alkaloid analogs and enantioseparation of natural and synthetic Cinchona Polysaccharide-based stationary chiral phases as efﬁcient tools fordiastereo- Journal of Pharmaceutical and Biomedical Analysis

(1)

ContentslistsavailableatScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

jo u r n al ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / j p b a

Polysaccharide-based chiral stationary phases as efﬁcient tools for diastereo- and enantioseparation of natural and synthetic Cinchona alkaloid analogs

Attila Bajtai

^a

, István Ilisz

^a,∗

, Róbert Berkecz

^a

, Ferenc Fülöp

^b

, Wolfgang Lindner

^c,∗

, Antal Péter

^a

aInstituteofPharmaceuticalAnalysis,InterdisciplinaryExcellenceCentre,UniversityofSzeged,H-6720,Szeged,Somogyiutca4,Hungary

bInstituteofPharmaceuticalChemistry,InterdisciplinaryExcellenceCentre,UniversityofSzeged,H-6720,Szeged,Eötvösutca6,Hungary

cDepartmentofAnalyticalChemistry,UniversityofVienna,Währingerstrasse38,1090,Vienna,Austria

a rt i c l e i nf o

Articlehistory:

Received14September2020

Receivedinrevisedform19October2020 Accepted20October2020

Availableonline29October2020

Keywords:

HPLC

Cinchonaalkaloids Diastereoseparation Enantioseparation

Polysaccharide-basedchiralstationary phases

a b s t ra c t

Inthisstudy,wepresentresultsobtainedonthediastereo-andenantioseparationofsomebasicnatural andsyntheticCinchonaalkaloidanalogsbyapplyingliquidchromatographic(LC)andsubcriticalﬂuid chromatographic(SFC)modalitiesonamyloseandcellulosetris-(phenylcarbamate)-basedstationary phasesusingn-hexane/alcohol/DEAorCO2/alcohol/DEAmobilephasesystems.Sevenchiralstationary phasesintheirimmobilizedformwereemployedtoexploretheirstereoselectivityforaseriesofclosely relatedgroupofanalytes.ThemostimportantcharacteristicsofLCandSFCsystemswereevaluated throughthevariationoftheappliedchromatographicconditions(e.g.,thenatureandcontentofthealco- holmodiﬁer,theconcentrationofadditives,temperature).ThecolumnsChiralpakICandIGturnedoutto bethebestinbothLCandSFCmodalities.Temperature-dependencestudyindicatedenthalpy-controlled separationinmostcases;however,separationcontrolledbyentropywasalsoregistered.

1. Introduction

Both natural and synthetic Cinchonaalkaloids have a rather complexstructuralpatternwithmorethanthirtyrepresentatives, amongthesequinine(QN),quinidine(QD),cinchonidine(CD),and cinchonine(CN)representthemaincomponents[1].Inadditionto thepharmaceuticalrelevanceofthemainalkaloidsandderivatives thereof,theymayalsoserveascatalystsinstereo-directedorganic synthesis[2],andaschiralselectorsinthecourseofthedevelop- mentofchiralstationaryphases(CSPs)forhigh-performanceliquid chromatography(HPLC)[3].Besides,quininehasalongtradition tobeappliedasaﬂavorcomponentinbitterbeverages.InTable1 thechemicalstructuresofnaturalandsyntheticCinchonaalkaloids appliedinthisstudyaredepicted.Withtheexceptionofracemic quinine,theotheranalytepairsQN/QD,DHQN/DHQD,CD/CN,and epi-QN/epi-QD arediastereomers toeach other,although often termedpseudo-enantiomers(Table1).QNandQDaswellasCDand CNarediastereoisomerictoeachotherwhich,inprinciple,eases

∗Correspondingauthors.

E-mailaddresses:ilisz@pharm.u-szeged.hu(I.Ilisz), wolfgang.lindner@univie.ac.at(W.Lindner).

their separation. Analytically, HPLC methods are often applied forthispurposeusingpreferentiallynon-chiralstationaryphases, assummarized earlierbyMcCalley [4]. Recently, capillaryelec- trophoretic[5]andHPLC-based[1,6]methodshavebeendescribed fortheseparationofthefourmajorCinchonaalkaloidsQN,QD,CD, andCNaswellasdihydroquinine(DHQN)anddihydroquinidine (DHQD).Alongthisline,thequantitativedeterminationofsixmajor alkaloidsimplementingsupercriticalﬂuidchromatography(SFC) hasrecentlyalsobeenreported[7].

Inprinciple,itisnotnecessarytouseCSPstoresolvethegiven fourpairsofdiastereoisomers(seeTable1).Nevertheless,itisstill ofinteresttoexaminetheirstereoselectivemolecularrecognition patterninthecontextoftheresolution oftheseanalytes.How- ever,fortheresolutionofenantiomericanalytes,theapplication ofaCSPismandatory.Aswehavehadaccesstoracemicquinine (rac.QN)producedthrougha novel,fullysyntheticway[8], we hadtheopportunitytoinvestigatethechromatographicresolu- tionofrac.QNversusthediastereomericpairQN/QDandoftheir epimers(9-epi-QN,9-epi-QD).Therefore,theimpactofthechiral environmentofthediversepolysaccharide-typeCSPsontheoverall stereoselectivityparameterscouldbeexplored.

https://doi.org/10.1016/j.jpba.2020.113724

(2)

Table1

StructureofnaturalandsyntheticCinchonaalkaloids.

Name Abbreviation Conﬁguration Structure

(+)Quinine (+)QN (+)(1R,3S,4R,8R,9S)

(-)Quinine (-)QN (-)(1S,3R,4S,8S,9R)

(+)Quinidine (+)QD (+)(1S,3R,4S,8R,9S)

(-)9-epi-Quinine (-)9-epi-QN (-)(1S,3R,4S,8S,9S)

(+)9-epi-Quinidine (+)9-epi-QD (+)(1S,3R,4S,8R,9R)

(-)Dihydroquinine (-)DHQN (-)(1S,3R,4S,8S,9R)

(+)Dihydroquinidine (+)DHQD (+)(1S,3R,4S,8R,9S)

(-)Cinchonidine (-)CD (-)(1S,3R,4S,8S,9R)

(+)Cinchonine (+)CD (+)(1S,3R,4S,8R,9S)

InanearlierpaperHoffmannetal.[9]studiedthestereoselective resolutionofQN/QDandCD/CNonachiralcationexchangerfol- lowedbyanotherrecentinvestigationbyBajtaietal.[10]applying chiralzwitterionicCSPs.Intermsofintermolecularlydriveninter- actionmechanismbetweenthechiralselector(SO)motifsofthe CSPsandthechiralanalytes(selectands,SAs),theionexchanger- type CSPs investigated previously and the polysaccharide-type CSPs examined in the present study (see Fig. S1) belong to structurallyentirelydifferentCSPswithrespecttoretentionand stereoselectivitycharacteristics.Inthisstudy,weaimedtotrace thespeciﬁcsofthepolysaccharideCSPsusedinnormal-phaseliquid chromatography(NP LC)andinSFCmodalitiesfortheseparation ofacloselyrelatedgroupofCinchonaalkaloids(seeTable1).

Polysaccharide-based PS-CSPs usually displaya great variety ofenantio-,diastereo-,andchemoselectivity[11–17].Avarietyof robustCSPswithverywidespectraofapplicationswereobtained viatheimmobilizationofsubstitutedamylose-andcellulose-based selectors[14,12–17].However,nodatawerefoundintheliter- aturefor theseparationofsomenaturalandsyntheticCinchona alkaloidsandtheirderivativesonPS-CSPs.Themainobjectiveof thepresentpaperistorevealsomegeneraltendenciesofasetof prominentPS-CSPs (Fig.S1)forthediastereo-andenantiosepa- rationofthesetofchiralanalytes(Table1)underNP-LCandSFC conditions.TheinvestigationofstereoselectivitycriteriaofSOsand SAswasinthefocusofthisstudy.Morespeciﬁcally,ofthethree isobaricpairsofstereoisomers,QN/QDand9-epi-QN/9-epi-QDare diastereomerictoeachother,characterizedbythechangeofonly thestereogeniccentersofC-8and C-9,whilethe1S,3R,and4S stereogeniccentersofthequinuclidineringresidueremaincon- stant. Forthetruly racemicQN, allﬁvestereogeniccenters are oppositetoeachother.ForthetwodiastereomericDHQN/DHQD andCD/CNpairs,thediastereomericbehavioroftheanalytesisthe sameasforQN/QD.SeeTable1formoredetails.

Methodological andexperimentalfactors,suchasthenature and theconcentrationofthedifferentmodiﬁers(alcohol,water, acidorbase)andmobilephasecompositionsinNP-LCandSFC,as

wellasthestructureofPS-CSPsandSAsandthetemperaturewere evaluatedonretention,selectivity,andresolutionofstereoisomers withspecialattentionwithrespecttoelutionsequences.

2. Materialsandmethods 2.1. Chemicalsandreagents

(–)QN, (+)QD, and (–)-1011-(–)DHQN were purchased from Buchler(Braunschweig, Germany). (+)-1011-DHQD, (+)-CN, and (–)-CDwere fromSigma-Aldrich(Vienna,Austria). RacemicQN [1:1 mixtureof (–)QN and (+)QN]wasa generous giftfrom N.

Maulidesynthesizedasdescribedin[8].C9-Epiquinine(–)epi-QN andC9-epiquinidine(+)epi-QDweresynthesizedasdescribedin [18](Table1).

Methanol(MeOH),ethanol(EtOH),andn-hexaneofHPLCgrade werepurchased fromVWRInternational(Radnor,PA,USA).The alcoholadditives1-propanol(1-PrOH),2-propanol(2-PrOH),the base additivediethylamine (DEA), and the acid additive acetic acid(AcOH)allanalyticalreagentgrades,werefromVWR.Liquid CO2 wasfromMesser(Budapest,Hungary).Ultrapurewaterwas obtainedfromUltrapureWaterSystem,PuranityTUUV/UF(VWR International).

2.2. Apparatusandchromatography

Threechromatographic systems were applied in this study.

TheﬁrstonewasaWatersBreezeapparatusconsistingofa1525 binary pump, a 487 dual-channel absorbance detector, a 717 plusautosampler,and acolumn thermostat.For datacollection Empower2datamanagersoftware(WatersCorporation,Milford, MA,USA)wasapplied.ALaudaAlphaRA8thermostat(LaudaDr.R.

WobserGmbh,Lauda-Königshofen,Germany)wasusedtoregulate columntemperature.

ThesecondliquidchromatographicsystemwasfromShimadzu, andit containeda low-pressurequaternary pump(LC-20AD), a

(3)

photodiodearraydetector(SPD-M20A),aModel7125injectorwith a 20-␮lloop (Rheodyne,Cotati,CA,USA), and LCSolutiondata acquisitionsystem(ShimadzuCorporation,TokyoJapan).Allexper- imentsinnormal-phasemode(NP)werecarriedoutunderisocratic conditionsataﬂowrateof0.6mLmin^–1andatacolumntemper- atureof25^◦C(ifnototherwisestated).

Thethirddevice,aWatersAcquityUltraPerformanceConver- genceChromatography^TM(UPC²,WatersCorporation)systemwas appliedforSFCstudieswithabinarysolventpump,anautosam- pler,abackpressureregulator,acolumnoven,andaphotodiode arraydetector.AnEmpower2softwarewasusedtosystemcontrol anddataacquisition.IneverycaseSFCwasperformedinisocratic modeataﬂowrateof2.0mLmin^–1andacolumntemperatureof 40^◦C(ifnototherwisestated).Theoutletpressurewassetat150 bar.ThemobilephasesappliedinSFCconsistedofliquidCO2and MeOH,EtOH,1-PrOHor2-PrOHindifferentratios(v/v)containing differentacidandbaseadditives.

Stocksolutionsoftheanalyteswerepreparedbydissolvingthe solidCinchonasamplesinMeOHor2-PrOHin1.0mgmL^–1concen- trationandfurtherdilutedwhennecessary.Aninjectionvolume of20␮LwasappliedinLCand7␮LinSFC.InLCthedead-time ofcolumns(t₀)wasdeterminedbyinjectionoftri-t-butylbenzene, whileinSFCmodetheﬁrstnegativesignalbyinjectingMeOHwas used.AnalytesweredetectedbytheirUVabsorptionat215–230 nm.

Polysaccharide-based columns amylose tris-(3,5- dimethylphenylcarbamate) (Chiralpak IA), amylose tris-(3-chlorophenylcarbamate) (Chiralpak ID), amylose tris-(3,5-dichlorophenylcarbamate) (Chiralpak IE), amylose tris-(3-chloro-4-methylphenylcarbamate) (Chiralpak IF), and amylosetris-(3-chloro-5-methylphenylcarbamate)(Chiralpak IG) aswellascellulosetris-(3,5-dimethylphenylcarbamate)(Chiralpak IB) and cellulose tris-(3,5-dichlorophenylcarbamate) (Chiralpak IC)allwiththesamesize(250mm×4.6mmI.D.,5-␮mparticle size)weregenerousgiftsfromChiralTechnologiesEurope(Illkirch, France).AllCSPsemployedinthisstudybelongtotheimmobilized PS-typecolumns.ThestructuresofselectorsarepresentedinFig.

S1.

3. Resultsanddiscussion

3.1. EffectsofmobilephasecompositioninNP-LCandinSFC

InthecaseofPS-CSPs,thegenerallyacceptedrecognitionmech- anismis basedontheinclusion ofchiral solutesintothechiral cavities ofthepolysaccharide-typeselectordrivenbyadditional attractiveforcessuchasH-bonding,dipole–dipoleand␲–␲inter- actions. Inaddition,therole ofsteric“hindrance” forthegiven SAs toenterdeep intothechiral grovesshouldalsobeconsid- ered[11–17].Toregulatetheoverallchromatographicretention, thenatureandconcentrationofanalcoholmodifierinbothnor- mal phase LC (NP LC) [19] and SFC [20] are often varied. To explorethepossibleeffectsofthealcoholmodifierinNP LCand SFC, two columnswere selected: the cellulose-based Chiralpak IC column (as the mosteffective CSP in thescreening process fortheinvestigatedconformationallyrestrictedanalytes)andthe amylose-basedChiralpakIEcolumn,bothpossessingthesamecar- bamatemodification(tris-3,5-dichlorophenylcarbamatemoiety)of thetwodifferentpolysaccharidebackbones.

Withvariationofthenatureofthealcoholforthestudiedana- lytes on the7 CSPsin NP-LC modality, a relatively similarbut slightincreaseink1wasregisteredintheEtOH<1-PrOH<2-PrOH sequencewiththeexceptionofepi-QN/epi-QD,whichwasretained moresigniﬁcantly(Fig.S2).Thebeststereoselectivityperformances (higher˛andR_S)couldgenerallybeachievedwith2-PrOHonChi-

ralpakIC.InSFCmodalityonthesamecolumns,slightlyhigherk₁ valueswereobtainedonmobilephasescontainingMeOHand2- PrOHcomparedtoEtOH(especiallyforepi-QN/epi-QD).For˛and R_SthevariationsobservedwerequitesimilartothoseinNP-LC;

namely,ChiralpakIEexhibitedmuchlesseffectivenessthanChi- ralpakIC.Asexpected,underNP-LCandSFCconditionsanincrease oftheapolarcharacterofthealcohol,i.e.,applyingalcoholswith alongerorbranchedchainusuallyresultedinanenhancedana- lyteretention,selectivity,andresolution,especiallyonChiralpak IC(Fig.S2).Itshouldbementioned,thatoppositebehaviorshad alsobeenreportedintheliterature[21,22].Thechangeinenantio- anddiastereoselectivityresultingfromchangingthealcoholmodi- ﬁerwaspreviouslyrationalizedasaresultofalterationofthesteric environmentofthechiralcavitiesbydifferentalcoholmodiﬁers dueto intra-andinter-molecular solvation effects [22].On the basisoftheseresults,furtherexperimentswerecarriedoutwith theapplicationofEtOHand2-PrOHinNP-LCaswellasMeOHand 2-PrOHinSFCconditions.

Foramorethoroughstudyoftheeffectsofalcoholconcentration onchromatographic parameters in NP LC,n-hexane/EtOH/DEA and n-hexane/2-PrOH/DEA (60/40/0.1–95/5/0.1 v/v/v) mobile phaseswereapplied.UnderSFCconditions,theMeOHand2-PrOH contentinliquidCO₂wasvariedfrom10to60v%(with20mM DEAinallexperiments).Theamylose-andcellulose-basedcolumns appliedearlierwiththesameselector(ChiralpakICandChiralpak IE,seeFig.S1)wereselectedforthisstudy.Regardingtheretentive characteristics,atypicalNPbehaviorwasobservedforbothNP LC andSFCmodalities:increasingtheratioofapolarn-hexaneorCO2

resultedinanincreasedk₁,especiallyfortheepi-QN/epi-QDpair (Fig.1).Itisnoteworthy,thatwiththeincreaseofthemobilephase polarity,thestrengthoftheH-bondsbetweentheanalytesandthe selectordecreasesandthesolubilityoftheanalytesinthemobile phaseincreases[23].

On Chiralpak IEthe stereoselectivity exhibited only a small enhancementinbothNP-LCandSFC modalitieswithincreasing n-hexaneorCO₂ content.However,onChiralpakICinNP-LCin n-hexane/EtOH/DEAmobilephaseaslightincrease,whereasinn- hexane/2-PrOH/DEAmobilephaseamoderateenhancementinthe

˛valuewasregistered.InSFConChiralpakIC˛generallyslightly decreasedordidnotchangesigniﬁcantlywithincreasingCO₂content.Theepi-QN/epi-QDpair,again,wasanexceptionshowinga slightincreaseonChiralpakICinCO₂/MeOH.Itshouldbenoted, thatslightlyhigher˛valueswereregisteredinCO₂/2-PrOHthanin CO2/MeOHmobilephases.

RegardingR_S values, in both NP-LC and SFC, theyincreased signiﬁcantly onChiralpak IC and slightly on Chiralpak IEwith increasingofn-hexaneorCO2 content(althoughanunexpected exceptionwasfoundinSFCmodalityforDHQN/DHQDonChiralpak IE,Fig.1).Itisworthmentioningthatthechangeinthechromato- graphic performance caused by the alcohol modiﬁer depended on the structure of the chiral selector as well; the cellulose- basedselectorfortheinvestigatedbasicanalytesoutperformedthe amylose-basedone.

Thealcoholmaybeincorporatedintothepolysaccharidestruc- ture, either into the cavities or between the polymer chains, affectingthetertiarystructureofthechiralpolymeritselfviasol- vationeffects[24].UnderLCconditions,themainadsorbingsites areconsideredtobethepolarcarbamateresidues[25] andthe differentinvolvementoftheNHandCOgroupsintheH-bonding processwerefoundtoberesponsibleforthedifferencesobserved inthestereoselectiveandenantioselectivebindingprocess[26].It isimportanttonote,thatinthisstudy10–60v%ofalcoholicmodi- ﬁerwasemployedunderSFCconditions,i.e.,SFCisoperatedunder subcriticalconditions,wheresigniﬁcantdeviationsduetothedif- ferenceofthesetvaluesandactualoperationalconditionscannot beexpected[27].Besidesaffectingthephysicalpropertiesofthe

(4)

Fig.1.Effectofmobilephasecompositiononchromatographicparameters,retentionfactor(k),separationfactor(˛)andresolution(RS)fortheseparationofCinchona alkaloidanalogsonChiralpakICandIEcolumnsinNP-LCandSFCmodalities.

Chromatographicconditions:columns,ChiralpakICandChiralpakIE;mobilephase,fornormalphase,n-hexane/EtOH/DEAandn-hexane/2-PrOH/DEA(60/40/0.1–95/5/0.1 v/v/v);forSFC,CO2/MeOH(40/60–90/10v/v)containing20mMDEAandCO2/2-PrOH(40/60–80/20v/v)containing20mMDEA;ﬂowrate,1.0mLmin^–1;detectionat220 nm;temperature,25^◦C;symbolsforChiralpakIC,䊏,QN/QD, ,DHQN/DHQD,,epi-QN/epi-QD,,CD/CNand ,rac.QN,forChiralpakIE,–,QN/QD;,DHQN/DHQD;, epi-QN/epi-QD;,CD/CNand♦,rac.QN.

eluent (e.g.,polarity,viscosity,anddensity),analterationofthe adsorptionlayershouldalsobeconsideredunderSFCconditions [27].

Besides thenatureand concentrationof thealcohol compo- nents,otheradditives(e.g.,acid,base,water)arealsofrequently appliedtomodify“chiral”resolutionondifferenttypesofCSPs.In SFC,beyondimprovementinpeakshape,additivesmayalsohave animpactonretention[28]andonthenumberoftheoreticalplates [29].Fig.2depictstheeffectsofAcOH(A)andH₂O(B)additiveson

thechromatographicperformanceofChiralpakICwitheluentsys- temsCO2/MeOH(60/40v/v)containing20mMDEA,and0.0–40 mMAcOH(A)or0.0–5.0v%H₂O(B)applyingtheQD/QNpairof diastereomersastestcompounds.Uponincreasingtheconcentra- tionoftheadditiveretentiondecreasedincaseofAcOHandslightly increasedincaseofwaterinparallelwithaminordecreaseof˛ andamarkedreductioninR_Svalues(Fig.2),whilethechangein theefﬁciencywasusuallyintherangeof10–15%(datanotshown).

Theslightchangecausedbytheadditionofwaterwasprobably

(5)

Fig.2.EffectofAcOHandH2OcontentonthechromatographicperformanceofChiralpakICforanalyteQD/QNunderSFCconditions.

Chromatographicconditions:mobilephase,CO2/MeOH(60/40v/v)containing20mMDEAandA,0mM,20mMand40mMAcOH;B,0.0v%,1.0v%,3.0v%and5.0v%H2O;

ﬂowrate,1.0mLmin^–1;detectionat220nm;temperature,25^◦C.

duetothelesspolarcharacteroftheCSPstudiedasobservedby Armstronget.al[29].

3.2. Effectofthestructureofpolysaccharide-typeselectors

The stereoselectivity characteristics of the studied Cinchona alkaloidsonthesevenPS-CSPswerestudiedwithdifferentmobile phase systems, and thecorresponding dataare summarized in Tables2–5.Duetothedifferenceinlinkageofglycopyranosemoi- eties inamylose-andcellulose-based selectors,theinteractions betweenanalyte andselectormaychangeresulting indifferent chromatographic behaviors,which mayeven affectthe elution orderoftheresolveddiastereomersandenantiomers.Thecontri- butionof thepolysaccharidebackbone canbeevaluatedbythe comparison of thechromatographicdata obtainedonthesame carbamateresidue;i.e.,tris-(3,5-dimethylphenylcarbamate)(Chi- ralpakIAvs.ChiralpakIB)andtris-(3,5-dichlorophenylcarbamate) (Chiralpak IE vs. Chiralpak IC), linked to amylose or cellulose,respectively.Tofacilitatethecomparisonbetweendifferent columns,chromatographicconditionswerekeptconstant,applying then-hexane/EtOH/DEAandthen-hexane/2-PrOH/DEA(80/20/0.1 v/v/v)mobilephasecompositionsinNP-LC.Inafewcaseswhena partialresolutionoccurred,theseparationwasfurtheroptimized.

DatasummarizedinTable2showthathigherretentioncouldbe obtainedonamylose-thanoncellulose-basedCSPsinmostcases.

Exceptions,however,doexistespeciallyformobilephasescontain- ing2-PrOH.

Itshouldbenoted,thatinNP-LC,despiteshorterretentiontimes, ChiralpakICprovedtobethemosteffectiveCSPintheseparationof thediastereomericpairsofbothCinchonaalkaloidsandofrac.QN.

InSFCmodalitytheeffectofthepolysaccharidebackbonewasalso evaluatedonthesamepairsofCSPsapplyingCO2/MeOHandCO2/2- PrOH(60/40v/v)mobilephasesallcontaining20mMDEA(Table3).

Regardingretention,theresultswithChiralpakIEandICweresimi- lartothedataobtainedinNP-LC;i.e.,theamylosebackboneoffered

higherretentions.ChiralpakIAandIB,inturn,gaveoppositeresults, withthecellulose-basedCSPprovidinghigherretentions.Among theabove-mentionedfourcolumnsunderSFCconditions,thebest separationswereachievedwithChiralpakIC,similartothosefound inNP-LC.

Thefundamentalstructuraldifferencesbetweenamylose-and cellulose-based tris-(3,5-dimethylphenylcarbamate) or tris-(3,5- dichlorophenylcarbamate) were found to be reﬂected in the stereochemicalrecognitionpatternsevenforsomediastereomeric analytes.Reversalofelutionorderbetweenamylose-andcellulose- basedCSPs containingthe same substituents wasregistered in NP-LCcontainingEtOHasmobilephaseforanalyteCN/CDonChi- ralpakIAvs.IB,andinmobilephasescontainingEtOHand2-PrOH onChiralpakIEvs.IC(Table4).Similarbehaviorswereregistered inSFCmodalityforQN/QDandCN/CDonChiralpakIEvs.IC(Table 2B).Examplesofreversedelutionordersofenantiomericanalytes onamylose-orcellulose-basedcolumnshavebeendescribedpre- viously[19].

Applyingconstantchromatographic conditions inNP-LC and SFCmodality,datalistedinTable2and3canprovideopportunity toevaluatetheeffectofthenatureofthesubstitutedphenylcar- bamatemoiety. BycomparingthedataobtainedwithChiralpak IAvs.IE(bothareamylose-basedCSPs),andIB vs.IC(bothare cellulose-basedCSPs),higherretentionscanclearlybeidentiﬁed forallanalytesonCSPswithtris-(3,5-dichlorophenylcarbamate) moiety. The higher ˛ and R_S values, observed generally in the case of CSPs with tris-(3,5-dichlorophenylcarbamate), suggest morepronouncedSO–SAinteractionsofthestudiedanalytes.This maybe attributed to a ␲–␲-typeinteraction increment of the acidicphenylcarbamateandthe␲-basic-typequinolinemoietyof Cinchonaalkaloids.However,inafewcases,inparticularforepi- QN/epi-QD,lower˛andR_S wereregisteredonChiralpakIEand ChiralpakICthanonChiralpakIAandChiralpakIB.Nevertheless, resultsshowedthatthetris-(3,5-dichlorophenylcarbamate)CSPs, inmostcases,outperformotherSOtypes.

(6)

Table2

EffectofEtOHand2-PrOHcontentinn-hexaneasbulksolventforthechromatographicdata,k1,˛,RSofCinchonaalkaloidsinnormalphasemodality.

Column Mobilephase k1,˛,RS QN/QD rac.QN DHQN/DHQD epi-QN/epi-QD CD/CN

IA

a

k1 0.59 0.59 0.60 2.47 0.76

˛ 1.18 1.00 1.00 1.00 1.45

RS 1.02 0.00 0.00 0.00 2.47

b

k1 0.60 0.64 0.62 2.45 0.77

˛ 1.19 1.00 1.00 1.00 1.00

RS 0.70 0.00 0.00 0.00 0.00

IB

a

k1 0.51 0.52 0.45 1.15 0.56

˛ 1.00 1.00 1.00 1.28 1.22

RS 0.00 0.00 0.00 1.12 0.47

b

k1 1.02 1.01 0.99 2.72 0.99

˛ 1.00 1.00 1.00 1.46 1.00

RS 0.00 0.00 0.00 1.26 0.00

IE

a

k1 7.12^∗∗ 1.45 1.40 4.57 3.00^∗

˛ 1.06^∗∗ 1.00 1.00 1.00 1.10^∗

RS 1.16^∗∗ 0.00 0.00 0.00 1.40^∗

b

k1 1.48 1.89 2.10 9.15 1.42

˛ 1.00 1.00 1.00 1.00 1.13

RS 0.85 0.00 0.00 0.00 1.25

IC

a

k1 0.73 0.82 0.86 3.61 0.83

˛ 1.23 1.19 1.20 1.00 1.22

RS 1.18 0.84 1.27 0.00 1.22

b

k1 1.05 1.02 1.09 15.48 0.97

˛ 1.52 1.56 1.48 1.00 1.80

RS 3.07 2.67 2.88 0.00 4.20

ID

a

k1 0.68 0.68 0.69 2.68 0.76

˛ 1.21 1.08 1.16 1.14 1.39

RS 1.49 0.22 0.94 1.62 2.55

b

k1 0.80 0.97 1.08 5.37 0.99

˛ 1.13 1.00 1.00 1.14 1.00

RS 0.18 0.00 0.00 1.45 0.00

IF

a

k1 0.99 0.99 0.96 4.17 1.15

˛ 1.00 1.00 1.00 1.00 1.18

RS 0.00 0.00 0.00 0.00 1.22

b

k1 1.37 1.41 1.53 5.73 1.18

˛ 1.00 1.00 1.00 1.03 1.24

RS 0.00 0.00 0.00 0.27 1.00

IG

a

k1 0.98 0.98 0.98 5.02 1.36

˛ 1.27 1.00 1.10 1.08 1.22

RS 2.27 0.00 0.22 1.25 2.14

b

k1 1.06 1.07 1.08 6.64 1.16

˛ 1.10 1.00 1.18 1.09 1.18

RS 0.46 0.00 0.48 1.28 1.64

Chromatographic conditions:columns, ChiralpakIA-IG; mobilephase, a, n-hexane/EtOH/DEA (80/20/0.1v/v/v), ^∗ n-hexane/EtOH/DEA (90/10/0.1v/v/v) and^∗∗n- hexane/EtOH/DEA(95/5/0.1v/v/v),b,n-hexane/2-PrOH/DEA(80/20/0.1v/v/v);ﬂowrate,1.0mLmin⁻¹;detection,230−250nm;temperature,25^◦C.

As mentioned earlier, the structure of SO may affect the sequenceofelutiontoo.Inthisstudy,areversalofelutionsequence was registered in NP-LC modality in EtOH-containing mobile phasesforanalytesQN/QDusingChiralpakIAvs.IE(Table4).The reversalofelutionsequencebychangingthechemicalstructure ofsubstituentsonthetris-(phenylcarbamate)moietywasalready indicatedinearlierpublications[19,30,31].

Theeffectofthepositionofthesubstituentsofthecarbamate moietyofthePS-CSPsonthechromatographicperformanceforthe givenanalyteswasinvestigatedbycomparisonofchromatographic dataobtainedonChiralpakIF,IG,andIDcolumnsinbothNP-LC andSFCmodalities(seeFig.1S).Ingeneral,higherretentionand better˛andR_SvalueswereobtainedonChiralpakIGthanonIF.

ChiralpakIDseemstobeasefﬁcientasChiralpakIGinNP-LCmodal- ityinthepresenceofEtOHasbulksolventcomponent,andless efﬁcientinallothercases.Thesecondarystructureofthetris-(3- chloro-5-methylphenylcarbamate)-basedCSP(ChiralpakIG)offers strongerretentiveinteractionswiththeanalytes.Regardingelu- tionsequencesofQN/QDandrac.QN,nochangewasregistered byalteringtheseCSPs.Incontrast,forepi-QN/epi-QDandCD/CNa reversalofelutionsequencewasregisteredinNP-LCinthepres- enceof2-PrOHaseluentconstituent.Thestrongdependence of thesequenceofelutionasafunctionofthemobilephase(eluent compositionandSOstructure),observedinallcases,drawsatten-

tiontotheimportanceofidentiﬁcationofeachpeakinthecase ofPS-CSPs.Forchiralion-exchangersinvestigatedpreviously,this scatteredtendencywasnotseen[10].

Intheenantio-anddiastereoseparationoftheinvestigatedCin- chona alkaloids in SFC modality, Chiralpak IC and IG columns performedsigniﬁcantlybetter.However,inNP-LC,the3-chloro- substitution in Chiralpak IC and ID promotes an increased stereodiscrimination but with lower effectiveness than in SFC modality.

Theenantio-anddiastereoselectivitiesofthesevenchiralselec- tors for the ﬁve pairs of Cinchona alkaloid stereoisomers are summarizedinTables4and5.Fortheseparationofenantiomers, consistencieshavebeennoticednumeroustimeswithanelution sequence(+)QN<(–)QN,butareversedelutionsequencewasreg- isteredonChiralpakIDandIFinSFCmodalityapplying2-PrOHas alcoholmodiﬁer.

Thesituationbecomesevenmorecomplicatedfortheresolu- tionofthediastereomericpairsQN/QD,DHQN/DHQD,CD/CN,and epi-QN/epi-QD,whereacleartrendcannotbeseen.Inboth NP- LCandSFCmodalitiesonthesevencolumns,theelutionsequence QN<QD,DHQN<DHQD,andCD<CNanditsreversalcanalso beobserved, as a clear indicationof thedifﬁculty to interpret enantioselectivityvsdiastereoselectivity(Tables2and3).Unex- pectedreversalsoftheelutionorderofthediastereomeric(often

(7)

Table3

ComparisonoftheeffectofMeOHand2-PrOHcontentinCO2asbulksolventforthechromatographicdata,k1,˛,RSofCinchonaalkaloidsinSFCmodality.

Column Mobilephase k1,˛,RS QN/QD rac.QN DHQN/DHQD epi-QN/epi-QD CD/CN

IA

a

k1 0.38 0.33 0.37 0.83 0.39

˛ 1.21 1.38 1.46 1.11 1.20

RS 1.16 1.91 2.43 0.74 1.03

b

k1 0.30 0.29 0.33 1.16 0.42

˛ 1.41 1.00 1.23 1.00 1.26

RS 1.43 0.00 0.79 0.00 0.65

IB

a

k1 0.41 0.41 0.40 1.47 0.50

˛ 1.00 1.00 1.00 1.00 1.07

RS 0.00 0.00 0.00 0.00 0.21

b

k1 0.52 0.52 0.54 1.55 0.69

˛ 1.00 1.00 1.00 1.18 1.00

RS 0.00 0.00 0.00 0.53 0.00

IE

a

k1 3.20 3.20 3.81 7.57 3.21

˛ 1.00 1.00 1.07 1.00 1.04

RS 0.00 0.00 0.91 0.00 0.32

b

k1 2.28 2.27 3.25 4.83 2.42

˛ 1.17 1.05 1.00 1.05 1.17

RS 1.68 0.40 0.00 0.28 1.79

IC

a

k1 1.87 1.66 1.89 7.52 2.19

˛ 1.22 1.38 1.43 1.09 1.47

RS 3.18 5.14 5.71 1.36 6.34

b

k1 0.90 0.80 0.97 13.86 1.23

˛ 1.53 1.73 1.66 1.00 1.44

RS 4.62 5.69 5.47 0.00 4.48

ID

a

k1 0.95 0.91 1.14 1.74 0.89

˛ 1.08 1.00 1.00 1.00 1.08

RS 0.91 0.00 0.00 0.00 0.73

b

k1 0.71 0.71 1.01 2.63 0.88

˛ 1.31 1.08 1.08 1.00 1.00

RS 2.89 0.00 0.83 0.00 0.00

IF

a

k1 1.14 1.08 1.27 2.95 1.12

˛ 1.00 1.00 1.00 1.00 1.08

RS 0.63 0.00 0.00 0.00 0.76

b

k1 0.70 0.72 0.85 2.62 0.91

˛ 1.34 1.13 1.28 1.00 1.13

RS 1.89 0.68 1.56 0.00 0.74

IG

a

k1 1.26 1.03 1.45 2.08 0.99

˛ 1.12 1.38 1.45 1.33 1.46

RS 1.32 3.84 4.68 3.81 4.37

b

k1 0.71 0.72 1.18 2.92 0.87

˛ 1.87 1.00 1.49 1.18 1.47

RS 5.98 0.00 4.20 2.49 3.82

Chromatographicconditions:columns,ChiralpakIA-IG;mobilephase,a,CO2/MeOH(60/40v/v)containing20mMDEA,b,CO2/2-PrOH(60/40)containing20mMDEA;ﬂow rate,2.0mLmin⁻¹;detection,215−230nm;temperature,40^◦C;backpressure,150bar.

termedpseudo-enantiomeric)pairscaneasilyhappenasafunction ofthecompositionofthemobilephase.Inthiscontext,itbecame particularlyinterestingthatonChiralpakIGapplying2-PrOHas eluentcomponentunderLCconditionstheelutionorders(+)DHQD

<(–)DHQNandCN<CDchangedinSFCmodalityto(–)DHQN<

(+)DHQDandCD<CN(Table4and5).Thisisanotherstrongindi- cation abouttherole of solvation oftheselector andselectand moietiesontheoveralldiastereoselectivity.

Representative chromatogramsfor theresolution of racemic quinineandfourdiastereomersofCinchonaalkaloidsinSFCmode aredepictedinFig.3andFig.S3onChiralpakICandIGapplying 2-PrOHandMeOHasalcoholmodifiers.Fullseparationandiden- tificationsoffivepairsofenantiomersanddiastereomerscanonly beachievedintwochromatographicruns.

3.3. Effectofthestructureofanalyte

Analytes, except DHQN/DHQD, possess an unsaturated side chain(vinylversusethylgroup)onthequinuclidinemoiety,which mayinﬂuenceinteractionsbetweenSAandCSP,despitesmalldif- ferencesinsizeandpolarity.SurveyingthedatainTables2and3 revealed, that,in general,at a givenmobile phasecomposition retentionsinNP LCorSFCmodalitiesdifferonlyveryslightly.In somecases,however,thedifferenceismoresigniﬁcant.Asaresult,

Fig.3.Chromatogramsforseparationofdiastereomersandenantiomersofnatural andsyntheticCinchonaalkaloidanalogs.

Chromatographicconditions:columns,A,ChiralpakIC;mobilephase,CO2/2-PrOH (70/30v/v)containing20mMDEA,B,ChiralpakIG;mobilephase,CO2/MeOH(70/30 v/v)containing20mMDEA;ﬂowrate,1.0mLmin^–1;detectionat220nm;temperature,25^◦C.

(8)

Bajtai,I.Ilisz,R.Berkeczetal.JournalofPharmaceuticalandBiomedicalAnalysis193(2021)113724

Table4

ElutionsequencesofCinchonaalkaloidsinn-hexaneasbulksolventcontainingEtOHor2-PrOHinnormalphasemodality.

Mobilephase:n-hexane/EtOH(80/20v/v)containing0.1%DEA

Column IA IB IE IC ID IF IG

QN/QD (-)QN<(+)QD – (+)QD<(-)QN (+)QD<(-)QN (-)QN<(+)QD – (-)QN<(+)QD

rac.QN – – – (+)QN<(-)QN (+)QN<(-)QN – –

DHQN/DHQD – – – (+)DHQD<(-)DHQN (-)DHQN<(+)DHQD – (-)DHQN<(+)DHQD

epi-QN/epi-QD – (+)epi-QD<(-)epi-QN – – (-)epi-QN<(+)epi-QD – (-)epi-QN<(+)epi-QD

CD/CN CD<CN CN<CD CD<CN CN<CD CD<CN CD<CN CD<CN

Mobilephase:n-hexane/2-PrOH(80/20v/v)containing0.1%DEA

QN/QD (-)QN<(+)QD – (-)QN<(+)QD (+)QD<(-)QN (-)QN<(+)QD – (-)QN<(+)QD

rac.QN – – – (+)QN<(-)QN – – –

DHQN/DHQD – – – (+)DHQD<(-)DHQN – – (+)DHQD<(-)DHQN

epi-QN/epi-QD – (+)epi-QD<(-)epi-QN – – (-)epi-QN<(+)epi-QD (-)epi-QN<(+)epi-QD (+)epi-QD<(-)epi-QN

CD/CN – – CD<CN CN<CD – CD<CN CN<CD

Chromatographicconditions:columns,ChiralpakIA-IG;ﬂowrate,1.0mLmin⁻¹;detection,230−250nm;temperature,25^◦C.

Table5

ElutionsequencesofCinchonaalkaloidsinCO2asbulksolventcontainingMeOHor2-PrOHinSFCmodality.

Mobilephase:CO2/MeOH(90/10v/v)containing20mMDEA

Column IA IB IE IC ID IF IG

QN/QD (+)QD<(-)QN – – (+)QD<(-)QN (-)QN<(+)QD – (+)QD<(-)QN

rac.QN (+)QN<(-)QN – – (+)QN<(-)QN – – (+)QN<(-)QN

DHQN/DHQD (+)DHQD<(-)DHQN – (+)DHQD<(-)DHQN (+)DHQD<(-)DHQN – – (+)DHQD<(-)DHQN

epi-QN/epi-QD (-)epi-QN<(+)epi-QD – – (+)epi-QD<(-)epi-QN – – (+)epi-QD<(-)epi-QN

CD/CN CN<CD CN<CD CN<CD CN<CD CN<CD CN<CD CN<CD

Mobilephase:CO2/2-PrOH(80/20)containing20mMDEA

QN/QD (-)QN<(+)QD – (-)QN<(+)QD (+)QD<(-)QN (-)QN<(+)QD (-)QN<(+)QD (-)QN<(+)QD

rac.QN – – – (+)QN<(-)QN (-)QN<(+)QN (-)QN<(+)QN –

DHQN/DHQD (-)DHQN<(+)DHQD – – (+)DHQD<(-)DHQN (-)DHQN<(+)DHQD (-)DHQN<(+)DHQD (-)DHQN<(+)DHQD

epi-QN/epi-QD – (+)epi-QD<(-)epi-QN (+)epi-QD<(-)epi-QN – – – (+)epi-QD<(-)epi-QN

CD/CN CD<CN – CD<CN CN<CD – CD<CN CD<CN

Chromatographicconditions:columns,ChiralpakIA-IG;ﬂowrate,2.0mLmin⁻¹;detection,215−230nm;temperature,40^◦C;backpressure,150bar.

8

(9)

theslightlylesspolarDHQN/DHQDanalytesaremoreretained.This isnotsurprising,butthemarkedincreaseofretentionofthepair ofepi-QN/epi-QDdiastereoisomersisstriking.Theincreasedreten- tionswereaccompaniedwithhigher˛andR_Svaluesonlyinafew cases.Aplausibleexplanationforthisbehaviorismostprobably relatedtotheconfigurationofthefivechiralcentersandtheresult- ingconformationofthestericallyrestrictedanalytes.Namely,for QN/QD,DHQN/DHQD,andCD/CNthechiralcentersofthequinucli- dineringareidentical[(1S),(3R),and(4S)],andonlythetwoother chiralcenters(C-8andC-9carbonatoms)aredifferent[(R)/(S)for QDand(S)/(R)forQN].Thereisasimilarsituationwithrespectto theabsoluteconfigurationofC-9in9-epi-QN[(9S)]and9-epi-QD [(9R)].Thishasastrongimpactontheconformationofmolecules withmultiplechiralcenters.

Inthepresentcase,onlytheoverallretentionbutnotthestere- ochemicaldifferentiationofthediastereomericanalytesisaffected markedlyundertheappliedchromatographicconditions.

3.4. Effectoftemperatureandthermodynamicparameters

The temperature dependence of retention and enantioselec- tivity may provide some valuable information on the chiral recognitionprocess[31–35].Keepinginmindthelimitations of theapproachappliedinthisstudy[33,32–35],thedifferenceinthe changeinstandardenthalpy(H^◦)andentropy(S^◦)forthe enantiomerswerecalculatedonthebasisofthevan’tHoffequation:

ln˛=−(H^◦)

RT +(S^◦) R

Table6

Thermodynamicparameters,(Hô),(Sô),Tx(Sô),(Gô),correlationcoefficients(R²),TisoandQvaluesofCinchona-alkaloidanalogsonChiralpakIA,IB,IC,IE columnsinnormal-phaseandSFCmodalities.

Analyte -(Hô)(kJ/mol) -(Sô)(J/(mol*K) Correlationcoefficients(R²) -Tx(Sô)298K(kJ/mol) -(Gô)298K(kJ/mol) Tiso(ôC) Q NP-LCmodality

ChiralpakIA

QN/QD 1.1 2.3 0.938 0.7 0.4 219 1.6

CD/CN −2.3 −7.9 0.993§ −2.4 0.2 14 0.9

ChiralpakIB

epi-QN/epi-QD 4.5 12.3 0.996 3.7 0.8 95 1.2

ChiralpakIE

QN/QD −0.4 −2.1 0.959 −0.6 0.2 −91 0.7

DHQN/DHQD −0.7 −3.0 0.998^♠ −0.9 0.2 −33 0.8

CD/CN −0.2 −1.55 0.931 −0.5 0.3 −170 0.4

ChiralpakIC

QN/QD 1.9 3.1 0.979 0.9 1.0 345 2.1

rac.QN 1.6 1.8 0.926 0.5 1.1 604 3.2

DHQN/DHQD −0.5 −4.9 0.979^♣ −1.5 1.0 30* 0.3

DHQN/DHQD 1.9 3.1 0.991^♠ 0.9 1.0 2.1

CD/CN −1.0 −8.2 0.963^♣ −2.4 1.4 29* 0.4

CD/CN 1.8 0.9 0.973^♠ 0.3 1.5 6.0

SFCmodality ChiralpakIA

QN/QD a 3.4 9.0 0.990 2.7 0.7 108 1.3

rac.QN 5.4 14.1 0.990 4.2 1.2 111 1.3

DHQN/DHQD 2.4 2.8 0.996 0.8 1.6 584 3.0

epi-QN/epi-QD −1.4 −5.4 0.997 −1.6 0.2 −15 0.9

ChiralpakIB

epi-QN/epi-QD b −4.2 −14.4 0.999 −4.3 0.1 20 0.9

ChiralpakIE

DHQN/DHQD c 0.5 1.0 0.994 0.3 0.2 217 1.7

epi-QN/epi-QD −1.9 −6.3 0.985 −1.9 4.4 24 0.9

ChiralpakIC

rac.QN c 1.0 0.4 0.997 0.1 0.9 >1000 10.0

DHQN/DHQD 1.0 0.4 0.993 0.1 0.9 >1000 10.0

epi-QN/epi-QD −5.5 −18.8 0.980 −5.6 0.1 18 0.9

Chromatographicconditions:column,ChiralpakIA,IB,IC,IE;mobilephase,inNP-LCmodality,n-hexane/2-PrOH/DEA80/20/0.1(v/v/v),inSFCmodality,a,CO2/MeOH80/20 (v/v)containing20mMDEA,b,CO2/MeOH90/10(v/v)containing20mMDEA,c,CO2/MeOH70/30(v/v)containing20mMDEA;inNP-LCmodality,temperaturerange,

§35−50^◦C,^♣7.5−30^◦C,^♠30−50^◦Cand^∗temperatureofpointofintersection(seeFig.S4);flowrate,inNP-LC,1.0mLmin⁻¹,inSFC,2.0mLmin⁻¹;backpressureinSFC,150 bar;detection,215−230nm;Tiso,temperaturewheretheenantioselectivitycancels;Q=(Hô)/Tx(Sô)298K.

whereRistheuniversalgasconstant,TistemperatureinKelvin, and˛istheapparentselectivityfactor.

Inordertoinvestigatetheeffectsoftemperatureonthechro- matographicparameters,avariabletemperaturestudywascarried outforﬁveanalytesinLC-NPandSFCmodalitiesonfourPS-CSPs (ChiralpakIA,IB,IE,andIC)inthetemperaturerange7.5–50^◦C inmobilephasesbasedonn-hexane/2-PrOH/DEA(80/20/0.1v/v/v) andCO2/MeOH(90/10,80/20and70/30v/v)allcontaining20mM DEA.TheexperimentaldataaresummarizedinTablesS1andTable S2.

Inbothmodalitiesonallfourcolumns,theretentioninmost casesdecreasedwithincreasingtemperature.Thetransferofthe analytefromthemobilephase tothestationary phase,in general,isanexothermicprocess.Asaresult,kand˛aswellasR_S decreasewithincreasingtemperature.However,regardingsepa- rationfactorandresolution,severalexceptionswereobserved.On ChiralpakIAforCD/CNinthetemperaturerange30–50^◦Candon ChiralpakICforDHQN/DHQDandCD/CNinthetemperaturerange 7.5–30^◦C,˛(andR_S)increasedwithincreasingtemperature(Table S2).Interestingly,inSFCmodalityforepi-QN/epi-QDonallinves- tigatedCSPs,kdecreased,but˛andRSincreasedwithincreasing temperatureinthestudiedtemperaturerange(TableS2).Apply- ingPS-CSPs,severalexampleswerereportedpreviously forthe increaseofselectivitywithincreasingtemperatureinchromato- graphicsystems[14,19,33,35].Itshouldbenoted,thatseveraltimes inbothmodalities,R_Sincreasedwithincreasingtemperatureinde- pendentlyofthechangeof˛.Thisphenomenoncanbeattributed totheenhancedkineticeffectuponincreasingtemperature.

From the chromatographic data van’t Hoff plots were con- structed.Asageneraltrend,ln˛vs.1/Tcurvesgaveclearlylinear