Visualisation of xanthan conformation by atomic force microscopy

Direct visual evidence obtained by atomic force microscopy demonstrates that when xanthan is adsorbed from aqueous solution onto the heterogeneously charged substrate mica, its helical conformation is distorted. Following adsorption it requires annealing for several hours to restore its ordered helical state.

jo u r n al h om ep age :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

Jonathan Moffata, Victor J Morrisb, Saphwan Al-Assafc, A Patrick Gunningb,∗

a Asylum Research an Oxford Instruments Company, Halifax Rd., High Wycombe, Buckinghamshire, HP12 3SE, UK

b Institute of Food Research, Norwich Research Park, Norwich, NR4 7UA, UK

c Hydrocolloids Research Centre, Institute of Food Science & Innovation, Faculty of Science & Engineering, University of Chester, Parkgate Road, Chester CH1

4BJ, UK

a r t i c l e i n f o

Article history:

Received 9 March 2016

Received in revised form 14 April 2016

Accepted 18 April 2016

Available online 20 April 2016

Keywords:

Atomic force microscopy

Xanthan

Structural conformation

Counterions

a b s t r a c t Directvisualevidenceobtainedbyatomicforcemicroscopydemonstratesthatwhenxanthanisadsorbed fromaqueoussolutionontotheheterogeneouslychargedsubstratemica,itshelicalconformationis distorted.Followingadsorptionitrequiresannealingforseveralhourstorestoreitsorderedhelicalstate Oncethehelixstatereforms,theAFMimagesobtainedshowedclearresolutionoftheperiodicitywith

avalueof4.7nmconsistentwiththepreviouslypredictedmodels.Inaddition,theimagesalsoreveal evidencethatthehelixisformedbyadoublestrand,aclarificationofanambiguityofthexanthan ultrastructurethathasbeenoutstandingformanyyears

©2016TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBYlicense

(http://creativecommons.org/licenses/by/4.0/)

1 Introduction

XanthanisabacterialpolysaccharideproducedbyXanthamonas

campestris (Garcia-Ochoa, Santos, Casas, & Gomez, 2000) The

polysaccharideconsistsofalinear␤(1,4)linkeddglucosecellulosic

backbonesubstitutedwitharegulartrisaccharidesidechain,

con-tainingtwomannose(Man)andaglucuronicacid(GlcA),attached

oneveryotherglucoseatC-3.Thechargedsidechainconsistsof

␤DMan(1-4)␤DGlcA(1-2)␣DMan(1-.Theterminalmannoseunits

maycontainapyruvicacidsubstituteandthe␣-linkedmannose

unitsmayhaveanacetylgroupatpositionO-6(Phillips&Williams,

2009).Anothertworecentstudieshaveshownthattherecanbe

moreheterogeneityofxanthan’srepeatunitthanwaspreviously

assumedintermsoftheratioofthechargegroupswithinxanthan’s

sidechainsdependinguponthefermentationconditions:Thereare

6differentpatternsofattachmentofpyruvateandacetategroups

tothepentasacchariderepeatunit,andtherelativeabundanceof

theseaffectsthestabilityoftheorderedstructure(Kool,Gruppen,

Sworn,&Schols,2013;Kool,Gruppen,Sworn,&Schols,2014).Itis

notclearwhetherthisheterogeneityarisesduetointra-or

inter-molecularsubstitution.Thechargedgroupsonthesidechainsplay

avitalroleinxanthan’saqueoussolubilityandalsoitsstructural

conformation(Phillips&Williams,2009).Inthepresenceof

stabil-isingcounterions,whichshieldtheintramolecularcharge–charge

∗ Corresponding author.

E-mail address: patrick.gunning@ifr.ac.uk (A.P Gunning).

interactions,thesidechainsfolddowncompactlyagainstthe back-boneleadingtotheformationofa5-foldorderedhelicalstructure (Norton,Goodall,Frangou,Morris,&Rees,1984).Theordered struc-tureismuchstifferthanthedisordered‘randomcoil’conformation

Inthehelicalstatexanthanhasapersistencelengthinexcessof

100nm,rankingitamongstthestiffestknownbiopolymers Pre-viousstudieshaveproventhatelectrostaticinteractionsbetween the charged groups within xanthan and screening counterions determineitsultrastructuralconformationinsolution(Matsuda, Biyajima,&Sato,2009;Bejenariu,Popa,Picton,&LeCerf,2010; Brunchi,Morariu,&Bercea,2014)

Traditionalmethodshavebeenusedwidelytoinvestigatethe molecularconformationofpolysaccharides.Opticalrotation, circu-lardichroism,differentialscanningcalorimetryandrheologyare convenientmethods for following thecourseof disorder–order andorder–disorder transitionsinresponse toexternal variables (temperature,ionicstrength,concentrationofspecificcationsand denaturants).X-rayfibrediffractionremainstheonlytechnique capableofcharacterisingorderedstructuresatatomicresolution, providedthatthechainsarewellenoughorientedandaligned,but atomicresolutionhasnotyetbeenachievedforxanthan

Theprinciplequestionaddressedbythepresentstudyisthat there has been considerable ambiguity for many years onthe detailofxanthan’ssecondarystructure(Morris,1998).Theinitial interpretationofX-rayfibrediffractiondatawasthatitformeda singlehelix (Moorhouse,Walkinshaw,&Arnott,1977).A subse-quentstudy(Okuyamaetal.,1980),carriedoutinresponsetothe

“twostrands=doublehelix”lobby,examinedpossibledouble-helix http://dx.doi.org/10.1016/j.carbpol.2016.04.078

0144-8617/© 2016 The Authors Published by Elsevier Ltd This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ).

J Moffat et al / Carbohydrate Polymers 148 (2016) 380–389 381

models.Itwasconcludedthat,onthebasisoftheX-rayevidence

alone,itwasnotpossibletoassignadouble-helixorsinglehelical

modelforxanthan

In terms of physical chemical studies the salt-induced

disorder–ordertransitionfollowedfirstorderratherthansecond

orderkinetics,whichsuggestedasinglehelix(Nortonetal.,1984)

Manygroups(Morris,1998)haveequatedobserveddimerization

ofxanthanwithdouble-helixformation;butthatisnotevidence

basedand itis potentiallyanoversimplifiedinterpretation.The

ambiguity with suchmethods is thefact that theyare

ensem-blemeasurements Thismeansthat theanalytical conclusionis

controlledby theratio of ordered todisordered states, sothat

theylackacertaindegreeofsensitivitycomparedto

microscop-icaltechniques,suchasatomicforcemicroscopy(AFM) AFMis

capableof visualisingthestructureofindividualmolecules.The

mainobjectiveofthisstudyistoprovidedirectevidenceonthe

natureofxanthan’ssecondarystructure.Theuniqueadvantageof

AFMisitsabilitytovisualisedirectlythetopologyofpolymer

net-worksundernear-nativeconditions,whichcanbeaverypowerful

complimentary technique to combine with ensemble

method-ologies Anintegral studyusingvarious biophysicaltechniques,

namely,AFM,gelpermeationchromatographywithmulti-angle

lightscattering(GPC-MALLS)andintrinsicviscositymeasurements

bycapillaryviscometryontheconformationofxanthan,

follow-ingvariousdifferenttreatments(heating,autoclaving,irradiation

andhighpressurehomogenisation),wasrecentlyreported(Gulrez,

Al-Assaf,Fang,Phillips,&Gunning,2012).Severalpolymer

param-etersderivedfromthesetechniques,suchastheradiusofgyration

(Rg),Mw,polydispersity,molarmassperunitcontourlengthofthe

rod(ML)andHugginsconstant(KH)werecorrelatedwellwiththe

resultsobtainedbyAFM.Itwaspossibletocorrelatetheheight

measurementsobtainedbyAFMwithvaluescloseto1000Dalton

pernanometre(Danm−1)and2000(Danm−1)assignedforsingle

anddoublehelix,respectivelyinagreementwithpreviousreports

whichsolelyreliedonlightscatteringmeasurements(Sato,Kojima,

Norisuye,&Fujita,1984;Sato,Norisuye,&Fujita,1984)

Further-more,usingpositively-chargedmica(coatedwithpoly-l-lysine)a

singlestrandmoleculewastrappedina‘randomcoil’conformation

(Gulrezetal.,2012).Thisisconsistentwiththewidelyagreedview

thatxanthanat lowconcentrationandnegligible ionicstrength

adopts‘randomcoil’conformation

The present study reveals new images of xanthan at

sub-molecular resolution revealing the fine detail of its secondary

structuredevelopment,whichhasenabledtheprocessofcharge

screeningtobeinvestigatedinamannerneverpreviouslyreported

2 Experimental

2.1 Atomicforcemicroscopy

Theatomicforcemicroscope(CypherAFM, AsylumResearch

Inc,anOxfordInstrumentscompany,SantaBarbara,CA,USA)was

operatedinACmodeinaqueousbufferscontainingdifferent

coun-terions.Buffer1:10mMHEPES3mMZnCl2pH5.3,andbuffer2:

10mMHEPES3mMNiCl2pH7.0(Sigma-AldrichChemical,Poole,

Dorset,UK) Oscillation of thecantilevers at their fundamental

resonant frequency wasdrivenusing ‘blueDrive’photo-thermal

excitation.Photothermalexcitationisanewtechnologydeveloped

byAsylum researchthatprovidesa morestable andcontrolled

formof cantilever oscillation.Thisis achieved bypositioning a

laserbeamwithwavelength425nmandmodulatedpowerdirectly

onto the cantilever’s bimetallic strip, as opposed to the

tradi-tionalpiezo-acousticmethod,whichmechanicallyoscillatesthetip

holdercausingmorepotentialdisturbanceofthesamples.Localised

heating bythe bluelaser causes thecantilever tobenddue to

thebimetallic strip effect and modulationof thepower causes theprobetooscillateatanaccuratelycontrolledfrequencyand amplitude.Thelaserpowermodulationfrequencywassetatthe fundamentalresonantfrequencyofthecantilever(1.37MHz)and thepowerlevel124.8␮Wsettogenerateanappropriatelysmall oscillationamplitude(∼1nm).Thefeedbackloopcontrolset-point wasalsokeptata verylowlevelofdampingofthecantilever’s free oscillation (∼5–10%) tominimise theloading forceon the molecules.TheAFMtipsusedwereArrowUHF-AuD(NanoWorld

AG,Neuchâtel,Switzerland).Scanratesweresetat1.5Hz 2.2 Preparationofxanthansolutions

Thexanthanusedinthisstudywasapowderedfoodgrade xan-thangum(KeltrolRD,CPKelco,Atlanta,GA,USA).‘RD’standsfor

areadilydissolvableproduct.Thestocksolutionwaspreparedata concentrationof1mgml−1inpurewater(18.2M).Thexanthan powderwasdispersedimmediatelyafteradditiontothewaterat

22◦Cbystirring.Itwasthenleftfor30mintohydratebefore heat-ingto95◦Cfor60mintocompletelydispersethemolecules.The stocksolutionwasallowedtocooltoroomtemperature(22◦C)and thendilutedto3␮gml−1intoeitherwater(method1,below),or theaqueousbuffers(method2,below).Thedilutedsolutionswere thenre-heatedto95◦Cfor60mintoreduceanyaggregationand allowedtocoolbackto22◦CpriortotheAFMimagingpreparation procedures

The additional heating step was merely to ensure that the samestructureandproportionsofsoluble/aggregatefractionsare presentinthetestmaterial(renaturedstate).Gulrezetal.(2012) investigated the effect of heat treatment on xanthan aqueous solutions(4mg/mL)dissolvedindistilledwater,whichwas sub-sequentlydilutedtocontain0.1MLiNO3prior toinjectioninto theGPC-MALLSsystem.Theydemonstratedthatheatingxanthan aqueoussolutionupto40minat85◦Cresultedinsimilar molec-ularweightparameters(i.e.weightaveragemolecularweight,% massrecoveryandpolydispersity).Furtherheatingupto60min resultedinanincreaseinthemassrecoveryandaslightincrease

inthemolecularweightasaresultofdisassociationoflarge aggre-gatesinitiallyretainedonthe0.45␮mfilter.Themolecularweight

isreducedtoalmosthalfwithfullmassrecoveryandanincrease

inthepolydispersity(from1.64to2.75)whenthedilutedsolution washeatedfor2at85◦C

2.1 AFMimagingpreparationprocedures Twomethodswereusedtophysisorbxanthanmoleculesonto freshly-cleavedmuscovitemica(AgarScientific,Cambridge,UK) 2.2.1 Method1:dropdeposition(includesdrying)

A3.5␮ldropletofxanthanataconcentrationof3␮gml−1in waterwasplacedontothefreshly-cleavedmicaandleftto evapo-rateatroomtemperature(22◦C).Whenfullydrythesamplewas thenplacedintotheliquidcelloftheAFMandimagedinthe aque-ousbuffersdescribedinSection2.1

2.2.2 Method2:in-situadsorption(no-drying)

100␮lofthebuffer-dilutedxanthansolution(3␮gml−1)was placed directlyintotheliquidcelloftheAFM,which contained freshly-cleavedmicaandthenimagedasdescribedinSection2.1

3 Results

Fig.1displaysanexampleofthedatathatwerealwaysobtained

attheearlystagesofimagingxanthaninaqueousbuffers,prepared

bybothmethods(Fig.1adropdeposition,Fig.1 in-situ adsorp-tion) The swirly white linesover one moleculein each image

Fig 1.Early stage images of xanthan on mica in aqueous buffer (a) Method 1, drop-deposited, imaged in buffer 1 (b) Method 2, in-situ adsorbed from and imaged in buffer

2 Bottom panels: Line profiles depict the heights of the features beneath the white lines in the images.

illustratesthelocationof thecross-sectionsthatareprofiled in

thegraphbeneaththeimages.Thereasontheyareswirlylinesis

toenablerepetitivequantificationoftheheightofthemolecules

Theaveragevalueinbothcasesis2.06±0.19nm.Measurementof

heightisthemostaccuratewaytoquantifythediameterof

poly-merswithAFM,aslateraldimensionsaresignificantlyoversized

byprobe-broadening(Morris,Kirby,&Gunning,2009).Thevalue

oftheheightofthischainisslightlylargerthanthepredictedwidth

(1.8nm)ofxanthaninthedoublehelicalform(Millane,1990).The

linearityofthechainsdemonstratesrigidity.Thisprovidesevidence

thatatthisearlystage,despitetherebeingnosecondarystructure

visiblewithinthechains,xanthanisnotina disordered

confor-mation.Previousresearchhasshownthatxanthaninadisordered

‘randomcoil’statehasasignificantlylowerchainheightthaninthe

orderedstateandalsoappearslesslinearduetoitslackofrigidity

(Gulrezetal.,2012).However,asmentionedabovetheinteresting

issueisthat,despitethefactthatthedimensionsandlinearityof

thechainsatthisearlystageareclosertotheorderedconformation

ofxanthan,noperiodicitywasvisiblealonganyofthechains

AfteracertainlengthoftimeintheliquidcelloftheAFMthe

heliceswerevisualisedineachofthesamplespreparedbythetwo

differentmethods(Fig.2aandb).Oncetheheliceshaveannealed

withintheirorderedstatethewidthofthechains(accurately

mea-suredasheightbyAFM)is1.6±0.16nm,whichismorecompactby

approximately0.4nmthanthosemeasuredpriortoannealing.This

heightvalueisreasonably consistentwiththewidthofxanthan

obtainedfromx-rayfibrediffractionmeasurements(1.6–1.8nm)

(Millane,1990).Inbuffer1there-conformationofthehelixtook

∼16hoursandinbuffer2ittook∼4hours.Thedifferencebetween

thebufferscontainingNi2+andZn2+isthepH(7.0&5.3

respec-tively)

Thevaluesoftheperiodicityobservedafterannealinghavebeen

quantifiedbylineprofiling alongthemolecules(Fig.3a andb)

Thisrelatestothepitchofthehelices, andthevaluesobtained

are4.67±0.29nminbuffer1(Zn2+)and,althoughinFig.3 the

lineprofileis noisier,theactualnumberofvisible helicalturns

givesasimilarvalue;4.75±0.11nminbuffer2(Ni2+).These

val-uesarefullyconsistentwiththepreviousx-rayfibrediffraction

data(Millane&Narasaiah,1990;Okuyamaetal.,1980).Thereis

asignificantdifferenceintermsoftheconsistencyofthe struc-turalorderbetweenthedifferentadsorptionmethodologiesused

Inthedropdepositedsamples(method1)manyofthechains dis-playendswherethehelicesareunravelled(Figs.2a,4,and5a–c) andtheoccasionalcasewithasmall,unravelledsectionofthehelix arrowedinFig.4.Theheightsofthehelicalsectionandthe unrav-elledsectionarequantifiedbythelineprofileinFig.4.Thehelical sectionhasaheightof1.6nm andtheunravelledsectionhasa heightof0.6nm,whichislessthanhalfthevalue.Thedifferencein theheightsconfirmsthatthetallersectionisnotasimple dimeriza-tionoftwosinglechains.Itisthereforeamorecomplexstructural arrangementasexpectedforahelix

Furtheranalysisofthisimagewiththeunravelledsectionofthe orderedstructureisdisplayedinFig.5.Thelengthofthe unrav-elledsectionis4.6nm(Fig.5a).ThelineprofilesinFig.5 &creveal thatthedifferenceintheheightsofeachsectionareconsistentwith thosemeasuredontheotherxanthanmoleculeinFig.4:helical sec-tion1.6nm(Fig.5b)andunravelled‘mid-section’0.6nm(Fig.5c) Theheightequivalenceandeliminationofperiodicityconfirmsfull disorderingoftheunravelled‘mid-section’.Thefactthatthelength

oftheunravelledsectionisnolongerthanthehelicalpitchobserved

inbothbuffers(Fig.3)indicatesthatitisunravellingofthehelix intotwodisorderedstrands

TheimagesinFig.6revealsfurtherevidenceforadoublehelical conformationofxanthan;oneoftheproposedhelicalstructures suggested to be formedin a dilute solution.The suggestion is thatxanthan’sdoublehelixdissociatesintotwosinglechainsat

adenaturingconcentration(≤1mgml−1)but,despiteitseeming inconsistent,thereisapossibilitythatsinglechainscanreconstruct theintramoleculardoublehelicalstructureinananti-parallel man-nerwithahairpinloopatoneendduringtherenaturationprocess (Matsudaetal.,2009).ThearrowinFig.6 showsthatoccasional predictedhairpinloopsdoindeedexistandasecondoneisshown

inFig.6c(analternativelycolouredversionofFig.2a)

Althoughtheconformationofthexanthanchangedwithtimeto thefullyorderedhelicalstate,Fig.7showsasetofimages demon-stratingthatthemoleculesremainedstablyattachedtothemica

J Moffat et al / Carbohydrate Polymers 148 (2016) 380–389 383

Fig 2.Resolution of the helical pitch of xanthan: (a) Method 1, drop-deposited, imaged in buffer 1 (b) Method 2, in-situ adsorbed from and imaged in buffer 2 Bottom panels: Line profiles depict the heights of the features beneath the white lines in the images.

invirtuallythesamepositionsoverthelongestinvestigatedtime

periodof16hoursintheliquidcellinbuffer1.Fig.7awasthelast

imagetakenduringtheinitialstageswhennoheliceswere

observ-ableandFig.7 wasthefirstimagetakenthefollowingdaywhen

theheliceswereobservable.Despitetherebeingasmallamount

ofdriftbetweenFig.7aandb(∼500nmtothetopright)thewhite

boxmarkershowsthematchingregionsinbothimageswithno

significantmovementofthemoleculesthemselves.Fig.7cshows

howconsistentthepositionofthemoleculesisinbothscansby

overlappingtheimages.Thelaterimage(Fig.7b)hasbeenplaced

ontopoftheearlierimage(Fig.7a)andsettoadifferentcolour

(green)withanopacityof42%sothatbothcanbevisualised

4 Discussion

Insolution,screeningofthexanthan’schargedgroupsby

coun-terionscanbeachievedrapidlyatanoptimalconcentrationofsalt

becausetheyaremobileandcansustainanequilibriumstate

Pre-viousresearchhasshownhowsensitivethehelicalconformation

ofxanthanistothelevelofsaltinthesolution(Bejenariuetal.,

2010;Brunchietal.,2014;Matsudaetal.,2009).Thetwo

aque-ousbuffersolutionsusedforimagingthexanthaninthepresent

studyincluded divalentcounterions (Ni2+ and Zn2+)at optimal

screeningconcentrationsfortwo purposes:Theprincipaloneis

tofacilitateadsorptionofxanthanontothemicasothatitcanbe

imagedinliquid.Withoutsufficientscreeningoftheelectrostatic

repulsionbetweennegativelychargedwatersolublemoleculesand

micathereisnopossibilityofsuccessfullyimagingmoleculesin

aqueousliquidsbecausetheywilldesorbfromthemicasurface,

eveniftheyhavebeenpreviouslydepositedbyairdrying.Thetwo

counterions(Ni2+,Zn2+)werediscoveredtobeoptimalforbinding

DNAtomicainaqueousbuffersduetotheirsmallionicradii(0.69

&0.74Årespectively),whichallowsthemtofitintothecavities

abovetherecessedhydroxylgroupsinthemicalattice(Hansma

&Laney,1996).Inaddition,Ni2+andZn2+haveanomalouslyhigh

enthalpiesofhydrationwhichisproposedtoenablethemtoform

strongcomplexeswithligandsotherthanwater(Hansma&Laney,

1996).Thesecondaspecttobeconsideredinthepresentstudyis

thatdivalentcounterionscausestabilisationofxanthan’s helical

conformationinsolutionatconcentrations≥1mM(Brunchietal.,

2014;Bejenariuetal.,2010).Thisiscrucialbecausethexanthan hadtobedilutedtoextremelylowconcentrations(≤3␮g/ml−1)to

ensurethatsub-monolayeradsorptionwasachievedonthemica surface.Athigherconcentrationsthemicasurfacebecomes over-crowdedwithmultilayersofxanthan,whichpreventsresolvingthe individualmolecules

Whenthexanthanmoleculesadsorbtothemica,whichhasa heterogeneoussurfacechargedistribution,thesituationisdifferent thaninsolution.Thelocalisedelectrostaticinteractionsthatoccur betweenthechargedgroupspresentonthepolysaccharidechain andthesolidmicasurfacearehighlyunlikelytobeoptimal,since

onthemicathespatialdistributionofthechargedgroupsarefixed

sotheycannotmoveandadapttothechargedgroupsonthe xan-thanmolecule.Theoreticalmodellingstudiesdemonstratedthat heterogeneouslychargedpolymersadsorbingtoheterogeneously chargedsurfacesdosobyadoptingtheirshape,interpretedas pat-ternrecognition(Chakraborty&Bratko1998;Golumbfskie,Pande,

&Chakroborty,1999).Ithasbeenpreviouslyreportedthat alter-ation ofxanthan solutionpHaffectsthetransitiontemperature (Bejenariuetal.,2010)althoughthepHdifferencesinvestigated

inthatstudyweresignificantlylarger(3,7&13)thaninthisstudy Anotherdifferencebetweenbuffers1&2isthewatersubstitution ratesofNickelcomparedtoZinc(Ni2+2.7×104/s,Zn2+5×108/s; Kobayashi,Nagayama,&Busujima,1998).Thisisthereforemore likelytobethereasonthatthere-annealingtimeofthexanthan helicesissignificantlydifferentinthebuffersusedinthepresent study

Thisclearlyhasaneffectonxanthan’sultrastructure,butthere aretwopotentialreasons.Theinitialandmoreobvious interpreta-tionofthetimerequirementtovisualisethehelicalperiodicitydue

totheslightstreakinessseenintheearlystageimages(specifically Fig.1a)isthatthexanthanmaynotbesufficientlystablyattached

tothemicaforhigh-resolutionimagingattheearlystages How-ever,thereisrelativelystrongevidencethatloosebindingmaynot

betheonlyreason.TheimagesinFig.7 showthatnearlyallof themoleculesremainedinpreciselythesamepositiononthemica overtheentireexperimenttimesotheymusthavebeenreasonably boundfromtheverystart.Thisisprobablyduetothecombination

ofthe‘patternrecognition’shapeadoptionofthemolecules bind-ingtothemostsuitablychargedregionsofthemicaandalsothe

Fig 3. Helical pitch of xanthan (a) Method 1, drop-deposited, in the presence of Zn 2+ , (b) Method 2, in-situ adsorbed in the presence of Ni 2+ Bottom panels: Line profiles depict the heights of the features beneath the black lines in the images.

J Moffat et al / Carbohydrate Polymers 148 (2016) 380–389 385

Fig 4.Method 1, drop deposited xanthan sample imaged in buffer 1 showing unravelling of the helices Bottom panel: Line profiles depict the heights of the features beneath the thin white line in the image.

strengthofthestiffpolymernetwork.Thisprovidesclearevidence

thatthemoleculesthatadoptedtheirconformationfrompartially

disorderedtothefullyorderedhelicalstateoverthetimeperiod

werethosewhichwereimagedduringtheinitialstages,andnota

setofotherxanthanmoleculesonthemica

Baseduponthisauniqueinterpretationofthelackofobservable

periodicityintheearlystageimagessuggeststhat,evenifitisin

afullyorderedconformationinsolution,xanthan’shelicalorderis

probablydistortedasitinitiallyadsorbstothemica.Thedistortion

ishoweverlimited;theheightmeasurementsandlinearity

illus-tratedinFig.1demonstratethatthedistortionisnotafullhelix

tocoiltransitionbutclearlyissufficienttoremoveanyobservable

periodicityalongthepolymerchain.Thisenablesinterpretationof

theheightreductionofthechainsfrom2nmattheinitialstage (Fig.1)to1.6nm(Figs.2,4,and5b).Helicalformationwas there-forelikelyduetoanannealingcompactionoftheslightlydistorted structure

Thereisafortunatebenefitfrommica’sdistortinginfluence; certainsectionsofthechains(Figs.2and4–6)donotfullyre-order whichenablesvisualisationofthecompositionofthehelix.Itis clearlydoublestranded.Thefactthatthelengthoftheunravelled sectioninFig.5aisnolongerthanthehelicalpitchobservedin bothbuffers(Fig.3)indicatesthatitisunravellingofthehelixinto twodisordered‘strands’.Althoughthisseemsobviousitprovides additional informationoninterpreting thenatureof thehelical structure,singleordoublehelical?Ifitwasasinglehelixitwould

Fig 5.Additional measurements of the xanthan molecule containing the partially unravelled section Line profiles (right panels) depict the heights and distances of the features beneath the red lines in the AFM images (left panels) (a) Profile of the gap created by the unravelled section, with blue markers (in both panels) labelling the transition zones (b) Profile across the helical section (c) Profile across the unravelled section (For interpretation of the references to colour in this figure legend, the reader

is referred to the web version of this article.).

notproducetwostrands.Unravellingofadoublehelixwould

pro-ducetwofullydisorderedstrands

Thevisualisation of thehairpins(Fig.6)providessignificant

assistanceintheinterpretationofthepartiallyunravelledmiddle

sectionoftheotherxanthanchaininFigs.4and5.Insummary,

thiscombinationofimagesprovidesfurtherdirectevidencethat

xanthan’sordered structure isa double helix.In addition,for a

doublehelixtobeformedbyasinglechainitwillwraparound

itselfin anti-parallelconformation.The imagesobtainedin this

studydemonstratethatxanthancan,byintra-molecular

associa-tion,formananti-paralleldoublehelix.Forthemajorityofimages

whichdonotshowhair-pinloopsthesimilarheightandpitch

sug-gesttheseareeitherparalleloranti-paralleldoublehelicesformed

byinter-molecularassociationoftwoxanthanchains

Thealternativethatthesemolecularstructurescouldbeformed

byassociationoftwosinglehelicesisunlikelyforthefollowing

rea-sons.Thestructurescontaininghairpinloopsarethesameasthose

thatdonotshowsuchloops.Instudiesoftherelatedxanthan-like

polysaccharideacetan(Kirby,Gunning,Morris,&Ridout,1995)it

waspossibletoimage,byAFM,side-by-sideassociationofhelices

inanalignedliquidcrystallinemonolayershowingtheexpected pitchandheightforthehelices.ThisshowsthatAFMwould distin-guishbetweenadoublehelixandpairedsinglehelices,sinceboth singlehelicesinthepairwouldneedtobindtothemica.Further,as discussedlater,althoughthereisastereo-chemicalbasisfor dimer-izationofchainstoformadoublehelix,thereisnostereo-chemical basisfordimerizationofsinglehelices,whichwouldrestrict aggre-gationtodimers,orexplainwhyitextends alongthecomplete lengthofthemolecules

In method 2, thesamples prepared byin-situ adsorption of thexanthanfromthebufferscontainingthedivalentcounterions, unravellingoftheheliceswasnotdetected(Figs.2 b,and3b) Thisreflectsthestatethatthexanthanislikelytobeinpriorto itsattachmenttomicainthedifferentmethods.Inmethod1(drop depositionfrompurewater)itislikelytobeinadisordered con-formationattheinitialstage,duetotheverylowconcentration

ofthexanthan,whichalsomeansthatthesolutionisverylowin ionicstrength.Asthewaterdropletevaporatesthepolysaccharide

J Moffat et al / Carbohydrate Polymers 148 (2016) 380–389 387

Fig 6. (a) Hairpin loop images: Helical xanthan molecules with unravelled ends (b) white box marked region electronically zoomed, and (c) second hair-pinned molecule.

concentration and ionic strength increases driving xanthan

towardsitshelicalconformation,butnotsurprisinglytherecanbe

manysectionsofthechainsthatdonothavetimeorthecorrect

conditionstofullyre-order.Inmethod2themoleculeswill

pre-dominantlybeinthehelicalconformationinthesolutionduetothe

optimalconcentrationofthedivalentcounterionsinthebuffer,and

alsotheheating/coolingstepsinthesamplepreparationprocedure

(Gulrezetal.,2012)beforetheyadsorbtothemica.Ascanbeseen

intheexamplesimagesinFigs.2b,and3bthisgreatlyreducesthe

probabilityofanyfullyunravelledsectionsoftheirconformation

If thesecondary structure of xanthan was one of theother

previous interpretations (Norton et al., 1984) based upon the

measurementsofthekineticsofconformationalordering,namely

dimerizationofsinglehelices,thentwostrandswouldalsobea

predictableoutcome.Lightscatteringandopticalrotationvalues

camefrompoint-bypointequilibriummeasurements,sothiswas

notakinetic“timelag”effectbutadifferenceinthetemperatures

neededtotriggertheonsetofconformational orderingandthe

increaseinmolecularweight.Similarly,whenthevariablewasnot

temperaturebutconcentrationofcadoxen,reductioninmolecular

weightbeganatsubstantiallylowercadoxenconcentrationthan

lossofconformationalorder.However,thatcouldnotbeattributed

toslowkineticsofordering, becausethestartingpoint wasthe

orderedsolid.Thatwasinterpretedasxanthaninitiallyforming

singlehelicesinthedisorder-ordertransitionand then

dimeris-ing,butthereisvisualevidenceinthisstudy,whichrevealsthat

xanthanisnotlikelytobeadimerisedsinglehelix.Amorerecent studyshowedthatwhenxanthansolutionistreatedbyhigh pres-surehomogenisationthereisasignificantdecreaseinthemolecular weightbutthemeasurementsofmolecularweightperunit con-tourlengthoftherod(ML)suggestsitisstilldoublehelical,and evenafterstorageofthesolutionfor3daysthemolecularweight parametershardlychanged(Gulrezetal.,2012)

Dimerizationwouldberesolvablealongtheentire‘fibre’ifthe singleheliceswereparallel-aligned.Theonlypotentialreasonthat dimerisedsingle helices couldbe visually concealed until they becomedisorderedwouldbeiftheywrapverycompactlyaround eachotherandhenceappearassingleratherthandoublefibres But in that case theheights of both the ordered sections and theseparated‘strands’wouldnotmatchthevaluesquantifiedin Figs.4and5.The‘strands’wouldnotdroptosuchanextentbecause themodelledwidthofxanthanasasinglehelixwasthesamevalue (1.6nm)asthat ofthedoublehelix (Nortonetal.,1984)andof coursethatmeanstheorderedsectionifitwascomposedoftwo wrappedsingleheliceswouldverylikelybetallerthan1.6nm Anotherhypothesisisthatfordimerisedsinglehelicesto sepa-rateanddisordertoreachthelowheightofthe‘strands’inallthe unravelledsectionsvisibleintheAFMimagesthenthelengthof thedisorderedgapinFig.5awouldbedifferentthanthemeasure

ofthepitchoftheorderedsectionifitwascomposedoftwosingle helicesthatravelaroundeachother.Therefore,thelengthofthis unravelled‘mid-section’potentiallyprovidesfurtherdirectvisual

Fig 7. (a) Example images of the stability of the xanthan molecules over a 16 hour time period in buffer 1 (a) at 1 hour, (b) at 16 hours, and (c) images overlaid (For interpretation of the references to colour in the text, the reader is referred to the web version of this article.)

evidencefavouringthatthesecondarystructureofxanthaninits

orderedstateisadoublehelix

AlthoughthisstudyisbasedsolelyuponAFMimagedatathe

conclusionisnotjustfromthevisualevidence.Thecombination

ofimageswithtopographicalquantificationhasenabledthe

inter-pretationoftheAFMdatainrelationtothepreviouspredictions

ofxanthan’ssecondarystructurefromalloftheothertechniques,

whichhavebeencarriedoutovermanyyears.Therefore,theheight

datafromalloftheorderedanddisorderedsectionsvisualisedin

theAFMimagesprovidesconfirmationthatthexanthanobserved

inthisstudyisdoublehelical.Notethatthesefactsarebasedon

theAFM-observedstructureswhichareofcourselimitedtothose

thatbindtomicaandnodimerisedsinglehelicalversionswere

detectedinanyoftheimagescaptured(n=41,typicalmolecules

perimage=50–250),butthere maystillbea possibilitythat in

solutionstochasticvariationsinxanthan’sconfirmationcanexist

duetomolecularmobility.Itisthereforepossiblethatthe

contro-versyoversingleordoublehelixformationmayhavearisendueto

experimentsdoneunderdifferentexperimentalconditionswhich

favouredintra-orinter-moleculeassociation:doublehelicescan formbyintra-molecularassociationofsinglechains

Itisinteresting toconsiderstereo-chemicalreasonswhythe helicalstructureiscomposedoftwointeractingchains.Thenature

of theprimary structure ofxanthan suggeststhat the distribu-tion of sidechains along the backbone resultsin anuncharged and a charged face of the cellulosic backbone Association of theunchargedfacesof thebackbone, anda twist into a 5-fold helix tooptimise thedistributionof charge onthehelix, could explaintheformationofthedoublehelix.Thiswouldbeconsistent withintra-molecularassociation(anti-parallel)orinter-molecular (anti-parallelorparallel)association.Suchamodelwouldbe con-sistent with the observed 6-fold helical complex with a pitch

of 5.6±0.1nm, formed between the xanthan-like polysaccha-rideacetanandtheglucomannankonjacmannan(Ridout,Cairns, Brownsey,&Morris,1998)andtheproposeddoublehelical struc-ture,which contains both a konjacmannan (uncharged) and a singleacetanchain(charged)withinthehelix(Chandrasekaran, Janaswamy,&Morris,2003).Acetanlikexanthanformsa5-fold helixwithapitchof4.8nm(Morris,Brownsey,Cairns,Chilvers,&

J Moffat et al / Carbohydrate Polymers 148 (2016) 380–389 389

Miles,1989).Thetransitiontoa6-foldhelixcouldresultfromthe

differentoptimiseddistributionofchargealongthehelical

com-plex

5 Conclusions

The ability of AFM to resolve polysaccharide molecules at

sub-molecularresolutionandthedistortingeffectofthe

heteroge-neouslychargedsubstrate,mica,hasprovidedthefirsteverdirect

visualevidence which confirms that theproposed anti-parallel

doublehelicalultrastructureofxanthancanbeformedthrough

intra-molecularassociation.Forthemajorityoforderedstructures,

whichdonotshowthepresenceofloops,bothanti-paralleland

parallelmodels forthe doublehelix formed byinter-molecular

associationarepossible.Thedatainthisstudyconfirmsthatthe

precisionoftheformationofxanthan’ssecondaryhelicalstructure

isindeedsensitivelydrivenbyanoptimisationofintra-molecular

chargescreening.TheAFMdatasuggeststhatxanthan’s

predomi-nantequilibriumstructuralconformationisadoublehelix

Acknowledgments

TheauthorsthankEdwinMorris(UniversityCollegeCork)for

discussionsonthescientificprinciplesofthepotentialvariations

inxanthan’sorderedstructuralarrangements.Thanksarealsodue

toNeilWilson(UniversityofWarwick)fortakingpartinsomeof

thesuccessfulimagingofxanthanhelicesattheRMSMMC2015

conferenceinManchester.Fundingforthisworkwasprovidedby

BBSRCthroughitscorestrategicgranttoIFR

References

Bejenariu, A., Popa, M., Picton, L., & Le Cerf, D (2010) Effect of concentration, pH

and temperature on xanthan conformation: a preliminary study before

crosslinking Revue Roumaine de Chimie, 55, 147–152.

Brunchi, C.-E., Morariu, S., & Bercea, M (2014) Intrinsic viscosity and

conformational parameters of xanthan in aqueous solutions: salt addition

effect Colloids and Surfaces B: Biointerfaces, 122, 512–519.

Chakraborty, A K., & Bratko, D (1998) A simple theory and Monte Carlo

simulations for recognition between random heteropolymers and disordered

surfaces Journal of Chemical Physics, 108, 1676–1682.

Chandrasekaran, R., Janaswamy, S., & Morris, V J (2003) Acetan:glucomannan

interactions: a molecular modeling study Carbohydrate Research, 338,

2889–2898.

Garcia-Ochoa, F., Santos, V E., Casas, J A., & Gomez, E (2000) Xanthan gum:

production, recovery, and properties Biotechnology Advances, 18, 549–579.

Golumbfskie, A J., Pande, V S., & Chakroborty, A K (1999) Simulation of

biomimetic recognition between polymers and surfaces Proceedings of the

National Academy of Sciences of the United States of America, 96, 11707–11712.

Gulrez, S., Al-Assaf, S., Fang, Y., Phillips, G O., & Gunning, A P (2012) Revisiting the conformation of xanthan and the effect of industrially relevant treatments Carbohydrate Polymers, 90, 1235–1243.

Hansma, H G., & Laney, G E (1996) DNA binding to mica correlates with cationic radius: assay by atomic force microscopy Biophysical Journal, 70, 1933–1939.

Kirby, A R., Gunning, A P., Morris, V J., & Ridout, M J (1995) Observation of the helical structure of the bacterial polysaccharide acetan by atomic force microscopy Biophysical Journal, 68, 360–363.

Kobayashi, S., Nagayama, S., & Busujima, T (1998) Lewis acid catalysts stable in water: correlation between catalytic activity in water and hydrolysis constants and exchange rate constants for substitution of inner-sphere water ligands J.

Am Chem Soc., 120, 8287–8288.

Kool, M M., Gruppen, H., Sworn, G., & Schols, H A (2013) Comparison of xanthans

by the relative abundance of its six constituent repeating units Carbohydrate Polymers, 98, 914–921.

Kool, M M., Gruppen, H., Sworn, G., & Schols, H A (2014) The influence of the six constituent xanthan repeating units on the order-disorder transition of xanthan Carbohydrate Polymers, 104, 94–100.

Matsuda, Y., Biyajima, Y., & Sato, T (2009) Thermal denaturation, renaturation, and aggregation of a double-helical polysaccharide xanthan in aqueous solution Polymer Journal, 41, 526–532.

Millane, R P (1990) Molecular and crystal structures of polysaccharides with cellulosic backbones In R Chandrasekaran (Ed.), Frontiers in carbohydrate research-2 (pp 168–190) London and New York: Eslevier Appled Science.

Millane, R P., & Narasaiah, T V (1990) X-ray fiber diffraction studies of a variant

of xanthan gum in which the sidechain terminal mannose unit is absent Carbohydrate Polymers, 12, 315–321.

Moorhouse, R., Walkinshaw, M D., & Arnott, S (1977) Xanthan gum molecular conformation and interactions In Sandford, P A Sandford, & A Laskin (Eds.), In extracellular microbial polysaccharides (Vol 45) (pp 90–102) San Francisco, CA USA: ACS Symp Ser.

Morris, V J (1998) Gelation of polysaccharides In S E Hill, D A Ledward, & J R Mitchell (Eds.), Functional properties of food macromolecules (2nd ed., Vol 45,

pp 143–226) Aspen Publishers ISBN0 7514 0421 7.

Morris, V J., Brownsey, G J., Cairns, P., Chilvers, G R., & Miles, M J (1989).

Molecular origins of acetan solution properties International Journal of Biological Macromolecules, 11, 326–328.

Morris, V J., Kirby, A R., & Gunning, A P (2009) Atomic force microscopy for biologists (2nd ed.) London: Imperial College Press ISBN-10: 184816467X (Chapter 2).

Norton, I T., Goodall, D M., Frangou, S A., Morris, E R., & Rees, D A (1984).

Mechanism and dynamics of conformational ordering in xanthan polysaccharide Journal of Molecular Biology, 175, 371–394.

Okuyama, K., Arnott, S., Moorhouse, R., Walkinshaw, M D., Atkins, E D T., & Wolf-Ullish, C H (1980) A D French, & K H Gardner (Eds.), Fiber diffraction methods (Vol 141) (pp 411–427) ACS Symposium Series http://dx.doi.org/10 1021/bk-1980-0141.ch026 Chapter 26, ISBN13: 9780841205895eISBN: 9780841207424

Phillips, G O., & Williams, P A (2009) Handbook of hydrocolloids (2nd ed.) Boston: CRC Press.

Ridout, M J., Cairns, P., Brownsey, G J., & Morris, V J (1998) Evidence for intermolecular binding between deacetylated acetan and the glucomannan konjac mannan Carbohydrate Research, 309, 375–379.

Sato, T., Kojima, S., Norisuye, T., & Fujita, H (1984) Double-stranded helix in dilute solution: further evidence Polymer Journal, 16, 423–429.

Sato, T., Norisuye, T., & Fujita, H (1984) Double-stranded helix in dilute solution: evidence from light scattering Polymer Journal, 16, 341–350.