lemak

Thesizeandinterfacialcompositionofmilkfatglobulesarekeyfactorscontrollingtriglyceridesbioavailabilityinsimulatedhumangastro-duodenaldigestionCyrielleGarcia a,b,ClaudineAntona b,c,BenoitRobertd,e,ChristelleLopez d,e,MartineArmand a,b,*aINSERM,U476,F-13385Marseille,FrancebAix-MarseilleUniversit,F-13385Marseille,FrancecINRA,UMR1260,F-13385Marseille,FrancedINRA,UMR1253ScienceetTechnologiedulaitetdeluf,F-35000Rennes,FranceeAgrocampusOuest,UMR1253STLO,F-35000Rennes,FrancearticleinfoArticlehistory:Received19March2013Accepted8July2013Keywords:CowmilkFatglobulesizeHumangastriclipasePancreaticlipaseLipidinterfacephysicochemicalpropertiesFattyacidcompositionabstractLipidsorganisationmightmodulatefattyacidbioavailabilityleadingtohealthimplications.Wedeter-minedwhetherthesizeandtheinterfacialcompositionofcowmilkfatglobulescouldaffecttriglyceridesdigestibility.Nativefatglobulesofvarioussizescoveredbytheirbiologicalmembrane(4.2 mm,largeLFG6.6 mm,smallSFG1.7 mm)orhomogenised heat-treated(0.3 mm)weredigestedingastricandduodenalconditionssimulatinghumanphysiology.Lipolysisextentswerecalculatedfromtheamountoffreefattyacidsgenerated,andthefattyacidcompositionoftheproductsoflipolysiswasdeterminedbyGCanalysis.SFGweremoreefcientlyhydrolysedthanLFGbygastric(13.3versus5.6%),gastricpluspancreatic(62.9versus48.7%)andpancreatic(79.6versus54.7%)lipases.AhigherlipidinterfaceareawithnativeSFG,thatmightincreaselipasesbindingsites,canexplaintheseresults.However,theho-mogenisation,whichmarkedlydecreasesfatglobulesizeincreasingconsequentlythelipid/waterinterfacearea,didnotimprovegastric(9%)orduodenal(64.5%)lipolysisprobablyduetoanimportantchangeinglobulesurfacecomposition(proteinsversusphospholipids).Interestingly,thesizeofthemilkglobule(SFGandHMversusNMandLFG)controlsthetypeofthefreefattyacidsgeneratedbythehumangastriclipase,palmiticversusoleicacid,suggestingadifferentorientationoftheaccessiblemixedtriglycerides.Moreover,thetypeofmonoglyceridesproducedfromSFGdigestioncouldbelessathero-geniccomparedtoLFG.Thesizeoffatglobulesgovernsgastricandduodenallipolysisextentwhenthecompositionoftheinterfaciallayerisappropriate.Itmightfurthercontrolfattyacidbioavailabilityimpactingongastricemptyingrateviathepreferentialreleaseofoleicacid,astrongstimulatorofCCK. 2013ElsevierLtd.Allrightsreserved.1.IntroductionAlthoughdairylipidsrepresentaboutonethirdofingestedd etaryfatinhumans(Astorgetal.,2004),themechanismsoftheirdigestionasafunctionoftheirorganisationarefarfrombeingelucidated.Lipidsaresecretedinmilkintheformofuniquebiologicalassembliescalledfatglobules(Keenan&Patton,1995;Lopez,2011).Theseglobulesarecomposedbyacoreoftri-glycerides,cholesterolandretinolesters,thatiscoatedbyabio-logicalmembranei.e.themilkfatglobulemembrane(MFGM)

2.Materialsandmethods2.1.CowmilksamplesoriginandpreparationRawwholecowmilkcallednativemilk(NM)waspurchasedfromadairyplant(Entremont,MontaubandeBretagne,France).Milksampleswerethermallytreated(73 Cfor20min)toinacti-vateendogenouslipases(Bertonetal.,2012).Nativelarge-sizedfat-globulesfractions(LFG)andsmall-sizedones(SFG)wereisolatedfromNMbymicroltrationaspreviouslydetailed(Bertonetal.,2012;Lopezetal.,2011).Homogenisedmilk(HM)wasobtainedfromNMusingatwo-stagehigh-pressurehomogeniser(20MPa,RannieLAB16/15,APVFrance,Evreux,France).Ultra-high-temperaturetreatment(UHT:140 Cfor4s,Actijouletubularsterilisator,Actini,EvianlesBains,France)wasappliedtoHM(HTMsamples).Severalbatchesweretested(9forNM,7forLFG,6forSFG,5forHM,4forHTM).Totalfatconcentrationofthesampleswasdeterminedusingareferencegravimetricmethod(ISO1211).2.2.Humangastricjuiceandpancreatin/bilecharacterisationGastricjuicescollectedfordiagnosticpurposesafterpentagas-tricstimulation(6 mg/kg)wereselectedfromadultsubjectswithnodiagnosedgastricpathologyandnoduodenalreux(Gastroenter-ologyDepartment,HpitalNord,Marseille,France).GastriclipaseactivitywasmeasuredontributyrinusingapH-stattitrator(Met-rohm,Herisau,Switzerland)atpH5.40and37 Caspreviouslydescribed(Armandetal.,1999).Pepsinactivitywasmeasuredat37 Cin10minatpH1.8froma2%haemoglobinsolutionasalreadypublished(Armandetal.,1995).Thechosengastricjuicesexpressednormalactivitiesforgastriclipase(70e97.5U/mLontributyrin,onelipaseunitcorrespondstothereleaseof1 mmolfattyacidpermin)(Armand,2007)andpepsin(533e869U/mL)(Armandetal.,1995).Adistilledwatersolutionofpowderfrompigpancreas(Pancreatin,Sigma,StLouis,MO,USA)waspreparedtomimicthepancreaticenzymes(1g/10mLsolutioncentrifugated5minat2000rpm).Thepancreaticlipaseactivityofthepreparation,measuredontribu-tyrinwithapH-stattitratoratpH8.0and37 Caspreviouslydescribed(Armandetal.,1999),was7200U/mL.Bilewascollectedfrompiggallbladder(AbattoirsdelEstaque,Marseille,France),anditsphospholipidsandbilesaltconcentrationsanalysedusingmethodspreviouslypublished(Armandetal.,1999)were19and235mM,respectively.2.3.Particlesizeandzeta-potentialmeasurementsThefatglobulesizedistributionsofthemilksamplesweremeasuredatzerotimeandafterthegastricstep(atacidicpHwithorwithoutgastricjuice)bylaserlightscatteringusingaMastersizer2000equippedwithtwolasersources(MalverninstrumentsInc,UK).Therefractiveindexesusedwere1.458and1.460formilkfatat633and466nm,respectively,and1.33forwater.About0.2mLofmilksamplesweredilutedin100mLofdistilledwaterdirectlyinthemeasurementcelloftheapparatusinordertoreach10%obscuration.ThesamplesobtainedafterthegastricstepweredispersedinSDS1%underagitationfor10minbeforemeasure-mentsinordertodissociatetheproteinnetworkformedatacidicpH.Themicellesofcaseinweredissociatedbyadding1mLof35mMEDTA/NaOHbufferpH7tothesamples.Granulometryparameterswerecalculatedbythesoftwaresuchasthevolume-weightedaveragediameter d43 ( nid4i= nid3i,where ni isthenumberoflipiddropletsofdiameter di),thevolume-surfaceaveragediameter d32 ( nid3i= nid2i),bothexpressedas mm,andthespecicsurfaceareaS(6.f/d32,where f isthevolumefractionofmilklipid)expressedasm2/goflipid.Theapparentzeta-potentialofmilkfatglobuleswascalculatedfromthemeasurementoftheirelectrophoreticmobilitydeterminedusingaZetasizer3000HS(Malvern)equippedwithpalladiumelectrodesandanavalanchephotodiodedetectoraspreviouslypublished(Bertonetal.,2012).Sampleswerepreparedbysuspending4e12 mLmilkin10mLmilkultraltrateandintroducedinthecapillarytubeformeasurement.2.4.DigestiontestprocedureGastricdigestionwasconductedinmicrotubesat37 Cand nalpH4.4underconstantandvigorousshaking(140rpm)for60min.The1mL-reactionmediumcontainedsodiumacetate50mM,CaCl26mM,NaCl150mM,asuitablevolumeofcowmilksampleproviding8mgtriglycerides(TG)andgastricjuice(2.2lipaseunits/mgTG).Atcompletionofthedigestiontime,chloroform/methanol2:1wasaddedtostopthereaction,ortesttubesweredirectlyusedforduodenaldigestion.Inthelatercase,thepHwasincreasedupto6.8usinganadequatebuffersolution(Tris80mM,NaCl150mM,CaCl2 10mM, nalconcentrations,pH8).Duodenaldigestionwasperformedat37 C,underconstantandvigorousshaking(140rpm)for30min,ina1.5mL nalvolumecontaining8mMbilesaltsfrompigbileandpancreatinsolution(22.5lipaseunits/mgTG).Atcompletionofthedigestiontime,chloroform/methanol2:1wasaddedinthetesttubestostopthereaction.Theconditionsusedi.e.quantitiesofenzymes,bilesaltconcentration,lipase-to-lipidsra-tios,volumeofconsumablemilktovolumeofgastricjuicesecretedunderstimulationduring1-h(ratio1.25),andpHvalueswereinC.Garciaetal./FoodHydrocolloids35(2014)494e504 495accordancewithintra-gastricandintra-duodenalconditionsdur-ingdigestioninhealthyhumans(Armand,Borel,etal.,1996,Armandetal.,1995,1999;Carrire,Barrowman,Verger,&Laudier1993;Pafumietal.,2002).SometesttubeswerepreparedforpHvericationduringdigestion(gastricphase, n 9,pH4.40 0.08;duodenalphase, n 11,pH6.79 0.08).2.5.LipidextractionandanalysisLipidsfromeachmilksampleandfromdigestiontestswereextractedfollowingamodiedFolchsmethod(Armand,Borel,etal.,1996,Armand,Hamosh,etal.,1996;Armandetal.,1999)usingchloroform/methanol2:1(v/v)(CarloErbaReagents,SDSPeypin,France).Onepartofthelipidextractwasdryedandresuspendedinisopropylalcoholtoquantifyfreefattyacids(FFA)usingenzymaticassay(kitNEFARandox,Crumlin,UK).TheoverallextentsoflipolysiswerecalculatedasthepercentofTGhydrolysed,from mmolesofFFAreleasedattheendofincubationtimes(FFAt)relativetothetotal mmolesofphysiologicallyreleasableFFA(sn-1andsn-3positions),basedontheinitialamountofTG,usingthefollowingequation:[FFAt/(TG 2)] 100,aspreviously(Pafumietal.,2002).FFApresentinmilksamplesbeforedigestionweresubstractedfromFFAtvalues.FFA,monoglycerides(MG),diglycerides(DG),andTGwereseparatedfromtheotherpartoflipidextractbyone-dimensionalthinlayerchromatographyonsilicagelplates(TLCsilicagel60withconcentratingzone20 20cm,Merck,Seram,Marseille,France)developedinchloroform/methanol/aceticacid(98:2:0.1,v/v/v)asalreadydescribed(Armand,Borel,etal.,1996,Armand,Hamosh,etal.,1996;Armandetal.,1999).TheresultingbandswerevisualizedbysprayingtheplatewithTNS(6-p-toluidino-2-naphthalenesulfonicacid,Fluka,SaintQuentinFallavier,France)andusingUVlight,thenindividualbandswerescrapedintoscrew-captubes,andlipidwereextracted.Fattyacidsfromwholelipidextracts,TG,DGandMGweretransesteriedusingsodiummeth-oxide0.5Minmethanol(Christopherson&Glass,1969),andFFAweremethylatedusingmethanolandtrimethylsilyldiazomethane(Aldai,Murray,Najera,Troy&Osoro,2005).Bothmethodswereperformedatroomtemperaturewithnoevaporationstepinordertoavoidthelossofshortandmediumchainfattyacids.FAMEwereanalysedbyGC(Clarus500, ameionizationdetector,Turbochromsoftware,PerkinElmer,Courtaboeuf,France)usingBPX70capillarycolumn(2 50m 0.32mmi.d.,0.25 mm lmthickness)(Sigma-Supelco,SaintQuentinFallavier,France).Carriergaswashydrogenandoventemperaturestartedat50 Cwitha10minhold,riseof10 C/minwasthenappliedto175 Cwitha27minhold,and nally4 C/minto215 Cwitha16minhold.FAMEwereidentiedbytheirretentiontimesonthecolumnusingstandards(GLC674,GLC81,Nu-check,USA)andexpressedasweightpercentoftotalfattyacids.2.6.ConfocallaserscanningmicroscopyThemicrostructuralanalysisofthemilksampleswasperformedusinganinvertedmicroscopeNIKONEclipse-TE2000-C1si(NIKON,ChampignysurMarne,France)allowingconfocallaserscanningmicroscopy(CLSM).Confocalexperimentswereperformedusinganargonlaseroperatingat488nmexcitationwavelength(emis-sionwasdetectedbetween500and530nm)aHeeNelaseroper-atingat543nmwavelengthexcitation(emissionwasdetectedbetween565and615nm).Theobservationswereperformedusingan 40(NA1.33)andan 100(NA1.4)oilimmersionobjectives.Triglycerideswerestainedwiththelipid-solubleNileRed uores-centdye(9-diethylamino-5H-benzo-alpha-phenoxazine-5-one,SigmaeAldrich,StLouis,USA)(orangecolour),phospholipidsintheMFGMwerelabelledusingRh-DOPE(N-lissaminerhodamineBsulfonyldioleoylphosphatidylethanolamine,AvantipolarlipidsInc.,Birmingham,England)(redcolour),andproteinswerestainedwithacridineorange uorescentprobe(AldrichChemicalCom-pany,Inc.,Milwaukee,USA)(greencolour).Theorganisationoffatglobuleswascharacterisedinallmilksamples.ThemicrostructureoftheNMandHMsampleswascharacterisedingastricconditionswithoutorwithgastricjuice(incubationatpH4,60min,37 C).2.7.StatisticalanalysisDataaregivenasmean SEMofthedifferentmilkstructures(severalbatchestestedforeachstructurecategory).Physicochem-icalparametersweredeterminedintriplicateforeachsample.Lipolysisassayswererunatleastinduplicateforeachmilkbatch.StatisticalanalyseswereperformedbyusinganonparametricKruskaleWallistest,anddifferencesbetweengroupswerecheckedusingtheManneWhitney U test.Differencesbetweenthemeanvalueswereconsideredsignicantif p < 0.05.AllanalyseswereperformedusingtheSTATVIEWsoftware(AbacusConcepts,Berkeley,CA,USA).3.Results3.1.MilksamplescharacteristicsThemicrostructure(Fig.1)andthephysicochemicalproperties(Table1)ofthenativeandhomogenisedmilkfatglobuleshavebeencharacterised.ConfocalmicrographsconrmedexpectedfatglobulesshapeinNM(Fig.1A,heterogeneoussize),LFGfraction(Fig.1B,largersize),SFGfraction(Fig.1C,smallersize),andinHM(Fig.1D,smallestsize).InNM,LFGandSFGfractions,milkfatglobuleswerecoveredbytheirnativebiologicalmembrane(MFGM)containingsphingomyelinerichdomainsandglycer-ophospholipids(Fig.1E)asrevealedpreviously(Lopezetal.,2011).Milkproteins(caseinsandwheyproteins,ingreencolourintheconfocalmicrographs)weredispersedintheaqueousphasesur-roundingnativefatglobules.(inthewebversion)InHM,partoftheproteinsfromtheaqueousphaserelocatedasadenselayeradsorbedaroundtheTGcoreofthefatglobulesinzoneswhereMFGMismissing(Fig.1F).Thefatglobulesofthedifferentmilkstructuresexhibitedsignicantlydifferentmeandiametersandconsequentlysignicantlydifferentlipidinterfaceareas(Table1).TheSFGwereabout2.4or3.7-foldsmallerthanNMglobulesorLFG,respectively,withanincreasedinterfaceareaofabout2.5-fold.Asexpected,homogenisationproceduredecreasedglobulessizeby13-or20-foldresultinginanabout20-foldincreaseinlipidinter-faceareaforHMandHTM,comparedtoNMorLFG.FatglobulesfromHMandHTMexhibiteda5-fold(d43)to8-fold(d32)lowersizeandan8.5-foldhigherinterfaceareathanSFG.Particlesizedistri-bution(Fig.2)showedthatSFG(Fig.2D)werericherin1 mm-diameterglobules,andhomogenisationdecreasedfatglobulesizeto0.1e1 mmglobules(Fig.2B),comparedtoNM(Fig.2A)andLFGfraction(Fig.2C).FatglobulesofNM,SFGandLFGfractionshadsimilarapparent z-potentialof 12.8 0.8mV, 13.0 0.8mVand 12.9 1.0mV,respectively,asalreadyreported(Bertonetal.,2012;Lopezetal.,2011).Theabsolute z-potentialvalueincreasedafterhomogenisationofmilk(20.5 0.7mV)asaresultofMFGMdisruptionandmilkcaseinsadsorptionatthesurfaceoflipiddroplets(Fig.1F)(Bertonetal.,2012;Lee&Sherbon,2002).TheUHTtreatmentoftheHMdidnotchangethe z-potential(20.1 0.7mV),howevercausingdenaturationofwheyproteinsandtheirinteractionwithcaseinmicellesatthesurfacelipiddroplets(Fig.1F)(Lee&Sherbon,2002).Thefattyacidcompositionofthemilksampleswasnotmodiedbyhomogenisationandheattreatments(datanotshown),butwasC.Garciaetal./FoodHydrocolloids35(2014)494e504 496signicantlydifferentfortheSFGcomparedtotheLFGfraction(5.9%shortchains, 6.14%myristicacid, 32.5%myristoleicacid, 11.73%palmitoleicacid, 15.3%stearicacid, 24.7%linoleicacid, 22.6%linolenicacid, 44.1%CLA)(Table1, onlinematerial)thatisrelativelyclosetodatapreviouslyreported(Fauquant,Briard,Leconte,&Michalski,2005;Lopez,2011;Lopezetal.,2011).3.2.Effectofthesizeofthemilkfatglobulesontheextentofgastric,duodenalandgastro-duodenallipolysisHumangastriclipasewasverysensitivetothesizeofthenativefatglobulesasSFGwerehydrolysedmoreefcientlythanLFG(138%increase, p 0.0001)(Fig.3A).Thiseffectwasalsoobservedforpancreaticlipasesingastro-duodenal(29%increase, p 0.0147)andduodenalconditions(45.5%increase, p 0.0019)buttoalesserextent(Fig.3B).ComparedtoNM,theTGfromLFGorSFGwere,respectively,signicantlyless(43%decrease)ormore(36%in-crease)lipolysedingastricconditions(Fig.3A).Comparedtohomogenisedfatglobules,theLFGwerealsolessefcientlyhydrolysedbythehumangastriclipase(39and42%decreasecomparedtoHTMandHM,respectively),meanwhileSFGweremorelipolysed(37and45%increasecomparedtoHMandHTM,respectively).Ingastro-duodenalphase,LFGwerelesshydrolysedthanNM,HMandHTM(16,14and22.6%decrease,respectively),whiletheoverallextentofSFGlipolysiswasnotsignicantlydifferenttoNM,HMandHTM(Fig.3B).Induodenalconditions,sowithnopriorgastricstep,LFGwerehydrolysedatasameextentthanNM,HMandHTMglobules,butSFGwere26e35%moreefcientlyhydrolysed.3.3.Effectofmilktechnologicaltreatmentsontheextentofgastric,duodenalandgastro-duodenallipolysisHomogenisation UHTdidnotaffecttheextentofdigestivelipolysis.TGfromNM,HMandHTMwerelipolysedtoabout9.5%bythehumangastriclipase,showingnoimpactofthefatglobulesizeandoftheinterfacialarea(Fig.3A).Samely,lipolysisingastro-duodenalorduodenalconditionsreachedcomparablevaluesranging57e64%and59e63%,respectively(Fig.3B).3.4.FattyacidcompositionofthedifferentclassesoflipidsgeneratedduringlipolysisstepsThegastriclipasegeneratedatthesametimeshort(range1.8e4.6%oftotalfattyacids),medium(6.3e10.1%)andlong(85.3e91.9%)chainFFA(Fig.4A)andwasalsoabletoproduce sn -2MG(Fig.4B).ThemaindifferencesinthecompositionofFFAgeneratedbetweenFat globules covered by aggregated milk proteins and residual MFGM F: Homogenized milk fat globules 0.15 -1.5 m; -potential = -16 to -20 mVE: Native milk fat globules 1.6 10 m; -potential = -13 mVGlycerophospholipids (PC, PE, PI, PS)Glycosylated proteinsTGSphingomyelin-rich domains in the MFGM: lipid raftsMFGMCaseinmicellesWhey proteinsCaseinsHomogenized and heat treated milk fat globules 01515m; potential= 16to 20mVSphingomyelin (SM)CholesterolGlycosphingolipidesTGGlycerophospholipids in the MFGM0.15- 1.5m; - = -16 to-20mVTGMFGMCaseinmicellesDenaturedwhey proteinsWhey protein casein complexButyrophilinThe MFGM: A trilayer of polar lipids and membrane proteinsFat globules covered by milk proteins and residual MFGM B A C D20 m 20 m 20 m 20 mTGFig.1. Microstructureofnativeversushomogenisedmilkfatglobules.ConfocallaserscanningmicrographsshowingthemicrostructureofnativewholemilkNM(A),ofcowmilkLFG(B)orSFG(C)fractionsobtainedfromwholenativemilkusinganoptimizedmulti-stagemicroltrationprocess,ofhomogenizedmilk(D),ofsphingomyelin-richdomainsandglycerophospholipidscontainedinnativebiologicalmembrane(MFGM)ofNM,LFGandSFGfractions(E),andofpartofmilkproteins(caseinsandwheyproteins)fromtheaqueousphaserelocatedasadenselayeradsorbedontotheTGcoreofthefatglobulesinzoneswhereMFGMismissing(F).Triglycerideslocatedinthecoreoffatglobuleswerelabelledusingnilered(orangecolour),milkproteinswerelabeledwithacridineorange(greencolour),andthephospholipidsintheMFGMwerelabeledusingRh-DOPE(redcolour).Scalebar 20 mm.(Forinterpretationofthereferencestocolourinthis gurelegend,thereaderisreferredtothewebversionofthisarticle).Table1Physicochemicalcharacteristicsofthecowmilksamples*.MilkstructureLipidconcentrationSizedistributionparameters(g/100g) d43 (mm) d32 (mm) S (m2/gfat)Native(NM)4.11 0.06a4.30 0.03a3.59 0.02a1.82 0.01aNativelarge-sizedfatglobules(LFG)8.82 1.25b6.56 0.07b3.98 0.06b1.64 0.03bNativesmall-sizedfatglobules(SFG)3.92 0.98a1.75 0.07c1.54 0.07c4.29 0.18cHomogenised(HM)4.09 0.12a0.34 0.02d0.18 0.01d37.08 1.23dHomogenised-UHT(HTM)4.20 0.05a0.35 0.02d0.18 0.01d36.22 1.14d*Valuesaremeans SEM,NM n 9batches(27measurements),LFG n 7batches(21measurements),SFG n 6batches(18measurements),HM n 5batches(15measurements),HTM n 4batches(12measurements).Meansinacolumnwithdifferentsuperscriptsletters(a,b,c,d)differsignicantly(p < 0.05).C.Garciaetal./FoodHydrocolloids35(2014)494e504 497milkstructureswereasignicanthigherlevelofbutyricacid(C4:0)withNM,alowerlevelofoleicacid(C18:1)oroftotalMUFAwithNMcomparedtoSFGfractionandHM,andahighertotalsaturatedFFA,mainlyimputabletopalmitic(C16:0)andstearic(C18:0)acidtakentogether,withtheLFGandNMcomparedtoSFGandHM(Fig.4A).BycalculatingtheratiobetweenindividualFFAreleasedingastricphaseandthecorrespondingFAininitialTG,wefoundthatthehumangastriclipaseactedmorepreferentiallyonestersbondsinvolvingcaprylicacid(C8:0)andstearicacid(ratio > 2),thenonpalmiticandbutyricacids(ratiocloseto1),andonlauricacid(C12:0)andC15:0(1 > ratio 0.5)(Fig.5).ConsideringtheTGintramolecularstructureofcowmilkreportedintheliterature(Table2)(Blasietal.,2008),ourdatashowedthatgastriclipaseisabletoreleasefattyacidspresentatthe sn -3position(i.e.butyricandcaprylicacids)butalso,andevenmoreefciently,thoselocatedatthe sn -1position(lauric,palmiticandstearicacids).Thefattyacidcompositionof sn -2MG(Fig.4B)isinconcordancewiththetypeoffattyacidpresentatthe sn -2positionofthecowmilkTG(Table2)(Blasietal.,2008).Consideringmilkstructures, sn -2MGprolewasalmostalwayssignicantlydifferentbetweenLFG(richerinlauric,palmitoleicandoleicacids)andSFG(richerinSFAi.e.palmiticandstearicacids),withintermediatevaluesforNM,HMandHTM.GastricDGandundigestedTGstillcontainedshortandmediumchainfattyacids(Fig.1AandB, onlinematerial).MainsignicantdifferencesinfattyacidcompositionofDGandundigestedTGwereobservedbetweentheLFGfractionandHTM,i.e.themilkstructuresextremelydifferentintermofbothfatglobulesizeandsurfacecomposition,withoverallinversechangesinfattyacidcompositioninDGversusTG.ConcerningspecicallytheSFGandLFGfractions,themaindifferencesintotalfattyacidcompositionfoundintheinitialTG(beforelipolysis)wereerasedafterthegastricdigestionstep(Table1, onlinematerial),exceptthatpalmiticacidlevelbecamesignicantlyhigherinSFG,andsimilarsignicantchangesinfattyacidsproportionwereobservedintheundigestedTGversusinitialTGforbothstructures(decreaseinC6:0,C8:0,C16:0,CLA,andincreaseinC12:0,C14:0,C14:1,C15:0,C16:1,C17:0,C18:1,C18:2n-6).Afterthewholegastro-duodenaldigestionwascompleted,thefattyacidcompositionofFFA(Fig.6A)andof sn -2MG(Fig.6B)wasmorehomogeneousbetweenthedifferentmilksamplesandwithinaspecic milkstructure(lowerSEMvalues)probablybecause97e98%ofthe sn -1and sn -3esterbondsofTGweredigested.Fig.3. Extentoftriglycerideslipolysisingastric(A)orgastro-duodenalandduodenal(B)conditionsmimickinghumanphysiologyfromnative(NM),homogenised(HM),homogenised-UHT(HTM)milksamplesandofnativeLarge-sized(LFG)andSmall-sized(SFG)fat-globulefractions.Theresultsrepresentmeans SEMof n 52mea-surements(NM), n 28(HM), n 24(HTM), n 48(LFG),and n 44(SFG)forgastricandgastro-duodenaldigestiontest,and n 20(NM,HM), n 16(HTM,LFG), n 12(SFG)forsingleduodenaldigestionstep.Barswithdifferentletters(a,b,c)aresignicantlydifferent(P < 0.05).Fig.2. Sizedistributionsoffatglobulesin(A)nativemilk(NM),(B)homogenisedmilk(HM),(C)nativelarge-sizedfatglobulefraction(LFG),and(D)small-sizedfatglobulefraction(SFG)atinitialtime,andintheconditionsofgastricdigestionasindicatedinfigure(n 3).C.Garciaetal./FoodHydrocolloids35(2014)494e504 498However,somesignicantdifferencespersisted:lessstearicacidgeneratedasFFAforthefatglobulesofsmallersizes(SFG,HM,HTM)comparedtofatglobulesoflargersizes(LFG)withaninter-mediatevalueforNM,andhigherlevelofoleicacid(sototalMUFA)andlowertotalSFAinthe sn -2MGforthefatglobulesofsmallersizes(SFG,HM,HTM)comparedtotheLFGfraction,NMbeingintermediate.3.5.EvolutionoftheorganisationofnativeversushomogenisedmilksamplesduringdigestivestepsThedecreaseofinitialneutralmilkpHtogastricacidpH4changedthefatglobulessizedistributionmoderatelyforNM(Fig.2A),LFG(Fig.2C)andSFG(Fig.2D),slightlyforHM(Fig.2B),andleadtotheformationofaproteinnetwork(ingreen)duetomilkcoagulation(Fig.7A,B,C).(inthewebversion)Themicro-structureofthegelsformedwascompletelydifferentbetweenNMandHMsamplesascharacterisedbyCLSM(Fig.7A,B,C).InNMcoagulum,fatglobulesformedaggregates(inred)duetoadecreaseinthechargeoftheMFGM(4.3 0.4mVatpH4)andwerelocatedmainlyinthepocketsofaqueousphase(inblack).(inthewebversion)InHM,thetinyfatglobulesweretightlyembeddedinFig.4. Fattyacidcompositionofthefreefattyacids(A)orofthe sn -2monoglycerides(B)generatedduringgastriclipolysisstep,fornativeLarge-sized(LFG)orSmall-sized(SFG)fat-globulefractions,andfornative(NM),homogenised(HM),homogenised-UHT(HTM)milksamples.Theresultsrepresentmeans SEMof n 5(LFG,SFG,NM,HM),and n 4(HTM).Foragivenfattyacid,barswithdifferentletters(a,b,c)aresignicantlydifferent(P < 0.05).Table2Rangesoffattyacidcompositionof sn -2MGgeneratedduringthegastricorthegastro-duodenaldigestionstepsofdifferentwholemilkstructures(NM,HM,HTM)comparedtodatafromtheliteratureonfattyacidcompositionanddistributionofthetriglyceridesincowmilk.Fattyacids(mol%)Cowmilk*sn -2MGgeneratedduringdigestionGastricGastro-duodenalTG sn -1 sn -2 sn -3TG sn -2 sn -2C4:0Butyric6.01.30.417.92.900C6:0Caproic2.90.30.98.21.800C8:0Caprylic1.70.30.24.71.20.01e0.350e0.04C10:0Capric3.41.11.48.73.11.0e1.71.4e1.8C12:0Lauric3.92.54.85.13.62.3e6.14.5e5.4C14:0Myristic13.111.722.88.212.17.0e16.218.8e19.9C16:0Palmitic31.646.844.112.036.336.2e41.041.0e42.2C16:1Palmitoleic1.81.52.51.92.71.9e2.32.4e2.7C18:0Stearic6.611.15.35.27.89.3e16.35.3e5.6C18:1Oleic19.221.515.725.121.114.6e19.915.7e16.7C18:2Linoleic1.91.71.92.71.90.9e3.01.6e2.6*Reportedfrom(Blasietal.,2008).C.Garciaetal./FoodHydrocolloids35(2014)494e504 499theproteinnetwork(ingreen).(inthewebversion)Whengastricjuicewasincubatedwithmilksamplesfor1hat37 C,weobservedaprofoundchangeinthefatglobulesizedistributionforallmilkstructuresthroughcoalescencephenomenonthatwasmorepro-nouncedforNMandLFGcomparedtoHMandSFG(Fig.2).ThroughCLSMapproach,wecharacterisedanincreaseofthesizeofonepartofthefatglobulesinNMandHMduetocoalescence,andobservedthattheproteinnetworkwasstillpresentinthegastricphase(Fig.7D,E,F).4.DiscussionTheimpactofthephysicochemicalpropertiesofcowmilk,atanativestateorafterbeingmodiedbytechnologicalprocessesusedinthedairyindustry,onmilkfatdigestibilityhasbeenquestionedforalongtimeandneverreallyanswered.Weaimedatgeneratingdatausefultoraiseconclusionsintermsofhumanhealthbyusingan invitro digestionmodelsimulatinghumanphysiologicalcon-ditionsthankstotheuseofhumangastricenzymes,wholepancreaticenzymesandbilesystemfrompig(supposedtobecloseenoughtohumanones),pHconditionsclosetotheonesfoundinthehumandigestivetractduringameal,andphysiologicalratiosoftriglycerides-to-lipasesamountsandofmilkvolume-to-gastricjuicevolume(Armand,Borel,etal.,1996,Armandetal.,1995,1999;Carrireetal.,1993;Pafumietal.,2002).WhenconsideringnativemilkfatglobulescoveredbytheirMFGM,theglobulesizeappearsclearlyasakeyfactorfortheoverallextentsoflipolysis;interestingly,despiteprofoundphysi-cochemicalchangesobservedduringthegastricdigestionstep,anegativecorrelationwasfoundbetweengastriclipolysisover60minandinitialfatglobulesizeregarding d43 (R 0.53,p 0.0159), d32 (R 0.58, p 0.0079),orspecicsurfacearea(R 0.57, p 0.0094).Thiswasalreadyshowninhealthyhumansforarticialfatglobulesfromtwoconventionallecithins-oilemulsionssizing0.7 mmversus10 mm(Armandetal.,1999).Itwasalsorecentlyshownthatthecatalyticefciencyofthehumanpancreaticrecombinantlipase(inthepresenceofcolipaseandbilesalts)isgreaterwithsmall(1.75 mm)thanlarge(6.6 mm)nativecowmilkfatglobules invitro (Bertonetal.,2012).Smallfatglobulespresentalargersurfaceareathatfacilitateslipasesaccesstosub-strate,andalsooffersagreaternumberofbindingsitesforgastricandpancreaticlipases,thatmightexplainthelipolysisefciency.Furthermore,theimpactispossiblymorevisibleduringthegastricdigestivestepbecausegastriclipaseactivityplateauedwhenaconcentrationof128e164 mmolesprotonatedlong-chainFFAperm2ofavailablelipidsurfaceareaisreached(Armandetal.,2004,1999),generatingparticlesonthesurfaceoftheglobulesthattrapgastriclipaseandpreventitsaccesstosubstrate(Pafumietal.,2002).Thus,foragivenamountofmilklipid,ahighersurfaceareawoulddelaythe inhibition phenomenonallowingtheaccumu-lationofmoreFFAmoleculesinthephospholipids-containingMFGMandwillleadtoahigherlipolysisextent.Proofsofsuchhypothesiscouldbeobtainedbyconductingkineticsstudiesoflipolysis.Whenconsideringhomogenised heat-treatedmilk,thefatglobulesizeisnolongerthemainphysicochemicalparametergoverninglipolysisextent.Infact,small-sizedfatglobulesfromhomogenisedmilkareless(gastricandsoleduodenalphases)orindentically(gastro-duodenalphase)lipolysedthansmall-sizednativefatglobules(SFG)inspiteoftheir5e8foldlargersize,andarelipolysedatasameextentthanwholenativemilkfatglobulesthatare12e20foldlarger(inalldigestiveconditions).Thisisinaccordancewithrecentliterature,sincethe invitro digestionrateofTGfromhomogenisedmilkwasshowntobeonly1.6e2foldhigherthanthosefromnativemilkbyrecombinanthumanpancreaticlipaseinpresenceofpurebilesaltsinanon-physiologicalcontext(Bertonetal.,2012;Berton,Sebban-Kreuzer,Rouvellac,Lopez,&Crenon2009).Furthermore,therateandthe nalextentoflipolysiswerecomparablebetweenrecombinanthomogenisedmilkandnativemilkinpresenceofpancreatin(wholepancreaticenzymes)andwholebileextract(Ye,Cui,&Singh,2010),asherein.Severalexplanationscanbeproposed.Homogenisationstepreducestremendouslythefatglobulesize,highlydamagestheMFGMrichinpolarlipidsandleadtotheadsorptionofmilkproteinsattheinterfaceasindicatedbythehighincreaseofthezetapotential(Bertonetal.,2012;Michalskietal.,2002).FatglobulesurfacebecomespartiallycomposedbyresidualfractionofMFGMandmostlybycaseinmicellesandwheyproteins(Lopez,2005;Michalski&Januel,2006;Ongetal.,2010)thatkeepTGdispersedasverysmall-sizedglobules.However,thebenecialimpactofincreasinglipidinterfaceareaby20foldonlipasesactivityisprobablyminoredbythemajorchangesoccurringatthesurfaceofFig.5. CalculatedratiobetweenindividualFFAgeneratedduringlipolysisandthecorrespondingFAininitialtriglyceridesfornativeLarge-sized(LFG)orSmall-sized(SFG)fat-globulefractions,andfornative(NM),homogenised(HM),homogenised-UHT(HTM)milksamples.Theresultsrepresentmeans SEMof n 5(LFG,SFG,NM,HM),andn 4(HTM).Foragivenfattyacid,barswithdifferentletters(a,b)aresignicantlydifferent(P < 0.05).C.Garciaetal./FoodHydrocolloids35(2014)494e504 500thefatglobulewherethelipasesmustadsorbtocomeintocloseproximitytotheTGsubstratecontainedwithintheglobules.Indeed,i)gastriclipaseishighlysensitivetothepresenceandtheclassofphospholipidsatthesurfacetobefullyactive(Favetal.,2007),ii)humannativemilkfatglobulesaremoredigestedinthestomachofprematurenewbornthanthetinyprocessedglobulescontainedininfantformula(Armand,Hamosh,etal.,1996)possiblyduetotheuniquephospholipid ngerprintoftheirsurface(Garciaetal.,2012),andiii)pancreaticcolipasethathelpsclassicpancreaticlipasetobindatthesurfaceexhibitsahighafnitytosurfacecoveredbyphospholipids,bilesaltsandFFA(Sugar,Mizuno,Momsen,&Brockman,2001;Wickham,Wilde,&Fillery-Travis,2002).Wethusprovidesupplementaryargumentsonthefactthatthephospholipid-richMFGMisahighlyefcientinterfacefordigestivelipolysis(Armand,Hamosh,etal.,1996;Bertonetal.,2012;Yeetal.,2010).Adirectproofofevidencecouldbeob-tainedbytestinghomogenisedmilkproducedwithaddingacon-troledamountofdairypolarlipidconcentratewithwell-denedphospholipidandsphingolipids(particularlytheamountofsphingomyelin)classes.Also,wheyproteinsandcaseinmicellesreplacingMFGM,mayformamorecompactadsorbedlayeraroundtheTGcorethatislesspenetrableforlipases.Additionally,thewholedairyproteinnetworkformedatacidicpH(gel)mightdelaygastriclipolysisbylimitingtheaccessofthegastriclipasetotheinterface.Moreover,surface-activecomponentspresentinhumangastricjuicecanmodifythefatglobuleorganisationbydisplacingFig.6. Fattyacidcompositionofthefreefattyacids(A)andsn2-monoglycerides(B)generatedduringthegastro-duodenallipolysisstepsofnativeLarge-sized(LFG)andSmall-sized(SFG)fat-globulefractions,andofnative(NM),homogenised(HM),andhomogenised-UHT(HTM)milksamplesinconditionsmimickinghumanphysiology.Theresultsrepresentmeans SEMof n 5(LFG,SFG,NMandHM)and n 4(HTM).Foragivenfattyacid,barswithdifferentletters(a,b)aresignicantlydifferent(P < 0.05).C.Garciaetal./FoodHydrocolloids35(2014)494e504 501someoftheoriginalmoleculesfromthefatglobulesurface,orbyinterpenetratingbetweentheexistingmolecules,orbyforminganinterfacialcomplexwithseveralotherlayersadsorbingontopoftheexistingsurface-activemoleculesattheinterface(GoldingandWooster,2010).Furthermore,structuralreorganisationsoftheinterfaceuponlipolysisbyhumangastriclipase,asherein,anduponproteolysisofproteinsbyhumangastricpepsin,presenthereinandasshowedbyothers(Gallieretal.,2012),leadtophysicalunstabilityoffatglobulesbothfromwholenativeandfromhomogenisedmilkexplainingtheobservedcoalescence.Induodenalstep,surface-activecomponentspresentinbile(bilesaltsandphospholipids)areabletodisplacesomeofthedairyproteinsfromthehomogenisedglobulesurfaceandpartoftheMFGMfromthenativeglobulesurface,andtherebyallowedthecolipase-pancreaticlipasesystemtoadsorbintocloserproximitytothelipidswhatevertheinitialsurfacecomposition(Mun,Decker,&McClements,2007;Yeetal.,2010).Attheendallchangesmayleadtoastandardisationintermsofmacro-andultra-structureofwholenativeandhomogenised heat-treatedcowmilkexplainingthecomparabledigestibilityofTG.WhateverthedynamicsoftheinterfacialprocessesoccurringduringdigestionoffatglobulesfromnativeorhomogenisedUHT-treatedwholemilk,theoverallleveloffatdigestionreachedaround9.5%after1-hofgastricstep,thattypicallyoccursinthestomachofhealthysubjectsforfatglobulessizing60e10 mm(Armand,Borel,etal.,1996,Armandetal.,1999),andwasabout60%ingastro-duodenalconditionsi.e.closetothemaximalphysiolog-icallevel,whichis66%(2/3esterbondsfromTGproducingtwoFFAandone sn -2MGpereachTGmolecule).Ithastobenoticedthatthemaximumextentofmilklipolysisbypancreaticlipasesisreachedquicker(30min)thaninotherworks(Bertonetal.,2012,2009;Yeetal.,2010)conductedinbiochemicalconditions,duetothefactthatoursystemisclosertophysiologicalratiobetweenlipidsandlipases.Gastriclipolysisstepisknowntobenecessaryforduodenallipolysisbyclassicpancreaticlipasetoproceed(Bernbck,Blckberg,&Hernell,1990).Thus,thefactthathereinduodenalstepalonewasasmuchasefcient,orevenhigherinsomecases,thangastro-duodenalstepsraisesthequestionoftheexactroleofthegastricstep.However,amixtureofpancreaticenzymes(classicpancreaticlipaseandcolipase,probablypancre-aticlipaserelatedproteins,phospholipase,carboxylesterlipase)wasprovidedbythepancreatin,usedtomimichumanpancreaticsecretion,anditishighlypossiblethatcarboxylesterlipase,inpresenceofbile(8mMbilesaltsconcentration),wasabletohydrolyseTGreleasingenoughamountofFFAtohelplipolysisbyclassicpancreaticlipasetoproceed(Lindstrm,Sternby,&Borgstrm,1988).Furthermore,weobservedhereinthatthegastricstepseemstoexertaspeciccontrolontheextentofduodenallipiddigestionespeciallyincaseswhenlipolysiscanundergotoohighlevelsi.e.leadingtotheproductionofFFAandglycerol(so3FFApereachTGmolecule)insteadofFFAand sn -2MG,thelatterbeingmorefavourableforlipidsynthesisbytheintestine(Phan&Tso,2006).Inaddition,someoftheFFA(caprylic,myristic,stearicandoleic)andthe sn -2MGreleasedduringthegastricdigestionstepareknowntoexertanti-microbialeffectandprotectthedigestivetractagainstinfections(Hamoshetal.,1999).ThespeciesofFFAreleasedduringbothgastricandduodenalphaseareconsistentwiththeTGstructureofthecowmilk.But,interestingly,theyseemalsolinkedtothesizeofthefatglobulesinthegastricstepsuggestingthatthehumangastriclipasedonotaccessthesamemixedTGmoleculesinthesamewaypossiblybecausetheyorganiseddifferentlywithinthefatglobulecore(cleardistinctionforfatglobulesofsmallersizesi.e.SFG,HM,HMTversusfatglobulesoflargersizesi.e.LFGandNM).Theresultingdifference,i.e.ahigherreleaseofoleicacidforthesmallestfatglobulesversusahigherreleaseofstearicacidforthelargestfatFig.7. Confocalmicrographsofnativemilk(A)andhomogenisedmilk(B,C),after1hincubationtimeatpH4and37 C,andafteradditionofhumangastricjuicefor1hat37 C(D:nativemilk,EandF:homogenisedmilk).Triglycerideswerelabelledusingnilered(orangecolour),milkproteinswerelabelledwithacridineorange(greencolour).A,B,D,E:scalebar 20 mm;C,F:scalebar 10 mm.(Forinterpretationofthereferencestocolourinthis gurelegend,thereaderisreferredtothewebversionofthisarticle.)C.Garciaetal./FoodHydrocolloids35(2014)494e504 502globules,mightbeofpotentialinterestforcontrollinggastricemptyingrateandthussatietyinhumanssinceoleicacidstimu-lateshighlythesecretionofCCKwhilestearicacidisapoorstim-ulus(Beradshalletal.,1989;Stewartetal.,2011).Thegastriclipaseisabletoreleasethefattyacidsatthe sn -3position,thatiscurrentlyknown(Armand,2007;Favetal.,2004),butnotsopreferentiallywhenconsideringthemixedTGofcowmilkasindicatedbyalowreleaseoffattyacids(C4:0,C6:0,C8:0andC10:0)mainlypositionedat sn -3(forabout92e78%oftheirtotalamounts)onthetriglyceridesbackbone(Blasietal.,2008).Inter-estingly,thehumangastriclipaseactsalsoatthe sn -1positionasitgeneratedmainlyplamiticandstearicacids(30e35%and35e50%oftotalFFAreleased,respectively)thatarereportedtobelocatedatsn -1for45.5%and52.6%oftotalrespectiveamountscomparedto11.6%and23.2%at sn -3(Blasietal.,2008).Whenthegastro-duodenaldigestioniscompleted,itappearsthattheSFGfractioncouldbeofspecicbenecialhealthinterestduetothecomposi-tionofthe sn -2MGgeneratedthatarericherinoleicacidandsobeingnoatherogeniccomparedto sn -2MGenrichedinsaturatedfattyacids(Favetal.,2004).5.ConclusionContrarytotheexistingparadigm,technologicalprocessessuchashomogenisation UHT-treatmentdonotimprovecowmilkfatdigestibilityascomparedtonativemilk.Microltrationmethodcouldbeusefulfordesigningfuturedairyproductsthatcouldbemore(small-sizedfatglobules)orless(large-sizedfatglobules)digestible.Thisrepresentspotentialinterestforsubjectswitheitherphysiological(infants,elderly)orpathological(pancreatitisandcystic brosis)pancreaticinsufciency,orneedingweightman-agement,respectively.Butthispointneedstobeconrmedbyhumanstudiesandinacontextofcomplexmealsandduration.Thegastriccompartmentrepresentsamoresensitivetargetintermsofphysiologicalsitetoactontoregulatethelipolysisextent,andpossiblyforcontrollingsatiety.Fatglobulesizeisnotthesoleimportantparametertomodulatefattyacidbioavailabilitythroughthedigestivestepsbecauseothersarealsoinvolved,suchasphospholipidsandproteinscompositionandexposurearoundtheglobulesurface.Theselectionofspecictypeandamountsofphospholipidsand/orproteinsfrommilkoriginappearsofhighinteresttocreatenewmatrices,includinginfantformulae,havingawell-controlleddigestionrate.ConictofinterestTheauthorshavedeclarednoconictofinterest.AcknowledgementsThisresearchwassupportedbytheFrenchNationalResearchAgency(ANR-AgenceNationaledelaRecherche)undertheProgrammeNationaldeRechercheenAlimentationetNutritionHumaine,andARILAITRecherches/CNIEL,projectAGILAIT ANR-06-PNRA-012 (2007-2010),theFrenchNationaleInstituteofHealth(INSERM),andRgionPACAetApplicationSantdesLipides(ASL,Hauterive)forCGthesis.ThelegalityofallauthorswascarefullycheckedaccordingtotheethicalprocedurepublishedbyAviesan,i.e.wecertifythatCG,CA,CLandMAhavecontributedsubstantiallytotheworkandhaveparticipatedinthewritingofthepaper.BRisco-authorofthepaperbecausehewasinchargeofthepreparationofthemilksamplesbyusingauniqueprocedurepatentedbyINRA.TheauthorsthankValrieBriard-Bion,EricBeaucherandFlor-enceRousseau(INRASTLO,Rennes,France)forthepreparationandcharacterisationofmilksamples,DrJacquesPeyrot(Gastroenter-ologyandhepatologydivision,HpitalNord,Marseille,France)forgenerousgiftofhumangastricjuice.VictorPirisiandDrIsabelleCrenon(UMRInsermU476/Inra1260/UniversitdelaMdi-terrane,FacultdeMdecineTimone,Marseille,France)areacknowledgedfor,respectively,GCmaintenanceandcoordinationofworkpackage4inAGILAITProject.AppendixA.SupplementarydataSupplementarydatarelatedtothisarticlecanbefoundat http://dx.doi.org/10.1016/j.foodhyd.2013.07.005.ReferencesAiqian,Y.E.,Singh,H.,Taylor,M.W.,&Anema,S.(2004).Interactionsofwheyproteinswithmilkfatglobulemembraneproteinsduringheattreatmentofwholemilk. Lait,84,269e283.Aldai,N.,Murray,B.E.,Najera,A.I.,Troy,D.J.,&Osoro,K.(2005).Derivatizationoffattyacidsanditsapplicationforconjugatedlinoleicacidstudiesinruminantmeatlipids. JournaloftheScienceofFoodandAgriculture,85,1073e1083.Armand,M.(2007).Lipasesandlipolysisinthehumandigestivetract:wheredowestand? CurrentOpinioninClinicalNutritionandMetabolicCare,10,156e164.Armand,M.,Borel,P.,Pasquier,B.,Dubois,C.,Senft,M.,Andr,M.,etal.(1996).Physicochemicalcharacteristicsofemulsionsduringfatdigestioninthehumanstomachandduodenum. AmericanJournalofPhysiology,34,G172eG183.Armand,M.,Hamosh,M.,Dipalma,J.S.,Gallagher,J.,Benjamin,S.B.,Philpott,J.R.,etal.(1995).Dietaryfatmodulatesgastriclipaseactivityinhealthyhumans.AmericanJournalofClinicalNutrition,62,74e80.Armand,M.,Hamosh,M.,Mehta,N.R.,Angelus,P.A.,Philpott,J.R.,Henderson,T.R.,etal.(1996).Effectofhumanmilkorformulaongastricfunctionandfatdigestionintheprematureinfant. PediatricResearch,40,429e437.Armand,M.,Hamosh,M.,Philpott,J.R.,Kovar,A.,Rosenstein,B.J.,Hamosh,A.,etal.(2004).Gastricfunctioninchildrenwithcystic brosis:effectofdietongastriclipaselevelsandfatdigestion. PediatricResearch,55,457e465.Armand,M.,Pasquier,B.,Andr,M.,Borel,P.,Senft,M.,Peyrot,J.,etal.(1999).Digestionandabsorptionof2fatemulsionswithdifferentdropletsizesinthehumandigestivetract. AmericanJournalofClinicalNutrition,70,1096e1106.Astorg,P.,Arnault,N.,Czernichow,S.,Noisette,N.,Galan,P.,&Hercberg,S.(2004).Dietaryintakesandfoodsourcesofn-6andn-3PUFAinFrenchadultmenandwomen. Lipids,39,527e535.Beardshall,K.,Frost,G.,Morarji,Y.,Domin,J.,Bloom,S.R.,&Calam,J.(1989).Saturationoffatandcholecystokininrelease:implicationsforpancreaticcarcinogenesis. Lancet,2(8670),1008e1010.Bernbck,S.,Blckberg,L.,&Hernell,O.(1990).Thecompletedigestionofhumanmilktriacylglycerol invitro requiresgastriclipase,pancreaticcolipase-dependentlipase,andbilesalt-stimulatedlipase. JournalofClinicalInvestiga-tion,85,1221e1226.Berton,A.,Rouvellac,S.,Robert,B.,Rousseau,F.,Lopez,C.,&Crenon,I.(2012).Effectofthesizeandinterfacecompositionofmilkfatglobulesontheirinvitrodigestionbythehumanpancreaticlipase:nativeversushomogenizedmilkfatglobules. FoodHydrocolloids,29,123e134.Berton,A.,Sebban-Kreuzer,C.,Rouvellac,S.,Lopez,C.,&Crenon,I.(2009).Individualandcombinedactionofpancreaticlipaseandpancreaticlipase-relatedproteins1and2onnativeversushomogenizedmilkfatglobules. MolecularNutritionandFoodResearch,53, 1592e1602.Blasi,F.,Montesano,D.,DeAngelis,M.,Maurizi,A.,Ventura,F.,Cossignani,L.,etal.(2008).Resultsofstereospecicanalysisoftriacylglycerolfractionfromdonkey,cow,ewe,goatandbuffalomilk. JournalofFoodCompositionandAnalysis,21,1e7.Carrire,F.,Barrowman,J.A.,Verger,R.,&Laugier,R.(1993).Secretionandcontributiontolipolysisofgastricandpancreaticlipasesduringatestmealinhumans. Gastroenterology,105,876e888.Christopherson,S.W.,&Glass,R.L.(1969).Preparationofmilkfatmethylestersbyalcoholysisinanessentiallynonalcoholicsolution. JournalofDairyScience,55,1289e1290.Fauquant,C.,Briard,V.,Leconte,N.,&Michalski,M.C.(2005).Differentlysizednativemilkfatglobulesseparatedbymicroltration:fattyacidcompositionofthemilkfatglobulemembraneandtriglyceridecore. EuropeanJournalofLipidScienceandTechnolology,107,80e86.Fav,G.,Coste,T.C.,&Armand,M.(2004).Physicochemicalpropertiesoflipids:newstrategiestomanagefattyacidbioavailability. CellularandMolecularBiology,50,815e883.Fav,G.,Lvque,C.,Peyrot,J.,Pieroni,G.,Coste,T.C.,&Armand,M.(2007).Modulationofgastriclipolysisbythephospholipidspecie:linktospeciclipase-phospholipidinteractionatthelipid/waterinterface? FASEBJournal,21.A1010-a.Gallier,S.,Ye,A.,&Singh,H.(2012).Structuralchangesofbovinemilkfatglobulesduringinvitrodigestion. JournalofDairyScience,95,3579e3592.C.Garciaetal./FoodHydrocolloids35(2014)494e504 503Garcia,C.,Lutz,N.W.,Confort-Gouny,S.,Cozzone,P.J.,Armand,M.,&Bernard,M.(2012).Phospholipid ngerprintsofmulkfromdifferentmammaliansdeter-minedby31PNMR:towardsspecicinterestinhumanhealth. FoodChemistry,135,1777e1783.Golding,M.,&Wooster,T.J.(2010).Theinuenceofemulsionstructureandstabilityonlipiddigestion. CurrentOpinioninColloid&InterfaceScience,15,90e101.Hamosh,M.,Peterson,J.A.,Henderson,T.R.,Scallan,C.D.,Kiwan,R.,Ceriani,R.L.,etal.(1999).Protectivefunctionofhumanmilk:themilkfatglobule. SeminarsinPerinatology,23,242e249.Keenan,T.W.,&Patton,S.(1995).Thestructureofmilk:implicationforsamplingandstorage.A.Themilklipidglobulemembrane.InR.G.Jensen(Ed.), Handbookofmilkcomposition (pp.5e50).California:AcademicPressInc.Lee,S.J.E.,&Sherbon,J.W.(2002).Chemicalchangesinbovinemilkfatglobulemembranecausedbyheattreatmentandhomogenizationofwholemilk.JournalofDairyResearch,69,555e567.Lindstrm,M.B.,Sternby,B.,&Borgstrm,B.(1988).Concertedactionofhumancarboxylesterlipaseandpancreaticlipaseduringlipiddigestioninvitro:importanceofthephysicochemicalstateofthesubstrate. BiochimicaBiophysicaActa,959,178e184.Lopez,C.(2005).Focusonthesupramolecularstructureofmilkfatindairyprod-ucts. ReproductionNutritionDevelopment,45,497e511.Lopez,C.(2011).Milkfatglobulesenvelopedbytheirbiologicalmembrane:uniquecolloidalassemblieswithaspeciccompositionandstructure. CurrentOpinioninColloid&InterfaceScience,16,391e404.Lopez,C.,Briard-Bion,V.,Mnard,O.,Beaucher,E.,Rousseau,F.,Fauquant,G.,etal.(2011).Fatglobulesselectedfromwholemilkaccordingtotheirsize:differentcompositionsandstructureofthebiomembrane,revealingsphingomyelin-richdomains. FoodChemistry,125,355e368.Michalski,M.C.,Cariou,R.,Michel,F.,&Garnier,C.(2002).Nativevs.damagedmilkfatglobules:membranepropertiesaffecttheviscoelasticityofmilkgels. JournalofDairyScience,85,2451e2461.Michalski,M.C.,&Januel,C.(2006).Doeshomogenizationaffectthehumanhealthpropertiesofcowsmilk? TrendsinFoodScience&Technology,17,423e437.Mun,S.,Decker,E.A.,&McClements,D.J.(2007).Inuenceofemulsiertypeonin vitrodigestibilityoflipiddropletsbypancreaticlipase. FoodResearchInter-national,40,770e781.Ong,L.,Dagastine,R.R.,Kentish,S.E.,&Gras,S.L.(2010).Theeffectofmilkprocessingonthemicrostructureofthemilkfatglobuleandrennetinducedgelobservedusingconfocallaserscanningmicroscopy. JournalofFoodScience,75,E135eE145.Pafumi,Y.,Lairon,D.,LecheneDeLaPorte,P.,Juhel,C.,Storch,J.,Hamosh,M.,etal.(2002).Mechanismsofinhibitionoftriacylglycerolhydrolysisbyhumangastriclipase. JournalofBiologicalChemistry,277,28070e28079.Phan,C.T.,&Tso,P.(2006).Intestinallipidabsorptionandtransport. FrontiersinBioscience,6,D299eD319.Stewart,J.E.,Seimon,R.V.,Otto,B.,Keast,R.S.,Clifton,P.M.,&Feinle-Bisset,C.(2011).Markeddifferencesingustatoryandgastrointestinalsensitivitytooleicacidbe-tweenleanandobesemen. AmericanJournalofClinicalNutrition,93(4),703e711.Sugar,I.P.,Mizuno,N.K.,Momsen,M.M.,&Brockman,H.L.(2001).Lipidlateralorganizationin uidinterfacescontrolstherateofcolipaseassociation. Bio-physicalJournal,81,3387e3397.Wickham,M.,Wilde,P.,&Fillery-Travis,A.(2002).Aphysicochemicalinvestigationoftwophosphatidylcholine/bilesaltinterfaces:implicationsforlipaseactiva-tion. BiochimicaBiophysicaActa,1580,110e122.Ye,A.,Cui,J.,&Singh,H.(2010).Effectofthefatglobulemembraneoninvitrodigestionofmilkfatglobuleswithpancreaticlipase. InternationalDairyJournal,20,822e829.C.Garciaetal./FoodHydrocolloids35(2014)494e504 504