DNA-stabilizedsilvernanoclustersasspecific,ratiometricfluorescentdopaminesensorsJackson T. Del Bonis-O’Donnell,a,†,* Ami Thakrar,a Jeremy Wain Hirschberg,a Daniel Vong,aBridgetN.Queenan,a-bDeborahK.Fygenson,candSumitaPennathuraaDepartmentofMechanicalEngineering,UniversityofCalifornia,SantaBarbara,CA,USA.bNeuroscienceResearchInstitute,UniversityofCalifornia,SantaBarbara,CA,USA.c Department of Physics and Program in Biomolecular Engineering, University of California,SantaBarbara,CA,USA.

*Correspondingauthor:jtdo@berkeley.edu† Current Address: Department of Chemical and Biomolecular Engineering, University ofCaliforniaBerkeley,Berkeley,CA,USA.ABSTRACTNeurotransmitters are small molecules that orchestrate complex patterns of brain activity.Unfortunately, there exist few sensors capable of directly detecting individualneurotransmitters. Those sensors that do exist are either unspecific or fail to capture thetemporal or spatial dynamics of neurotransmitter release. DNA-stabilized silver nanoclusters(DNA-AgNCs)areanewclassofbiocompatible, fluorescentnanostructuresthathaverecentlybeendemonstratedtoofferpromiseasbiosensors.Inthiswork,weidentifytwodifferentDNAsequenceswhichformdopamine-sensitivenanoclusters.WedemonstratethateachsequencesupportstwodistinctDNA-AgNCscapableofprovidingspecific,ratiometricfluorescentsensingof dopamine concentration in vitro. DNA-Ag nanoclusters therefore offer a novel, low-costapproachtoquantificationofdopamine,creatingthepotentialforreal-timemonitoringinvivo.Keywords:Neurotransmitter;dopamine;biosensor;DNA-stabilizedsilvernanoclusters.INTRODUCTIONNeurotransmitters (NTs) are small molecules used in signaling pathways throughout thenervous system. The dynamic patterns of neurotransmitter signaling are critical to brainfunction and tightly regulated, both temporally and spatially. Improper release, reuptake,and/or degradation of neurotransmitters has been implicated in a wide range ofneurodevelopmental, neurological and neurodegenerative disorders, from epilepsy1 todepression2 to Parkinson’s disease.3Monitoring patterns of neurotransmitter release in thebrainwouldprovide fundamental insights intobrain functionandcould serveasadiagnostictool for maladaptive mental states and diseases. However, the ability to quantify specificendogenousneurotransmittersislimited.Current methods to detect neurotransmitters are predominantly electrochemical. Certainmolecules,includingthebiogenicamines(whichincludedopamine,epinephrine,serotonin,andascorbic acid) can be oxidized by an applied potential, allowing nearby microelectrodes todetectthechemicalreactionasanelectricalsignal.4Thetimeresolutionofthesemethodsisasfastastheelectrochemicalreaction,butthespecificityislow:theappliedpotentialwilloxidize

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

any biogenic amine present. Constant-potential amperometry has been adapted to increaseperformance; the resulting fast-scan cyclic voltammetry is now coupledwithmicrodialysis todetectthebiogenicamines5–7(aswellasoxygenandpH8)morereliablyinvivo.However, while electrochemical methods provide the millimeter spatial and millisecondtemporalresolutionadequateforneuro-signalsensing,9their lackofspecificityseverely limitssignal interpretation. Furthermore, the family of electrochemical techniques is simply notapplicable to other small molecules, including the major neurotransmitters, glutamate andGABA. Further research is needed to enable specific and reliable detection of individualneurotransmitterswithappropriatespatiotemporalresolution.

The use of optically active nanomaterials for specific binding and fluorescent detection is anemergingandpromisingapproachtoinvivosensingofbiomolecules.10–12However,widespreadutilitywillbelimitedifcomplexdesignorcostlysynthesisisrequiredforeachbiologicaltarget.DNA-stabilized fluorescent silver nanoclusters (DNA-AgNCs) are a new class of fluorescentnanostructureformedwhenaparticularsequenceofsingle-strandedDNAstabilizesafew-atomcluster of silver.13,14 DNA-AgNCs offer certain key advantages as biosensors.DNA-AgNCs arerelativelycheapandeasy tosynthesize,andarebiocompatible.15The fluorescenceofaDNA-AgNC is largely determined by the stabilizing DNA sequence and therefore readilymanipulated16–18.PerturbationstothelocalenvironmentorDNAconformationofDNA-AgNCscanproducemeasureablechangesintheiremissiveproperties.16,19ThesecombinedpropertieshaveledtotheuseofDNA-AgNCprobesasnovelfluorescentsensorsformetalionsandDNAsequences incomplexbiologicalmatrices.20–22WetheorizedthatDNA-stabilizedAgNCprobescould serve as fluorescent sensors for biologically relevant small molecules, such asneurotransmitters.HerewedemonstratethatDNA-AgNCscanbespecificallysensitivefordopamine.WeidentifytwodistinctDNA-AgNCswhich showpreference for dopamineover other neurotransmitters.These two DNA probes enable two classes of dopamine sensing: one provides a “turn-on”fluorescentsensorinthepresenceofdopamineandtheotherprovidesa“turn-off”fluorescentsensor. Importantly, we determined that these two DNA sequences each produced silvernanocluster populations with distinct spectral populations. The nanocluster populationsresponded differentially to dopamine, providing a ratiometric measure of dopamineconcentration in vitro. We validated that these DNA-AgNC probes were responsive todopamine, but insensitive to other common neurotransmitters, related catecholamines, andknown competitors. We therefore present DNA-AgNCs as novel, biocompatible sensors ofdopamine with potential for quantitative, ratiometric, optical in vivo sensing ofneurotransmitters.RESULTS&DISCUSSIONWe hypothesized that DNA-stabilized silver nanoclusters could be used as sensors of smallmolecule neurotransmitters. We selected six DNA ‘template’ sequences (sequences S1-S6,Table1)knowntostabilizesilveratomsintofluorescentnanoclusters,16,17,23,24andsynthesizedAgNCs on each using a standard approach: addition of AgNO3 followed by reduction using

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

NaBH4.We first determined the basal fluorescence of theDNA-AgNC created by each probe(Figure 1a). In these initial screens, we used UV excitation (260 nm) because it universallyexcites fluorescent DNA-AgNCs regardless of their specific visible excitation spectra.25 In theabsence of neurotransmitter, all six template sequences produced bright, stable DNA-AgNCswith emission peaks between 500-800 nm (Figure 1a) consistent with previous reports (seeFigureS1andSupportingInformationforspectraldetails).To identify potential DNA-AgNC sensors for neurotransmitter (NT) detection, we screened alibraryofeightwell-studiedNTs(acetylcholine,L-asparticacid,dopamine,epinephrine,GABA,L-glutamicacid,glycine,andhistamine,seeSupplementaryFigureS5 forchemicalstructures)against the six DNA-AgNCs (Figure 1b-c). We used the peak emission of each probe(representedaswhite) togauge thechange in intensityelicitedby thepresenceofNTeitherduringorafternanoclustersynthesis.WefirsttestedwhetheremissionintensitywasalteredbythepresenceofNTduringsynthesis(Figure1b).OfalltheNTstested,dopamineproducedthegreatest changes in DNA-AgNC fluorescence. When nanoclusters were assembled in thepresenceofdopamine,samplescontainingtemplatesS1-3fluorescedless,whiletemplatesS4-6fluoresced more than DNA-AgNCs assembled in the absence of neurotransmitter. We thendeterminedwhetherthefluorescenceofassembledDNA-AgNCnanoclusterswasalteredbytheaddition of NTs after synthesis (Figure 1c). Most DNA-AgNCs showed mild reductions influorescencewhenNTwasadded.However,whendopaminewasadded, the fluorescenceofassembledDNA-AgNCsformedusingsequencesS1,S2,S3andS5decreasedprofoundly.OfthesixDNA-AgNCstested,twoshowedpromiseforsensordevelopment.Thefluorescenceof nanoclusters assembledusingDNA sequence S4 increaseddramatically in thepresenceofdopamine during synthesis, though itwas largely unaffected by otherNTs (Figure 1b, boxedcolumn S4). Conversely, the fluorescence of assembled DNA S5 nanoclusters declineddramatically uponadditionof dopamine, butwasunchangedbyotherNTs (Figure1c, boxedcolumn S5). The fluorescence emission ofDNA S2-AgNCs andDNA S3-AgNCswere similar toDNA S5-AgNCs, but DNA S5-AgNCs exhibited a stronger quenching response and was moreselective to dopamine. We therefore focused our efforts on the two strongly dopamine-responsiveDNAsequences,S4andS5.DNAS4silvernanoclustersasratiometric,“turn-on”dopaminesensorsOur initial screens identified neurotransmitter-responsive DNA-AgNCs using universal UVexcitationat260nmandmonitoringemissionatthespecificwavelengthinthevisiblespectrumwhere intensitywas greatest in the absence ofNT. For our focused studies,we tailored theexcitation wavelength to the individual construct and collected emission spectra. We firstturnedourattentiontothesilvernanoclustersassembledbyDNAsequenceS4.CollectingafullemissionspectrumfromtheDNAS4-AgNCsusing525nmexcitationrevealedanemissionpeakat625nmthatmorethandoubledinintensitywhendopaminewaspresent(Figure2c)butwasbarelychangedbytheothercompoundstested(Figure2b).TheS4templatewasalsoexceptionalinthatitsexcitationwithUV(260nm)illuminationclearlyrevealedtwospectrallydistinctclusters(Figure2d).Whenilluminatedwithvisiblelightattheir

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

respectivepeakexcitationwavelengths,oneclusterhasapeakemissionwavelengthof625nm(asreportedpreviously23)andtheotherhaspeakemissionat730nm(notpreviouslyreported)(Figure 1a, Figure 2d). The presence of dopamineduring nanocluster formation dramaticallyincreasesthe625nmpeak,whiledecreasingthe730nmpeak(Figure2d).Wesystematicallytestedtheeffectofdopamineconcentrationontheformationofthesetwospectrally distinct populations. When DNA-AgNC were templated with sequence S4 in theabsence of dopamine, the formation of the near infrared-emitting (730 nm) populationwasfavored(Figure2d-e).However,asdopamineconcentration increasedfrom1µMto100µM,nIR fluorescence (730 nm emission) decreased approximately 66% and was effectivelyquenchedat1mMdopamine(Figure2e). Incontrast, redfluorescence(collectedat615nm)doubled in intensity asdopamine concentration increased from1µM to100µM (Figure2e)andwasnotquencheduntil10mMdopamine.Itshouldbenotedthatthefluorescenceofbothconstructsdecreasedabove100µMdopamineand,atthesehighconcentrations,weobservedsilver nanoparticle formation and aggregation, which may have been induced directly bydopamine,12 or by competition for silver between the bases and the excess chloridepresentfromusingdopamine-HClsaltinourpreparation.Although nanoparticle formation and aggregation prevent quantitative interpretation atdopamineconcentrationsover100µM(Figure2e,FigureS3),atconcentrationsbelow100µM,thetwoemissionpeakstogetherprovideareliable,ratiometricsignal(Fred/FnIR)thatchangesbynearlytwoordersofmagnitude(FigureS3),agreaterdynamicrangethaneithersignalaloneand comparable to the range of dopamine concentrations it reflects.We therefore proposethat the relative fluorescence of the two spectrally distinct and differentially dopamine-sensitiveDNAS4-AgNCpopulations(Figure2a)canprovidearobust,quantitativeandsensitivemeasureofdopamineconcentrationinvitro.DNAS5silvernanoclustersasratiometric,“turn-off”dopaminesensorsWe next turned our attention to nanoclusters stabilized by DNA sequence S5. In our initialscreens,weobservedthat,whenexcitedwithUVlight(260nm),S5-AgNCshadapeakemissionwavelengthof600nm(Figure1a)andafluorescenceintensitythatwaslargelyunaffectedbytheadditionofmostneurotransmitters(Figure1c,S5column).However,whendopaminewasadded,DNAS5-AgNCsexhibiteda~80%decreaseinfluorescenceintensity(Figure1c).Basedontheseobservations,wescannedexcitationwavelengthswhileholdingemissionfixedat600nmandidentified535nmasthepeakvisibleexcitationwavelength.Wethenre-assessedtheDNAS5-AgNC response to neurotransmitters. Encouragingly, addition of dopamine reducedfluorescenceatthe600nmemissionpeakevenmoreunder535nmexcitationthanunder260nm excitation, diminishing by nearly two orders of magnitude (96% decrease, Figure 3b-c).Meanwhile,changes influorescence inresponsetotheotherNTswere lessthan11%(Figure3b).Uponcloserinspection,wenoticedthisresponsetodopaminewasactuallyacompoundeffect:under260nmexcitationaddingdopaminetoS5-AgNCsbothdecreasedemissionat600nmandshifted the peak emission from 600 nm to 560 nm (Figure 3d). Full excitation and emission

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

spectraoftheDNAS5-AgNCsrevealedtwoclusterpopulationswithdistinctspectralproperties:onewithapeakemissionat650nm(580nmexcitation)andanotherwithapeakemissionat575 nm (495 nm excitation) (Figure S4). These populationswere not resolved using 260 nmexcitation(Figure1a,Figure3d)andtheirsumproducedanapparentemissionpeakat600nm.Weexaminedthedopaminesensitivityofthesetwospectrallydistinctnanoclusterpopulationsgenerated by DNA sequence S5 individually. The shorter wavelength population (495 nmexcitation,575nmemission)remainedstableupto100µMdopamineandwasonlyquenchedathigherconcentrationswhenaggregationwasobserved(Figure3e,greencircles).Bycontrast,dopamineeffectivelyquenchedthefluorescenceofthelongerwavelengthpopulation(580nmexcitation,650nmemission)intherangeof5-50µMdopamine(Figure3e,redtriangles).Weconcludethatthequench+shiftphenomenonobservedwithUVexcitation(Figure3d)wasdueto theselectivequenchingof the longerwavelengthpopulationby themoderateamountsofdopamineusedinourinitialscreen(50µM,schematicinFigure3a).Wethereforeproposethatthe relative fluorescence of these two spectrally distinct S5-AgNC populations, under visibleexcitation, can be used as a fluorescent, ratiometric (F575/F650) and quantitative “turn-off”sensor(F575:invariantcalibration,F650:indicator)ofdopamineconcentrationinvitro.SpecificityofsilvernanoclustersfordopamineoverknownhomologuesandcompetitorsEncouragingly, both S4- and S5-templated silver nanoclusters were highly responsive todopamine and relatively insensitive to the other neurotransmitters tested. However, wewondered if the sensors would be specific to dopamine over more closely related smallmolecules. We therefore interrogated both S4- and S5-stabilized AgNCs with catecholamineprecursors andmetabolites of dopamine (L-tyrosine, phenylalanine, L-DOPA, andDOPAC), aswellaswithsmallmoleculesknowntointerferewithdopaminesensing(L-tryptophan,anduricacid)(seeSupplementaryFigureS6forchemicalstructures).10,26Wetestedwhethertheserelatedmoleculeswerecapableofenhancingtheformationofvisible(615nmemission)fluorescentnanoclustersbyDNAsequenceS4.Under525nmillumination,synthesisofS4-AgNCsinthepresenceofdopamineagainproducedarobustincreasein615nmemission. However, synthesis in the presence of the other related molecules failed tosignificantlyenhancefluorescentemission(Figure4a).Wefurthertestedwhethertheseothermolecules were capable of quenching S5-AgNC fluorescence. Again, we observed that theaddition of dopamine to existing S5-AgNC caused robust quenching of fluorescence (600 nmemission, 535 nm excitation) while S5-AgNC was unaffected by addition of other testedmolecules(Figure4b).Weconcludethatcatecholamineprecursors,dopaminemetabolites,andknown dopamine competitors do not significantly alter the fluorescence of these two DNA-AgNCs.TobetterunderstandthemechanismofdopamineinteractionwithDNA-silvernanoclusters,weinvestigatedtheroleplayedbythecatecholgroupintheinteractionsbetweendopamineandDNA-AgNCs. We compared the dopamine response of S5- and S4- templated AgNCs to thecorrespondingcatecholresponse(Figure4c-d).Again,weaddeddopamine,catechol,orcontrolsolution(buffer)priortonanoclustersynthesisforconstructS4andafternanoclustersynthesis

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

forconstructS5.Forbothsensors,catecholmodulatednanocluster fluorescence inamannersimilartodopamine,albeittoalesserextent(Figure4c-d).Likedopamine,catecholincreasedtheformationofvisible(615nmemission)nanoclusters,whileslightlydecreasingtheformationof nIR nanocluster (730 emission) (Figure 4c). Catechol also induced the quench+shiftphenomenon,thoughtoalesserdegreethandopamine(Figure4d).However,dopaminehadamore pronounced ability to prevent nIR nanocluster assembly (Figure 4c). This similaritysuggests the catechol group is critical to the perturbation of DNA-AgNC fluorescence bydopamine.However,othermoleculeswe tested containinga catechol group (namelyDOPACandL-DOPA)inducedessentiallynofluorescentchange(Figure4a-b).Wehypothesizethatselectivityfordopamineisbasedonthesmallsizeofitsalkylaminegroup.Allothercatecholaminestestedcontainedabulkieralkylgroup(seeSupplementaryFigureS6forchemicalstructures),whichcouldstericallyhinderinteractionswiththeDNA-AgNCcomplex.WepresumethatdirectinteractionsbetweendopamineandtheDNAstrandperturbtheDNA-silver nanocluster conformation.16 Dopamine binding to template S4 before nanoclustersynthesismay destabilize the conformation necessary to form the near infrared population.Similarly,dopaminecoulddestabilizethe650nmnanoclusterpopulationformedbysequenceS5byperturbingtheconformationoftheDNA.However,dopaminecouldalsopromotedirectconversion from one species to anotherwith a different emissionwavelength. Alternatively,dopaminemay irreversibly interactwith theDNAbases, resulting incleavageor formationofadducts.Inpreviousreports,quinonesformedbytheoxidationofdopaminewereobservedtoquench the fluorescence of certain AgNCs, potentially through irreversibleDNA damage.27–30SuchmodificationstotheDNAbaseswouldproducechangesinfluorescencebyeliminatingAg+bindingdomainsororbital interactionsbetweenDNAbasesandtheAgNC,orbyperturbingaconformation critical to cluster formation or the continued stability of a cluster aftersynthesis.12,28,29,31 Elucidating the mechanism of dopamine action (and especially itsreversibility) will be critical to establishing this technique as a viable option for in vivomonitoring.We conclude that DNA-stabilized AgNCs can serve as selective, sensitive, ratiometricfluorescent sensors of dopamine concentration. Specifically, we identify stabilizing DNAsequences for fluorescent silver nanoclusters that support two complementary dopaminedetection approaches in vitro: one (S4) provides a “turn-on” sensor that reports on thepresenceofdopamineduringDNAS4-AgNCsynthesisandtheother(S5)providesa“turn-off”sensorthatisquenchedbytheadditionofdopaminetopre-formedDNAS5-AgNCs.BothDNAsequencesproduce twodistinctemissionpeaks, therebyenablingquantificationofdopamineconcentrationusingratiometricfluorescentsensing.Ratiometric sensorsoffer importantadvantagesoverconventional fluorescent sensors,whichreport on the presence of a targetmolecule through a fractional change in overall emissionintensity,∆ƒ=(F-F0)/F0,atasingleemissionwavelength32.Inconventionalfluorescentsensing,themagnitudeofthesignalisoftensmallandbothbaselineandtarget-modulatedfluorescencearesusceptibletointerference(duetonon-uniformconcentrationdistribution,photobleaching,aggregation or nonspecific interactionswith the environment, such as local pH changes,33,34)

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

that complicates interpretation. Ratiometric fluorescent sensors, with their two spectrallydistinct fluorescent signals, can be more responsive and less susceptible to interferences,especiallyifthefluorophoresarechemicallyrelatedand/orrespondsimilarlytoenvironmentalconditions.DNAS5-AgNCs’ratiometricsignalistypicalinthatoneoftheemissionschangesinresponsetotarget(dopamine)whiletheotherremainsinvariant,andsocanbeusedtoscaleornormalize the indicator signal. DNA S4-AgNCs’ ratiometric signal offers enhanced sensitivitybecauseitstwoemissionshaveoppositedependencyondopamine.With proper calibration, ratiometric sensing can report not merely the presence of anddirectional change in a target analyte, but also its concentration. Basal dopamineconcentrations in relevant brain areas, including the extracellularmatrix of the striatum, arecommonly 1nM-1µM, though they can be higher in the vicinity of the synapse shortly afterrelease.7,35–38 While the synthesis-dependent signal from DNA S4-AgNCs is sensitive in thisrange(10nM-100µM),thesensitivityandworkingrangeofthepost-synthesisresponsive(andtherefore in vivo compatible) DNA S5-AgNC nanosensor (1-100 µM) falls outside thesecommonly measured values. Ratiometric fluorescent sensors are generally challenging toengineer,34 but the sequence modularity of DNA-AgNCs offers a flexible new toolkit.Improvements in sensitivity might be achieved through directed mutation or automatedscreensofDNAsequenceandnanoclustersynthesisspace.Giventhedocumentedsuccessesofsuch screens39 as well as the ability to engineer FRET pairs,40 there is a real possibility ofdiscovering clusters with sufficient sensitivities and temporal responses to enable real-time,quantitativeimagingofneurotransmittersinvivo.In summary, we present DNA-silver nanoclusters as novel, low-cost, easy-to-assemble,biocompatible reporters that offer selective, optical, quantitative sensing of dopamineconcentration in vitro.With furtherdevelopment,DNA-stabilizedAgNCsmayenable spatiallydistributed,real-time,quantitative,opticalmonitoringofdopamineinvivo.FIGURES

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

Figure1.ChangesinDNA-AgNCfluorescenceinthepresenceofneurotransmitter.Top panels: Schematic representation of experimental approach. (a) We synthesized silvernanoclusters(AgNCs)using6DNAsequences(S1-S6)anddeterminedthefluorescenceemissionspectraofeachnanocluster.Fortheseinitialscreens,weusedUVexcitation(260nm,purple)because it universally excites fluorescent DNA-AgNCs regardless of their specific visibleexcitationspectra.25Bottom:FluorescenceemissionspectraofeachDNA-AgNC.DNAsequencesS2andS3hademissionspectrasimilartothoseofS5andwerethusomittedforclarity.(b-c)Peak fluorescence emission response of the 6 DNA-AgNCs (x-axis, [DNA] = 12.5 µM) in thepresenceof 8 neurotransmitters (NTs, y-axis, [NT] = 50µM). Baseline fluorescence for eachsequence(white)wasestablishedbyexcitingeachDNA-AgNCintheabsenceofNT.Sensitivityof thenanoclusters toNTwas assayedunder two conditions: neurotransmitterswere addedeither before (b) or after (c) the assembly of fluorescent nanoclusters. Relative fluorescencewasthenquantifiedforeachsequence:redindicatesincreasedwhileblueindicatesdecreasedfluorescence in thepresenceofNTrelativetobaselineDNA-AgNCfluorescenceforeachDNAsequence.Uniquely,DNAS4nanocluster fluorescencedramatically increasedwhendopaminewaspresentduringnanocluster synthesis (boxed columnS4and schematic inb). Conversely,DNA S5 nanocluster fluorescence dramatically decreased when dopamine was added aftersynthesis(boxedcolumnS5andschematicinc).

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

Figure2.DNAS4silvernanoclustersactsasratiometric,“turn-on”dopaminesensors.(a) Schematic summary of experimental results. DNA sequence S4 produces two spectrallydistinct AgNC populations. Formation of the long-wavelength (nIR) population (640 nmexcitation,730nmemission,top)isfavoredintheabsenceofdopamine,whileformationoftheshort-wavelength(visible)population(535nmexcitation,615nmemission,bottom)isfavoredin thepresenceofdopamine. (b)Change in fluorescenceemission intensity (∆ƒ= [F-F0]/F0)ofAgNCssynthesizedusingDNAS4inthepresenceofeightdifferentNTs([DNA]=12.5µM,[NT]=50 µM). Fluorescence emission was collected at 610 nm using 525 nm excitation (peakexcitation for the S4-AgNCs). Error bars indicate the standard deviation of n=3 replicateexperiments.(c-d)FluorescenceemissionspectraofS4-AgNCsassembledintheabsence(gray)andpresence(blue)ofdopamineunder525nm(c)and260nm(d)excitation.Inthepresenceof dopamine, intensity at the lowerwavelength emission peak (615 nm)more than doubleswhileintensityatthehigherwavelengthpeak(730nmemission)dropsacomparableamount.(e)FluorescenceemissionresponseofthetwospectrallydistinctAgNCsgeneratedbyS4whendifferent concentrations of dopamine are added prior to cluster synthesis. Orange triangles:535 nm excitation, 615 nm emission nanoclusters; red circles: 640 nm excitation, 730 nmemissionnanoclusters.OneofthetwoAgNCsspecies(730nmemission,red)wassignificantlyquenched by moderate amounts of dopamine, while the other species (615 nm emission,orange)wasenhancedbymoderateamountsofdopamine.

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

Figure3.DNAS5silvernanoclustersactsasratiometric,“turn-off”dopaminesensors.(a) Schematic summary of experimental results. DNA sequence S5 produces two spectrallydistinctAgNCpopulations.Fluorescenceofthelong-wavelengthpopulation(580nmexcitation,650nmemission,top)isquenchedbydopamineinthe1-10µMrange,whilethatoftheshort-wavelength population (495 nm excitation, 575 nm emission, bottom) is not significantlyaffected by dopamine concentrations up to 50 µM. (b) Change in fluorescence emissionintensity(∆ƒ=[F-F0]/F0)ofassembledDNAS5-stabilizedAgNCsuponadditionofeightdifferentNTs([DNA]=12.5µM,[NT]=50µM).Fluorescenceemissionwascollectedat600nmusing535nm excitation (peak visible excitation for the S5-AgNCs). Error bars indicate the standarddeviationofn=3 replicateexperiments. (c-d)FluorescenceemissionspectraofassembledS5-AgNCswithout(black)andwithdopamine(blue)under535nm(c)and260nm(d)illumination.Note the decrease in overall intensity and the shift in peak emission to shorterwavelengthsunder UV illumination (d). (e) Fluorescence emission response of the two spectrally distinctAgNCsgeneratedbyS5fordifferentconcentrationsofdopamineaddedafterclustersynthesis.Greencircles:495nmexcitationand575nmemission;redtriangles,580nmexcitationand650nm emission. Only one of the two AgNCs species (650 nm emission, red) was significantlyquenchedbymoderateamountsofdopamine.[DNA]=1.25µM.

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

Figure 4. Preference of DNA-silver nanocluster sensors for dopamine over catecholaminesandmetabolites.(a-b)Changeinfluorescenceemissionintensity,∆ƒ=(F-F0)/F0,ofnanoclustersassembledbytemplates(a)S4or(b)S5inthepresenceofdopamine,dopaminehomologuesorknowncompetitors([DNA]=12.5µM,[NT]=50µM).Bothconstructsshowedhighspecificityfordopamine. (c-d) Fluorescenceemissionspectraofnanoclustersassembledby templateS4(c) or S5 (d) in thepresenceofdopamine, catecholorNT-free solution.Note:As inpreviousexperiments,smallmoleculeswereaddedbeforesynthesisofS4(excitation:525nm,emission:615nm)(a,c)andaftersynthesisofS5(excitation:535nm,emission:600nm)(b,c).

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

METHODSChemicals.Unless otherwisenoted, chemicalswerepurchased fromSigmaAldrich (St. Louis,MO, USA): Dopamine hydrochloride, Acetylcholine chloride (≥99% TLC), Histaminedihydrochloride (≥99.0% AT), γ-Aminobutyric acid (GABA) (BioXtra, ≥99%), L-Aspartic Acid(reagent grade, ≥98% HPLC), Glycine (from non-animal source, meets EP, JP, USP testingspecifications, suitable for cell culture, ≥98.5%), Epinephrine hydrochloride, L-Glutamic acid(ReagentPlus®, ≥99% HPLC), Uric acid (≥99%, crystalline), 3,4-Dihydroxy-L-phenylalanine (L-DOPA)(≥98%TLC),L-tryptophan(reagentgrade,≥98%HPLC)),L-Tyrosine(BioUltra,≥99.0%NT),3,4-Dihydroxyphenylacetic acid (DOPAC) (98%), L-Phenylalanine (reagent grade, ≥98%), (−)-Norepinephrine(≥98%,crystalline).Oligonucleotides. DNA oligos (detailed in Table 1) were purchased from Integrated DNATechnologies,Inc.(Coralville, IA,USA)withstandarddesalting,hydratedwithMilliQdeionizedwater(DI),andstoredat-20°Cinstockconcentrationsofbetween100µMand1mM.Sodiumphosphate buffer was prepared by titrating solutions of sodium phosphate monobasic andsodiumphosphatedibasic(FisherScientific,Waltham,MA,USA)toapHof7.3andcompositeconcentrationof200mM.DNA-containingsolutionswerepreparedbycombining25µLof100µMDNAsolutionwith50µLof20mMsodiumphosphatebuffer,heatingat90°C for5min,thensnapcoolingbyrunningthesampletubeundercoolDIwaterfor10to20seconds.Table1.Oligonucleotidessequences# Sequence ReferenceS1 CGTGTCCCCCCCCGATTTTTATCGGCCCCCCACACG J.T.DelBonis-O’Donnell,S.PennathurandD.K.

Fygenson,Langmuir,2016,32,569–576.S2 CGTGACTGTACCCCCCCCCCCCTACAGTCACGS3 GCACAGGCTACCCCCCCCCCCCTAGCCTGTGCS4 ACCCGAACCTGGGCTACCACCCTTAATCCCC J.Sharma,R.C.Rocha,M.L.Phipps,H.Yeh,K.A.

Balatsky,D.M.Vu,A.P.Shreve,J.H.Werner,J.S.Martinez,C.L.AlamosandL.Alamos,Nanoscale,2012,4,4107–10.

S5 GGGTTAGGGTCCCCCCACCCTTACCC T.Li,L.Zhang,J.Ai,S.DongandE.Wang,ACSNano,2011,6334–6338.

S6 TTCACCGCTTTTGCCTTTTGGGGACGGATA D.Schultz,K.Gardner,S.S.R.Oemrawsingh,N.Markešević,K.Olsson,M.Debord,D.BouwmeesterandE.Gwinn,Adv.Mater.,2013,25,2797–2803.

Cluster synthesis and neurotransmitter screening. Neurotransmitter and screening librarychemicalswerepreparedas10-100mMaqueousstocksolutionsat4°C.Silvernitrate(AgNO3,99.99%),sodiumborohydride(NaBH4,99.99%,pellets)andsodiumhydroxide(NaOH,>97.0%,pellets) solutionswere prepared fresh for synthesis of silver nanoclusters (AgNC). Dopaminesolutions were prepared fresh daily, and protected from light. Note: While someneurotransmittersweresuppliedastheirHClsalt,weobservednosystematicresponseofAgNCfluorescencetothepresenceofHClduringourinitialscreens,indicatingthattheresponsesweobservedresultedfromtheneurotransmitterspeciesandnotinteractionbetweenAg+andCl-.ForMethod1(effectofNTonassemblyoffluorescentDNA-AgNCs,Figure1b),50µLof100µMsolutionsofeachneurotransmitter(orDI)wereaddedtoDNAsolutions(75µL),andallowedto

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

incubateforfifteentotwentyminutesatroomtemperature.Tocreatenanoclusters,12.5µLof0.16mMsilvernitratewereaddedtoeachsample,followedby12.5µLoffreshlyprepared0.8mMsodiumborohydridedilutedin100µMsodiumhydroxide.Solutionswerevortexedandlefttoincubateat4°Cfortwohours.Then20µLofeachDNA-AgNC/neurotransmittersolutionwastransferred to a 384-well plate (Product #3821BC, Corning, Inc.) for fluorescencemeasurements. The final concentrationofDNAwas16.7µM.ForMethod2 (effectofNTonassembledDNA-AgNCfluorescence,Figure1c),12.5µLof1.6mMAgNO3wasaddedtotheDNAsolution and allowed to incubate in the dark for 10-20 minutes. Next, 12.5 µL of freshlyprepared 0.8 mM NaBH4 diluted in 100 µM NaOH was added to the DNA-Ag solution. Thesolutionwasvortexedandlefttoreactinthedarkfor16-20hoursat4°C.10µLsamplesoftheresulting DNA-AgNCs were placed in a 384-well plate (Product #3821BC, Corning, Inc.) andcombinedwith10µLofa100µMneurotransmittersolutionorDIwaterasacontrol.Thefinalconcentration of DNA was 12.5 µM. Each sample was prepared in triplicate. The well plateincubatedfor2-4hoursinthedarkat4°Cuntilmeasurementsweretaken.Fluorescence detection. Fluorescence excitation and emission spectra of each sample weremeasured using a Tecan Infinite 200Pro plate reader with Tecan i-control software (TecanGroupLtd.)between1and5hoursaftereitherchemicalreductionoradditionofNTs.Stepsizeandbandwidthweresetto2nmand20nmforemissionscans,andto5nmand5-10nmforexcitation scans, respectively. All scanswere performed using TopMode. The gainwas keptconstantat105(whenusing260nmexcitation)orat75(whenusingvisibleexcitation).Emptywellswereprocessedtocheckforautofluorescenceofthewellplate.ForeachDNA-AgNC,werecorded the proportional change in fluorescence, ∆ƒ = (F-F0)/F0, of a single emissionwavelengthinthepresenceofeachNT(Figure1b-c). Weusedthepeakemissionintensityatpeak visible excitation in the absence of NTs (F0) and the emission intensity at the sameexcitationandemissionwavelengthsinthepresenceofNTs(F).

SAFETY.Nounexpected,new,and/orsignificanthazardsorrisksassociatedwiththereportedworkwereencountered.

SUPPORTINGINFORMATION.Supporting Information is available free of charge on the ACS Publications website.FluorescenceemissionspectraofallDNAsequences(SupplementaryFigure1);distinctspectralpopulationsforsequenceS4(SupplementaryFigure2)andsequenceS5(SupplementaryFigure4);ratiometricresponsestodopamineconcentration(SupplementaryFigure3);andstructuralformulas for neurotransmitters and small molecules in the screening library (SupplementaryFigures5and6)areprovided.ABBREVIATIONS.DNA-AgNC=DNA-stabilizedsilvernanoclusterDOPAC=3,4-DihydroxyphenylaceticacidL-DOPA=3,4-Dihydroxy-L-phenylalanineGABA=γ-AminobutyricacidNT=neurotransmitter

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

AUTHORINFORMATIONCorrespondingAuthorjtdo@berkeley.eduPresentAddresses†DepartmentofChemicalandBiomolecularEngineering,UniversityofCalifornia,Berkeley,CA,USAORCIDJacksonTravisDelBonis-O’Donnell:0000-0002-9135-2102BridgetNQueenan:0000-0002-1739-3424Authorcontributions J.T.D.O.,D.K.F.andS.P.designed the research. J.T.D.O.,A.T, J.W.H.andD.V. conducted experiments and performed data analysis. J.T.D.O., B.N.Q., D.K.F. and S.P.preparedandwrotethemanuscript.Fundingsourcesandconflictsofinterest.TheauthorsdeclarenocompetingfinancialinterestsandacknowledgesupportfromNationalScienceFoundationgrantNSF-CHE-1213895.Thisworkwas partially supported by the Institute for Collaborative Biotechnologies through grantsW911NF-09-0001andW911NF-12-1-0031fromtheU.S.ArmyResearchOffice.ThecontentoftheinformationdoesnotnecessarilyreflectthepositionorpolicyoftheU.S.Government,andno official endorsement should be inferred. The authors also acknowledge the use of theBiologicalNanostructuresLaboratorywithintheCaliforniaNanoSystemsInstitute,supportedbythe University of California, Santa Barbara, and the University of California Office of thePresident.ACKNOWLEDGMENTTheauthorsthankS.Helmyforhelpfuldiscussions.REFERENCES(1)Morimoto,K.,Fahnestock,M.,andRacine,R.J.(2004)Kindlingandstatusepilepticusmodelsofepilepsy:rewiringthebrain.ProgNeurobiol73,1–60.(2)Duman,R.S.,Aghajanian,G.K.,Sanacora,G.,andKrystal,J.H.(2016)Synapticplasticityanddepression:newinsightsfromstressandrapid-actingantidepressants.Nat.Med.22,238–249.(3)Lotharius,J.,andBrundin,P.(2002)Pathogenesisofparkinson’sdisease:dopamine,vesiclesandα-synuclein.Nat.Rev.Neurosci.3,932–942.(4)Kissinger,P.,Hart,J.,andAdams,R.(1973)Voltammetryinbraintissue--anewneurophysiologicalmeasurement.BrainRes55,209–213.(5)Park,J.,Takmakov,P.,andWightman,R.M.(2011)Invivocomparisonofnorepinephrineanddopaminereleaseinratbrainbysimultaneousmeasurementswithfast-scancyclicvoltammetry.J.Neurochem.119,932–944.(6)Bunin,M.A.,Prioleau,C.,Mailman,R.B.,andWightman,R.M.(1998)Releaseanduptakeratesof5-hydroxytryptamineinthedorsalrapheandsubstantianigrareticulataoftheratbrain.J.Neurochem.70,1077–87.

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

(7)Venton,B.J.,Zhang,H.,Garris,P.A.,Phillips,P.E.M.,Sulzer,D.,andWightman,R.M.(2003)Real-timedecodingofdopamineconcentrationchangesinthecaudate-putamenduringtonicandphasicfiring.J.Neurochem.87,1284–1295.(8)Venton,B.J.,Michael,D.J.,andWightman,R.M.(2003)CorrelationoflocalchangesinextracellularoxygenandpHthataccompanydopaminergicterminalactivityintheratcaudate-putamen.J.Neurochem.84,373–81.(9)Adams,K.L.,Puchades,M.,andEwing,A.G.(2008)InVitroElectrochemistryofBiologicalSystems.Annu.Rev.Anal.Chem.1,329.(10)Kruss,S.,Landry,M.P.,VanderEnde,E.,Lima,B.M.a,Reuel,N.F.,Zhang,J.,Nelson,J.,Mu,B.,Hilmer,A.,andStrano,M.(2014)Neurotransmitterdetectionusingcoronaphasemolecularrecognitiononfluorescentsingle-walledcarbonnanotubesensors.J.Am.Chem.Soc.136,713–724.(11)Wang,H.-B.,Zhang,H.-D.,Chen,Y.,Huang,K.-J.,andLiu,Y.-M.(2015)Alabel-freeandultrasensitivefluorescentsensorfordopaminedetectionbasedondouble-strandedDNAtemplatedcoppernanoparticles.SensorsActuatorsBChem.220,146–153.(12)Lin,Y.,Chen,C.,Wang,C.,Pu,F.,Ren,J.,andQu,X.(2011)SilvernanoprobeforsensitiveandselectivecolorimetricdetectionofdopamineviarobustAg-catecholinteraction.Chem.Commun.(Camb).47,1181–3.(13)Petty,J.T.,Zheng,J.,Hud,N.V,andDickson,R.M.(2004)DNA-templatedAgnanoclusterformation.J.Am.Chem.Soc.126,5207–12.(14)Gwinn,E.G.,Schultz,D.E.,Copp,S.M.,andSwasey,S.M.(2015)DNA-ProtectedSilverClustersforNanophotonics.Nanomaterials5,180–207.(15)Choi,S.,Dickson,R.M.,andYu,J.(2012)Developingluminescentsilvernanodotsforbiologicalapplications.Chem.Soc.Rev.41,1867–91.(16)DelBonis-O’Donnell,J.T.,Pennathur,S.,andFygenson,D.K.(2016)ChangesinSpectraandConformationofHairpinDNA-StabilizedSilverNanoclustersInducedbyStemSequencePerturbations.Langmuir32,569–576.(17)Li,T.,Zhang,L.,Ai,J.,Dong,S.,andWang,E.(2011)Ion-TunedDNA/AgFluorescentNanoclustersasVersatileLogicGates.ACSNano5,6334–6338.(18)Li,J.,Zhong,X.,Zhang,H.,Le,X.C.,andZhu,J.-J.(2012)Binding-inducedfluorescenceturn-onassayusingaptamer-functionalizedsilvernanoclusterDNAprobes.Anal.Chem.84,5170–4.(19)DelBonis-O’Donnell,J.T.,Vong,D.,Pennathur,S.,andFygenson,D.K.(2016)AuniversaldesignforaDNAprobeprovidingratiometricfluorescencedetectionbygenerationofsilvernanoclusters.Nanoscale8,14489–14496.(20)Obliosca,J.M.,Liu,C.,andYeh,H.-C.(2013)FluorescentsilvernanoclustersasDNAprobes.Nanoscale5,8443–61.(21)Liu,J.(2014)DNA-stabilized,fluorescent,metalnanoclustersforbiosensordevelopment.TrACTrendsAnal.Chem.58,99–111.(22)Latorre,A.,andSomoza,Á.(2012)DNA-mediatedsilvernanoclusters:synthesis,propertiesandapplications.Chembiochem13,951–8.(23)Sharma,J.,Rocha,R.C.,Phipps,M.L.,Yeh,H.,Balatsky,K.A.,Vu,D.M.,Shreve,A.P.,Werner,J.H.,Martinez,J.S.,Alamos,C.L.,andAlamos,L.(2012)ADNA-templatedfluorescentsilvernanoclusterwithenhancedstability.Nanoscale4,4107–10.(24)Schultz,D.,Gardner,K.,Oemrawsingh,S.S.R.,Markešević,N.,Olsson,K.,Debord,M.,

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint

Bouwmeester,D.,andGwinn,E.(2013)EvidenceforRod-ShapedDNA-StabilizedSilverNanoclusterEmitters.Adv.Mater.25,2797–2803.(25)O’Neill,P.,Gwinn,E.,andFygenson,D.K.(2011)UVExcitationofDNA-StabilizedAg-ClusterFluorescenceviatheDNABases.J.Phys.Chem.C115,24061–24066.(26)Wightman,R.M.,May,L.J.,andMichael,A.C.(1988)DetectionofDopamineDynamicsintheBrain.Anal.Chem.60,769A–793A.(27)Liu,X.,Wang,F.,Niazov-Elkan,A.,Guo,W.,andWillner,I.(2013)ProbingbiocatalytictransformationswithluminescentDNA/silvernanoclusters.NanoLett.13,309–14.(28)Lévay,G.,andBodell,W.J.(1993)DetectionofdopamineDNAadducts:potentialroleinParkinson’sdisease.Carcinogenesis14,1241–1245.(29)Zahid,M.,Saeed,M.,Yang,L.,Beseler,C.,Rogan,E.,andCavalieri,E.L.(2011)Formationofdopaminequinone-DNAadductsandtheirpotentialroleintheetiologyofParkinson’sdisease.IUBMBLife63,1087–93.(30)Cavalieri,E.L.(2002)Catecholortho-quinones:theelectrophiliccompoundsthatformdepurinatingDNAadductsandcouldinitiatecancerandotherdiseases.Carcinogenesis23,1071–1077.(31)Liu,J.,Wang,Z.-H.,Luo,G.-A.,Li,Q.-W.,andSun,H.-W.(2002)TheInteractionofDNAwithDopaminebySpectroscopicandElectrochemicalMethods.Anal.Sci.18,751–755.(32)Chen,T.-W.,Wardill,T.J.,Sun,Y.,Pulver,S.R.,Renninger,S.L.,Baohan,A.,Schreiter,E.R.,Kerr,R.a,Orger,M.B.,Jayaraman,V.,Looger,L.L.,Svoboda,K.,andKim,D.S.(2013)Ultrasensitivefluorescentproteinsforimagingneuronalactivity.Nature499,295–300.(33)Lee,M.H.,Kim,J.S.,andSessler,J.L.(2015)Smallmolecule-basedratiometricfluorescenceprobesforcations,anions,andbiomolecules.Chem.Soc.Rev.44,4185–4191.(34)Doussineau,T.,Schulz,A.,Lapresta-Fernandez,A.,Moro,A.,Körsten,S.,Trupp,S.,andMohr,G.J.(2010)Onthedesignoffluorescentratiometricnanosensors.Chem.-AEur.J.16,10290–10299.(35)Owesson-White,C.A.,Roitman,M.F.,Sombers,L.A.,Belle,A.M.,Keithley,R.B.,Peele,J.L.,Carelli,R.M.,andWightman,R.M.(2012)Sourcescontributingtotheaverageextracellularconcentrationofdopamineinthenucleusaccumbens.J.Neurochem.121,252–262.(36)Daberkow,D.P.,Brown,H.D.,Bunner,K.D.,Kraniotis,S.A.,Doellman,M.A.,Ragozzino,M.E.,Garris,P.A.,andRoitman,M.F.(2013)AmphetamineParadoxicallyAugmentsExocytoticDopamineReleaseandPhasicDopamineSignals.J.Neurosci.33,452–463.(37)Dreyer,J.K.,Herrik,K.F.,Berg,R.W.,andHounsgaard,J.D.(2010)InfluenceofPhasicandTonicDopamineReleaseonReceptorActivation.J.Neurosci.30,14273–14283.(38)Ngo,K.T.,Varner,E.L.,Michael,A.C.,andWeber,S.G.(2017)MonitoringDopamineResponsestoPotassiumIonandNomifensinebyinVivoMicrodialysiswithOnlineLiquidChromatographyatOne-MinuteResolution.ACSChem.Neurosci.8,329–338.(39)Copp,S.M.,Bogdanov,P.,Debord,M.,Singh,A.,andGwinn,E.(2014)BasemotifrecognitionanddesignofDNAtemplatesforfluorescentsilverclustersbymachinelearning.Adv.Mater.26,5839–45.(40)Schultz,D.,Copp,S.M.,Markešević,N.,Gardner,K.,Oemrawsingh,S.S.R.,Bouwmeester,D.,andGwinn,E.(2013)Dual-ColorNanoscaleAssembliesofStructurallyStable,Few-AtomSilverClusters,AsReportedbyFluorescenceResonanceEnergyTransfer.ACSNano7,9798–9807.

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted October 18, 2017. . https://doi.org/10.1101/205591doi: bioRxiv preprint