SupportingInformationTuningCoiledCoilStabilitywithHistidine-MetalCoordination IsabellTunn,aAlbertoS.deLéon,aKerstinG.Blank,a*MatthewJ.Harringtona,b*

a. I.Tunn,Dr.A.S.deLéon,Dr.K.G.Blank,Dr.M.J.HarringtonMaxPlanckInstituteofColloidsandInterfaces,ScienceParkPotsdam-Golm,14424Potsdam,Germany,*E-mail:Kerstin.Blank@mpikg.mpg.deb. Dr.M.J.Harrington,McGillUniversity,801SherbrookeSt.West,Montreal,QuebecH3A0B8,Canada,*E-mail:Matt.Harrington@mcgill.ca

TableofContents

TableofContents......................................................................................................................11Ramanspectroscopy..............................................................................................................22CDspectroscopy.....................................................................................................................23AFM-basedsingle-moleculeforcespectroscopy(SMFS)........................................................44Hydrogelpreparationandoscillatoryshearrheology..........................................................115References............................................................................................................................136Authorcontributions............................................................................................................13

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

2

1Ramanspectroscopy

Allcoiledcoil(CC)peptideswereobtainedfromacommercialsupplier(CenticBiotech).They

were synthesized using standard solid phase peptide synthesis methods and purified via

HPLC(purity>95%).Ramanmicro-spectroscopywasperformedtostudytheCCsecondary

structureandHis-Ni2+coordination.ThepeptidesHA4andHB4(nocysteine)weredissolved

inultrapurewatertoaconcentrationofapprox.5mgml-1.Thepeptidesweremixedina1:1

molarratiotoyield0.5mMHA4HB4.Forinvestigatingmetalcoordination,NiCl2wasaddedto

obtainafinalconcentrationof1mMandthepHwasraisedto8usingNaOH.HClwasadded

to obtain an acidic pH of ~4. The solution (5µl) was dried on a coverslip. For the

measurements of the resulting peptide films, a confocal Raman microscope (alpha300,

WITec) was used. The microscope was equipped with a piezo scanner (P-500, Physik

Instrumente) and a 50x objective (Nikon, NA 0.6). A linearly polarized laser (λ= 532nm,

Oxxius)wasfocusedontothesamplewithapolarizationangleof0°andnoanalyzerinthe

light path. The Raman scattered light was detected on a thermoelectrically cooled CCD

detector(DU401A-BV,Andor)withanintegrationtimeof2sand30accumulations.Spectra

fromfivedifferentpositionsonthesamplewerecollectedandaveraged.Themeasurement

andsubsequentdataanalysiswereperformedwiththesoftwareScanCtrlSpectroscopyPlus

(Version 1.38,WITec) and Project FOUR (Version 4.1WITec).OPUSwas used for baseline

correction(rubberbandmethod,linear,1pt),andthespectrawerenormalizedtotheamide

Ipeakat1654cm-1.

2CDspectroscopy

CDspectrawererecordedtoinvestigatethesecondarystructureandthethermalstabilityof

theindividualpeptidesaswellastheCCintheabsenceandpresenceofNiCl2.Peptidestock

solutionswerepreparedinultrapurewaterandtheirmolarconcentrationwasdetermined

with amino acid analysis. The stock solutions were mixed in a 1:1 molar ratio and

subsequentlydilutedin10mMPIPPS,137mMNaCl,2.7mMKCl,pH8.1(PIPPSBS)toobtain

afinalpeptideconcentrationof50µM.FormeasurementsinthepresenceofNi2+,NiCl2was

added to theHA4HB4 solution (1:3His:Ni2+) beforedilutionwithPIPPSBS. TheCD spectra

wereacquiredwithaChirascanqCDspectrometer(AppliedPhotophysics),equippedwitha

Peltiertemperaturecontroller(QuantumNorthwest).Aquartzcuvette(Hellma)withapath

3

length of 1mmwas used. All spectrawere recorded from200nm to 250nmwith a step

resolutionof1nmandabandwidthof1nm.Theintegrationtime(timeperpoint)wassetto

1snm-1.Thespectrawererecordedatatemperatureof20°C(threerepeats).Forbaseline

correction, the buffer spectra were recorded, using identical parameters, and subtracted

fromthepeptideandCCspectra.Themolarellipticity [q] [degcm2dmol-1]wascalculated

accordingtoEquation1:

𝜃 = 100 ∙ 𝜃𝑐 ∙ 𝑑

(𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛1)

withtheellipticityq [deg],theconcentrationc[M]andthepathlengthd[cm].1

Thermalunfoldingofthesampleswasperformedwithtemperaturerampsfrom4to90°C,

withaheating rateof1°Cmin-1. Full spectrawere recorded in1°C steps,using the same

parametersasusedforthesinglespectrawithashortenedintegrationtimeof0.7snm-1.All

measurements were performed in triplicate, unless stated otherwise. The melting

temperaturewasdetermined froma global fit to all spectra in the range from205nm to

250nm, using Global3 software (v.1.6.0.0, Applied Photophysics). For fitting, a single

transitionwasassumedandpre-andpost-transitionbaselinecorrectionwasapplied.

Fig.S1.CDspectraandthermalunfoldingoftheHis-modifiedCCandtheindividualpeptides.A)Fullspectraat20°C. B) Thermal unfolding of the peptides. For visualization, the values at the minimum at 222nm wereextractedfromallspectrameasured.Thetotalpeptideconcentrationwas50µMin10mMPIPPSBS,pH8.1.The data for HA4HB4 are the average of threemeasurements. The individual peptides weremeasured onlyonce.

Fig.S1A shows the full CD spectra of the individual peptides and of CC HA4HB4with and

withoutNi2+. TheCD spectrumof theCC shows two characteristicminima at 208nmand

4

222nm, indicatinga-helical secondary structure.2 The ratiobetween theminima r222/208 is

1.02,showingthatHA4andHB4areformingastableCC.3WhenNi2+isadded(1:3His:Ni2+),

thespectrumsuggestsahigherCCstability,evidencedbyanincreaseofr222/208to1.05anda

decrease of the absolute molar ellipticity. The individual peptides are largely unfolded.

ThermaldenaturationrevealedthattheindividualpeptideshaveameltingtemperatureTm

smaller than 4°C (Fig.S1B). HA4HB4 shows a thermal stability ofTm=76.4± 0.9°C,which

increases to 80.2±0.6°C in the presence of Ni2+. The values represent the mean ± the

standarderrorofthemean(SEM),n=3.

Fig. S2.CD spectra and thermal unfolding of the reference CC A4B4without His. A) Full spectra at 20°C. B)ThermalunfoldingoftheCC.Forvisualization,thevaluesattheminimumat222nmwereextractedfromallspectrameasured.Thetotalpeptideconcentrationwas50µMin10mMPIPPSBS,pH8.1.Thedatarepresentthemean±SEMofthreemeasurements.

CD spectroscopywas alsoperformedwith the referenceCCA4B4 todemonstrate that the

presenceofNi2+alonehasnoinfluenceontheconformationandstabilityoftheCC(Fig.S2).

A4B4showsthecharacteristicminimafora-helicalsecondarystructure.Thethermalstability

of A4B4 is identical in the absence (Tm=83.8±0.3°C) and presence of Ni2+

(Tm=85.0±0.1°C).

3AFM-basedsingle-moleculeforcespectroscopy(SMFS)

SurfaceandcantileverpreparationforSMFSwereperformedasdescribed inZimmermann

et al.4 The cantilevers (MLCT, Bruker) were cleaned in an UV-ozone cleaner (BioForce

Nanoscience)for20min.Foramino-silanization,thecantileversweresubmergedinpure3-

aminopropyldimethylethoxysilane(APDMES)(ABCR)for10minatroomtemperature(RT)

5

andwashedwithisopropanolandultrapurewater.Thecantileverswerecuredfor30minat

80°C.Thecoverslips(MenzelGläser)weresonicatedfor10mininisopropanoland3x5min

inwater,followedbyUV-ozonecleaningfor15min.Amino-silanizationwascarriedoutina

solutionof1%(v/v)ADPMESinabsoluteethanolfor1h.Afterwashing3xwithisopropanol

andwater,thecoverslipswerecuredat80°Cfor1h.Bothsurfacesweretreatedinparallel

for the next steps. First, the NHS-ester of the heterobifunctional NHS-PEG-maleimide

(10kDa, Rapp Polymere) was coupled to the amino-functionalized surface. The PEG was

dissolvedto30mMinboratebuffer(50mMH3BO3/Na2B4O7,pH8.5)andincubatedonthe

surfaces for 1h in a humidity chamber at RT. The surfaceswerewashed intensivelywith

wateranddriedundernitrogenflow.ThepeptidesHA4andHB4,carryingacysteineatthe

desired terminus, were dissolved to 1.5mM in coupling buffer (50mM Na2HPO4, 50mM

NaCl, 10mM EDTA, pH7.2) and stored at -20°C. Peptide stock solutions were diluted to

0.5mMincouplingbuffer.PeptideHB4wasincubatedonthecantileverandpeptideHA4was

incubatedonthecoverslipfor1hinahumiditychamberat4°C.Finally,thesurfaceswere

washedintensivelywithPIPPSBStoremovenon-covalentlyboundpeptides.

TheSMFSmeasurementswereperformedwithaForceRobot300(JPKInstruments)atRTin

PIPPS BS. Cantilever Cwith a nominal spring constant of 0.01Nm-1was used. The spring

constantwascalibratedusingthethermalnoisemethod5afterfinalizingthemeasurement.

Thevaluesrangedfrom0.014to0.025Nm-1.Therelativeset-pointforcewas~0.08nNand

theretractspeedwasvaried(50,200,400,1000,2500,5000nms-1;4000nms-1foronedata

set with Ni2+). Measurements were performed on a 10x10µm grid with 8x8 points. To

investigate the effect of Ni2+ coordination on CC stability, 1mM NiCl2 was added in the

buffer.

The data were analyzed using the JPKSPM data processing software (Version5.0.68, JPK

Instruments).TheforcecurvesshowthecharacteristicforceextensionbehaviorofthePEG

linkersusedtoimmobilizetheCC.Notethatthetotallengthoftwo10kDaPEGlinkers(one

on the surface and one on the cantilever) is approx. 130nm,6whereas the length of the

folded CC is only 4nm. The PEG linkers are used as an internal control, since their force

extension behaviour can be fitted with the wormlike-chain (WLC)model7 (Equation 2) in

ordertoobtainthepersistencelengthandthecontourlength:8-9

6

𝐹 𝑥 =𝑘6𝑇𝑙9

1

4 1 − 𝑥𝑙<

= −14+𝑥𝑙<

(𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛2)

wereFistheforce,lpthepersistencelength,lcthecontourlength,xtheend-to-enddistance

ofthepolymerchain,kBtheBoltzmannconstantandTthetemperature.

Onlyforcecurvesthatdisplayedasingleruptureevent(visual inspection),werefittedwell

withtheWLCmodelandpossessedacontourlengthlargerthan90nm(takingintoaccount

thepolydispersityofthePEG)wereconsideredforfurtheranalysis.Thesecriteriaefficiently

eliminate force distance curves displaying non-specific binding events, which frequently

possessmuchshortercontourlengthsorcannotbefittedwiththeWLCmodel.

Fig.S3.Rupture forceand loadingratehistogramsofHA4HB4 foraretractspeedof400nms-1.A)MeasurementwithoutNi2+.ThemostprobableruptureforceFRis32.3pNandthemostprobableḞis317pNs-1.B)Measurementwith1mMNi2+.Themost probableFR is 44.4pN at a loading rate Ḟ of 431pNs

-1. C)MeasurementwithNi2+ afterwashingwith 10mMEDTA.ThemostprobableFRis31.7pNataloadingrateḞof230pNs

-1.ThemostprobablevaluesofFRandḞwereobtainedfromaGaussianfittothehistograms(dashedline);nrepresentsthenumberofforcecurvesincludedineachhistogram.

The rupture forces (FR) and the loading rates (Ḟ=dF/dt) obtained from the JKPSPM data

processing software were plotted into histograms. For each retract speed, the most

probable rupture force and the most probable loading rate were determined using a

7

Gaussian fit to the corresponding histogram (logarithmic plot for Ḟ) in IGOR Pro6.37.

Representative rupture forcedistributionsobtained in thepresenceandabsenceof1mM

Ni2+andafterwashingwith10mMEDTAareshowninFig.S3(retractspeedof400nms-1).

Thehistogramsshowthattheruptureforceincreasedby~10pNinthepresenceofNi2+.This

stabilizationisreversedafterwashingthecantileverandsurfacewith10mMEDTA.Itshould

be noted that the histogram measured in the presence of Ni2+ is wider than the

correspondinghistogramsintheabsenceofNi2+orafteraddingEDTA.Weattributethisto

thepresenceofmultiple species, includingCCswhere zero,oneor twohelices containan

intactmetalcoordinationbridge.Fig.S4andS5showonerepresentativedatasetobtained

fromdynamicSMFSmeasurementsperformedintheabsenceandpresenceof1mMNi2+.A

summary of the results from three independentmeasurements performedwith different

cantileversandsurfacesisgiveninTab.S1.

8

Fig.S4.RepresentativedatasetofadynamicSMFSmeasurementofHA4HB4,performedintheabsenceofNiCl2.A)Ruptureforce histograms obtained at the indicated retract speeds. B) Corresponding loading rate histograms (plottedlogarithmically).ThedashedlinerepresentsaGaussianfittothehistogramstoobtainthemostprobableruptureforceandloadingrate;nrepresentsthenumberofforcecurvesincludedineachhistogram.

9

Fig.S5.RepresentativedatasetofadynamicSMFSmeasurementofHA4HB4,performedinthepresenceof1mMNiCl2.A)Rupture force histograms obtained at the indicated retract speeds. B) Corresponding loading rate histograms (plottedlogarithmically).ThedashedlinerepresentsaGaussianfittothehistogramstoobtainthemostprobableruptureforceandloadingrate;nrepresentsthenumberofforcecurvesincludedineachhistogram.

10

Tab.S1. SummaryofalldynamicSMFSdataofHA4HB4,measured in theabsenceand thepresenceof1mMNiCl2.DatashowninFig.S3andS4aremarkedinbold.*Datameasuredataretractspeedof4000nms-1.

cantilever 1 2 3

sample speed[nms-1]

FR[pN] Ḟ[pNs-1] n FR[pN] Ḟ[pNs-1] n FR[pN] Ḟ[pNs-1] n

HA4HB4 50 27.8 30 334 28.7 49 145 24.7 32 164

200 32.3 129 207 31 202 95 28.1 134 92

400 33.8 298 251 31.7 461 230 32.3 317 170

1000 37.9 924 272 36.2 1606 63 34.1 926 254

2500 42.5 3231 126 40.8 5006 54 42.8 3047 72

5000 50.7 10051 88 49.4 11715 33 - - -

HA4HB4+Ni2+

50 34.4 32 341 32.9 34.4 132 36.8 33.8 204

200 38.5 159 172 43.6 190 150 39.7 133 55

400 41.4 351 227 44.4 431 131 43.6 295 124

1000 43.4 1084 232 50.7 1281 122 47.2 953 126

2500 54.6 4037 76 53.1 4134 154 49 2691 99

5000 56.4* 4513* 28 - - - 56.2 6488 60

Foreachofthethreecantilevers,thevaluesofthemostprobableruptureforcewereplotted

against the logarithmof thecorresponding loadingrate.Thedatawas fittedwiththeBell-

Evans model10 (Equation3) to determine the extrapolated thermal off-rate at zero force

(koff)andthepotentialwidthΔxatagiventemperatureT(25°C):

𝐹@ 𝐹 =𝑘6𝑇∆𝑥

∙ ln𝐹 ∙ ∆𝑥

𝑘6𝑇∙ 𝑘DEE(𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛3)

Theobtainedfitvalues,koffandΔx,aswellastheirmean±SEMaresummarizedinTab.S2.

In thepresenceofNi2+,HA4HB4 is almostoneorderofmagnitudemore stable (lower koff)

than in theabsenceofNi2+. In contrast,Δx isnotaffectedsignificantlyby thepresenceof

Ni2+.

11

Tab.S2.ResultsoftheBell-EvansfitforHA4HB4,measuredintheabsenceandthepresenceofNi2+.

Sample HA4HB4 HA4HB4+Ni2+

Parameter koff[s-1] Δx[nm] koff[s-1] Δx[nm]

1 6.9·10-3 1.10 4.8·10-3 0.94

2 8.7·10-3 1.16 2.4·10-3 0.98

3 17.5·10-3 1.08 0.3·10-3 1.18

Mean±SEM 11.0±3.3·10-3 1.11±0.02 2.5±1.3·10-3 1.03±0.07

4Hydrogelpreparationandoscillatoryshearrheology

Thecysteine-containingpeptideswereusedtosynthesizeHA4HB4crosslinkedPEGhydrogels,

based on maleimide-functionalized star-shaped polyethyleneglycol (sPEG). For hydrogel

synthesis,theindividualpeptideswerefirstdissolvedinPBS(10mMNa2HPO4,2mMKH2PO4

137mM NaCl, 2.7mM KCl, pH7.4) in a concentration of 10mgml-1. To ensure that all

peptides are available for reaction with sPEG-maleimide, possible disulfide bonds were

reducedwithPierce™ImmobilizedTCEPDisulfideReducingGel(ThermofisherScientific)for

1.5hat4°Conamixer(2000rpm).ThereducedpeptideswerecoupledtosPEG(40kDa,4-

arm;JenkemTechnology)separatelytoobtainsPEG-HA4andsPEG-HB4.A1.2-foldto1.5-fold

excess of peptide was used to ensure that all arms of sPEG are functionalized. After

incubating the reaction mixture for 30min (RT, 800rpm), the excess of peptide was

removed using ultrafiltration (molecular weight cut-off 10kDa). During ultrafiltration (5x,

14000gfor10min)thebufferwasexchangedtoultrapurewater.ThepurifiedsPEG-peptide

conjugates were lyophilized and redissolved in PIPPS BS (pH8.1) to a concentration of

0.5mM. sPEG-HA4 and sPEG-HB4 were mixed in a ratio of 1:1 to induce CC formation.

Hydrogelsformedin lessthan1min.Airbubblesentrappedinthehydrogelwereremoved

by centrifugation for 2min at 2000g. To study the effect of Ni2+ coordination, NiCl2 was

addedinaratioof1:1His:Ni2+.Thereversibilityofmetalioncoordinationwasinvestigated

byaddingEDTAtoafinalconcentrationof10mM.

12

Fig.S6.AmplitudesweepsoftheCCcrosslinkedhydrogels(angularfrequencyof10rads-1).A)Measurementperformedinthe absenceofNi2+. Inset: picture of thehydrogel. B)Measurement performed in thepresenceofNi2+ (4mMNiCl2). C)Measurementperformedinthepresenceof4mMNiCl2and10mMEDTA.Theamplitudesweepswereconductedfrom1%to1000%stainandbackfrom1000%to1%totestforself-healing.

Theresultinghydrogelswerecharacterizedwithstrain-controlledoscillatoryshearrheology

(MCR 301, Anton Paar), using a 12mm diameter cone-plate geometry (CP12-1, angle 1°,

AntonPaar).Thegapwidthwasadjustedto0.02mm.Therheometerwasequippedwitha

temperature controlled hood (25°C) to prevent evaporation. To determine the linear

viscoelastic range, amplitude sweeps were performed ranging from 1% to 1000% shear

strain (Fig.S6) and vice versa. During the amplitude sweeps, the frequency was kept

constantat10rads-1.Afterashortrestingtimeof2min,afrequencysweepwasconducted

from0.001or0.002rads-1to100rads-1.Duringthefrequencysweeps,thestrainamplitude

waskept constantata strainof1%,which lies in the linearviscoelastic range. Frequency

sweepswereusedtoobtaininformationaboutthedynamicviscoelasticproperties(Fig.4in

themaintext).Therelaxationtimetofthenon-covalentcrosslinkscanbeobtainedfromthe

crossoverpointofG’ andG” in the frequencysweep.Tab.S3 containsanoverviewof the

13

relaxation times obtained from three independent measurements at each different

condition.

Tab.S3.Relaxationtimetof theCC-crosslinkedsPEGhydrogels.Shownarethevaluesobtainedfromthree independentmeasurements,includingthemean±SEM.

Sample nometal 1:1His:Ni2+ Ni2++EDTA

Parameter t[s] t[s] t[s]

1 72 270 50

2 135 286 33

3 70 217 48

Mean±SEM 92±12 258±36 44±4

5References1. S.M.Kelly,T.J.JessandN.C.Price,Biochim.Biophys.Acta,2005,1751,119.2. S.Y.Lau,A.K.TanejaandR.S.Hodges,J.Biol.Chem.,1984,259,13253.3. C.Aronsson,S.Dånmark,F.Zhou,P.Öberg,K.Enander,H.SuandD.Aili,Sci.Rep.,

2015,5,14063.4. J.L.Zimmermann,T.Nicolaus,G.NeuertandK.Blank,Nat.Protoc.,2010,5,975.5. J.L.HutterandJ.Bechhoefer,Rev.Sci.Instrum.,1993,64,1868.6. F.Oesterhelt,Rief,M.,Gaub,H.E.,NewJ.Phys.,1999,1,6.1.7. J.N.MilsteinandJ.-C.Meiners,inEncyclopediaofBiophysics,ed.G.C.K.Roberts,

SpringerBerlinHeidelberg,Berlin,Heidelberg,2013,pp.2757.8. F.Kienberger,V.P.Pastushenko,G.Kada,H.J.Gruber,C.Riener,H.SchindlerandP.

Hinterdorfer,SingleMol.,2000,1,123.9. T.HugelandM.Seitz,Macromol.RapidCommun.,2001,22,989.10. E.Evans,Annu.Rev.Biophys.Biomol.Struct.,2001,30,105.

6Authorcontributions

I.T., A.S.L, K.G.B. andM.J.H. designed the experiments. I.T. carried out and analyzed the

experiments. A.S.L. set up the rheology experiments and contributed to analyzing and

interpreting the hydrogels results. I.T., A.S.L., K.G.B andM.J.H. discussed the results. I.T.,

K.G.B.andM.J.H.wrotethemanuscript.