protein surface functionalisation as a general strategy ... · details of the proteins, their...

Supportinginformationfor:

ProteinsurfacefunctionalisationasageneralstrategyforfacilitatingbiomimeticmineralisationofZIF-8

NatashaK.Maddigana,AndrewTarziaa,DavidM.Huanga,ChristopherJ.Sumbya,StephenG.Bella*,PaoloFalcaroab*andChristian.J.Doonana*

a.DepartmentofChemistryandtheCentreforAdvancedNanomaterials,TheUniversityofAdelaide,Adelaide,SouthAustralia5005,Australia.Email:christian.doonan@adelaide.edu.au.

b.InstituteofPhysicalandTheoreticalChemistry,GrazUniversityofTechnology,Stremayrgasse9,Graz8010,Austria.

Electronic Supplementary Material (ESI) for Chemical Science.This journal is © The Royal Society of Chemistry 2018

Contents

1. Materials 3

2. Proteinsurfacemodificationreactions 3

3. Timecoursebiomimeticmineralisationstudies 4

4. PowderX-raydiffraction(PXRD)data 6

5. Scanningelectronmicroscopy 8

6. UV-visiblespectra 9

7. Additionalcomputationalmethods 10

8. Additionalcomputationalresults 11

9. References 16

1. Materials

TableS1.Detailsoftheproteins,theirsources,productcodesandtheProteinDataBank(PDB)codesfortheproteinsinvestigatedinthisresearch.

Protein Source ProductCode PDBFileUseda

Pepsin Porcinegastricmucosa

P6887 4pep1

Bovineserumalbumin(BSA) Bovine A9418 4f5s2Lipase,CandidaantarcticaLipaseB(CALB)

Aspergillusoryzae 62288 1tca3

Catalase BovineLiver C9322 3re84Peroxidasefromhorseradish(HRP)

Horseradish 77332 1w4w5

Myoglobin Equineskeletalmuscle

M0630 2frf6

Haemoglobin Human H7379 2dn27Trypsin Porcinepancreas T4799 1s818Lysozyme Eggwhite AstralScientific

(LDB0308)2vb19

aProteinstructuresarefromthesameorganismfromwhichtheproteinsampleissourced.

2.Proteinsurfacemodificationreactions

SchemeS1.Surfacemodificationreactions.SuccinylationandacetylationreactionslowerthepIofaproteinbymodificationofexposedaminegroups.Aminationreactionscapcarboxylgroupswithafreeamine,thusincreasingthepI.

3. Timecoursebiomimeticmineralisationstudies

Haemoglobin

HaemoglobinSuccinylated

HaemoglobinAcetylated

Myoglobin

MyoglobinSuccinylated

MyoglobinAcetylated

Figure S1. Sequential photographs of Haemoglobin (Hb), succinylated haemoglobin (HbSucc),acetylatedhaemoglobin(HbAc),myoglobin(Mb),succinylatedmyoglobin(MbSucc),andacetylatedmyoglobin(MbAc)samples(2mgprotein)immediatelyaftermixingofthemIM(160mM)andzincsolutions (40mM) (T=0) until immediately prior to centrifugation andwashing (T=16 hours). Theunmodifiedhaemoglobinandmyoglobinsamplesremainclearuponadditionofthezincsolution,bothyieldingminimalproductafter16hours.ThesuccinylatedandacetylatedformsofbothproteinscauseimmediateprecipitationofZIF.

BSA

Pepsin

BSAAminated

PepsinAminated

FigureS2.Sequentialphotographsofunmodifiedandaminatedbovineserumalbumin(BSA)andpepsin(2mgprotein)immediatelyaftermixingofthemIM(160mM)andzincsolutions(40mM)(T=0)untilimmediatelypriortocentrifugationandwashing(T=16hours).TheunmodifiedBSAandpepsinsamplesgaveimmediatebiomineralizationuponadditionofthezincsolution.TheaminatedBSAandpepsinsamplesshowadramaticreductioninprecipitationyieldingonlyminimalproductafter16hours.

4. PowderX-raydiffraction(PXRD)data

FigureS3.PowderX-raydiffractionpatternsofbiomimeticallymineralisedZIFsamplesofunmodifiedproteinsmadeunderstandardconditions(4:1mIM:Zn2+).Datacollectedondriedsamplesafterwasheswithwaterandethanol.Unmodifiedhaemoglobin,myoglobin,lysozyme,andtrypsindidnotyieldsufficientproductforPXRDanalysis.ThesimulatedpatternrelatestoZIF-8.

FigureS4.PowderX-raydiffractionpatternsofbiomimeticallymineralisedZIFsamplesofHbSuccandMbSucc(top)andHbAcandMbAc(bottom)madeunderstandardconditions(4:1mIM:Zn2+).Datacollectedondriedsamplesafterwasheswithwaterandethanol.AminatedBSAandpepsin,didnotyieldsufficientproductforPXRDanalysis.ThesimulatedpatternrelatestoZIF-8.

5. Scanningelectronmicroscopy

FigureS5.HbSucc@ZIF-8(left)andMbSucc@ZIF-8(right)after16hoursfromthebeginningofthebiomimeticmineralizationreaction;therhombicdodecahedralmorphologycanbeobserved.

1µm 1µm

6. UV-visiblespectra

FigureS6.UV-visiblespectraofthesupernatantremovedaftercentrifugationofthebiocompositesampleswheretheproteinwasmyoglobin(Mb),succinylatedmyoglobin(MbSucc),andacetylatedmyoglobin(MbAc)(left)andhaemoglobin(Hb),succinylatedhaemoglobin(HbSucc),andacetylatedhaemoglobin(HbAc)(right).UnmodifiedhaemoglobinandmyoglobinformedminimalproductandthereforeshowalargeSoretabsorbanceintheremovedsolution,indicatingthattheproteinhasnotbeenimmobilised.Inboththesuccinylatedandacetylatedvariants,theabsorbancehasdecreasedthusindicatingthattheproteinhasbeenremovedfromsolutionandincorporatedintoZIF-8asitformed.

FigureS7.UV-visiblespectraofthewashingsofsuccinylatedhaemoglobinZIF-8samples.Thesupernatant(red)wasobtainedaftercentrifugationoftheproductandshowsnoevidenceofproteinremaininginsolution.SDSwashes1(blue)and2(pink)showtheappearanceofthehaemoglobinabsorbancepeak,indicatingthatsomeproteinwassurfaceboundhadbeenwashedoff.AftertheSDSwashestoremovesurfaceboundproteinshowednofurtherprotein,theZIF-8samplewasdissolvedincitricacidbuffer(pH6)containingEDTA(20mM)andtheabsorbancewasmeasuredtoshowpresenceofencapsulatedprotein.

7. Additionalcomputationalmethods

7.1. Calculationoftheaveragehydropathicindex

Thehydropathicindexisameasureofanaminoacidsequenceshydropathicity.Negativevaluesimplyan overall hydrophilic protein,whereas positive values imply an overall hydrophobic protein. ThehydropathicindexforaproteinsequencewascalculatedusingtheKyteandDoolittlescaleofresiduehydropathicity,10whichquantifiesthehydropathicityofeachresidue,andtheBiopythonmodule.11,12Asinglevalueisreported,whichisthesumofthehydropathicindicesofallresiduesdividedbythelengthofthesequence.7.2. PreparationofPDBfilesandcalculationofproteinchargestate

CrystalstructureswereobtainedfromtheProteinDataBank13foreachprotein(PDBaccessioncodesgiveninTableS1).Whereavailableaproteinstructureassociatedwiththesameorganismastheexperimentalsourcewasobtained.EachPDBfilecomeswithoneormorepeptidechain,whereeachchainrepresentsaseparatesequenceofaminoacidsinthecrystalstructure.ForBSA,onlythefirstpolypeptidechaininthePDBfilewasusedbecausethisproteinisexpectedtoexistasamonomerinsolution.InallothercasesallchainsinthePDBfilewereused.Heteroatoms(non-naturalaminoacidresiduesorligands),boundionsorwatermoleculesintheproteinstructureswereremoved.

PROPKA3.014,15wasusedtoestimatethepKaofeachionisableresidueineachproteinstructureusingahighlyefficient,empiricalmethod.PROPKAuseseffectivepotentialstocalculatethetotalenvironmentalperturbationtothefreeenergyofprotonationduetomovingtheionisableresiduefromwaterintothe3Denvironmentoftheprotein.TheresultantfreeenergywasusedtodeterminetheshiftintheknownpKaofeachresidueduetotheproteinenvironment.WehaveconfirmedthatsimilarresultsareobtainedforthecalculatedpKa’susingthemoresophisticatedDELPHIPKA16toassignatomchargesandprotonationstates(resultsnotshown).DELPHIPKAusesavariabledielectriccoefficientwithintheproteinandthefreeenergydifferencebetweentheprotonatedanddeprotonatedstateofeachionisableresiduewithinthe3Dstructure(usingaPoisson–Boltzmann-basedapproachtocalculatethefreeenergydifference)toobtainthepKaforeachresidue.ThecalculatedpKaofeachionisableresidue,givenbyPROPKA,wasthenusedtocalculatethe3DmodelpIofeachproteinusingtheHenderson–Hasselbachequation.

Beforeanalysingeachcrystalstructure,thePDB2PQRsoftware17,18wasusedtoaddmissingheavyatoms,tomakesuretherewerenooverlappingatomsinthestructure,toprotonatethestructurebasedonthepKa’scalculatedbyPROPKAandthegivenpH(wherearesidueisprotonatedifitspKaisgreaterthanthegivenpH)andtoassignchargesandradiifromtheAMBER19forcefieldtoeachatom.WenotethattheAMBERforcefieldincludedwithPDB2PQRdoesnotcontainchargeparametersforresiduesincertainprotonationstatesderivedbyPROPKA(forexample,aneutralN-terminusstateisnotsupportedbytheforcefieldprovidedbyPDB2PQR)andthereforesomeresidueswillalwaysexistintheirpH7state.

8. Additionalcomputationalresults

8.1. Proteinmetrics

InFigureS8weshowthecalculatedaveragehydropathicindexofthesequenceofeachprotein.TheresultsindicatethattheproteinsthatseedZIF-8growthhavehydropathicitiesthatoverlapcompletelywithproteinsthatdonotseedZIF-8growth.ThisfindingfurthersupportsthedominantnatureofelectrostaticinteractionsindeterminingZIF-8formation,whichallowsfortheuseofsuchsimplifiedscreeningmethods.

FigureS8.Categoricalscatterplotoftheaveragehydropathicityofthepeptidesequencesforallproteins.ClosedcirclesareproteinsthatformZIF-8andopencirclesareproteinsthatdonotformZIF-8.

8.2. ComparisonofpIfromsequencemodels,3Dmodelsandexperiments

FigureS9showscategoricalscatterplotsofthecalculatedpIsfromthe3Dstructure(obtainedfromPROPKA3.0)andpeptidesequence(obtainedfromBiopython)ofeachprotein,whichshowsthatbothcalculationmethodspredictZIF-8growthreasonablyaccurately.ParityplotsofthepIsobtainedfrombothcalculationmethodsaswellasacomparisonbetweenthepIscalculatedfromthe3DproteinstructureandthereportedpIs(Table1)arealsoshown.Importantly,reasonableagreementbetweenthetwocalculationmethodswasobtained,indicatingthatthemuchsimplersequence-basedmodelcanbeusedwithoutadverselyaffectingpredictionaccuracy.WenotethatsomeexperimentalpIsarereportedasarangeofvalues,andsoerrorbarsareincludedinFigureS9d,forwhichtheuncertaintyencompassesthereportedrangeandthepIisthemeanofthereportedrange.

FigureS9.CategoricalscatterplotsofthecalculatedpI(a)fromthe3Dmodeland(b)sequencemodelofallproteins.ParityplotscomparingthecalculatedpIfromthe3Dmodeland(c)thesequencemodeland(d)thereportedpIvalues(they=xlineisshown).ErrorbarsrepresentrangesofpIvaluesreportedfromexperiments.ClosedcirclesareproteinsthatformZIF-8andopencirclesareproteinsthatdonotformZIF-8.

8.3. Experimentalzincionenhancement

FigureS10showsacategoricalscatterplotoftheenhancementofzincionscalculatedfromtheexperimentalzetapotentialsinTable1usingEquation4.Experimentalzetapotentialsgivereasonableapproximationsofthesurfaceelectrostaticpotentialofeachproteininsolutionand,therefore,aproteinsabilitytoenhancezincionconcentrationsnearthesurfaceand,hence,seedZIF-8growth.BasedonFigureS10,asurfacezincionenhancementof>10,whichisazincionconcentrationof0.4M,leadstoZIF-8formationunderexperimentalconditions.

FigureS10.CategoricalscatterplotsofthezincionenhancementscalculatedfromtheexperimentalzetapotentialsatpH11forallproteins.ClosedcirclesareproteinsthatformZIF-8andopencirclesareproteinsthatdonotformZIF-8.

8.4. Zincionenhancementfrom3Dmodel

Thecalculatedaveragesurfacepotentialsfromthe3Dmodelofeachproteinshowreasonableagreementwiththeexperimentalzetapotentials(FigureS11).Themaindiscrepanciesaretheoverestimationoftheaveragesurfacepotentialcomparedwithexperimentalzetapotentialsforveryhighlychargedproteins,suchasBSA,catalaseandpepsin.Thisresultisnotunexpected,giventheuseofthelinearisedPoisson–Boltzmannequation,whichbreaksdowninregimesofhighzeta

potential(|𝜁| > %B'()*

≈ 12mV).Theunderestimationoftheaveragesurfacepotentialcompared

withexperimentalzetapotentialsforlipaseandHRPislikelyaresultofexperimentalimpurities.Bothproteinsareexpectedtobeglycosylated,20,21whichisknowntoaffectzetapotentialmeasurements,22whereasthecalculationsusednon-glycosylatedstructures.Additionally,HRPcouldbeamixtureofdifferentiso-enzymeswithvastlydifferentelectrostaticproperties.23WenotethatbothproteinshavereportedpIsthatspanabroadrangeofvalues(Table1),indicatingabroadrangeofelectrostaticpropertiesfordifferentsamples.

Ourcalculationmethodologyusedstatic3DstructuresofeachproteinobtainedfromX-raycrystallography,whichareunlikelytoberepresentativeoftheproteinstructureinsolutionatapHof11.AthighpHs,thepresenceofhigh-chargeregionswouldleadtorepulsionandadegreeofunfolding,whichsuchasimplemodelcouldnottakeintoaccount.Wealsonotethatithasbeenshownpreviouslythattheinteriorofaproteinhasahighlyvariabledielectriccoefficientandassumingaconstantdielectriccoefficient,aswehave,cangiverisetoerrorsnearthesurfaceofproteins.24Furthermore,bytakingtheaveragesurfacepotentialtobeequaltotheexperimentalzetapotentialforaheterogenousproteinsurfaceweassumedthattheelectricdoublelayersurroundingtheproteinisthincomparedwiththesizeoftheproteinandthatthelinearisedPoisson–Boltzmannequationapplies,whichmaynotalwaysbethecaseforthesystemsstudied(discussedabove).25Thesemi-quantitativeagreementwithexperimentinmostcasesisveryencouraging,consideringtheapproximationsinthecalculations.

FigureS11.Parityplotscomparingthecalculatedsurfacepotentialfromour3Dmodelandexperimentalzetapotentialsat(a)pH7and(b)pH11forallproteins(they=xlineisshown).ClosedcirclesareproteinsthatformZIF-8andopencirclesareproteinsthatdonotformZIF-8.

FigureS12showsacategoricalscatterplotofthecalculatedaveragesurfacepotentialsatpH9andpH11forallproteins.Theseresultssupporttheexperimentalfindingsandshowareasonableabilitytopredictaprotein’spropensitytoseedZIF-8formation.ResultsatpH9andpH11areshownastheinitialsolution(beforezincionsareadded)isatapproximatelypH11,butuponzincionaddition,thepHquicklydecreases toaround9, likelybecauseof ZIFnucleation.26 Finally, FigureS13 shows thevariationinthesurfacepotentialaroundallproteinsatpH11calculatedfromour3Dmodel.

FigureS12.Categoricalscatterplotsofthecalculatedsurfacepotentialfromthe3Dmodel(a)atpH9and(b)pH11forallproteins.ClosedcirclesareproteinsthatformZIF-8andopencirclesareproteinsthatdonotformZIF-8.TheshadedregionhighlightstheapproximateboundaryofthezetapotentialintheexperimentsforproteinsthatdoanddonotseedZIF-8growth.

FigureS13.Surfacepotentialsurroundingallproteinscalculatedfromour3DmodelatpH11.LipaseandHRPareoutliersbasedonouranalysis.

9. References

1. A.R.Sielecki,A.A.Fedorov,A.Boodhoo,N.S.AndreevaandM.N.G.James,J.Mol.Biol.,1990,214,143-170.

2. A.Bujacz,ActaCrystallogr.D,2012,68,1278-1289.3. J.Uppenberg,M.T.Hansen,S.PatkarandT.A.Jones,Structure,1994,2,293-308.4. N.Purwar,J.M.McGarry,J.Kostera,A.A.PachecoandM.Schmidt,Biochem.,2011,50,

4491-4503.5. G.H.Carlsson,P.Nicholls,D.Svistunenko,G.I.BerglundandJ.Hajdu,Biochem.,2005,44,

635-642.6. D.M.Copeland,A.S.Soares,A.H.WestandG.B.Richter-Addo,J.Inorg.Biochem.,2006,

100,1413-1425.7. S.-Y.Park,T.Yokoyama,N.Shibayama,Y.ShiroandJ.R.H.Tame,J.Mol.Biol.,2006,360,

690-701.8. T.R.Transue,J.M.Krahn,S.A.Gabel,E.F.DeRoseandR.E.London,Biochem.,2004,43,

2829-2839.9. J.Wang,M.Dauter,R.Alkire,A.JoachimiakandZ.Dauter,ActaCrystallogr.D,2007,63,

1254-1268.10. J.KyteandR.F.Doolittle,J.Mol.Biol.,1982,157,105-132.11. T.HamelryckandB.Manderick,Bioinformatics,2003,19,2308-2310.12. P.J.A.Cock,T.Antao,J.T.Chang,B.A.Chapman,C.J.Cox,A.Dalke,I.Friedberg,T.

Hamelryck,F.Kauff,B.WilczynskiandM.J.L.deHoon,Bioinformatics,2009,25,1422-1423.13. H.M.Berman,J.Westbrook,Z.Feng,G.Gilliland,T.N.Bhat,H.Weissig,I.N.Shindyalovand

P.E.Bourne,NucleicAcidsRes.,2000,28,235-242.14. C.R.Søndergaard,M.H.M.Olsson,M.RostkowskiandJ.H.Jensen,J.Chem.Theory

Comput.,2011,7,2284-2295.15. M.H.M.Olsson,C.R.Søndergaard,M.RostkowskiandJ.H.Jensen,J.Chem.Theory

Comput.,2011,7,525-537.16. L.Wang,L.LiandE.Alexov,Proteins:Struct.,Funct.,Bioinf.,2015,83,2186-2197.17. T.J.Dolinsky,J.E.Nielsen,J.A.McCammonandN.A.Baker,NucleicAcidsRes.,2004,32,

W665-W667.18. T.J.Dolinsky,P.Czodrowski,H.Li,J.E.Nielsen,J.H.Jensen,G.KlebeandN.A.Baker,Nucleic

AcidsRes.,2007,35,W522-W525.19. J.Wang,P.CieplakandP.A.Kollman,J.Comput.Chem.,2000,21,1049-1074.20. O.Spadiut,L.Rossetti,C.DietzschandC.Herwig,ProteinExpr.Purif.,2012,86,89-97.21. I.Høegh,S.Patkar,T.HalkierandM.T.Hansen,Can.J.Bot.,1995,73,869-875.22. M.Kreuß,T.StrixnerandU.Kulozik,FoodHydrocolloids,2009,23,1818-1826.23. M.C.Hoyle,PlantPhysiol.,1977,60,787.24. L.Li,C.Li,Z.ZhangandE.Alexov,J.Chem.TheoryComput.,2013,9,2126-2136.25. F.Fogolari,P.Zuccato,G.EspositoandP.Viglino,Biophys.J.,1999,76,1-16.26. M.Jian,B.Liu,R.Liu,J.Qu,H.WangandX.Zhang,RSCAdv.,2015,5,48433-48441.