Isolation and characterization of acetylated glucuronoarabinoxylan from sugarcane bagasse and straw

Sugarcane bagasse and straw are generated in large volumes as by-products of agro-industrial production. They are anemerging valuable resource for the generationofhemicellulose-basedmaterials and products, since they contain significant quantities of xylans (often twice as much as in hardwoods).

jo u r n al h om ep a g e :w w w e l s e v i e r c o m / l o c a t e / c a r b p o l

a Pulp and Paper Laboratory, Department of Forestry Engineering, Federal University of Vic¸ osa, Av P H Rolfs, S/N, Campus, 36570-900 Vic¸ osa, Minas

Gerais, Brazil

b Department of Fibre and Polymer Technology, KTH, Royal Institute of Technology, Teknikringen 56-58, SE-100 44 Stockholm, Sweden

c Division of Glycoscience, School of Biotechnology, KTH, Royal Institute of Technology, AlbaNova University Center, SE-106 91 Stockholm, Sweden

d CICECO-Aveiro Institute of Materials, Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal

e Wallenberg Wood Science Center, Department of Fibre and Polymer Technology, KTH, Royal Institute of Technology, SE-100 44 Stockholm, Sweden

a r t i c l e i n f o

Article history:

Received 26 April 2016

Received in revised form 6 September 2016

Accepted 7 September 2016

Available online 9 September 2016

Keywords:

Acetylated xylan

Arabinoxylan

Sugarcane bagasse

Sugarcane straw

Linkage analysis

1 H NMR spectroscopy

a b s t r a c t

Sugarcanebagasseandstrawaregeneratedinlargevolumesasby-productsofagro-industrialproduction Theyareanemergingvaluableresourceforthegenerationofhemicellulose-basedmaterialsandproducts, sincetheycontainsignificantquantitiesofxylans(oftentwiceasmuchasinhardwoods).Heteroxylans (yieldsofca20%basedonxylosecontentinsugarcanebagasseandstraw)weresuccessfullyisolatedand purifiedusingmilddelignificationfollowedbydimethylsulfoxide(DMSO)extraction.Delignification withperaceticacid(PAA)wasmoreefficientthantraditionalsodiumchlorite(NaClO2)delignificationfor xylanextractionfrombothbiomasses,resultinginhigherextractionyieldsandpurity.Wehaveshown thattheheteroxylansisolatedfromsugarcanebagasseandstrawareacetylatedglucuronoarabinoxylans (GAX),withdistinctmolecularstructures.BagasseGAXhadaslightlylowerglycosylsubstitutionmolar ratioofAraftoXylpto(0.5:10)and(4-O-Me)GlpAtoXylp(0.1:10)thanGAXfromstraw(0.8:10and 0.1:10respectively),butahigherdegreeofacetylation(0.33and0.10,respectively).Ahigherfrequency

ofacetylgroupssubstitutionatposition␣-(1→3)(Xyl-3Ac)thanatposition␣-(1→2)(Xyl-2Ac)was confirmedforbothbagasseandstrawGAX,withaminorratioofdiacetylation(Xyl-2,3Ac).Thesizeand molecularweightdistributionsfortheacetylatedGAXextractedfromthesugarcanebagasseandstraw wereanalyzedusingmultiple-detectionsize-exclusionchromatography(SEC-DRI-MALLS).Light scatter-ingdataprovidedabsolutemolarmassvaluesforacetylatedGAXwithhigheraveragevaluesthandid standardcalibration.Moreover,thedatahighlighteddifferencesinthemolarmassdistributionsbetween thetwoisolationmethodsforbothtypesofsugarcaneGAX,whichcanbecorrelatedwiththedifferent Arafandacetylsubstitutionpatterns.Wehavedevelopedanempiricalmodelforthemolecularstructure

ofacetylatedGAXextractedfromsugarcanebagasseandstrawwithPAA/DMSOthroughtheintegration

ofresultsobtainedfromglycosidiclinkageanalysis,1HNMRspectroscopyandacetylquantification.This knowledgeofthestructureofxylansinsugarcanebagasseandstrawwillprovideabetterunderstanding

oftheisolation-structure-propertiesrelationshipofthesebiopolymersand,ultimately,createnew pos-sibilitiesfortheuseofsugarcanexylaninhigh-valueapplications,suchasbiochemicalsandbio-based materials

©2016ElsevierLtd.Allrightsreserved

∗ Corresponding authors at: Wallenberg Wood Science Center, Department of

Fibre and Polymer Technology, KTH, Royal Institute of Technology, SE-100 44

Stock-holm, Sweden.

E-mail addresses: franvila@kth.se (F Vilaplana), olena@kth.se (O Sevastyanova).

1 Introduction

A growing demandfor themore effective utilization of lig-nocellulosic biomass has led to greater interest in the use of agro-industrialresidues,includingsugarcanebagasseand sugar-canestraw(Bian,Peng,Xu,Sun,&Kennedy,2010;Canilhaetal.,

2012;Carvalhoetal.,2015;Pandey,Soccol,Nigam,&Soccol,2000; http://dx.doi.org/10.1016/j.carbpol.2016.09.022

0144-8617/© 2016 Elsevier Ltd All rights reserved.

Sug-arcane,whichisasourceofbothsugar(sucrose)andethanol,is

oneofthemostimportantindustrialcropsinBrazil.Accordingto

theBrazilianSugarcaneAssociation(UNICA,2016),the2015/2016

harvestestimateforSouth-CentralBrazilwillresultinthe

produc-tionof618milliontonsofsugarcanebiomass.Sugarcanebagasse

(stalks)andstraw(tipsandleaves)eachrepresentapproximately

14%oftheplantandaregeneratedinlargeamountsasthemain

agriculturalwastefromthesugarcaneindustry(Conab,2014)

Typ-ically,bagasseisburnttoproducesteam,whilestrawisdepositedat

theharvestingsites.However,beinglignocellulosicresidues,there

isgreatpotentialfortheuseofsugarcanebagasseandstrawfor

theproductionofpulpandsecond-generationethanol,aswellas

fortheirconversionintobio-chemicalsand bio-basedmaterials

(Canilhaetal.,2012;Cardona,Quintero,&Paz,2010;Oliveiraetal.,

2013;Pandeyetal.,2000)

Sugarcanebagasseandstraw,similartootherannualand

peren-nialplants,containlargequantitiesofhemicelluloses–sometimes

upto50%oftheirchemical composition(Carvalhoetal.,2015)

Thishighrelativecontentofhemicelluloses,especiallyxylan,

con-stitutesanexcellentbasisfortheextractionandvalorizationofsuch

hemicellulosicfractions(Ebringerová&Heinze,2000;Ebringerová,

Hromádková,&Heinze,2005).Therehasbeenanincreasing

inter-estintheexploitationofxylanbiopolymersaspotentialresources

forthedevelopmentofnewmaterialsandproducts(Bosmansetal.,

2014;Ebringerová&Heinze,2000;Höije,Sternemalm,Heikkinen,

Tenkanen,&Gatenholm,2008;Littunen etal.,2015;Peng,Ren,

Zhong,&Sun,2011),whichisespeciallytrueforxylansthatare

readilyavailableasby-productsoftheforestandagriculture

indus-tries(Egüésetal.,2014;Svärdetal.,2015).However,thestructural

heterogeneityofxylansisastronglimitingfactor,astheydifferin

termsoftheirchemicalcompositionandstructuralpatternsamong

differentbiomasssources,andevenbetweendifferenttissuesand

developmentalstagesof thesame plant(Ebringerová&Heinze,

2000;Stephen,1983).Inordertogainabetterunderstandingofits

potentialapplication,agreaterknowledgeofxylanstructureand

theisolation-structure-propertiesrelationshipisneeded(Littunen

etal.,2015;Köhnke,Östlund,&Brelid,2011;Mikkelsen,Flanagan,

Wilson,Bacic,&Gidley,2015)

Typically, the backbone in xylan is formed by ␤-(1→

4)-d-xylopyranosyl(Xylp) units, withpossible glycosyl substitutions

inpositionsC-2and/orC-3,andwithacertainnumberofacetyl

groups The main side groups in the xylan backbone are

l-arabinofuranose(Araf), d-glucopyranosyluronicacidunits (GA)

and4-O-methyld-glucuronicacidunits(4-O-MeGlcA).Other

sub-stitutions in xylan can also occur, but they are less abundant

(Ebringerováetal.,2005;Evtuguin,Tomás,Silva,&Neto,2003)

Glucuronoarabinoxylan(GAX)andarabinoxylan(AX)aretypicalfor

grasses,inwhichbranchesofarabinose,GA,4-O-MeGlcAandacetyl

groups(Ac)inthebackboneofxylosecanbeobserved(Ebringerová

&Heinze,2000;Ebringerováetal.,2005)

Themolecularstructure ofxylans in sugarcanebagasse and

strawisexpectedtobesomewhatdifferent,bothfromeachother

andfromthatinotherbiomasses.Ourpreviousworkshowedthat

theamountofxylanfoundinsugarcanebagasseandstrawisat

leasttwicethatfoundin hardwoods cultivatein tropicalareas,

althoughithadalowercontentofuronicacidunitsandacetyl

sub-stitutions(Alvesetal.,2010;Carvalhoetal.,2015).Thexylansin

sugarcanebagasseareconsideredpartiallyacetylated

l-arabino-(4-O-methylglucurono)-d-xylans,whereglucuronicandarabinose

units are linked at O-2 and O-3, respectively, of internal

␤-d-xylopyranosylunitsinthebackbone(Peng,Ren,Xu,Bian,Peng,&

Sun,2009;Shietal.,2012).However,littleinformationonthe

sub-stitutionpatternsofacetylgroupsinsugarcanebagasseandstraw

xylansiscurrentlyavailable,whichlimitstheirpotentialuseinthe

developmentofxylan-basedmaterialsandproducts

Fig 1.Working plan for delignification, xylan isolation and chemical and structural characterization of xylan.

Ouraimwastoinvestigatethedifferencesinthechemical struc-tureofnativeacetylatedxylansfromsugarcanebagasseandstraw

Wehaveisolatedintactacetylatedheteroxylansfromsugarcane bagasseandstrawbyextractingtheperaceticacid(PAA)orsodium chlorite(NaClO2)holocellulosesusingdimethylsulfoxide(DMSO) TheuseofDMSOresultedinpartialxylan extraction,but deliv-eredxylanfractionswithmolecularstructuresthatresembledthe structureofthenativexylan.Theisolatedxylanswerethoroughly characterized using glycosidic linkage analysis, Fourier trans-forminfraredspectrometry(FTIR),1Hnuclearmagneticresonance spectroscopy (1H NMR) and multiple-detection size-exclusion chromatography(SEC)(usingbothstandardanduniversal calibra-tion).Basedonthisinformation,empiricalstructuralformulasfor bothtypesofxylanareproposed.Suchknowledgeisrequiredfor

abetterunderstandingofthechemicalreactivityofthesexylan speciesduringchemicalprocessingandmodificationandforthe creationofnewpossibilitiesfortheuseofsugarcanexylan biopoly-mersinmaterialsandproducts

2 Experimental

2.1 Materials Therawmaterials,5-montholdsugarcane(cultivarRB867515) bagasse(stalksafterfragmentationandpressing)andstraw(leaves andtips),weresuppliedbytheCenterforSugarcane Experimenta-tion(Oratórios,MinasGeraisState,Brazil).Thesugarcanebagasse andstrawwereconvertedtosawdust(<35mesh)byusingaWiley millbenchmodel

Thechemicalsusedwereethanol96%(VWR,France),toluene 99.8%(SigmaAldrich,USA),ethylenediaminetetraceticacid(EDTA) 99%(SigmaAldrich,USA),peraceticacid(PAA)39%(SigmaAldrich, USA),sodiumhydroxide(NaOH)pellets analyticalgrade(Merck Milipore,Germany),acetone99.5%(VWR,France),aceticacid100% (VWR,France),sodiumacetate99%(Merck,USA),sodiumchlorite (NaClO2)80%(AlfaAesar,Germany),dimethylsulfoxide(DMSO) 99% (VWR, France), formic acid 98/100% (VWR, England) and methanolHPLCgrade(FisherChemicals,UK)

2.2 Isolationofacetylatedheteroxylan Theisolationprocedureforacetylatedheteroxylanisdepictedin Fig.1.Inbrief,sugarcanebagasseandstrawwereconvertedto

preparedfromtheextractives-freesawdustbydelignificationwith

NaClO2orPAA.Heteroxylanwasextractedfromtheholocelluloses

byDMSO,precipitatedinethanol,centrifuged,purifiedanddried

Thexylansyieldswereca19–22%and3–4%(basedonxylose

con-tent)fromthose presentedinrawmaterialsforPAA/DMSOand

NaClO2/DMSO isolation procedures correspondently Acetylated

heteroxylanwasanalyzedusing monosaccharideand glycosidic

linkageanalyses,size-exclusionchromatography(SEC),1Hnuclear

magneticresonancespectroscopy(1HNMR)andFouriertransform

infraredspectrometry(FTIR)

2.2.1 Extractivesremoval

Biomass sawdust (<35mesh) was extracted with

ethanol/toluene1:2(v/v)for12hinaSoxhlexextractor(Shatalov,

Evtuguin,&Neto,1999;Sunetal.,2004).Extractives-freesawdust

wasair-driedandstoredinairtightplasticbagsatroom

tempera-turepriortouse.Themoistureoftheextractives-freesawdustwas

determinedaccordingtoTAPPIT264cm-07

2.2.2 Delignification

2.2.2.1 PAAdelignification PriortodelignificationwithPAA,the

extractives-freebagasseandstrawsawdustwastreatedwith0.2%

(w/v) EDTA at90◦C for 1h, withconstant stirring,in order to

remove the metal ions and prevent decomposition of the PAA

(Brienzo, Siqueira,& Milagres, 2009) Thedelignification ofthe

biomasseswasperformedwith10gofextractives-freesawdust

treatedwith500mLof 10%(v/v)PAA at pH3.5(adjusted with

sodiumhydroxidesolution),at85◦Cfor30min,withconstant

stir-ring.Afterthetreatment,thesolutionwascooledinanicebath

anddilutedtwicewithwater.Theholocellulosewascollectedona

porousglassfilterP2(porosity100),washedwith5Lofwarm

dis-tilledwaterand,soonthereafter,with50mLofacetone/ethanol1:1

(v/v)(Evtuguinetal.,2003).Theholocellulosewasdriedatroom

temperature(24◦C)andstoredinairtightcontainers

2.2.2.2 Sodiumchloritedelignification.10gofextractives-free

saw-dustwastreatedwith388mLofwater,15mLofaceticacid100%,

72mLofsodiumacetate30%(w/v)and55mLofsodiumchlorite

30%(w/v), at75◦C for 30min,withconstant stirring.Afterthe

treatment,theholocellulosewascollectedonapolystyrene

mem-brane(porosity60␮m),washedwith5Lof distilledwater and,

soonthereafter,with100mLofacetone(Magaton,Piló-Veloso,&

Colodette,2008).Theholocellulosewasdriedatroomtemperature

(24◦C)andstoredinairtightcontainers

2.2.3 Isolationofxylans

Asampleof6gofholocellulose(PAA-holocelluloseorNaClO2

-holocellulose)wastreatedwith130mLofDMSO,at24◦Cfor24h,

undernitrogenatmosphereandwithconstantstirring(Hägglund,

Lindberg,&McPherson,1956).Afterthetreatment,thesuspension

wasfilteredthroughapolystyrenemembrane(porosity60␮m)and

washedwith∼20mLofdistilledwater.Thesupernatantliquidwas

addedto600mLofethanolatpH3.5(adjustedwithformicacid)and

leftfor12hat4◦C(Magatonetal.,2008).Theprecipitated

hemi-celluloseswereisolatedbycentrifugation(10minat4500rpm)and

washed5timeswithmethanol(Evtuguinetal.,2003).Thexylans

weredriedinavacuumovenfor72hat30◦C

The total yield was estimated gravimetrically based onthe

amountofstartingmaterial(extractives-freebiomasses).Thexylan

yieldsafterDMSOisolationwerecalculatedtakinginto

consider-ationtheamountofxyloseintheextractives-freebiomassesand

thexylansamples,aswellasthetotalyield

2.3 Compositionalandstructuralcharacterization 2.3.1 Monosaccharidecompositionanalysis The monosaccharides composition was determined by acid hydrolysisfollowed bychromatographicanalysis.Total polysac-charide depolymerization of the sugarcane fractions, including glucosefromcrystallinecellulose,wasachievedbysulfuric hydrol-ysis.Samples(4mg)werekeptinaglasstubewith0.25mLof72% sulfuricacidfor3hatroomtemperature.Then,deionizedwater wasaddedtodilute thesolutiontoapprox.1,2-1,3molL−1 sul-furicacid,andthetubeswereincubatedat100◦Cfor3h.Uronic acidsinhydrolysatesweredeterminedcolorimetricallywith m-phenylphenolusingaknownprocedurepublishedbySelvendran, Verne, and Faulks (1989) The monosaccharide composition of theisolatedheteroxylanfractionscontainingmorelabile uronic acids(glucuronicandgalacturonicacid)wasdeterminedbyacidic methanolysis(Appeldoorn,Kabel,VanEylen,Gruppen,&Schols,

2010;Bertaud,Sundber,&Holmbom,2002).Freeze-driedsamples (1mg)wereincubatedwith1mLof2molL−1HClindrymethanol for5hat100◦C.Subsequently,thesampleswereneutralizedwith pyridine,driedunderastreamofair,andfurtherhydrolyzedwith

2molL−1trifluoroaceticacid(TFA)at121◦Cfor1h.Thesamples wereagaindriedunderastreamofairanddissolvedinH2O Thehydrolysed monosaccharidesby both sulfurichydrolysis andacidicmethanolysis,wereanalyzedusinghighperformance anionexchangechromatographywithpulsedamperometric detec-tion(HPAEC-PAD) usinganICS-3000system(Dionex)equipped withaCarboPacPA1column(4×250mm,Dionex).Inositolwas addedtoallsamplesasaninternalstandardpriortohydrolysis Allexperimentswereperformedintriplicate.Twodifferent gra-dientswereusedfortheseparationandquantificationofneutral (Fuc,Ara,Rha,Gal,Glc,Man,andXyl)andacidic(GalA,GlcA,and 4-O-MeGlcA)monosaccharides,asreportedinourpreviousstudy (McKeeetal.,2016).Quantificationof4-O-MeGlcAwasperformed usingtheanalyticalresponsefactorforGlcAwiththeappropriate correctionasreportedinChongetal.(2013)

2.3.2 Klasonlignin The Klason lignin content in theholocellulose samples was determined gravimetrically to be insoluble residue after acid hydrolysiswith72%sulfuricacid,accordingtotheTAPPI222om-02 standardmethod(TAPPI,2011)

2.3.3 Acetylcontentanddegreeofacetylation The acetyl content of the xylan samples was determined after alkaline hydrolysis with NaOH at 70◦C overnight using highperformanceliquidchromatography(HPLC)withUV detec-tion (Voragen, Schols, & Pilnik, 1986) The HPLC instrument (Dionex–Thermofisher,CA,USA)wasequippedwithaUVdetector (Rezex,210nm)anda ROA-Organicacidcolumn(300×7.8mm; Phenomenex,Torrance,CA,USA).Separationswereperformedat

aflowrateof0.5mLmin−1,using2.5mmolL−1H2SO4asamobile phaseat50◦C.Thedegreeofacetylation(DA)wasdeterminedfrom theacetylcontentinthexylansamplesaccordingtoEq.(1)

DA= 132×%acetyl

Macetyl×100

−

Macetyl−1

where:DAisthedegreeofacetylation,%acetylistheacetyl con-tentdeterminedbyanalysis,Macetylistheacetylmolecularweight (43gmol−1)and132gmol−1isthemolecularweightof anhydrox-ylopyranose(Xuetal.,2010)

2.3.4 Glycosidiclinkageanalysis Freeze-driedxylanfractions(1mg,threetechnicalreplicates) wereswelledinanhydrousDMSOfor16hat60◦Candmethylated

com-pletemethylation(CiucanuandKerek,1984).Thesampleswere

thenhydrolyzedwith2molL−1TFAat121◦Cfor2h,reducedwith

sodiumborohydride (NaBH4) and acetylated withacetic

anhy-dride in pyridine (Albersheim, Nevins, English, & Karr, 1967)

Theobtainedpermethylatedalditolacetates(PMAAs)were

sepa-ratedandanalyzedusingagaschromatographer(HP-6890,Agilent

Technologies)coupledtoanelectron-impactmassspectrometer

(HP-5973,AgilentTechnologies) ona SP-2380 capillary column

(30m×0.25mm i.d.; Sigma–Aldrich) with a temperature range

increasingfrom160◦Cto210◦Catarateof1◦C/minusingHeas

car-riergas.Theretentiontimesandfragmentationmassspectrafrom

thePMAAswerecomparedwiththoseofthereference

polysaccha-rides(wheatarabinoxylan,Megazyme,Ireland)andwithavailable

data(Carpita&Shea,1989).Thequantificationwasbasedonthe

carbohydratecompositionandtherelativemolar-responsefactor

ofeachcompound(Carpita&Shea,1989),asdetectedbyGC–MS

2.3.5 Size-exclusionchromatography

Themolarmassdistributionsofthexylanextractsfromthe

sug-arcanebagasse andstraw wereanalyzedusing asize-exclusion

chromatographer (SECcurity 1260, Polymer Standard Services,

Mainz,Germany)coupledinseriestoamultiple-anglelaserlight

scattering detector (MALLS; BIC-MwA7000, Brookhaven

Instru-mentCorp.,NewYork)andarefractiveindexdetector(SECcurity

1260,PolymerStandardServices,Mainz,Germany).SECanalyses

wereperformedataflowrateof0.5mLmin−1usingdimethyl

sul-foxide(DMSO,HPLCgrade,Sigma-Aldrich,Sweden)with0.5%w/w

LiBr(ReagentPlus)asamobilephase,usingacolumnsetconsisting

ofaGRAMPreColumn,100and10000analyticalcolumns

(Poly-merStandardsServices,Mainz,Germany)thermostattedat60◦C

Priortothe analyses, thexylanswere dissolveddirectlyin the

SECeluentfor16hat60◦C.Standardcalibrationwasperformed

by theinjection of pullulanstandards of known molar masses

providedbyPolymerStandardsServices(PSS,Mainz, Germany)

TheSECelutionvolumeswerethenconvertedintohydrodynamic

volumes(Vh)usingtheMark-Houwinkequation,asreportedby

VilaplanaandGilbert(2010).TheMark-Houwinkparametersfor

pullulanin DMSO/LiBr (0.5wt%) are K=2.427×10-4dLg−1 and

a=0.6804(KramerandKilz,PSS,Mainz,privatecommunication)

Thedifferential refractive index increment (dn/dc)for pullulan

inDMSO/LiBr0.5%wasconsideredas0.0853mLg−1(Kramerand

Kilz,PSS,Mainz,privatecommunication).Datacollectionandlight

scatteringcalibrationwasperformedusingtheWinGPCsoftware

(PolymerStandardsServices,Mainz,Germany),withadn/dcvalue

of 0.0881mLg−1 (calculated for sugarcane xylan in DMSO/LiBr

0.5%;SupplementarymaterialFig.S3).TheSECweight

distribu-tion,w(logVh), and thesize dependenceof theweight-average

molecularweight, ¯MW(Vh),werecalculatedusingadditional

math-ematical procedures presented elsewhere (Vilaplana & Gilbert,

2010) The macromolecular size distributions are presented in

termsofhydrodynamicradius(Rh),withVh=4··R3

h.Four

dif-ferentxylanconcentrationsbetween1.0–4.0gL−1wereinjectedfor

eachsample,whichresultedinfourreplicatesforthecalculation

ofthenumber-average( ¯Mn)andweigh-averagemolarmass( ¯MW)

Themolarmassdistributionsandtheaveragemolarmassvalues

forthexylanextractsfromsugarcanebagassehavebeenvalidated

andcomparedwiththeinjectionofareferencearabinoxylan(AX)

fromwheatendosperm(Megazyme,Ireland)

2.3.6 1HNMRspectroscopy

1HNMRspectrawereregisteredonaBrukerAVANCE300

spec-trometeroperatingat300.13MHzat298K.Thexylanwasdissolved

inD2O(ca2%w/w)andthesodium

3-(trimethylsilyl)propionate-d (TMSP,ı0.00)wasusedasinternalstandard.Theacquisition

parametersfortheprotonspectrawereasfollows:12.2␮spulse width(90◦),18srelaxationdelay,and300scanswerecollected These conditions guarantee that quantitative information was obtainedfromtheNMRmeasurements

2.3.7 Fouriertransforminfraredspectrometry The Fourier transform infrared (FTIR) spectra (wavelength 4000–600cm−1)were recorded usinga Perkin-Elmer Spectrum

2000FTIRspectrometer(Waltham,MA,USA) equippedwithan attenuatedtotal reflectance(ATR)system(Spectac MKIIGolden GateCreecstoneRidge,GA,USA).Thespectrawereobtainedfrom drysamplesusing16scansataresolutionof4cm−1andat inter-valsof1cm−1atroomtemperature.Origin9.1softwarewasused forthespectraevaluation

3 Results and discussions

3.1 Yieldandchemicalcompositionofisolatedxylan

Inourpreviouswork,sugarcanebagasseandstrawwere chemi-callycharacterized(Carvalhoetal.,2015).Theresultsthereof,based

onthecompletemassbalance,aresetoutintheSupplementary materialFig.S1.Bagasseandstrawwereshowntocontain signif-icantquantitiesofhemicellulosesandpectins,expressedasother sugars(thesumofxylose,galactose,mannose,arabinose,uronic acids andacetyl groups) Thehighrelative content of hemicel-lulosesintheserawmaterialsconstitutes asoundbasis forthe extractionandvalorizationofsuchfractionsfromsugarcane

In thepresent work,delignificationmethods using PAA and NaClO2wereusedpriortotheextractionofxylanwithDMSO.Using thePAAdelignificationprocess,thelossofdrymatter(including lignin,ashand,toalesserextent,polysaccharides)fromthe ini-tialamountofextractives-freebiomasseswas24.8%and25.7%for bagasseandstraw,respectively.Thelossofdrymatterusingthe NaClO2 delignificationprocess was15.4% and27.7%for bagasse and straw, respectively The heteroxylans yield (based on the xylosecontentinsugarcanebagasseandstraw)wasca19–22%for PAA/DMSOisolationprocedure,whileonlyca3–4%ofxylanwas extractedfromNaClO2hollocellulose(Table1)

Inapreviousstudy,almost5timeshigherxylanyieldswere alsoobservedforeucalyptuswhenusingthePAA/DMSOextraction process,incomparisonwiththeNaClO2/DMSOprocess(Evtuguin

etal.,2003).Such resultwasexplainedbythehigherdegreeof delignificationwithperaceticacidandthesimultaneousbreaking

oflignin-xylanetherbonds

TheFTIRspectraobtainedforxylanextractedfrombagasseand strawweretypicalforxylans,asshowninFig.2:asharpbandat

1039cm−1,whichisduetotheC O,C CstretchingorC OH bend-ingin thesugarunits(Chaikumpollert,Methacanon, &Suchiva,

2004)andtheexpectedbandsbetween1175and1000cm−1(Sun

etal.,2004).The␤-glycosidiclinkagesbetweenthexyloseunits wereevidencedbythepresenceofasharpbandat897cm−1(Gupta, Madan,&Bansal,1987).Thebandat3350–3330cm−1corresponds

tothehydroxylstretchingvibrationsofxylans,aswellasthewater involved in the hydrogenbonding, and the bandat 2920cm−1 representsC Hstretchingvibrations(Sunetal.,2004).Theband

at1160cm−1indicatesthepresenceofarabinoseresidues(Egüés

etal.,2014).Thepresenceofacetylgroupsinthexylanwas con-firmedbytheabsorptionat1734cm−1,whichisduetotheC O stretching(Bian et al.,2010).The bandat 1241cm−1, which is alsoduetotheC Ostretching,andthebandat1370cm−1,which

isdue totheC CH3 stretching,werealsoconfirmed(Xuetal.,

2010).Theabsorptionat1628cm−1isattributedprincipallytothe waterabsorbedbyxylans(Kaˇcuráková,Belton,Wilson,Hirsch,& Ebringerová,1998).Theweakbandat1510cm−1,whichisdueto

Table 1

Isolation yields for the hollocellulose, DMSO extracts and xylan (based on the xylose content in the sugarcane bagasse and straw).

Fig 2.FTIR spectra for xylan samples extracted from bagasse by PAA/DMSO

pro-cess (spectrum a), from bagasse by NaClO 2 /DMSO process from (spectrum b), from

straw by PAA/DMSO process (spectrum c) andfrom straw by NaClO 2 /DMSO process

(spectrumd).

thearomaticskeletalvibration,indicatesthepresenceofasmall

amountoflignininthexylansamples(Sunetal.,2004)Duringthe

isolationprocedures,theglycosidiclinkagescanbedisruptedand

thehydroxylgroupscanbeoxidized,resultingintheformationof

ketonecarbonylgroups,seenasbandsataround1720cm−1inFTIR

spectra(Magatonetal.,2008;Sun&Tomkinson,2002).Theabsence

ofsuchsignalsinthespectraweobservedinthepresentworkfor

xylansextractedfrombagasseandstrawrulesoutanyoxidation

reactionsduringthedelignificationandisolationprocedures

Theevolutionofthexylanisolationprocesswasmonitoredby

notingthemonosaccharidecompositionoftheextractives-freeraw

material,thecontentofhollocelluloses(afterdelignificationwith

bothPAA andNaClO2)andthecomposition oftheDMSOxylan

extracts(Fig.3).Ascanbeobserved,xyloseisthemain

compo-nentintheDMSOextractsforbothinvestigatedmaterials,which

evidencesthesuccessfulisolationofthexylanusingtheproposed

procedures.Theratiobetweenglucoseandxyloseisquitesimilar

fortheextractives-freebiomassesandbothtypesofholocelluloses

(delignifiedbyPAAorNaClO2).However,thecompositionofthe

DMSOextractsisolatedfromvarioustypesofholocellulosesisquite

different.ThetotalamountofuronicmoietiesinthePAA/DMSO

samplesweredeterminedcalorimetricallyinsugarshydrolysates

afterSaemanhydrolysis(Saeman,1945)andprovedtobefairly

similar(1.4%inbagasseand1.8%instraw).Thesugaranalysisof

thePAA/DMSOsamplesconfirmedthepresenceof a significant

amountofarabinoseand 4-O-methylglucuronicacid(MeGlcA),

inaddition tothexylose, inthebagasse and straw, whichwas

alsoseen in theFTIRspectra (Fig.2).Thus heteroxylans inthe

sugarcanebagasseandstrawweretypicalglucuronoarabinoxylans

(GAX).Glucose,galactoseandgalacturonicacidwerealsopresent

inthePAA/DMSOsamples,althoughinmuchsmalleramounts.This

canbeattributedtotheminorpresenceofmixed-linkage␤-glucans

andpectinpolysaccharidecomponents,whichisconfirmedbythe

resultsoftheglycosidiclinkageanalysis

The NaClO2/DMSO-extracted sampleshad lower xylose con-tentthanthoseobtainedafterthePAA/DMSOisolationprocedure ThecompositionoftheNaClO2/DMSO-extractedsamples,together withthelowyieldsobtained,suggestthatthisisolationprocess wasless efficient and less selectivetowards xylan Arelatively highcontentofglucose,andaconsiderableincreaseintheamount

ofgalactose,werefoundintheNaClO2/DMSO-extractedsamples Thisfurtherindicatesthattheisolationofnon-xylan hemicellu-loses,suchasmixed-linkage␤-glucans,andotherpecticmaterial occurredduringtheNaClO2/DMSOprocess.Mostlikely,the acces-sibilityofDMSOinthecellwallswasverylow,duetoinsufficient delignification,andamixtureofsuchpolysaccharideswasremoved onlyfromthefibersurface

3.2 Glycosidiclinkageanalysisinisolatedxylans

Inordertoinvestigatethesubstitutionpatternsoftheisolated xylanfractions,linkage(methylation)analysiswasperformedon the PAA/DMSOand NaClO2/DMSO extracts (Table2).From the results of linkage analysis, it can be deducedthat the general structureofthexylansfrombothbagasseandstrawconsistsofa linearbackboneof(1→4)-linked␤-d-xylopyranosylunits(Xylp), partially O-3 substituted with l-arabinofuranosyl (Araf) units andO-2substitutedessentiallywith4-O-methyl-d-glucuronosyl units (MeGlcpA) 4-O-methyl-d-glucuronic acid residues were previously detected in sugarcane bagasse and straw by acidic methanolysis(Carvalhoetal.,2015).Itisworthmentioningthat glycosidiclinkageanalysisisunabletoidentifytheXylpunits mod-ifiedbyacetylation,duetotheextremealkalineconditionsapplied duringthemethylationofthesamples.Therelativeratioof 3,4-Xylpversus2,4-Xylpresiduesinthelinkageanalysisevidencedthat thesubstitutionatO-3was3–5foldhigherthansubstitutionsat O-2in thexylan backbone,which as wellmatches therelative amountsofterminalAraf(t-Araf)andMeGlcA,respectively.Only tracesofdoublesubstituted2,3,4-Xylpresidueswerefoundinthe PAA andNaClO2 treatedxylans,in contrasttoreports onother grassxylans,suchascereals,wherethesedoublesubstitutionsare relativelyabundant(Heikkinenetal.,2013).TheamountsofGAX presentin thedifferentextractscanbecalculated directlyfrom therelativeabundanceofthelinkagesinvolvedinthestructure, i.e., t-Araf, t-Xylp, 4-Xylp, 2-Xylp, 2,4-Xylp 3,4-Xylp, and 2,3,4-Xylp These resultsconfirm once again thepurity of the xylan extractsobtainedafterthePAA/DMSOtreatment,comparedwith thoseobtainedaftertheNaClO2/DMSOprocedure.Indeed,when thetwoisolationtreatmentswerecompared,itwasevidentthatthe chlorite-extractedxylanscontainedasignificantamountofother contaminatingpolysaccharides,consistentwiththesugar compo-sitionresults.TheGAXcontentwasapproximately85–95%inthe PAA/DMSOextractedsamples,whereas 75–85%GAX puritywas obtainedusingtheNaClO2/DMSOprocedure Thepresenceof 5-Araf,2-Araf,3-Arafand2,5-Arafunits,aswellas3-Galpandother branchedGalpunits,pointedtothepresenceofarabinanand ara-binogalactan,mainlyintheNaClO2/DMSOextracts.Theextraction

ofmixed-linkage(1→3),(1→4)-␤-glucaninboththestrawand bagassesamplesisevidencedbythepresenceof3-Glcpand4-Glcp Thesedifferencesinpurityaremorepronouncedinthestrawxylans thaninthebagassesamples

Fig 3.Sugar composition of bagasse (A) and straw (B) for extractives-free biomasses, PAA-holocellulose, NaClO 2 -holocellulose and xylan isolated by PAA/DMSO and NaClO 2 /DMSO.

Table 2

Monosaccharide composition and glycosidic linkage analysis of the xylan extracted from sugarcane bagasse and straw.

Notes: 1 Relative abundance (%mol) of the different linkages corrected by the results of monosaccharide composition by acid methanolysis; 2 Monosaccharide composition as calculated by acid methanolysis; 3 Total GAX content calculated from the specific linkages (t-Xylp + 4-Xylp + 2-Xylp + 2 × 2,4-Xylp +2 × 3,4-Xylp + 3 × 2,3,4-Xylp); 4 Araf/Xylp ratio calculated based on the monosaccharide composition; 5 (4-O-Me)GlcpA/Xylp ratio calculated based on the monosaccharide composition; 6 t:bp ratio for GAX calculated based on the amount of terminal sugars (t-Araf and 4-O-MeGlcA) versus the relative amount of branching points (2,4-Xylp and 3,4-Xylp); 7 u:m(Xylp) ratio for GAX based on the ratio of unsubstituted Xylp units (4-Xylp) and mono-substituted Xylp units (2,4-Xylp and 3,4-Xylp) Standard deviation ( .); n.d: not determined.

Thelinkagepatternsobservedinthestrawandbagassexylans

werequitesimilar,withthestrawxylanhavingaslightlyhigher

degreeofArafbranching(Araf/Xylp)thanthebagasse xylan.To

thebestofourknowledge,thisisthefirsttimethatlinkage

anal-ysishas been performedon PAA-extracted sugarcanestraw or

bagasse.Previouslyreportedlinkageanalysisonalkali-extracted

xylanfromsugarcanebagassealsoshowedtheabsenceofdouble

substitutionsinthistypeofxylan,butindicateda2–3timesgreater

degreeofarabinosylbranchingatO-3(Banerjee,Pranovich,Dax,& Willfor,2014;Mellinger-Silvaetal.,2011).Thesedifferencesmight

berelatedeithertodeacetylationinalkalineconditions,whichmay haveenhanced theextractionof largerGAX populationswitha higherAraf/Xylpratio,ortothehighersolubilityofacetylatedxylan fractionsinDMSO,whichmayhaveresultedinasmallerdegree

ofarabinosylbranching.TheamountofMeGlcAsubstitutionsas

sugarcanebagasseandstrawxylans

UsingthePAA/DMSOprocess,theformallydeterminedmolar

ratiobetween terminalresidues (t-Araf) andbranch units

(2,4-Xylp, 3,4-Xylp and 2,3,4-Xylp), hereinafter referred to as t:bp,

was 1.14for bagasse and 1.05 for straw, which indicates that

nounder-methylationoccurredduringtheanalysis.Ontheother

hand,theNaClO2/DMSOprocessrevealedlessconformityinthe

t:bpratioforbagasse(1.05)andstraw(0.80).Thislowerratioin

theNaClO2/DMSOxylansamplesmightindicateinsufficient

delig-nificationduringthechloriteprocess,which willunderestimate

theamountofterminalunitsnotreflectedinthelinkageanalysis

(Jeffries,1991)

3.3 Characterizationofthestructureofacetylatedxylanby1H

NMR

1HNMRanalysiswasusedtoconfirmtheintramolecular

struc-ture of xylan obtained by linkageanalysis and toidentify and

quantify the position of acetyl groups in the xylan backbone

The PAA/DMSOextracted xylans from bagasse and straw were

selected for this more detailed structural investigation due to

theirhigheryieldand higherpurity Theextractedxylanswere

completely(bagasse)or almostcompletely (straw)solubleD2O

usedforNMRmeasurements.The1HNMRspectraofxylan

sam-plesareshowninFig.4.Theassignmentsforprotonresonances

weredoneusingexistingdatabasesforglucuronoxylans(Cavagna,

Deger,&Puls,1984;Evtuguinetal.,2003;Hoffmann,Kamerling,&

Vliegenthart,1992;IzydorczykandBiliaderis,1992Magatonetal.,

2008;Marques,Gutiérres,delRío,&Evtuguin,2010;Sun,Cui,Gu,

&Zhang,2011)andarepresentedinTable3.The(1→4)-linked

␤-d-xylopyranosyl internal units of the backbone were clearly

identified.Theintegralofprotonsfromacetylgroups(CH3 CO-)

(␦H2.00–2.25)showedgoodbalancetoamountsofacetylsin

acety-latedxylopyranoseunitsidentifiedusingcharacteristicprotonsin

correspondingstructures.TheseacetylatedstructureswereXylp

unitsacetylatedatO-3(anomericprotonintegralat␦H4.54–4.62),

atO-2(anomericprotonandH-2integralsat␦H4.54–4.62)or

2,3-di-O-acetylated(H-3integralat␦H5.11–5.15).TheXylpunitsO-3

acetylatedandO-2substitutedbyMeGlcAresidueshavebeenalso

detected(H-3integralat␦H5.05–5.08).MostpartofterminalAraf

unitswere␣-(1→3)-linkedtomono-substitutedXylpunitsas

fol-lowedfromthecharacteristic anomericprotonresonanceat␦H

5.39(Izydorczyk&Biliaderis,1992).Thisfact,however,notexclude

completely the possibility for mono-substituted ␣-(1→2)-Araf

anddi-substitutedXylpunitswith␣-(1→3)and␣-(1→2)-linked

Arafunitsfoundpreviouslyinwheatarabinoxylan(Izydorczyk&

Biliaderis, 1992).However,theresultsof theglycosidic linkage

analysis(Table2)showedthatthesedi-substitutedwithArafXylp

unitsbranchesoccurredinaverylowfrequencyinxylanfromboth

bagasseandstraw

Thepresenceof␣-(1→2)-linkedMeGlcpAtoXylpunitswas

confirmedbythepresenceofanomericprotonsincorresponding

structuresat␦H5.28–5.29andverynarrowsignalatca␦H 3.45

assignedtoprotonsinmethoxylgroupsofMeGlcpAinheteroxylans

(Shatalovetal.,1999;Shietal.,2012)

3.4 Quantificationofcontentandpositionofacetylgroupsin

sugarcaneacetylatedheteroxylans

Thequantificationofthetotalnumberofreleasedacetylgroups

bysaponificationandsubsequentHPLCanalysis,togetherwiththe

assignmentofthepositionsoftheacetylgroupsby1HNMR,allows

forthefullcharacterizationoftheacetylcontent(in%),degreeof

acetylation(DAc)andacetylationpatternintheisolatedacetylated

arabinoxylans fromsugarcanebagasseand straw (Table4).The

acetylcontent inthePAA/DMSO-extracted xylansfrombagasse and straw accountsfor8.7% and 2.4%,respectively, whichis in agreementwiththeacetylcontentsinextractives-freebiomasses reportedinourpreviouswork(Carvalhoetal.,2015).Theacetyl contentinstrawissignificantlylowerthaninbagasse,which evi-dencesthedivergentacetylationpatternindifferentplanttissues, depending on their developmental stage and function (Gille & Pauly,2012).TheacetylcontentandDAcaremarkedlylowerfor thesamplesextractedaftertheNaClO2/DMSOprocess,whichagain indicatesthelowerefficiency,andlowerselectivitytowards acety-latedGAX,ofthisprocess

3.5 Averagemolecularweightandmolecularweightdistribution The size and molecular weight distributions for the xylans extracted from the sugarcane bagasse and straw were ana-lyzed using multiple-detection size-exclusion chromatography Themolarmass averagevaluesand sizedistributions, obtained

bybothstandardcalibration(usingpullulanasalinearcalibrant) andabsolutecalibration(bylightscattering),werecomparedfor

4replicatesatdifferentconcentrationsbetween1.0–4.0gL−1.The number-andweight-averagemolarmassesoftheextractedxylan samplesareshowninTable5.Thereproducibilityofthefour repli-catesisverygoodasdemonstratedbythelowstandarddeviation showninTable5,whichreinforcestheaccuracyoftheprocedure TheSECweightdistribution,w(logVh),andthesizedependence

oftheweight-averagemolecularweight ¯MW(Vh), presentedasa functionofthehydrodynamicradius(Rh),isshowninFig.5.The procedureemployedforthecalibrationofthehydrodynamicsize fromtheSEC,usingtheMark-Houwinkequation,ispresentedinthe SupplementarymaterialFig.S2,whiletherawSECchromatograms obtainedwithadifferentialrefractiveindexandlight scattering (at90◦)arepresentedintheSupplementarymaterialFig.S4.The procedurewasvalidatedforareferencewheatendosperm arabi-noxylan(AX),whichexhibitedsimilarmolarmassdistributionsand averagemolarmassvaluesasthosereportedbyprevious publica-tionsinaDMSO/LiBrsolventsystem(Pitkänenetal.,2009;Shelat, Vilaplana,Nicholson,&Gibert,2010).Themolecularsolubilityof thexylansamplesintheSECeluent(DMSO/LiBr)canbeverifiedby theabsenceofaggregationspikesinthelightscatteringsignalat90◦ (SupplementarymaterialFig.S3).Aggregationofarabinoxylanhas beenreportedinaqueousconditions(Pitkänenetal.,2009,2011); however,theuseofapolarorganicsolventsuchasDMSOwith lithiumsaltsashydrogen-bonddisruptors(such asLiBr)should contributetoanenhancedsolubilityandtheabsenceofaggregation phenomenaduringSECseparations,whichwouldbiasthemolar massdeterminations(Shelatetal.,2010;Vilaplana&Gilbert,2010) TheSECweightdistributions,w(logVh),forthedifferentsugarcane samplesexhibitmonomodalsizeprofiles.Thisconfirmsthe homo-geneousmacromolecularpopulationsinsuchextracts,whichare relatedmostlytoacetylatedGAXmacromolecules.Theamountof otherpolysaccharideimpurities(mainlymixed-linkage␤-glucans and arabinogalactans,asreportedbyglycosidiclinkageanalysis

inTable2)shouldnotinfluencethemolarmassdistributionsand averagemolarmassvalues

Theaveragemolarmassvaluesfortheextractedxylansamples that wereobtained usingstandard and light scattering calibra-tionshowinterestingandmarkeddifferences.Theaveragemolar massvaluesforthesugarcanebagasseandstrawxylansthatwere obtainedusingastandardcalibrationareintherangeof26–40kDa, whicharequitesimilartothevaluespreviouslyobtainedfor sug-arcane bagasse xylans(Sunet al., 2004; Peng etal., 2009)and forstrawxylansfromothergrasses(e.g.,wheatstraw)(Persson, Ren,Joelsson,&Jönsson,2009).However,thelightscatteringdata providedmolarmassvaluesforthesugarcanebagasseandstraw thatarelargerthanthose obtainedfromastandard calibration,

Fig 4.1 H NMR spectra of acetylated xylan from bagasse and straw extracted by PAA/DMSO process Designation for the structural fragments is presented in Table 3 *solvent impurities.

Table 3

1 H NMR chemical shifts for structural units of acetylated xylan from bagasse and straw obtained by PAA/DMSO process.

n.d Not detected or non-existent The designations used were as follows: Xyl (isol.) is non-acetylated Xylp in the backbone isolated from other acetylated Xylp units; Xyl (Xyl-Ac) denotes non-acetylated Xylp linked with neighboring acetylated Xylp; Xyl-3Ac corresponds with 3-O-acetylated Xylp; Xyl-2Ac is 2-O-acetylated Xylp; Xyl-2,3Ac denotes 2,3-di-O-acetylated Xylp; Xyl-3Ac-2GlcA, MeGlcA 2-O-linked and 3-O-acetylated Xylp; ␣Ara-3Xyl corresponds with terminal arabinose linked to O-3 of monosubstituted xylose; MeGlcA denotes terminal MeGlcA residue linked to O-2 in monosubstituted Xylp and also in 3-O-acetylated Xylp.

andalsoindicatedsomedifferencesbetweenthetwoextraction

methods.Standardcalibrationprovidesrelativemolarmasseswith

respecttoalinearmacromolecularstandard(inourcase,

pullu-lan).DuetothefactthatSECseparatesmacromoleculesbasedon

sizeorhydrodynamicvolume,Vh(Jonesetal.,2009)(which,for

SEC,is knowntobeproportional totheproductoftheintrinsic

viscosityandthenumber-averagemolarmass,inaccordancewith

theuniversalcalibrationtheoryestablishedbyHamielec,Ouano,

andNebenzahl(1978)andKostanski,Keller,andHamielec(2004),

andnotmolarmass,theaveragemolarmassvaluesand distribu-tionsobtainedbystandardcalibrationusingalinearpolymerwill underestimatetheabsolutevaluesforasubstituted/branched poly-mer.DuringSECelution,linearmacromoleculeswillhaveamore elongatedhydrodynamicconformationandwilleluteearlierthan macromoleculeswithasimilarmolarmassbutmoresubstituted andcompactstructure.Insuchcases,itisnecessarytouselight scat-teringdetectiontoobtainabsolutemolarmassdeterminations.The averagemolarmassvaluesobtainedbythelightscatteringmethod,

Table 4

Acetyl content, degree of acetylation (D Ac ) and acetylation pattern in sugarcane xylans.

a The acetyl content was obtained by saponification and HPLC analysis of the released acetic acid.

b The degree of acetylation was calculated according to Xu et al (2010)

c The relative acetylation pattern (%) was calculated by the integration of the corresponding 1 H NMR peak areas for the Xyl-2Ac, Xyl-3Ac and Xyl-2,3Ac signals Standard deviation (); n.d: not determined.

Table 5

Number-average molar mass ( ¯ M n ), weight-average molar mass ( ¯ M W ) and dispersity (D) for the extracted xylan samples from sugarcane bagasse and straw using standard and light scattering calibration obtained from SEC-DRI-MALLS The average molar mass values are compared with a reference AX from wheat endosperm.

¯

Standard deviation ( .).

Fig 5. SEC weight distribution, w(log V h ), and the size dependence of the weight-average molecular weight, ¯ M W (V h ) as a function of the hydrodynamic radius (R h ) obtained after SEC-DRI-MALLS: (a) sugarcane bagasse GAX, (b) sugarcane straw GAX and (c) wheat endosperm AX.

therefore, providemore accurateinformation onthe molecular

propertiesofxylansisolatedfromsugarcanebagasseandstraw.The

SECweightdistributionsw(logVh)fortheGAXsamplesextracted

withPAA/DMSOfrom bagasse and strawshow similarprofiles,

beingthedistributionsforbagasseslightlyshiftedtowardslarger

macromolecularsizes.Thesizedependenceoftheabsolutemolar

massfromlightscattering ¯MW(Vh)shows similarlylargervalues

forbagassethanstraw,whichcorrespondswiththereportedSEC

distributions(Fig.5).Thisbehavioragreesaswellwiththelarger

averagemolarmassvaluesforthesugarcanebagassexylans

com-paredtostrawusingbothstandardandlightscatteringcalibration

(Table5).Thesesmalldifferencesinmolarmassbetweenbagasse

andstrawcanbeattributedtothelargerdegreeof

polymeriza-tioninthexylanbackbone,in agreementwiththeresultsfrom

linkageanalysis.Asimilartrendcanbeobservedusingboth

stan-dardcalibrationandlightscatteringdataforthesamplesisolated

byNaClO2/DMSOextraction,wherethesugarcanebagassexylans

exhibit higherabsolutemolar massvalues than xylansisolated

fromsugarcanestraw

However, largerdifferences in the averagemolar mass

val-ues and the size dependence of the weight-average molecular

weight ¯MW(Vh)canbeobservedforbothsugarcanebagasseand

strawwhencomparingthesamplesextractedbythePAA/DMSO methodwiththoseextractedbytheNaClO2/DMSOmethod.Both extraction procedures provide xylan samples with similar SEC weightdistributions,w(logVh),butslightlyshiftedtowardslarger macromolecularsizesforthecaseofPAA/DMSOextractedxylans However, the absolute molar masses are markedly higher for thexylansextractedbythePAA/DMSOmethodcomparedtothe NaClO2/DMSOmethod.Theselargermolarmassescanbeattributed

toa moreselectivedelignificationprocessbyPAA compared to NaClO2, which induces minimized xylan degradation The dif-ferent macromolecularconformations forxylansextracted from bothmethodscanbeattributedtothelargerGAXcontentinthe PAA/DMSO samples, the content of Araf substitutions and the larger degree of acetylation, which results in a more compact conformationincertainsizes(orhydrodynamicvolumes)ofsuch macromolecularpopulations.Fromtheseintegratedresults,itcan

beinferredthatacetylationandglycosyl(ArafandMeGlcA) sub-stitutionplayasignificantroleinfine-tuningthemacromolecular conformation ofthe acetylated GAXsextracted from sugarcane bagasseandstraw,which,inturn,affectstheelutionprofilesduring SECelutionandlightscatteringdetection

Fig 6.Empirical structure of xylan isolated by the PAA/DMSO process from bagasse (A) and straw (B).

Table 6

Empirical structure of acetylated GAX from bagasse and straw isolated by PAA/DMSO process The relative abundance has been calculated integrating the results from glycosidic linkage analysis, acetylation content by HPLC and NMR.

3.6 Empiricalstructuresofacetylatedarabinoxylanfrombagasse

andstraw

Theresultsofthe glycosidiclinkageanalysis,1H NMR

spec-troscopyandacetylquantificationwereintegratedsoastodevelop,

forthefirsttime,anempiricalmodelforthemolecularstructureof

acetylatedglucuronoarabinoxylan(GAX)isolatedfromsugarcane

bagasseandstrawwithPAA/DMSO.Theseempiricalxylan

struc-tures,togetherwithnomenclatureoftheintramolecularmotifs,are

setoutinFig.6andTable6,respectively

The acetylated GAX from sugarcane bagasse and sugarcane

strawwere showntobestructurallydifferentfromeach other

AcetylatedGAXfromsugarcanebagassehasamolarratiobetween

thexyloseunits,O-2linkedglucuronicunitsandtheO-3linked

ter-minalarabinoseunitsof10:0.1:0.5.Bagassexylancontained0.33

acetylgroupsperxyloseunit,withca53%oftheacetylgroupsbeing

observedatpositionO-3ofthexylose,ca37%atpositionO-2and

10%atpositionsO-2andO-3ofthesamexyloseresidue

AcetylatedAGXfromsugarcanestraw,ontheotherhand,has

amolarratiobetweentheXylpunits,O-2linkedglucuronicunits

andtheO-3linkedterminalarabinoseunitsof10:0.1:0.8,which

indicatesthat strawxylanis slightlymore branchedthan xylan

frombagasse,withpreferenceforArafsubstitutioninposition

O-3.However,strawxylanhassignificantlyloweracetylation(0.10

acetylgroupsperxyloseunit),withca68%oftheacetylgroups

beingobservedatpositionO-3ofthexylose,21%atpositionO-2of thexyloseandca11%atpositionsO-2andO-3ofthesamexylose residue

4 Conclusions

Two different extraction procedures (PAA/DMSO and NaClO2/DMSO) were compared for the isolation of acetylated AGXfromsugarcanebagasseandstraw.Ingeneral,thePAA/DMSO methodresulted in greater efficiencyand selectivity.This mild isolationmethodology,togetherwithdetailedstructuralanalyses, providesevidenceoftheintramolecularAraandacetylsubstitution patterninsugarcanexylans

Wesuccessfullydevelopedanempiricalmodelforthemolecular structureofacetylatedglucuronorabinoxylan(GAX)extractedfrom sugarcanebagasseandstraw,integratingtheresultsfrom methy-lationglycosidiclinkageanalysis,H1NMRspectroscopyandacetyl quantification

WehavefoundthatGAXfromsugarcanebagassediffers struc-turallyfromthatofsugarcanestraw.BagasseGAXhadaslightly lowerglycosylsubstitutionmolarratioofAraftoXylp(0.5:10)than xylanfromstraw(0.8:10),butahigherdegreeofacetylation(0.33 and0.10forbagasseand straw,respectively).Theacetylgroups wereattachedpredominantlytopositionsO-3(53%),O-2(37%)and O-2,3(10%)oftheXylpunitsinbagasseGAX,andtopositionsO-3