enzyme by several orthogonal methods Investigating thermal stability based on the structural changes oflactase Biotechnology Reports

(1)

Investigating thermal stability based on the structural changes of lactase enzyme by several orthogonal methods

Márton Király

^a

, Borbála Dalmadi Kiss

^a

, Péter Horváth

^b

, László Drahos

^c

, Arash Mirzahosseini

^b

, Gyula Pálfy

^d,e

, István Antal

^a

, Krisztina Ludányi

^a,

*

aDepartmentofPharmaceutics,FacultyofPharmacy,SemmelweisUniversity,HÅgyesEndreu.7.,1092,Budapest,Hungary

bDepartmentofPharmaceuticalChemistry,FacultyofPharmacy,SemmelweisUniversity,HÅgyesEndreu.7.,1092,Budapest,Hungary

cMSProteomicsResearchGroup,ResearchCentreforNaturalSciences,MagyarTudósokkörútja2.,H-1117,Budapest,Hungary

dLaboratoryofStructuralChemistryandBiology,InstituteofChemistry,EötvösLorándUniversity,PázmányP.sétány1/A,1117,Budapest,Hungary

eProteinModelingGroupHAS-ELTE,InstituteofChemistry,EötvösLorándUniversity,1538,Budapest,P.O.B.32,Hungary

ARTICLE INFO

Articlehistory:

Received20April2021 Accepted24May2021

Keywords:

NMR Proteomics β-Galactosidase Aggregation Massspectrometry

ABSTRACT

Thermalstabilityoflactase(β-galactosidase)enzymehasbeenstudiedbyavarietyofphysico-chemical methods.β-galactosidaseis themain active ingredientof medicationsforlactoseintolerance.Itis typicallyproducedindustriallybytheAspergillusoryzaeﬁlamentousfungus.Lactasewasusedasa model to help understand thermal stability of enzyme-type biopharmaceuticals. Enzyme activity (hydrolyzationoflactose)ofβ-galactosidasewasdeterminedafterstoringthesolidenzymesubstanceat varioustemperatures.Forabetterunderstandingoftherelationshipbetweenstructureandactivity changeswedeterminedthemassandsizeofthemoleculeswithgelelectrophoresisanddynamiclight scatteringanddetectedaggregationprocesses.Abottom-upproteomicprocedurewasusedtodetermine theprimaryaminoacidsequenceandtoinvestigatechangesintheN-glycosylationpatternoftheprotein.

NMRandCDspectroscopicmethodswereusedtoobservechangesinhigherorderstructuresandto revealrelationshipsbetweenstructuralandfunctionalchanges.

1.Introduction

Biologicals produced bybiotechnology are becoming widely used. Qualityassurance and stability testing of these complex protein type molecules demand extensive knowledge of their structure in order to provide effective and safe medications.

Stabilitytestingofimmunoglobulin-basedbiologicalsisrelatively wellestablished,butforotherbiologicals,likelarge,enzymetype drugsithasnotyetbeendeﬁned.Stabilityandqualityassessments areusuallybasedonsimpleactivitytests,withoutanyinformation onpossiblechanges inthestructure orin theposttranslational modiﬁcationsoftheenzyme.Stressmeasurementsoftheenzyme are usually performed in the form of a solution, although the formulationsandtechnologicalprocessesaremainlypresentinthe solidform.

β-galactosidase,alsocalledlactaseisahydrolyticenzymethat decomposeslactose,adisaccharidefoundindairyproducts.Italso catalysesthereverseofthereaction,calledtransglycosylation[1].

β-Galactosidasesofvariousorigins belongtofamiliesofglycosyl hydrolasesproducedbyalmosteverylivingorganism.Thelackor decreasedproductionofthisdigestiveenzymeinthehumansmall intestine results in the inability to properly break down and assimilatemilkcontainingproducts.Thesedairyproductsplayan important role in gastronomy in many parts of the world, irregularityofitsmetabolism(insufficienthydrolysis oflactose) mayresultinaseriousdeteriorationofthequalityoflife[2,3].The prevalence ofthe lactoseintoleranceis over 60% amongst the humanpopulation,causing health problemsof varyingseverity [4].Thetreatmentofthisdiseaseinvolvestheadministrationof various medications containing lactase. These pharmaceutical products:tablets,chewing-tablets,capsules,solutionsetc.usually containone specific streamof lactase derivedfrom Aspergillus oryzae (Yellow koji mold), also called Tilactase [5]. The beta- galactosidaseproducedbythismedicallyandindustriallyimpor- tantfilamentousfungusisarelativelybigandcomplexglycopro- tein, which contains 1009 amino acids, and has an estimated molecular mass of 109kDa without the post translational modifications[6].

Enzymemoleculeslikelactasehavemanydifferentstructural variantswithdifferentpropertiesandresistancetoexternaleffects

*Correspondingauthor.

E-mailaddress:ludanyi.krisztina@pharma.semmelweis-univ.hu(K.Ludányi).

https://doi.org/10.1016/j.btre.2021.e00637

ContentslistsavailableatScienceDirect

Biotechnology Reports

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

(2)

(pH, temperatureetc.), dependingon its source.[7]. Structural differencesbehindthisphenomenonhavenotbeenexploredyet, butitseemslikelythattheglycanstructuresplayanimportantrole inthestabilityandefﬁciencyoftheprotein[8–11].Enzymetype activeingredientsareusuallysensitivetopH,moisture,pressure, temperature, applied excipients and reactants (metal ions) occurring during manufacture. There is a possibility that they maylosetheireffectivenessbeforeorduringoraladministration, although various technologiestryto overcomethese disadvan- tages[12–15].Forceddegradationstudiesareoftenperformedon active pharmaceutical ingredients to determine and predict possible degradationpathways under variousstress conditions.

Alargevarietyofdegradationprocessesmayresultinthelossof functionorlossofefﬁciency,andtheseareverycomplextoanalyse [16].Currently,therearenoexact[17]guidelinesfortheindustry deﬁninghowtoperformstabilitytestingincase ofproteintype activeingredients.[18].

Despitethepharmaceuticalandcommercialrelevanceofthe lactasecontainingproducts[19],ourknowledgeislimitedabout the relationship between their structure and activity. In the presentstudywewilluseacombinationofanalyticaltoolsinorder todiscoverstructuralchangescausedbystressconditions,which inﬂuence enzyme activity and function. The lactase enzyme in pulverizedformwillbestudied(i.e.,inthesolidform),whichis typicalinandproductionstorage.Bottom-upproteomicprocedure utilizingnanoUHPLC-MSmethodwasusedtoobservethepossible alterations intheprimary aminoacidsequence andchanges in PTMs, such as N-glycan structures. Beside mass spectrometric measurementsSDS-PAGE,DLS,NMR,CDmethodswereappliedto observe thepossiblechanges in theprimary and in thehigher orderedstructuresandtolookforpossibleaggregation.NMRand CDtechniqueswereusedtodeterminealterationsinthesecondary and tertiary structures. Parallel ortho-nitrophenyl-β-galactoside (ONPG)basedactivity studieshavebeenperformedtoestablish structure-activityrelationship.

The main objective of the research is to reveal structural changesofthelactaseenzymeunderheat-stressconditions.Thisis importantbothforunderstandingfundamentalaspectsofenzyme activitychanges under storage,and forpractical considerations suchasimprovingdosing,manufacturing,shippingandstorageof biopharmaceuticals,likelactasecontainingproducts[20].

2.Materialsandmethods 2.1.Materials

The powder of β-galactosidase enzyme (Opti-lactase A-50 powder) derived from Aspergillus oryzae was purchased from Optiferm Gmbh(Oy-Mittleberg,Germany). O-nitrophenyl-D-galactopyranoside (ONPG), MU-Gal (4-Methylumbelliferyl β-D-galactopyranoside)andthebroadrangeSDS–PAGEmolecularweight standardswereobtainedfromVWRInternational(Radnor,PA,US).

ForthetrypticdigestionTrypsinGold,MassSpectrometryGrade (Promega Corporation,Madison, WI, US)and Rapigest(Waters, Milford,MA,US)wereused.Ultrapurewaterobtainedfromthe MilliRO-hMilliQ-system(MerckMillipore,Burlington,MA,US)was usedthroughoutallmeasurements.Allotheringredientslikethe reagents, solvents,eluentswere obtainedfrom SigmaChemical Company(St.Louis,MO,US).

2.2.Stressstabilitytests

Theessenceofthestressstabilitytestistomodelthebehaviour of the sample underextreme conditions, suchas hightemper- aturestodetermineandmodelextremestorageandproduction parameters[21],andpredictprocessesoccurringamongnormal

conditions.Duringthetemperature-inducedstresstestthenative proteinpowdersamples(1000mgeach)wereincubatedatseveral temperaturesfordifferenttimeperiodsandwerekeptinlocked glassvials.Theheatloadtookplaceinastabilitytestchamberkept and controlled at a constant temperature and humidity. Two controlsampleswerestoredatroomtemperatureandrefrigerated.

TheexactexperimentalconditionsaresummarisedinTable1.

2.3.Activitymeasurement

The hydrolysing activity of theexamined enzyme stored at varioustemperatures forvarioustime periodslistedabovewas estimatedusing o-nitrophenyl-β-D-galactopyranoside (ONPG) as substrate[22].ONPGisacolourlesssubstance,whichiscleavedby β-galactosidasetogalactoseando-nitrophenol(ONP),thereaction isindicatedbyayellowcolourchange.Thereagentwassolvedin pH=4.5phosphate-citrate(McIlvaine)buffersolution[23].5ng/

mLconcentrationsamplesweremadewithMilliQwaterandthe twosolutionsweremixedinaratioof1:3[24,25].Thesamples wereincubatedinwaterbathat37Cfor30min.Aliquotswere withdrawnfromthereactionmixtureandpipetteddirectlyinto 1mLNa2CO3 solution tostop the enzymatic hydrolysisand to producethemaximumpossibleabsorptionduetotheliberatedo- nitrophenolate ion. The solution was left to cool to room temperatureandtheabsorbanceofo-nitrophenolwasdetermined by UV–vis spectroscopy using an ATi Unicam UV2 UV/VIS Spectrometer (UNICAM,Budapest,Hungary)at420nm inthree parallelmeasurementswithﬁvedifferentstressedsamples.

2.4.DLS

Theaveragediameteroftheproteinsinthedifferentsamples wasdeterminedbydynamiclightscattering(DLS)usingZetasizer NanoZS^TM(MalvernPanalyticalLtd.,Malvern,UK).Absorbancesof a sample havebeen investigatedbeforeeach particlesizing by single beam UV–vis spectrophotometer, (8453 type; Agilent Technologies,USA)at

l

^max⁼⁶³³^nm,^to^determine ^the^correct

amount of sample scatterings vs. absorbances. Samples were illuminatedbyaHeNelaser(wavelengthof632.8nm,4.0mW),an avalanchephotodiode(APD)detectionanglewas173(backscatter mode). Protein samples were dissolved in MilliQ water in the concentration of 1mg/mL to reduce nonlinearity effects on measurements by increased viscosity of solvent with higher concentrations.Measurementsweremadeperforming15runsof 10seach,anequilibrationtimeofatleast5minwassetbeforethe measurementstarted.Sizemeasurementsweremadeintriplicate and wereusedimmediately afterpreparation.The experiments were carried out at a constant room temperatureof 25C. All solutions were analysed without prior ﬁltration to assess if aggregateswerepresent upto10,000nm (detectionlimit).The diameterofthesolvedmoleculesisrelatedtothepropertiesofthe solvent, so refractive index and viscosity were tested. The refractive index was measuredby a clinical refractometer, the viscositybyaBrookﬁledAmetekDVEviscometer(Middleboro,MA, US)anditprovedtobe1.031and0.8991cP.

Table1

Experimentlayoutofthermalstabilitystresstests.

Conditions StorageTime

40C 1day 3days 5days 1week 2weeks

60C 1day 3days 5days 1week 2weeks

80C 1day 3days 5days 1week

Control1(25Croomtemp.) 1day 2weeks

Control2(5 8Crefrigerated) 1day 2weeks

20Cdeepfrozen 1day 2weeks

(3)

2.5.Electrophoreticanalysis

SDS-PAGEwascarriedoutusinga10%Tris-Glycingelanda verticalNovexMinigelelectrophoreticsystem.100

m

^L^of¹^mg/mL

lactasesolutionwasmixedto100

m

^L^SDS^sample^buffer^(62.5^mM

Tris–HCl,2%SDS,25%glycerol,0.01%bromophenolblue,5% mercapto-ethanol or100mM DTT,pH=6.8)denaturedat80C for5min.15

m

^Lôfêach^sample^wasûsedând^theelectrophoresis was carried out at 35mA and 150V for 90min until the bromophenol blue has reached the lower edges of the gel.

Protein bands were visualised by staining with Coomassie Brilliant Blue R-250 in 50 % methanol, and 10 % acetic acid stainingsolution,distainedbyasolutionof5%methanoland10% acetic acid for overnight. Broad-range SDS–PAGE molecular weightstandardspurchasedfromVWRwereusedasmolecular mass standards (Myosin 200kDa, β-Galactosidase 120kDa, Bovine Serum Albumine 91kDa, Glutamic dehydrogenase 62kDa,Ovalbumin47kDa,CarbonicAnhydrase37kDa,Myoglo- bin28kDa,Lysozyme19kDa,Aprotinin9kDa).

FortheZymography(correlationbetweenactivityandprotein position)thegelissubjectedtoﬂuorometricanalysisusingMU-Gal (4-Methylumbelliferylβ-D-galactopyranoside)totestforβ-galactosidase activityof the separated bands. Thisgel is soaked for 15min in 100mM sodium acetate buffer, pH=5.0 with lateral shaking,100rpm.Then,itisincubatedwiththeMU-Galsolutionat 37Cfor 15min.Thezymogram gelis analyzedunderUVlight,

l

⁼³⁶⁵^nm ^to ^detect ênzyme âctivity ^due ^to^the ^release ôf ^4-

methylumbelliferone.Thisstepshouldbeperformedquickly,as the ﬂuorescent product diffuses within the gel, obscuring the bands.

NotethatwhileSDSandreducingagentareusedintheSDS- PAGEgelandrunningbuffer,non-denaturingelectrophoresiswas carried outwiththeomissionofSDSfromthegelrunningand loading buffers, and the sample was pre-treated only under moderatetemperature.

2.6.NMR

DatawereacquiredfromtwosampleswithtwodifferentNMR methods:1D¹Hand2Ddiffusion-orderedspectroscopy(DOSY) NMR.Thesampleswere:lactasestandard100mgsolvedinMilliQ water, 10 % D2O, 1% DSS (standard) added, pH=7.45. Heated lactaseat80Cfor14days100mgsolvedinMilliQwater,10% D2O, 1% DSS (standard) added, pH=6.96. Before the NMR measurements the samples were filtered five times with Eppendorf centrifuge tube equipped with a membrane of 10kDa cut. After the filtration, the samples were diluted to 500

m

^L ^with ^MilliQ ^Water ^then ^measured immediately. NMR measurementswereperformedonaBrukerAvanceIII700MHz spectrometerequippedwitha5mmProdigyTCIH&F-C/N-D,z- gradientprobeheadoperatingat700.05MHzfor¹H.Allspectra wereprocessedwithBrukerTOPSPINsoftware.¹HNMRmeasure- mentsDOSYexperimentswereperformedat14Cinaccordance withapreviousNMRstudywhichcomparesdiffusionconstants with the size of proteins [26], using the stebpgp1s19 pulse sequence with water suppressionfor diffusionmeasurements.

The lengthsofdiffusion delaysand pulseswereoptimizedto

d

=5ms,

D

⁼³⁰⁰^ms. ^The ^strength ^of ^the ^diffusion ^gradient ^was

linearlyincrementedin32equalsteps,varyingbetween5%and95

%ofitsmaximumvalue;theappliedmaximumgradientstrength was45.4G/cm.Thenumberofscanswasadjustedforeachsample toobtainreliableS/Nratios(16for ¹Hand 40 or160forDOSY experiments). Each measurement was repeated at least three times.Toinvestigatethediffusionconstantssignalswerechosen fromseveralregionsinthealiphaticprotonrange,andthedecay wasprocessedwith2DFouriertransform.

2.7.CD

Circulardichroism(CD)measurementswereperformedona JascoJ-815CD spectrometerequippedwitha thermostable cell holder.TemperaturewascontrolledandmonitoredwithaPeltier typeJascoCDF-426Lthermostatunit.Thesubstrateanddifferent stressed enzyme mixtures were measured together for kinetic studies,todeterminethereductionofactivityintime.Thepossible manifestationinsecondarystructurechangesduetothesupposed temperatureinduceddegradationwasalsointendedforanalyza- tion.Conformationalchangesinthesecondarystructureofprotein were monitored in the region of 190–250nm with a protein concentration of 50

m

^g/mL ^(0.60

m

^mol) ⁱⁿ ^a ^quartz ^cuvette

(Hellma) with a path length of 1mm. The scanning speed, bandwidth and data pitch were set to 50nm/min, 1nm and 0.5nm,respectively,whilethenumberofaccumulationsforeach recordedspectrumwassetto3(within600HTvoltagerange)and averagedtogetthecompletespectrum.

TheactivitymeasurementswerefollowedbyCDspectroscopy alternatively, where the ligand ONPG hydrolysis could be measured selectively online, and the kinetic curve could be acquired directly. Kinetic measurements were performed as follows: ONPG (0.050g / 10mL) +1

m

^L ^enzyme ^stock ^solution

(10mg/mL)dissolvedin3mLofMcIlvainecitrate-Na2HPO4buffer, pH=4.6,temperature37C0.05.Detectionwavelengthwasset at356nm(ONPGCDmaximum).

2.8.PeptideandN-glycosylationanalysis

Forthesequenceandglycanidentiﬁcationbottom-upanalysis wasused.Fivedifferentparallelstressedsampleswereidentiﬁed toinspectthepermanencelevelofthechanges.Allreagentswere used in relevant volume, comparable to theprotein level. The samplescorrespondingto100

m

^g^of^protein^were^solvedⁱⁿ^MilliQ

water.FortheS-Sbridgereductionammoniumbicarbonatebuffer atpH8.0containing200mMofDTTandRapigestwasadded.The samplesweresuccessivelyincubatedat60Cfor30min,andthen cooledat roomtemperature. Alkylationof cysteineresidues by adding 200mM IAA in 100mM ammonium bicarbonate was carried out in the dark at room temperature for 30min. The sampleswereleft tobedigested for90minat 37C byadding trypsin1:50w/wratioconsideringthetotalproteincontent.Lastly, sampleswereacidiﬁedwithcc.formicacidtostopthedigestion.In theendthesampleswerecentrifugedat13447G-forcefor10min andthephaseswerecarefullyseparated.Thetrypsindigestionwas carriedoutinsolutionwithoutTFAinordertoavoidtheMSsource contamination. The mass spectrometric measurements started rightafterthepreparationasfollows:6

m

^L^sample^was^subjected

tonanoLC-MS/MS analysisusing a Dionex Ultimate3000 RSLC nanoLC(Dionex,Sunnyvale,CA,USA)coupledtoaBrukerMaxisII Q-TOF (Bruker Daltonics, Bremen, Germany) via CaptiveSpray nanoBoosterionsource.PeptideswereseparatedonanAcquityM- ClassBEH130C18 analyticalcolumn(1.7

m

^m,⁷⁵

m

^m250mm Waters,Milford,MA) usinggradientelution(4–50%eluentBin 120min)followingtrappingonanAcclaimPepMap100C18,5

m

^m,

100

m

^m20mm(ThermoFisherScientiﬁc,Waltham,MA)trap column.SolventAconsistedofwatercontaining0.1%formicacid, while SolventB was acetonitrile containing 0.1 % formic acid.

Spectra were collectedusing a ﬁx cycle time of 2.5s and the following scan speeds: MS spectra at 3Hz, while CID was performedonmultiplychargedprecursorsat16Hzforabundant ones and at 4Hz for the ones with low abundance. Internal calibrationwasperformedbyCompassDataAnalysissoftware4.3 (BrukerDaltonics,Bremen,Germany)basedoninfusingsodium- formate. Data were processed by the Byonic software (Protein Metrics,Cupertino,CA,USA).Peptideswereidentiﬁedbysearching

(4)

againsttheAspergillusOryzae(W5ZSH9_ASPOZ)in theUNIprot database.GlycanswerequantiﬁedwiththeGlycoPatternsoftware [27].

3.Resultsanddiscussion

The enzyme activity and various structural aspects of the lactasesamplestoredundervariousconditions(Table1)hasbeen studied in detail. No observable change was found among the control and the frozen samples, so these will not be further discussed.Asexpected,thelargestandmostsigniﬁcantchangesin enzyme activity and property was observed at the highest temperaturesandlongestperiods.Structuralchangescorrespond- ingtotheobservedchangeinenzymeactivityhavebeenstudiedby a number of analytical methods. In the following discussion structureoflactase,andinparticularchangesinitsstructurehave beenevaluatedusinganumberofanalyticalmethods.Changesin the molecularmass has been studied by SDS-PAGE and NMR.

Changes in themolecularsize and polydispersity werestudied usingDLS.Informationaboutthestructurechangeswereobtained byMSandCDspectrometry.

Theseresultswillbepresentedbelowindetail.

3.1.Activity

Thechangesinlactaseactivityduetothermalstressareshown inFig.1.Thisclearlyshowsthatenzymeactivityisconsiderably reduced underheat stress.Degradationof thestressedenzyme sampleswascalculatedcomparedtotheactivityofthestandard sampletakenas100%.Enzymeactivityofthesamplestoredat 80Cforoneweekdecreasedto64%thatofthecontrolsample.

This is,however, a smallerdecreasethan occursis solution(in whichcaseactivitydecreasedtoca.40%thatofthecontrolsample) [7,28]. It is important, that after dissolving the heat-stressed sample, lactase activityslowly regeneratesin time (inhoursor days,from64to80%thatoftheoriginalactivity).Thisisshownin Fig.2.Thissuggestschangesinthesampleduetoheat-stressare partiallyreversible.

3.2.DLS

Havingobservedchangesinlactaseactivityduetoheat-stress, wehavetriedtounderstandthecorrespondingchangesinprotein structure.First,wehaveperformeddynamiclightscattering(DLS) measurements to characterize molecular size and possible

aggregation. We have determined the average molecular size (hydrolytic diameter using a hard sphere model) and the polydispersityindex.Note,polydispersityindicatesthevariability ofthemolecularsizebasedontheDLSdistribution[29,30].The DLSdistributionisshowninFig.3,boththatofthestandardand theheat-stressedsample.Thehydrolyticdiameteroftheparticles inthestandardenzymesolutionwasbetween9and10nmineach measurement, and polydispersity index was less than 0.2, suggesting a narrow distribution. These indicate, that under normal conditions the enzyme in solution is in the form of monomers.Fig.1showsthatheatstress(whichdecreasesenzyme activity)increasestheaverageparticlesize,indicatingaggregation andproteinunfolding.Dissolvingtheheat-stressedsample,itwas shownthatenzyme activityslowlyrecovers (Fig.2).In parallel, particle size decreases, indicating that aggregation (and/or unfolding)ispartiallyreversible.

Increaseofparticlesize issmall atlow temperaturesand is likely due to proteinunfolding. At higher temperatures (80C) large,2 3-foldincreaseinaverageparticlesizecanbeobserved.

Fig.3showsthatinthiscaseaslargeasmicrometersizedparticles alsoappear,clearlyindicatingproteinaggregation.Thisﬁgurealso showsthattheﬁrstpeakintheparticledistributionoftheheat- stressedsample iswiderthanthatofthestandardsample.This smallerchangeinparticlesizesuggestsproteinunfolding.

Fig.1.Theeffectoftemperatureofheatstressontheβ-galactosidaseactivity(n=3) andtheaveragehydrolyticdiameter(n=4)oftheparticlesinthestandardand differenttemperaturestressedenzymesolutions,immediatelyaftersolvation.

Fig.2.Changesinactivity(n=3)andaveragemolecularsize(n=4)overtimeinthe 80Cstressedlactasesampleinsolution.

Fig.3.Sizedistributionofthestandardandheatstressedproteindeterminedby DLSmeasurement.

(5)

3.3.SDS-PAGE

AccordingtotheinformationintheUNIprotdatabase[6],the weightofβ-galactosidasefromAspergillusoryzaeis109kDa.This informationisbasedontheaminoacidsequenceanddoesnottake into account possible structural modiﬁcations. Our preliminary MALDI-TOFmeasurementindicatesamolecularmassof123kDa, whichsuggestsextensivepost-translationalmodiﬁcations(PTMs).

We havestudiedchanges in themolecularmassof thecontrol sampleandheatstressedsamplesstoredat40,60and80Cforone weekusingSDS-PAGE.TheresultisshowninFig.4.

In the ﬁrst column a standard protein ladder was used to determinewherethedifferentmassesofmaterialsarelocatedon thegel.Thereisnovisiblechangeinmolecularweightinthecaseof 40C,60Cheatstressedsamples,butat80Cabandappearedat higher(approximately500kDa)molecularweight,fromwhichit canbeconcludedthattheproteinstartstoaggregateand likely formstri-ortetramericstructures,assuggested[31].Itcanalsobe seenonthegelthattheloadedsampledrewacontinuousstrip, presumably due tothepresence oflarger butnot well-deﬁned particlesincomparisontothemaincomponent.Boththestandard andstressedsamplecontainmoleculeswithlowermolecularmass at90,60and40kDa(Fig.4).Thepresenceofthesebandssuggests contaminantsordegradationproducts.Asthepresenceofthese bandsdidnotchangeduetoheatstress,thesewerenotfurther studied.

3.4.NMR

TheNMRmeasurementsprovidedinformationonthestructure andsizeoftheprotein.Thelargemolecularsizeandheterogeneous composition of the lactase samples caused some practical difficulties regarding NMR measurements. For this reason, the classicalStejksal-Tannerequationfittingdidnotgiveusefulresults, as regression always resulted in a multicomponent fitting.

Informationonvarioussizemolecularcomponentsinthelactase sample(i.e.smallmolecules,intactlactaseandlactaseaggregates wereobtainedusing2DDOSYspectroscopy).Thisrevealedthat multiplediffusionconstantscanbedetectedinthesampleswith

differencesbothinsizeanddistribution.Thespeciﬁeddiffusion constantsandthecalculatedmassesareextrapolated(basedonthe workofDudásandBodor[26])aresummarizedinTable2.

The27and 58kDasizerangesinthemeasurementsarenot causedbythethermal stress;theyare presentinboth samples fromthebeginninglikepresentedearlierinFig.4.Thesearelikely tobecontaminantsor degradationproducts in thecommercial lactasesample. Theparticlesobservedfollowingheat stressare likelydi-andtetrameraggregatesofthelactasemolecule(257and 550kDa).

BoththeDOSYand¹H-NMRpatternschangeovertimeinthe samples.BasedontheDOSYspectra,thesampleswerefoundtobe mixtures characterized by multiple diffusion constants. The moleculesizesdeterminedbasedonDOSYaresignificantlylarger inthestressedsamples(Table2).Thedissolvedproteinaggregates areunstableamongthesecircumstancesinaccordancewiththe formerobservations.AdefinitealterationinDOSYpatterncouldbe observedovertimeincaseoftheheatstressedsamplesdissolved forNMRmeasurement.Interestingly,thefractionofDOSYsignals correspondingtotheproteinaggregatesdecreasesovertime,and after72htheDOSYpatternoftheheatedsamplesfinallyreaches that of the untreated sample. This indicates that aggregation occurs in the solid state during heat stress, but it is mostly reversibleinsolution(Fig.5).

Theintensityofsharpsignalsofsmallmoleculesissigniﬁcantly growing in the¹H spectra of the heated samples (Fig.6).The increaseintheintensityofthesmallmoleculesignalsinthe¹H- NMRspectraispresumablyassociatedwiththesmallcarbohydrate moleculesarisingduetothedegradationoftheglycanchainsfrom theprotein.Glycandegradation hasbeen conﬁrmedbytheMS study,andwillbediscussedbelowindetail.

3.5.Circulardichroism

Lactase activity changes were observed/conﬁrmed by CD kineticstudies,whereenzymekineticswasstudiedbyobserving bothlactaseand itssubstrate.Directmonitoringofthereaction wasaccomplishedbydualCD/UVdetection.Itwaspossible,asthe amountofsubstrates, which formedasa resultofthecatalytic reaction of ONPG can be selectively measured in the solution, becausegalactoseischiraland theortho-nitrophenolisagood chromophore. During the hydrolysis, the disruption of the glycosidic bond causes thedisappearanceof CD signal and the strengtheningofUVsignal.TheCDcurveselectivelymeasuresthe hydrolysingsubstrate,whiletheincreaseintheUVcurveindicates the formation of dinitrophenol (Fig. 7). Results obtained from enzymekineticCDstudiescorrelatewellwithactivitymeasure- ments(Fig.1).

Thepossibleconformationalchangesinthesecondarystructure werealsoanalysed.Thedifferencesbetweenthespectramaybe causedbythedifferentinitialconcentrations,butifwecalculate theDissymmetryFactor(g),whichistheratiooftheCDandtheUV signals and a dimensionless parameter,then the concentration differences disappear. Furthermore, if there is a slight but signiﬁcant difference in either the CD or the UV spectrum, it

Fig.4. SDS-PAGEelectropherogramoftheStandardandheatstressedenzymes.The contentsofthelanesfromlefttorightarethefollowing:Lane1:molecularmass markers,massindicatedalongside.Lane2:Standardβ-gal.enzyme.Lane3-5:Heat stress.

Table2

Thediffusionconstantvaluesofthedifferentparticlesinthestandardand80C stressedsamplesandtheequivalentmasses.

Lactasestandard: Stressedsample(80C):

6.4710¹¹m²s ¹(27kDa) N/A

4.86¹⁰ ¹¹^m²^s ¹⁽⁵⁸^kDa) ^4.88¹⁰¹¹^m²^s ¹⁽⁵⁷^kDa) 3.6310 ¹¹m²s ¹(124kDa) 3.6610¹¹m²s ¹(121kDa)

N/A 2,7510¹¹m²s ¹(257kDa)

N/A 2.75¹⁰ ¹¹^m²^s ¹⁽⁵⁵⁰^kDa)

(6)

Fig.5.The2DDOSY1HNMRspectraofthesamesampledetected2h(left)and72h(right)afterdissolvingthesolidsamplesinwater.TheDOSYproﬁlebelongingtothe stressedsample(red)returntothestandard’sproﬁle(blue)after72h.

Fig.6. The¹Hspectrumofthestandardandstressedsamples.

Fig.7.LactoseandONPGCD(Upper)andUV(Bottom)kineticcurvesofsameconcentrationat37C.

(7)

appearsontheg-spectrum[32].ThefollowingFig.8showstheCD andUVspectraofsamplesontheleftandtheg-spectraofstandard andtreatedlactaseat80Contherightside.Theintensitiesdiffer duetothedifferentconcentrations.However,thevalue'g'(CD/UV ratiospectrum),whichisindependentofconcentration,doesnot differ signiﬁcantlyin thecase of thestandardand thesamples storedat40and60C,whichmeanstheresidualproteinstilldoes not loseitsstructure. Achangein theintensityof g-spectrum indicatessomestructuraldifferencescomparingthestandardand thetreatedsampleat80C.Realevidenceforthestructuralchange istheratioofthepeaksat241/222.4nm,wherethevalueis1.086 forthestandardand1.307forthetreatedsample.

3.6.PeptideandN-glycosylationanalysis

Thelactaseenzymes(thecontrolandheat-stressedsamples) were studied,after tryptic digestion,using LC–MS/MS analysis.

Peptides were identiﬁed using the Byonicsoftware. Very high, 95.02%sequencecoveragewasobservedforthecontrolsample, while91.4%forthe80Cstressedsample.Recentstudiesonother proteins [9,33] suggest glycans hasa largeinﬂenceonenzyme activity.Forthisreason,wehavestudiedN-glycosylationoflactase

indetail,for whichverylittleinformationisavailable. Wehave found8differentglycosylationsites; each wassubstitutedbya large number of glycans (Table 3). The glycans contained N- acetylhexosamine, hexose, sialic acid and fucose residues, and mosthadahigh-mannosestructure(Figs.9and10).Note,ourLC– MSanalysisallowedsite-speciﬁcglycosylationanalysis(i.e.tohave informationontheglycansobservedatthevariousglycosylation sites).Mostglycans wereso-called complexand highmannose glycans.

Heat-stress caused a large (and irreversible) change in glycosylation.Thiswasrevealedbyquantitativemassspectromet- ricanalysis.Forquantitativecomparison5replicatesofthecontrol sample,and5replicatesofthe80C heatstressedsamplewere analysed.Theglycopeptidesfoundinbothsampleswereselected for further investigation. A t-test was used to identify those glycans,thatshowedsigniﬁcant(p=0.05)differencebetweenthe controlandtheheat-stressedsample(Table3).Theglycosylation pattern(relativeglycoformabundances)fortwoglycosylationsites are shown in Figs. 9 and 10. The glycopeptide in Fig. 9 (NLTTGVYTDTSDLAVTPLMGDSPGSFFVVR) shows a very large changeduetoheat-stress:Thelargestchangeoccurredinglycan size,inparticularthenumberofhexoseunitspresent.Duetoheat- Fig.8.CDandUVspectraoffreshlypreparedlactasesamples(left)andg-spectraoflactasestandardandlactasestoredat80C(right).

Table3

Thetableshowstheexaminedpeptides,thenumberoffoundglycans,thenumberofstatisticallysigniﬁcantdifferencesbetweenthesamples.

Peptide Numberoffoundglycans SigniﬁcantDifferences

VNGTLR 25 21

SNVTIIEGSDSGIVSTR 26 20

GWDVPLYFNFGNNTQAAR 13 5

LPTSAGNLTIPQLEGTLSLNGR 24 17

LKLPTSAGNLTIPQLEGTLSLNGR 20 12

IHVVDYNVSGTNIIYSTAEVFTWK 19 19

NLTTGVYTDTSDLAVTPLMGDSPGSFFVVR 16 16

GAHLDGADLHLTADFNATTPIEVIAPTGAK 5 5

(8)

stresstheaveragenumberofhexoseunitsdecreasedformca.8to 5.Intheotherexample(VNGLTR)thechangeinglycancomposi- tion was not so large, but still observable, and statistically signiﬁcant(Fig.10andTable3).Notethatthebuildupofcomplex glycansoccursinlivingcells,sotheirdegradationisirreversible.

4.Conclusion

Our investigations indicate various changes occur when submittingthesolid enzymetoheat stress.Heat stress causes signiﬁcant decrease in enzyme activity. However, after Fig.9. SectionoftheglycosylationpatternofNLTTGVYTDTSDLAVTPLMGDSPGSFFVVR,comparingtheaverageabundanceoftheglycansofthestressedandstandardsamples ontheNLTTGVYTDTSDLAVTPLMGDSPGSFFVVRglycopeptide.(N:N-acetylhexosamine,H:hexose,S:sialicacid,F:fucose).

Fig.10.SectionoftheglycosylationpatternofVNGLTR,comparingtheaverageabundanceoftheglycansofthestressedandstandardsamplesontheVNGLTRglycopeptide.

(N:N-acetylhexosamine,H:hexose,S:sialicacid,F:fucose).

(9)

dissolution,inseveralhours,enzymeactivitysomewhatrecovers, suggestingtheactivitydecreaseispartlyreversible.Thiscanbe explained by reversible aggregation of lactase molecules, suggested by the results of DLS and NMR measurements. In thecaseoforalmedication,partialrecoveryofenzymeactivityis notrelevant,aslactasemoleculesdonotspendsufﬁcienttimein dissolved form in the body. The aminoacidsequence andthe secondarystructureoftheenzymedidnotshowanychangedue toheatstress.Lactaseisheavilyglycosylated,andheatstressdid changeglycosylation signiﬁcantly.LC–MSexperimentsrevealed thatvariouspeptide-boundglycanchainswereshorteneddueto heat stress. These are irreversible changes, changeing the physico-chemical properties of the enzyme and may increase tendencyofaggregation.

Thesestudiessuggestthatbothstructuralchangesinthehigher order structure of the protein and changes due to altered glycosylation are responsible for the decrease of activity.

Understanding behaviourunderheat stressmayhelp todesign better manufacturing, shipping, and storage conditions and improve our understanding of different types of lactases, like Kluyveromyceslactisusedindairyindustryorevenhumanlactase.

The methodology developed here may serve as a model for investigatingotherenzymesofpharmaceuticalinterest.

CRediTauthorshipcontributionstatement

MártonKirály:Writing-originaldraft,Visualization,Formal analysis, Investigation, Methodology. Borbála Dalmadi Kiss:

Writing-review&editing,Methodology,Conceptualization.Péter Horváth: Investigation, Writing - original draft, Visualization.

László Drahos: Software, Formal analysis, Writing - review &

editing. Arash Mirzahosseini: Funding acquisition, Writing - originaldraft.GyulaPálfy:Investigation,Visualization,Writing- review & editing. István Antal: Resources. Krisztina Ludányi:

Conceptualization, Supervision, Funding acquisition, Writing - review&editing,Methodology.

DeclarationofCompetingInterest

The authors declare that they have no known competing ﬁnancial interests or personal relationships that could have appearedtoinﬂuencetheworkreportedinthispaper.

Acknowledgements

TheresearchwasﬁnancedbytheHigherEducationInstitutional Excellence Programmeof the Ministryof Human Capacities in Hungary,withintheframeworkofthemolecularbiologythematic programmeoftheSemmelweisUniversity.

AppendixA.Supplementarydata

Supplementarymaterialrelatedtothisarticlecanbefound,in the online version, at doi:https://doi.org/10.1016/j.btre.2021.

e00637.

References

[1]S.Zárate, M.H. López-Leiva, Oligosaccharideformation duringenzymatic lactosehydrolysis:aliteraturereview,J.FoodProt.53(1990)262–268,doi:

http://dx.doi.org/10.4315/0362-028X-53.3.262.

[2]T.M. Bayless, E. Brown, D.M. Paige, Lactase non-persistence and lactose intolerance,Curr.Gastroenterol.Rep.19(2017)23,doi:http://dx.doi.org/

10.1007/s11894-017-0558-9.

[3]B.Misselwitz,D. Pohl,H.Frühauf,M.Fried,S.R.Vavricka,M.Fox,Lactose malabsorptionandintolerance:pathogenesis,diagnosisandtreatment, UnitedEur.Gastroenterol.J.1(2013)151–159,doi:http://dx.doi.org/10.1177/

2050640613484463.

[4]C.L.Storhaug,S.K.Fosse,L.T.Fadnes,Country,regional,andglobalestimatesfor lactosemalabsorptioninadults:asystematicreviewandmeta-analysis, LancetGastroenterol.Hepatol.2(2017)738–746,doi:http://dx.doi.org/

10.1016/S2468-1253(17)30154-1.

[5]S.Saqib,A.Akram,S.A.Halim,R.Tassaduq,Sourcesofβ-galactosidaseandits applicationsinfoodindustry,3Biotech7(2017)79,doi:http://dx.doi.org/

10.1007/s13205-017-0645-5.

[6]Y.Ito,T.Sasaki,K.Kitamoto,C.Kumagai,K.Takahashi,K.Gomi,G.Tamura, Cloning,nucleotidesequencing,andexpressionofthebeta-galactosidase- encodinggene(lacA)fromAspergillusoryzae,J.Gen.Appl.Microbiol.48 (2002)135–142,doi:http://dx.doi.org/10.2323/jgam.48.135.

[7]L.F.Atyaksheva,O.S.Pilipenko,E.S.Chukhrai,O.M.Poltorak,Similarityofand differencesbetweenthemechanismsofthermalinactivationofβ- galactosidasesofdifferentorigins,Russ.J.Phys.Chem.A,FocusChem.82 (2008)864–869,doi:http://dx.doi.org/10.1134/S0036024408050300.

[8]N.G.Jayaprakash,A.Surolia,Roleofglycosylationinnucleatingproteinfolding andstability,Biochem.J.474(2017)2333–2347,doi:http://dx.doi.org/10.1042/

BCJ20170111.

[9]P.Goettig,Effectsofglycosylationontheenzymaticactivityandmechanisms ofproteases,Int.J.Mol.Sci.17(2016)1969,doi:http://dx.doi.org/10.3390/

ijms17121969.

[10]D.Shental-Bechor,Y.Levy,Effectofglycosylationonproteinfolding:aclose lookatthermodynamicstabilization,Proc.Natl.Acad.Sci.U.S.A.105(2008) 8256–8261,doi:http://dx.doi.org/10.1073/pnas.0801340105.

[11]R.J.Solá,K.Griebenow,Effectsofglycosylationonthestabilityofprotein pharmaceuticals,J.Pharm.Sci.98(2009)1223–1245,doi:http://dx.doi.org/

10.1002/jps.21504.

[12]I.Wagner,Z.K.Nagy,P.Vass,C.Fehér,Z.Barta,T.Vigh,P.L.Sóti,A.H.Harasztos, H.Pataki,A.Balogh,G.Verreck,I.VanAssche,G.Marosi,Stableformulationof protein-typedruginelectrospunpolymericﬁberfollowedbytabletingand scaling-upexperiments,Polym.Adv.Technol.26(2015)1461–1467,doi:http://

dx.doi.org/10.1002/pat.3569.

[13]P.Vass,E.Hirsch,R.Kóczián,B.Démuth,A.Farkas,C.Fehér,E.Szabó,Á.

Németh,S.K.Andersen,T.Vigh,G.Verreck,I.Csontos,G.Marosi,Z.K.Nagy, Scaled-upproductionandtabletingofGrindableelectrospunﬁberscontaining aprotein-typedrug,Pharmaceutics.11(2019)329,doi:http://dx.doi.org/

10.3390/pharmaceutics11070329.

[14]B.S.Dionizio,D.V.deBabos,D.H.Souza,E.R.Pereira-Filho,Evaluationofthe QualityofFormulationsContainingLactaseβ-Galactosidase)EmployingGel ElectrophoresisandCellPhone,(2019).

[15]M.J.Hernaiz,D.H.G.Crout,Immobilization/stabilizationonEupergitCoftheβ- galactosidasefromB.circulansandanα-galactosidasefromAspergillus oryzae,EnzymeMicrob.Technol.27(2000)26–32,doi:http://dx.doi.org/

10.1016/S0141-0229(00)00150-2.

[16]D.Kishore, S.Kundu,A.M.Kayastha,Thermal,chemical andpH induced denaturationofamultimericβ-galactosidaserevealsmultipleunfolding pathways,PLoSOne7(2012)e50380,doi:http://dx.doi.org/10.1371/journal.

pone.0050380.

[17]H.Khan,M.Ali,A.Ahuja,J.Ali,Stabilitytestingofpharmaceuticalproducts- comparisonofstabilitytestingguidelines,Curr.Pharm.Anal.6(2021)142–

150.n.d.https://www.ingentaconnect.com/content/ben/cpa/2010/00000006/

00000002/art00006.

[18]A.Hawe,M.Wiggenhorn,M.vandeWeert,J.H.O.Garbe,H.-C.Mahler,W.

Jiskoot,Forceddegradationoftherapeuticproteins,J.Pharm.Sci.101(2012) 895–913,doi:http://dx.doi.org/10.1002/jps.22812.

[19]N.Stourman,J.Moore,Analysisoflactaseinlactoseintolerancesupplements, Biochem.Mol.Biol.Educ.46(2018)652–662,doi:http://dx.doi.org/10.1002/

bmb.21185.

[20]L.P.L. Lim, E. Garnsey, M. Gregory, Product and process innovation in biopharmaceuticals:anewperspectiveondevelopment,R&DManag.36 (2006)27–36,doi:http://dx.doi.org/10.1111/j.1467-9310.2006.00413.x.

[21]C.P.Chan,Forceddegradationstudies:currenttrendsandfutureperspectives forprotein-basedtherapeutics,ExpertRev.Proteomics13(2016)651–658,doi:

http://dx.doi.org/10.1080/14789450.2016.1200469.

[22]R.Kwapiszewski,J.Szczudlowska,K.Kwapiszewska,M.Chudy,Z.Brzozka, Determinationofacidβ-galactosidaseactivity:methodologyandperspectives, IndianJ.Clin.Biochem.29(2014)57–62,doi:http://dx.doi.org/10.1007/

s12291-013-0318-z.

[23]J. Lederberg, The beta-d-galactosidaseof Escherichia coli, strain K-12, J.

Bacteriol.60(1950)381–392.https://www.ncbi.nlm.nih.gov/pubmed/

14784466.

[24]B. Rotman, Measurement of activity of single molecules of beta-D- galactosidase,Proc.Natl.Acad.Sci.U.S.A.47(1961)1981–1991,doi:http://

dx.doi.org/10.1073/pnas.47.12.1981.

[25]W.Li, X.Zhao,S.Zou,Y.Ma,K.Zhang,M.Zhang,Scanningassayofβ- galactosidaseactivity,Appl.Biochem.Microbiol.48(2012)603–607,doi:

http://dx.doi.org/10.1134/S0003683812060075.

[26]E.F.Dudás,A.Bodor,Quantitative,diffusionNMRbasedanalytical toolto distinguishfolded,disordered,anddenaturedbiomolecules,Anal.Chem.91 (2019)4929–4933,doi:http://dx.doi.org/10.1021/acs.analchem.8b05617.

[27] O.Ozohanics,L.Turiák,A.Puerta,K.Vékey,L.Drahos,High-performanceliquid chromatographycoupledtomassspectrometrymethodologyforanalyzing site-speciﬁcN-glycosylationpatterns,J.Chromatogr.A1259(2012)200–212, doi:http://dx.doi.org/10.1016/j.chroma.2012.05.031.

[28]A.Bosso,L.R.I.Morioka,L.F.dosSantos,H.H.Suguimoto,Lactosehydrolysis potentialandthermalstabilityofcommercialβ-galactosidaseinUHTand

(10)

skimmedmilk,FoodSci.Technol.36(2016)159–165,doi:http://dx.doi.org/

10.1590/1678-457X.0085.

[29]Z.Yu,J.C.Reid,Y.-P.Yang,Utilizingdynamiclightscatteringasaprocess analyticaltechnologyforproteinfoldingandaggregationmonitoringin vaccinemanufacturing,J.Pharm.Sci.102(2013)4284–4290,doi:http://dx.doi.

org/10.1002/jps.23746.

[30]K. Shiba, T. Niidome,E. Katoh, H. Xiang,L. Han, T. Mori, Y. Katayama, Polydispersityasaparameterforindicatingthethermalstabilityofproteinsby dynamiclightscattering,Anal.Sci.Int.J.Jpn.Soc.Anal.Chem.26(2010)659–

663,doi:http://dx.doi.org/10.2116/analsci.26.659.

[31]C.Sagné,M.F.Isambert,J.P.Henry,B.Gasnier,SDS-resistantaggregationof membraneproteins:applicationtothepuriﬁcationofthevesicular

monoaminetransporter,Biochem.J.316(Pt3)(1996)825–831,doi:http://dx.

doi.org/10.1042/bj3160825.

[32]B.R.Baker,R.L.Garrell,g-Factoranalysisofproteinsecondarystructurein solutionsandthinﬁlms,FaradayDiscuss.126(2004)209–222,doi:http://dx.

doi.org/10.1039/B305291E.

[33]G.T.Beckham,Z.Dai,J.F.Matthews,M.Momany,C.M.Payne,W.S.Adney,S.E.

Baker,M.E.Himmel,Harnessingglycosylationtoimprovecellulaseactivity, Curr.Opin.Biotechnol.23(2012)338–345,doi:http://dx.doi.org/10.1016/j.

copbio.2011.11.030.