J Sci Food Agric 1998, 77, 327È333

Gastrointestinal or Simulated In Vitro DigestionChanges Dietary Fibre Properties and theirFermentationChristine Hoebler,* Fabienne Guillon, Anthony Fardet, Christine Cherbutand Jean-Luc Barry

Institut National de la Recherche Agronomique, Rue de la Ge� raudiere, BP 71627, 44316 Nantes Cedex 03,France

(Received 27 February 1997 ; revised version received 21 July 1997 ; accepted 22 October 1997)

Abstract : This study evaluated the e†ect of digestion on the chemical and physi-cochemical characteristics of dietary Ðbre and on its behaviour during fermenta-tion. Three dietary Ðbre sources (wheat bran, barley bran and beet Ðbre) wererecovered from ileal cannulated pigs after in vivo digestion and prepared by invitro enzymatic treatment simulating digestion. Raw substrates and Ðbre residueswere analysed for their chemical and physicochemical properties as well as theirpotential fermentation by human colonic bacteria. In vitro and in vivo treatmentsled to insoluble residues, enriched in cell wall polysaccharides, with similar cellwall sugar composition and physicochemical properties. Degradations of cellwall polysaccharides with losses of sugar residues occurred mainly after in vivodigestion, especially for pectins from beet Ðbre and b-glucans from barley bran.Solubilisation of b-glucans removed highly fermentable substrates for further fer-mentation. For beet Ðbre, removal of pectins led to increased hydration proper-ties and faster fermentation of cell-wall polysaccharides. Enzymatic treatmentsimulated correctly the passage of Ðbre through the digestive tract, modifying thecell-wall matrix and predisposing the Ðbre to further fermentation. 1998 SCI.(

J Sci Food Agric 77, 327È333 (1998)

Key words : dietary Ðbre ; cell-wall polysaccharides ; particle size ; hydrationproperties ; digestion ; fermentation

INTRODUCTION

The physiological e†ects of dietary Ðbre depend largelyon its fermentation in the colon (Cherbut et al 1991 ;Cummings and Macfarlane 1991 ; Scheppach 1994).Many factors can a†ect the extent of Ðbre fermentationand the production of short-chain fatty acids (SCFA),some of which relate to the host (Michel and MacFar-lane 1996) and colonic Ñora (Barry et al 1995), whileothers depend on Ðbre properties. Fermentation ofdietary Ðbre has been widely studied using in vitro batchsystems, which seem to provide accurate evaluation ofthe fermentability of a large number of substrates andallow the role of their physicochemical properties to bedetermined (Mortensen et al 1988 ; Bourquin et al 1992 ;

* To whom correspondence should be addressed.

Au†ret et al 1993 ; Salvador et al 1993 ; Barry et al1995). These systems are based on incubation of Ðbreswith human faecal inoculum in anaerobic conditions(Barry et al 1989). However, before their arrival in thehindgut, Ðbres are subjected to the peculiar luminalenvironment (enzymes, pH, temperature, water activity,minerals, bile acids, etc) of the upper digestive tract.These conditions can lead to irreversible changes in theintrinsic characteristics of dietary Ðbre (chemical com-position, particle size) or reversible modiÐcations ofphysicochemical properties. During intestinal digestion,starch and proteins are hydrolysed and removed fromraw materials. The acidic or neutral conditions prevail-ing in the digestive tract can partially solubilise pecticsubstances (arabinose, galactose, uronic acid) Millardand Chesson 1984 ; Graham et al 1986 ; McBurney et al1988). Non-starch glucose losses are also observed, pre-

3271998 SCI. J Sci Food Agric 0022È5142/98/$17.50. Printed in Great Britain(

328 C Hoebler et al

sumably attributable to bacteria colonising the terminalileum (Millard and Chesson 1984 ; Graham et al 1986,Jensen and Jorgensen 1994). As all these changes in thechemical composition and physicochemical propertiesof dietary Ðbre are likely to a†ect the rate and extent ofits fermentation in the colon (Guillon et al 1992, 1995 ;Au†ret et al 1993), they need to be taken into accountby in vitro fermentation tests. Thus, studies on thebehaviour of Ðbre during fermentation require an invitro Ðbre puriÐcation procedure that can simulate thechemical and physicochemical modiÐcations of Ðbreduring its transit through the upper gut.

The purpose of this study was to compare the charac-teristics of raw Ðbre substrates and the correspondingdietary Ðbre obtained after digestion in vivo (transitthrough the upper digestive tract of pigs) or by in vitroenzymatic treatment. Three characteristics (chemicalcomposition, physicochemical properties and fermenta-tive properties) were studied relative to raw substratesand residues obtained after in vivo and in vitro digestion.The three substrates used (beet Ðbre, wheat bran andbarley bran) di†ered in their chemical, physicochemicaland fermentative properties.

EXPERIMENTAL

Materials

The Ðbre sources used in this study were of commercialorigin. Coarse wheat bran (WB) was purchased from LaCana, (Ancenis, France), barley bran (BB) was donatedby Whestove (Arques, France), and sugar beet Ðbre (BF)was obtained from Agro-industrie Recherche et De� vel-oppement (Paris, France). The chemical properties ofraw substrates and insoluble fractions recovered after invitro preparation and at terminal ileum of pigs are givenin Table 1.

In vivo digestion

Three pigs (40È50 kg) were Ðtted with a permanent T-shaped cannula inserted into the ileum anterior to theterminal ileo-caecal junction. Animals were housed indi-vidually and fed twice daily (7 :30 am and 4 pm) with a

diet composed of 500 g of dry matter mixed with 1É5litres of water before feeding. The basal diet was semi-synthetic (drum-dried wheat starch 720 g kg~1, casein177 g kg~1, maize oil 50 g kg~1 and minerals53 g kg~1). Animals were adapted to an experimentaldiet for 3 days before digesta samples were collected.The experimental diet was supplemented with Ðbre(60 g kg~1 of dietary Ðbre on a dry matter basis) byreducing the basal diet by an equivalent amount ofstarch (660 g kg~1). On the day of collection of digesta(day 4), animals received a meal containing 120 g kg~1DM. Each pig was tested successively on the threeÐbres, in random order, allowing a 3-day period eachtime for adaptation to the new Ðbre.

On the day of collection, pigs were strapped in astanding position to avoid disturbing the Ñow of digestathrough the cannula. They were then fed a meal supple-mented with 120 g Ðbre kg~1, refusals were removedand weighed. Plastic bags attached to each cannulawere changed every hour from 8:30 am to 10 :30 pm andweighed. pH was immediately recorded and materialscollected every hour were separately stored at] 4¡C.Between the 5th and 8th hour of collection, wasHgCl2added to fraction of digesta for subsequent analysis ofSCFA. Digestive contents collected over a 14 h periodwere pooled and stored at ]4¡C before treatment. Afraction of digesta was centrifuged (10 min at5000 rpm), and the pellets were immediately washed bytwo centrifugations with 100 ml of water. They werekept overnight at ]4¡C and then extensively washedwith water on a screen (80 lm), weighed, freeze-driedand stored for further analysis.

In vitro enzymatic treatment

Fibre sources (10 g) were treated with commercial enzy-matic mixtures (Alcalase 0É6L, Termamyl 120L, NovoNordisk A/S, Denmark) to remove non-Ðbre constitu-ents, as previously reported (Hoebler et al 1991). Afterenzymatic incubations, Ðbre materials were fractionatedby Ðltration through an 80-lm screen. The residueswere washed twice with 500 ml of distilled water,weighed and freeze-dried. In vitro enzymatic treatmentswere done in triplicate for each Ðbre material.

TABLE 1Chemical composition (g kg~1 DM) of the raw Ðbre materials (RF), insoluble fractions recovered after in vitroenzymatic preparation (EF) or at terminal ileum (IF) of pigs fed diets containing dietary Ðbre (mean^ SD, n \ 3)

W heat bran Barley bran Beet Ðbre

RF EF IF RF EF IF RF EF IF

Crude protein 126 109 ^ 2 57 ^ 1 115 112 ^ 2 67 ^ 5 39 44 ^ 7 15 ^ 1Starch 125 15 ^ 9 28 314 21^ 8 52 NDa ND NDAshes 70 38^ 26 47 ^ 5 56 25 ^ 1 65 ^ 6 35 68 ^ 39 52 ^ 3

a ND, not determined.

Digestion and dietary Ðbre properties 329

In vitro fermentation procedure

For the fermentation experiment, the three residuesrecovered from in vitro enzymatic studies were mixed inequal parts. The residues collected at the terminal ileumof three pigs were also pooled. The raw materials andthe in vivo and in vitro residues were fermented by ahuman faecal inoculum according to the procedure ofBarry et al (1989), as modiÐed by Guillon et al (1995).All samples were run twice in the same fermentationexperiment. Fermentation residues were analysed foracidic and neutral cell wall sugars.

Chemical analyses

Chemical analyses were performed on raw materialsand residues of Ðbre recovered after in vitro enzymatictreatment (EF) or from digesta of the three pigs (IF).

All results were calculated on a dry matter basis.Ashes was determined by weighing after 5 h at 550¡C.Crude protein (N ] 5É7 for WB, N] 6É0 for BB andN ] 6É25 for BF) was determined by the Kjeldahlmethod. Starch was estimated by the method of Kar-kalas (1985). SCFA concentrations were quantiÐed bygasÈliquid chromatography (GLC) (Jouany 1982).Neutral cell wall sugars were transformed into alditolacetates for GLC analysis (Hoebler et al 1989). Uronicacid residues were determined by an automated colori-metric method (Thibault 1979), and mixed-linked b-glucans were analysed by the enzymatic method ofMcCleary and Codd (1991).

Physicochemical characterisations were determinedon raw materials and the residues obtained separatelyby three in vitro or in vivo treatments of digestion andthen mixed together. Particle size distribution wasanalysed on dried Ðbre residues (5 g), as described byGuillon et al (1995). Swelling capacity was measured bythe bed volume technique (Kuniak and Marchessault1972), as adapted by Guillon et al (1995). Water reten-tion capacity (WRC) was measured by a dialysismethod using a known suction pressure, according toRobertson and Eastwood (1981) as modiÐed by Guillonet al (1995). Tests were run in six replicates.

Calculations and statistical analyses

The amount of individual cell wall sugars was expressedas anhydrosugars. Cell-wall glucose was estimated bydeducting starch content from the total glucose content.Cellulose was calculated as the di†erence between cell-wall glucose and glucose measured as b-glucan. Therecovery of individual cell wall sugars in Ðbre residuesobtained after in vitro enzymatic treatment was calcu-lated as :

S(EF)] EFS(RF)] RF

] 100

where S is the individual sugar and and areS(EF) S(RF)the sugar concentrations, respectively, in Ðbre residue(EF) from in vitro enzymatic treatment and raw Ðbrematerial (RF). The terms RF and EF refer, respectively,to the amount of Ðbre dry matter used before andrecovered after in vitro enzymatic digestion.

The recovery of sugars found after in vivo digestionwas calculated using xylose as a marker (Graham et al1986) ; and was respectively the xyloseXyl(RF) Xyl(IF)concentrations in the raw material and in Ðbre residueof ileal digesta.

S(IF)] Xyl(RF)S(RF)] Xyl(IF)

] 100

The values are expressed as means^ SEM. Statisticaltests were performed on fermentative variables byone-way analysis of variance (ANOVA). When e†ectswere signiÐcant, means were compared by StudentÏst-test. Homogeneity of variance was assumed. All calcu-lations were run using StatViewTM SE software (AbacusConcepts Inc, Calabasas, CA, USA) on a Power Macin-tosh computer.

RESULTS

E†ects of in vivo and in vitro enzymatic preparation onchemical and physicochemical characteristics of Ðbreresidues

Cereal brans and BF di†ered in their chemical composi-tion (Table 1). Cereal brans contained 20È40% of pro-teins and starch, and cell-wall sugars accounted for lessthan half of the dry matter (90% of total sugars consist-ing of xylose, glucose and arabinose) (Table 2). In BB,glucose was derived from both cellulose and b-glucan(22% of total glucose). Beet was low in starch andprotein, containing mainly cell-wall polysaccharides(68% of dry matter), with arabinose, glucose and uronicacid as major sugars (accounting for 88% of totalsugars).

The starch present in the raw cereal substrates (125and 314 g kg~1 DM, respectively) was mainly elimi-nated during in vitro and in vivo digestion (from 15 to50 g kg~1 DM). The removal of proteins by in vitroenzymatic treatment was not very efficient. Residualprotein amounts were lower in Ðbre residues recoveredfrom pig digesta (around 50 g kg~1 DM) than after invitro treatment (up to 112 g kg~1 DM). Both in vivoand in vitro enzymatic removal of starch and proteinsproduced residues containing similar high concentra-tions of cell-wall polysaccharides (about 650 g kg~1 forWB, 600 g kg~1 for BB and 750 g kg~1 for BF) (Table2). The rate of xylose recovery in residues accounted for95È100%, conÐrming the low xylose loss during guttransit and justifying its use as marker of sugarsrecovery (Graham et al 1986). For cereal brans, the

330 C Hoebler et al

TABLE 2Cell wall sugar composition of the raw Ðbres (RF), the insoluble fractions recovered after in vitro enzymatic preparation (EF)

or at terminal ileum (IF) of pigs fed diets supplemented with dietary Ðbre and % recovery of cell wall sugars (mean^ SD)

W heat bran Barley bran Beet Ðbre

RF EF IF RF EF IF RF EF IF

Cell wall sugar compositiona (g kg~1 of dry matter)Arabinose 87 129 ^ 23 158 ^ 15 51 126 ^ 4 133 ^ 8 188 199 ^ 11 233 ^ 9Xylose 162 242 ^ 42 279 ^ 9 95 237 ^ 8 258 ^ 2 12 14 ^ 1 16 ^ 1Uronic acid 15 15^ 3 23 ^ 2 15 21 ^ 2 15 ^ 2 181 150 ^ 24 195 ^ 27Glucose 123 234^ 23 172 ^ 11 120 189 ^ 8 138 ^ 5 231 285 ^ 16 251 ^ 14b-glucan NDb ND ND 34 8 ^ 0 22 ^ 8 ND ND ND

Total sugars 410 654 ^ 84 652 ^ 36 302 604 ^ 4 590 ^ 13 682 728 ^ 22 776 ^ 38

% recovery of cell wall sugars (calculated with xylose as marker for IF)Arabinose 94 ^ 10 103 ^ 7 97 ^ 8 92 ^ 6 79 ^ 12 85 ^ 7Xylose 96 ^ 10 103 ^ 8 95 ^ 14Uronic acid 68 ^ 6 81 ^ 9Total glucose 123 ^ 7 80 ^ 3 70 ^ 15 48 ^ 2 111 ^ 19 88 ^ 1

Total sugars 90 ^ 6 80 ^ 2 61 ^ 10 64 ^ 1 79 ^ 19 77 ^ 4

a Expressed in anhydrosugar.b ND, not determined.

arabinose/xylose ratio was not modiÐed in residuesrecovered after both treatments, whereas the glucosecontent of BB residues dropped after both in vitro andin vivo digestion. The EF residue was very low in b-glucan, suggesting that selective loss (77%) of this poly-saccharide could occur by solubilisation or chemicaldegradation during in vitro enzymatic treatment. Therelative proportion of BF sugars was di†erent in EFand IF. The pH and the temperature conditions of thein vitro enzymatic procedures involved losses of 10 and30%, respectively, of arabinose and uronic acid residuesin BF. In ileal digesta (IF), concentrations of pecticsugars (arabinose, uronic acid) were higher than in EF,whereas cellulosic glucose concentration was lower inIF than EF. The recovery of glucose, as calculated bythe xylose marker, indicated that about 12% of cellulosefrom BF was lost during passage through the digestivetract. Apparently, pectin content was partly degraded orsolubilised in the digestive tract (25% of galactose anduronic acid and 13% of arabinose). Actually, amountsof SCFA were measured in digesta (24È74 mM litre~1 ofdigesta) recovered at the terminal pig ileum, indicatingthat bacterial contamination may induce fermentationin the ileum and cause cell-wall polysaccharide degrada-tion.

Swelling and water retention capacities were aboutthree times as high for raw BF than for cereal bran(Table 3). Transit through the digestive tract and invitro enzymatic treatment produced a signiÐcantincrease in the swelling capacity and water retentioncapacities of Ðbres, mainly for BF residues (Table 3). Nochanges in the particle size distribution of cereal bransand BF were observed after in vivo or in vitro digestion.

Fermentative behaviour

The fermentation patterns of native Ðbre proceeded dif-ferently from that of digestion residues, being morerapid for cereal bran and slower for BF (Table 4). Lossof total sugars was greater for native WB at 12 h offermentation than for the two digestion residues (EF,IF), due to extensive glucose degradation. Contrary tothe situation for cereal brans, the rate of fermentationfor the total sugars of native BF was signiÐcantly lowerthan that of digestion residues. At 6 h of fermentation,uronic acid was the lowest fermented sugar of raw BF.Removal of the soluble BF fraction signiÐcantlyincreased the fermentability of pectin sugars after 6 and12 h, as compared to that of raw BF. Cellulose degrada-tion was similar for native BF and residues (EF, EI).

The two digestive treatments (in vivo vs in vitro) led tosimilar rates of Ðnal fermentation (24 h) of the residues,regardless of the substrate considered (Table 4).However, at 12 h of WB fermentation, residues from invitro digestion were more slowly degraded than after invivo digestion because of the poor fermentation of cellu-lose and the more progressive degradation of xylose andarabinose polymers. The fermentation pattern for xyloseand arabinose di†ered in digestion residues of WB andBB. When the digestion residues (EF, IF) of BF werecompared, the fermentability of total cell wall sugarswas not signiÐcantly di†erent. The loss of arabinose anduronic acid was slightly higher for IF than EF, prob-ably because of its higher concentration of these sugars.For both digestion residues, pectic substances(arabinose, uronic acid) were more fermented than cellu-losic materials and almost totally degraded.

Digestion and dietary Ðbre properties 331

TABLE 3Physicochemical properties of the raw Ðbres (RF), the insoluble residues recovered after in vitro enzymatic preparation (EF) or at

terminal ileum (IF) of pigs fed diets containing dietary Ðbre

W heat bran Barley bran Beet Ðbre

RF EF IF RF EF IF RF EF IF

Swelling capacitya (ml g~1) 7É6 ^ 0É1 9É4 ^ 0É5 10É7 ^ 0É1 5É4 ^ 0É7 9É2 ^ 0É6 11É1 ^ 0É3 18É4 ^ 0É4 26É8 ^ 0É8 34É7 ^ 1É0WRCb (g g~1) 3É1 ^ 0É3 3É8 ^ 0É3 3É5 ^ 0É4 2É3 ^ 0É4 2É9 ^ 0É5 2É5 ^ 0É6 4É6 ^ 0É4 5É8 ^ 0É5 6É2 ^ 0É6Geometric mean 523^ 245 577 ^ 246 593^ 256 521^ 168 603^ 179 655^ 191 510^ 222 573 ^ 248 637^ 233particle sizec (lm)

a The results were the mean of 5 determinations ^ SD.b The results were the mean of 3 determinations ^ SD.c The results were the mean of 2 determinations ^ geometric SD.

DISCUSSION

Beet Ðbre and cereal bran were chosen as Ðbre sourcesfor this study because they are representative of the twokinds of dietary Ðbre found in the human diet. BeetÐbre is rich in pectic substances and cellulose andsimilar to the Ðbre found in fruit and vegetables. It isfree of starch and proteins and thus provides the idealreference substrate to study the inÑuence of in vivo or invitro treatment on the fermentation of beet cell-wallpolysaccharides. In order to compare the fermentationpattern of dietary Ðbre before and after digestive treat-

ments, we focused our study on insoluble Ðbrematerials, which are considered to be the main limitingfraction in Ðbre fermentation. It was previouslyobserved that the fermentation of isolated poly-saccharides is not representative of intact cell-wall fer-mentation (Bourquin et al 1992). The fermentativepattern of dietary Ðbre is more variable in insolubleÐbre as polysaccharides are integrated into a complexcell-wall structure which controls the physicochemicalproperties of Ðbre (macroporosity, particle size) andlimits their accessibility to hydrolytic enzymes and bac-teria (Guillon et al 1992). In our experiment, the

TABLE 4Degradation of wheat bran, barley bran and beet Ðbre cell wall polysaccharides after 6, 12, 24 of fermentationa of raw Ðbre (RF),

insoluble Ðbre fractions recovered after in vitro enzymatic treatment (EF) or in vivo digestion (IF)

Substrates T ime of T otal sugars Arabinose Xyloseincubation

(h) RF EF IF RF EF IF RF EF IF

Wheat bran 6 20 ^ 4 3 ^ 5 15 ^ 2 13^ 2 16^ 4 16 ^ 2 1^ 6 2^ 4 6 ^ 212 32 ^ 1* 12 ^ 0** 28 ^ 1*** 23^ 2* 15^ 1** 21 ^ 0* 21^ 2* 16^ 2** 36 ^ 2**24 48 ^ 0 42 ^ 6 44 ^ 1 29^ 1* 20^ 3** 25 ^ 1 62^ 1 47^ 11 62 ^ 0

Barley bran 6 25 ^ 1 13 ^ 4 5 ^ 7 25^ 0 21^ 3 10 ^ 7 8^ 13 19^ 2 ([2)^ 612 36 ^ 9 34 ^ 0 32 ^ 2 36^ 8 42^ 1 41 ^ 3 27^ 0 35^ 3 33 ^ 224 56 ^ 3 42 ^ 5 43 ^ 2 59^ 1 49^ 4 53 ^ 2 43^ 2 49^ 2 46 ^ 2

Beet Ðbre 6 20^ 3* 54 ^ 1** 64 ^ 7** 27^ 4* 69^ 2** 85 ^ 3***12 62 ^ 1* 86 ^ 1** 79 ^ 6** 68^ 2* 91^ 0** 94 ^ 2**24 82 ^ 2* 96 ^ 4** 92 ^ 2 86^ 2* 98^ 2** 98 ^ 1**

Substrates T ime of Glucose Uronic acidincubation

(h) RF EF IF RF EF IF

Wheat bran 6 38 ^ 1* ([8)^ 8** 20 ^ 1*12 42 ^ 1* 1 ^ 4** 29 ^ 0**24 41 ^ 1* 15 ^ 0** 35 ^ 1***

Barley bran 6 25 ^ 3 4 ^ 4 6 ^ 712 39 ^ 6 26 ^ 1 22 ^ 324 55 ^ 3* 33 ^ 5** 29 ^ 3**

Beet Ðbre 6 25^ 2 42 ^ 2 47 ^ 14 5^ 4* 56^ 2** 66 ^ 4**12 63 ^ 1 80 ^ 3 64 ^ 9 55^ 2* 88^ 0** 84 ^ 5**24 80 ^ 3 93 ^ 7 89 ^ 5 80^ 2* 99^ 1** 93 ^ ***

a % apparent disappearance ^ SEM (two sets of fermentation, each time in duplicate).b For each cell wall sugar and each time of incubation means within the same row with a di†erent number of asterisks di†er by ANOVA (P\ 0É05).

332 C Hoebler et al

recovery of Ðbre at the end of the ileum of cannulatedpigs was a means of obtaining insoluble Ðbres subjectedto the true digestive conditions prevailing in the gastro-intestinal tract. We chose to collect insoluble Ðbreduring a Ðxed period (14 h) to obtain materials repre-sentative of Ðbre passing through the gastrointestinaltract and to limit the heterogeneity of Ðbre from di†er-ent transits in the digestive tract. The pooling of digestafor a 14 h period was considered to approximate asample representative of 24 h collection. This minimisedthe errors resulting from circadian variations in digestacomposition (Livingstone et al 1980). The insoluble Ðbreresidues obtained after both digestion treatmentsshowed comparable chemical characteristics, withsimilar polysaccharide content (a coefficient of variationof less than 23% for each monosaccharide, Table 2).Therefore, the di†erent residue preparations were mixedand considered as the mean representing the residuescollected from the three pigs and the three in vitroexperiments.

The in vitro treatment was intended to simulateprotein and starch hydrolysis in the digestive tract. Thefraction of nitrogen retained in the in vitro residue wasprobably due to insufficient proteolysis in substrateswith large particle size (Brillouet et al 1988). Yet theamount of these residual proteins was low and probablydid not modify the production of SCFA (IC4 and IC5)released during protein fermentation (MacFarlane et al1992). During the fermentation of raw cereal Ðbre, alarge proportion of SCFA arose from starch fermenta-tion, whereas end-product fermentation of cereal resi-dues resulted only from degradation of cell-wallpolysaccharides and residual proteins (Bourquin et al1992). Accordingly, the loss of cell-wall sugars was theonly parameter used to estimate the fermentativepattern of Ðbre before and after digestion.

Digestive treatments have led to various forms ofdegradation or solubilisation of cell-wall poly-saccharides, depending on their sugar composition andbehaviour relative to the pH and temperature condi-tions of the particular treatment involved (Guillon et al1992 ; Au†ret et al 1993). Xylose and glucose are usuallya†ected very little (Au†ret et al 1993). Thus, the cellu-lose, xylan and arabinoxylan present in WB were nearlyall recovered (82È104%) at the end of the small intestine(Millard and Chesson 1984 ; Graham et al 1986 ; Bach-Knudsen and Hansen 1991). Conversely, the main cellwall polysaccharides of BB, similar to those in oat Ðbre,were di†erent from those of WB. Moreover, partialsolubilisation of arabinoxylans and b-glucan from BBoccurred during transit through the upper gut (74% ofb-glucan was recovered in the duodenum and 30% atthe terminal ileum; Bach-Knudsen et al 1993). The lossof b-glucan corresponded to the soluble form present inBB (about two-thirds of total b-glucans, or 52% of totalglucose ; Bach-Knudsen and Hansen 1991, Bach-Knudsen et al 1993) ; b-glucans, because of its high solu-

bility, are easily degradable substrates for the bacteriapermanently colonising this part of the gastrointestinaltract (70È97% for b-glucan, Graham et al 1986 ; Bach-Knudsen and Hansen 1991). In our study, the loss ofb-glucan during digestive treatments did not lead tomodiÐcation of physicochemical properties but mainlyto the removal of a highly fermentable substrate.Actually, microbial degradation of polysaccharidesoccurred in the small intestine of the cannulated pig,producing SCFA (24È74 lM) and involving a furtherdecrease in the fermentation of barley residues recov-ered from pig digesta. All modiÐcations of cell wallpolysaccharides occurring during digestive treatment,added to the di†erences of chemical structures of poly-saccharides can modify the fermentation pattern of cellwall polysaccharides as observed for the fermentation ofarabinoxylans of WB and BB residues. Some fractionsof BF pectin, arabinan and arbinogalactan are moresusceptible to removal from the Ðbre matrix, regardlessof the pH and temperature in the digestive tract orduring enzymatic hydrolysis (Au†ret et al 1991, 1993 ;Guillon et al 1992). Despite the di†erent pH and tem-perature conditions, in vivo and in vitro digestion of BFled to similar pectin solubilisation linked to an increasein hydration properties. This removal of the solubleÐbre fraction increased the rate of degradation of cell-wall pectic and cellulosic polysaccharides at the begin-ning of fermentation by favouring the accessibility ofother insoluble cell-wall polysaccharides (pectins,cellulose). During the fermentation of native BF, thesame losses of pectic substances were only reached after12 h of fermentative treatment. From these results, itwould seem that a proportion of amorphous materials(arabinan, arabinogalactan, pectin) protects the remain-ing cell-wall polysaccharides against further degrada-tion. The removal of some pectins leads to an increasein hydration properties and Ðbre porosity by creatingempty spaces in the BF cell-wall matrix (Au†ret et al1993), favouring microbial degradation. Thus, digestivetreatments in our study involved chemical and physi-cochemical modiÐcations, resulting in quicker fermenta-tion of BF.

In vitro as well as in vivo digestion led to the removalof non-dietary Ðbre components (starch, proteins) whichinterfere quantitatively in the fermentation pattern ofÐbre substrates. Although the intensity of in vivo and invitro digestive treatments was slightly di†erent depend-ing on the substrate, they resulted in insoluble residueswith similar chemical composition and physicochemicalproperties to those of native substrates. Moreover, theloss of easily degradable polysaccharides (b-glucan,pectin) during both types of digestion modiÐed fermen-tation qualitatively by increasing the rate and extent ofsugar loss. Thus, in the study of Ðbre fermentation, it isimportant to use Ðbre with the chemical nature andphysical arrangement of polysaccharides at entry intothe colon. Enzymatic treatment correctly reproduces

Digestion and dietary Ðbre properties 333

Ðbre modiÐcations arising in the digestible tract andappears to be a suitable means of preparing a largenumber of puriÐed Ðbres for further in vitro fermenta-tion studies.

ACKNOWLEDGEMENTS

The authors wish to thank Mr F de Monredon for hisassistance in the particle size determinations.

REFERENCES

Au†ret A, Barry J L, Thibault J F 1991 In vitro degradation ofchemically treated sugar beet Ðbres by human faecal bac-teria. Food Hydrocoll 5 41È44.

Au†ret A, Barry J L, Thibault J F 1993 E†ect of chemicaltreatments of sugar beet Ðbre on their physico-chemicalproperties and their in vitro fermentation. J Sci Food Agric61 195È203.

Bach-Knudsen K E, Hansen I 1991 Gastrointestinal implica-tions in pigs of wheat and oat fractions : 1 Digestibility andbulking properties of polysaccharides and other major con-stituents. Br J Nutr 65 217È232.

Bach-Knudsen K, Jensen B B, Hansen I 1993 Digestion ofpolysaccharides and other major components in the smalland large intestine of pigs fed on diets consisting of oat frac-tions rich in b-glucan. Br J Nutr 70 537È556.

Barry J L, Chourot J M, Bonnet C, Kozlowski F, David A1989 In vitro fermentation of neutral monosaccharides byruminal and human faecal Ñora. Acta V et Scand 86 (suppl)93È95.

Barry J L, Hoebler C, MacFarlane G T, MacFarlane S,Mathers J C, Reed K A, Mortensen I, Norgaard I, RowlandI R, Rumney C J 1995 Estimation of dietary Ðbre : a Euro-pean interlaboratory study in vitro. Br J Nutr 8 303È322.

Bourquin L D, Titgemeyer E C, Garleb K A, Fahey G C 1992Short-chain fatty acid production and Ðber degradation byhuman colonic bacteria ; e†ects of substrates and cell wallfractionation procedures. J Nutr 122 1508È1520.

Brillouet J M, Rouau X, Hoebler C, Barry J L, Carre� B, LortaE 1988 A new method for determination of insoluble cellwall and soluble nonstarchy polysaccharides from plantmaterials. Agric Food Chem 36 969È979.

Cherbut C, Salvador V, Barry J L, Doulay F, Delort-Laval J1991 Dietary Ðbre e†ects on intestinal transit in man: aninvolvement of their physico-chemical and fermentativeproperties. Food Hydrocoll 5 15È22.

Cummings J H, MacFarlane G T 1991 The control and conse-quences of bacterial fermentation in the human colon. JAppl Bacteriol 70 443È459.

Graham H, Hesselman K, Aman P 1986 The inÑuence ofwheat bran and sugar-beet pulp on the digestibility ofdietary components in a cereal-based pig diet. J Nutr 116242È251.

Guillon F, Barry J L, Thibault J F 1992 E†ect of autoclavingsugar-beet Ðbre on its physico-chemical properties and its invitro degradation by human faecal bacteria. J Sci FoodAgric 60 69È79.

Guillon F, Renard C M G C, Hospers J, Thibault J F, BarryJ L 1995 Characterisation of residual Ðbres from fermenta-tion of pea and apple Ðbres by human faecal bacteria. J SciFood Agric 68 251È529.

Hoebler C, Barry J L, David A, Delort-Laval J 1989 Rapidhydrolysis of cell wall polysaccharides and simpliÐed quan-

titative determination of neutral monosaccharides by gas-liquid chromatography. J Agric Food Chem 37 360È367.

Hoebler C, Barry J L, David A, Kolzowski F 1991 Enzymaticpreparation and analysis for in vitro fermentation studies.Food Hydrocoll 5 35È40.

Jensen B B, Jorgensen H 1994 E†ect of dietary Ðber on micro-bial activity and microbial gas production in variousregions of the gastrointestinal tract of pigs. Appl EnvironMicrobiol 60 1897È1904.

Jouany J P 1982 Dosage des acides gras volatils (AGV) et desalcools dans les contenus digestifs, les jus dÏensilage, les cul-tures bacte� riennes et les contenus de fermenteurs anae� -robies. Sci Aliment 2 131È144.

Karkalas J 1985 An improved enzymic method for the deter-mination of native and modiÐed starch. J Sci Food Agric 351019È1027.

Kuniak L, Marchessault R H 1972 Study of cross-linkingreaction between epichlorhydrin and starch. Die Starcke 4110È116.

Livingstone R M, Baird B A, Atkinson T, Crofts R M J 1980Circadian variation in the apparent digestibility of dietsmeasured at the terminal ileum in pigs. J Agric Sci Camb 94399È405.

MacFarlane G T, Gibson C R, Beatty J H, Cummings J H1992 Estimation of short-chain fatty acid production fromprotein by human intestinal bacteria based on branched-chain fatty acid measurements. FEMS Microbiol Ecol 10181È88.

McBurney M I, Thompson L U, Cu† J, Jenkins D J A 1988Comparison of ileal effluents, dietary Ðbres and whole foodsin predicting the physiological importance of colonic fer-mentation. Am J Gatroenterol 83 536È540.

McCleary B V, Codd R 1991 Measurement of (1-3), (1-4)-b-D-glucan in barley and oats : a streamlined enzymatic pro-cedure. J Sci Food Agric 55 303È312.

Michel C, MacFarlane G T 1996 Digestive fate of solublepolysaccharides from marine macroalgae : involvement ofthe colonic microÑora and physiological consequences forthe host. J Appl Bacteriol 80 349È369.

Millard P, Chesson A 1984 ModiÐcation to swede (Brassicanapus L.) anterior to the terminal ileum of pigs : some impli-cations for the analysis of dietary Ðbre. Br J Nutr 52 583È594.

Mortensen M, Holtug K, Rasmussen H S 1988 Short-chainfatty acid production of mono- and disaccharides in a fecalincubation system: implication for colonic fermentation ofdietary Ðber in humans. J Nutr 118 321È325.

Robertson J A, Eastwood M A 1981 A method to measure thewater-holding properties of dietary Ðbre using suction pres-sure. Br J Nutr 46 247È255.

Robertson J A, Murison S D, Chesson A 1986 Loss of selectedwater-insoluble polysaccharides and component neutralsugars from swede (Brassica napus (cv. danestone) andcereal bran measured during digestion in the pig caecum. JSci Food Agric 37 359È365.

Robertson J A, Murison S D, Chesson A 1992 Particle sizedistribution and solubility of dietary Ðbre in swede-baseddiets during gastrointestinal transit in the pig. J Sci FoodAgric 58 197È205.

Salvador V, Cherbut C, Barry J L, Bertrand D, Bonnet C,Delort-Laval J 1993 Sugar composition of dietary Ðbre andshort-chain fatty acid production during in vitro fermenta-tion by human bacteria. Br J Nutr 70 189È197.

Scheppach W 1994 E†ects of short chain fatty acids on gutmorphology and function. Gut 1 (suppl) S35ÈS38.

Thibault J F 1979 Automatisation du dosage des substancespectiques par la me� thode au me� ta-hydroxydiphe� nyl.L ebensm W iss T echnol 12 247È251.