surrounding ants Oxidación de

FACULTAD DE FARMACIA Sterols oxidation: effect of heating, unsaturation degree of the surrounding lipids and presence of antioxidaa ants Oxidación de esteroles: efecto del calentamiento, grado de insaturación de la matriz lipídica y presencia de antioxidantes Blanca Barriuso Estebann Pamplona, Juniode 2015

FACULTA AD DE FARMACIAA Sterols oxidation: effect of heating, unsaturation degree of the surrounding lipids and presence of antioxidaa ants Oxidación de esteroles: efecto del calentamiento, grado de insaturación de la matriz lipídica y presencia de antioxidantes Memoria presentadaa por Dña. Blanca Barriuso Esteban para aspirar al grado de Doctor por la Universidad de Navarra El presente trabajo ha sido realizado bajo la direcciónde la Dra. Diana Ansorena Artieda y la Dra. Iciar Astiasarán Anchía en el Departamento de Ciencias de la Alimentación y Fisiología y autorizamos su presentación ante ell Tribunal que lo ha de juzgar. En E Pamplona, Junio de 2015 Dra. Diana Ansorena Artiedaa Dra. Iciar Astiasarán Anchía

Facultad de Farmacia Departamento de Ciencias dee la Alimentación y Fisiología La directora del Departamento, Dra. Diana Ansorena Artieda hace h constarr que el presente trabajo de investigación ha sido realizado por Dña. Blanca Barriuso Esteban, en el Departamento de Ciencias de la Alimentación y Fisiología de laa Facultad de Farmacia de la Universidad de Navarra. Dra. Diana Ansorena Artieda En Pamplona, Junio de 2015

AcknowledgementsAgradecimientos En primer lugar, me gustaría agradecer a mis directoras de tesis, Dra. Ansorena y Dra. Astiasarán,sudedicación,interésyconfianza.Estoymuyagradecidaporlaoportunidadque mebrindarondeunirmeasugrupodeinvestigación. AlaUniversidaddeNavarra,porlosmediostécnicosyhumanosproporcionadosparami formacióncomoinvestigadora. AlaAsociacióndeAmigosdelaUniversidad,porlaayudaeconómicaduranteestoscuatro años. Al Plan Investigador Universidad de Navarra (PIUNA), por la contribución económica al proyecto AlBancoSantander,porlaconcesióndelabecademovilidadpararealizarunapartedemi investigaciónencampinas(brasil). ToDrBragagnolo,forgivingmetheopportunitytoworkinherlaboratoryforfourmonths. ToallthemembersoftheDepartmentofFoodScienceoftheUniversityofCampinas(Brazil), especiallytolilianmariutti,hugosouzaandnoraneidheart,fortheiradviceandassistance duringmystayincampinas. Atodoslosquehancontribuidoaestetrabajo,especialmenteaÍñigoNavarroyAlfredoGea. Alostécnicosdelaboratorio,LuisJáureguiyGwenaëlleCeniceros,porsuasistenciatécnica. Atodasmiscompañerasdeldepartamento,poreltrabajoyporeldescansocompartidos.Esta tesisnohabríasidoposiblesinellas. Amifamiliayamigos,porsucomprensiónyapoyoincondicional.

Amispadresyhermano

Abstract Dietarysterolsarenutritionallyinterestingcompoundswhichcanundergooxidationreactions duringfoodmanufactureandstorage,aswellasinthehumanbody.theiroxidationproducts areassociatedwiththedevelopmentofhighlyprevalentnoninfectiousdiseases.therefore,it isrelevanttoevaluatetheparticularfactorswhichaffectsteroldegradationandoxysterols formationinfoods. Inthiscontext,thepresentworkaimedtoassesstheeffectofheating treatment,unsaturationdegreeofthesurroundinglipidsandalsopresenceofantioxidants,on sterolsdegradationandoxidesformation.modelsystems(whicharevaluabletoolstoavoid the influence of interferences), as well as some food applications, were used in the experimentaldesigns. Ourstudyconcludedthat:sterolsthermooxidationisamultifactorialprocesswhichstrongly dependsontimetemperaturecombination,producingahighsteroloxidationalreadyfromthe beginningoftheprocess.moreover,thepresenceandunsaturationdegreeofthelipidmatrix, aswellasthepresenceofphenolicsandtocopherols,significantlyprotectedsterolsfrom oxidationinourmodelsystems.theadditionofplantextractsinfoodstuffstoachievethis samegoalappearedtobeapromisingstrategywhenthesensoryaspectsandcharacteristicsof thesamplearetakenintoaccount.inaddition,wecouldestablishthatthemonitoringofsterol oxidationthroughthemeasurementoftheoxidesgeneratedisacomplexissue,hencea scientificconsensustoachieveastandardizedmethodologyisstillneeded.

Resumen Losesterolesdietéticossoncompuestosinteresantesdesdeelpuntodevistanutricional,que pueden sufrir reacciones de oxidación durante el procesado y el almacenamiento de alimentos,asícomoenelorganismo.susproductosdeoxidaciónestánrelacionadosconel desarrollodeenfermedadesnoinfecciosasdealtaprevalencia.porlotanto,esimportante estudiarlosfactoresqueafectanaladegradacióndeesterolesyalaformacióndeoxiesteroles enlosalimentos.enestecontexto,elpresentetrabajotratódedeterminarelefectodel tratamientotérmico,delgradodeinsaturacióndelamatrizlipídicaydelapresenciade antioxidantes, en la degradación de esteroles y la formación de óxidos. En los diseños experimentales se emplearon sistemas modelo (muy útiles para evitar la influencia de interferentes),ysellevaronacaboalgunasaplicacionesenalimentos. Nuestroestudioconcluyóque:latermooxidacióndeesterolesesunprocesomultifactorialque depende en gran medida de la combinación tiempotemperatura, produciendo una alta oxidacióndeesterolesdesdeeliniciodelcalentamiento.además,lapresenciaygradode insaturacióndelamatrizlipídica,asícomolapresenciadefenólicosytocoferoles,protegió significativamente a los esteroles de la oxidación en nuestros sistemas modelo. La incorporación de extractos de plantas en alimentos para conseguir este mismo objetivo, resultóserunaestrategiaprometedorasisetienenencuentalosaspectossensorialesylas características del alimento. Además, se constató que el seguimiento de la oxidación de esterolespormediodeladeterminacióndelosóxidosgeneradosesunacuestióncompleja. Portanto,seríaesencialunconsensocientíficoparaalcanzarunametodologíaestandarizada.

Index

Index INTRODUCTION...1 1.Lipids...3 2.Sterols...5 2.1Chemicalstructureandproperties...5 2.2Presenceinfoods...6 2.3Effectsintheorganism...7 3.Steroloxidationproducts...9 3.1Chemicalstructureandproperties...9 3.2Presenceinfoods...10 3.3Effectsintheorganism...10 4.Steroloxidationprocess...13 4.1Routesofoxysterolsformation...13 4.2Relevanceofexogenousformation...15 4.3Modelsystemsasausefulexperimentaltool...18 JUSTIFICATIONANDOBJECTIVES...19 EXPERIMENTALDESIGN...23 MATERIALANDMETHODS...27 1.Samplespreparation...29 1.1Modelsystems...29 1.2Beefpatties...29 1.3Tunapatties...30 2.Moisturedetermination...30 3.Lipidextraction...30 3.1Quantitativedeterminationbysoxhletextraction...30 3.2Quantitativedeterminationbychloroform:methanolextraction...30 3.3Qualitativedeterminationbychloroform:methanolextraction...31 4.Fattyacidsdetermination...31 5.Sterolsdetermination...33 6.SOPsdetermination...38 7.Otheroxidationparameters...43 7.1TBARS...43 7.2PV...43 7.3Hexanalcontent...43

8.Antioxidantcapacityandcompounds...44 8.1Totalphenoliccompounds...44 8.2ORAC...44 8.3Rosmarinicacid...45 8.4VitaminE...45 8.55caffeoylquinicacidandotherphenoliccompounds...45 9.Sensorialanalysis...46 10.Statisticalanalysis...46 RESULTS...47 ResultsI.Paper1: Areviewofanalyticalmethodsmeasuringlipidoxidationstatusin foods:achallengingtask...49 ResultsII.Poster1: Determinationofcholesteroloxidationproductsinfoods: improvementofcosttimeefficiency...73 ResultsIII.Paper2: Interlaboratoryharmonizationtrial...77 ResultsIV.Paper3: Sterolsheating:Degradationandformationoftheirringstructure polaroxidationproducts...83 ResultsV.Paper4: RoleofMelissaofficinalisincholesteroloxidation:Antioxidanteffect inmodelsystemsandapplicationinbeefpatties...99 ResultsVI.Paper5: Solanumsessiliflorum(manacubiu)antioxidantprotectiveeffect towardscholesteroloxidation:influenceofdocosahexaenoicacid!...119 ResultsVII.Poster2: ProtectiveeffectofaSolanumsessiliflorum(manacubiu)extract intunapatties...131 ResultsVIII.Paper6: Cholesterolandstigmasterolwithinasunfloweroilmatrix: Thermaldegradationandoxysterolsformation...135 ResultsIX.Paper7: Unsaturatedlipidmatricesprotectplantsterolsfromdegradation duringheatingtreatment...149 GENERALDISCUSSION...167 CONCLUSIONS...183 LISTOFABBREVIATIONS...187 REFERENCES...191 DISSEMINATIONOFRESULTS...221

Introduction

Introduction 1.LIPIDS Lipids are defined as substances insoluble in water yet soluble in organic solvents. They comprise a wide variety of compounds, namely waxes, fatty acids (and their glycerides), phospholipids, sphingolipids, tocopherols and steroids, among others. Most of them are importantcomponentsofanimalandplantlivingcells,wheretheirmainbiologicalfunctions includeenergystorage,maintainingthestructureofcellmembranesandcellsignaling. Inrelationtotheseimportantbiologicalfunctions,lipidscontainedinfoodsplayakeyrolein humannutritionandhealth. Amongfattyacids,anintakewhichisrichinsaturatedfattyacids(SFA)hasbeenwidely associatedwithharmfuleffectsintheorganism,mostlycardiovasculardiseases(cvd)(krauss etal.,2000).bycontrast,adietrichinmonounsaturatedfattyacids(mufa)promotesa healthylipidbloodprofile,improvesbloodpressure,modulatesthesensitivitytoinsulinand glycemiclevelsandcontributestopreventobesity,henceimprovingthemetabolicsyndrome andreducingtheriskofcvd(ros,2003;gillinghametal.,2011).oleicacid,asthemain representativeofthiskindoffattyacids,isthusresponsibleofmostofthebeneficialeffects attributed to olive oil consumption, characteristic of the Mediterranean Diet. Regarding omega3fattyacids(polyunsaturatedfattyacidspresentingthefirstdoublebondatposition3 fromtheendofthecarbonchain),havebeenshowntoreduceplasmatriglycerides,heartrate andarterialpressure,aswellaspromotingantiinflammatoryeffectsandimprovingmental disorders, among other benefits (De Henauw et al., 2007; SánchezVillegas et al., 2007; MozaffarianandWu,2011).Themainomega3PUFAfromplantoriginislinolenicacid,and thosemostrepresentativeofmarineoriginareeicosapentaenoicacidanddocosahexaenoic acid,bothcharacteristicofalgaeandfish. Dietaryhydrophobicvitaminsalsohaveabeneficialimpactonhumanhealth.First,vitaminA (andcarotene,asprovitamina)consumptionpreventsvisionproblemsandkeratinizationof the mucosae of several organic systems (Combs, 2001). Vegetables such as carrots and pumpkins,aswellasliverorfisharethemainsourcesofvitamina.inaddition,theintakeof vitamindcontributestothepreventionofbonediseasesbyincreasingcalciumabsorption, CVD,multiplesclerosis,diabetesandcancer(Zittermann,2003;BischoffFerrarietal.,2006; Viethetal.,2007).Fattyfish,liveranddairyproductsarethemainfoodsourcesofvitaminD, apartfromsolarexposure.moreover,vitamineconsumptionhasbeenreportedtocounter theeffectsofeffectcvd(stampferetal.,1993;rimmetal.,1993),althoughitsantioxidant propertiesinhumansdonotseemtobeveryeffective(robertsetal.,2007;gazianoetal., 3

Introduction 2009).Nevertheless,anexcessiveintakeofvitaminsistoxicandcancausemoderatetosevere healthproblems(ochoaandmataix,2009;mataixanddelahiguera,2009). Steroidsareanothergroupoffoodlipidsthathasimportantbiologicaleffectsinhumans. Amongthisgroup,cholesterolisessentialformaintainingmembranefluidity,althoughan excessiveintakeisagainharmful,increasingtheriskofcvd.ontheotherhand,phytosterols which do not perform any specific function in human organisms reduce cholesterol absorptionintheintestine,decreasingplasmaldllevels(seesection2.3). Allthesebiologicalbenefits,togetherwiththeincreaseinpublicinterestinhealthissues,have promotedthedevelopmentoflipidenrichedfunctionalproducts.inthissense,formulations increasingunsaturatedfatsindetrimentofsfaarecommonlyappliedinmeatanddairy products(berasategietal.,2011;rodríguezcarpenaetal.,2012a).hydrophobicvitaminsare alsousuallyaddedintodairyproductsandfruitbeverages(petrogiannietal.,2013;delavariet al.,2015)andthenumberofphytosterolenrichedproductshavealsoseenasharpincrease overthelastdecade(seesection2.2). However,lipidscanundergooxidationreactionsunderfavorableconditionscharacterizedby oxygen (or any other oxidant) availability, incidence of light and temperature. This lipid oxidationcantakeplaceinfoodsbeforeconsumptionorwithintheorganism.regardlessof the environment where the oxidation takes place, the process itself is harmful not only becauseitdamagescellfunctionsbydestroyingthelipids,butalsobecauseitresultsinthe productionofhydroperoxides,alcoholsandcarbonylcompounds,whichhavebeenrelatedto cytotoxicityandmutagenicity(uchidaetal.,2000;delrioetal.,2005;gueraudetal.,2010; Otaeguietal.,2010),aswellasspoiledorganolepticquality.Severaltypesoflipids(fattyacids, phospholipids,vitaminsandsterols)havebeenreportedtobedegradedthroughoxidation reactionsandtoproducetheabovementionedtoxicsubstances. Differentapproacheshavebeenadoptedtoavoidorcounteractlipidoxidationbothinfood and in vivo. First, oxygen availability, light exposure and temperature are controlled in foodstuffsbymeansofappropriatepackagingconditions.second,thecontrolofcookingand otherprocessesarecriticalpoints.andlast,controloverprooxidantenvironmentsbothin foodandbiologicaltissuesareexercisedbymeansoftheadditionofantioxidantsinthe formulationandtheincorporationofantioxidantsinthediet,respectively.thesestrategiesare sometimesappliedtolimittheoxidationoflipidsingeneral,andoccasionallytolimitthe oxidationofaparticularkindoflipid,suchassterols. 4

Introduction Consideringthewidespreadpresenceoflipidsbothinfoodstuffsandinthehumanbody,a large number of research studies has as their aim, the determination of lipids and, in particular,thedeterminationoftheiroxidationproducts.however,thegreatdiversityof oxidizedcompoundsandthecomplexityofcertainbiologicalandfoodmatrices,aswellasthe possibilityofmultipletechnicalapproaches,makeitdifficulttoestablishuniversalmethodsfor determiningthestatusoflipidoxidation. Inparticular,inthecaseofsteroloxidationproducts(SOPs),itisworthhighlightingthattheir analysisislaboriousandexpensive,andhencetheoptimizationofthemethodologyremainsa centralobjective.timeandcostefficiencyarefactorstobeconsidered,withoutdisregardfor reliabilityofresearchresults.inthissense,theformationofartefacts(oxysterolsnotpresentin thesample)duringthelaboratorypreparationprocessisamajorconcernsincebothsterols and oxysterols are normally present in the samples and the former in relatively higher amounts.besides,therearesensitivitydifficultiesassociatedwiththedeterminationoftrace levels.amongthevariousmethodsavailabletoanalyzesops,thegeneralprocedureinvolves thefollowingsteps:lipidextraction,saponification,solidphaseextraction,derivatizationand chromatography. Furthermore, research groups have shown substantial variances in the executionofthesesteps.aninterlaboratoryharmonizationofthemethodologiesisanurgent issuesincecertainparameters(suchastemperature,oxygenexposureorcontactwithalkaline solutions)playacrucialroleinartefactgenerationandthesensitivityisnotablyaffectedby chromatographicconditions(griffithsetal.,2013,georgiouetal.,2014). 2.STEROLS 2.1Chemicalstructureandproperties Sterols are unsaponifiable lipids, whose chemical structure is characterized by a cyclopentanophenanthreneringwithahydroxylgroupinpositionc3andasidechaininc17, asillustratedinfigure1.particularsterolsdifferfromeachotherintermsofthesubstitutions ofthesidechain.cholesterolisthemainsterolofanimalorigin,incomparisontothemore than250plantsterols(usuallynamedasphytosterols)thathavebeenidentified.sitosterol, campesterolandstigmasterolarethemostabundantcompounds,representingmorethan 95%ofthewholephytosterolscontentinfood. The different substituents in the molecule provide a graduationofits hydrofobicity, and therefore,intheabsorptionpropertiesofthesterolinorganisms.therefore,phytosterolsare, ingeneral,lesspolarthancholesterol.sterolscanbefoundasfreeoresterifiedmolecules, withafattyacidattachedtotheirhydroxylgroup. 5

Introduc 6 2.2Pre Cholest mostfo varyde than40 esterifi recomm 15045 Phytost vegetab produc Gupta thoseo AWest 2004;K anycho awaren develop wasint ofmarg snackb tion Figure1.C esenceinfoo terolisacon oodsderived ependingon 00mg/100g edwith the mendsanin 0mgcholest terolsareth bles and ve ctsto70150 etal.,2011 ofcampester terndietpr Kuhlmannet olesterollow nessregardi pmentofph troducedinf garines,yog bars(kuhlma Chemicalstr ods nstituentof dfromanim nthefoodstu inanimalen efatty acids takeofless terolperday heplantste egetable oils 00mg/100g ). Sitoste rolandstigm ovidesbetw tal.,2005;e weringeffect nghealtha hytosterolen Finland,the hurts,milk,s annetal.,2 ucturesofch thecellmem alssuchasm uff,ranging ntrails(more softhefoo than300m y(andersson erolscounte s. Typical co ginvegetabl rolisgener masterol. ween100an Escurriolet ts(seesectio ndfunctiona nrichedfood varietyoffo saladdressin 2005;Ozeran holesterolan mbraneofa meat,eggs, fromlessth eirasetal.,20 odmatrix.w mg/dayofch netal.,2004 erparts,occu oncentration leoils(philli allypresent nd400mgp al.,2010).t on2.3)is23 alfoods,th dproducts.s oodstuffson ngs,vegetab ndkirmaci, ndsomecom nimaltissue fishanddai han10mg/1 011).Choles Whilethe A holesterol,th 4;Escurriolet urringnatur ns range fro psetal.,20 infoodat phytosterols Therequired 3g/day.Give elastdecad Since1995, offerhasgro bleoils,fruit 2010;Alema mmonphytos s,andhence ryproducts. 100ginfresh terolusually AmericanHe heeuropean tal.,2010). allyinfruits om 150 mg 002;Marango higherconc perday(an phytosterol entheincrea dehasseen whenthefi ownsignifica juices,bake anyetal.,20 sterols eitispresen Concentrat hmilk,tom ymanifestsit eartassocia nmeanintak s,nuts,cere g/100g in fr onietal.,20 centrationst nderssonet lintaketoh aseinconsu aconsidera irstsuchspr antlyinthef eryproducts 012a;Botelh ntin ions more tself tion keis eals, resh 010; than tal., have mer able read orm and oet

Introduction al.,2014;).theeuropeancommitteehasauthorizedthedistributionofmostoftheseproducts intheeuropeanmarket(eurlex,online).thisenrichmentinphytosterolshasenabledthe increase in their intake above the clinically important levels for its cholesterollowering propertiestotakeeffect.moreover,thefoodmatrix,thesterolformandtheirdistributionin severalservingsthroughouttheday,mayaffectthemagnitudeoftheldlreductionachieved (Cliftonetal.,2004;Guptaetal.,2011;Shaghaghietal.,2014). Phytosterols to be added as ingredients in enriched foods are generally extracted from byproductsfromwoodpulpinthepaperindustryorfromvegetableoils.plantstanolsare mainly produced by hydrogenation of plant sterols (Brufau 2008). They can be found in powder,microencapsulated,emulsifiedoresterified.astheyformcrystalsthatareinsolublein wateranddifficulttodisperseinfat(sharma2005),phytosterolsareusuallyaddedtofood productsintheiresterifiedform.thisprocessmakesthemmoresolubleindietaryfatand enhancestheirdispersionintheintestine,therebypromotingtheirefficacy(katanetal.,2003). However,hypocholesterolemiceffectivenesshasrecentlybeenfoundtobehigherforwater dispersablephytosterolsthanforesterifiedphytosterolsaddedtoyogurts(shaghaghietal., 2014).Ratiosofthephytosterolmixturewhichmustbeusedforenrichmentisdeterminedby legislation,limitingthemaximumamountforeachsterol(eurlex). Thedevelopmentofnewformulationsofenrichedproductshasbeenaccompaniedbythe consequent legislation regarding scientific substantiation of their health effects and their safety.in2009,theeuropeansafetyassociation(scientificopinion,efsaq200900530, EFSAQ200900718) accepted a claim related to plant sterols and lower/reduced blood cholesterolandreducedriskofheartdiseaseinphytosterolenrichedfoodproducts.thefood anddrugadministrationhadpreviously(2000)authorizedaclaimonthesameissue(federal RegistrationofSeptember8,200065FR54686).Asithappenswithmostfunctionalfoods,the properscientificvalidationoffunctionalclaimsstillremainsthecriticalissue. 2.3Effectsintheorganism Cholesterol Cholesterolinbloodandtissuescomesmainlyfromendogenousformationandtoalesser extentfromdietarycholesterol.consideringdietarycholesterol,around5080%isabsorbedin theintestine(bosneretal.,1999)mainlythroughinclusioninmixedmicelles(containingbile acidsandphospholipids)butalsobyspecifictransporters(altmanetal.,2004).severalabc typetransporterstakecontroloftheexcretionofcholesterolbacktotheintestinallumen. Onceintheenterocyte,cholesterolisesterifiedbyACATandtransferredintoQuilomicronsin 7

Introduction ordertobesenttothebloodtorrent.quilomicronsbringtgintotheadiposetissueand become enriched in cholesterol. Similarly, VLDL which are also formed by combining cholesterolandtg,bringtgtomuscularandskeletaltissue,givingrisetoldl,enrichedin cholesterol.cholesterolhomeostasisisregulatedbylxr(liverxreceptor)andsrebps(sterol Regulatory Element Binding Proteins), by means of modulations in intestinal absorption, biosynthesis,hdlactivityandcholesterolexcretion(fiévetamdstaels,2009). Highcholesterollevelshavebeensteadilyassociatedwithseveralchronicdiseases,mainly cardiovasculardiseases,suchasmetabolicsyndrome(d Adamoetal.,2014;Gilbertetal., 2014). Phytosterols Phytosterolscannotbesynthetizedendogenouslyandtheirlimitedpresenceintheorganism (around 2 order of magnitude less than cholesterol) is completely of dietary origin. The absorptionrateforthesecompoundsislessthan5%,whichisreportedtobemuchlowerthan thatofcholesterol.thisisprimarilyrelatedtotheirloweraqueoussolubilityandslower transferencetomixedmicelles(ostlundetal.,2002;matsuokaetal.,2010;alemanyetal., 2013a). Theircholesterolloweringeffectwasfirstlyassociatedwithphytosterolsinthe1950s,and sincethen,theinterestonthesubjecthasgrownoverthelastdecades.nowadays,duetothe vastinformationprovidedbymorethan100clinicaltrials,anintakeof23g/dayofplant sterolsisgenerallyacceptedtoreduceplasmaldlcholesterolaround10%(lawetal.,2000; Katanetal.,2003;Abumweisetal.,2008;Wuetal.,2009;Demontyetal.,2009;Talatietal., 2010).Currentdatasuggestthattriglyceridelevelsarealsoreducedalthoughnoeffectis observedinhdlcholesterol(theuwissenetal2009;baumgartneretal.,2013;demontyetal., 2013;Langellaetal.,2014).Thestructuralsimilarityofphytosterolsandcholesterolaccounts fortheirsimilarmetabolicpathwaysandexplainstheirlipidloweringeffect(vonbergmanet al.,2005).severalpossiblemechanismshavebeensuggested(trautweinetal.,2003;smetet al.,2012),mainly:1)physicalcompetitionforspaceinmixedmicellesbetweencholesteroland phytosterols(matsuokaetal.,2010);2)higheravailabilityofphytosterolsintheintestinedue to better hydrolyzation by enzymes (Gupta et al., 2011); 3) Upexpression of ABCtype transportersasaresultoftheaccumulationofphytosterolsintheenterocyteduetotheirpoor esterificationbyacat(platetal.,2005). Besidestheircholesterolloweringeffect,someinvitroandinvivostudiesshowpromising resultswithrespecttoantiinflammatory,antipyretic,antidiabetic,immunoregulatorandanti 8

Introduction carcinogenicpropertiesofphytosterols(woyengoetal.,2009;brulletal.,2009;cillaetal., 2015). Ontheotherhand,highphytosterolplasmalevelshavebeenassociatedwithatherosclerosis andcardiovasculardisease(assmanetal.,2006) ).Thus,patientswithphytosterolemia a diseasecharacterizedbyhighabsorptionandlowexcretionofphytosterolspresentahigher riskofsufferingthiskindofpathologies(weingartneretal.,2014).furthermore,phytosterols may replace not only cholesterol from the core of the mixedm micelles, but also other compoundspresentinthemicelles,suchaslipophilicvitamins( Katanetal.,2003). 3.Steroloxidationproducts(SOPs) 3.1Chemicalproperties Asanyotherlipid,sterolscanundergooxidationrenderingothercompounds.Sterolstructure issusceptibleofoxidationinthedoublebondofthesterolring,aswellasinotherpositionsof itssidechain(ryanetal.,2009),classifyingoxysterolsintotwocategories:thoseoxygenated onthesterolring(mainlyattheposition7)andthoseoxygenatedonthesidechain(mainlyat positions24,25and27).asaresult,hydroperoxidesareformedatfirst,andsecondary Figure2.Chemicalstructuresofcommonoxysterols(Hovenkampetal.,2008) 9

Introduction oxidation products (alcohols, epoxides and carbonyls) later on. Dimers, oligomers and polymersofunoxidizedandoxidizedformsofsterolsarealsoformed,mainlyinadvanced stagesofoxidation.steroloxidationproductsareusuallyknownassopsoroxysterols;when the oxidation products are derived from cholesterol their designation is COPs or oxycholesterols,andwhenderivedfromphytosterols,theyarecalledpopsoroxyphytosterols. 3.2Presenceinfood Oxycholesterolsarecommonlyfoundinanimalfoodstuffscontainingnotableamountsof cholesterol.concentrationsfoundinaselectionofresearcharticlesrangefrom0.1to50µg/g inmeat,from0.7to30µg/ginfish,from3to290µg/gineggandeggderivedproductsand from 1 to 260 µg/g in dairy (Echarte et al., 2004; Otaegui et al., 2010; Derewiaka and Obiedzinski, 2012). The estimated common daily oxycholesterol intake is 3 mg/day (Hovenkampetal.,2008). Oxyphytosterolshavebeenidentifiedinavarietyofvegetablefoods,includingvegetableoils, margarines,frenchfries,milk,coffeebeans,wheatflour,fruitjuicesandinfantformulas (Otaegui et al., 2010; Alemany et al., 2012a; Derewiaka and Obiedzinski, 2012). Concentrationsrangebetween1and60µg/ginvegetableoils.Consideringadailyintakeof40 gofoil,aconsumptionofaround2mg/daycanbeestimated. Phytosterolenrichedfoodsmaybeanimportantdietarysourceofoxyphytosterols,compared tononenrichedproducts.commerciallyavailablenonenrichedandenrichedspreadscontain upto13and46µg/gofthesecompounds,respectively(conchilloetal.,2005).consideringan estimateddailyintakeof15gofspread,oxyphytosterolconsumptionwouldincreasefrom 0.195mgto0.7mgperday. 3.3Effectsintheorganism SOPspresentinfoodareabsorbedandincorporatedintotheorganismthroughdiet,asseveral invitroandinvivostudieshaveassessed,usingdosesofaround100500ppminthediet, mainlywithrodents(staprans2000;andoetal,2002;tomoyorietal.,2004,sotorodríguezet al.,2009;liangetal.,2011;plat2014).nonhumanprimatessufferedadverseeffectsafter consuming a diet containing oxidized cholesterol, compared to the control diet group, indicatingaprobableabsorptionofcops(deushietal.,2011).inhumanstudies,increasesin plasmalevelsforupto3001600µg/dlhavebeendetectedafterintakesof3400mgcops withinpotato,salami,cheeseandpowdered eggs(emanueletal.,1991; Linseisen1998; Staprans et al., 2003). The absorption, distribution and excretion are supposed to be accomplishedthroughsimilarmechanismsassterols(hovenkampetal.,2008;brownand 10

Introduction Jessup,2009;Terunumaetal.,2013).Theabsorptionratiosrangearound220%forCOPsand 2050%forPOPs(Alemanyetal.,2013a).Nevertheless,thisabsorptioncannotalwaysreflect anestimationoftheplasmalevels,sincesopscanalsobegeneratedwithintheorganism(see section4.1). Regardlessoftheirorigin,thepresenceofbothCOPsandPOPsinplasmaandtissueshasbeen extensivelyrelatedtoanumberofbiologicaleffects(polietal.,2009;sotteroetal.,2009; Otaeguietal.,2010;O Callaghanetal.,2014,Alemanyetal.,2014).Whilstthereisbroad biologicalresearchonoxycholesterols,theamountofbiologicalresearchonoxyphytosterolsis morerecentandlimited,andthemajorityofstudiescomparecopsandpops. Ontheonehand,SOPs(mainlythosegeneratedfromautooxidation)havebeenshownto upregulate the expression of various proinflammatory molecules, including adhesion molecules,growthfactors,cytokinesandchemokines(leonarduzzietal.,2005;lemaire Ewingetal.,2005;Masciaetal.,2010;Alemany2013b).Conversely,oxysterolsoriginating fromenzymaticsteroloxidationproduceanantiinflammatorysignallinginmacrophages (Olkkonen,2012). Ontheotherhand,invitroandinvivocytotoxiceffectshavebeenwidelyreportedforboth COPsandPOPs,althoughtheformerpresentmuchhighercytotoxicitylevels(Adcoxetal., 2001;Meynieretal.,2005;Maguireetal.,2003;Roussietal.,2007;O Callaghan2010;Kenny etal.,2012;vejuxetal.,2012;alemanyetal.,2012b;biasietal.,2013).nevertheless,recent studies have also supported the cytotoxic effects of campesterol, stigmasterol and sitosteroloxides(koschutnigetal.,2009;o Callaghanetal.,2010;O Callaghanetal.,2013). Besides, COPs and POPs activate cell death signalling (including apoptosis) by different routes(ryanetal.,2005;roussietal.,2005).amongthedifferentsopsstudied,7hydroxy, 7ketoandtriolderivativesarethemostcytotoxicones.Thepotentialuseofoxysterolsas chemotherapeuticdrugsisanemergingresearchlinewhichdeservesfurtherattentionsince selectivecytotoxicityhasbeenfoundinsomecases(carvalhoetal.,2011;segalaetal., 2013).Moreover,somegenotoxiceffectshavebeenshownbyCOPs(Osada,2002)but studiesassessedwithpopsfailedtofoundmutagenesis(maguireetal.,2003;koschutniget al.,2010). SOPsinplasmaandtissuesarerelatedtooxidativestressbytwofeedbackmechanisms. First,theirpresencecontributestotheoveralloxidativestatusincells(Koschutnigetal., 2009;O Callaghanetal.,2010).Second,anoxidantenvironmentenhancesinsituSOPs 11

Introduction formation (Vaya et al., 2013). But their cytotoxicity has not been counterbalanced by antioxidantsinsomestudies(ryanetal.,2005;o Callaghanetal.,2010;Baumgartner2013). Takentogether,alltheseSOPinducedeffectssuggesttheirpotentialimportanceintheonset ofchronicdiseasesinwhichoxidativestress,inflammationandcelldeathappeartobe involved,suchasatherosclerosisandneurodegenerativediseases. MenéndezCarreñoetal.(2011)foundhighcorrelationbetweenCOPslevelsandCVDrisk factorsinhumans.specialattentionmustbepaidtotheatherogeniceffects.wideresearch isavailableonthepromotionofatheromatousplaquedevelopmentbycopsinanimaland humanstudies(stapransetal.,2003;larssonetal.,2006;chenetal.,2009;chalubinskyet al.,2013).particularly,7hydroxy,7ketoandtriolderivativesarethemainoxycholesterols involved.thefirstinvivoexperimentssearchingforpopsproatherogenicactivityreported noeffect(andoetal.,2002;tomoyorietal.,2004)butrecentstudiespointouttoacertain proatherogenicity (Liang et al., 2011; Yang et al., 2013; Plat et al., 2014). Other hypercholesterolemicrelatedpathologies(suchasdiabetesandhyperlipidemia)alsoresult inhighoxycholesterolplasmalevels(aboetal.,2000;arcaetal.,2007).theinvolvementof SOPswithpathologiesofthecentralnervoussystemincludesoptical,psycriaticandaged related diseases. Particularly, patients with visual abnormalities, depression, fatigue, Alzheimer and Parkinson have shown elevated levels of oxysterols (mainly 24 hydroxycholesteroland27hydroxycholesterol)incertaintissuesandfluids(xuetal.,2012; Shichirietal.,2013;Leonietal.,2013;Björkhemetal.,2013;Freemantleetal.,2013). Besides,high7hydroxycholesterollevelswerefoundinsamplesfrompatientswithlung cancer. TheuseofSOPsasbiomarkersinsomeofthesepathologiesisapromisingstrategytoallow anearlierdiagnosis,asmanyofthesestudiessuggest. Finally,oxyphytosterolsappeartoimprovecholesterolhomeostasis.Themechanisminvolves upregulationoftheexpressionofabcfamilygenesthroughlxractivation,thusinhibiting intestinalcholesterolabsorption(engelandschubert,2005;platetal.,2005).otherlipid metabolismrelatedparametershaveshowntoimprovebydietarysops(suzukietal.,2002; Ikedaetal.,2006).Theyhavealsobeenrelatedtomodulationoftheimmunesystem (Kimuraetal.,1995)andcertainhormonalactivity(ChristiansonHeikaetal.,2007). 12

Introduction 4.Steroloxidationprocess 4.1Formationmechanisms SOPscanoccurbothendogenously(invivo)andexogenously(exvivo).SOPspresentinfoods areundoubtedlyformedexogenously.conversely,sopsinplasmacanbeattributedtoboth endogenous(invivooxidativetransformationfromsterols)andexogenous(oxidationinfood andlaterabsorptionfromthediet)sources.whilstprocessesinvolvedincholesteroloxidation arewellknown,detailedknowledgeisstilllackingonphytosteroloxidation;however,current data suggest that both kinds of sterol oxidation products are formed following similar pathways. Two main mechanisms have been suggested: enzymatic and nonenzymatic. Generally,ringoxygenatedsterolstendtobeformednonenzymatically,whereassidechain oxygenatedsterolsusuallyhaveanenzymaticorigin.nonetheless,25hydroxyand7hydroxy canbeformedbybothpathways(romerandgarti,2006;brownandjessup,2009).non enzymaticpathwaycomprisesautooxidationandphotooxidationandtakesplacebothinvivo andexvivo. TheinitialreactionsinsterolautooxidationprocessstartwhenanallylichydrogenatC7is abstracted,generatingafreeradical.thisonecanreactwithmolecularoxygentoforma7 peroxylradical,whichisstabilizedbyhydrogenabstractionproducingthemorestable7 hydroperoxides (Brown and Jessup, 2009; Iuliano et al., 2011). These compounds can decompose, yielding epimeric 7hydroxysterols and 7ketosterols. The epimers 5,6 epoxysterolsareformedviaabimolecularinteractionofintactsterolandhydroperoxides.5,6 epoxysterols can be further converted to 3,5,6triol through hydration in an acidic environment(lampietal.,2002;saynajokietal.,2003;grandgirardetal.,2004;ryanetal., 2009).Whereasthegenerationofsidechainautooxidationproductsisnotascommonasthat of the ring structure, 20/24/25/27cholesterol hydroperoxides and their decomposition productshavebeenreported.reportsonsidechainoxidationproductsofphytosterolsare,on thecontrary,limited;tracelevelsof24hydroxyand25hydroxyderivativesofphytosterols havebeenidentifiedinheatedvegetableoils(smith,1981;lampietal.,2002;johnssonand Dutta,2005). SimilarPOPsmaybegeneratedthroughphotooxidation(Synajokietal.,2003;Zhangetal., 2006).Themechanismisnotfreeradicalmediatedandinvolvestheincorporationofsinglet oxygenspeciesdirectlyeitheronthe56doublebondoronthec7positiontogivethe corresponding hydroperoxides, which can further decompose to the previously named compounds. 13

Introduction Whereasautooxidationisenhancedathightemperaturesduetothehydrogenabstraction process,photooxidationispromotedbylightorpresenceofphotosensitizerssincetheyfavor singletoxygenformation. Sidechainoxidationisbelievedtobeduemainlytoenzymaticreactions.CytochromeP450 monooxygenases, dehydrogenases, epoxidases, hydroxylases and oxidases are involved in cholesteroloxidation.24,25and27hydroxycholesterols,amongothers,aregeneratedby specificenzymes(björkhemetal.,1998;lundetal.,1998;russel2000;bodinetal.,2001; Javitt2002;Björkhemetal.,2007;Bretillonetal.,2007).Thesameenzymesandroutesare presumably also involved in phytosterol oxidation. Alkyl groups at position C24 enable stereospecific 24Shydroxylation and might limit the formation of 25 or 27 hydroxyphytosterols. sterol25 sterol25oo sterol25ooh sterol25oh sterol7 sterol7oo sterol7ooh sterol7oh sterol5,6epoxy sterol3,5,6triol sterol7keto sterol5ooh sterol7ooh sterol7oh sterol steroltrioxolane sterol5,6epoxy sterol3,5,6triol sterol7oh sterol24oh sterol25oh nonenzymaticroutes enzymaticroutes sterol27oh Figure3.MainenzymaticandnonenzymaticroutesofSOPsformation. Extendeddegreesofoxidation(achievedatlongtermand/orhightemperaturetreatments) leadtodegradationofsopsandformationofoligomersandpolymers(struijsetal.,2010; Sosinskaetal.,2014;Derewiakaetal.,2015). 14

Introduction 4.2Relevanceofexogenousformation ConsideringthattheabsorptionofSOPsfromthediethasbeendemonstrated,andgiventheir potentiallyharmfuleffectsforhumanhealth,adeeperstudyofthefactorsaffectingthe formationofdietarysopsisessential.thecontentanddistributionofsopsinfoodsdependon foodcomposition,industrialprocessing,storageconditionsandculinaryprocess.amongthem, particularfactorscouldbeidentifiedasfollows:heating,air,light,lipidsurroundingmatrix, antioxidantsandwater. Heating The sterol oxidation process is directly related to the temperature. The higher the temperature,thefastersteroldegradationandsopsformation,andhigherconcentrationsare reached.particularbehaviorsstronglydependontheconditionsapplied,butintheoverall,it couldbestatedthattemperaturesbelow120 ChardlypromoteSOPsformation,whereas temperaturesover180 CproduceaveryintensegenerationofSOPs(Zhangetal.,2005; Kemmoetal.,2005;Seckinetal.,2005;Soupasetal.,2007;Yenetal.,2010;Derewiakaetal., 2015). The influence of the heating time is undoubtedly important, too. After longterm heatingtreatments,oxidationissohighthatsopsmaymedegraded.thisdecreaseinsops levelsisobservedatdifferenttimesdependingontheconditionsapplied.cookingconditions andindustrialprocesseshaveshowntoinducesopsformationandsubsequentdegradation throughheatingtreatments(menéndezcarreñoetal.,2008;azadmarddamirchianddutta, 2009;Broncanoetal.,2009;MazalliandBragagnolo,2009;Pikuletal.,2013;Liraetal.,2014; Zardettoetal.,2014). Air Oxidativereactionscannotoccurunlesssufficientelementaloxygenisavailableinthemedium. ThecontentonSOPSofseveralfoodstuffsstoredunderdifferentpackagingconditionshas been monitored, concluding that atmospheres poor in oxygen significantly improved the preservationoftheproducts(bosellietal.,2012;penkoetal.,2015).storagetemperatureis alsocrucialforsteroloxidativestability(gawrysiakwitulskaetal.,2012;botelhoetal.,2014; Rudzinskaetal.,2014). Light Ontheotherhand,lightisknowntobeafreeradicalreactionsinducer,aswellasforsinglet oxygenspecies.thus,asignificantincreaseinsopslevelshavebeenextensivelyfoundin vegetableoils,dairyproducts,eggs,meatandfishafterexposuretonaturalorartificiallight (Zhangetal.,2006;Bosellietal.,2012;Cardeniaetal.,2013;HernándezBecerraetal.,2014). 15

Introduction Photooxidation depends mainly on exposure duration, although very intense treatments couldfurtheroxidizepops.moreover,certainsubstancesnaturallyoccurringinfoodssuchas rivoflavin,chlorophyllorporphyrin,canactasphotosensitizers,increasingphotooxidation (Wanasundaraetal.,1998;Chienetal.,2003).Nevertheless,theuseofalternativeprotective packagingandlightingconditionsduringcommercialretailstoragecanefficientlyprevent sterolphotooxidation. Water Thepresenceofwater,eitherwithinthefoodorintheatmosphere,adverselyaffectssterols, asseveralstudieswithoilshaverecentlyshown(cercacietal.,2007;gawrysiakwitulskaetal., 2012). Unsaturationdegreeofthesurroundinglipids Duringthelast15years,therehasbeenconsiderableevidenceoftheinfluenceofthelipid unsaturationdegreeontheintensityofsteroloxidativeprocesses,despitenoconsensuson thematterhasbeenachievedyet.whereassomeauthorshavefoundaprotectiveeffect (Chienetal.,2003),someothershaveobservedaprooxidanteffectoftheunsaturatedlipids surroundingthesterol(lehtonenetal.,2012).timeandtemperatureconditionshavealso beenproposedaspossiblecriticalfactorsonthebehaviorofsterolswithinunsaturatedlipids (Soupasetal.,2004;Xuetal.,2011).Therefore,moreresearchisneededtoclarifythis question. Antioxidants Protectionagainststeroloxidationhasbeenassociatedwiththeantioxidantcapacityoffoods (Xuetal.,2009;Tianetal.,2011).Thisprotectionisusuallyattributedtophenoliccompounds andtocopherols,compoundsnaturallypresentinfruits,seedsandvegetableoils(xuetal., 2001;Chienetal.,2006;Palozzaetal.,2008).Hence,theadditionofnaturalantioxidantsis understoodasaninterestingtooltoavoidplantsterolloss,aswellasformationoftoxiccops andpops.inthissense,plantextractshavebeenextensivelytestedinfoods,achievingvery successfulresults(rodríguezcarpenaetal.,2012b;dasetal.,2012;figueiredoetal.,2014). Amongthewidevarietyofplantscontainingantioxidantcompoundsandpotentiallyapplicable in foods, two of them have been selected in this work: Melissa officinalis and Solanum sessiliflorum. 16

Introduction Melissaofficinalis(Lemonbalmormelisa)isamedicinalplantusuallyconsumedasinfusion duetoitsrecognizedbeneficialeffectsmainlytowardsthecentralnervoussystemandthe digestivesystem.recentstudieshavereportedits antiproliferativeeffectsuponcoloncancer cellsanditsantioxidanteffectstowardslipidoxidationbothinvivoandinfoods(lópezetal., 2009;Encaladaetal.2011;Berasategietal.,2011). Thus,itisapotentiallyinterestingplantfor the use against sterol oxidation. The major phenolic compound found in this plant is rosmarinicacid,followedbyotherphenolicacidssuchascaffeic,syringicandchlorogenic (Hoyos,2009). MelissaofficinallisRosmarinicacid Solanumsessiliflorum(manacubiu)isafruitnativetotheAmazonianregionandconsumed mostly as salad, juice or jelly. It is traditionally used formedicinalm purposes due to its hypoglycemic and hypocholesterolemic activity,and it alsoexhibits antigenotoxiceffects (Pardo,2004;Hernandesetal.,2014).Besides,itshowshighantioxidantactivity(Rodrigueset al.,2013).thusitisalsoapotentiallyinterestingplantforitsuseagainststeroloxidation.the majorphenoliccompoundfoundinthisfruitis5caffeoylquinicacid,anditalsocontainsgreat amountsofcarotenoids. Solanumsessiliflorum 5Caffeoylquinicacid 17

Introduction 4.3Modelsystemsasausefulexperimentaltool Foodsareusuallycomplexmatriceswhereinterferencesamongseveralcomponentsmay hamperaclearviewaboutthemechanismsofsteroloxidation.therefore,modelsystemsarea veryusefultooltoevaluateseparatelythefactorsthatexertaninfluenceinthisprocess, avoidingtheambiguityfrominterferencesamongthem.thus,adeeperunderstandingofthe underlyingmechanismsisenabledandkineticcurvescanbedeterminedeasily.theeffectof severalantioxidantsandlipidmatricesagainststeroloxidationhasbeentestedinmodel systems(chienetal.,2006;palozzaetal.,2008;xuetal.,2009;yenetal.,2011;kmieciketal., 2011;Xuetal.,2011;Lehtonenetal.,2012;Ansorenaetal.,2013).Adiversityofexperimental approachescanbefound:fromfullymodelledstudieswhereonlychemicalstandardsareused ascomponentsoftheexperiments,tointermediatemodelsystems,wherechemicalstandards aremixedwithinfoods.somemathematicalmodelsforsterols degradationandoxysterols formationhavebeenobtainedfromthiskindofstudies(chienetal.,1998;huandchen,2002; Ansorenaetal.,2013). 18

Justificationandobjectives

Justificationandobjectives Takingintoaccountthestateofartinthefieldoflipidandsteroloxidation,itcanbestated that: a) Thereisoverwhelmingmethodologyforlipidoxidationanalysis.Particularly,sterols andoxysterolsdeterminationresultsincomplex,laboriousandexpensiveprocedures. b) Dietarysterolsarenutritionallyinterestingcompoundswhichcanundergooxidation reactionsduringfoodmanufactureandstorage,aswellasintheorganism.their oxidationproductsareassociatedwiththedevelopmentofhighlyprevalentnon infectiousdiseases.therefore,itisrelevanttoevaluatethefactorsthataffectsterol degradationandoxysterolsformationinfoods. c) Modelsystemsarevaluabletoolstoseparatelyevaluatefactorsexertinganinfluence insteroloxidation,avoidingtheambiguitywhichnormallyresultsfrominterferences amongthem. Consequently,inthepresentwork,thefollowingobjectiveswereaimed: 1. Tooptimizethemethodologyforoxysterolsanalysis. 2. To monitor the behavior of cholesterol and three major plant sterols (sitosterol, campesterolandstigmasterol)duringheatingat180 C,assessingsterolsdegradation andoxysterolsformation. 3. Toevaluatetheinfluenceoftheunsaturationdegreeofthesurroundinglipidsin sterolsdegradationandoxysterolsformationunderheatingconditions. 4. Toevaluatethepotentialprotectiveeffectofdifferentnaturalantioxidantsonsterol degradationandoxysterolsformationunderheatingconditions,bothinmodeland foodsystems. 21

Justificaciónyobjetivos Teniendoencuentaelconocimientoactualsobeoxidacióndelípidosydeesteroles,sepuede afirmarque: a) Lametodologíadeanálisisdeoxidaciónlipídicaesmuyvariada.Particularmente,la determinación de esteroles y oxiesteroles requiere procedimientos complejos, laboriososycaros. b) Los esteroles dietéticos son compuestos interesantes desde el punto de vista nutricional, que pueden sufrir reacciones de oxidación durante el procesado y el almacenamiento de los alimentos, así como en el organismo. Sus productos de oxidaciónserelacionanconeldesarrollodeenfermedadesnoinfecciosasdealta prevalencia. Por lo tanto, es importante evaluar los factores que afectan a la degradacióndeesterolesyalaformacióndeoxisterolesenalimentos. c) Lossistemasmodelosonútilesparaevaluarporseparadolosfactoresinfluyentesenla oxidacióndeesteroles,evitandolaambigüedadresultantedelasinterferenciasentre ellos. Porlotanto,enelpresentetrabajo,losobjetivosfueronlossiguientes: 1. Optimizarlametodologíadeanálisisdeoxiesteroles. 2. Estudiar el comportamiento de colesterol y tres esteroles vegetales mayoritarios (sitosterol, campesterol y estigmasterol) durante el calentamiento a 180 C, determinandoladegradacióndeesterolesylaformacióndeoxiesteroles. 3. Evaluarlainfluenciadelgradodeinsaturacióndelamatrizlipídicaenladegradación deesterolesylaformacióndeoxiesterolesduranteelcalentamiento. 4. Evaluarelpotencialefectoprotectordediferentesantioxidantesnaturalessobrela degradacióndeesterolesylaformacióndeoxiesterolesduranteelcalentamiento, tantoensistemasmodelocomoenalimentos. 22

Experimentaldesign

Thefollowingdiagramrepresents thedifferentitemsstudiedalongthewholeresearchperiod.the experimentalsetsarerepresentedincoloredboxesandthedisseminationofresultsishighlightedinblue. FOODSYSTEMS MODELSYSTEMS ANALYTICALGOALS LIPIDMATRIX 18:0 18:1 18:2 18:3 campesterol stigmasterol+ sitosterol Paper7 stearate oleate linoleate linolenate cholesterol stigmasterol LIPIDOXIDATION STEROLOXIDATION HEATING 180 C 0360min +sunfloweroil Paper6 Reviewofmethods Methodoptimization Methodinterlabcomparison cholesterol campesterol stigmasterol sitosterol ANTIOXIDANTS Paper3 Paper1 Poster1 Paper2 22:6 VitaminE Manacubiu Phenolics Melisa tuna+manacubiu Poster2 cholesterol+manacubiu+dha Paper5 cholesterol+melisa beef+melisa+oliveoil Paper4 25 Paper4

Materialandmethods

Materialandmethods 1.Samplepreparation 1.1Modelsystems Foreachmodelsystem,thecorrespondingmixtureofsterols,sunfloweroil,FAMEorplant extractswassolvedinchloroform.aliquotsweretransferredtoopenglasstubes(15x100 mm)andevaporatedunderastreamofn 2. Then,thetubeswereplacedinthetermbloc, previouslyheatedat180 C.Afterthecorrespondingheatingtime,theywerecooleddownin anicebath,solvedinchloroformandkeptat20 Cuntilanalysis. Table1.Amounts(mg)ofsterols,plantextractsandlipidsusedinthedifferentmodelsystems Experiment Cholesterol Campesterol Stigmasterol Sitosterol Othercompounds Paper3 2.5 0.08 1.34 0.93 Paper4 20 Melisaextract(0.4) Paper5 1 Manacubiuextract(0.5) DHA(1) Paper6 1.2 1.2 Sunfloweroil(240) Paper7 0.7 0.36 1.3 FAME(240) 1.2Beefpatties Meatwasconvenientlydoublemincedandallpattiesweighed80g.Twotypesofpattieswere formulated:simplepatties(withoutemulsion)andemulsioncontainingones(includedanoilin water emulsion). In each case, patties with and without an aqueous extract of Melissa officinallis(melisa)wereprepared.simplepattiescontained79.2gmeatand0.8gcommon salt.for simple with melisa patties,saltwassubstituted withenrichedsalt (previously preparedbymixtureandhomogenizationwiththem.officinalisextract:16gsalt+64,80,104, 200,600or800mgmelisaextract).Formulationoftheemulsioncontainingpattiesconsisted of75.2gofmeat,0.8gsaltand4gofanoilinwateremulsion.tomaketheemulsion,52.63g ofextravirginoliveoilwasslowlyaddedto42.1gwater(containing5.3gsoyaprotein),while continuouslyhomogenizingwithanultraturrax.for emulsionwithmelisapatties,melisa extract(250,300or400mg)wasaddedtothewaterphaseoftheemulsionbeforemixingwith oil. Mixtureofingredientswascompressedwithaconventionalburgermakeruntilacompacted andhomogenizedpattywasobtained(80g,8.6cmdiameterand1.5cmthickness). Forthedifferenttypesofmeatpatties,fourindependentbatcheswereprepared,eachone containing4patties(twotokeeprawandtwoforcooking).pattieswereputinapreheated 29

Materialandmethods ovenat185 Cfor12min,reaching65 Cofinternaltemperature.Justafterthecooking process,theywerecooleddownfor10min,weighted,minced,andstoredat 20 Cunder vacuumuntiltheanalysis. Themelisaextractwasobtainedbyheating50gofleaveswith500mLofdistilledwaterat100 Cduring30min.Theprocesswasrepeatedtwiceandthesolutionwaslyophilized(García ÍñiguezdeCirianoetal.,2010b). 1.3Tunapatties Tunawasmincedwithaconventionalfoodmincer.Allpattiescontained50gtuna,0.5gsalt and2.5mlofanaqueoussolution.thisaqueoussolutioncontained0,0.02or0.1mg/mlofan aqueous Solanum sessiliflorum (manacubiu) extract. The ingredients were homogenized manuallyandintroducedinapreheatedgriddleat180 Cfor7min(3.5mineachsideofthe patty),reaching72 Cofinternaltemperature.Justafterthecookingprocess,theywere cooleddownfor10minandminced.hexanaldeterminationwascarriedoutthesamedayof cooking.forcholesterol,copsandlipidcontentdeterminations,sampleswerestoredat20 C undervacuumuntiltheanalysis. Manacubiufruitswerelyophilizedbeforeextraction.Fiftygramsoflyophilizedmanacubiu fruitwerehomogenizedwithultrapurewaterinavortexfor5minandcentrifugedat20000g at10 C.Theaqueouslayerwaslyophilized. 2.Moisturedetermination TheAOACofficialmethodwasusedformoisturedetermination(AOAC,2002a).5goffood samplewerehomogenizedwithsandand5mlethanol.sampleswereleftat100 Cuntil constantweight. 3.Lipidextraction 3.1Lipidextraction(I)Quantitative Extraction with petroleum ether by the Soxhlet method was applied for a quantitative determinationofthetotalfatcontent,accordingtotheaoacofficialmethod(aoac,2002b). 3.2Lipidextraction(II)Qualitative Extractionwithchloroform:methanol,asproposedbyFolchetal.(1957)wasfollowedwith slightmodifications.samples(120g)werehomogenizedwith300mlofchloroform:methanol during3min,andcentrifugedat10000rpmfor20minat010 C.Thesolidresiduewasadded with 100 ml chloroform and again homogenized, centrifuged and filtered. Both filtered 30

Materialandmethods solutionsweremixedand100mlofkcl0.88%wereadded.thesolutionwasshakenandthen separationofphaseswasallowed.chloroformphasewasrecovered,andevaporatedina rotavapor.thismethodwasappliedtobeefpatties. 3.3Lipidextraction(III)Qualitative Extractionwithchloroform:methanol,asproposedbyFolchetal.(1957)wasfollowedwith slight modifications. Ten grams of sample were homogenized with 100 ml of chloroform:methanolduring2min,andpouredintoadecantationfunnelthroughfiltration withpaper.thesolidresiduewasaddedwith50mlofthesolvents mixtureandagain homogenized,andfiltered.finally,thesolidresiduewasaddedwith25mlofthesolvents mixtureandagainhomogenized,andfiltered.40mlofkcl0.74%wereaddedtothefunnel. Thesolutionwasshakenandthenseparationofphaseswasallowed.Aftertherecoveryofthe organiclayer,25mlmoreofkcl0.74%wereadded.shakingandseparationwascarriedout againandthesolventwasevaporatedinarotavapor.thismethodwasappliedtotunapatties. 4.Fattyacidsdetermination Forthedeterminationinsunfloweroil,analiquotofsample(correspondingto0.2gofoil)was transferredtoaroundbottomflaskandchloroformwasevaporatedunderastreamofn 2. Fattyacidmethylesters(FAME)werepreparedbyderivatizationwithBorontrifluoride/ Methanol,andtheiridentificationandquantitationwasperformedbyCGFID,asdescribedin Ansorenaetal.(2013b). ForFAMEmodelsystemsnoderivatizationwasrequiredasthefattyacidsusedwerealready methylated.thefirstfractionrecoveredfromnh 2 SPEpurificationofthesamples(asdetailed below)wasevaporated,resolvedinheptane(2ml)andinjected(0.5µl)inthegcfid,as describedinansorenaetal.(2013b). Forthemanacubiumodelsystem,docosahexaenoicacidwasconvertedintoitsmethylester accordingtojoseph&ackman(1992)andanalyzedwithagaschromatograph(gc2010 model,shimadzu)equippedwithafusedsilicacpsil88capillarycolumn100mx0.25mm, 0.20µmandflameionizationdetector.Chromatographicconditionsweredescribedindetail bysanchoetal.(2011). 31

Materialandmethods GasChromatographFlameIonizationDetector(PerkinElmerClarus500): Column:SP2560(100mx0.25mmx0.20µm) Carriergas:H 2,2.15mL/min Oventemperaturesprogram: 175 Cduring10min Slope1:10 C/minupto200 C Slope2:4 C/minupto220 C 220 Cduring15min Injectortemperature:250 C Volumeofsampleinjected:0.5µL,splitratio=120 Detectortemperature:260 C Identificationofthecompoundswascarriedoutbycomparisonwiththeretentiontimesof theirpurestandards.quantitationwasperformedbyinternalstandardcalibrationcurves, usingmethylheptadecanoateastheinternalstandard. Table1.Retentiontimesandcalibrationcurvesofthefattyacidsmethylesters FattyAcidMethylEsters T R (min) Calibrationcurve Palmitic 9,45 y=0,9805x0,0204 tpalmitoleic 10,22 y=0,9469x0,0012 Palmitoleic 10,58 y=0,9497x0,0019 Estearic 12,47 y=0,9983x0,0002 Elaidic 13,10 y=0,9850x0,0009 Oleic 13,40 y=1,0012x0,0071 Vaccenic 13,45 y=1,0694x0,0124 tlinoleic 14,22 y=0,9538x0,0019 ctlinoleic 14,45 y=1,0241x0,0028 tclinoleic 14,53 y=1,0758x0,0062 Linoleic 14,75 y=0,9961x0,0007 linolenic 15,77 y=0,9260x0,0005 Eicosaenoic 15,88 y=1,0522x0,00004 linolenic 16,34 y=0,9200x0,0004 Aof FAME Where:y ; A is x mg mg FAME is 32

Materialandmethods 5.Sterolsdetermination Bygaschromatography Formodelsystemsamples,analiquot(correspondingtoapproximately0.2mgofsterol)was transferredtoatubeand5cholestane(2mg/mlinhexane)wasadded.thesolventwas evaporatedundergentlenitrogenstream.sampleswerederivatizedtotrimethylsilyl(tms) ethersaccordingtoamodifiedversionofthemethoddescribedbyduttaandappelqvist (1997).FourhundredmicrolitresofTriSilreagentwereaddedtoeachsampleandtheywere keptat60 Cfor45mininawaterbath.Thesolventwasevaporatedunderastreamof nitrogenandthetmsetherderivativesweresolvedinhexaneforgaschromatography.510 mlofhexanewereaddedwhengcmsdwasaimedtobeused,and0.4mlifgcfidwasthe analyser.thesesolutionswerefiltratedwithasyringeandafilter(0.45µm)andpouredtoa glassvial,beforethechromatographicanalysis. BeefpattiessamplesrequiredprevioussaponificationandextractionaccordingtoKovacsetal. (1979).Briefly,3gofsamplewereweigthedandaddedwith1mL5cholestane(2mg/mLin chloroform).then,20mlofethanol(95%)and5mlkoh(50%)wereaddedandthemixture washeatedto50 Cduring1h.Whenthesamplewascooleddown,13mLofdistilledwater wereaddedandtheextractionwithhexanewasperformed(20mlx6times).finally,the solventwasevaporatedintherotavaporat35 C.DerivatizationtoTMSetherswasmadeasin themodelsystem.gcfidwasusedfortheanalysis. Forserumsamples,thesamesaponificationandextractionproceduresasforbeefpatties wereapplied.differentvolumesofsampleweretakenforcholesterol(0.05ml)andforplant sterols (0.3 ml) determination. Derivatization to TMSethers was made as in the model system.anddifferentfinalvolumesofhexanewerealsoaddedtothederivatisedsamples beforechromatographicanalysis:5mlforcholesteroland0.4mlforplantsterols.gcmsd (68905973)wasusedfortheanalysis. 33

Materialandmethods GaschromatograpyMassspectrometer(Agilent68905973): Column:19091S433HP5ms5%PhenylMethylSiloxane(30mx250mx0.25m) Carriergas:He,1mL/min Oventemperaturesprogram: 85 Cduring0.5min Curve1:50 C/minupto290 C Curve2:0.5 C/minupto298 C Injectortemperature:280 C Volumeofsampleinjected:1µL,splittlessmode Transferlinetothedetector:280 C Sourcetemperature:230 C Electronimpact:70eV Detectortemperature:300 C Massinterval:50.00550.00uma Detectionmode:SCAN Peakidentificationwasbasedoncomparisonoftheirmassspectrawiththespectraofthe Wileylibraryandalsowiththoseobtainedfromtheliterature.Acomparisonoftheirretention timeandmsfragmentswiththoseofstandardpurecompoundswasalsodone. Aninternalstandardmethodwasusedforquantitation,with5cholestaneastheinternal standard (is). Cholesterol and 5cholestane quantitation was made using total ion chromatograms,whileplantsterolswerequantifiedusingextractionchromatograms,onthe basis of the amount of a specific ion for each peak (343, 484, 357, for campesterol, stigmasterolandsitosterol,respectively),andtakingintoaccounttherelativeabundanceof eachionwithineachcompound(berasategietal.,2012). Table3.Retentiontimes,characteristicionsandmodeofquantitationofsterolsinchromatographicanalysis Compound t R (min) Characteristicions(m/z) Quantitation 5cholestane(is) 7.80 217,357,372 Totalarea(A) cholesterol 9.95 329,353,368,458 Totalarea campesterol 11.07 343,367,382 343(6.62%)* stigmasterol 11.56 355,394,484 484(3.05%) sitosterol 12.21 357,381,396,486 357(6.56%) *ionusedforintegration(abundanceoftheion) Aof ion Totalareaofeachplantsterolwascalculatedasfollows:totalA * 100 abundance 34

Materialandmethods Table4.Internalstandardcalibrationcurvesofsterols Compound Calibrationcurve R 2 cholesterol y=1.1903x0.0288 0.9937 campesterol y=1.1348x0.008 0.9923 stigmasterol y=1.0346x0.0156 0.9905 sitosterol y=1.1257x0.0324 0.9974 total Where:y Aof A is sterol ;x mg mg sterol is GaschromatographFlameIonizationDetector(AutosystemPerkinElmer): Column:HP1(30mx0.25mmx0.1m) Carriergas:H 2,10mL/min, Oventemperaturesprogram: 265 Cduring8min Injectortemperature:300 C Volumeofsampleinjected:0.5µL,splitratio=20 Detectortemperature:300 C Onlysomeofthecholesterolsampleswereanalysedwiththisequipment.Identificationwas performedbycomparisonoftheretentiontimeofthepurestandard.quantificationwas performedbyintegrationoftheareasofthepeaksobtained,usinganinternalstandard calibration curve, with 5cholestane as the internal standard (y = 1.0817x0.0285; R 2 =0.9993). 35

Materialandmethods GaschromatographMassspectrometer(Agilent6890N5975): Column:VF5msCP89475%PhenylMethylSiloxane(50mx250mx0.25m) Carriergas:He,1mL/min Oventemperaturesprogram: 85 Cduring0.5min Curve1:50 C/minupto290 C Curve2:0.05 C/minupto291 C Injectortemperature:250 C Volumeofsampleinjected:1µL,splittlessmode Sourcetemperature:230 C Electronimpact:70eV Detectortemperature:150 C Massinterval:50.00550.00uma Detectionmode:SCANandSIM Peakidentificationwasbasedoncomparisonoftheirmassspectrawiththespectraofthe Wileylibraryandalsowiththoseobtainedfromtheliterature.Acomparisonoftheirretention timeandmsfragmentswiththoseofstandardpurecompoundswasalsodone. Aninternalstandardmethodwasusedforquantitation,with5cholestaneastheinternal standard(is).sterolquantitationwasperformedbyselectedionmonitoring(sim)analysis:for eachstageoftime,adifferentionwasselectivelydetectedandquantitated. Table5.Retentiontimesandcharacteristicionsofsterolsinchromatographicanalysis Compound t R (min) Characteristicions(m/z) 5cholestane(is) 13.0 217 *,357,372 cholesterol 17.7 329,353,368,458 campesterol 20.5 343,367,382 stigmasterol 21.1 355,394,484 sitosterol 23.2 357,381,396,486 * Ionsinbolddenotetheionusedforintegration Table6.Internalstandardcalibrationcurvesofsterols Compound Calibrationcurve R 2 cholesterol y=0.3928x0.005 0.9972 campesterol y=0.3188x+0.0253 0.9961 stigmasterol y=0.1033x+0.0037 0.9995 sitosterol y=0.2977x0.0415 0.9993 Where:y Aion Aion sterol is x mg mg sterol is 36

Materialandmethods ByHPLC Forthemanacubiumodelsystem,eachsamplewasdissolvedwith1mLhexane:2propanol (97:3,v:v),filteredthrougha22µmfilter,beforethechromatographicanalysis. Fortunapatties,previoussaponificationandextractionwererequired,whichwereperformed asinsaldanhaetal.(2008).tenmlofkoh20%inethanol(90%)wereaddedto1gofsample, andkept22hinabsenceoflightunderagitation.then,5mlofwaterwereaddedtothe samples,andextractionwithhexanewascarriedout(10mlx4times).afterwards,the sampleswerewashedwithwater(5mlx3times),driedwithna 2 SO 4,andthesolventwas evaporated in the rotavapor. Chromatographic analysis was performed as in the model system. HPLCUVRI(Shimadzu) Column:NovaPackCNHP(300mmx3.9mm,4µm) Mobilephase:hexane:2propanol(97:3,v:v)ataflowrateof1mL/min,30min Injectortemperature:280 C Volumeofsampleinjected:60µL(loop=20µL) Cholesterolidentificationwasmadebycomparisonofitsretentiontimewiththatofthepure standard. Quantitationwasdonebyexternalstandardization,usingtheareasfromtherefractiveindex detector,asinmariuttietal.(2008),(y=183186x4595;r 2 =0.9948). 37

Materialandmethods 6.SOPsdetermination Bygaschromatography Forthemodelsystems,analiquotofthesample(seesection1.1)wastransferredtoatube (corresponding to approximately 2 mg sterol), and added with 19hydroxycholesterol as internalstandard(1mlofa20µg/mlhexane:isopropanolsolution). Forthebeefpatties,previouslipidextraction,coldsaponificationandextractionwasrequired. Approximately0.5gofthepreviouslyextractedfat(asreportedbyFolchetal.,1957)was weightedinaflaskcontaining10mlofkoh1minmethanoland1ml19hydroxycholesterol (20 g /ml in hexane:isopropanol 3:2) and kept at room temperature for 20 h. Three extractionswithdiethylether(10ml)wereperformed.thewholeorganicextractwaswashed withwater(3x5ml)andfilteredthroughanhydroussodiumsulphate.thenitwasrecovered in a roundbottom flask, and the solvent was evaporated under a stream of nitrogen. PurificationbyNH 2 SPE,derivatizationtotrimethylsilylethersandanalysisbyGCMSwere performedfollowingthesameprocedureasinthemodelsystem(rosesallinetal.,1995; MenéndezCarreñoetal.,2008). Forserumsamples,asimplerlipidextractionwascarriedout.Chloroform/methanol2:1(9mL) was added to a tube containing 1 ml of serum and 0.1 ml of 19hydroxycholesterol (0.02mg/mlinhexane:isopropanol3:2).Shakefor1min,centrifugeat4000rpmfor15min and separate in a decantation funnel. Subsequent saponification, extraction, purification, derivatizationandanalysiswereperformedaspreviouslydetailedforbeefpatties. 38

Materialandmethods TwodifferentSPEpurificationprocedureswereused: a) AsdescribedindetailinGuardiolaetal.(1995),thesamplesdilutedin5mLof hexanewereappliedtoaspesilicacartridge,previouslyequilibratedwith5mlof hexane.thecartridgewassubsequentlytreatedwith10mlofhexane:diethyl ether (95:5, v/v), 30 ml of hexane:diethyl ether (90:10, v/v), and 10 ml of hexane:diethylether(80:20,v/v).steroloxidationproductswerefinallyeluted fromthespecartridgewith10mlofamixtureofacetone/methanol(60:20,v/v). Thesolventwasevaporatedinrotaryevaporatorunderwarmwaterbath(35 C). b) AsdescribedindetailinRoseSallinetal.(1995),samplesweredilutedin400µL hexane:ethyl acetate (95:5, v/v) and transferred to NH 2 SPE cartridge. The cartridgewassubsequentlytreatedwith8mlofhexane:ethylacetate(95:5,v/v) and10mlofhexane:ethylacetate(90:10,v/v).steroloxidationproductswere finallyelutedfromthespecartridgewith10mlofacetone/methanol.thesolvent wasevaporatedunderastreamofn 2. Thesamplesolutionsofsteroloxidationproductswerederivatizedtotrimethylsilyl(TMS) ethersaspreviouslydescribedforsterols,beforethechromatographicanalysis. GCMS(Agilent6890N5975): Column:VF5msCP89475%PhenylMethylSiloxane(50mx250mx0.25m) Carriergas:He,1mL/min Oventemperaturesprogram: 75 Cduring0.5min Slope1:30 C/min(or20 C/min )upto250 C Slope2:8 C/minupto290 C Slope3:0.05 C/minupto292 C Injectortemperature:250 C Volumeofsampleinjected:1µL,splittlessmode Transferlinetemperature:280 C Sourcetemperature:230 C Electronimpact:70eV Quadrupoletemperature:150 C Massinterval:50.00600.00uma Detectionmode:SCANandSIM theslowerslope(20 C/min)wasappliedinsomeexperiments 39

Materialandmethods Table7.RetentiontimesandcharacteristicionsofSOPs Compound t R (min) Characteristicions(m/z) Abundance(%) 7hydroxycholesterol 22.7(23.6) 456457 * 458 12.0 19hydroxycholesterol(is) 22.8(25.61) 353366 10.5 7hydroxycampesterol (25.73) 470471472 11.0 7hydroxystigmasterol (26.10) 482483484 12.0 7hydroxycholesterol 23.9(26.9) 456457458 3.4 7hydroxysitosterol (27.94) 484485486 11.0 5,6epoxycholesterol 25.4(28.4) 384474445357 2.0 5,6epoxycholesterol 25.8(28.9) 384474445366 1.5 7hydroxycampesterol (29.74) 470471472 3.1 cholestanetriol 27.9(30.94) 403456471546 6.7 7hydroxystigmasterol (29.88) 482483484 3.5 5,6epoxycampesterol (31.74) 370383398488 1.3 5,6epoxycampesterol (32.28) 398380488 11.0 7hydroxysitosterol (32.54) 484485486 3.0 5,6epoxystigmasterol (32.59) 253382410500 1.5 25hydroxycholesterol 30.0(33.1) 131456546271 2.7 5,6epoxystigmasterol (33.13) 253392410500 0.8 campestanetriol (34.84) 417418470560 3.6 7ketocholesterol 31.1(34.14) 367416472 9.0 5,6epoxysitosterol (35.14) 384394412502 0.14 stigmastanetriol (35.62) 429253482572 3.0 5,6epoxysitosterol (35.78) 394397412502 0.9 7ketocampesterol (38.74) 486381487396 4.8 sitostanetriol (38.75) 431432484574 1.4 7ketostigmasterol (39.96) 357359498347 4.3 7ketositosterol (43.70) 395500510410 6.9 * Ionsinbolddenotetheionusedforintegration Retentiontimesinparenthesisarethosefromthemethodwiththeslowerslope 40

Materialandmethods Peakidentificationwasbasedoncomparisonoftheirmassspectrawiththoseobtainedfrom theliteratureand,onlyinthecaseofcops,theirretentiontimesandmsfragmentswiththose ofstandardpurecompounds. Aof ion Totalareaofeachplantsterolwascalculatedasfollows:totalA * 100 abundance Quantitation was based on an internal standard method (19hydroxycholesterol). It was performedusingselectedionmonitoring(sim)analysis.foreachstageoftime,differentions wereselectivelydetectedand,consequently,extractionchromatogramwasusedtointegrate thecorrespondingpeakareas. Table8.InternalstandardcalibrationcurvesofCOPs Compound Calibrationcurve R 2 7hydroxycholesterol y=0.999x0.0256 0.9994 7hydroxycholesterol y=1.2394x0.0134 0.9996 5,6epoxycholesterol y=0.4132x0.005 0.9982 5,6epoxycholesterol y=0.4256x+0.0022 0.9976 Cholestanetriol y=1.4217x0.0123 0.9996 25hydroxycholesterol y=1.1149x0.0044 0.9996 7ketocholesterol y=0.26870.0133 0.9997 total Aof Where:y A is COP mg ;x mg GiventhelackofavailablePOPsstandardsandtheirdemonstratedsimilaritytoCOPsresponse, COPscalibrationcurveswerealsousedtodeterminePOPscontent. COP is 41

Materialandmethods ByHPLC Forboththemanacubiumodelsystemandthetunapatties,COPsdeterminationwasmade followingthesameprocedureasforsterolsdetermination. QuantificationinHPLCUVRIwasdonebyexternalstandardization,asinMariuttietal.(2008). Table9.RetentiontimesanddetectorusedforCOPsdetermination Compound t R (min) Detector 5,6epoxycholesterol 9.0 RI 5,6epoxycholesterol 10.2 RI 7ketocholesterol 14.8 UV 7hydroxycholesterol 20.7 UV 7hydroxycholesterol 21.8 UV Table10.ExternalstandardcalibrationcurvesofCOPs Compound Calibrationcurve R 2 5,6epoxycholesterol y=97.299x78.639 0.9988 5,6epoxycholesterol y=65.814x99.497 0.9988 7ketocholesterol y=8531.1x858.1 0.9999 7hydroxycholesterol y=5876.2x820.48 0.9999 7hydroxycholesterol y=4627.6x1378.3 0.9999 Where:y Aof COP ;x mgcop TheidentificationofCOPswasconfirmedbyHPLCAPCIMS/MSusingthechromatographic conditionsdescribedindetailbyzardettoetal.(2014)andthemsconditionspreviously optimizedbymariuttietal.(2008). Table11.RetentiontimesandmassspectrometrydataofCOPsidentification. Compound t R (min) [M+H] + (m/z) Fragmentions(m/z) 7hydroxycholesterol 5.5 nd 385[M+H18] +,367[M+H1818] + 7ketocholesterol 5.7 401 383[M+H18] +,365[M+H1818] + 7hydroxycholesterol 5.8 nd 385[M+H18] +,367[M+H1818] + 5,6epoxycholesterol 7.0 403 385[M+H18] +,367[M+H1818] + 5,6epoxycholesterol 7.6 403 385[M+H18] +,367[M+H1818] + 42

Materialandmethods 7.Otheroxidationparameters 7.1TBARS TBARSvaluesweredeterminedinsunfloweroilaccordingtothemethoddescribedbyPoyato etal.(2013).briefly,analiquotofsample(correspondingtoapproximately0.25gofoil)was transferredtoatubeandthesolventwasevaporatedunderastreamofn 2. Distilledwater(0.5 ml),bht(20µl,1%)andthetbarsreagent(2ml)wereaddedtothesampleandvortexed, placedinaboilingwaterbathfor15minandthencooleddowninanicebathtoroom temperature.cyclohexanone(4ml)andammoniumsulphate(1ml,4m)wereaddedtothe mixtureandvortexed.themixturewascentrifugedat1300gfor10minutes.theabsorbance wasmeasuredat532nminafluostaromegaspectrofluorometricanalyzer.forbeefpatties, previous lipid extraction was performed, according to Folch et al (1957), as explained previously.acalibrationcurvewasmadewithtetraethylpropaneasexternalstandard(y= 938.82x+0.0037;R 2 =0.9991).Resultswereexpressedinmgofmalondialdehyde(MDA)/Kg sample. 7.2PV PeroxidesValue(PV)wasanalysedinsunfloweroilandFAMEmodelsystemfollowingthe methodofshantaanddecker(1994)withslightmodifications.briefly,analiquotofsample (correspondingtoapproximately10mgoffat)wastransferredtoatubeandthesolventwas evaporated under a stream of N 2. The residue was dissolved in 5 ml of a mixture butanol:methanol,(2:1).scnnh 4 (30%indistilledwater,25µL)wasaddedandtubeswere vortexedfor4s.then,asolutionoffecl 2 (36mMinHCl,25µL)wasaddedandtubeswere vortexed. After 15 min, absorbance was measured at 510 nm in a FLUOStar Omega spectrofluorometricanalyzer.acalibrationcurvewithcumenehydroperoxidewasdonefor quantification(y=5.878x+0.0322;r 2 =0.9963).ResultswereexpressedasmeqO 2 /Kgsample, beingthedatatheaverageof2measurementsperreplicate. 7.3Hexanalcontent HexanaldeterminationwascarriedoutintunapattiesaccordingtoSouzaetal.(2014).Briefly, anaqueousdilutionwasperformedin10gsample,extractedbyspmeandinjectedinagcms (GCMSQP2010UltraShimadzu). 43

Materialandmethods 8.Antioxidantcapacityandspecificbioactivecompounds 8.1Totalphenoliccompounds TotalPhenolicContent(TPC)wasdeterminedinmelisaaqueousextractasdescribedinPoyato etal.(2013).a12mgextractsamplewassolvedin10mlwater.reagentsweremixed:237µl distilledwater,3µlsamplesolution,15µloffolinciocalteu sreagent,and45µlof20% sodiumcarbonateanhydroussolution.after2hinthedark,theabsorbancewasmeasuredat 765nminaFLUOStarOmegaspectrofluorometricanalyzer.Foroilsamplestheprocedurewas thesamebutpreviousphenolextractionwasperformed,asdescribedinpoyatoetal.(2013). A calibration curve with gallic acid was done for quantification (y = 0.3318x+0.0053; R 2 =0.9995).TPCwasexpressedasµggallicacid/mgsample(extractoroil). 8.2ORAC AntioxidantcapacityinthemelisamodelsystemwasassessedbymeansoftheORACmethod, accordingtotheproceduredescribedinouetal.(2001),withslightmodifications.analiquot ofsample(correspondingtoapproximately0.25mgcholesterol)wasevaporatedundera streamofnitrogen.phosphatebuffer(1ml)andchloroform(300µl)wereadded.then,the sampleswerevortexedfor20sandcentrifugedat4000rpmfor10min.atotalof0.5mlof theaqueouslayerwastakenandkeptinthedarkuntilanalysis.a0.5mstocksolutionof Troloxwaspreparedin10mMphosphatebuffer,anddividedinto1mLaliquots,whichwere storedat20ºcuntiluse.anewsetofstocktroloxvialswastakenfromthefreezerdailyfor thepreparationofthecalibrationcurveandthequality controls(12.5and50µm).the phosphatebuffersolutionwasusedasblank,todissolvethetroloxqualitycontrolsandto preparethesamples.toconducttheoracassay,analiquotofthesample(40µl)and120µl ofthefluoresceinsolution(132.5nm)wereaddedtothe96wellblackplate.themicroplate wasequilibrated(5min,37ºc),andthenthereactionwasinitiatedbytheadditionofaaph (40 µl, 300 mm); readings were obtained immediately, in a FLUOStar Omega spectrofluorometricanalyzer.acalibrationcurvewithtroloxwasdoneforquantification(y= 0.7293x+5.4373;R 2 =0.9923).Theresultswereexpressedasmgtroloxequivalent/gsample. 44

Materialandmethods 8.3Rosmarinicacid Rosmarinicacidcontentwasdetermintedinthemelisamodelsystem.Analiquotofsample (correspondingtoapproximately0.1mgmelisaextract)wasevaporatedunderastreamof nitrogen.ultrapurewater(1ml)andhexane(1ml)wereadded.thesamplewasvortexedfor 20sandcentrifugedat1300gfor6min.Theupperlayerwasdiscardedandtheprocesswas repeatedtwomoretimes.theaqueouslayerwasfilteredthrougha0.20µlmembranefilter andanalyzedusingthechromatographicconditionsdescribedingarcíaiñiguezdecirianoet al.(2010b).perkinelmeruvvislambda200seriesequippedwithaphotodiodearraydetector Series200PDAwasused.Briefly,inaC18column,andataflowrateof0.8mL/min,a gradientofacidifiedwater:acetonitrilewasapplied(startingat90:10;changingto70:30for 20min;andreturningto90:10in7min).Theprofileswererecordedat280nm.Acalibration curvewasdoneforquantification(y=10000000x;r 2 =0.9977).Theresultswereexpressedas mgrosmarinicacid/gsample. 8.45caffeoylquinicacidandotherphenoliccompounds Theidentificationandquantificationofthephenoliccompoundsofthemanacubiuextract (MCE)wascarriedoutaccordingtoRodriguesetal.(2013). 8.5VitaminE ThetocopherolcontentwasdeterminedinsunfloweroilbyHPLCUVanalysisaccordingto themethoddescribedbyberasategietal.(2012).briefly,analiquotofsample(corresponding to approximately 0.2 g oil) was transferred to a volumetric flask and chloroform was evaporated under a stream of N 2. tocopherol acetate (0.1 ml, 10 mg/ml solved in methanol)wasaddedasinternalstandardandtheflaskwasfilledupto10mlwithpreviously warmed(30ºc)supergradienthplcgrademethanol.dilutionwasvortexedfor30secand filteredwith0.20 mfilter.thesample(10µl)wasinjectedintothehplcsystemandan isocraticelutionwithmethanol/water(97:3)at1.5ml/minflowwasperformed.uvspectra wererecordedat295nmonaperkinelmeruvvislambda200seriesequippedwitha photodiodearraydetectorseries200pda,usingananalyticalprecolumn(3.8mmx8mmwith 4mmx3mmofC18cartridges;Phenomenex,)andaLC18column(150mmx3.9mm,4m; Waters).Identificationoftocopherolwasdoneusingtheretentiontimeofthepurestandard compound and its characteristic UV spectra. The quantification was performed using an internalcalibrationcurvepreviouslypreparedwithtocopherolacetateastheinternalstandard (y=7.3925x 0.0261;R 2 =0.9932).TheresultswereexpressedasmgvitaminE/100gsample. 45

Materialandmethods 9.Sensoryanalysis Triangleanalysiswascarriedoutintheexperimentinwhichbeefpattieswereelaboratedwith varyingamountsofmelisaextractandoliveoilemulsion.eachpanellistwaspresentedwith threesamples,twoofwhichwereidentical,andaskedtoindicatewhichonedifferedfromthe others.thisprocesswasrepeatedseveraltimes,onceforeachdifferentconcentrationof extracttested.thenumberofcorrectanswersforeachtypeofcomparisonwasdetermined. AccordingtoISO4120:2004,fora9memberpanel,thedifferencebetweensampleswas significantifthenumberofcorrectanswerswas6(p<0.05). 10.Statisticalanalysis SPSS15.0,Stata12andStatGraphicswereusedforthestatisticalanalysis.Toevaluatethe statisticaldifferencesbetweentwosamples,studenttandkruskalwallistestswereused.to evaluatethestatisticaldifferencesamongseveralsamples,anovaandmannwhitneyutests wereused,withastatisticallevelofsignificanceof0.05.tukeybandbonferroniposthoc comparisonstestswerealsoapplied.pearsonandspearman scoefficientswerecalculatedto determinethecorrelationbetweentwovariables.themathematicalmodelswereadjusted usingnonlinearregressions. 46

Results

ResultsI Paper1 Areviewofanalyticalmethodsmeasuringlipid oxidationstatusinfoods:achallengingtask

Results EuropeanFoodResearchandTechnology(2013)236(1),115 Areviewofanalyticalmethodsmeasuringlipidoxidationstatusinfoods: achallengingtask BlancaBarriuso 1,IciarAstiasarán 1,DianaAnsorena 1 1 DepartmentofNutritionandFoodScience,PhysiologyandToxicology.FacultyofPharmacy,UniversityofNavarra Abstract Lipidoxidationanalysisinfoodsamplesisarelevanttopicsincecompoundsgeneratedinthe processarerelatedtoundesirablesensoryandbiologicaleffects.astheprocessiscomplex anddependsonthetypeoflipidsubstrate,oxidationagentsandenvironmentalfactors, proper measurement of lipid oxidation remains a challenging task. A great variety of methodologieshavebeendevelopedandimplementedsofar,bothfordeterminingprimary oxidationproductsandsecondaryoxidationproducts.mostcommonmethodsandclassical procedures are described, including peroxide value, TBARS analysis and chromatography. Some other methodologies such as chemiluminescence, fluorescence emission, Raman spectroscopy,infraredspectroscopyormagneticresonance,provideinterestingandpromising results,soattentionmustbepaidtothesealternativetechniquesintheareaoffoodlipid oxidationanalysis. Keywords:Fatoxidation;Hydroperoxides;Secondarylipidoxidationproducts;TBA;Hexanal 51

Results 1.Introduction Lipidoxidationinfoodsconstituteacomplexchainofreactionsthatfirstlyyieldsprimary products (peroxides), that, when exposed to extended oxidation conditions, give rise to secondaryoxidationproducts,includingaldehydes,ketones,epoxides,hydroxycompounds, oligomersandpolymers.mostofthemproduceundesirablesensoryandbiologicaleffects (MárquezRuizetal.,2007;Kanner,2007).Therefore,itscontrolisofgreatimportance. Lipidoxidationoccursviadifferentpathways:radicalmechanism(knownasautoxidation), singlet oxygen mediated mechanism (known as photooxidation) and also the enzymatic oxidationhasbeendescribed,catalyzedbylipoxigenases.thisreviewwillbefocusedonthe nonenzymaticroutes.bothautoxidationandphotoxidationgiverisetoidenticalorsimilar peroxides,differingjustsometimesinpositionandstereoisomerism.thefirstmechanism requiresaninitialactivationenergyfortheremovalofahydrogenatom,soitisenhancedby hightemperaturesandpresenceofdoublebonds.thelatteristriggeredbythehighlyreactive singletoxygenspecie,whichisformedbyexcitationoftripletmolecularoxygen,underlight exposureandpresenceofphotosensitizers(choeandmin,2006;minandboff,2002). The first compounds formed during oxidation process are peroxides, especially hydroperoxides;hencetheyarecalledprimaryoxidationproducts.despitebeingintermediate compounds of lipid oxidation process, they are relatively stable (depending on the lipid structure),andcanbeusedtoassesslipidoxidationstatusinfoodsamples,providingnottoo advancedautoxidationisdevelopedinthesample.becauseofthisintermediatecharacteristic, temperature conditions during analysis must be controlled to avoid hydroperoxide decomposition,andadditionofantioxidantisoftenrequired. Hydroperoxidesusuallysufferfurtheroxidationtogivesecondaryoxidationproducts.Silvagni etal.(2010)proposedanalternativekineticmodelwherethealdehydesaregeneratednot only via direct degradation of hydroperoxides but from peroxyl radicals through an independentpathway.thismechanisminvolvesabimolecularreactiontoformintermediate tetraoxides,whichareunstableathightemperaturesanddecomposetogivealkoxylradicals. The wide variety of secondary oxidation products to which oxidation gives rise includes aldehydes,ketones,epoxides,hydroxycompounds,oligomersandpolymers.amongthem, bothvolatileandnonvolatilecompoundscanbefound,suchashexanalormalondialdehyde (MDA),respectively,asmainrepresentatives. 52

Results Evaluatinglipidoxidationstatusisachallengingtaskduetoanumberofreasons.Firstly, differentcompoundsareformeddependingonthetime,extentofoxidationandmechanism involved.therefore,choosingjustoneparametertoanalysetheoxidativestatusisrather difficultanditisfrequentlymoreconvenienttocombinedifferentmethods.besides,asstated byeymardetal.(2009),notonlynatureandcompositionoflipidasthesubstrateofthe reaction have an impact on lipid oxidation process, but also type and concentration of proteins, antioxidants and prooxidants present in the food matrix, as well as its physicochemicalcharacteristics.inmeatsamples,richardsanddettmann(2003)suggested thatratesoflipidoxidationmaydependontherelativeabilityofhaemoglobinsfromdifferent animalspeciestopromoteit.chenetal.(2010)proposedthatcolloidalstructuresformedby phospholipidsinvegetableoilscouldhaveanimpactontheoxidativestabilityoffoodoils. Lipidoxidationwasobservedtobedelayedinfishsausagesaftertheadditionofseveral antioxidants(maqsoodetal.,2012).milksamplesoxidationhasbeenrecentlystudiedinthe presenceofcatechinsandascorbicacid(mun,2011).ontheotherhand,eachmethodallowsa numberofdifferentexperimentalconditions,andthis,togetherwiththelackofuniformity amonglaboratories,leadsto(atleastforthemomentunavoidable)dissimilarresults.finally, mostoftheoxidationcompoundsarepronetobefurtherdegraded,whichprovidesanadded sourceofdivergence.therefore,aprecisecontroloftheexperimentalproceduremustbe kept. Relatedtolipidoxidationinfoodsamples,otherassessmentscanbealsoperformed.Onthe onehand,determinationofparametershighlyindicativeoflipiddeteriorationandsubsequent enhancedsusceptibilitytooxidation(suchashydrolysisoftriglycerides)isverycommon.on theotherhand,measuringthetimerequiredbyasampletoachieveacertainoxidativelevel through artificially promoting oxidation is another valid procedure to evaluate lipid susceptibilitytooxidation(and/oroxidationstability).however,thisreviewwillonlyfocuson methodsdeterminingtheactualandcurrentlipidoxidationofasample,discardingprocedures assessinghydrolyticstatusandthoseinvolvinginductionofoxidativedegradation,sincethey arenotproperlyindicatorsofoxidationstatusbutofoxidativesusceptibilityandstability, respectively. Thisreviewwilldescribetraditionalmethodstodeterminebothprimaryandsecondarylipid oxidation products in foods, from spectroscopic to chromatographic techniques. Their characteristics, advantages and limitations will be pointed out. Then, alternative methodologiesdevelopedduringlastdecadeswillalsoberevisedinordertoprovidethe 53

Results completeoversightofpossibleoptions.table1summarizesthemaincharacteristicsofthe methodsdescribedinthisreview. 2.Primaryoxidationproducts 2.1Peroxides Hydroperoxidesredoxpropertiesarethebaseofsomeofthekeymethodsappliedintheir determination.anumberofreagentscanbeoxidizedbyhydroperoxides,includingsimple inorganicions,suchasiodideorferrousion. Thesemethodsusuallyrequiresubsequent complexationtoimprovethesensitivity. 2.1.1Volumetricmethod Amongthedifferentmethodsproposedfortheanalysisofperoxides,theiodometryhasbeen the most conventional and widespread method mainly due to the simplicity of the experimentalprocedure.althoughtheprocedurerequirespriorlipidextraction,rapidand easilyunderstandableresultsareprovided. Inacidicmedium,hydroperoxidesandotherperoxidesreactwiththeiodideiontogenerate iodine,whichistitteredusingasodiumthiosulfatesolution,inthepresenceofstarch.the AOACoffersanofficialmethodsince1965(AOAC,2000).Accordingtothismethod,Peroxide Value(PV)isconsideredtorepresentthequantityofactiveoxygen(inmeq)containedin1kg oflipidandwhichcouldoxidizepotassiumiodide. Itshowshoweversomedrawbacks,mainlyderivedfromtheiodidehighsusceptibilityto oxidation in the presence of molecular oxygen and accelerated by light exposure. Also spontaneoushydroperoxideformationcanoccur(whichwouldleadtooverestimation)and absorptionofiodinebyunsaturatedfattyacids(whichwouldleadtounderestimation)(sunet al.,2011).moreover,itrequiresanhydroussystemstoavoidinterferenceproblems,forwhat lipidextractionisrequired,andthisprocedurestageincreasesthecontactwithoxygen.in addition,theperoxidevaluedeterminationdoesnotgivearealmeasureoftheoxidative degradation,sinceperoxidesareusuallyfurtherdegraded,sosimultaneousmeasurementof secondaryproductswouldbeappropriate. 54

Results 2.1.2VISUVspectroscopicmethods Aswellasthevolumetricmethod,spectroscopiconesarerathersimpleandaremoderately sensitive, reliable, and reproducible when carried out under standardized conditions. However,theyarehighlyempiricalastheymeasurecomplexmixturesofoxidizedmolecules. Inaddition,theyaregenerallyworkintensiveanduselargeamountsofsolventsandreagents thatmightbehazardous(kamaleldinandmin,2010). Ferrousoxidationmethod Theferrousoxidationmethodfordeterminationofperoxidecontentissimplertousethan iodometry.themainreasonisthelowersensitivityofferrousiontospontaneousoxidationby oxygeninair,ascomparedtohighsusceptibilitytooxidationofiodidesolutions.itconsistsof oxidationoffe(ii)tofe(iii),mediatedbyhydroperoxidereductioninacidicconditionsandin thepresenceofthiocyanateorxylenolorange(inthislatercase,methodisknownasfox). Thesetwocompoundsprovidethespectrophotometricproperties,astheyformcomplexes withtheferricion,givingmaximumabsorbancepeaksat500nmand560nmrespectively, whichcanbemeasuredwithauvvisspectrophotometer(eymardetal.,2009;shanthaand Decker,1994;Bouetal.,2008;Verardoetal.,2009;Chotimarkonetal.,2009;Sorensenetal., 2010).However,neitherofthemethodsisfreefromcomplications(Nielsenetal.,2003).The thiocyanatemethodrequireslargeamountsofsolvent,andasforthefox,itdetectsinasmall rangeofperoxidesconcentrationsandmolarabsorptivityoftheferrilxylenolorangecomplex varieswithdifferentproceduresofmakingthedye.nuchietal.(2009)concludedthatfox results (from degradation of fat for feed uses) correlated better with other oxidation parametersthantraditionaliodometry. Iodideoxidationmethod A spectrophotometric iodidedependant method has also been set to determine hydroperoxidecontent.inthismethodology,notsocommonlyused(watanabeetal.,2010), thelipidsampleisplacedinanacidicsolution,whichisthenmergedwithiodide.thelipid hydroperoxideoxidisesiodidetoiodine.then,generatediodineandiodide(inexcess)reactto givetriiodideanion,whichisdetectedspectrophotometricallyat350nm.bloomfield(1999) usedfe(ii)asacatalyst.theclosedconditionspreventinterferencefromatmosphericoxygen andtheshortreactiontimeminimisesinterferencefromsidereactions. 55

Results 2.1.3Chromatography Methodologies explained up to here are in general quite simple regarding theory base, implementationoftheprocedureandulteriorinterpretationofthedata,presentinglowto moderateselectivityandsensitivity,though.ontheotherhand,chromatographictechniques arefarmoreaccurate,sensibleandspecificforthecompoundininterest,allowingbetter identification of individual products. Indeed, their implementation for hydroperoxides determinationinsteadofthatofvolumetricandspectroscopicmeasurementsisgrowingup more and more over the last years. As an unavoidable consequence, chromatographic methods usually require long or meticulous experimental work, precise control of the experimentalconditionsandthedataprocessingpresentssomecomplexity. Liquidchromatography High Performance Liquid Chromatography (HPLC) is being recently used to determine hydroperoxides.thismethodishighlysensitiveandprettyversatileconsideringbothcolumn and detector properties, allowing to analyze compounds with different characteristics of volatility,molecularweightorpolarity.ontheotherhand,samplepreparationisfrequently tediousandusuallyrequireslipidextraction.zebandmurkovic(2010)foundtheisocratic HPLCESIMSausefulmethodfortheidentificationandcharacterizationofoxidizedspeciesof triacylglycerols(tags),i.e.monoandbishydroperoxides.gotohetal.(2011)developeda methodformeasuringtheperoxidevalueincoloredlipidsonthebasisofthereactionwith triphenylphosphine,formingacompoundwhichabsorbsat260nm.samplethenunderwent HPLC separation and UV detection. Ferrous oxidation mediated methods have also been adaptedtohplcseparation(sugino,1999).specifichydroperoxidesgeneratedfromsterols can also be assessed by liquid chromatography. Saynajoki et al. (2003) determined stigmasterolhydroperoxidesbymeansofanormalphasecolumnandtwotypesofdetectors (UVandfluorescence). Gaschromatography Gaschromatographycoupledtomassspectrometry(GCMS)canalsobeusedfortheanalysis oflipidhydroperoxides,butduetotheirthermolability,previousreductionisneeded.this, along with the prior lipid extraction and subsequent derivatization step, makes it a cumbersomeandtimeconsumingmethod(lagardaetal.,2003). 56

Results 2.2Conjugateddienes/trienes Hydroperoxideformationfrompolyunsaturatedfattyacidsisgenerally(over90%ofthecases) accompaniedbystabilizationoftheradicalstateviadoublebondrearrangement(electron delocalization), which gives rise to conjugated dienes and trienes. These relatively stable compoundsabsorbintheuvrange(235nmand270nmrespectively)andthisabsorptioncan bemeasuredbyspectrophotometrictechniquestoassessoxidationlevel(laguerreetal., 2007;ShahidiandZhong,2005).Thistechniqueissimpleandrapidbutnotaswidespreadas determinationofperoxidesdeterminations,probablybecauseitcanleadtounderestimation sinceoleicacidhydroperoxides,containinglessthantwodoublebonds,cannotbedetected. Ontheotherhand,overestimationispossibleifconjugateddoublebondsarepresentinthe original fatty acid. Furthermore, it is not suitable for oils that have been heated under conditionsthatdecomposehydroperoxidesbecauseinterferencemayoccurwithabsorptionof carbonylcompounds(frankel,1998).evenso,anumberofstudieshaveusedthemforthe monitoringoflipidoxidationduringheatingtreatments,especiallyinvegetableoils(maggioet al.,2011;karouietal.,2011;moralesetal.,2003).correlationbetween235nmabsorption valuesandperoxidevalueshasbeenreported(wanasunduraetal.,1995). 3.SecondaryOxidationProducts Lipidprimaryoxidationproductscangenerate,ifsubmittedtofurtheroxidationconditions, secondaryoxidationproducts,includingaldehydes,ketones,epoxides,hydroxycompounds, oligomers and polymers. These compounds show a wide variety of physicochemical properties,differingmainlyinvolatility,polarityandmolecularweight.mostrelevantgroups ofcompoundswillbecommented(aldehydes,volatilesandpolymers),aswellasaparticular moleculeveryfrequentlyusedasoxidationmarker(malondialdehyde). 3.1Malondialdehyde Malondialdehyde(MDA)isoneofthemostabundantlygeneratedaldehydesduringsecondary lipidoxidationanditisprobablythemostcommonlyusedasoxidationmarker,too. 3.1.1UVVisSpectroscopy ThemostwidelyemployedmethodfordeterminationofMDAisthespectrophotometric determinationoftheredfluorescentmdathiobarbituricacid(mdatba)complex. ReactionoccursbyattackofthemonoenolicformofMDAontheactivemethylenegroupsof TBA,atlowpHandhightemperature,givingthementionedchromophorewhichoffersa maximumabsorbancepeakat532nm.reactionkineticsdependsontheconcentrationoftba 57

Results solution,temperatureandph(fernándezetal.,1997).severalvariationsofmdatbamethod exist,withdifferentprocedurescurrentlyperformedinfoodanalysis:directheatingofthe sample,sampledistillation,lipidextractionwithorganicsolventsoraqueousacidextraction, followedbyacidreactionwithtba.generalprocedureusuallyconsistsofhomogenizationand centrifugation at acidic medium (usually provided by trichloroacetic acid) and posterior reactionwithtbaathightemperatures(around90100 C).Nevertheless,thereisquitealot ofvariabilityinreactionconditions,suchasheattreatmentexposuretime;toillustrateit: Berasategietal.,Peirettietal.,Jungetal.andJongbergetal.(2012;2011;2011;2011)left mixturereactatboilingwaterbathfor15,20,30and40minutes,respectively.ontheother hand,trichloroaceticsolutionconcentrationshavealsobeenreportedtobedifferent(from3% to15%w/v)amongworks(maqsoodetal.,2012;leygonieetal.,2011). TraditionalspectrophotometricTBAtesthasbeencriticisedforsomereasons.Firstly,TBAis not selective to MDA, since it also reacts with many other compounds, such as other aldehydes,carbohydrates,aminoacidsandnucleicacids(salihetal.,1987),interferinginthe TBAassayandresultinginconsiderableoverestimation,aswellasvariabilityintheresults.This iswhyitisalsoknownastbareactivesubstancesmethod(tbars).thereisalsoariskof underestimatingtheresponsesincemalondialdehydecan,underinvivoconditions,formlinear orcyclicalschiffbases,orevencrosslinkedbonds,withlysineandargininefromproteins.so poorquantificationsensitivityandpoormolecularspecificityandselectivitycanbeattributed tothismethod.furthermore,thehightemperatures(95 100 C),extendedincubationtimes andstrongacidicconditionscommonlyrequiredforthereactionofmdawithtbamaycause anartifactualperoxidationofsampleconstituentseveninthepresenceofaddedantioxidants. Notefinallythatmalondialdehyde,whichismainlyformedfromlinolenicacidoxidation,does notoccurinotheroxidizedlipids(especiallywhenonlyonedoublebondispresent,i.e.,oleic acid).so,itisoftenaminorsecondaryoxidationproduct,spoilingtheroleoflipidoxidation markerroleusuallyassumedforthiscompound. Despitethementionedlimitations,conventionalspectrophotometricMDATBAmethodsare preferredbecauseoftheirsimplicity.infact,ithasbeenrecentlysuggestedasamoreaccurate andsensitiveparameterinassessmentofoxidativedeteriorationthanpanisidinetestand hexanaldetermination(nuchietal.,2009;pignolietal.,2009). 3.1.2Chromatography Toovercomesomeoftheselimitations,moreadvancedchromatographicdeterminationshave beendeveloped.thesetechniquesprovide,asinthecaseofhydroperoxidesmeasurement 58

Results (section2.1.3)moreaccuracy,sensitivityandspecificityformda.harderexperimentalwork, andcertainlevelofcomplexityindataprocessingarethedrawbacks. Someofthem(StalikasandKonidari,2001;Jardineetal.,2002;delasHerasetal.,2003;Cesa, 2004;Seljeskogetal.,2006;Mendesetal.,2009)involvetheformationofMDATBAcomplex, purificationbychromatography(gcorhplc)andsubsequentdetectionbyms,uvvisor fluorometricdetector.andsomeothersusederivatizationofmdainsteadofreactionwith TBA,inordertoobtainadetectablecompound.Reactionwith2,4dinitrophenylhydrazine (DNPH) or pentafluorophenylhydrazine and conversion into pyrazole and hydrazone derivatives are the most commonly used procedures with HPLC separation and spectrophotometric/fluorometricdetection(mendesetal.,2009;mateosetal.,2005;ichinose etal.,1989).ontheotherhand,conversionintotetramethylacetalormethylpyrazoleismore commonwithgcseparation,withflameionizationdetector(fid)ornitrogen/phosphorus specificdetector(ichinoseetal.,1989). Mendesetal.(2009)andMarcincaketal.(2006) comparedtwohplcseparationmethodsformdadetermination(mdatbaandmdadnph adduct)withthetraditionalspectrophotometricmdatbatest,insamplesofchilledfishand pork. The methods were fast, simple, sensitive and stable and presented overall better performance (based on accuracy, specificity and recovery levels) than the traditional spectrophotometricmdatbatest,althoughmdadnphshowedarelativelyhighlimitof detectionandalowerreproducibilityatlowermdacontentsinstandardsandsamples. 3.2Othersecondaryoxidationcompounds 3.2.1UVVisSpectroscopy AnumberofotheraldehydesapartfromMDAaregeneratedduringlipidsecondaryoxidation. Thespectroscopicmethodusedthepanisidinevalue(PAV)todetecttheirpresenceeven whenitisoneoftheoldestmethodsforevaluatingsecondarylipidoxidation,especiallyinthe analysis of animal fats and vegetable oils. It provides useful information on carbonyl compounds, especially nonvolatile unsaturated aldehydes (such as 2alkenals and 2,4 dienals)becauseitisbasedonthereactivityofthealdehydecarbonylbondonthepanisidine aminegroup,leadingtotheformationofaschiffbasethatabsorbsat350nm.thepanisidine valueisdefinedas100timestheabsorbanceofasolutioncontaining1goffatin100mlof solvent.itisconsideredaverysimpleandrapidmethodology.pvandpavallowcalculating totaloxidation.thisparameter(totaloxidation)combinesevidenceaboutthepasthistoryand presentstateofanoil,soitallowstoestimatetheoverallextentofoxidationinthefood(sun etal.,2011). 59

Results PAVhasbeenrecommendedasagoodcontrolparameterforsecondaryoxidationcontrol sinceitcorrelateswellwithperoxidescontent(foxandpv),tbaandvolatilealdehydes analysis(nuchietal.,2009;tompkinsandperkins,1999).intheresearchfield,ithasremained alittlebackward,infavourofothertechniques(poullietal.,2009). Itiswellknownthatthecolorimetricresponsewithpanisidinevariesaccordingtotheextent ofaldehydeunsaturation.hence,atidenticalconcentrations,theresponseismoreintense withdiunsaturatedaldehydesthanwithmonounsaturatedaldehydes,whichinturnaremore sensitive than saturated aldehydes. Moreover, panisidine reacts with all aldehydes, irrespectiveoftheirorigin.thisisespeciallythecaseforsomephenolcompoundsofvirgin olive oil, such as decarboxymethyloleuropeine dialdehyde, which could interfere in the assessment. Finally, studies on correlations between PAV and the organoleptic quality highlightedtheefficacyofthistestformeasuringoxidationinmanydifferentlipids.however, thesecorrelationsmayvarymarkedlybetweenlipidsandalsoaccordingtotheprevailing oxidationconditions.cautionisthusrequiredwheninterpretingthisindex(laguerre,2007). 3.2.2Chromatography Anumberofothercompoundsapartfromcarbonylsaregeneratedduringlipidsecondary oxidation. Concerningfattyacids,theycansufferoxidationasfreeform,withintriacylglycerolsorbonded tophospholipids).theirsecondaryoxidationproductscanbeassessedbyhplc(rovelliniand Cortesi,2004).However,whilethistechniquemaybeusefultoobtainafingerprintofthe oxidationstatusofthesample,onlyaminorityofsignalscanbeattributedunequivocallytoa specificcompoundbecauseseparationisnotgoodenough.betterquantitativeanalysiscanbe carriedoutbymeansofgcfidandgcmsafterderivatizationintomethylesters(aguirreet al.,2010).developmentoflditofmsandesims(schilleretal.,2002;calvanoetal.,2005; Simasetal.,2010)hasmeantagreatstepforwardinthisfield. EventhoughSterolOxidationProducts(generallyknownasSOPs)presentlowlevelsinfoods, theyshowanumberofharmfuleffectsintheorganism(otaeguiarrazolaetal.,2010),soa significantnumberofstudieshavefocusedtheirattentionintheiranalysis.experimental procedure involves lipid extraction, saponification, purification, derivatization and chromatographicanalysis.thatdeterminationischallenginginmanyways:artifactgeneration, verylowconcentrations,matrixeffects,incompleteidentificationandreporting,tonoteafew (Guardiolaetal.,2004;BuschandKing,2009). GCMSisthemostaccurateandcommonly applied quantification method for this kind of compounds (Johnsson and Dutta, 2006; 60

Results MenéndezCarreñoetal.,2008b;UbhayasekeraandDutta,2009;DerewiakaandObiedzinski, 2010;Xuetal.,2011).Clarianaetal.(2011)foundthistechniquebetterthanGCFIDinastudy performed with pork meat. Due to the necessity of a derivatization process and the impossibilityofanalysingthermolabilemolecules,someliquidchromatographymethodshave beenrecentlydeveloped(kemmoetal.,2008;mazalliandbragagnolo,2009;matsunagaetal., 2009).However,liquidchromatographyshowslowerresolutionthangaschromatography,and thebestwaytoovercomethisproblemiscouplingittoamassspectrometerdetector,which inthiscaseisquitecomplexandstillhasnotbeenwellsolved.anewfastgcmsmethodhas beenrecentlydevelopedandappliedtocholesteroloxidationproductsanalysis,givinghighly promising results (Cardenia et al., 2012). Satisfactory resolution, good repeatability and sensitivity,togetherwiththeconsequentreductionofthetimeofanalysisandconsumables makeitavalidalternativetoconventionalgcms. 3.3Volatiles Underthisgroupofsecondaryoxidationproductsagreatdiversityofcompoundshasbeen included, presenting very different functional groups: aldehydes, ketones, alcohols, short carboxylicacidsandhydrocarbons.theyallsharethepropertyofgivingfrommoderatetohigh smells and are related to rancidity in sensorial tests. Measurement of these secondary oxidation products is of great importance, since their formation closely relates to the deteriorationofflavour.someofthesevolatilecompoundsarehighlyspecifictotheoxidative degradationofaparticularpolyunsaturatedfattyacidfamily:propanalisthemainmarkerof oxidationofn3fattyacids,whilehexanalandpentanalaremarkersofoxidationofn6fatty acids.bothpropanalandhexanalareoftenusedasindicatorsoflipidoxidationinfoods becausetheycanbemeasuredinthesampleheadspaceandtheirlackofdoublebondsmakes themmorestabletowardsoxidationthanunsaturatedaldehydes.nevertheless,hexanalis morefrequentlymeasuredasitsformationishigherthanthatofmostsecondaryoxidation products,apartfromafewexceptions.however,measuringtheextentofoxidationwithjust oneortwomarkersisarathercoarseapproach,somethodsinvolvingassessmentoflargeset ofcompoundsshouldbepromoted(laguerre,2007). Gas chromatography is the preferred method to quantify volatile molecules and mass spectrometry detection contributes to identify them. Different methods may be used to recovervolatileoxidationcompoundsbeforechromatographicanalysis,including:(a)solvent extractionand(b)headspace(hs)techniques. 61

Results (a)althoughliquidliquidextractionsarenotverysuitabletorecoverthevolatilecontent (becausetheyarelong,laboriousandrequireasolventevaporationstep,whichleadsto substantialvolatilecompounddegradation),novelvariantshavebeenrecentlyproposedto overcomesomeoftheselimitations.noteespeciallysimultaneousdistillationextraction(sde) andreducedpressuresteamdistillationextraction(rpde).bothallowtoobtaincompoundsof relativelyhighboilingpoint,butwithrpdeevaporationisreachedwithlowertemperatures, avoidingpossibleartefactformation(varletetal.,2007).sdeandrpdeshowtheadvantageof beingabletoextracthighquantitiesoftargetcompoundssincethevolatilefractionsgenerally havehighsolubilityinorganicsolvents(liuetal.,2010;ningetal.,2011).moreover,ferhatet al. (2007) developed a microwave energymediated extraction method. Liquidliquid extractionsarethepreferredrecoveringmethodswheneverthesamplesrequirederivatization stepprevioustochromatographicanalysis(hplcandgc).dnph,benzyloximeandthiazolidine derivativesarethemostfrequentlyusedcompoundstoimprovestabilityand/ordetectionby visibleultraviolet spectrometry, flameionization, nitrogenphosphorous and mass spectrometrydetection(varletetal.,2007b). (b) HS analysis can be performed by static headspace (SHS), dynamic purgeandtrap headspace(dhs)orheadspacesolidphasemicroextraction(hsspme)techniques.allofthem arepriortogaschromatographyanalysis. InSHSmethod,thesampleisplacedinanairtightvial.Mostcompoundsthatarevolatileatthe analysistemperatureevaporatefromtheliquidorsolidfractionandpassintotheoverhead gashs.atequilibrium,analiquotisharvestedandinjectedonthegccolumn.thismethodis relativelyinexpensiveandeasytouse,itdoesnotrequiresolventextractionandcanbe automated.however,asequilibriumisestablishedbetweenthevolatilecompoundsinthehs andthoseremaininginthesample,onlylowquantitiesofcompoundsareactuallyrecovered, whichlimitsthesensitivity.theincreaseintheextractiontemperaturecouldincreasethe volatilizationofthetargetcompoundsandthusincreasethequantitiesrecovered,butthe temperaturemustbekeptaslowaspossibleinordertominimizegenerationofnewoxidation productsand/orthermaldegradationofoxidationmarkers.anumberofauthors(joaquinet al.,2008;vieiraetal.,2012)haveappliedthismethodinfoodsamplesanalysis. Onthecontrary,DHStechniquedoesnotrequiretheestablishmentofequilibrium:thesample iscontinuallypurgedbyinertgastoextractvolatilecompounds.then,thegaseffluentpasses through a porous polymer trap that collects volatile analytes. Among all available trap materials, tenax is the most commonly used. As volatiles contained in the sample are 62

Results constantlyreleasedandtrapped,ahighconcentrationofcompoundsareinjectedonthegc column. Despite its high sensitivity, the instrumentation is complex and expensive, thus increasingthesourcesoferror(trapdrying,traptransfer,purgingefficiency,etc.)anditisin generaltermsslowerthanshs.nevertheless,severalstudieshavehighlightedtheefficacyof DHSGCinassessingtheoxidativestatusofdifferentfoodmatrix(NielsenandJacobsen,2009; HaarandJacobsen,2008). In SPME analysis, volatile compounds make a first equilibrium between sample and HS, followedbyasecondonebetweenthehsandthecontactfibre(whichiscoatedwithahighly adsorbantpolymericfilm).finally,thefibreisintroducedinthegcinjector.thismethod providesmanyadvantagesoverotherones,includingeasymanipulationandexperimentalset up,shortsamplingtimes,easyautomationandhighsensitivity(iglesiasetal.,2007).anumber of authors have applied this method for food lipid oxidation determinations (Haar and Jacobsen,2008;IglesiasandMedina,2008).Itsmaindrawbackisthatfibredegradationand contaminationoccursquiterapidly,thusreplacementisrequiredperiodically. Recentcomparativestudiesperformedwithallthesemethodsforcaptureofvolatilecontent leadtotheconclusionthateachonepresentsitsshortcomingsandadvantages(shuetal., 2010;Prosenetal.,2010),butHSSPMEisbeingusedtoanincreasingextentonaccountofits mostpromisingresults. 3.4Oligomers/Polymers Duringextendedoxidation,alipidiccompoundcanbelinkedtogetherwithotheroneor severalones,givingrisetodimers,oligomersorpolymers.simultaneousanalysisofoxidized formsoftriacylglycerolsandtheiroligo/polymersisverycommontoassesslipidoxidation progress.monomersareveryreactiveandhighlycorrelatewithperoxidevalue,sotheycould give information about the primary oxidation level of a sample. On the contrary, triacylglycerols oligopolymers are rather stable compounds, being considered as good indicatorsofsecondaryoxidationstatus(bilanciaetal.,2007;gomesetal.,2012). High Performance Size Exclusion Chromatography (HPSEC) has demonstrated to provide satisfactoryresultsintheanalysisofthiskindofoxidationproducts.itallowsseparationand subsequentidentificationandquantificationofmoleculesaccordingtotheirmolecularweight. Itisusuallyperformedonpolarcompounds,soitrequiresapreviouspurificationofthepolar lipid fraction, which is usually done by silica gel column chromatography. Some studies (MarquezRuizetal.,2007;Summoetal.,2010;Caponioetal.,2011)havedemonstratedthe usefulnessofhpsecinthedeterminationofthelevelsoftheoxidativedegradationofavariety 63

Results offoodsamples,andparticularlythatofrefinedvegetableoils,whosetechnologicalprocess involves quality deterioration. Morales et al. (2010) applied it for the determination of advancedoxidationinvegetableoilsthroughthedetectionoffattyacidspolymers.oligomers formationduringthermooxidationofphytosterolshasalsobeenreported(struijsetal.,2010; MenéndezCarreñoetal.,2010;Rudzinskaetal.,2009;Rudzinskaetal.,2010)bymeansof HPSECanalysis. 4.Alternativemethodologies Theprevioustechniquesareeithertooempiricalorhighlydependantonseveralexperimental factors,suchastechnicianskill,lightexposureandatmosphericoxygen,apartfromthefactof beingtimeconsuming.toavoidtheselimitations,variousmethodologieshavebeenproposed asgoodalternativesinanalysisofbothprimaryandsecondaryoxidationproducts.theyare basedondirectspectroscopicanalysesofsamples,suchasmagneticresonance,fluorescence andvibrationalspectroscopy,andonchemiluminescentproperties.asgeneralgoodpoints, preliminarytreatmentisminimalorunnecessary,lowamountofsampleisrequiredandhighly specificresultsareobtained. 4.1Chemiluminescence Certain chemcal reactions generate electromagnetic radiation. This emission of energy is knownaschemiluminescence(cl)anditcanbeappliedtodetectandquantifycompoundsof interest.however,lightintensityisverylow(ultraweakclisaccompaniedduringoxidationof hydrocarbonsandlipids(navasandjimenez,1996)),solightamplifiersshouldbeintroduced toincreaseit.oneofthemostcommonlyusedoneistheluminol.theluminolenhanced chemiluminescenceinvolvesoxidationofluminolinbasicsolutiongeneratingafreeradical intermediatewhichreactswithfluxofoxidizingagents(activefreeradicals)presentinthe system,e.g.lipidhydroperoxides.thisleadstoformationofluminolderivedproductinexcited state,whicheventuallyreturnstogroundstateemittingstrongbluelightat430nm(roginsky andlissi,2005).differentversionsofthismethoddifferinthetypeofactivefreeradical producedandthewayoffreeradicalproductionaswellasindetailsoftheprocedure. Robinsonetal.(1997)suggestedtheadditionofpiodophenoltoprovidemoreintensive, prolonged,andstablelightemissionascomparedtothetraditionalluminolsystem.more recently,anewchemiluminescencemethodinnonaqueousmediumclwasdevelopedto detectlipidperoxidesinvegetableoils(szterkandlewicki,2010),presentinggoodcorrelation withspectrophotometricpvanalysis. 64

Results Bajetal.(2009)discoveredthatpartialexclusionofoxygenfromthereactionmediumstrongly influencedthelightintensityoftheluminolreaction,andtheeffectisdependentonthe oxidant analyzed, so an alternative mechanism was suggested for some oxidant species. Besides,theystatedthattheoxygenconcentrationalwaysaffectsthereproducibilityofthe results, so equilibrating the working solutions with oxygen or air should always lead to improvedresults. TheattractivefeaturesofCLmethodsaretheirhigherquickness(takingonlyafewminutes), sensitivity (picomol levels have been assessed), low sample requirements, low cost and simplicityascomparedwithothermethods(rolewskietal.,2009).asforshortcomingsofthis kindofmethods,firstofall,thekinetictheoryandmechanismforchemicalprocessesresulting inclisnotknownindetail.thismaymeanproblemswithdatainterpretation.furthermore, thismethodisnotspecifictothelipids(otheroxidizingagentsalsogivesignal);butthis opportunitycanbeseizedtoestimatetheoveralltotaloxidantstatusofthesample. BuntingandGray(2003)developedanautomatedflowinjectionchemiluminescencesystem formeasuringlipidhydroperoxideconcentrationsinoilsandfoundgoodagreementwitha traditionaliodometrictitrationassay,whatcoulddenotetheusefulnessofclmethodsto assesslipidprimaryoxidation;andalsoinvegetableoils,yangetal.(2010)foundasimilar trendfortbarsandclmeasurementsduringoxidation. 4.2Fluorescencespectroscopy When a compound is irradiated with an electromagnetic energy source, some of their electronspromotefromtheirfundamentalstatetoanexcitedone,andsubsequentlythey returntotheiroriginalstate,reemittingtheenergypreviouslyabsorbed.nevertheless,certain compoundscanlosesomeofthatenergyasheat,whatallowstheirelectronstoreturntoa higherlevelthantheoriginalone,soemittedlightisinthiscaselowerthantheabsorbedone. This phenomenon is named as fluorescence, and compounds presenting this property, fluorescents.beamoflightisusuallyfromtheuvrangeandemittedenergyistypically,but not necessarily, from the visible range. It can be used in analytical chemistry for both qualitative and quantitative determinations, as well as in isolated and coupled to chromatographyequipments. Regardingfoodfield,itsimplementationisgrowingupmoreandmore(KarouiandBlecker, 2011).Thefreeaminogroupsofproteinscanreactwithaldehydesfromlipidperoxidationor reducing sugars to give Schiff bases. These compounds present a high colour intensity (browning)andcharacteristicfluorescencespectra(excitationandemissionwavelengths,and 65

Results fluorescenceintensity)accordingtothetypeofproteinandadduct.althoughitssensitivityis high,excitationandemissionwavelengthmaximavarydependingonthefoodsampleandthe procedurefollowed.theyrangefrom250nmto500nmforexcitation,andfrom280nmto 600nmforemission(Poullietal.,2009:Tironietal.,2009;Elmnasseretal.,2008;Gatellieret al.,2009).manyauthorshaveusedtheabilityoftheseschiffbasestoemitfluorescenceto monitorthermaloxidativeprocesses,especiallyindairyproducts(schambergerandlabuza, 2007;Dalsgaardsetal.,2011),meat(Gatellieretal.,2007;Chelhetal.,2007),fish(Naseriet al.,2011;nguyenetal.,2012)andoils(barrettetal.,2011),butfluorescencemethodologies arestillpoorlydocumentedinfoodlipidoxidationanalysis.bothgatellieretal.(2009)and Nguyenetal.(2012)foundahighcorrelationbetweenfluorescentpigmentsandTBARSof meat and fish products, which demonstrated that the interaction between proteins and aldehyde products of lipid oxidation is mainly involved in the production of fluorescent pigmentsandthesearegoodmarkersoflipidoxidation. AdifferentimplementationoffluorescentpropertieswasdevelopedbyAndersenetal.(2008) withacheesesample.theymeasuredthefluorescenceofthephotosensitizersinvolvedinthe lipidoxidationmechanismofthecheeseandusedthespectratosuccessfullypredictthe contentofvolatilecompounds. 4.3Infraredspectroscopy Infrared(IR)spectroscopyisalsoknownasaveryhelpfulwaytostudylipiddegradationunder oxidativeconditions(kongandsingh,2011),particularlysinceitisaneasy,rapid,economical andnondestructivetechnology.itisbasedonthedeterminationoffundamentalvibrational transitionsofaparticularcompoundandinvolvestheabsorptionofdiscreteenergylevelsfrom theirregion.thesediscreteenergylevelsarecharacteristicofeachofatomatomlinkage,so studyingtheirspectrumcanprovideenoughinformationtofindoutthenatureofthe analyzed compound. Mathematical tools, such as Fourier Transform (FT) or chemometric methods,permitdataprocessing.continuousageingmonitoringcanbecarriedoutwiththis methodology, although for the moment, most of the works have been assessed in discontinuousway.someadvanceshaverecentlybeenperformedregardingtechnological devices(garcíagonzálezandvandevoort,2009). IRhasbeenappliedtomeasuretheperoxidevalueinoxidizedlipids(Guillénetal.,2007)and differenceswerefoundintheirspectraoffreshandagedoils(christyetal.,2003;rusaket al.,2003);soirspectracanbeusedtocharacterizetheagingofvariousedibleoils(yangetal., 2005;Muiketal.,2007;leDreauetal.,2009;Wangetal.,2011;Beltránetal.,2011).The 66

Results investigationoftheftirspectraofthetreatedoilsrevealedthatthemicrowaveheatingofoils (MoharamandAbbas,2010)causedsignificantchangesintheintensitiesoftheirabsorption bandsandproducednoshiftsinthepositionofthebands.thesechangeswereattributedto thereductionin18:2and18:3fattyacidscontentduetotheoxidation. Ithasalsobeenusedfortheanalysisofedibleoils(Belhajetal.,2010),horsemackerelpatties (Giménezetal.,2011)andcannedtomatojuice(Rubioetal.,2010),incombinationwithother analyticalmethodswhichleadtosimilarconclusions,andthereforeprovidingmarkerbandsto improvetheunderstandingofchemicalchangestakingplaceduringprocessingandstorage. 4.4Ramanspectroscopy Ramanspectroscopyalsodetectsfundamentalvibrationaltransitionsalthough(contraryto infraredspectroscopy)notbymeansofdirectenergyabsorption,butthroughanenergy (originatedfromauv,visibleorirlaser)scattering:promotiontoavirtualvibrationalstate andsubsequentrelaxationtoafundamentalvibrationalstatedifferentfromtheoriginalone. Therefore, Raman and IR spectroscopy are complementary techniques and provide complementarystructuralinformationaboutmolecules.actually,onlysomemoleculesshow Ramanscatteringproperties,andmostofthemataverysmallintensity,soquitesophisticated andexpensiveopticaldetectionequipmentsarerequired.thisreducesitspracticalusetoa fewcases.indeed,itisstillverysparinglyusedinthefoodfield,inspiteofitsinteresting characteristics,whichincludebeingnondestructive,fast,relativelyinexpensive,noninvolving chemical products, requiring very little sample preparation, being highly sensitive to unsaturations and poorly sensitive to water (Reid et al., 2003; Herrero, 2008). Two instrumental methods can be employed with Raman spectroscopy: confocal Raman spectroscopy with a powerful laser in visible range and Fourier Transform Raman spectroscopy.mostoftheapplicationsonoilshavebeenperformedbythelater(korifietal., 2011).However,aportableRamanspectrometerhasbeenrecentlydeveloped(Guzmánetal., 2011),which,ontheotherhand,showslowerresolutionthanclassicones.Zhangetal.(2010) reportedthefirstproofofconceptstudyofsurfaceenhancedramandetectionofatbamda adductusingsilvernanoparticlesastheserssubstrate Ramanspectroscopyresultsandoxidationlevelswererelatedinlipidsextractedfromseveral meatandfishproducts(herrero,2008;sarkardeiandhowell,2007).inlinewithperoxide valuesrises,ramanspectradatashowedanincreaseinparticularbandsandregionsofthe spectra of oils extracted which could be attributed to alterations in lipids structure. Furthermore,Ramanspectroscopycouldbeanalternativetogaschromatographicfattyacids 67

Results analysis,sinceitsuccessfullypredictedtotalunsaturationandindividualcompoundsseveral meatproducts(beattieetal.,2006).salmonramanspectra(herreroetal.,2009)indicated differences in the fat fraction (as well as in protein fraction) in coldsmoked products. Regardingvegetableoilsstudies,Muiketal.(2005)detectedformationofaldehydesand conjugateddoublebondsystems,aswellasisomerizationofcistotransdoublebonds.the timedependentintensitychangesincertainramanbandswerecomparedtoconventional parametersusedtodeterminetheextentofoxidationinoils,suchasanisidinevalueandk 270, andshowedgoodcorrelation.elabassyetal.(2009)assessedfattyacidcontentinoliveoil. Zhangetal.(2010)developedamethodtodetermineMDAinamodelsystembymeansofthis technique.theyfoundthatitwasselectiveandspecificformdatbaadductsintermsof differential spectra and high response versus adducts formed by TBA and other TBARS differentfrommda.besides,theyachievedbettersensitivitythaninworksusinguvvisor fluorescencedetectors.sometimes,reductionofcarotenoidscontentmeasuredbyraman spectroscopyhasbeenusedtomonitorlipidoxidationprocess(kathriveletal.,2008). SimultaneousanalysisoftheoxidationofedibleoilshasbeenalsoperformedbyInfraredand Raman techniques (Muik et al., 2007). These techniques led to improved information comparedtoisolatedanalysisconcerningassignmentofpeaks,andtherefore,compounds formedduringoxidation. 4.5MagneticResonance ThebasisofNuclearMagneticResonance(NMR)reliesonthepropertyofcertainatomsof absorbing and reemitting energy in the presence of a strong magnetic field due to the excitation of their atomic nuclei. This energy is at a specific resonance frequency which dependsonthestrengthofthemagneticfieldandonthemagneticpropertiesoftheparticular isotopeoftheatominstudy.theenergyabsorptionsoftheatomicnucleiareaffectedbythe nuclei of surrounding molecules, which cause small local modifications to the external magneticfield.promisingresultsareobtainedbythisalternativemethodologyconsidering reliabilityandspecificityofthedatasincetheyprovideanaccuratefingerprintofthesample.it doesnotrequireextensivemanipulationofthesample,thuspreservingmolecularintegrity, andallowingdetectionofallthesubstancespresentinthesampleatthesametime.this,in additiontoitshighsensitivityevenincomplexmatrices,highlightsthenecessityofimproving andspreadingitsuse.however,thatisaveryexpensivemethodologyandrequiresspecial skillstointerpretthespectra.theuseof 1 Hand 13 CNMRspectroscopyinfood,appliedby differentresearchgroups(guillénandruiz,2008;elhajjoujietal.,2008;dybviketal.,2008; 68

Results Tyletal.,2008;Colzatoetal.,2011;Scanoetal.,2011;GuillénandUriarte,2012a;Guillénand Uriarte,2012b;AlonsoSalcesetal.,2011),hasprovedtobeveryusefulinevaluatingthe oxidativestatusofthelipidfraction,aswellasinprovidinginformationonthenature(main functionalgroups)andconcentrationofthecompoundsfound(i.e.hydroperoxides,carbonyl compoundsanddienes).itisconsideredavaluabletoolforquantificationofoxidationoffood lipids(namaletal.,2007),andgoodcorrelationwithconventionalanalysissuchastbahas beenreported(deoliveiraetal.,2011). SeveralmultidimensionalNMRtechniqueshavebeendevelopedinlastyears(correlational spectroscopy,nuclearoverhausereffectspectroscopy,diffusionorderedspectroscopy ).They allowabetterassignmentthantheonedimensionalspectra,improvingthecharacterizationof foodlipidsamples(scanoetal.,2011;hatzakisetal.,2011).however,themaindifficulty derivedfromtheapplicationofthesetoolsisthehightimerequiredfortheacquisition. ThebasisofElectronParamagneticResonance(EPR)isthesameasthatofNMRbutinthis case,energyexcitesspinsofsingleelectrons.so,onlymoleculespresentingsingleelectrons (thatis,radicals)haveeprspectra.ithasbeenusedtodetectoxidantintermediatespeciesin foodmatrices(szterketal.,2011;huvaereetal.,2011).however,theseradicalsshowquite shortlivesunlessverylowtemperaturesareguaranteed(huvaereetal.,2011;geoffroyetal., 2000; KamalEldin and Min, 2010). In an attempt to avoid this problem, some recently developedmethodologiesdealwiththedetectionofunstablefreeradicals.amongthem,spin trappingtechniquesallowtheindirectdetectionoflipidderivedradicalsbyformationofstable spinadductsthatcanaccumulateindetectableconcentrations.thisway,bothidentification and quantification of these intermediates is possible.traps are not radicalspecific, neverthelessparticulartrapsareconsideredmoreorlessusefulfortrappingparticularradicals. CompoundssuchasPBN(phenyltertbutylnitrone)andDMPO(5,5dimethyl1pyrrolineN oxide)arefrequentlyusedforthatpurpose(szterketal.,2011;papadimitriouetal.,2006). Combinedapplicationofbothmethodologies(NMRandEPR)isofgreatinterest.Inthissense, Silvagnietal.(2010)usedtheminastudyinvestigatingthekineticsofthermallyinducedlipid peroxidationofpeanutoil.theuseofeprallowedthemtodeterminetheprimaryalkyl radicals,andprovidedanestimationoftheradicalgenerationrate;whereasbymeansofnmr, simultaneouslydetectionofprimaryandsecondaryoxidationproductswasperformed,thus allowingamoredetailedkineticinvestigation. 69

Results 5.Conclusion Differentkindofcompoundscanbeusedaslipidoxidationmarkersinfoodsamples,among whichhydroperoxidesandavarietyofaldehydesarethemostcommonones.eachoneof themisindicativeofaparticularstateofoxidation,sochoosingjustoneparametertoanalyse theoxidativestatusisratherdifficultanditisfrequentlymoreconvenienttocombinedifferent methods.therefore,analystmustchoosecarefullythemostadequateforhispurpose,taking intoconsiderationthemostsuitablemoleculesandexperimentalconditionsrequiredineach case. First general decision is whether determining primary or secondary oxidation compounds,consideringmainlytheextentofoxidation.afterwards,precisionrequiredand characteristicsofthefoodmatrixmustbeconsideredtofollowonemethodologyoranother.a variety of conventional and alternative methodologies have been developed and implemented. Considering the later, they have been proven to provide interesting and promisingresults,soattentionmustbepaidtothesealternativetechniquesintheareaoffood lipidoxidation. 6.Acknowledgements We thank the Programa ConsoliderIngenio 2010 CARNISENUSA CSD 200700016, the ProyectoAGL200801099/ALI (MinisteriodeCienciaeInnovación),and PlanInvestigador delauniversidaddenavarra (PIUNA)fortheircontributiontothefinancialsupportofthis work.b.barriusoisgratefulto CátedraTomásPascualSanzUniversidaddeNavarra andto AsociacióndeAmigosdelaUniversidaddeNavarra forthegrantsreceived. 70

Table1.Characteristicsofthedifferentmethodsforanalysisoflipidoxidationinfoodsreviewedinthisarticle. Method Analyte Sample preparation Amountof sample Sensitivity Specificity Cost Limitations Titration Peroxides MediumShort 1g Mediumlow Mediumlow Low UvVis a spectroscopy Chromatography Peroxides, *Conjugated dienes/trienes, *MDA, aldehydes Peroxides, MDA,SOPs, volatiles, oligomers Medium 500mg Medium Medium Low Long 1100mg Highveryhigh (dependingon thedetector) Highveryhigh (dependingon thedetector) Chemiluminiscence Peroxides Short 1200mg High Medium Low Fluorescence IR b spectroscopy Ramanscattering Nuclearmagnetic resonance Electron paramagnetic resonance a Ultravioletvisible b Infrared Aldehydesand volatiles Peroxides, unsaturations, MDA Peroxides, unsaturations, MDA Peroxides, aldehydes, dienes High Veryshort 1050mm 2 Veryhigh High Medium Veryshortnone 240mg Mediumhigh High Medium Veryshortnone 1050mm 2 Mediumhigh High Low Veryshortnone 10200mg High Veryhigh Radicals Veryshortnone 100900mg High High Very high Very high Reagentssusceptibleto oxidation AbsorptionbyUFA Drynessrequired Highamountofsolvents Lowconcentration range Variabilitydependingon thedye *Insensitivetooleicacid Laboriousexperimental procedureanddata processing Unknownmechanisms Lightamplifiers required Variabilityin wavelenghts Nonaqueoussolutions required Somemoleculesare inactive Complexdata interpretation Complexdata interpretation Most relevant references AOAC,2000 Bou,2008; Maggio,2011; Berasategi, 2012;Nuchi, 2009 MárquezRuiz, 2007; Zeb,2010; Mendes,2009; Derewiaka, 2010 Rolewski,2009 Gatellier,2007 Yang,2005 Muik,2005 Tyl,2008; Namal,2007 Szterk,2011; Geoffroy,2000

ResultsII Poster1 Determinationofcholesteroloxidationproductsin foods:improvementofcosttimeefficiency

ResultsIII Paper2 Interlaboratoryharmonizationtrial

Results (Underpreparation) Interlaboratoryharmonizationtrial Otherauthors,BlancaBarriuso 1,DianaAnsorena 1, IciarAstiasarán 1,otherauthors 1 DepartmentofNutrition,FoodScience,PhysiologyandToxicology.FacultyofPharmacy,UniversityofNavarra 79

Results Introduction SterolandSOPsanalysisarecomplexprocedures,usuallyinvolvingthefollowingsteps:lipid extraction, saponification, purification by solid phase extraction, derivatization and chromatographicanalysis.agreatvariabilityintheaccomplishmentofthesestepsisnowadays found among the different research groups. An interlaboratory harmonization of the methodologieswasthenconsideredofinterest.thus,theaimofthisworkwastocomparethe analyticalresultsobtainedbyupto17differenteuropeanlaboratoriesinthedeterminationof sterolsandoxysterols.foursterols(cholesterol,campesterol,sitosterol,sitostanol)andtwo oxycholesterols(7hcand7hc)weredeterminedintwoserumsamples(aandb).samples werepreparedbythe ReferenceInstituteforBioanalytics(RfB),inBonn(Germany),and shippedatroomtemperaturetothelaboratoriesincludedinthestudy. Resultsandpreliminarydiscussion Figure1showsthedifferentconcentrationsobtainedforeachofthecompoundsanalysedby thegroupsincludedinthestudy.differentcoloursforthespotsindicatetheapplicationof different methodologies. A great dispersion of the data could be observed for all the compounds,asthehighcoefficientsofvarianceconfirmed(table1).thisdispersionwasnot dependent on the methodology applied since data were not grouped according to the differentmethodologiesapplied.thiscouldindicatethatparticularstepsoftheexperimental proceduresarekeyfactorsandshouldbetakenintoaccounttoclassifytheresults.thus, furtherinformationwasrequiredandthedetailedprotocolsofeachlaboratorywerenecessary tobecollected. Moreover,oxycholesterolslevelsinBserumsamplewere,forsomelaboratories,outoforder comparedtotheotherlaboratoriesresults.thiswasattributedtoapossibleoxidationofthe sampleduringthetransport,asaconsequenceofpoorpreservationconditions.someauthors highlightedtheliophylizationprocesstowhichsamplesweresubjectedbeforethedelivery,as alikelycauseofthesampledeterioration. Allthesequestionsarenowbeingconsideredandadeeperdiscussion,pointingoutfinal conclusions,isstillunderpreparationbythecoordinatorofthestudy. 80

Results cholestero campesterol -sitosterol sitostanol 7-HC 7-HC Figure1.Concentrationofcholesterol,campesterol,sitosterol,sitostanol,7HCand7HC,obtainedforserum samplesaandb,indifferentresearchlaboratories.differentspotcolordenotesdifferentmethodologiesapplied. 81

Results Table1.Statisticalparametersforthedeterminationofa)cholesterol,b)campesterol,c)sitosterol,d)sitostanol, e)7hcandf)7hc,obtainedforserumsamplesaandbinalltheresearchlaboratories. Compound Number of participants SampleA SampleB Mean Sd Cv Mean Sd Cv Cholesterol(g/L) 10 1.68 0.422 25.1 2.11 0.455 21.5 Campesterol(µg/dL) 16 597 321 53.8 760 410 54.0 sitosterol(µg/dl) 17 334 335 100 401 353 88.0 Sitostanol(µg/dL) 8 223 421 489 240 425 177 7HC(µg/dL) 7 39.7 27.2 68.5 48.8 28.9 59.2 7HC(µg/dL) 7 47.6 40.3 84.7 62.3 52.1 83.6 82

ResultsIV Paper3 Sterolsheating:Degradationandformationoftheirring structurepolaroxidationproducts

Results FoodChemistry(2012),135,706712 Sterolsheating:Degradationandformationoftheirringstructurepolar oxidationproducts BlancaBarriuso a,aneotaeguiarrazola a,maríamenéndezcarreño a,b, IciarAstiasarán a,dianaansorena a a DepartmentofNutrition,FoodScience,PhysiologyandToxicology.FacultyofPharmacy,UniversityofNavarra b DepartmentofBiochemistryandCellBiology,FacultyofVeterinaryMedicine,UniversityofUtrecht Abstract Cholesterolandphytosterolscansufferoxidationunder heating conditionstogivesterol oxidationproducts(sops),knownbytheirtoxiceffects.thispaperstudiedthedegradationof cholesterolandthreeplantsterolsduringa360minheatingtreatment(180 C).Theformation and further degradation of SOPs was also analysed by GCMS. Results revealed a sterol susceptibilitytodegradationaccordingtothefollowingdecreasingorder:campesterol sitosterol stigmasterol > cholesterol. Their degradation curve fit (R 2 = 0.9070.979) a logarithmicmodel.steroloxidationproductsincreasedtheirconcentrationduringthefirst5 10minandthereafter,theirdegradationratewashigherthantheirformationrate,resultingin a decrease over time. Irrespective of the sterol from which they had derived, 7keto derivativespresentedthehighestlevelsthroughouttheentireprocess,andalsosopswiththe sametypeofoxidationfollowedasimilardegradationpattern(r=0.900.99). Keywords:Cholesterol;Plantsterols;Heatingstability;Degradation;Kineticmodels Highlights 1.Steroldegradationat180 Cwas:campesterolsitosterolstigmasterol>cholesterol. 2.Sterolsdegradationcurvefit(R 2 =0.9070.979)alogarithmicmodel. 3.ThehighestSOPsconcentrationwasachievedat510min. 4.7ketoderivativespresentedthehighestlevelsthroughoutthe360mintreatment. 5.SOPswiththesametypeofoxidationfollowedasimilardegradationpattern. 85

Results 1.Introduction Cholesterolandplantsterolscansufferautoxidationtogivesteroloxidationproducts(SOPs), namedcops(cholesterol)andpops(phytosterols),respectively.someofthesecompounds havebeendemonstratedtoexertharmfuleffectsintheorganism,includingatherosclerosis, cytotoxicityandmutagenesis(larssonetal.,2006;roussietal.,2007;o'callaghanetal.,2010). The presence of polar SOPs has been widely reported in a great variety of foods, from vegetableoilstoporkmeat(otaeguiarrazolaetal.,2010).particularattentionhasbeenpaid tothedevelopmentofphytosterolenrichedfunctionalfoodsduetotheirmuchhighersterol contentcomparedtoconventionalfoods.asaconsequence,finaloxysterolarealsoincreased, reaching up to 35fold compared to the nonenriched foods (Conchillo et al., 2005; MenéndezCarreñoetal.,2008a).Assessingtheperspectivesofplantsterolsenrichedfood, particularlyfocusingontheoccurrenceofplantsteroloxidationproductsisofgreatinterest (GarcíaLlatas and RodríguezEstrada, 2011). Not only their formation but also their degradationpatternsshouldbestudiedindepthtopredicttheirlevelsinfoods.inorderto avoidinterferencesofthefoodmatrixandidentifytheirnetinfluenceontheprocess,model systemsarecommonlyusedsothatabettercomprehensionofthefactorsgoverningthisissue can be obtained and a better control of certain food processing conditions might be consequentlyproposed.furthermore,thereisalsowideevidenceofthecapabilityofthese compoundstobeabsorbedfromthediet(stapransetal.,2005).giventheabovementioned harmfuleffectsinhumancelllines,theirformationshouldbeminimisedandtheresponsible factorsshouldbestudiedindetail. Inthissense,factorssuchashightemperatureandexposuretooxygenorlightareresponsible foroxysterolsformation.although7ketooxiderivativesarenormallythemostabundantones asaconsequenceofheattreatments,7hydroxy,7hydroxy,5,6epoxy,5,6epoxyand 5,6,7triol derivatives are usually analysed, too (Lampi et al., 2002; Xu et al., 2011). Degradationofbothsterolsandoxysterolsoccursover150 C,givingrisetofragmented phytosterolmolecules,volatilecompoundsandoligomers(rudzinskaetal.,2009;struijsetal., 2010).Onehundredeightydegreecelsiusisthetemperaturecommonlychoseninmostofthe studiestoevaluatesterolthermalsusceptibilityasitrepresentsthemoreusualtemperature appliedinfryingculinaryprocesses.atthistemperature,menéndezcarreñoetal.(2010) found a progressive decrease in stigmasterol content and formation and subsequent degradationofsopsafteronehourofheating.nevertheless,thissopsdegradationhasbeen reportedtostartatdifferentmomentsoftheheatingprocess(kemmoetal.,2005;xuetal., 2011)dependingontheexperimentalprocedure,evenwhensametemperatureisapplied. 86

Results Comparisonbetweenoxysterolsderivedfromdifferentinitialsterolsisamatterofgreat interest, which could contribute to complete a general overview. Some studies involving parallel monitoring of several sterols have been carried out (Grandgirard et al., 2004; MenéndezCarreño,etal.,2012;GonzálezLarenaetal.,2011).However,fewofthemassessa discussedcomparisonofrelativedegradationdata(cercacietal.,2006;menéndezcarreñoet al.,2008a)andthemajorityusedalownumberofsterols. There are some studies concerning modelling with regression equations of cholesterol degradationaswellascopsformation(chienetal.,1998;chienetal.,2006;yenetal.,2010). However, to our knowledge, no study has dealt with plant sterols or SOPs degradation modellingandthiswouldproviderelevantinformationtoestimatetheiractuallevels. Therefore,theaimofthisworkwastostudythebehaviourofthreephytosterols(sitosterol, stigmasterolandcampesterol),aswellastheformationanddegradationpatternofsixring structurepolaroxidationproductsat180 C,comparingthemtothoseofcholesterol.Non linear regression models for the degradation curves of both sterols and oxysterols were designed. 2.Materialandmethods 2.1Reagents Cholesterol, 5cholestane and commercial mixtures of sitosterol, campesterol and stigmasterol were purchased from SigmaAldrich Chemical (Steinhei, Germany). 19 hydroxycholesterol was purchased from Steraloids (Wilton, NH, USA). Trisil reagent was obtained from Pierce (Rockford, IL, USA). Acetone, chloroform, diethyl ether, methanol, hexane and 2propanol were obtained from Panreac (Barcelona, Spain). Hexane for gas chromatographyanddichloromethaneforgaschromatographywerefrommerck(whitehouse Station,NJ,USA).SeppackVac6ccsilica1gcartridgeswereobtainedfromWaters(Milford, USA). 2.2Heatingofsterolsamples Thermooxidationofsterolstandardswasdoneat180 Cforvarioustimedurations:0,5,10, 20,30,60,90,120,180and360min.Forthethermooxidation,0.5mLofcholesterolstandard solution(5mg/ml)wasaddedinto20openglassvials(15x100mm).halfofthesampleswere usedfortheanalysisofsterolsandhalfofthesamplesforthedeterminationofsops.the solventwasevaporatedundergentlenitrogenstream.subsequently,thevialswereplaced open(allowingenoughoxygendisposal)inthetembloc(pselecta,spain)previouslystabilized at180 C.Afterthecorrespondingtime,vialsweretakenoutfromheat.Sameprocedurewas 87

Results applied to the phytosterol standards solution (5 mg/ml; 3.29% campesterol, 0.40% campestanol, 53.81 % stigmasterol, 37.33% sitosterol and 5.18% sitostanol). Then, the samplesweremaintainedatroomtemperaturefor20min,exceptfor5and10minsamples, whichwerecooledinicefor5minbeforeacclimatisation.afterheating,samplespresentedan oilyappearance.theexperimentwasperformedinquadruplicate. 2.3Sterolanalysis Inordertogetasimilarconcentrationthanthatofphytosterolsmixturesamples,cholesterol heatedsampleswereredissolvedwith5mlhexane/2propanol(3:2,v/v)and1ml was transferredintoanewtube.subsequently,0.1mlof5cholestane(2mg/ml)wasaddedto eachcholesterolandphytosterolheatedsample.thesolventwasevaporatedundergentle nitrogenstream. Bothcholesterolandphytosterolheatedsampleswerederivatizedtotrimethylsilyl(TMS) ethersaccordingtoamodifiedversionofthemethoddescribedbyduttaandappelqvist (1997).FourhundredmicrolitresofTriSilreagentwereaddedtoeachsampleandtheywere keptat60 Cfor45mininawaterbath.Thesolventwasevaporatedunderastreamof nitrogenandthetmsetherderivatesweresolvedin10mlofhexaneforgaschromatography. Fourhundredmicrolitresofthissolutionwerefiltratedwithasyringeandafilter(0.45µm) andpouredtoaglassvialpriortogcmsanalysis. GaschromatographyMassspectrometry(GCMS)analysiswasperformedonaHPHewlett Packard6890GCcoupledtoaHP5973MassSelectiveDetector.TheTMSetherderivativesof cholesterolandphytosterolstandardswereseparatedonacapillarycolumnagilent19091s 433HP5ms5%PhenylMethylSiloxane(30mx250mx0.25mfilmthickness)(Agilent,CA, USA).Oventemperatureconditionshadpreviouslybeenoptimisedinordertoachieveproper separationoftheindividualcompounds.theprogrammestartedattemperatureof85 C, heatedto290 Catarateof50 C/minand,finally,increasedto298 Catrateof0.5 C/min. Highpurityheliumwasusedasacarriergasataflowrateof1mL/min.Theinletpressureused was9.64psi.theinjectortemperaturewas280 Candthesampleswereinjected(1µL)ina splitlessmode. Peakidentificationwasbasedoncomparisonoftheirmassspectrawiththespectraofthe Wileylibrary(HPCHEM,Wiley,275,6 th ed.)andalsowiththoseobtainedfromtheliterature.in somecases,acomparisonoftheirretentiontimeandmsfragmentswiththoseofstandard purecompoundswasalsodone.aninternalstandardmethodwasusedforquantification, with5cholestaneastheinternalstandard.cholesteroland5cholestanequantificationwas 88

Results madeinscan,whileplantsterolswerequantifiedusingselectedionmonitoring(sim)analysis onthebasisoftheamountofaspecificionforeachpeak(343,484,357,forcampesterol, stigmasterolandsitosterol,respectively),andtakingintoaccounttherelativeabundanceof eachion(berasategietal.,2012).calibrationcurveswerepreviouslybuilt.fortheintegration AgilentMSDProductivityChemStationforGCandGC/MSSystemsDataAnalysisApplication wereused. 2.4Steroloxidationproductsanalysis Theidentificationandquantificationofsteroloxidationproductswasperformedaccordingto thevalidatedmethodofmenéndezcarreñoetal.(2008b). Firstly,1mLofinternalstandard(20g/mLof19hydroxycholesterol)wasaddedtotheheated samples. SPE was used to separate SOPs from nonpolar and midpolar products. The purificationofoxysterolsallowsobtainingclearchromatograms.thespewasmadeaccording totheproceduresdescribedindetailinguardiolaetal.(1995).thetesttubescontainingthe samples diluted in 5 ml of hexane were applied to a SPE silica cartridge, previously equilibratedwith5mlofhexane.thecartridgewassubsequentlytreatedwith10mlof hexane:diethylether(95:5,v/v),30mlofhexane:diethylether(90:10,v/v),andwith10mlof hexane:diethylether(80:20,v/v).steroloxidationproductswerefinallyelutedfromthespe cartridge with 10 ml of a mixture of acetone/methanol (60:20, v/v). The solvent was evaporatedinrotaryevaporatorunderwarmwaterbath(35 C). Thesamplesolutionsofsteroloxidationproductswerederivatizedtotrimethylsilyl(TMS) ethersaspreviouslydescribedforsterols. GCMSanalysiswasperformedonaHewlettPackard6890NGCcoupledtoa5975Mass SelectiveDetector.TheTMSetherderivativesofsteroloxideswereseparatedonacapillary columnagilentcp8947varianvf5ms5%phenylmethylsiloxane(50mx250mx0.25m filmthickness).oventemperatureconditionswereasfollows:initialtemperatureof75 C, heatedto250 Catarateof30 C/min,increasedto290 Catrateof8 C/min,andfinally,it wasraisedto292 Catarateof0.05 C/min.Highpurityheliumwasusedasacarriergasata flowrateof1ml/min.theinletpressureusedwas9.08psi.theinjectortemperaturewas250 Candthetransferlinetodetector280 C.Thesamples(1µL)wereinjectedinsplitlessmode. Peakidentificationwasmadefollowingthesameprocedureasforsterols.SOPsquantification wasalsobasedonaninternalstandardmethod(19hydroxycholesterol).itwasperformed usingselectedionmonitoring(sim)analysis.foreachstageoftime,differentions(33)were selectively quantified and, consequently, extract ion chromatogram had to be used to 89

Results integratethecorrespondingpeakareas.giventhelackofavailablepopsstandardsandtheir demonstrated similar to COPs response, COPs calibration curves assessed by Menendez Carreñoetal.(2008b)werealsousedtodeterminePOPscontent.FortheintegrationAgilent MSDProductivityChemStationforGCandGC/MSSystemsDataAnalysisApplication(Agilent Technologies,Inc.,CA,U.S.A.)wereused. 2.5Statisticalanalysis Forthestatisticalanalysisofthedata,SPSS15.0programme(SPSS,Inc.,Chicago,IL,U.S.A.) wasused.meanandstandarddeviationofdataobtainedfromeachreplicatewerecalculated. Forthemathematicalmodellingofthedegradationofsterolsandtheiroxides,thenonlinear regressionanalysiswasused(frombeginninginthecaseofsterolsandfromthemomentof themaximumachievedinthecaseofsops).fortheevaluationofthesignificantdifferencesof theamountsofsterolsandsteroloxidesalongtimeandamongdifferentsterols,onefactor ANOVAwithTuckeyposthocmultiplecomparisons(p<0.05)wasapplied.Finally,correlations betweenoxysterolsofthesameoxidationpattern(concerningfunctionalgroupandposition) butfromdifferentsteroloriginwereassessedbymeansofpearson scorrelationtest. 3.Results 3.1Sterolstudy Amountsofremainingcholesterolandphytosterolsafterthedifferentheatingtimes(0360 min)areshownintable1.significantdifferenceswereobservedamongeveryheatingtime duringthefirst2030min,whenadrasticdropwasdetected.thereafter,smalldifferences werefound.degradationreachedaround5560%oftheinitialsterolcontent exceptfor cholesterol,41.80%duringthefirst5minoftreatment(table2).after30minheating,around 88%and74.71%oftheinitialsterolshadalreadybeendegraded,forphytosterols(meanvalue) andcholesterol,respectively,andaround90%and79.64%respectivelyafter90min.atthe endoftheheattreatmentallsterolsweredegradeduptoaround95%oftheirinitiallevel. Differentnonlinearregressionmodelswereassayedtopredictthelossofsterols(logarithmic, inverseandexponential),withlogarithmicmodelshowingthehighestr 2 forallcases(0.907, 0.972, 0.953 and 0.979 for cholesterol, campesterol, stigmasterol and sitosterol, respectively). The plots and corresponding equations are shown in Figure 1. Cholesterol showedthehighestfirstconstantnamedas(48.720)comparedtotheotherequations (39.78641.474). 90

Results 3.2SterolOxidationProducts Theevolutionof24differentSOPswasfollowedinthisstudy.7hydroxy,7hydroxy,5,6 epoxy,5,6epoxy,trioland7ketoderivativesofeachsterolwereanalysed(tables36).the totalamountofcompoundsderivedfromeachsterolwasalsocalculated(figure2).maximum levelsforeachsteroltotaloxideswere73.79,106.53,49.75and98.38µg/mg,forcholesterol (10min),campesterol(10min),stigmasterol(5min)and sitosterol(5min),respectively. Afterthistime,therewasasignificantandprogressivedecreasefortotaloxysterols,reaching minimalamounrs,around6%ofthemaximumconcentrationattheendoftheprocess.higher levelsoftotalcampesterolandsitosteroloxidesweredetectedcomparedtothoseachieved bycholesterolandstigmasterolderivativesthroughoutthewholeheatingprocess. ThemostabundantSOPformedwas,byfar,7ketosterol(irrespectiveofthetypeofsterol), reaching64.28%,65.93%,53.85%and67.68%ofthetotalamountofoxidesderivedfrom cholesterol,campesterol,stigmasteroland sitosterol,respectively,atthemomentofthe maximumtotalsopsconcentration,whichwassetat510mintreatment.atrendtoincrease inall7ketoplantsterolderivativeswasobservedatminute60,althoughnotstatistically significant. 7Hydroxy and 7hydroxy were the next derivatives formed from the quantitative point of view. While 7hydroxy derivatives were more abundant than hydroxy ones at the initial points of analysis, as the treatment progressed, hydroxy formation is favoured over alpha s. Sterol epoxides level is much smaller than 7 hydroxysterols atthebeginning(especiallyinthecaseofcampesterolandstigmasterol),but higher at the end, ie, 8.09 and 21.11 µg/mg of 5,6/epoxycampesterol and 7/ hydroxycampesterolrespectivelyat5min,comparedto5.89and0.52µg/mgat90min.triol derivativesweretheonesformedinaloweramount,evenfindingnocampestanetriol. Asasimilarbehaviourcouldbeobservedamongthesamekindofoxysterolsofdifferentsterol origin(table1andfigure2),correlationwasalsostudiedamongthem,showingpearson s Coefficientsbetween0.90and0.99inmostcases. Amongthedifferentnonlinearregressionmodelsassayed(Table7),andconsideringtheR 2 values,theoneswhichbestfitthedegradationwerefor7and7hydroxy,aninversemodel; for5,6epoxy,alogarithmicmodel;andfortherestofcompoundsanexponentialmodel exceptfor5,6epoxycholesterol,whichfitalogarithmicmodel.intheoverallview,totalchol ox,totalcamox,totalstigmaoxandtotalsitooxfollowedanexponentialtypedegradation. Concerningthedegradationrateconstants,7hydroxycompoundsshowedadegradation rate constant around twice that of 7hydroxy, except for cholesterol samples. Among 91

Results exponentialregressionadjustments,rateconstantswereinallcasesrangingbetween0.010 and 0.004. Values were rather uniform within the same type of oxidation mechanism, regardlessofthesteroloriginexceptfor5,6epoxycholesterol,theconstantsoftriolbeing thelowest(inabsolutevalue)andthatof5,6epoxythehighest. 4.Discussion Comparedtootherstudiesperformedbothinmodelorrealsystems(Xuetal.,2009;Yenetal., 2010;MenéndezCarreñoetal.,2010),theheattreatmentconditionsappliedinourstudy seemedtobemoredestructive,sincethedegradationpercentagesofsterolswere,ingeneral, higher. Around 55% degradation of cholesterol and sitosterol standards and 38% stigmasterolstandardhavebeenpreviouslyreportedafter1hheatingat180 C(Xuetal., 2009;MenéndezCarreñoetal.,2010),comparedtomorethan70%notedforthepresent study.differentheatingtemperature/exposuretimeandotherexperimentalconditions(such asinitialsterolamount)couldbebehindthisdiversityofresults.intheirwork,menéndez Carreñoetal.(2010)heatedthestigmasterolstandardsamplesintoglassvialsplacingthem intoanovenwheretheenergytransfertakesplaceviaconvectionheating.incontrast,sterol standardswereheatedinthepresentstudybyusinganelectronicheatingdevice.thus,the transferofenergyhereoccursbetweentwoobjectsthatareinphysicalcontact(conduction heating).then,asithasbeensuggestedinseveralpreviousworks,thesteroldegradation patterncandiffersignificantlyaftertheapplicationofdifferentheatingtechnologies(chienet al.,1998). Thecomparisonamongthedifferentdegradationpercentagesofsterolsshowedagreater susceptibility to oxidation for phytosterols than for cholesterol, during first 120 min. Moreover, nonlinear regression equation parameters for remaining sterol content also suggested a rather lower degradation intensity of cholesterol compared to plant sterols, attendingtothevalueofthefirstconstant,higherforcholesterolthanforplantsterolsas different initial amounts of sterols were used, interferences during oxidation could have occurredsincethermooxidationissignificantlydependentonthesampleandtovolumeratio (Lampietal.,2002).Inthissense,cholesterolparticleswouldbelessexposedtooxidation becauseofthehighestinitialamountsused(2.5mg).inthiscase,somecontroversialdatacan beobservedindifferentworks:menéndezetal.(2008a)andcercacietal.(2006)found cholesteroltobemorepronetooxidativedegradationthan sitosterol,whereasxuetal. (2009)reportednodifferencesindegradationratesamongsterols.Onbasisonakineticstudy, Yen et al. (2010) and Chien et al. (1998) proposed first order equations for cholesterol degradation.lowerheatingtemperature(150 C insteadof180 C)resultedinaslower 92

Results degradationprocess,whichcouldhaveledtotheadjustmentofafirstorderequationwith theseparametersinsteadofalogarithmicone. RegardingtheaccumulationofSOPs,maximumlevelsachievedwereinaccordancetothe valuesfoundintheliteratureforcholesteroland sitosteroloxides,butslightlylowerfor campesterolandstigmasteroloxides(xuetal.,2011;yenetal,2010;kemmoetal.,2008; Lampietal.,2009).ItisimportanttonotethatthemaximumSOPslevelswerereachedat510 min,whereasotherstudiesfoundmaximumlevelsat60minorlongertimesatthesame temperature(180 C)(Xuetal.,2011;Menéndezetal.,2010).Thesedatapointedoutthatour degradationstartedearlierandhigherconcentrationscouldnothavebeenachieved.the subsequentdrasticdropwasinaccordancetowhatxu,zhang,prinyawiwatkulandgodber (2005)found,whenheatingcholesterolat175 C.Nevertheless,comparedtootherstudies (Kemmoetal.,2005;Xuetal.,2011;Chienetal.,2006;Yenetal.,2010;MenéndezCarreñoet al.,2010),ourdegradationoccursmuchearlier,probablyduetothedifferentexperimental conditionsappliedmentionedabove.changingprocessingtemperatureshaveanimportant effect on the formation of oxysterols, but other experimental conditions are likely to contributetodifferentialresults,iedifferentinitialsampleamounts,purificationproceduresor chromatographictechniquesapplied.polymericproducts,steradiensandbothnonpolarand midpolarcompoundshavebeenreportedaspossibledegradationproductsformedduring extremeheatingconditions(menéndezcarreñoetal.,2010;rudzinskaetal.,2009;lampiet al.,2009;lerckerandestrada,2002). 7Hydroxyand7ketooxidesareusuallyexpectedtobethemajorSOPs(Grandgirardetal., 2004;GonzálezLarenaetal.,2011),atleastduringthefirststagesofoxidation.Dominancyof epimeramong7hydroxycompoundsispossiblyduetostearichindranceofthehydroxyl groupatposition3(kemmoetal.,2005;smith,1987);otherauthorshavealsoobservedthis trend(lampietal.,2002;xuetal.,2005;soupasetal.,2007).thetrendtoincreaseinall7 keto plant sterol oxides observed at minute 60 (Tables 36), might be attributed to the conversionfrom7hydroxyderivatives(rudzinskaetal.,2009;kemmoetal.,2008).negligible levelsoftriolcompoundshavepreviouslybeenreported(conchilloetal.,2005;lampietal., 2002),onlybyheatingtreatments. Regardingthehigherlevelsofcampesteroland sitosterolderivativescomparedtothatof cholesterol and stigmasterol ones, this behaviour could be attributed either to a faster oxidationofthecampesterolandsitosteroltogiverisetothecorrespondingoxideseitherto aslowerdegradationoftheseoxides.thelowerinitialamountsofcampesterol(0.081mg)and sitosterol(0.916mg)couldbebehindthisbehaviour(lampietal.,2002),sinceitwouldhave 93

Results beenoverexposedtooxygen.inaddition,thepresenceofanextradoubleboundintheside chainofstigmasterolatposition22maypossiblyhaveaffectedtotherateofoxidationofthis compound. Thus, stigmasterol presented the lowest total maximum amount of oxidation productsespeciallyaffecting7ketostigmasterolformationincomparisontotheothersterols. Nevertheless,furtherresearchwouldbenecessarytoconfirmthishypothesis. Consideringcomparativegraphsandkineticcurves,itcanbestatedthatoxysterolsfolloweda tendencyaccordingtothetypeofoxidation,andregardlessofthesterolorigin.however,a particularcasecanbehighlighted:5,6epoxycholesterolpresenteditsbestadjustmentfora logarithmicmodel,whereasplantsterolsanalogousoxideswereadjustedtoanexponential one.differentsidechaincouldbebehindthisbehaviour,althoughfurtherstudieswouldbe requiredtomakemoreaccurateconclusions.drasticdropof7hydroxyderivativeslevels (Tables 36) could be related to their adjustment to an inverse model, instead of an exponentialone,asintheothercases. In conclusion, our results revealed a sterol susceptibility to degradation following this decreasingorder:campesterol sitosterol stigmasterol>cholesterol.regardingsterol oxidation products, the levels increased during the first 510 min and thereafter, their degradationratewashigherthantheirformationrate.sopsdegradationseemedtodepend onthemolecularstructureoftheoxidisedcompound,irrespectiveofthesterolfromwhich theywerederived. 5.Acknowledgements We thank the Programa ConsoliderIngenio 2010 CARNISENUSA CSD 200700016, the ProyectoAGL200801099/ALI (MinisteriodeCienciaeInnovación),and PlanInvestigador delauniversidaddenavarra (PIUNA)fortheircontributiontothefinancialsupportofthis work.b.barriusoisgratefulto CátedraTomásPascualSanzUniversidaddeNavarra andto AsociacióndeAmigosdelaUniversidaddeNavarra forthegrantsreceived. 94

Results Table1.Degradationofsterolstandardsduring360minutethermooxidationat180 C.Meanandstandard deviation(n=4)arerepresentedwiththestatisticalanalysis.differentletterswithineachcolumndenotesignificant differencesamongheatingtimes(p<0.05). Time(min) Remainingsterols(mg) cholesterol campesterol stigmasterol sitosterol 0 2.41 ±0.05 g 0.08 ±0.01 e 1.32 ±0.10 d 0.92 ±0.07 f 5 1.40 ±0.16 f 0.03 ±0.01 d 0.58 ±0.07 c 0.35 ±0.03 e 10 1.07 ±0.21 e 0.02 ±0.003 c 0.48 ±0.06 c 0.23 ±0.03 d 20 0.88 ±0.02 d 0.01 ±0.001 b 0.25 ±0.02 b 0.15 ±0.02 c 30 0.61 ±0.01 c 0.01 ±0.001 ab 0.16 ±0.02 ab 0.10 ±0.01 bc 60 0.55 ±0.02 bc 0.01 ±0.001 ab 0.13 ±0.003 a 0.10 ±0.01 bc 90 0.49 ±0.02 bc 0.01 ±0.001 ab 0.12 ±0.003 a 0.09 ±0.01 bc 120 0.38 ±0.01 b 0.01 ±0.001 ab 0.10 ±0.004 a 0.08 ±0.00 ab 180 0.16 ±0.01 a 0.01 ±0.001 a 0.09 ±0.002 a 0.07 ±0.001 ab 360 0.13 ±0.01 a 0.01 ±0.001 a 0.05 ±0.002 a 0.04 ±0.001 a Table2.Degradationpercentagesofcholesterol,campesterol,stigmasterolandsitosterolduring360minute thermooxidationat180 C.Withineachrow,differentlettersdenotesignificantdifferences(p<0.05)amongsterols. Time(min) Degradationpercentages(%) cholesterol campesterol stigmasterol sitosterol 5 41.80 ±7.23 a 58.06 ±5.83 b 56.12 ±5.84 ab 61.45 ±3.41 b 10 55.74 ±9.75 a 71.56 ±389 b 63.47 ±4.71 ab 74.68 ±3.65 b 20 63.58 ±0.75 a 82.68 ±1.83 b 80.87 ±1.78 b 83.89 ±1.73 b 30 74.71 ±0.66 a 87.06 ±1.31 b 88.30 ±1.38 b 88.72 ±1.14 b 60 76.99 ±0.75 a 87.16 ±1.22 b 90.40 ±0.26 c 89.31 ±0.79 c 90 79.64 ±0.76 a 88.14 ±0.40 b 90.87 ±0.26 d 89.73 ±0.53 c 120 84.37 ±0.35 a 89.60 ±0.27 b 92.20 ±0.27 d 91.42 ±0.28 c 180 93.25 ±0.53 c 90.89 ±0.35 a 93.22 ±0.17 c 92.40 ±0.15 b 360 94.51 ±0.54 a 94.34 ±0.20 a 96.39 ±0.14 b 95.79 ±0.14 b Table3.Concentrationofcholesteroloxidationproducts(µg/mg).Withineachcolumn,differentlettersdenote significantdifferencesamongheatingtimes(p<0.05). Time(min) cholesteroloxides(g/mgcholesterol) 7hydroxy 7hydroxy 5,6epoxy 5,6epoxy triol 7keto 0 nd nd 0.08±0.01 a 0.25±0.01 a nd nd 5 5.26±0.35 c 3.60±0.36 c 3.85±0.30 c 3.62±0.36 c 0.26±0.05 de 35.83±5.57 ef 10 7.05±0.83 d 6.83±0.74 d 5.42±0.77 d 6.79±1.12 e 0.26±0.05 de 47.43±6.80 g 20 2.76±0,39 b 5.64±1.01 e 3.21±0.37 c 5.14±0.20 d 0.28±0.04 de 41.86±1.77 fg 30 0.78±0.15 a 1.49±0.11 b 1.46±0.25 b 3.88±0.45 c 0.25±0.02 de 34.40±2.65 ef 60 0.55±0.02 a 0.93±0.01 ab 0.72±0.08 ab 3.03±0.23 b 0.31±0.04 e 31.18±4.63 e 90 nd nd 0.09±0.02 a 0.68±0.13 a 0.24±0.01 de 21.22±2.80 d 120 nd nd 0.04±0.01 a 0.40±0.08 a 0.20±0.03 cd 17.07±2.55 cd 180 nd nd 0.02±0.001 a 0.23±0.02 a 0.12±0.004 bc 10.76±0.64 bc 360 nd nd 0.01± 0.001 a 0.12±0.01 a 0.05±0.01 ab 3.32±0.41 ab 95

Results Table4.Concentrationofcampesteroloxidationproducts(µg/mg).Withineachcolumn,differentlettersdenote significantdifferencesamongheatingtimes(p<0.05). Time(min) campesteroloxides(g/mgcampesterol) 7hydroxy 7hydroxy 5,6epoxy 5,6epoxy triol 7keto 0 nd nd 0.38±003 a 0.001±0.000 a nd 0.09±0.03 a 5 14.06±2.32 c 7.05±0.25 d 0.79±0.11 ab 7.30±0.28 ef nd 70.23±7.12 e 10 16.85±1.59 c 7.19±1.24 d 1.49±0.20 c 10.13±0.94 g nd 70.87±8.74 e 20 8.30±0.71 b 4.66±0.95 c 1.47±0.03 bc 7.61±0.80 f nd 52.01±6.41 d 30 1.89±0.31 a 2.52±0.27 b 124±0.17 bc 6.23±0.66 de nd 43.25±0.27 cd 60 0.89±0.14 a 1.11±0.03 ab 1.54±0.26 c 4.91±0.57 cd nd 52.96±2.60 d 90 0.20±0.01 a 0.33±0.06 a 1.62±0.30 c 4.27±0.61 c nd 38.71±2.32 c 120 nd 0.12±0.02 a 1.47±0.30 bc 2.65±0.37 b nd 33.19±4.17 bc 180 nd 0.11±0.02 a 1.33±0.17 bc 2.09±0.26 b nd 24.18±4.00 b 360 nd 0.06±0.005 a 0.44±0.05 a 0.59±0.10 a nd 7.90±0.50 a Table5.Concentrationofstigmasteroloxidationproducts(µg/mg).Withineachcolumn,differentlettersdenote significantdifferencesamongheatingtimes(p<0.05). Time(min) stigmasteroloxides(g/mgstigmasterol) 7hydroxy 7hydroxy 5,6epoxy 5,6epoxy triol 7keto 0 0.08±0.01 a 0.03±0.004 a 0.01±0.001 a nd nd 0.03±0.01 a 5 9.96±1.46 c 5.01±0.25 d 4.77±0.08 d 2.79±0.55 de 0.44±0.08 bc 26.79±4.33 f 10 11.15±1.92 c 5.02±0.76 d 4.69±0.85 d 3.44±0.40 e 0.87±0.13 f 23.61±2.74 f 20 6.20±0.75 b 3.37±0.63 c 3.94±0.67 d 2.44±0.29 cd 0.81±0.07 ef 17.09±2.59 e 30 1.33±0.23 a 1.74±0.11 b 2.13±0.36 c 1.88±0.19 bc 0.70±0.03 def 13.83±0.59 de 60 0.51±0.06 a 0.68±0.03 ab 1.44±0.11 bc 1.53±0.14 b 0.65±0.11 de 16.16±2.42 e 90 0.15±0.02 a 0.25±0.05 a 1.05±0.10 abc 1.38±0.26 b 0.61±0.11 cd 13.04±1.22 de 120 0.06±0.01 a 0.09±0.02 a 0.59±0.07 ab 0.57±0.07 a 0.46±0.02 bc 8.85±0.15 cd 180 0.06±0.01 a 0.07±0.008 a 0.45±0.09 ab 0.45±0.09 a 0.31±0.05 b 6.86±0.75 bc 360 0.01±0.00 a 0.02±0.002 a 0.07±0.01 a 0.10±0.02 a 0.05±0.01 a 2.51±0.11 ab Table6.Concentrationof sitosteroloxidationproducts(µg/mg).withineachcolumn,differentlettersdenote significantdifferencesamongheatingtimes(p<0.05). Time(min) sitosteroloxides(g/mgsitosterol) 7hydroxy 7hydroxy 5,6epoxy 5,6epoxy triol 7keto 0 0.06±0.01 a 0.02±0.00 a nd 0.001±0.003 a nd 0.04±0.01 a 5 12.27±1.69 c 6.86±0.25 d 5.80±0.56 d 6.05±0.58 e 0.82±0.13 e 66.58±12.73 f 10 13.94±2.44 c 7.04±1.10 d 5.58±0.67 d 8.36±0.64 d 0.56±0.06 d 59.17±11.12 ef 20 6.02±1.14 b 4.74±0.86 c 4.17±0.83 c 5.88±0.68 d 0.42±0.06 c 42.53±6.59 de 30 1.31±0.09 a 2.37±0.17 b 1.87±0.20 b 4.55±0.40 c 0.38±0.03 bc 34.95±0.48 cd 60 0.60±0.08 a 1.01±0.03 ab 0.67±0.07 ab 3.79±0.29 c 0.43±0.03 c 41.76±6.28 d 90 0.18±0.08 a 0.52±0.18 a 0.30±0.03 a 3.31±0.61 c 0.43±0.05 c 33.39±3.23 cd 120 0.07±0.008 a 0.17±0.02 a 0.08±0.004 a 1.39±0.17 b 0.34±0.02 bc 23.35±0.49 c 180 0.08±0.004 a 0.14±0.02 a 0.06±0.01 a 0.87±0.05 ab 0.28±0.03 b 18.94±2.93 bc 360 0.01±0.001 a 0.02±0.00 a 0.01±0.001 a 0.23±0.03 a 0.11±0.02 a 6.35±0.25 ab 96

Results Table7.RegressionmodelandparametersofthekineticequationsofallSOPs. Compound Regressionmodel R 2 a k 7hydroxychol inverse a 0.956 0.672 73.678 7hydroxycam inverse 0.965 1.439 179.080 7hydroxystig inverse 0.947 0.993 121.065 7hydroxysito inverse 0.946 1.363 149.086 7hydroxychol inverse 0.868 0.410 81.721 7hydroxycam inverse 0.937 0.204 77.478 7hydroxystig inverse 0.945 0.163 54.600 7hydroxysito inverse 0.941 0.170 75.919 5,6epoxychol logarithmic b 0.772 7.275 1.419 5,6epoxycam 5,6epoxystig logarithmic 0.912 7.146 1.296 5,6epoxysito logarithmic 0.890 8.486 1.635 5,6epoxychol logarithmic 0.890 10.824 1.997 5,6epoxycam exponential c 0.960 8.296 0.008 5,6epoxystig exponential 0.949 2.790 0.010 5,6epoxysito exponential 0.954 6.856 0.010 triolchol exponential 0.918 0.350 0.005 triolcam exponential triolstig exponential 0.947 1.043 0.008 triolsito exponential 0.846 0.551 0.004 7ketochol exponential 0.976 45.397 0.007 7ketocam exponential 0.953 64.773 0.006 7ketositg exponential 0.935 21.658 0.006 7ketosito exponential 0.912 6.493 0.007 totalchol exponential 0.936 58.094 0.008 totalcam exponential 0.959 135.203 0.007 totalstig exponential 0.926 39.901 0.008 totalsito exponential 0.928 81.567 0.007 a y=a+k/t b y=a+k.log(t) c y=a.e kt 97

Results a)cholesterol b)campesterol 100 80 y = 48.720 6.219 log t R 2 = 0.907 100 80 y = 41.267 6.452 log t R 2 = 0.972 60 60 40 40 %Remainingsterol 20 0 0 100 c)stigmasterol 100 80 200 300 400 y = 41.474 6.645 log t R 2 = 0.953 20 0 0 100 d)sitosterol 100 80 200 300 y = 39.786 6.556 log t R 2 = 0.979 400 60 60 40 40 20 20 0 0 0 100 200 300 400 0 100 200 300 400 Time(min) Figure1.Mathematicalmodellingofthedegradationkineticofsterolstandardsduring360minutethermo oxidationat180 C.Remainingpercentagesofa)Cholesterolb)Campesterolc)Stigmasterold)sitosterol µg/mgsterol 120 100 80 60 40 cholesterol campesterol stigmasterol sitosterol 20 0 0 100 200 300 400 min Figure2.GraphicrepresentationsoftotalSOPsduring thermooxidationupto360minfordifferentsterolorigin 98

ResultsV Paper4 RoleofMelissaofficinalisincholesteroloxidation: Antioxidanteffectinmodelsystemsandapplicationin beefpatties

Results FoodResearchInternational(2015)69,133140 RoleofMelissaofficinalisincholesteroloxidation:antioxidanteffectin modelsystemsandapplicationinbeefpatties BlancaBarriuso a,dianaansorena a,mariaisabelcalvo b, RitaYolandaCavero c,iciarastiasarán b a DepartmentofNutrition,FoodScienceandPhysiology,FacultyofPharmacy,UniversityofNavarra b DepartmentofPharmacyandPharmaceuticalTechnology,FacultyofPharmacy,UniversityofNavarra c DepartmentofPlantBiology,FacultyofScience,UniversityofNavarra Abstract Cholesteroloxidationproducts(COPs)constituteaknownhealthriskfactor.Theantioxidant effectofalyophilizedaqueousmelissaofficinalisextractagainstcholesteroldegradationand COPsformationduringaheatingtreatmentwasevaluatedinamodelsystem(180 C,0180 min)ataratioof2mgextract/100mgcholesterol.furthermore,theplantextractwas subsequentlyaddedtobeefpattiesaloneorincorporatedwithinanoilinwateroliveoil emulsiontoassessitseffectivenessduringcooking.melisaextractprotectedcholesterolfrom thermaldegradationinthemodelsystem,yieldinghigherremainingcholesterolandlower COPsvaluesthroughoutthewholeheatingprocess.MaximumtotalCOPswereachievedafter 30and120minofheatingforcontrolandmelisacontainingsamples,respectively.Incooked beefpatties,eventhoughtheoliveoilemulsionwasusedasflavormaskingapproach,melisa extractoffflavorlimitedthemaximumdosewhichcouldbeadded.atthesedoses(65µg/g and150µg/gwithoutandwiththeemulsion,respectively),noadditionalprotectiveeffectof melisaovertheuseoftheemulsionwasfound.additionofnaturalextractsintofunctional foodsshoulddefinitivelytakeintoaccountsensoryaspects. Keywords:cholesteroloxidation,antioxidant,modelsystem,lemonbalm, oddflavor,beef patties Highlights 1AlyophilizedMelissaofficinalisextractprotectedcholesterolfromthermaldegradation. 2MelisaextractinhibitedCOPsformationat180 Cupto180min. 3Melisaextractoffflavorrestrictedtheviabledosetobeaddedintobeefpatties. 4Noprotectiveeffectwasfoundinmeatpattiesatsensoryacceptabledoses. 5Anoliveoilinwateremulsionexertedantioxidanteffectinbeefpatties. 101

Results 1.Introduction CholesterolOxidationProducts(COPs)havebeenrelatedtoseveraldiseases(atherosclerosis, neurodegenerativediseases,mutagenicandcarcinogeniceffects,etc)(otaeguiarrazolaetal., 2010;Biasietal.,2013).Theycanbeformedendogenouslyandalsoabsorbedfromthediet.As cholesterolispresentinavarietyofanimalfoodsamples,thermaloxidation,photooxidation andautooxidationcantakeplace,compromisingthesafetyoffood.thus,minimizingthe formationofcopsleadstosaferfood. Theincorporationofantioxidantshasbeenproposedasagoodstrategyforpreventingsterol oxidation.promisingresultshavebeenreportedinmodelsystems(xuetal.,2009;yenetal., 2010;Kmieciketal.,2011),usingdirectapplicationofantioxidantsoncholesterolandalsoon fatmatrices,suchastriglyceridesandlard.butylhydroxytoluene,conjugatedlinoleicacid, tocopherol,quercetin,greenteacatechinsandrosemaryextracts,amongothers,aresomeof thetestedantioxidants.ahigherinterestonnaturalantioxidantsthanonsyntheticonesis nowadaysincreasingamongindustriesandresearchers,giventheirsimilarorevenhigher activity(xuetal.,2009;kmieciketal.,2011)andtheirassumedsaferandhealthierproperties. Melissaofficinalisisamedicinalplant,usuallytakenasinfusion,withavarietyofbeneficial effects, i.e. antidepressive, anxiolytic, antitumoral, neurobiological and it has also been involvedintheregulationoflipidemicdisordersandinthepreventionofoxidativedamage (Encaladaetal.,2011;Fazlietal.,2012;Junetal.,2012;Taiwoetal.,2012;Lópezetal.,2009). Itshighantioxidantcapacity,duetothepresenceofphenoliccompounds,mainlyrosmarinic acid(barrosetal.,2013),hasinduceditsaddition,mainlyasextracts,infoodstopreventlipid oxidationforbothfunctionalandtechnologicalpurposes(fazlietal.,2012,petrovicatel., 2012,Berasategietal.,2011;GarcíaIñiguezdeCirianoetal.,2010a;Poyatoetal.,2013). However,toourknowledge,thepotentialinhibitoryeffectofthisplantagainstcholesterol oxidationandformationofcholesteroloxidationproductshasnotbeenevaluatedyet. Whenselectinganaturalantioxidantandtheconcentrationtobeaddedtofoodstuffs,sensory impactontheproduct(suchasflavororcolor)shouldbeconsideredtoachievedesiredtraits (Karreetal.,2013).Theseattributesaredeterminantsofwhetheraconsumerwillpurchasea specifictypeofmeatornot(goodsonetal.,2002).whenusingmelisa,sensoryaspectshave beenevaluatedondifferentmeatderivatives,givingrisetoproductsinwhichnosensory problemswerenoticedwhenusingupto686µg/ginthecaseofdryfermentedsausages (GarcíaIñiguezdeCirianoetal.,2010a)andupto965µg/ginthecaseofcookedpork sausages(berasategietal.,2011).itisworthnotingthatthesearemeatderivativeswithahigh 102

Results content of sensory potent spices, that might mask its contribution to potential negative effects. Consideringallthis,theaimofthisstudywastoevaluatetheantioxidantprotectiveeffectofa lyophilized aqueous M. officinalis extract against cholesterol degradation and cholesterol oxidationproductsformationinamodelsystem.oncetheeffectivenessofmelisainthemodel systemwasprobed,theapplicationofthisextracttoafoodsystem(beefpatties)wascarried outinordertoassessitseffectivenessasantioxidantatdosesthatweresensoryacceptable. 2.Materialandmethods 2.1Reagents Cholesterol,5cholestane,thiobarbituricacid,trolox,AAPHandfluoresceinsodiumsaltwere purchased from SigmaAldrich Chemical (Steinhei, Germany). 19hydroxycholesterol was obtained from Steraloids (Wilton, NH, USA). Trisil reagent was obtained from Pierce (Rockford, IL, USA). Acetone, chloroform, ethyl acetate, methanol, hexane, 2propanol, hydrochloric acid, cyclohexanone, trichloroacetic acid, potassium chloride, potassium hydroxide,anhydroussodiumsulfateandsodiumphosphatewereobtainedfrompanreac (Barcelona, Spain). Hexane for gas chromatography and dichloromethane for gas chromatographywerefrommerck(whitehousestation,nj,usa).stratanh 2 (55µm,70A) 500mg/3mLSolidPhaseExtractioncartridgeswereobtainedfromPhenomenex(Torrance, USA).M.officinalisdriedleaveswerepurchasedfromPlantaronS.L.(Barcelona,Spain).Beef meat was purchased in a minor local distributor, and showed Ternera de Navarra PGI (ES/PGI/0005/0130). 2.2PreparationandcharacterizationofM.officinalisextract AqueousextractofM.officinaliswaspreparedasdescribedinGarcíaÍñiguezdeCirianoetal. (2010b).Briefly,50gofdriedleaveswereweightedandaddedto500mLofdistilledwater, preheatedat100 C.Themixturewassubjectedtorefluxfor30minatthetemperatureabove. Extractionprocesswasrepeatedwith500mLofdistilledwaterandbothextractswerejoined, filteredandlyophilized.determinationofitsrosmarinicacidcontentwasperformedbyhplc UVasdescribedinGarcíaÍñiguezdeCirianoetal.(2010b).Resultswereexpressedasmg rosmarinicacid/glyophilizedmelisaextract.totalphenoliccontent(tpc)wasdeterminedas describedinpoyatoetal.(2013).a12mgextractsamplewassolvedin10mlwater.reagents weremixed:237µldistilledwater,3µlsamplesolution,15µloffolinciocalteu sreagent, and45µlof20%sodiumcarbonateanhydroussolution.after2hinthedark,theabsorbance was measured at 765 nm in a FLUOStar Omega spectrofluorometric analyzer (BMG 103

Results Labtechnologies,Offenburg,Germany).TPCwasexpressedasµggallicacid/mgsample (extractoroil). 2.3Modelsystem 2.3.1Heatingofsamples Thermooxidationofcholesterolwasdoneat180 Cforvarioustimedurations:0,10,20,30, 60,120and180min.Forthethermooxidation,4mLofcholesterolstandardsolution(5 mg/mlhexane)wasaddedintoopenglassvials(15x100mm).thesolventwasevaporated undergentlenitrogenstream.subsequently,thevialswereplacedopen(allowingenough oxygendisposal)inthetembloc(pselecta,spain)previouslystabilizedat180 C.Afterthe correspondingtimes,vialswereremovedfromthetemblocandplacedinicefor10min.the residuewassolvedin4mlhexaneandstoredat20 Cuntilanalysis.Thesameprocedurewas applied to the M. officinaliscontaining samples with the following differences: 2 ml of standard solution (10 mg/ml cholesterol and 0.2 mg/ml melisa extract in a trichloromethane:methanol (2:1) mixture) was aliquoted, dried and heated as previously described.theexperimentwasperformedinquadruplicate,withheatingtreatmentsdonein fourdifferentdays. 2.3.2Cholesteroldetermination 50µLwasaliquotedfromeachsample(cholesterolorcholesterol+melisa)and100µLofthe internal standard, 5cholestane (2 mg/ml, hexane:2propanol, 3:2), was added. Chromatographic analysis, identification and quantification were performed according to Conchilloetal.(2005). 2.3.3CholesterolOxidationProductsdetermination Firstly,250µLwasaliquotedfromeachsampleand1mLof19hydroxycholesterol(20g/mL, hexane:2propanol,3:2)asinternalstandardwasaddedtotheeachaliquot.nh 2 SPEwasused toseparatecopsfromnonpolarandmidpolarproducts,assuggestedbyrosesallinetal. (1995).COPswerefinallyelutedinacetone,whichwasfurtherevaporatedunderastreamof nitrogen (35 C). Samples were then derivatized to trimethylsilyl (TMS) ethers. Chromatographicanalysis,identificationandquantificationwereperformedaccordingtothe validatedmethodofmenéndezcarreñoetal.(2008b).sevendifferentcopsweredetermined: 7hydroxycholesterol (7HC), 7hydroxycholesterol (7HC), 5,6cholesterol epoxide (5,6CE), 5,6cholesterol epoxide (5,6CE), 3,5,6cholestanetriol (CT), 25 hydroxycholesterol(25hc),and7ketocholesterol(7kc). 104

Results 2.3.4Antioxidantcapacityalongtheheatingprocess AntioxidantcapacitywasassessedbymeansoftheORACmethod,accordingtotheprocedure described in Ou et al., 2001, with slight modifications. Cholesterol and melisa extract containing sample was aliquoted (50 µl) and evaporated under a stream of nitrogen. Phosphate buffer (1 ml) and chloroform (300 µl) were added. Then, the samples were vortexedfor20sandcentrifugedat4000rpmfor10min.atotalof0.5mloftheaqueous layerwastakenandkeptinthedarkuntilanalysis.a0.5mstocksolutionoftroloxwas preparedin10mmphosphatebuffer,anddividedinto1mlaliquots,whichwerestoredat20 C until use. A new set of stock Trolox vials was taken from the freezer daily for the preparationofthecalibrationcurveandthequalitycontrols(12.5and50µm).thephosphate buffersolutionwasusedasblank,todissolvethetroloxqualitycontrolsandtopreparethe samples.toconducttheoracassay,analiquotofthesample(40µl)and120µlofthe fluoresceinsolution(132.5nm)wereaddedtothe96wellblackplate.themicroplatewas equilibrated(5min,37 C),andthenthereactionwasinitiatedbytheadditionofAAPH(40µL, 300mM);readingswereobtainedimmediately,inaFLUOStarOmegaspectrofluorometric analyzer (BMG Labtechnologies, Offenburg, Germany). The results were expressed as mg troloxequivalent/gsample. 2.3.5Rosmarinicacidcontentalongtheheatingprocess Cholesterol and melisa containing sample was aliquoted (1 ml) and evaporated under a streamofnitrogen.ultrapurewater(1ml)andhexane(1ml)wereadded.thesamplewas vortexedfor20sandcentrifugedat1300gfor6min.theupperlayerwasdiscardedandthe processwasrepeatedtwomoretimes.theaqueouslayerwasfilteredthrougha0.20µl membrane filter (Millipore, USA) and analyzed using the chromatographic conditions describedingarcíaiñiguezdecirianoetal.(2010b).briefly,inac18column,andataflowrate of0.8ml/min,agradientofacidifiedwater:acetonitrilewasapplied(startingat90:10; changingto70:30for20min;andreturningto90:10in7min).theprofileswererecordedat 280nm.Theresultswereexpressedasmgrosmarinicacid/gsample. 2.4Foodsystem The experimental designapplied to thispartof theworkis presentedinfigure1.four different meat patty formulations were assessed, namely simple (S), simple+melisa (SM), emulsioncontaining(e)andemulsioncontaining+melisa(em). 105

Results TYPES OF PATTIES SIMPLE S SIMPLE + MELISA SM EMULSION E EMULSION + MELISA EM sensory analysis chemical analysis S SM (65 µg/g) E EM (150 µg/g) RAW COOKED RAW COOKED RAW COOKED RAW COOKED Figure1.Experimentaldesignforbeefpattiesstudy. 2.4.1Totalphenoliccontentinextravirginoliveoil TheprocedurewasthesameasforTPCinthemelisaextractbutpreviousphenolextraction wasperformed,asdescribedinpoyatoetal.(2013). 2.4.2Meatpattypreparation Allthepattiescontainedleanbeefmeat(TerneradeNavarraPGI,ES/PGI/0005/0130).Meat wasconvenientlydoublemincedandallpattiesweighed80g. S pattiescontained79.2g meat and 0.8 g common salt. For SM patties, salt was substituted with enriched salt (previouslypreparedbymixtureandhomogenizationwiththem.officinalisextract:16gsalt+ 64,80,104,200,600or800mgmelisaextract).Formulationof E pattiesconsistedof75.2g ofmeat,0.8gsaltand4gofanoilinwateremulsion.tomaketheemulsion,52.63gofextra virgin olive oil was slowly added to 42.1 g water (containing 5.3 g soya protein), while continuouslyhomogenizingwithanultraturrax.for EM patties,melisaextract(250,300or 400mg)wasaddedtothewaterphaseoftheemulsionbeforemixingwithoil. Mixtureofingredientswascompressedwithaconventionalburgermakeruntilacompacted andhomogenizedpattywasobtained(80g,8.6cmdiameterand1.5cmthickness). 2.4.3Cookingprocedure Forthedifferenttypesofmeatpatties,fourindependentbatcheswereprepared,eachone containing4patties(twotokeeprawandtwoforcooking).pattieswereputinapreheated 106

Results ovenat185 Cfor12min,reaching65 Cofinternaltemperature.Justafterthecooking process,theywerecooleddownfor10min,weighted,minced,andstoredat 20 Cunder vacuumuntiltheanalysis. 2.4.4M.officinalisextractaddition:sensoryevaluation Thedeterminationoftheadequatequantityofmelisaextracttobeaddedtothemeatpatties (SMandEM)wasdonethroughsensoryanalysiswith9semitrainedpanelists,bymeansofa trianglesensoryanalysisoncookedsamples.panelistsweretrainedbyallowingthemtotaste beef patties in which different doses of melisa were added, in order to help in the identificationofitstaste.thecomparisonsmadewere:svssmandevsem.ineachcase, differentamountsofm.officinalisextractwereaddedtotherespectivetypeofpatty(500, 375, 125, 65, 50 and 40 µg/g patty for SM and 200, 150 and 125 µg/g patty for EM). Additionally,differencesbetweenSandEwereanalyzedtotakeintoaccountthepotential effectoftheemulsion,evenwithouttheextract,overthesensoryevaluationoftheproduct. Foreverycomparison,eachpanelistwaspresentedwiththreesamples,ofwhichtwowere identical,andaskedtoindicatewhichonedifferedfromtheothers.thisprocesswasrepeated severaltimes,onceforeachdifferentconcentrationofextracttested.thenumberofcorrect answersforeachtypeofcomparisonwasdetermined.accordingtoiso4120:2004,fora9 memberpanel the difference between samples was significant if the number of correct answerswas6(p<0.05). 2.4.5Moisture,fatandcholesterolcontent AOACofficialmethodswereusedformoistureandtotalfatquantitativedetermination(AOAC, 2002a,b).Thedeterminationofcholesterolwassimilartothatofthemodelsystemsamples, butpreviousextractionwasmadeaccordingtokovacsetal.(1979). 2.4.6TBARSdetermination TBARSvaluesweredeterminedonpreviouslyextractedfataccordingtothemethoddescribed bypoyatoelal.(2013).theabsorbancewasmeasuredat532nminafluostaromega spectrofluorometricanalyzer(bmglabtechnologies,offenburg,germany). 2.4.7CholesterolOxidationProductsdetermination Approximately0.5gofthepreviouslyextractedfat(asreportedbyFolch,J.,Lees,M.,Stanley, G.H.S., 1957) was weighted in a flask containing 1M KOH in methanol and 1 ml 19 hydroxycholesterol(20g/mlinhexane:isopropanol3:2)andkeptatroomtemperaturefor 20htocompletethecoldsaponification.Theunsaponificablematerialwasextractedwith 107

Results diethylether(3x10ml).thewholeorganicextractwaswashedwithwater(3x5ml)and filteredthroughanhydroussodiumsulfate.thenitwasrecoveredinaroundbottomflask,and thesolventwasevaporatedunderastreamofnitrogen.purificationbynh 2 SPE,derivatization totrimethylsilylethersandanalysisbygcmswereperformedfollowingthesameprocedure asinthemodelsystem(rosesallinetal.,1995;menéndezcarreñoetal.,2008b). 2.5Statisticalanalysis Forthestatisticalanalysisofthedata,Stata12program(SataCorpLP,Texas,U.S.A.)wasused. Meanandstandarddeviationofdataobtainedfromeachreplicatewerecalculated.Forthe evaluationofthesignificantdifferencesoftheamountsofcholesterol,cholesteroloxidesand TBARSalongtimeandamongdifferentsamples,onefactorANOVAwithBonferroniposthoc multiplecomparisons(p<0.05)wasapplied. 3.Resultsanddiscussion 3.1.Modelsystem 3.1.1.EffectofM.officinalisextractoncholesteroldegradation Figure2showsthepercentageoftheremainingcholesterolthroughouttheheatingprocessof cholesterolheatedwithandwithoutmelisa(2mgmelisa/100mgcholesterol).asignificant dropwasnoticedforsampleswithoutmelisaextract(control)after10minofheating,when thepercentageofremainingcholesterolwas66%,whereaswithmelisaitremainedat93%. Degradationcontinuedfor50moreminutes.Asithasbeenpreviouslyfoundinstudiesdealing withneatcholesterolthermalstability(barriusoetal.,2012;ansorenaetal.,2013a),thefirst stagesofheating(1020min)werealsocriticalat180 C.Throughoutthewholeprocess,the valueswerealwayslower(p<0.05)forthecontrolthanforthetreatedsamples,reaching23 and69%attheendofheating(180min),respectively.soitcanbestatedthatm.officinalis extract,atthedoseappliedinthisstudy,protectedcholesterolfromthermaldegradation. Knownantioxidantssuchasgreenteacatechinsandquercetin(200ppm)havepreviously demonstratedtheireffectivenessduringcholesterolheatingat180 C,where,after30min, around 60 and 95 % ofinitial cholesterol were found in control and antioxidanttreated samplesrespectively(xuetal.,2009).yenetal.(2010),using5%conjugatedlinolenicacidin cholesterol,alsofoundasignificantdecreaseincholesteroldegradation(54vs67%,for control and treatment). However, a study using rosemary extract showed no significant differencesincampesteroldegradationafter4hat180 C,althoughdifferencesamongtotal steroloxidationproductscontentweredetected(kmieciketal.,2011). 108

Results Thecholesteroldegradationcurvepresentedamuchhigherslopeforcholesterolalonethan formelisacontainingsamplesduringthefirst10min,butverysimilarslopescouldbeobserved thereafterforbothsamples.thiscoulddenoteahighprotectiveeffectofthemelisaduringthe first10minandslowerprotectionthereafter.accordingly,theantioxidantcapacityvalues (ORAC)foundforthemodelsystemthatincludedthemelisaextractwerereducedafterthe first10min.figure3showsthatapproximatelyhalfoftheantioxidantcapacityinitiallynoticed inthemodelsystemwaslostafter10minofheating,decreasingfrom43.11to23.71mgtrolox /gsample. Thehighcontentofphenoliccompoundsintheextractmatrix(TPCwas356µggallicacid/mg extract)couldexplainitsantioxidanteffect.asthemajorantioxidativecompoundinthiswater melisa extract was rosmarinic acid (123 mg / g extract), monitorization of its remaining concentrationduringtheheatingprocesswasalsodone(figure3).asimilardecreasingcurve asthatoforacdeterminationwasobserved,withadecreaseofaround50%afterthefirst10 min.thus,ahighcorrelationbetweentheantioxidantcapacitylossandrosmarinicacidloss wasnoticed(pearsonr=0.9517).noantioxidantcapacitywasnoticedwhencholesterolwas heatedalone,exceptafter180min,where0.60mgtrolox/gsamplewasdetected,meaninga 5.1%oftotalORACvalueatthispoint.Therefore,theprotectiveeffectobservedformelisa extractinthecurrentstudywasmainlyattributedtoitshighrosmarinicacidcontent,whichis acompoundknownbyitsantioxidantcapacity(erkanetal.,2008).nevertheless,evenifits contributionshouldbeveryimportant,othercompoundsfoundintheextract(showingpeaks muchsmallerbutnotquantitated)couldbealsoresponsiblefortheantioxidantproperties owingtosynergisticeffects,asitisstatedinmironetal.(2013). 3.1.2.EffectofM.officinalisextractoncholesteroloxidationproductsformation COPswereprogressivelyformedduringheatinguntiltheyachievedamaximum,andthentheir concentrationstartedtodecrease,followingadifferentpatterndependingonthetypeofcop andsample(figure4).formationofcopswasquickandhighinthecontrolsample.at10min, 94µgoftotalCOPspermginitialcholesterolwasformedinthecontrolsample,whereas practicallynocopswereformedinthemelisatreatedsample.thisbehaviorisinaccordance with data from cholesterol degradation, where the best antioxidant effectiveness was recordedduringthefirst10minoftreatment. ThemelisacontainingsamplecontinuedyieldingCOPsforalongertime,sincetheprocesswas retardedinrespecttothecontrol.consequently,themaximumcopslevelwasachievedat30 and120minforcontrolandtreatedsamples,respectively,yielding142.97and93.03µg/mgin 109

Results controlandinmelisatreatedsamples.similartimes(10and20min)wererequiredinprevious studiestoreachmaximumcopslevelsinneatcholesterolsamplesat180 C(Barriusoetal., 2012;Ansorenaetal.,2013a).Therefore,itcanbestatedthatM.officinalisextractsinhibited cholesteroloxidationproductsformationbybothdelayingtheirappearanceanddecreasing theirformationrate. COPsformationhasbeenpreviouslyreportedtobedepletedinthepresenceofphenolic compoundssuchasgreenteacatechinsandquercetin(xuetal.,2009)fromaround12%to lessthan5%withrespecttotheinitialcholesterolcontentafter30minat180 C.Inthe currentstudy,atthesametemperaturetimeconditions,similarreductionwasfound:from31 to 11 % cholesterol oxidation. In general, better results have been observed for natural antioxidantsthanforsyntheticonesregardingsteroloxidationproductsinmodelsystems(xu etal.,2009;kmieciketal.,2011). TotalCOPsbehaviorwassignificantlyaffectedby7ketocholesterol(Fig4g),whichwasthe most abundant COP among those analyzed. It was followed by epoxy and hydroxyl compounds,withtriolatnegligiblelevels(fig4e),asexpected,givingthelackofwaterinthe medium(lampietal.,2002).25hydroxycholesterolonlysufferedasmallincrease(fig4f) whichwasalsoexpectedduetothesterolchainlowerlikelihoodtooxidizeintheabsenceof enzymes. 3.2.Foodsystem:meatpatties 3.2.1.IncorporationoftheM.officinalisextractintomeatpattiesandsensoryevaluation Themelisaextractdoseusedinthemodelsystemwas2mgmelisa/100mgcholesterol.To extrapolatethisconcentrationtothefoodmatrix(meatpatty)ithastobeconsideredthat cholesterolisnottheonlylipidcompoundsusceptibletooxidationinthisfoodstuff.taking intoaccountthisfact,theconcentrationchosenwas2mgmelisa/100mglipidfraction,which correspondedto500µgmelisa/gmeatpatty.whenthesensoryevaluationwasperformedon thesemeatpatties,anunpleasanttastewasclearlydetectedbypanelists.therefore,sensory evaluationofmeatpattysamplescontainingdecreasinglevelsofmelisaextract(sm)was carried out until a nondetectable concentration of melisa was noticed. The comparison betweenthecontrolpatty(s)andthedifferentmelisacontainingpatties(sm)inthetriangle sensorytest(table1a)revealedthatpanellistswereabletodetectsignificantdifferenceswith dosesover65µg/gpatty. Inordertocomparetheantioxidantefficiencyofmelisaextractwiththatofarecognized potentantioxidantinmeatpatties(rodríguezcarpenaetal.,2012b),beefpattiescontaining 110

Results extravirginoliveoil(e)wereprepared.besides,pattiescontainingbothextravirginoliveoil andmelisaextract(em)werealsopreparedtocheckforpossibleadditionalorsynergistic effectsofmelisaextractandoliveoil.thetastyandflavorfulpropertiesofoliveoilwould efficientlymaskmelisaoddflavorandwouldpermittoenhancemelisadoseinpatties.oliveoil wasappliedthroughoilinwateremulsion,wheremelisaextractwassolvedwithinthewater phase.thistechnologyhasbeensuccessfullyappliedpreviouslybyourgroupinothermeat products(garcíaíñiguezdecirianoetal.,2010b;berasategietal.,2011)andithasalsobeen usedbyotherauthors(lópezlópezetal.,2010)forimprovingthenutritionalpropertiesof lipidfractionofnewmeatproductformulations.inthiscase,theemulsionwasagoodsystem toincludehigheramountsoftheantioxidant,sincedirectcontactwithtastebudsandmelisais avoided. Ithastobepointedoutthatthepercentageatwhichtheemulsionwaspresentinthe formulation(5%)didnotmodifythetypicalsensorypropertiesofbeefpatties,aspanelists werenotabletodiscriminatebetweensandesamples(p<0.05). Then,increasingconcentrationsofmelisavehiculizedthroughtheemulsion(EM)wereadded andsensorytestswereperformedfacingthemtoesamples(table1b).asitwashypothesized, resultsledtotheconclusionthatthelevelofundetectablemelisaextractwasabletobe increasedupto150µg/ginemulsioncontainingpatties(em),comparedto65µg/ginpatties wheremelisawasnotvehiculizedwithinanemulsion(sm). HigherdosesofM.officinalisextracts(965and686µg/g)thanthoseusedinthecurrentstudy havebeenpreviouslyaddedtomeatproductswithoutnoticingsensoryproblemsberasategi etal.(2011)withbolognatypeproductsandgarcíaiñiguezdecirianoetal.(2010a)with fermentedproducts.nevertheless,allthesestudiesdealtwithsamplesrichinaromasand flavorsfromgarlicorredpepper,whichcaneasilymaskthesensoryoddflavormelisanotes. Thiswasnotthecaseofourfreshbeefpatties(containingaquitesimpleformulation:beef meat,water,oliveoil,soyaproteinandsalt). 3.2.2.EffectofM.officinalisextractonlipidoxidation Lipidoxidationinrawandcookedconditionswasassessedforthefollowingfourtypesof patties:simple,withoutandwith65µg/gmelisaextract(sandsm)andemulsioncontaining sampleswithoutandwith150µg/gmelisaextract(eandem).overalllipidoxidationresults areshowninfigure5andcholesteroloxidationwasmonitoredthroughcopsdetermination (Table2).Rawsamplesdidnotshowsignificantdifferencesamongthefourtypesofpatties, 111

Results presentingmeanvaluesaround0.1mgmda/kgandfrom538to913µgcops/00gdry sample. CookingsignificantlyincreasedtheTBARSinsimplepattiesbutnoefficientprotectionofthe melisaextractwasdetected(ssm).thesamebehaviorwasobservedforcops:theirvalues significantly increased after cooking in all four types of patties, but similar values were reportedforcookedsimplepattieswithandwithoutmelisaextractaddition(ssm). Comparisonbetweensimplepattiesandthosethatwereincorporatedwitholiveoilemulsion allowedustoconcludethattheemulsionprotectedfromlipidoxidation,probablyduetothe highphenoliccontentofoliveoil(143ppmgallicacid).potentialantioxidantpropertiesofsoy proteincontainedintheemulsion(bloukasetal.,1997)couldalsobebehindthisbehavior. When the higher dose of melisa (150 ppm) was used within the olive oil emulsion, no additionalprotectiveeffectofmelisawasobservedoverthatoftheoliveoilneitherfortbars norforcopsvalues(0.18vs0.18mgmda/kgand1030vs972µgcops/100gdrysamplefore vsem,respectively). Theconcentrationwasashighasthesensoryqualityallowedguaranteeinggoodflavor,soit had to be concluded that melisa extract was not efficient in these conditions. These unfavorableresultswereprobablycausedbythelowdoseofmelisausedinthecurrent experiment:sensoryrequirementshaveforcedthedecreaseoftheconcentrationofmelisa extractbelowthelevelatwhichantioxidanteffectscanbeobservedinthemeatsystem. Besides, rosmarinic acid could have disappeared during cooking, decreasing the extract antioxidantcapacity,asitoccursinthemodelsystem(figure3). Anumberofstudieshavereportedmoresuccessfulresultsindifferentmeatpattiesenriched withhighphenolicextracts.rodríguezcaprenaetal.,(2012a);sampaioetal.,(2012)and Duthieetal.,(2013)obtainedfrom20to85%ofreductiononTBARSvalues,anddetectedup to1.8mgmda/kg.inchickenandporkpatties,mariuttietal.(2011),rodríguezcarpenaetal. (2012a)andKarwowskaetal.(2014)reportedreductionsfrom2to7foldasaconsequenceof antioxidantenrichment(fromsage,mustardandavocado),andthecopscontentsranged between90and1350µg/100g,probablyduetothehigherdosesapplied.however,notallof them assessed sensory evaluation of the products, which is critical for acceptability. KarwowskaandDolatowski(2014),usingporkmeataddedwithmustardseed,didnotfindany protectionagainstmdaformationattheendofthecookingprocedure(asinthecurrent study),butonlyafter12daysofstorage. 112

Results Inthecurrentstudy,M.officinaliswasmuchmoreeffectiveinthemodelsystemthaninmeat patties.thisisfrequentlyreportedincomparativestudies:greenteacatechins,tocopheroland quercetinweremuchmoreeffectivewhencholesterolwasheatedalonecomparedtowhen theexperimentwasmadewithinlard(xuetal.,2009).thislowerefficiencyfoundinfoodlike systemscanberelatedtothepresenceofothersurroundinglipids,whichcanactasprotective factorsthemselves(yenetal.,2010;yenetal.,2011;ansorenaetal.,2013). Inconclusion,M.officinalisaqueousextractprotectedcholesterolfromoxidationinthemodel system,butnoprotectiveeffectwasfoundinmeatpattiesatsensoryacceptabledoses. Therefore,attentionshouldbepaidtosensoryconsiderationsintheevaluationofnatural extractsasasourceofbioactivecompoundsinfoods.besides,newtechnologiesforthe incorporation of these possible ingredients should be developed, such as the use of encapsulatedstructuresorgelledemulsions. 4.Acknowledgements Wethank PlandeInvestigaciónUniversidaddeNavarra (PIUNA)forthefinancialsupport.B. Barriusoisgratefulto AsociacióndeAmigosdelaUniversidaddeNavarra forthegrant received. Remaining cholesterol (%) a b cholesterol cholesterol+melisa 100 c cd 80 e de e 60 b c c 40 d d e 20 0 0 50 100 150 min Figure2.Remainingcholesterol(%)duringthe heating process in model system. Different letters for each sample denote statistical differencesalongtime(p<0.05).ateverytime ofanalysis,studentttestcomparedbothtypes ofsamples(*p<0.05;p<0.01; p<0.001) mg trolox / g sample 60 50 40 30 20 10 0 a b a bc b b b b ORAC c b rosmarinic acid de 0 50 100 150 min e 2.5 2.0 1.5 1.0 0.5 0.0 Figure 3. Antioxidant capacity (ORAC determination) and rosmarinic acid content duringtheheatingprocessinthemodelsystem (cholesterol+melisasample).differentletters foreachsampledenotestatisticaldifferences alongtime(p<0.05). b b mg rosmarinic acid / g sample 113

Results 15 15 c c c 10 5 0 bc c bc c a * b a a bc d ab 0 50 100 150 200 * e 10 bc e d b 5 * c c d a a a b 0 a 0 50 100 150 200 b 40 de e de 40 µg/mgcholesterol 30 20 10 cd c b a 0 a a e * * d c b c 0 50 100 150 200 30 20 10 0 c c b a c b a a c c c e d d 0 50 100 150 200 1.2 1.0 0.8 0.6 0.4 0.2 0.0 e d d d a b c b c d e d a a 0 50 100 150 200 1.2 1.0 0.8 0.6 0.4 0.2 0.0 d c b c a a b a d cd cd e e d 0 50 100 150 200 70 60 50 40 30 20 10 0 a a b c d c b a c e 0 50 100 150 200 b e b d cholesterol Figure4.COPsinthemodelsystemduringtheheatingprocessa)7HC,b)7HC,c)5,6EC,d)5,6EC,e)CT,f)25HC, g)7kc,h)totalcops.differentlettersforeachsampledenotestatisticaldifferencesalongtime(p<0.05).ateverytime ofanalysis,studentttestcomparedbothtypesofsamples(*p<0.05;p<0.01; p<0.001). 150 120 90 60 30 a a 0 cd b time (min) d c b a d e bc * f 0 50 100 150 200 cholesterol+melisa d b 114

Results Table1.Scoresoftrianglesensoryanalysis.Comparisonsbetweena)SandSMpatties;b)EandEMpatties. a) SvsSM 500µg/g 375µg/g 125µg/g 65µg/g 50µg/g 40µg/g Correctreplies 9*** 9*** 7* 4ns 2ns 0 Incorrectreplies 0 0 2 5 7 9 b) EvsEM 200µg/g 150µg/g 125µg/g Correctreplies 9* 4ns 1ns Incorrectreplies 7 11 6 Figure5.TBARSofbeefpatties(S,SM,EandEM).Foreachtypeofpatty,thefilledbarindicatesrawsampleandthe stripedbarindicatescookedsample.differentlettersdenotesignificantdifferencesamongsamples(p<0.05). Table2.COPsconcentration(µg/100gdrysample)inrawandcookedbeefpatties(S,SM,EandEM).Different lettersdenotesignificantdifferencesamongsamples(p<0.05). mg MDA / Kg 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 a c a c 1 a ab ab S-raw SM-raw E-raw M-raw Sraw Sc SMraw SMc Eraw Ec EMraw EMc 7HC 76.26a 155.21b 72.84a 168.06b 55.24a 69.26a 52.66a 62.21a 7HC 113.81a 234.76b 106.17a 242.25b 80.83a 113.25a 83.96a 105.89a 5,6EC 286.57ab 529.20c 260.49a 533.72c 144.91a 370.90a 240.51a 351.49bc 5,6EC 55.21ab 57.69a 56.79a 62.60a 20.53c 64.57b 37.30b 64.59a CT 22.55a 21.76a 29.63ab 13.43a 33.73b 20.63a 19.61a 24.84a 25HC 11.38a 13.01b 11.73a 9.94a 15.54c 12.84b 12.84b 14.11c 7KC 347.79ab 761.47d 335.80ab 700.05d 187.43a 378.54a 270.87a 349.25bc TotalCOPs 913.58ab 1773.10d 873.46ab 1716.46d 538.22a 1030.01c 717.75a 972.39bc b S-c SM-c E-c M-c 115

Results 116 Figure1 Figure2 Numbers 25HC,8 a) b) c) d) (SupMat).HPL 2(SupMat).GC sindicatetheid :7KC. 1 LCchromatogra Cchromatogra dentifiedcops; 2 mofmelissao amsofbeefpa ;1:7HC,2:1 3 5 4 officinalisextrac atties.a)sraw 19hydroxychole 5 6 Rosmar ct. wb)scooked esterol,3:7h 6 7 inic acid c)emcooked HC,4:5,6EC,5 8 7 dandd)ecoo 5:5,6EC,6:C oked. CT,7:

Results Table1(SupMat).Moisture(%),lipid(%)andcholesterol(mg/100g)contentofthebeefpatties. Sraw Sc SMraw SMc Eraw Ec EMraw EMc moisture 73.16±0.83 66.05±0.42 73.36±0.95 65.88±0.61 71.58±0.06 65.83±0.39 71.53±0.05 66.32±0.21 lipid 2.47±0.16 3.12±0.14 2.61±0.25 3.30±0.14 4.53±0.22 6.09±0.12 4.70±0.21 5.58±0.13 cholesterol 50.30±1.34 54.08±2.04 51.05±1.60 54.22±1.95 47.63±0.66 54.70±1.27 45.04±1.32 54.32±1.65 117

ResultsVI Paper5 Solanumsessiliflorum(manacubiu)antioxidant protectiveeffecttowardscholesteroloxidation: influenceofdocosahexaenoicacid

Results PlantFoodsforHumanNutrition(underrevision) Solanumsessiliflorum(manacubiu)antioxidantprotectiveeffect towardscholesteroloxidation:influenceofdocosahexaenoicacid BlancaBarriuso a,lilianreginabarrosmariutti b,dianaansorena a, IciarAstiasarán a,neurabragagnolo b a DepartmentofNutrition,FoodScienceandPhysiology,FacultyofPharmacy,UniversityofNavarra b DepartmentofFoodScience,FacultyofFoodEngineering,UniversityofCampinas Abstract Harmfulhealtheffectshavebeenattributedtocholesteroloxidationproducts(COPs).Factors thatmodulatetheirformationinfoodsarelight,oxygen,heat,andfoodmatrix(suchas antioxidantscontentorunsaturationdegreeoflipids),amongothers.theobjectiveofthis workwastoassestheeffectivenessofanextractobtainedfromsolanumsessiliflorum(mana cubiu)(mce)asapotentialinhibitorofcholesteroloxidationunderheatingconditions.the influenceoffreedhapresenceinthesystemwasalsoevaluated.resultsshowedthatmce inhibitedcholesteroldegradation(44%vs18%withoutandwithmce,respectively)and reduced9foldcopsformationintheabsenceofdha.however,whendhawaspresent,the MCEwasnoteffectivetowardscholesteroloxidation.Inthiscase,MCEshoweditsantioxidant effectprotectingdhafromdegradation(89%vs64%).antioxidantpropertiesofthissolvent freenaturalextractmakemceapotentialgoodingredientinfoodproductscontaininghighly polyunsaturatedlipids. Keywords:oxysterols,docosahexaenoicacid,oxidation,naturalantioxidants Highlights: 1. Manacubiu aqueous extract inhibited cholesterol degradation and reduced COPs formationinamodelsystem. 2. Manacubiuaqueousextractwasnoteffectivetowardscholesteroloxidationinthe presenceofdha. 3. PresenceofDHAinhibitedcholesteroldegradation. 4. Refrigerationstorageupto72hdidnotshowanyeffectonlipidoxidation. 121

Results 1.Introduction Cholesterolchemicalstructuremakesitaneasytooxidizemolecule,leadingtotheformation ofoxysterols.theseoxidationproducts,usuallynamedascholesteroloxidationproducts (COPs),havebeenrelatedtoseveraldiseases(atherosclerosis,neurodegenerativediseases, mutagenicandcarcinogeniceffects,etc)(otaeguietal.,2010).theyhavebeenfoundina varietyofanimalfoodsamples(otaeguietal.,2010),andsomestudieshavepointedouttheir potential absorption through the diet (Meunier et al., 2003; Baumgartner et al., 2013). Therefore,minimizingtheirformationduringfoodmanufacturing,processingand/orcooking isofgreatinteresttoreducehealthrisk.useofadditivesoroxygenrestrictionmethodsis commonlyappliedforthatpurpose. Inthissense,agrowinginterestinnaturalantioxidantsfoundinplantsisnoticed,notonly becauseoftechnologicalreasons,butalsoduetotheirpotentialabilitytosuppressoxidative stressandrelateddiseases.regardingtheirusefulnessinfoodsystems,successfulcasesin controllingcholesteroloxidationhavebeenreportedinseveraltypesofmatrices(mariuttiet al.,2011;sampaioetal.,2012;karkwowskaetal.,2014).manacubiu(solanumsessiliflorum)is afruitnativetoamazonia,whichpossessantioxidantpropertiesattributedtothepresenceof carotenoidsandphenoliccompoundsinitscomposition(rodriguesetal.,2013). Ontheotherhand,theinterestinhighlyunsaturatedfattyacidshasrecentlyincreaseddueto theirhealthrelatedproperties.particularly,longchainomega3polyunsaturatedfattyacids have demonstrated cardiovascular disease lowering effects (Mozaffarian et al. 2011). Nevertheless,inadequatemanufacturingandcookingconditionscanleadtosomelossinthe contentoftheseinterestingcompounds.anumberofstudieshavedealtwiththeprevention offattyaciddegradationthroughantioxidantaddition,afterdifferentcookingandstorage conditions(sampaioetal.,2012;bhaleetal.,2007;sanchoetal.,2011).apossibleinteraction betweencholesteroloxidationandthesurroundingfattyacidshasbeenproposedbyseveral authorsasafactorthatmodulatescholesteroloxidationsusceptibility,althoughnoconsensus onthesubjecthasbeenfound(soupasetal.,2004;xuetal.,2009;ansorenaetal.,2013a). Foodsareusuallycomplexmatriceswhereinterferencesamongseveralcomponentsmay hamperaclearviewaboutthemechanismsofcholesteroloxidation.therefore,modelsystems areaveryusefultooltoevaluateseparatelythefactorsthatexertaninfluenceinthisprocess. Avarietyofantioxidants(Xuetal.,2012;Chienetal.,2006;Kmiecketal.,2011),andlipid matrices(ansorenaetal.,2013a;lehtonenetal.,2012)havebeentestedinmodelsystems. 122

Results Considering the exposed above, the aim of this study was to evaluate the antioxidant protectiveeffectofasolanumsessiliflorumlyophilizedaqueousextractagainstcholesterol degradation and cholesterol oxidation products formation in a model system, with and withoutthepresenceofdocosahexaenoicacid(dha).furthermore,theeffectofrefrigeration storagewasalsoevaluated. 2.Materialandmethods 2.1Materialandreagents Manacubiufruits(~21Kg)wereacquiredatCEAGESP(SãoPauloGeneralWarehousingand Centers Company, São Paulo, Brazil). Cholesterol, 22Rhydroxycholesterol, 22S hydroxycholesterol, 20hydroxycholesterol, 25hydroxycholesterol, 5,6epoycholesterol, 5,6epoxycholesterol,7ketocholesterolandDHAstandardswerepurchasedfromSigma Aldrich.7hydroxycholesteroland7hydroxycholesterolwerepurchasedfromSteraloids. The purity of the standards was at least 95% as determined by HPLC or GC analyses. Chloroform and methanol were purchased from Synth. Chromatographic grade hexane (minimum63%nhexane)and2propanolwerepurchasedfrompanreac. 2.2Manacubiuextractpreparationandcharacterization Manacubiufruitswerelyophilizedbeforeextraction(Rodriguesetal.,2013).Fiftygramsof lyophilizedmanacubiufruitwerehomogenizedwithultrapurewaterinavortexfor5minand centrifugedat20000gat10 C.Theaqueouslayerwaslyophilizedduring120hat92 Cbelow 40 µhg (Liobras, São Paulo, Brazil). The identification and quantification of the phenolic compoundsofthemanacubiuextract(mce)wascarriedoutaccordingtorodriguesetal. (2013). 2.3Samplepreparationandheating Stocksolutionsofcholesterol(1mg/mLinchloroform),DHA(1mg/mLinchloroform)andMCE (2.5mg/mLinmethanol)wereprepared.Fourtypesofsampleswereprepared:cholesterol alone,cholesterolwithmce,cholesterolwithdha,cholesterolwithdhaandmce.aliquotsof cholesterolsolution(1ml)werepouredintesttubes.fordhaandmcecontainingsamples,1 mland0.2mlofthecorrespondingstocksolutionwasadded,respectively.solventwas evaporatedunderastreamofn 2 anduncappedtubeswereplacedinadryblock(marconi, Brazil)at180 C.After7minheating,tubesweretakenoutandintroducedintoanicewater bathfor4minandcapped.then,theywerekeptinthefridge(4 C)for72horinthefreezer( 30 C)untilanalysis.Theexperimentwascarriedoutintriplicate. 123

Results 2.4CholesterolandCOPsdetermination Eachsamplewasdissolvedwith1mLhexane:2propanol(97:3,v:v),analysedbyHPLCUVRI, followingthesameprocedureasinmariuttietal.,(2008).identificationofthecompoundswas confirmedbyhplcapcims/ms,asinandzardettoetal.,(2014).chromatographicandms andms/msdataareshownintable1. 2.5DHAdetermination DocosahexaenoicacidwasconvertedintoitsmethylesteraccordingtoJoseph&Ackman (1992)andanalyzedwithagaschromatograph(GC2010model,Shimadzu,Kyoto,Japan) equippedwithafusedsilicacpsil88capillarycolumn100mx0.25mmi.d.,0.20umfilm thickness (Chrompack, Middelburg, The Netherlands) and flame ionization detector. ChromatographicconditionsweredescribedindetailbySanchoetal.(2011). 2.6Statistics ThedataobtainedwereanalyzedbymeansofthesoftwareStata12(SataCorpLP,Texas, U.S.A.).Fortheevaluationofthesignificantdifferencesamongtheamountsofcholesteroland COPsofdifferentsamples,onefactorANOVAwithBonferroniposthocmultiplecomparisons (p<0.05)wasapplied.fortheevaluationofthesignificantdifferencesbetweentheamounts ofcholesterol,copsanddhaafter0and72hofstorageunderrefrigeration,studentttest wasapplied. 3.Resultsanddiscussion 3.1MCEproperties Theprofileofphenoliccompounds(Figure1)intheaqueousmanacubiuextractshowedthat 5caffeoylquinicacid(5CQA)wasthemainone,representing2.48mg/gextract.Thisisa slightlylowervaluethanthatobtainedinamethanolwaterextractofthesamefruit(4.49 mg/gextract)(rodriguesetal.,2013),wheretwoothercompoundswerealsofound:bisand trisdihydrocaffeoylspermidine. In the current extract, a small amount of bis dihydrocaffeoylspermidinewasdetected,butitwasbelowthelimitofquantification.despite itslowerphenoliccontent,theaqueousextractwasinteresting,since5cqahasdemonstrated highantioxidantcapacity.thiscompoundisusuallyfoundinhighamountsincoffee,especially greencoffee,orcoffeeextracts,whichhavebeenappliedbothinmodelandfoodsystemsto protectthemfromoxidationortoincreasetheirantioxidantcapacity(budrynetal.,2014;lin etal.,2015).ontheotherhand,thisextractwassafe,environmentallyfriendlyandpotentially 124

Results applicableinfoodstuffs,sinceitwasfreefromorganicsolvents.hence,theaqueousmcewas selectedfortheexperimentscarriedoutinthiswork. 3.2EffectofMCEoncholesterolandDHAdegradation Theinitialamountofcholesterol(beforeheating)was1.05mg(Table2)inallsamples.All samplesshowedasignificantdecreaseinthecholesterolcontentafterheating(presenting valuesbelow0.82mginallcases).asitcanbeobservedinfigure2,higheramountsof remainingcholesterolwerefoundwhenmcewasaddedtothesample(80%),comparedto theremainingamountpresentwhencholesterolwasheatedalone(55%).thisreductionin cholesterol degradation was attributed to the high content in 5CQA of MCE and its antioxidantcapacity.similarreductionsincholesteroldegradationhavealsobeennoticedin othermodelsystemsusingphenoliccompoundssuchasgreenteacatechinsandquercetin(xu etal.,2009;chienetal.,2006). Ontheotherhand,similarvaluesofremainingcholesterol(around20%)werefoundforthe twotypesofsamplesthatincludeddhainthemixture,regardlessthepresenceofmce.as comparedtocholesterolalone,thepresenceoffreedhaenhancedcholesteroldegradation, andmcecouldnotcounteractthiseffect.therefore,mceseemedtoprotectcholesterolfrom oxidationintheabsenceofdha,butnotinthepresenceofthisfattyacid. TheinitialamountofDHA(beforeheating)was1.00mg(Table2).DHAremainingcontentafter thethermaltreatmentwasalsoanalyzed(fig3).bothmcelackingandcontainingsamples showedasignificantdecreaseinthedhacontentafterheating(0.11and0.36mgremaining, respectively).heatingofdhaalone(withoutcholesterolnormce)resultedin11.53±3.05% ofremainingcompound,socholesterolhadnoeffectondhathermaldegradationgiventhat thepresenceofcholesterolinthemixtureyieldedthesameremainingamount(11%).results showedthatdhacontentwasmuchhigherinthepresenceofmce(36%),comparedtothe previouslymentionedremainingamountfoundintheabsenceoftheextract(11%).similar protectiveeffectsofnaturalextractsagainstdhadegradationhavebeenalsoreportedin studiesdealingwithfishmeatballsandfishoil(bhaleetal.,2007;sanchoetal.,2011).hence, itcouldbeassumedthatmceantioxidantpropertiesweredevotedtoprotectdhafrom degradation,lesseningtheprotectiveeffecttowardscholesteroldegradation. 125

Results 3.3EffectofMCEonCOPsformation COPscontentwasmuchhigherincholesterolalonesample(227µgCOP/mgcholesterol)than inthepresenceofmce(25µgcop/mgcholesterol),asitcanbeobservedintable3.onthe otherhand,similarvalues(around87µgcop/mgcholesterol)werefoundforbothmce containingandmcelackingsampleswhendhawaspresentinthemedium.soagain,asit occurredwithcholesteroldegradation,mceseemedtopreventfromcopsformationinthe absence of DHA, but not in the presence of this compound. COPs formation has been previouslyreportedtobedepletedinthepresenceofphenoliccompounds(xuetal.,2009; Barriusoetal.,2015). NineCOPswereanalyzedandonlyfivewerefoundinthesamples.Fromthese,7KCwasthe mainoneinmostcases,followedbyecand7hc(table3).anumberofstudiesdealing withcholesteroloxidationinmodelsystemshavereportedthisprofileofcops(xuetal.,2009; Barriusoetal.,2015;RodríguezEstradaetal.,2014;Derewiakaetal.,2015).Thedominanceof isomerwassupportedbythesterichindranceatc3position.interestingly,whencomparing cholesterolandcholesterol+mcesamples,whereasa90%reductionin7kcwasnoticed;only a40%reductionwasreportedin7hc,becomingthemaincompound.sotheantioxidant extractseemedtoshowdifferentialbehaviortowardsindividualcopsformation.inthissense, reactionratemightbesloweddowninthepresenceofmce,remainingas7hcforlonger timebeforestartingtheconversioninto7kc.similarly,kmiecikandcoworkers(2009)found differences among sterol oxides distribution depending on the antioxidant applied. This selective inhibition towards individual derivatives could be related to the differences in chemicalstructure,thatcouldhampercertainpositionstobeattackedand,hence,certain oxidationderivativestobeformed. 3.4EffectofDHAoncholesteroldegradationandCOPsformation CholesteroldegradationwashigherwhenheatedwithinDHAthanwhenheatedalone,asit can be observed in figure 3. The presence of a lipid unsaturated surrounding has been reportedtoprotectcholesterolfromoxidation(ansorenaetal.,2013a;barriusoetal.,in press).thisdiscordancecouldberelatedtothehigherratiocholesterol:lipidmatrixusedinthe currentstudy(1:2)comparedtothoseones(1:100).higheramountsofcholesterolcouldhave hamperedthephysicalprotectionandfavouredcholesterolinteractionwithhighlyoxidated DHA.Thisway,Lehtonenandcoworkers(2012),usingcholesterylesters(stechiometry1:1) foundhigherlevelsofoxidationincholesteryllinoleatethaninfreecholesterol(0.17%and 0.084%),whichwasattributedtothelinoleatedoublebondslikelihoodtoradicalformation. 126

Results Additionally,usingfreeDHAascomparedtotriglycerides(mainconstituentsofthematrixin Ansorenaetal.(2013a)andBarriusoetal.(inpress))makesalsoanimportantdifference regardingphysicalprotection,chemicalgroupinteractionandviscosity,whicharekeyfactors intheprocess(rodríguezestradaetal.,2014). On the other hand, even though cholesterol degradation was enhanced by DHA, COPs formationwaslowerthanintheabsenceofdha,denotingthattheroutesofcholesterol oxidation were different. Consequently, the oxidation products formed were different, probablyoligomersorvolatilecompounds(derewiakaetal.,2015;sosinskaetal.,2014).itwas alsopossiblethatreactionratesforcopdegradationwerehigherthanforcopformationin thepresenceofdha,givingrisetotheaforementionedcompounds.previousstudieshave shownnocorrelationbetweenthesterolsdegradationandtheoxidesformed(derewiakaet al.,2015;oehrletal.,2001). 3.5Effectofrefrigeratedstorage Storageunderrefrigerationconditions(4 C,72h)modifiedneithercholesterollevelsnorCOPs concentrationinmostcases,exceptfortwosamples.chol+dhasampleslightlydecreasedits contentincops,possiblyduetodegradationofthecompounds(derewiakaetal.,2015).dha levelssufferednochangesalongthetimeeither.thisbehaviourwasattributedtothelackof wateroranyothersolventsinthesamples,whatretardedtheoxidationprocesses. Inconclusion,MCEprotectedagainstcholesteroldegradationandCOPsformationwhenthere wasnootherlipidcompoundinthesystem,butnotinthepresenceofdha.ontheother hand,dhawaseffectivelyprotectedfromoxidationbymceaddition.consideringthatit impliesasolventfreeextractionprocess,thismanacubiuextractcouldbeapotentialgood ingredientinfoodproductscontaininghighlypolyunsaturatedlipids. 4.Acknowledgements We are grateful to the PIUNA (Plan de Investigación de la Universidad de Navarra) and MinisteriodeEconomíayCompetitividad(AGL201452636P)fortheircontributiontothe financialsupportofthiswork.b.barriusoacknowledgesbancosantanderandasociaciónde AmigosdelaUniversidaddeNavarraforthegrantsreceived.Wearegratefulto Redde ExcelenciaConsolider PROCARSE(AGL201451742REDC).N.BragagnolothanksFAPESP(grant 2013/064891)andCNPqforfinancialsupport. 127

Results 200 5-caffeoylquinic acid Detector response at 280 nm (mau) 100 0 N1,N5or N5,N10- bis(dihydrocaffeoyl) spermidine 0 10 20 30 40 50 60 70 time (min) Figure1.ChromatogramobtainedbyHPLCDADofthephenoliccompoundsfromtheaqueousmanacubiuextract. Remaining cholesterol (%) 120 100 80 60 40 20 a b b a b b a b b a c b unheated heated - 0 h stored heated - 72 h stored 0 chol chol+mce chol+dha chol+dha+mce Figure2.Remainingpercentageofcholesteroloftheunheatedsampleandthefourheatedsamplesafter0and72h storage.differentlettersforeachsampledenotestatisticaldifferencesamongtheunheated,the0hstoredandthe 72hstoredsamples. 120 100 a a Remaining DHA (%) 80 60 40 b b unheated heated - 0 h stored heated - 72 h stored 20 b c 0 chol+dha chol+dha+mce Figure3.RemainingpercentageofDHAoftheunheatedsampleandthetwoheatedsamplesafter0and72h storage.differentlettersforeachsampledenotestatisticaldifferencesamongtheunheated,the0hstoredandthe 72hstoredsamples. 128

Results Table1.ChromatographicandmassspectrometrycharacteristicsofcholesteroloxidesobtainedbyHPLCMS/MS. Cholesterol oxide t r (min) [M+H] + (m/z) Fragment ions (m/z) 22R-hydroxycholesterol 3.7 nd 385 * [M+H-18] +, 367 [M+H-18-18] + 22S-hydroxycholesterol 4.2 nd 385 * [M+H-18] +, 367 [M+H-18-18] + 20-hydroxycholesterol 4.4 nd 385 * [M+H-18] +, 367 [M+H-18-18] + 25-hydroxycholesterol 4.7 403 385 [M+H-18] +, 367 [M+H-18-18] + 7-hydroxycholesterol 5.5 nd 385 * [M+H-18] +, 367 [M+H-18-18] + 7-ketocholesterol 5.7 401 383 [M+H-18] +, 365 [M+H-18-18] + 7-hydroxycholesterol 5.8 nd 385 * [M+H-18] +, 367 [M+H-18-18] + 5,6-epoxycholesterol 7.0 403 385 [M+H-18] +, 367 [M+H-18-18] + 5,6-epoxycholesterol 7.6 403 385 [M+H-18] +, 367 [M+H-18-18] + nd:notdetected.* Insourcefragmentation. Table2.CholesterolandDHAcontent(mg)ofunheatedcholesterolandDHA,andthefourheatedsamplesduring storageat4 Cfor0and72h. unheated chol chol+mce chol+dha chol+dha+mce 0 h 72 h 0 h 72 h 0 h 72 h 0 h 72 h cholesterol 1.05 aa 0.56 c 0.54 C ns 0.82 b 0.81 B ns 0.22 d 0.22 D ns 0.25 d 0.29 D * DHA 1.00 aa - - - - 0.11 c 0.07 C * 0.36 b 0.41 B ns Differentlowercaselettersdenotestatisticaldifferencesamongtheunheatedsampleandtheheatedsamplesafter 0h.Differentcapitallettersdenotestatisticaldifferencesamongtheunheatedsampleandtheheatedsamplesafter 72h. ns:nonsignificantlydifferentcontentbetween0and72hwithineachtypeofsample. *:significantlydifferentcontentbetween0and72hwithineachtypeofsample. Table3.Cholesteroloxidationproducts(µg/mgcholesterol)contentoftheunheatedsampleandthefoursamples duringstorageat4 Cfor0and72h. unheated chol chol+mce chol+dha chol+dha+mce 0 h 72 h 0 h 72 h 0 h 72 h 0 h 72 h 7-HC nd 28.90 c 27.78 C ns 17.07 b 9.55 B * 6.47 a 0.18 A * 10.42 ab 0.18 A * 7-HC nd 48.25 c 47.17 C ns 1.84 a 4.74 A * 13.01 b 11.06 B ns 13.25 b 15.04 B ns 5,6-EC nd 61.06 c 63.12 C ns 1.21 a 1.11 A ns 20.78 b 15.29 B * 20.11 b 18.49 B ns 5,6-EC nd 29.22 c 31.81 C ns 0.65 a 0.59 A ns 15.33 b 10.78 B * 12.58 b 10.88 B ns 7-KC nq 59.65 c 68.37 C ns 3.26 a 7.97 A * 31.10 b 25.71 B * 30.78 b 30.64 B ns Total COPs - 227.07 c 238.24 C ns 24.03 a 23.96 A ns 86.69 b 63.02 B * 87.14 b 75.22 B ns Differentlowercaselettersdenotestatisticaldifferencesamongheatedsamplesafter0h.Differentcapitalletters denotestatisticaldifferencesamongheatedsamplesafter72h. nd:notdetected(detectionlimit:7hc=0.98µg/mg,7hc=0.46µg/mg,ec=4.99µg/mg,ec=3.67µg/mg) nq:notquantitated(quantificationlimit:7kc=1.01µg/mg) ns:nonsignificantlydifferentcontentbetween0and72hwithineachtypeofsample. *:significantlydifferentcontentbetween0and72hwithineachtypeofsample 129

ResultsVII Poster2 ProtectiveeffectofaSolanumsessiliflorum(mana cubiu)extractintunapatties