

implications of the degradation efficiency. However, further animal studies need to be

performed to investigate these effects on growth performance, microbiota etc.

Corn DDGS comprises the highest potential for improvement as this substrate was

highly resistant to enzymatic degradation compared with wDDGS. The markedly higher

content of polysaccharide cross-linkages in cDDGS is the most likely cause limiting the

degradation. However, further studies need to be conducted in order to investigate this

hypothesis e.g. characterization of residual fractions recalcitrant to hydrolysis.

Identification of the compositional constituents in DDGS most limiting for the enzymatic

hydrolysis is essential in order to further improve the effect of enzymatic addition, as

enzymes targeting these factors may be applied.

The effects of enzyme addition in combination with other technologies such as acid or

alkali treatments may also lead to an increased degradation.

107

REFERENCES

108

REFERENCES

109

1. Avelar, E., R. Jha, E. Beltranena, M. Cervantes, A. Morales, and R.T. Zijlstra, The effect of feeding wheat distillers dried grain with solubles on growth performance and nutrient digestibility in weaned pigs. Anim. Feed Sci. Technol., 2010. 160(1–2): p. 73-77.

2. Lyberg, K., J. Borling, and J.E. Lindberg, Characterization and nutrient evaluation of wet and dried wheat distillers’ grain for growing pigs. Anim. Feed Sci. Technol., 2013. 186(1–2): p. 45-52.

3. Liu, K., Chemical composition of distillers grains, a review. J. Agric. Food Chem., 2011. 59(5): p. 1508-26.

4. Bothast, R.J. and M.A. Schlicher, Biotechnological processes for conversion of corn into ethanol. Appl. Microbiol. Biotechnol., 2005. 67(1): p. 19-25.

5. Stein, H.H. and G.C. Shurson, Board-invited review: the use and application of distillers dried grains with solubles in swine diets. J. Anim. Sci., 2009. 87(4): p. 1292-303.

6. Olukosi, O.A. and A.O. Adebiyi, Chemical composition and prediction of amino acid content of maize- and wheat-Distillers’ Dried Grains with Soluble. Anim. Feed Sci. Technol., 2013. 185(3): p. 182-189.

7. Cozannet, P., M. Lessire, C. Gady, J.P. Métayer, Y. Primot, F. Skiba, and J. Noblet, Energy value of wheat dried distillers grains with solubles in roosters, broilers, layers, and turkeys. Poult. Sci., 2010. 89(10): p. 2230-2241.

8. Fastinger, N.D. and D.C. Mahan, Determination of the ileal amino acid and energy digestibilities of corn distillers dried grains with solubles using grower-finisher pigs. J. Anim. Sci., 2006. 84(7): p. 1722-1728.

9. Belyea, R.L., K.D. Rausch, and M.E. Tumbleson, Composition of corn and distillers dried grains with solubles from dry grind ethanol processing. Bioresour. Technol., 2004. 94(3): p. 293-8.

10. Whitney, M.H., G.C. Shurson, L.J. Johnston, D.M. Wulf, and B.C. Shanks, Growth performance and carcass characteristics of grower-finisher pigs fed high-quality corn distillers dried grain with solubles originating from a modern Midwestern ethanol plant. J. Anim. Sci., 2006. 84(12): p. 3356-3363.

11. Woyengo, T.A., E. Beltranena, and R.T. Zijlstra, NONRUMINANT NUTRITION SYMPOSIUM: Controlling feed cost by including alternative ingredients into pig diets: A review. J. Anim. Sci., 2014. 92(4): p. 1293-1305.

12. Izydorczyk, M.S. and C.G. Biliaderis, Cereal arabinoxylans: advances in structure and physicochemical properties. Carbohydr. Polym., 1995. 28(1): p. 33-48.

13. Carpita, N.C. and D.M. Gibeaut, Structural models of primary cell walls in flowering plants: consistency of molecular structure with the physical properties of the walls during growth. Plant. J., 1993. 3(1): p. 1-30.

14. Surget, A. and C. Barron, Histologie du grain blé (histology of the wheat grain). Ind. Cér., 2005. 145: p. 3-7.

15. Britannica, E. corn: layers and structures of corn kernel. 2014 [cited 2014 22 November]; Available from: http://www.britannica.com/EBchecked/media/162/The-outer-layers-and-internal-structures-of-a-kernel-of.

16. Saulnier, L., P.-E. Sado, G. Branlard, G. Charmet, and F. Guillon, Wheat arabinoxylans: Exploiting variation in amount and composition to develop enhanced varieties. J. Cereal Sci., 2007. 46(3): p. 261-281.

17. Knudsen, K.E., Fiber and nonstarch polysaccharide content and variation in common crops used in broiler diets. Poult. Sci., 2014. 93(9): p. 2380-93.

110

18. Allerdings, E., J. Ralph, H. Steinhart, and M. Bunzel, Isolation and structural identification of complex feruloylated heteroxylan side-chains from maize bran. Phytochemistry, 2006. 67(12): p. 1276-1286.

19. Ralph, J., J.H. Grabber, and R.D. Hatfield, Lignin-ferulate cross-links in grasses: active incorporation of ferulate polysaccharide esters into ryegrass lignins. Carbohydr. Res., 1995. 275(1): p. 167-178.

20. Bunzel, M., J. Ralph, F. Lu, R.D. Hatfield, and H. Steinhart, Lignins and ferulate-coniferyl alcohol cross-coupling products in cereal grains. J. Agric. Food Chem., 2004. 52(21): p. 6496-6502.

21. Piber, M. and P. Koehler, Identification of dehydro-ferulic acid-tyrosine in rye and wheat: evidence for a covalent cross-link between arabinoxylans and proteins. J. Agric. Food Chem., 2005. 53(13): p. 5276-84.

22. Bunzel, M., Chemistry and occurrence of hydroxycinnamate oligomers. Phytochemistry Rev., 2010. 9(1): p. 47-64.

23. Ishii, T., Isolation and characterization of a diferuloyl arabinoxylan hexasaccharide from bamboo shoot cell-walls. Carbohydr. Res., 1991. 219: p. 15-22.

24. Saulnier, L. and J.-F. Thibault, Ferulic acid and diferulic acids as components of sugar-beet pectins and maize bran heteroxylans. J. Sci. Food Agric., 1999. 79(3): p. 396-402.

25. Allerdings, E., J. Ralph, P.F. Schatz, D. Gniechwitz, H. Steinhart, and M. Bunzel, Isolation and structural identification of diarabinosyl 8-O-4-dehydrodiferulate from maize bran insoluble fibre. Phytochemistry, 2005. 66(1): p. 113-124.

26. Bunzel, M., E. Allerdings, J. Ralph, and H. Steinhart, Cross-linking of arabinoxylans via 8-8-coupled diferulates as demonstrated by isolation and identification of diarabinosyl 8-8(cyclic)-dehydrodiferulate from maize bran. J. Cereal Sci., 2008. 47(1): p. 29-40.

27. Dobberstein, D. and M. Bunzel, Separation and Detection of Cell Wall-Bound Ferulic Acid Dehydrodimers and Dehydrotrimers in Cereals and Other Plant Materials by Reversed Phase High-Performance Liquid Chromatography With Ultraviolet Detection. J. Agric. Food Chem., 2010. 58(16): p. 8927-8935.

28. Jilek, M.L. and M. Bunzel, Dehydrotriferulic and Dehydrodiferulic Acid Profiles of Cereal and Pseudocereal Flours. Cereal Chem., 2013. 90(5): p. 507-514.

29. Mutwil, M., S. Debolt, and S. Persson, Cellulose synthesis: a complex complex. Curr. Opin. Plant Biol., 2008. 11(3): p. 252-7.

30. Festucci-Buselli, R.A., W.C. Otoni, and C.P. Joshi, Structure, organization, and functions of cellulose synthase complexes in higher plants. Bra. J. Plant Physiol., 2007. 19: p. 1-13.

31. Selvendran, R.R., The plant cell wall as a source of dietary fiber: chemistry and structure. Am. J. Clin. Nutr., 1984. 39(2): p. 320-37.

32. Grimm, A., E. Krüger, and W. Burchard, Solution properties of β-D-(1, 3)(1, 4)-glucan isolated from beer. Carbohydr. Polym., 1995. 27(3): p. 205-214.

33. Johansen, H.N., K.E. Bach Knudsen, P.J. Wood, and R.G. Fulcher, Physico-Chemical Properties and the Degradation of Oat Bran Polysaccharides in the Gut of Pigs. J. Sci. Food Agric., 1997. 73(1): p. 81-92.

34. Widyaratne, G.P. and R.T. Zijlstra, Nutritional value of wheat and corn distiller’s dried grain with solubles: Digestibility and digestible contents of energy, amino acids and phosphorus, nutrient excretion and growth performance of grower-finisher pigs. Can. J. Anim. Sci., 2007. 87(1): p. 103-114.

REFERENCES

111

35. Gibreel, A., J.R. Sandercock, J. Lan, L.A. Goonewardene, A.C. Scott, R.T. Zijlstra, J.M. Curtis, and D.C. Bressler, Evaluation of value-added components of dried distiller's grain with solubles from triticale and wheat. Bioresour. Technol., 2011. 102(13): p. 6920-7.

36. Belitz, H.D., W. Grosch, and P. Schieberle, Food chemistry. 4th revised and extended edition. Food chemistry. 4th revised and extended edition.2009: Springer, Berlin. 670-675.

37. Bach Knudsen, K.E., Carbohydrate and lignin contents of plant materials used in animal feeding. Anim. Feed Sci. Technol., 1997. 67(4): p. 319-338.

38. Bedford, M.R., Mechanism of action and potential environmental benefits from the use of feed enzymes. Anim. Feed Sci. Technol., 1995. 53(2): p. 145-155.

39. Saulnier, L., J. Vigouroux, and J.-F. Thibault, Isolation and partial characterization of feruloylated oligosaccharides from maize bran. Carbohydr. Res., 1995. 272(2): p. 241-253.

40. Yang, J., M.X. Maldonado-Gómez, R.W. Hutkins, and D.J. Rose, Production and in Vitro Fermentation of Soluble, Non-digestible, Feruloylated Oligo- and Polysaccharides from Maize and Wheat Brans. J. Agric. Food Chem., 2013.

41. Renger, A. and H. Steinhart, Ferulic acid dehydrodimers as structural elements in cereal dietary fibre. Eur. Food Res. Technol., 2000. 211(6): p. 422-428.

42. Bunzel, M., J. Ralph, J.M. Marita, R.D. Hatfield, and H. Steinhart, Diferulates as structural components in soluble and insoluble cereal dietary fibre. J. Sci. Food Agric., 2001. 81(7): p. 653-660.

43. Bach Knudsen, K.E., M.S. Hedemann, and H.N. Lærke, The role of carbohydrates in intestinal health of pigs. Anim. Feed Sci. Technol., 2012. 173(1–2): p. 41-53.

44. Varel, V.H. and J.T. Yen, Microbial perspective on fiber utilization by swine. J. Anim. Sci., 1997. 75(10): p. 2715-22.

45. Jensen, B.B. and H. Jørgensen, Effect of dietary fiber on microbial activity and microbial gas production in various regions of the gastrointestinal tract of pigs. Appl. Environ. Microbiol., 1994. 60(6): p. 1897-1904.

46. Bach Knudsen, K.E. and N. Canibe, Breakdown of plant carbohydrates in the digestive tract of pigs fed on wheat- or oat-based rolls. J. Sci. Food Agric., 2000. 80(8): p. 1253-1261.

47. Bach Knudsen, K.E. and H. Jørgensen, Impact of wheat and oat polysaccharides provided as rolls on the digestion and absorption processes in the small intestine of pigs. J. Sci. Food Agric., 2007. 87(13): p. 2399-2408.

48. Bach Knudsen, K.E., B.B. Jensen, J.O. Andersen, and I. Hansen, Gastrointestinal implications in pigs of wheat and oat fractions. 2. Microbial activity in the gastrointestinal tract. Br. J. Nutr., 1991. 65(2): p. 233-48.

49. Dierick, N.A., I.J. Vervaeke, D.I. Demeyer, and J.A. Decuypere, Approach to the energetic importance of fibre digestion in pigs. I. Importance of fermentation in the overall energy supply. Anim. Feed Sci. Technol., 1989. 23(1–3): p. 141-167.

50. Anguita, M., N. Canibe, J.F. Pérez, and B.B. Jensen, Influence of the amount of dietary fiber on the available energy from hindgut fermentation in growing pigs: Use of cannulated pigs and in vitro fermentation. J. Anim. Sci., 2006. 84(10): p. 2766-2778.

51. Just, A., J. Fernández, and H. Jørgensen, The net energy value of diets for growth in pigs in relation to the fermentative processes in the digestive tract and the site of absorption of the nutrients. Livest. Prod. Sci., 1983. 10(2): p. 171-186.

112

52. Giusi-Perier, A., M. Fiszlewicz, and A. Rerat, Influence of diet composition on intestinal volatile fatty acid and nutrient absorption in unanesthetized pigs. J. Anim. Sci., 1989. 67(2): p. 386-402.

53. Noblet, J., H. Fortune, X.S. Shi, and S. Dubois, Prediction of net energy value of feeds for growing pigs. J. Anim. Sci., 1994. 72(2): p. 344-54.

54. Knudsen, K.E., B.B. Jensen, and I. Hansen, Digestion of polysaccharides and other major components in the small and large intestine of pigs fed on diets consisting of oat fractions rich in beta-D-glucan. Br. J. Nutr., 1993. 70(2): p. 537-56.

55. Bach Knudsen, K.E. and I. Hansen, Gastrointestinal implications in pigs of wheat and oat fractions. 1. Digestibility and bulking properties of polysaccharides and other major constituents. Br. J. Nutr., 1991. 65(2): p. 217-32.

56. Glitso, L.V., G. Brunsgaard, S. Hojsgaard, B. Sandstrom, and K.E. Bach Knudsen, Intestinal degradation in pigs of rye dietary fibre with different structural characteristics. Br. J. Nutr., 1998. 80(5): p. 457-68.

57. Ivarsson, E., R. Andersson, and J.E. Lindberg, Molecular weight distribution of soluble fiber fractions and short chain fatty acids in ileal digesta of growing pigs. J. Anim. Sci., 2012. 90(Supplement 4): p. 65-67.

58. Le Goff, G. and J. Noblet, Comparative total tract digestibility of dietary energy and nutrients in growing pigs and adult sows. J. Anim. Sci., 2001. 79(9): p. 2418-27.

59. Le Goff, G., J. van Milgen, and J. Noblet, Influence of dietary fibre on digestive utilization and rate of passage in growing pigs, finishing pigs and adult sows. J. Anim. Sci., 2002. 74: p. 503-515.

60. Le Gall, M., A. Serena, H. Jørgensen, P.K. Theil, and K.E. Bach Knudsen, The role of whole-wheat grain and wheat and rye ingredients on the digestion and fermentation processes in the gut – a model experiment with pigs. Br. J. Nutr., 2009. 102(11): p. 1590-1600.

61. Schulze, H., P. van Leeuwen, M.W. Verstegen, J. Huisman, W.B. Souffrant, and F. Ahrens, Effect of level of dietary neutral detergent fiber on ileal apparent digestibility and ileal nitrogen losses in pigs. J. Anim. Sci., 1994. 72(9): p. 2362-8.

62. Schulze, H., P. van Leeuwen, M.W. Verstegen, and J.W. van den Berg, Dietary level and source of neutral detergent fiber and ileal endogenous nitrogen flow in pigs. J. Anim. Sci., 1995. 73(2): p. 441-8.

63. Mariscal-Landín, G., B. Sève, Y. Colléaux, and Y. Lebreton, Endogenous Amino Nitrogen Collected from Pigs with End-to-End Ileorectal Anastomosis is Affected by the Method of Estimation and Altered by Dietary Fiber. J. Nutr., 1995. 125(1): p. 136-146.

64. Leterme, P., A. Théwis, P.v. Leeuwen, T. Monmart, and J. Huisman, Chemical Composition of Pea Fibre Isolates and their Effect on the Endogenous Amino Acid Flow at the Ileum of the Pig. J. Sci. Food Agric., 1996. 72(1): p. 127-134.

65. Stanogias, G. and G.R. Pearce, The digestion of fibre by pigs. 3. Effects of the amount and type of fibre on physical characteristics of segments of the gastrointestinal tract. Br. J. Nutr., 1985. 53(3): p. 537-548.

66. Jin, L., L.P. Reynolds, D.A. Redmer, J.S. Caton, and J.D. Crenshaw, Effects of dietary fiber on intestinal growth, cell proliferation, and morphology in growing pigs. J. Anim. Sci., 1994. 72(9): p. 2270-8.

67. Wenk, C., The role of dietary fibre in the digestive physiology of the pig. Anim. Feed Sci. Technol., 2001. 90(1–2): p. 21-33.

68. Kerr, B.J. and G.C. Shurson, Strategies to improve fiber utilization in swine. J. Anim. Sci. Biotechnol, 2013. 4(1): p. 11.

REFERENCES

113

69. Bedford, M.R., Exogenous enzymes in monogastric nutrition — their current value and future benefits. Anim. Feed Sci. Technol., 2000. 86(1–2): p. 1-13.

70. Zijlstra, R.T., A. Owusu-Asiedu, and P.H. Simmins, Future of NSP-degrading enzymes to improve nutrient utilization of co-products and gut health in pigs. Livest. Sci., 2010. 134(1–3): p. 255-257.

71. Hovel, K., D. Shallom, K. Niefind, V. Belakhov, G. Shoham, T. Baasov, Y. Shoham, and D. Schomburg, Crystal structure and snapshots along the reaction pathway of a family 51 alpha-L-arabinofuranosidase. EMBO J., 2003. 22(19): p. 4922-32.

72. Subramaniyan, S. and P. Prema, Biotechnology of Microbial Xylanases: Enzymology, Molecular Biology, and Application. Crit. Rev. Biotechnol., 2002. 22(1): p. 33-64.

73. Prade, R.A., Xylanases: from biology to biotechnology. Biotechnol. Genet. Eng. Rev., 1996. 13: p. 101-31.

74. Sunna, A. and G. Antranikian, Xylanolytic enzymes from fungi and bacteria. Crit. Rev. Biotechnol., 1997. 17(1): p. 39-67.

75. Kulkarni, N., A. Shendye, and M. Rao, Molecular and biotechnological aspects of xylanases. FEMS Microbiol. Rev., 1999. 23(4): p. 411-56.

76. Beg, Q.K., M. Kapoor, L. Mahajan, and G.S. Hoondal, Microbial xylanases and their industrial applications: a review. Appl. Microbiol. Biotechnol., 2001. 56(3-4): p. 326-38.

77. Cantarel, B.L., P.M. Coutinho, C. Rancurel, T. Bernard, V. Lombard, and B. Henrissat, The Carbohydrate-Active EnZymes database (CAZy): an expert resource for Glycogenomics. Nucleic. Acids Res., 2009. 37(suppl 1): p. D233-D238.

78. Paes, G., J.-G. Berrin, and J. Beaugrand, GH11 xylanases: Structure/function/properties relationships and applications. Biotechnol. Adv., 2012. 30(3): p. 564-92.

79. Collins, T., C. Gerday, and G. Feller, Xylanases, xylanase families and extremophilic xylanases. FEMS Microbiol. Rev., 2005. 29(1): p. 3-23.

80. Lagaert, S., A. Pollet, C.M. Courtin, and G. Volckaert, β-Xylosidases and α-l-arabinofuranosidases: Accessory enzymes for arabinoxylan degradation. Biotechnol. Adv., 2014. 32(2): p. 316-332.

81. Biely, P., M. Vršanská, M. Tenkanen, and D. Kluepfel, Endo-β-1,4-xylanase families: differences in catalytic properties. J. Biotechnol., 1997. 57(1–3): p. 151-166.

82. Beaugrand, J., G. Chambat, V.W. Wong, F. Goubet, C. Remond, G. Paes, S. Benamrouche, P. Debeire, M. O'Donohue, and B. Chabbert, Impact and efficiency of GH10 and GH11 thermostable endoxylanases on wheat bran and alkali-extractable arabinoxylans. Carbohydr. Res., 2004. 339(15): p. 2529-40.

83. Broekaert, W.F., C.M. Courtin, K. Verbeke, T. Van de Wiele, W. Verstraete, and J.A. Delcour, Prebiotic and Other Health-Related Effects of Cereal-Derived Arabinoxylans, Arabinoxylan-Oligosaccharides, and Xylooligosaccharides. Crit. Rev. Food Sci. Nutr., 2011. 51(2): p. 178-194.

84. Gibson, G.R., H.M. Probert, J. Van Loo, R.A. Rastall, and M.B. Roberfroid, Dietary modulation of the human colonic microbiota: updating the concept of prebiotics. Nutr. Res. Rev., 2004. 17(2): p. 259-275.

85. Van Craeyveld, V., E. Dornez, U. Holopainen, E. Selinheimo, K. Poutanen, J.A. Delcour, and C.M. Courtin, Wheat Bran AX Properties and Choice of Xylanase Affect Enzymic Production of Wheat Bran-Derived Arabinoxylan-Oligosaccharides. Cereal Chem. J., 2010. 87(4): p. 283-291.

86. Fierens, E., S. Rombouts, K. Gebruers, H. Goesaert, K. Brijs, J. Beaugrand, G. Volckaert, S. Van Campenhout, P. Proost, Christophe M. Courtin, and Jan A. Delcour, TLXI, a novel

114

type of xylanase inhibitor from wheat (Triticum aestivum) belonging to the thaumatin family. Biochem. J., 2007. 403(Pt 3): p. 583-591.

87. Gebruers, K., K. Brijs, C.M. Courtin, K. Fierens, H. Goesaert, A. Rabijns, G. Raedschelders, J. Robben, S. Sansen, J.F. Sørensen, S. Van Campenhout, and J.A. Delcour, Properties of TAXI-type endoxylanase inhibitors. Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics, 2004. 1696(2): p. 213-221.

88. Juge, N., F. Payan, and G. Williamson, XIP-I, a xylanase inhibitor protein from wheat: a novel protein function. Biochim. Biophys. Acta., 2004. 1696(2): p. 203-11.

89. Bedford, M.R. and A.J. Cowieson, Exogenous enzymes and their effects on intestinal microbiology. Anim. Feed Sci. Technol., 2012. 173(1–2): p. 76-85.

90. Chesson, A., Feed enzymes. Anim. Feed Sci. Technol., 1993. 45(1): p. 65-79. 91. de Vries, S., A.M. Pustjens, H.A. Schols, W.H. Hendriks, and W.J.J. Gerrits, Improving

digestive utilization of fiber-rich feedstuffs in pigs and poultry by processing and enzyme technologies: A review. Anim. Feed Sci. Technol., 2012. 178(3–4): p. 123-138.

92. Widyaratne, G.P., J.F. Patience, and R.T. Zijlstra, Effect of xylanase supplementation of diets containing wheat distiller's dried grains with solubles on energy, amino acid and phosphorus digestibility and growth performance of grower-finisher pigs. Canadian J. Anim. Sci., 2009. 89(1): p. 91-95.

93. Yanez, J.L., E. Beltranena, M. Cervantes, and R.T. Zijlstra, Effect of phytase and xylanase supplementation or particle size on nutrient digestibility of diets containing distillers dried grains with solubles cofermented from wheat and corn in ileal-cannulated grower pigs. J. Anim. Sci., 2011. 89(1): p. 113-123.

94. Jones, C.K., J.R. Bergstrom, M.D. Tokach, J.M. DeRouchey, R.D. Goodband, J.L. Nelssen, and S.S. Dritz, Efficacy of commercial enzymes in diets containing various concentrations and sources of dried distillers grains with solubles for nursery pigs. J. Anim. Sci., 2010. 88(6): p. 2084-91.

95. Emiola, I.A., F.O. Opapeju, B.A. Slominski, and C.M. Nyachoti, Growth performance and nutrient digestibility in pigs fed wheat distillers dried grains with solubles-based diets supplemented with a multicarbohydrase enzyme. J. Anim. Sci., 2009. 87(7): p. 2315-2322.

96. Hansen, B., Determination of Nitrogen as Elementary N, an Alternative to Kjeldahl. Acta. Agric. Scand., 1989. 39(2): p. 113-118.

97. Stoldt, W., Vorschlag zur Vereinheitlichung der Fettbestimmung in Lebensmitteln Fette und Seifen/(Fette, Seifen, Anstrichmittel), 1952. 54: p. 206-207.

98. AOAC, Official methods of analysis2006: Assoc. Off. Anal. Chem., Arlington, VA. 99. Givens, D.I., J.L. De Boever, and E.R. Deaville, The principles, practices and some future

applications of near infrared spectroscopy for predicting the nutritive value of foods for animals and humans. Nutr. Res. Rev., 1997. 10(01): p. 83-114.

100. Bakalli, R.I., G.M. Pesti, and R.D. Etheridge, Comparison of a Commercial Near-Infrared Reflectance Spectroscope and Standard Chemical Assay Procedures for Analyzing Feed Ingredients: Influence of Grinding Methods. J. Appl. Poult. Res., 2000. 9(2): p. 204-213.

101. Sohn, M., D.S. Himmelsbach, F.E. Barton, II, C.A. Griffey, W. Brooks, and K.B. Hicks, Near-infrared analysis of ground barley for use as a feedstock for fuel ethanol production. Appl. Spec., 2007. 61(11): p. 1178-1183.

102. Pekel, A.Y., E.O. Çakır, M. Polat, K. Çakır, G. İnan, and N. Kocabağlı, Correlations between chemical assays and near-infrared reflectance spectroscopy for nutrient components and

REFERENCES

115

correlations between nutrients and color scores of distillers dried grains with solubles. J. Appl. Poult. Res., 2013. 22(4): p. 814-824.

103. Englyst, H., H.S. Wiggins, and J.H. Cummings, Determination of the non-starch polysaccharides in plant foods by gas-liquid chromatography of constituent sugars as alditol acetates. Analyst, 1982. 107(1272): p. 307-318.

104. Theander, O. and P. Åman, Studies on dietary fibres. 1. Analysis and chemical characterization of water-soluble and water-insoluble dietary fibres. Swe. J. Agric. Res., 1979. 107(1272): p. 307-318.

105. Theander, O. and E.A. Westerlund, Studies on dietary fiber. 3. Improved procedures for analysis of dietary fiber. J. Agric. Food Chem., 1986. 34(2): p. 330-336.

106. Barberousse, H., O. Roiseux, C. Robert, M. Paquot, C. Deroanne, and C. Blecker, Analytical methodologies for quantification of ferulic acid and its oligomers. J. Sci. Food Agric., 2008. 88(9): p. 1494-1511.

107. Boisen, S. and J.A. Fernández, Prediction of the total tract digestibility of energy in feedstuffs and pig diets by in vitro analyses. Anim. Feed Sci. Technol., 1997. 68(3–4): p. 277-286.

108. Scholten, R.H.J., C.M.C. van der Peet-Schwering, M.W.A. Verstegen, L.A. den Hartog, J.W. Schrama, and P.C. Vesseur, Fermented co-products and fermented compound diets for pigs: a review. Anim. Feed Sci. Technol., 1999. 82(1–2): p. 1-19.

109. Rouau, X. and A. Surget, A rapid semi-automated method for the determination of total and water-extractable pentosans in wheat flours. Carbohydr. Polym., 1994. 24(2): p. 123-132.

110. Scientific, T. BCA Protein Assay Reagent (bicinchoninic acid). 2014 [cited 2014 17 November]; Available from: http://www.piercenet.com/product/bca-protein-assay.

111. Bach Knudsen, K.E., H.N. Lærke, S. Steenfeldt, M.S. Hedemann, and H. Jørgensen, In vivo methods to study the digestion of starch in pigs and poultry. Anim. Feed Sci. Technol., 2006. 130(1–2): p. 114-135.

112. Zijlstra, R.T., R. Jha, A.D. Woodward, J. Fouhse, and T.A.T.G. van Kempen, Starch and fiber properties affect their kinetics of digestion and thereby digestive physiology in pigs. J. Anim. Sci., 2012. 90(Supplement 4): p. 49-58.

113. Schürch, A.F., E.W. Crampton, S.R. Haskell, and L.E. Lloyd, The use of Chromic Oxide in Digestibility Studies with Pigs Fed Ad Libitum in the Barn. J. Anim. Sci., 1952. 11(2): p. 261-265.

114. Shewry, P.R., M.J. Hawkesford, V. Piironen, A.-M. Lampi, K. Gebruers, D. Boros, A.A.M. Andersson, P. Åman, M. Rakszegi, Z. Bedo, and J.L. Ward, Natural Variation in Grain Composition of Wheat and Related Cereals. J. Agric. Food Chem., 2013. 61(35): p. 8295-8303.

115. Cromwell, G.L., K.L. Herkelman, and T.S. Stahly, Physical, chemical, and nutritional characteristics of distillers dried grains with solubles for chicks and pigs. J. Anim. Sci., 1993. 71(3): p. 679-86.

116. Spiehs, M.J., M.H. Whitney, and G.C. Shurson, Nutrient database for distiller's dried grains with solubles produced from new ethanol plants in Minnesota and South Dakota. J. Anim. Sci., 2002. 80(10): p. 2639-45.

117. Liu, K., Particle size distribution of distillers dried grains with solubles (DDGS) and relationships to compositional and color properties. Bioresour. Technol., 2008. 99(17): p. 8421-8.

116

118. Batal, A.B. and N.M. Dale, True Metabolizable Energy and Amino Acid Digestibility of Distillers Dried Grains with Solubles. J. Appl. Poult. Res., 2006. 15(1): p. 89-93.

119. Han, J. and K. Liu, Changes in Composition and Amino Acid Profile during Dry Grind Ethanol Processing from Corn and Estimation of Yeast Contribution toward DDGS Proteins. J. Agric. Food Chem., 2010. 58(6): p. 3430-3437.

120. Belyea, R., T. Clevenger, V. Singh, M.E. Tumbleson, and K. Rausch, Element concentrations of dry-grind corn-processing streams. Appl. Biochem. Biotechnol., 2006. 134(2): p. 113-128.

121. Batal, A. and N. Dale, Mineral composition of distillers dried grains with solubles. J. Appl. Poult. Res., 2003. 12(4): p. 400-403.

122. de Vries, S., A.M. Pustjens, M.A. Kabel, S. Salazar-Villanea, W.H. Hendriks, and W.J.J. Gerrits, Processing Technologies and Cell Wall Degrading Enzymes To Improve Nutritional Value of Dried Distillers Grain with Solubles for Animal Feed: an in Vitro Digestion Study. J. Agric. Food Chem., 2013. 61(37): p. 8821-8828.

123. Bunzel, M., A. Schüßler, and G. Tchetseubu Saha, Chemical Characterization of Klason Lignin Preparations from Plant-Based Foods. J. Agric. Food Chem., 2011. 59(23): p. 12506-12513.

124. Bunzel, M., J. Ralph, P. Brüning, and H. Steinhart, Structural Identification of Dehydrotriferulic and Dehydrotetraferulic Acids Isolated from Insoluble Maize Bran Fiber. J. Agric. Food Chem., 2006. 54(17): p. 6409-6418.

125. Gallardo, C., L. Jiménez, and M.T. García-Conesa, Hydroxycinnamic acid composition and in vitro antioxidant activity of selected grain fractions. Food Chem., 2006. 99(3): p. 455-463.

126. Rose, D. and G. Inglett, A Method for the Determination of Soluble Arabinoxylan Released from Insoluble Substrates by Xylanases. Food Anal. Meth., 2011. 4(1): p. 66-72.

127. Faulds, C.B., G. Mandalari, R.B. Lo Curto, G. Bisignano, P. Christakopoulos, and K.W. Waldron, Synergy between xylanases from glycoside hydrolase family 10 and family 11 and a feruloyl esterase in the release of phenolic acids from cereal arabinoxylan. Appl. Microbiol. Biotechnol., 2006. 71(5): p. 622-9.

128. Morgavi, D.P., K.A. Beauchemin, V.L. Nsereko, L.M. Rode, T.A. McAllister, A.D. Iwaasa, Y. Wang, and W.Z. Yang, Resistance of feed enzymes to proteolytic inactivation by rumen microorganisms and gastrointestinal proteases. J. Anim. Sci., 2001. 79(6): p. 1621-1630.

129. Van Soest, P.J., Use of detergents in the analysis of fibrous feeds. I. Preparation of fibre residues of low nitrogen content. J. AOAC, 1963. 46: p. 825-829.

130. Van Soest, P.J. and R.H. Wine, Use of detergents in the analysis of fibrous feeds. IV. Determination of plant cell-wall constituents. J. AOAC, 1967. 50: p. 50-55.

131. Petersen, B.O., F. Lok, S. Meier, Probing the structural details of xylan degradation by real-time NMR spectroscopy. Carbohydr. Polym., 2014. 112: p. 587-594.