High-efficiency optogenetic silencing with soma … · Anion-conducting channelrhodopsins could therefore relieve constrains 70 imposed by ion-pumping rhodopsins. The naturally-occurring

1

2

3

High-efficiencyoptogeneticsilencingwithsoma-targetedanion-4

conductingchannelrhodopsins5

6

7

8

MathiasMahn,LihiGibor,KatayunCohen-KashiMalina,PritishPatil,YoavPrintz,ShirOring,RivkaLevy,9

IlanLamplandOferYizhar10

11

DepartmentofNeurobiology,WeizmannInstituteofScience,Rehovot,Israel12

13

14

15

Correspondingauthor:16

OferYizhar,Ph.D.17

DepartmentofNeurobiology18

WeizmannInstituteofScience19

234Herzlst,Rehovot761000120

Israel21

Tel.+972-8-934695722

[email protected] 23

peer-reviewed) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not. http://dx.doi.org/10.1101/225847doi: bioRxiv preprint first posted online Dec. 8, 2017;

http://dx.doi.org/10.1101/225847

Abstract24

Optogenetic silencing allows time-resolved functional interrogation of defined neuronal populations.25

However, the limitationsof inhibitoryoptogenetic tools impose stringent constraintsonexperimental26

paradigms.Thehighlightpowerrequirementoflight-drivenionpumpsandtheireffectsonintracellular27

ion homeostasis pose unique challenges, particularly in experiments that demand inhibition of a28

widespreadneuronalpopulationinvivo.Guillardiathetaanion-conductingchannelrhodopsins(GtACRs)29

are promising in this regard, due to their high single-channel conductance and favorable photon-ion30

stoichiometry.However,GtACRsshowpoormembranetargetinginmammaliancells,andtheactivityof31

suchchannelscancausetransientexcitationintheaxonduetoanexcitatorychloridereversalpotential32

inthiscompartment.Hereweaddressbothproblemsbyenhancingmembranetargetingandsubcellular33

compartmentalization of GtACRs. The resulting GtACR-based optogenetic tools show improved34

photocurrents,greatlyreducedaxonalexcitation,highlightsensitivityandrapidkinetics,allowinghighly35

efficientinhibitionofneuronalactivityinthemammalianbrain.36

37

Introduction38

Perturbationofneuronalactivityisafundamentalaspectofneuroscienceresearch,oftenusedtogain39

insight into the functional rolesofparticularbrain regions, circuits and cell types 1.Optogenetic tools40

havegreatlyenhancedtheprecisionwithwhichsuchmanipulationscanbeperformed2,providingboth41

temporal precision and cell-type specificity to experiments aimed at defining the roles of individual42

neuralcircuitcomponents inneuralcomputationoranimalbehavior.Currentoptogeneticapproaches43

forsilencingofneuronsaremainlybasedonthelight-activatedmicrobialrhodopsinshalorhodopsin3,4,44

archaerhodopsin 5andcruxhalorhodopsin 6.Theseproteinspump ionsacross theneuronalmembrane45

with millisecond kinetics, independently of the electrochemical gradient. These tools allow neuronal46

silencingwithprecisetemporalonsetandoffset5,7,8.However,ion-pumpingrhodopsinspossessseveral47

characteristics that impose substantial constraints on the experimental paradigm and complicate the48

interpretation of experimental outcomes. These limitations become evenmore pronounced in cases49

whereneuronalsilencingisrequiredforextendedperiodsoftime.Theunfavorablestoichiometryofone50

transportedionforeachabsorbedphotonnecessitatescontinuousilluminationathighlightpower.The51

resultingtissueheating9,10andphototoxicity11restricttheopticallyaddressablebrainvolumethatcan52

beefficientlysilenced.Furthermore,ion-pumpingmicrobialrhodopsinsexhibitadeclineinphotocurrent53


http://dx.doi.org/10.1101/225847

amplitudesofupto90%withinaminuteofillumination,leadingtoreducedsilencingefficacyovertime5412,8,13.Becauseoftheirinsensitivitytoelectrochemicalgradients,ion-pumpingmicrobialrhodopsinscan55

shift the concentrationsof intracellular ions tonon-physiological levels. In the caseof halorhodopsin,56

this can lead toaccumulationof chloride in theneuron, inducing changes in the reversal potential of57

GABAergic synapses 14. While in the case of archaerhodopsin this can increase the intracellular pH,58

inducing action potential-independent Ca2+ influx and elevated spontaneous vesicle release 13.59

Furthermore,thehyperpolarizationmediatedbyion-pumpingactivitytogetherwiththefastoffkinetics60

canleadtoanincreasedfiringrateuponterminationoftheillumination6,13. 61

Anion-conducting channelrhodopsins (ACRs), a newly established set of optogenetic tools 15,16,17, are62

distinctfromion-pumpingrhodopsinsintwomajoraspects:first,theycanconductmultipleionsduring63

each photoreaction cycle. This increased photocurrent yield per photon makes channelrhodopsins64

superiorintermsoftheiroperationallight-sensitivity.Second,conductingionsaccordingtothereversal65

potential, ACRs are more likely to avoid non-physiological changes in ion concentration gradients. A66

light-gatedchlorideconductancewillshuntmembranedepolarization,whichcanbeusedtoeffectively67

clamp the neuronal membrane potential to the reversal potential of chloride, given that the ion68

permeabilityissufficientlyhigh.Anion-conductingchannelrhodopsinscouldthereforerelieveconstrains69

imposed by ion-pumping rhodopsins. The naturally-occurring anion-conducting channelrhodopsins70

(nACRs) from the cryptophyte algaGuillardia theta 16 are particularly interesting in this regard. These71

channelrhodopsins,namedGtACR1andGtACR2,havenear-perfectanionselectivityandproducelarge72

photocurrentsinmammaliancells,owingtoahighersingle-channelconductancethanthatoftheknown73

cation-conductingchannelrhodopsins16,18.WhileGtACRswereshowntoinhibitbehaviorinthefruitfly7419,20and larval zebrafish 21, theyhavenotyetbeenapplied tomammaliansystems,most likelydue to75

poor membrane targeting and complex activity in the axonal compartment. To overcome these76

limitationsofGtACRsandtherebyofoptogeneticinhibitioningeneral,wegeneratedseveralmembrane77

targeting-enhanced GtACR variants, converging onto soma-targeted GtACR2 (stGtACR2), a fusion78

construct thatcombinesGtACR2withaC-terminal targetingmotif fromthesoma-localizedpotassium79

channelKv2.122.WedemonstrateherethatstGtACR2showsincreasedmembranetargeting,extremely80

high anion photocurrents and reduced axonal excitation, making it the most effective tool for81

optogeneticinhibitionatthecellsomatodate.82

83


http://dx.doi.org/10.1101/225847

Results84

GtACR2efficientlysilencesneuronalactivityinvitroandinvivo85

To determine the utility of ACRs for silencing of neurons we first expressed the three previously-86

described blue light-activated ACRs, GtACR2 16, iC++ 17, and iChloC 15, in cultured rat hippocampal87

neurons by adeno associated virus (AAV)-mediated gene transfer.Whole-cell patch-clamp recordings88

fromGtACR2-expressingneurons showed reliableoutwardphotocurrents (Fig. 1a) in response to47089

nmfullfieldlightpulses.Thephotocurrentafter1sofcontinuousillumination(stationaryphotocurrent)90

ofGtACR2expressingneuronsclampedto-35mVwassignificantlyhigherthanthatoftheengineered91

ACRs(eACRs)iC++andiChloC(628.5±61.8pA,330.2±37.9pA,and136.3±21.4pA,respectively;Fig.92

1b). Given the poormembrane targeting and intracellular accumulation ofGtACR2 (Fig. 1c), the high93

single-channelconductanceofGtACR216 is likelythecauseforthehighphotocurrentsobservedinthe94

whole-cell recordings.While thenativeGtACR2 seems tooutperform thepreviously-describedeACRs,95

the above findings suggest that improved membrane targeting of GtACR2 would greatly facilitate96

silencing of neuronal activity. Importantly, this should allow efficient silencing by using significantly97

lowerlightpower,enablingoptogeneticcontrolofalargerbrainvolume.98

To verify this prediction, we characterized the efficiency of GtACR2-mediated inhibition in awake,99

behavingmicewithextracellularrecordings(Fig.1d-g).Toquantifytheefficiencyofsilencinginalarge100

cortical volume, we recorded from mice expressing GtACR2 in pyramidal neurons in the medial101

prefrontal cortex (mPFC). Mice were implanted with movable fiberoptic-coupledmicrowire arrays in102

whichelectrodeswereplaced500µmbelowtheopticalfibertip(Fig.1d).Usinganalyticalmodelingof103

lightscatteringandabsorptioninbraintissue23weestimatedthelightpowerdensityatthepositionof104

the extracellular recording site to be 0.11% of the light power density exiting the optical fiber105

(SupplementaryFig.1).Single-unitrecordingsshowedasignificantreductioninneuronalfiringratesin106

responsetoa5s460nm lightpulse (Fig.1e-h).Theserecordingsshowedthatoutof100singleunits107

recorded (n = 2mice, 12 recording sites), 43% and 35% significantly reduced their firing rate at the108

secondhighestandhighesttestedlightpowerdensities(0.5and1mWmm-2,respectively;Fig.1f).Half109

ofthesilencedunitsshowedasignificantfiringratereductionalreadyatthelowesttestedlightpower110

density(125µWmm-2),withsomeunitscompletelysilenced(Fig.1e,h).Atthehighest light intensity,111

silencedunitsshowedpronouncedreboundactivityupon lightpulsetermination (Fig.1g).Notably,all112

lightpowerdensitiesusedwerewellbelowthenecessarylightpowersforinvivooptogeneticinhibition113


http://dx.doi.org/10.1101/225847

usingmicrobialion-pumps8,6,indicatingthatGtACRscanserveaspotentinhibitoryoptogenetictoolsfor114

somaticsilencinginmammalianneurons.115

ActivationofACRsintheaxonalcompartmentinducesantidromicactionpotentials116

We have previously shown that GtACR1 can induce vesicle release from thalamocortical projection117

neuronsuponilluminationoftheiraxonalterminalsintheacutebrainslice13.Malyshevandcolleagues118

24demonstratedasimilareffectinGtACR2-expressingcorticalpyramidalneuronsinacutebrainslices.119

ToverifythatthisexcitatoryactionofGtACRsisnotanartifactoftheacutebrainslicepreparation,we120

evaluatedtheexcitatoryeffectofACRsintwoseparatepreparations:inculturedhippocampalneurons121

and in awake, freely moving mice. Whole-cell patch-clamp recordings in current-clamp mode from122

GtACR2-expressing cultured neurons revealed that during a 100 ms long illumination pulse, GtACR2123

reliably inhibited action potential (AP) generation (Fig. 2a). However, this was often associated with124

what appeared to be an attenuated AP shortly after light onset (‘escaped AP’, Fig. 2a, inset).When125

recordedinvoltage-clampmode,escapedAPsweremeasuredin6outof12testedGtACR2-expressing126

neuronsinresponseto1mslightpulsesatsaturatinglightpower(4.5mWmm-2).Theseescapedspikes127

occurredevenwhentherecordedphotocurrentwasanoutwardcurrent,expectedtohyperpolarizethe128

somaticmembrane(Fig.2b,upperleft).Wethenaskedwhethertheselight-evokedantidromicAPsare129

specific to naturally-occurring chloride-channels, or a general feature of light-evoked chloride130

conductance in the axon. We therefore tested two engineered anion-conducting channelrhodopsins131

underthesameexperimentalconditions.APswereevokedinresponseto1msand1slonglightpulses132

in 2 / 13 and 7 / 12 iC++ expressing neurons, respectively, showing that this effect is not specific to133

GtACRs(Fig.2b,middle).NoAPswereevokediniChloCexpressingcells(n=22andn=15,for1msand134

1s longlightpulses,respectively;Fig.2bright), likelybecauseoftheoverallsmallerphotocurrentswe135

observed incellsexpressingthisconstruct (Fig.1b).To furthertest thehypothesis thatACRactivation136

mightbedepolarizing in someneuronal compartments,weapplied spatially-restricted laserpulses to137

thesomaorneuritesofculturedhippocampalneuronsduringwhole-cellpatch-clamprecordings.Light138

pulsesdirectedat the somausingagalvanometricmirror system (seeOnlineMethods) inducedsmall139

hyperpolarizingordepolarizingphotocurrents,while lightpulsesdirected toneuritesof the samecell140

evokedantidromicspikes(Fig.2c).141

Theserecordingswereperformedinculturedneuronsafterat least14days invitro,astageatwhich142

intracellular chloride concentrations should reach theadult state 25 as theexpressionof theneuronal143

potassiumchlorideco-transporterKCC226isfullyup-regulated27.Nevertheless,chloridehomeostasisof144


http://dx.doi.org/10.1101/225847

neuronalculturesmaydifferduetoreducedconcentrationsofKCC2regulatorssuchasinsulin28inthe145

culturemedium.WethereforetestedwhetheractivationofGtACR2could leadtoaxonalexcitation in146

vivo.Werecorded fromGtACR2-expressingmice (Fig.1d)usingmovable fiberoptic-coupledmicrowire147

drivesatthesiteofAAVinjection(Fig.1d-h).Inthesameanimals,weimplantedasecondopticalfiber148

terminatingatthenucleusaccumbens(NAc),aprominentprojectiontargetofthemPFC29 (Fig.2d). In149

responsetobrieflightpulsestothemPFC(5mspulsewidth,460nmat1mWmm-2,whichcorresponds150

to28.8mWatthefibertip),APswereevoked(Fig.2e,g)inthesameAAV-expressingregionthatshowed151

significantsilencingduring5s lightpulses (Fig.1d-h).Moreover, similar5ms lightpulsesdeliveredto152

the NAc led to the induction of short-latency APs in themPFC (Fig. 2f-g), presumably due to axonal153

excitationandantidromicpropagation.154

Insummary,lightstimulationofGtACR2-expressingaxonsledtoAPsinhippocampalneuronsinvitro,in155

thalamocorticalprojectionneuronsinacutebrainslices13andinstriatum-projectingcorticalneuronsin156

awake, behaving mice. We therefore concluded that chloride-mediated axonal depolarization is a157

generalphenomenonthatcouldconfoundtheanalysisofoptogeneticsilencingexperiments.158

OverexpressionofKCC2reducesGtACR2-mediatedantidromicactionpotentials 159

Based on these findings, we reasoned that if GtACR-mediated antidromic spiking is indeed due to a160

positivelyshiftedchloridereversalpotential in theaxon,decreasing theaxonalchlorideconcentration161

should reduce the probability of antidromic spike generation. The chloride extruder KCC2 26 is up-162

regulated inneuronsduringdevelopment, leading tohighendogenousKCC2protein levels in somatic163

and dendritic membranes 30, but does not localize to axons 31,32,33. We first tested whether164

overexpression of KCC2 leads to its localization to the axonal compartment in cultured hippocampal165

neurons.We co-transfected neurons with expression vectors encoding the green fluorescent protein166

mNeonGreen (GFP) and KCC2, or with GFP alone as control. We then labeled the neurons with167

antibodies against the dendrite-specific microtubule-associated protein-2 (MAP2) and KCC2.168

Overexpression of KCC2 led to strong KCC2 immunoreactivity in dendrites and somata of transfected169

neurons(Fig.3a),comparedtoendogenousexpressionlevels(Fig.3a,whitearrow).Toquantifyaxonal170

KCC2 levels,axonsweredetectedasneurites thatareGFP-positiveandMAP2-negative (Fig.3a, zoom171

in).WhilemeanaxonalKCC2intensitywasnotsignificantlydifferentbetweenyoung(7daysinvitro)and172

mature(16daysinvitro)hippocampalcultures,KCC2overexpressionledtoa6.6±1.1foldhigheraxonal173

KCC2signal(Fig.3b,ctrlyoungvs.ctrlmature:P=0.20;ctrlyoungvs.KCC2:P=5*10-7;ctrlmaturevs.174

KCC2:P=9.8*10-3).KCC2expressiondidnotsignificantlyshiftthechloridereversalpotentialmeasured175


http://dx.doi.org/10.1101/225847

inthesoma(Fig.3c)ortheactionpotentialinitiationthreshold(rheobase,Fig.3d),butindeedledtoa176

significantreductionofGtACR2-evokedantidromicspiking(1mslightpulseswidth,470nmat4.5mW177

mm-2;Fig.3e).Thisresultprovidesfurthersupportforthenotionthatthechloridereversalpotentialin178

theaxonisdepolarizingunderphysiologicalconditions.179

180

Soma-targetingofGtACR2increasessomaticphotocurrentsandreducesaxonalexcitation181

ToovercomethetwomaincaveatsofGtACRsinrespecttotheirutilityasoptogeneticinhibitorytools,182

namelytheirpoormembranetargetingandtriggeringofantidromicspikesintheaxonalcompartment,183

we designed several new GtACR2 variants with altered membrane targeting sequences (Fig. 4a).184

Addition of the ER export and trafficking signals from the mammalian inward rectifying potassium-185

channelKir2.1waspreviouslyshowntoreduceintracellularaggregationofthechloride-pumpNpHR34,35.186

Indeed,fusionofthesesequencestoGtACR2(eGtACR2)ledtoreducedintracellularaccumulationinvivo187

(Fig. 4c, 2nd column). To reduce antidromic spike generation, GtACR2 should be removed from the188

axonalcompartment.Toachievesoma-specificlocalizationofGtACR2,wereplacedtheERexportsignal189

with the soma-targetingmotif of the soma and proximal dendrite localized voltage-gated potassium-190

channel Kv2.1 36,37, which was previously shown to enhance soma-localized expression of191

channelrhodopsin-2 38,22.Hypothesizingthatdestabilizing theproteinbyaddingaproteindegradation-192

promotingproline(P),glutamicacid(E),serine(S),andthreonine(T)richsequence(PEST39)couldlimit193

the effective lifetime of membrane-resident channels diffusing along the axon, thereby potentially194

further restricting GtACR2 protein levels outside the somatic compartment. To test these new viral195

constructs, we infused AAVs encoding them unilaterally into the mPFC of mice, leading to strong196

fluorescence at the injection site aswell as sparse labeling along the injection needle track (Fig. 4b).197

Soma-targetedGtACR2(stGtACR2;Fig.4c,3rdcolumn)aswellasthedestabilizedstGtACR2-PEST(Fig.4c198

,4thcolumn)showedimprovedmembranetargeting,strongsoma-associatedfluorescenceandreduced199

neuritefluorescence(Fig.4d).Functionalcharacterizationofthesoma-targetedconstructsbywhole-cell200

patch-clamprecordings intheacutebrainsliceshoweda2.6-fold increase instationaryphotocurrents201

comparedtountargetedGtACR2,leadingtoaveragephotocurrentsofmorethan2nAwhencellswere202

clamped to -35mV (Fig. 4e). Improvedmembrane targeting alone strongly increased the antidromic203

spikegenerationprobability(Fig.5a;eGtACR2),whilesoma-targetingnotonlyincreasedphotocurrents204

(Fig. 4e) but also decreased the probability of inducing antidromic spikes in cultured hippocampal205

neurons(Fig.5a).DestabilizingstGtACR2usingthePESTsequenceledtoalesspronouncedreductionin206


http://dx.doi.org/10.1101/225847

antidromicspikegenerationcomparedtostGtACR2(Fig.5a).Photocurrentswerequantifiedinthesame207

neuronstoverifythatthereducedprobabilityofantidromicspikegenerationisnotduetodifferencesin208

peakphotocurrentsofthedifferentconstructsinculturedneurons(Fig.5b).Incontrasttothestationary209

photocurrentsinacutebrainsliceexperiments(Fig.4e)peakphotocurrentsinculturedneuronsdidnot210

differ significantly between constructs, pointing to a lowermembrane targeting efficiency in cultured211

neuronsoraninfluenceoftheshortervirusincubationtime.Nevertheless,itfollowsthatthedramatic212

reductioninantidromicspikingforstGtACR2isnotduetolowerphotocurrents.213

We next asked whether stGtACR2 and stGtACR2-PEST would show improved performance through214

reducedlight-evokedsynapticreleasefromlong-rangeprojectingaxons.WeinjectedAAVsencodingthe215

GtACR2variantstogetherwithasecondAAVencodingacell-fillingfluorophoreunilaterallytothemPFC,216

to allow for visualization of the axons of cortico-cortical projection neurons in the contralateral217

hemisphere (Fig. 5c). During acute brain slice preparation from thesemice, the corpus callosumwas218

severed, separating the somata of the transduced cortico-cortical projecting neurons from their axon219

terminals in the contralateralmPFC. Conductingwhole-cell patch-clamp recordings frompostsynaptic220

neuronsinareaswithfluorescently-labeledaxonscontralateraltotheinjectionsitethereforeallowedus221

tocharacterizetheisolatedeffectofGtACR2activationontheaxonalcompartment.Bluelightpulsesled222

to reliably evoked EPSCs in slices expressing thenon-targetedGtACR2 (Fig. 5d). In contrast, the EPSC223

amplitudewasdramaticallyreducedinslicesexpressingthesoma-restrictedstGtACR2(Fig.5d,e;473.4±224

153.4pAvs.34.1±9.5pAforGtACR2andstGtACR2,respectively).Theincreasedsomaticphotocurrents225

of stGtACR2, together with the near-elimination of antidromic spiking and neurotransmitter release,226

makeitahighlyefficienttoolforoptogeneticinhibition.227

stGtACR2mediatedBLAinhibitionpreventsfearextinctionlearning228

ToverifytheutilityofstGTACR2-mediatedoptogeneticinhibitioninawake,behavinganimals,wechose229

touse this tool for suppressingbasolateral-amygdala (BLA) activityduring extinctionof auditory-cued230

fearconditioning,awell-establishedformofassociative learning40.TheBLAplaysacentralrole inthe231

acquisition aswell as extinction of the conditioned freezing response. Based on previouswork 41,we232

hypothesizedthattemporally-preciseinhibitionoftheBLAduringthedeliveryofconditionedstimuliin233

extinction training would suppress the formation of extinctionmemory.We bilaterally injectedmice234

withAAVencodingstGtACR2orafluorophore-onlycontrolvectorintotheBLAandimplanted200µm-235

diameteroptical fibers above the injection sites (Fig. 6a; Supplementary Fig. 2). Following3weeksof236

recovery,miceunderwentfearconditioningincontextA(Fig.6b).Bothgroups(stGtACR2,n=8;control,237


http://dx.doi.org/10.1101/225847

n=8)showedincreasedfreezingduringacquisition(ctrl:from3.8±1.9%to39.5±8.1%;stGtACR2:from238

3.5±1.5% to29.6±3.5%, ScheirerRayHare testH=52.91,P=3.5*10-10)withno significantdifference239

between groups (ctrl vs. stGtACR2: Scheirer Ray Hare test H = 0.11, P = 0.74), suggesting that BLA240

activity isnotalteredmerelybyexpressionofstGtACR2(Fig.6c).Totestforfearrecallandextinction,241

miceunderwentextinctiontrainingtwodays later inadifferentcontext fromthat inwhichtheywere242

fearconditioned(contextB).Theextinctionprotocolconsistedof twenty30s tonepresentationsthat243

werepairedwithblue lightdelivery(447nm;5mWfromeachfibertip).Micewerethentested inan244

extinctionretrievaltestthefollowingdayinwhichtheyweresubjectedtotwentyCSpresentations,but245

no light was delivered (Fig. 6b, right). During this test, stGtACR2 mice showed higher freezing rates246

duringCSpresentation (ctrl vs. stGtACR2: ScheirerRayHare testH = 4.30, P = 3.8*10-2), but freezing247

levels were indistinguishable from control mice during the inter-tone intervals (ctrl vs. stGtACR2:248

ScheirerRayHaretestH=3.6*10-2,P=0.85),indicatingthatfearextinctionwaspreventedbystGtACR2249

mediatedBLA inhibitionduringCS presentation. Sierra-Mercado and colleagues 41, previously showed250

that inhibition of the BLA by muscimol injection prior to fear extinction interfered with extinction251

learning. Our results extend these findings, demonstrating that temporally-precise inhibition of BLA252

activityonlyduringCSpresentationusingstGtACR2caninterferewithextinctionlearning.Insummary,253

ourexperiments indicate that stGtACR2 isapowerful inhibitoryoptogenetic tool, allowing temporally254

precisesilencingofneuronalpopulationsinvivo.255

Discussion256

We took amembrane targeting approach to allow the utilization of the high-conductanceGuillardia257

thetaanion-conductingchannelrhodopsins16asanoptogenetictoolinmammalianneurons.Whilethese258

naturally-occurringchannelrhodopsinsshowedgreatpromiseduetotheirhighlyefficientphotocurrents259

and light sensitivity 16, andhaveproveneffective in silencingdrosophila and zebrafish neurons 19,20,21,260

they have seen little use in mammalian neuroscience applications. This was mainly due to poor261

membranetargetingandtocomplexeffectsonaxonalexcitability13,24.Ourfindingsindicatedthateven262

in its non-targeted form, GtACR2 can efficiently silence neurons in the medial prefrontal cortex of263

behaving mice. These results were consistent with the high photocurrent amplitudes recorded in264

neuronsexpressingGtACR2,comparedwithcellsexpressingtheengineeredACRs iC++andiChloC17,15.265

Given the high single channel conductance, favorable photon-ion stoichiometry, and high light266

sensitivity, the light power density for neuronal inhibition with GtACR2 is at least one order of267

magnitude lowerthanthatofothersilencingopsins8,42.However,despite itsapparenthighefficacy,a268


http://dx.doi.org/10.1101/225847

significant portion of the protein seemed to reside in intracellular compartments, where it cannot269

contribute to functional photocurrents. Furthermore, activationofGtACR2 inour recordingswas also270

associatedwithantidromicspikingatlightonsetwhenilluminatingboththeproximalanddistalaxons.271

WehavepreviouslyobservedGtACR2-mediatedtriggeringofsynapticrelease inthalamocorticalaxons27213, consistentwith recent reports of antidromic spiking in layer 2/3 pyramidal neurons 24 in the slice273

preparation. In this study,weobservedGtACR2-mediatedantidromic spiking in culturedhippocampal274

neurons,incortico-corticalneuronsintheacutesliceandincortico-striatalaxonsofbehavingmice.Our275

findingsindicatethataxonalexcitationbyachlorideconductanceisageneralphenomenon,andcould276

reflectadepolarizedreversalpotentialforchlorideintheaxonalcompartment.Whilesucheffectshave277

been previously reported, for example in hippocampalmossy fibers 43,44, cerebellar 45 and brain-stem278

axons 46,47, systematic evaluation of the phenomenon has been previously restricted to axons that279

naturallyexpressGABA-Areceptors.280

To determine whether elevated chloride concentration in the axon could indeed lead to GtACR2-281

mediated axonal excitation,we co-expressed the KCC2 transporterwithGtACR2 in cultured neurons.282

The endogenous KCC2 transporter, which is expressed in mature neurons and is known to be283

responsible for extruding chloride from the somatodendritic compartment 48, is known to be absent284

from theaxon 31,32,33, potentially permitting ahigher chloride concentration in this compartment.Our285

findingthatoverexpressionofKCC2resultedinasignificantdecreaseinlight-inducedantidromicspiking286

indicatesthatACR-mediatedantidromicspikingcouldindeedbetheresultofasmallerchloridegradient287

in the axon, even in adult neurons. While this antidromic spiking phenotype would probably not288

interferewithlong-terminhibitionexperiments(minutesandupward),itmightbeaconfoundingfactor289

whentemporally-precise (millisecond-scale) inhibition is required.FutureworkcouldcombineGtACR2290

stimulation with red-shifted chloride indicators to directly examine changes in chloride levels in the291

axonalcompartmentduringACR-mediatedchlorideconductance.292

Most importantly,ourstudydemonstrates that thesoma-targetedvariantsofGtACR2show improved293

membraneexpressioninthesomatodendriticcompartment,andoffersuperioranionphotocurrentsfor294

high-efficiency optogenetic silencing of neurons in the mammalian brain. Current optogenetic295

experimentsoften involveasparsely labeledpopulationofneuronsthataredistributedacrossa large296

brain tissue volume 49,50. Efficient silencing of such widely-distributed neuronal populations require297

continuous activity of the inhibitory optogenetic tool 51, placing considerable constraints related to298

tissueheatingandphotodamage52,9.OurdataindicatethatstGtACR2canprovideaneffectivemeansof299


http://dx.doi.org/10.1101/225847

performingsuchchallengingexperiments,dueto its intrinsicallyhighconductance,which increases its300

effective light sensitivity in expressing neurons. With increasing distance from the fiber tip, the301

wavelength dependent transmittance becomes increasingly relevant. For instance, multiplying the302

actionspectraofGtACR1andGtACR216with theanalyticallymodeled light transmittancecurve 23 for303

braintissue(SupplementaryFig.1)revealedthatexcitationofGtACR2with480nmorofGtACR1with304

510nmwouldprovideoptimal light-mediatedsilencingat500µmdistancefromtheopticfibertip. In305

experiments that require optogenetic manipulation of functionally- but not anatomically-segregated306

neuronalpopulations,stGtACR2mightbecombinedwithred-shiftedtoolssuchasC1V153,Chrimson54307

orReaChR55.Red-shiftedcalciumsensors56,57couldalsobeusedincombinationwithstGtACR2dueto308

itsminimalresponsivityatwavelengthsabove560nm.Notably,bothstGtACR1andstGtACR2arealso309

highly advantageous for multiphoton single-cell silencing experiments 58,59 owing to their somatic310

restriction22andhigh-amplitudephotocurrents.311

In summary,we have demonstrated thatmembrane targeting and somatodendritic restriction of the312

naturally-occurring anion-conducting GtACR2 address two independent constraints of this313

channelrhodopsin, greatly improving photocurrents and minimizing axonally-generated antidromic314

action potentials. We were able to achieve high-efficiency neuronal silencing with the optimized315

stGtACR2 and demonstrated its efficacy for temporally-precise inhibition of amygdala activity during316

extinctionlearning.317

Acknowledgments318

We thank M. Segal (Weizmann Institute of Science) for reagents. We thank R. Zwang for help with319cloning.We thank theYizhar laboratorymembers forcommentson themanuscript.Weacknowledge320support(toO.Y.)fromtheHumanFrontierScienceProgram,aEuropeanResearchCouncilstartinggrant321(ERC-2013-StG OptoNeuromod 337637), the Adelis Foundation, the Lord Sieff of BrimptonMemorial322Fund and the Candice Appleton Family Trust. O.Y. is supported by the Gertrude and Philip Nollman323CareerDevelopmentChair.324

AuthorContributions325

M.M.andO.Y.designedthestudywithinputfromI.L.;M.M.designedtheconstructs,performedinvitro326electrophysiology and imaging experiments. K.CKM. performed in vitro neuronal recordings. L.G.327performedandanalyzed invivo electrophysiology recordings.P.P.performedandanalyzedbehavioral328experimentsundertheguidanceofY.P.andM.M.;S.O.performedhistologyandimagingonbehavioral329mice. R.L. prepared neuronal cultures and viral vectors.M.M. andO.Y. analyzed and interpreted the330resultsandwrotethemanuscript.331


http://dx.doi.org/10.1101/225847

References332

1.Wurtz,R.H.,Usingperturbationstoidentifythebraincircuitsunderlyingactivevision.Phil.Trans.R.Soc.B370,20140205(2015).

2.Deisseroth,K.,Optogenetics.Naturemethods8,26-29(2011).

3.Han,X.&Boyden,E.S.,Multiple-coloropticalactivation,silencing,anddesynchronizationofneuralactivity,withsingle-spiketemporalresolution.PloSone2,e299(2007).

4.Zhang,F.etal.,Multimodalfastopticalinterrogationofneuralcircuitry.Nature446,633-639(2007).

5.Chow,B.Y.etal.,High-performancegeneticallytargetableopticalneuralsilencingbylight-drivenprotonpumps.Nature463,98-102(2010).

6.Chuong,A.S.etal.,Noninvasiveopticalinhibitionwithared-shiftedmicrobialrhodopsin.Natureneuroscience17,1123-1129(2014).

7.Inoue,K.etal.,Alight-drivensodiumionpumpinmarinebacteria.Naturecommunications4,1678(2013).

8.Mattis,J.etal.,Principlesforapplyingoptogenetictoolsderivedfromdirectcomparativeanalysisofmicrobialopsins.Naturemethods9(2),159-172(2011).

9.Stujenske,J.M.,Spellman,T.&Gordon,J.A.,ModelingtheSpatiotemporalDynamicsofLightandHeatPropagationforInVivoOptogenetics.Cellreports12(3),525-534(2015).

10.Arias-Gil,G.,Ohl,F.W.,Takagaki,K.&Lippert,M.T.,Measurement,modeling,andpredictionoftemperatureriseduetooptogeneticbrainstimulation.Neurophotonics3(4),045007(2016).

11.Cardin,J.A.etal.,Targetedoptogeneticstimulationandrecordingofneuronsinvivousingcell-type-specificexpressionofChannelrhodopsin-2.Natureprotocols5,247-254(2010).

12.Groma,G.I.&Dancshazy,Z.,HowManyMFormsarethereintheBacteriorhodopsinPhotocycle?Biophysicaljournal50(2),357-366(1986).

13.Mahn,M.,Prigge,M.,Ron,S.,Levy,R.&Yizhar,O.,Biophysicalconstraintsofoptogeneticinhibitionatpresynapticterminals.Natureneuroscience19(4),554-556(2016).

14.Raimondo,J.V.,Kay,L.,Ellender,T.J.&Akerman,C.J.,Optogeneticsilencingstrategiesdifferintheireffectsoninhibitorysynaptictransmission.Natureneuroscience15,1102-1104(2012).

15.Wietek,J.etal.,Animprovedchloride-conductingchannelrhodopsinforlight-inducedinhibitionofneuronalactivityinvivo.Scientificreports5,14807(2015).

16.Govorunova,E.G.,Sineshchekov,O.A.,Janz,R.,Liu,X.&Spudich,J.L.,Naturallight-gatedanionchannels:Afamilyofmicrobialrhodopsinsforadvancedoptogenetics.Science349,647-650(2015).


http://dx.doi.org/10.1101/225847

17.Berndt,A.etal.,Structuralfoundationsofoptogenetics:Determinantsofchannelrhodopsinionselectivity.ProceedingsoftheNationalAcademyofSciences113,822-829(2016).

18.Sineshchekov,O.A.,Govorunova,E.G.,Li,H.&Spudich,J.L.,Gatingmechanismsofanaturalanionchannelrhodopsin.ProceedingsoftheNationalAcademyofSciences112,14236-14241(2015).

19.Mohammad,F.etal.,Optogeneticinhibitionofbehaviorwithanionchannelrhodopsins.Naturemethods14(3),271-274(2017).

20.Mauss,A.S.,Busch,C.&Borst,A.,OptogeneticNeuronalSilencinginDrosophiladuringVisualProcessing.ScientificReports7,13823(2017).

21.Mohamed,G.A.etal.,Opticalinhibitionoflarvalzebrafishbehaviourwithanionchannelrhodopsins.BMCbiology15,103(2017).

22.Baker,C.A.,Elyada,Y.M.,Parra,A.&Bolton,M.M.,Cellularresolutioncircuitmappingwithtemporal-focusedexcitationofsoma-targetedchannelrhodopsin.eLife5(2016).

23.Yona,G.,Meitav,N.,Kahn,I.&Shoham,S.,Realisticnumericalandanalyticalmodelingoflightscatteringinbraintissueforoptogeneticapplications.eneuro3,ENEURO--0059(2016).

24.Malyshev,A.Y.etal.,ChlorideconductinglightactivatedchannelGtACR2canproducebothcessationoffiringandgenerationofactionpotentialsincorticalneuronsinresponsetolight.Neuroscienceletters640,76-80(2017).

25.Ganguly,K.,Schinder,A.F.,Wong,S.T.&Poo,M.-m.,GABAitselfpromotesthedevelopmentalswitchofneuronalGABAergicresponsesfromexcitationtoinhibition.Cell105,521-532(2001).

26.Payne,J.A.,Stevenson,T.J.&Donaldson,L.F.,MolecularcharacterizationofaputativeK-ClcotransporterinratbrainAneuronal-specificisoform.JournalofBiologicalChemistry271,16245-16252(1996).

27.Khirug,S.etal.,DistinctpropertiesoffunctionalKCC2expressioninimmaturemousehippocampalneuronsincultureandinacuteslices.EuropeanJournalofNeuroscience21,899-904(2005).

28.Kelsch,W.etal.,Insulin-likegrowthfactor1andacytosolictyrosinekinaseactivatechlorideoutwardtransportduringmaturationofhippocampalneurons.TheJournalofneuroscience:theofficialjournaloftheSocietyforNeuroscience21(21),8339-8347(2001).

29.Gabbott,P.L.A.,Warner,T.A.,Jays,P.R.L.,Salway,P.&Busby,S.J.,Prefrontalcortexintherat:Projectionstosubcorticalautonomic,motor,andlimbiccenters.TheJournalofComparativeNeurology492,145-177(2005).

30.Zhu,L.,Lovinger,D.&Delpire,E.,CorticalneuronslackingKCC2expressionshowimpairedregulationofintracellularchloride.Journalofneurophysiology93(3),1557-1568(2005).

31.Hübner,C.A.etal.,DisruptionofKCC2revealsanessentialroleofK-Clcotransportalreadyinearly


http://dx.doi.org/10.1101/225847

synapticinhibition.Neuron30(2),515-524(2001).

32.Szabadics,J.etal.,ExcitatoryeffectofGABAergicaxo-axoniccellsincorticalmicrocircuits.Science(NewYork,N.Y.)311(5758),233-235(2006).

33.Báldi,R.,Varga,C.&Tamás,G.,DifferentialdistributionofKCC2alongtheaxo-somato-dendriticaxisofhippocampalprincipalcells.TheEuropeanjournalofneuroscience32(8),1319-1325(2010).

34.Gradinaru,V.,Thompson,K.R.&Deisseroth,K.,eNpHR:aNatronomonashalorhodopsinenhancedforoptogeneticapplications.Braincellbiology36(1-4),129-139(2008).

35.Gradinaru,V.etal.,Molecularandcellularapproachesfordiversifyingandextendingoptogenetics.Cell141(1),154-165(2010).

36.Trimmer,J.S.,ImmunologicalidentificationandcharacterizationofadelayedrectifierK+channelpolypeptideinratbrain.ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica88(23),10764-10768(1991).

37.Lim,S.T.,Antonucci,D.E.,Scannevin,R.H.&Trimmer,J.S.,AnoveltargetingsignalforproximalclusteringoftheKv2.1K+channelinhippocampalneurons.Neuron25(2),385-397(2000).

38.Wu,C.,Ivanova,E.,Zhang,Y.&Pan,Z.-H.,rAAV-mediatedsubcellulartargetingofoptogenetictoolsinretinalganglioncellsinvivo.PloSone8(6),e66332(2013).

39.Rogers,S.,Wells,R.&Rechsteiner,M.,Aminoacidsequencescommontorapidlydegradedproteins:thePESThypothesis.Science(NewYork,N.Y.)234(4774),364-368(1986).

40.Tovote,P.,Fadok,J.P.&Lüthi,A.,Neuronalcircuitsforfearandanxiety.Naturereviews.Neuroscience16(6),317-331(2015).

41.Sierra-Mercado,D.,Padilla-Coreano,N.&Quirk,G.J.,Dissociablerolesofprelimbicandinfralimbiccortices,ventralhippocampus,andbasolateralamygdalaintheexpressionandextinctionofconditionedfear.Neuropsychopharmacology:officialpublicationoftheAmericanCollegeofNeuropsychopharmacology36(2),529-538(2011).

42.Wietek,J.etal.,Anion-conductingchannelrhodopsinswithtunedspectraandmodifiedkineticsengineeredforoptogeneticmanipulationofbehavior.ScientificReports7,14957(2017).

43.Alle,H.&Geiger,J.R.P.,GABAergicspill-overtransmissionontohippocampalmossyfiberboutons.JournalofNeuroscience27,942-950(2007).

44.Ruiz,A.,Campanac,E.,Scott,R.S.,Rusakov,D.A.&Kullmann,D.M.,PresynapticGABAAreceptorsenhancetransmissionandLTPinductionathippocampalmossyfibersynapses.Natureneuroscience13,431-438(2010).

45.Pugh,J.R.&Jahr,C.E.,AxonalGABAAreceptorsincreasecerebellargranulecellexcitabilityandsynapticactivity.TheJournalofneuroscience:theofficialjournaloftheSocietyforNeuroscience31(2),


http://dx.doi.org/10.1101/225847

565-574(2011).

46.Turecek,R.&Trussell,L.O.,Presynapticglycinereceptorsenhancetransmitterreleaseatamammaliancentralsynapse.Nature411(6837),587-590(2001).

47.Price,G.D.&Trussell,L.O.,Estimateofthechlorideconcentrationinacentralglutamatergicterminal:agramicidinperforated-patchstudyonthecalyxofHeld.JournalofNeuroscience26,11432-11436(2006).

48.Kaila,K.,Price,T.J.,Payne,J.A.,Puskarjov,M.&Voipio,J.,Cation-chloridecotransportersinneuronaldevelopment,plasticityanddisease.NatureReviewsNeuroscience15,637-654(2014).

49.Liu,X.etal.,Optogeneticstimulationofahippocampalengramactivatesfearmemoryrecall.Nature(2012).

50.Kim,C.K.,Adhikari,A.&Deisseroth,K.,Integrationofoptogeneticswithcomplementarymethodologiesinsystemsneuroscience.NatureReviewsNeuroscience18,222-235(2017).

51.Wiegert,J.S.,Mahn,M.,Prigge,M.,Printz,Y.&Yizhar,O.,Silencingneurons:tools,applications,andexperimentalconstraints.Neuron95,504-529(2017).

52.Lee,J.H.etal.,GlobalandlocalfMRIsignalsdrivenbyneuronsdefinedoptogeneticallybytypeandwiring.Nature465(7299),788-792(2010).

53.Yizhar,O.etal.,Neocorticalexcitation/inhibitionbalanceininformationprocessingandsocialdysfunction.Nature477(7363),171-178(2011).

54.Klapoetke,N.C.etal.,Independentopticalexcitationofdistinctneuralpopulations.Naturemethods11(3),338-346(2014).

55.Lin,J.Y.,Knutsen,P.M.,Muller,A.,Kleinfeld,D.&Tsien,R.Y.,ReaChR:ared-shiftedvariantofchannelrhodopsinenablesdeeptranscranialoptogeneticexcitation.Natureneuroscience16(10),1499-1508(2013).

56.Dana,H.etal.,Sensitiveredproteincalciumindicatorsforimagingneuralactivity.eLife5(2016).

57.Inoue,M.etal.,Rationaldesignofahigh-affinity,fast,redcalciumindicatorR-CaMP2.Naturemethods12(1),64-70(2015).

58.Prakash,R.etal.,Two-photonoptogenetictoolboxforfastinhibition,excitationandbistablemodulation.NatureMethods9,1171-1179(2012).

59.Packer,A.M.etal.,Two-photonoptogeneticsofdendriticspinesandneuralcircuits.NatureMethods9,1202-1205(2012).

60.Liu,Y.etal.,OptogenSIM:a3DMonteCarlosimulationplatformforlightdeliverydesigninoptogenetics.Biomedicalopticsexpress6(12),4859-4870(2015).


http://dx.doi.org/10.1101/225847

61.Chugh,B.P.etal.,Measurementofcerebralbloodvolumeinmousebrainregionsusingmicro-computedtomography.NeuroImage47(4),1312-1318(2009).

62.Grimm,D.,Kay,M.A.&Kleinschmidt,J.A.,Helpervirus-free,opticallycontrollable,andtwo-plasmid-basedproductionofadeno-associatedvirusvectorsofserotypes1to6.Moleculartherapy:thejournaloftheAmericanSocietyofGeneTherapy7(6),839-850(2003).

63.Graham,F.L.&Eb,A.J.,Anewtechniquefortheassayofinfectivityofhumanadenovirus5DNA.Virology52(2),456-467(1973).

64.Anikeeva,P.etal.,Optetrode:amultichannelreadoutforoptogeneticcontrolinfreelymovingmice.Natureneuroscience15(1),163-170(2011).

65.Grünwald,D.,Shenoy,S.M.,Burke,S.&Singer,R.H.,Calibratingexcitationlightfluxesforquantitativelightmicroscopyincellbiology.Natureprotocols3(11),1809-1814(2008).

66.Pugh,J.R.&Jahr,C.E.,ActivationofaxonalreceptorsbyGABAspilloverincreasessomaticfiring.TheJournalofneuroscience:theofficialjournaloftheSocietyforNeuroscience33(43),16924-16929(2013).

333

334


http://dx.doi.org/10.1101/225847

Figures 335

336

Figure1.GtACR2efficientlysilencesneuronalactivityinvivo337

(a-c) Characterization of ACRs in cultured rat hippocampal neurons. (a) Sample whole-cell voltage-clamp338photocurrentrecordingofaGtACR2-expressingcell illuminated(470nm)with increasinglightpowerdensity.(b)339Comparisonofstationaryphotocurrentsofbluelight-sensitiveACRs(Vhold=-35mV,currentafter1sofcontinuous340illumination). Neurons expressing GtACR2 (n = 12) showed the highest photocurrents compared with neurons341expressing iC++ (n=12)and iChloC (n=15). F(2,36)=36.92,P=1.9x10-9 (c)Representative imageofGtACR2342localization.Green:GtACR2,red:cytoplasmicRFP,blue:nucleus.(d-h) InvivoquantificationofGtACR2mediated343neuronal silencing efficiency. (d) Schematic of experimental paradigm. Extracellular recordings from the mPFC344wereperformedwithamovableoptrodefollowinginjectionofAAV2/1encodingGtACR2intothemPFC.(e)Two345representativerasterplotsofunitssignificantlyreducingtheirfiringrateduring5sofilluminationwithbluelight346(460nm).Eachlightpowerwastested10times.Whiletheunitdepictedontheleftshowsagradedresponseto347increasing light powers, the unit depicted on the right is completely inhibited even at the lowest tested light348power.(f)Numberofunitsthatsignificantlyreducedtheirfiringratecomparedto5spre-lightperiod,dependent349on the tested lightpowers. (g)Normalized firing rate (FR/pre-light-FR,100msbins)ofall significantly silenced350units.(h)Quantificationofg.Inbandh,resultsarepresentedasmeans(±SEM).351


http://dx.doi.org/10.1101/225847

352

Figure2.ActivationofGtACR2intheaxonalcompartmentinducesactionpotentialsinvitroandin353vivo354

(a-b)Characterizationoflight-evokedspikinginACR-expressingculturedhippocampalneurons(470nmat4.5mW355mm-2). (a) Representative whole-cell current-clamp recording of a GtACR2-expressing cell silenced by light356application. Inset: strongly attenuated spike occurring shortly after light onset. (b) Representative whole-cell357voltage-clamprecordingofescapedactionpotentialsinresponseto1mslightpulses.Piechartsdepictthenumber358of neuronswith induced spikes for the three tested light-gated chloride channels:GtACR2, iC++ and iChloC. (c)359Illuminationofdistalneurites inducesspiking inculturedneurons.Schematicdepicting theoutlineofaGtACR2-360expressing neuron overlaid with the locations of laser illumination spots. Shown are whole-cell voltage-clamp361responses to spatially-restricted illumination at the indicated locations. (d-g) In vivo extracellular recording362followingGtACR2expressioninthemPFC.(d)SchematicoftheimplantationallowingforilluminationoftheNAc,a363downstreamtargetofthemPFC,whilerecordinginthemPFC.(e)SingleunitsrecordedinthemPFC,showingrapid364light-evokedresponsesduringa20mstime-windowstartingwitha5mslightpulse(1mWmm-2

,correspondingto36528.8mWatthefibertip).Unitsarearrangedfromtoptobottomaccordingtotheirmeanfirstspikelatencyacross36620trials.Piechartsdepictthenumberofneuronswithsignificantlyincreasedspikerates.(f)mPFCunitsshowing367significantlyincreasedfiringratesinresponsetoilluminationoftheNAc.Unitsaresortedbymeanspikelatency.368Lightpoweratthefibertip:28.8mW.Piechartasin(e).(g)Percentofunitswithincreasedfiringrateinresponse369to5mslightpulsesofincreasinglightpower.370

371


http://dx.doi.org/10.1101/225847

372

Figure3.OverexpressionofKCC2reducesantidromicspikinginculturedhippocampalneurons373

(a)ImagesfromanKCC2overexpressionexperiment.EndogenousKCC2isexpressedinthesomaticcompartment,374while overexpression of KCC2 led to increased expression in axonal projections. Cultured hippocampal neurons375weresparselytransfectedeitherwithGFPalone(ctrl)orGFPandKCC2.Neuronswerethenfixedandstainedfor376MAP2andKCC2.Topimagesshowarepresentativeregionofinterestwithoneoverexpressingcell inthecenter.377The arrow indicates a neuronal cell body expressing endogenous KCC2 levels at 16 days in vitro (DIV). Bottom378images depict MAP2-expressing dendrites and a single MAP2-negative axon (arrow), which is positive for379overexpressedKCC2basedon itsanti-KCC2fluorescence. (b)QuantificationofaxonalKCC2 immunofluorescence380for immature (DIV7) andmature control neurons (DIV16) and neurons overexpressing KCC2, normalized to the381averageaxonalKCC2signal in immatureneurons(DIV7).AxonalKCC2fluorescence issignificantlyhigher inKCC2382overexpressingcultures.(Kruskal–WallisHtest,H(2,44)=29.26,P<10-4;ctrl:nDIV7=11,nDIV16=10,KCC2:n=23)383(c-e) Physiological properties and light-evoked spiking in cultured hippocampal neurons expressing either only384GtACR2,orco-expressingKCC2.(c)EffectofKCC2overexpressiononthe IV-curve.Thereversalpotentialdidnot385differ significantly (Students t-test = 1.5, GtACR2: n = 14, GtACR2+KCC2: n = 18, P = 0.15 two-tailed). (d)386Comparison of the minimal current injection to induce an action potential (rheobase). KCC2 overexpressing387neuronsdidnotdifferfromGtACR2onlyexpressingneurons(Studentst-test=0.5,GtACR2:n=21,GtACR2+KCC2:388n=22,P=0.7two-tailed).(e)KCC2overexpressionsignificantlyreducedthelikelihoodofGtACR2mediatedaction389potential generation. (Mann–Whitney U = 146.5, nGtACR2 = 21, nGtACR2+KCC2 = 22, P = 4 * 10

-2). All results are390presentedasmeans(±SEM). 391


http://dx.doi.org/10.1101/225847

392

Figure4.TargetingGtACR2totheneuronalsomaleadstoenhancedphotocurrentamplitude393

(a) Schematic of different targeting approaches. (b) Image showing the fluorescence resulting from AAV2/1394mediatedcytosolicfluorophoreexpressioninthemPFC.Transductionismostdenseattheinjectionsite(indicated395bythelowerdashedbox)andsparsealongtheinjectionneedletrack(upperdashedbox).(c)Highermagnification396imagesof theareas indicated inb. c-I:Zoom inon the injectionsite. c-II:Zoom inon themoredorsal regionof397sparseexpression.c-III:Highermagnificationofc-II.stGTACR2andstGtACR2-PESTshowenrichmentatthesoma.398(d) Quantification of soma restriction by normalizing mPFC layer 1 fluorescence by the mean fluorescence399measuredattheinjectioncenter.Targetingreducesrelativelayer1fluorescence(n=2pergroup).(e)Lightpower400density dependence of stationary photocurrent inwhole-cell patch-clamp recordings of neurons in acute brain401slices. Inset:Representativewhole-cellvoltage-clamprecording.Thestationaryphotocurrentwasdefinedas the402photocurrentattheendofa1slightpulse.Thefitwasperformedpercellwiththeeffectivelightpowerdensity403for50%photocurrent(EPD50)asfreeparameter.stGtACR2andstGtACR2-PESThaveasignificantlyhighermaximal404stationaryphotocurrent thanGtACR2 (F(2,23)=11.84,P=2.9x10-4;GtACR2:n=8,stGtACR2:n=10,stGtACR2-405PEST:n=8).Resultsarepresentedasmeans(±SEM).406


http://dx.doi.org/10.1101/225847

407

Figure5.TargetingGtACR2tothesomatodendriticcompartmentattenuatesaxonalexcitation408

(a)Comparisonoftheincidenceofantidromicspikestriggeredby1mslightpulsesinvirallytransducedcultured409hippocampal neurons. eGtACR2 has a significantly increased AP incidence, while stGtACR2 decreases the410occurrenceofAPs.(Kruskal–WallisHtest,H(3,97)=47,P<10-4;nGtACR2=20,neGtACR2=22,nstGtACR2=22,n411stGtACR2-PEST = 33) (b) Peak photocurrents did not differ significantly between the constructs, showing that412reducedAPincidenceinstGtACR2transducedneuronsdoesnotstemfromsmallerphotocurrents.(F(3,82)=0.83,413P = 0.48). (c) Schematic of the experimental setup to characterize GtACR2 triggered axonal neurotransmitter414release inacutebrainslices.Virusencodingacytosolic fluorophorewasco-injectedwiththeGtACR2variants to415allow for visualization of the axon terminals of transduced cortico-cortical projection neurons. Contralateral416neurons in areas with high fluorescence intensity were recorded. (d) Representative traces of excitatory post-417synapticcurrentsinresponseto1mslightpulses(470nm,at4.5mWmm-2).(e)Quantificationofthelightevoked418post-synaptic current amplitude. Soma targeting led to significant reduction in light evoked EPSCs amplitudes419(F(2,22)=5.54,P=1.13*10-2).Allresultsarepresentedasmeans(±SEM).420

421


http://dx.doi.org/10.1101/225847

422

423Figure6.Silencingofcue-associatedBLAactivityusingstGtACR2suppressesextinctionofcued424

freezing.425

(a)MicewerebilaterallyinjectedwitheYFPorstGTACR2-encodingvirusandimplantedwithopticfiberstargeting426theBLA.Bottom:representativeimageofstGtACR2expressionintheBLA.(b)stGtACR2andcontrol(eYFP)mice427weresubjectedtoauditoryfearconditioning(day1,conditioning),extinctiontraining(day3,earlyandlate428extinction)andextinctionrecall(day4,earlyandlaterecall).Duringauditoryfearconditioningmicewere429submittedtofivetone(CS)-shock(US)presentationsincontextA.Onday3twenty30stone(CS)presentations430werepairedwithlight(447nmat5mWexitingthefibertip)incontextB.Onday4extinctionrecallwastestedby431twenty30stonepresentationsincontextB.(c)PercentageoffreezingduringpresentationoftheCS(toprow)and432the30spriortoCS(bottomrow).Therightcolumndepictsthemeanpercentageoffreezingduringtheearlyand433latephasesofthetrialsforthetwogroups.MeanfreezinglevelsforctrlandstGtACR2mice,duringtheearlyrecall434phase,were28±4%and54±6%,respectively;thedistributionsinthetwogroupsdifferedsignificantly(Scheirer435RayHaretestH=4.30,P=3.8*10-2).Allresultsarepresentedasmeans(±SEM).436

437


http://dx.doi.org/10.1101/225847

438

439

440

441

442

443

444

445

446

447

448

449

450

Supplementaryfigure1.FunctionalactionspectraofGtACR1andGtACR2basedonmultiplicationwith451analyticallightspreadestimation452

Lightpropagationfromaflat-cleavedopticalfiberwithinbraintissuewasestimatedbyanalyticalmodeling23,using453graymatterparametersasestimatedbyLiuetal.60:Wavelengthdependentbrainscatteringcoefficiecient:µs(λ)=454(2.37*(λ/500nm)-1.15)/(1-g). Absorption is estimated as: µa(λ) = B*S*µa(HbO2(λ)) + B*(1-S)*µa(Hb(λ)) +455W*µa(H2O(λ)). Blood oxygen saturation: S (62%). Estimated percentage of water in brain tissue: W (65%).456Percentage of blood in the brain tissue (B). Cerebral cortex (PFC) blood volume excluding major vessels was457estimated as 4.6 % according to Chugh et al. (2009) 61. Wavelength dependent blood and water absorption458coefficients fromomlc.orgwereused. (a)Contour linesofestimated lightpowerdensities resulting from4mW459lightcoupledtoanopticalfiber.Attherecordingsite(500µmbelowand100µmlateraltotheopticalfibercenter)460thelightpowerdensitydropsto~0.11%ofthelightpowerdensityatthefibersurface.(b)ComparisonofGtACR1461andGtACR2activationefficiencywithinbrain tissueby including thewavelengthdependent transmittance.Top:462relative quantum efficiency as reported in Govorunova et al. (2015) 16. Middle: Wavelength dependent463transmittanceat electrode recording site,modeledas in a. Bottom:Relative activationefficiencynormalizedby464maximalactivationofGtACR1.Thehigherscatteringandabsorptionat470nmshiftthemostefficientexcitation465wavelengthto480nmforGtACR2.Thehigherbloodabsorptioncoefficientat520nmresultsin510nmbeingthe466mostefficientwavelengthforGtACR1activationatthisdistancefromtheopticfiber.Accordingtothisestimation467GtACR1usedat510nmallowsfor14%lowerlightpowerscomparedtoGtACR2excitedat480nm,makingGtACR1468themore efficient tool when only a single wavelength is needed. However, GtACR1 will cause non-permissive469activationintheloweraswellasthehigherwavelengthranges,thereforeonlyGtACR2allowsforthecombination470withothercurrentlyavailableoptogeneticactuatorsorreporters.471

472


http://dx.doi.org/10.1101/225847

473

Supplementaryfigure2.SummaryofopticalfiberplacementofBLAsilencing474

In fixed brain sections obtained from fear extinction learning experiment animals, fiber placement was475determined, and stGtACR2 expression in the BLA close to the fiber was verified. Horizontal lines mark the476approximatefiberfaceposition.477

478


http://dx.doi.org/10.1101/225847

ExperimentalProcedures479

ProductionofrecombinantAAVvectors480

HEK293cellswereseededat25%-35%confluence.Thecellsweretransfected24hlaterwithplasmids481

encodingAAVrep,capandavectorplasmidfortherAAVcassetteexpressingtherelevantDNAusingthe482

PEImethod62.Cellsandmediumwereharvested72haftertransfection,pelletedbycentrifugation(300483

g), resuspended in lysis solution ([mM]: 150NaCl, 50 Tris-HCl; pH8.5withNaOH) and lysedby three484

freeze-thawcycles. The crude lysatewas treatedwith250Ubenzonase (Sigma)per 1mlof lysate at485

37°Cfor1.5htodegradegenomicandunpackagedAAVDNAbeforecentrifugationat3000gfor15min486

to pellet cell debris. The virus particles in the supernatant (crude virus)were purified using heparin-487

agarose columns, eluted with soluble heparin, washed with phosphate buffered saline (PBS) and488

concentratedbyAmiconcolumns.Viralsuspensionwasaliquotedandstoredat–80°C.Viraltiterswere489

measuredusing real-timePCR.AAVvectorsused for intracranial injectionshadgenomic titers ranging490

between8.6*1010and2*1011genomecopiespermilliliter(gc/ml).Wheredirectlycomparedvirustiters491

were matched by dilution to the lowest concentration. AAV vectors used for neuronal culture492

transductionwereadded4daysaftercellseeding.Thetiterwasmatchedtofinalmediumconcentration493

of1.1*108gc/ml.AlloftheAAVexpressionconstructsdescribedinthisstudywillbeavailablefreelyon494

Addgenetofacilitatetheutilizationofthesenewtoolsbytheneurosciencecommunity.495

Thefollowingviruseswereusedinthisstudy:496

AAV2/1.hSyn1.GtACR2-eGFP.WPRE,AAV2/1.CamKIIα.GtACR2-ts-Fred-Kv2.1.WPRE,497

AAV2/1.CamKIIα.GtACR2-ts-Fred-ER.WPRE,AAV2/1.CamKIIα.GtACR2-ts-Fred-Kv2.1-PEST.WPRE,498

AAV2/1.CamKIIα.TagRFP-T.WPRE,AAV2/1.CamKIIα.eYFP.WPRE,AAV2/1.CamKIIα.iC++-eYFP.WPRE,499

AAV2/1.hSyn.iChlOC-eGFP.WPRE.500

Primaryhippocampalneuronculture501

Primary cultured hippocampal neuronswere prepared frommale and female P0 Sprague-Dawley rat502

pups(Envigo).CA1andCA3wereisolated,digestedwith0.4mgml−1papain(Worthington),andplated503

ontoglass coverslipspre-coatedwith1:30Matrigel (Corning).Culturedneuronsweremaintained ina504

5% CO2 humidified incubator with Neurobasal-A medium (Invitrogen) containing 1.25% fetal bovine505

serum(FBS,BiologicalIndustries),4%B-27supplement(Gibco),2mMGlutamax(Gibco)andplatedon506

coverslips in a24-well plateat adensityof 65,000 cells perwell. To inhibit glial overgrowth,200µM507

fluorodeoxyuridine(FUDR,Sigma)wasaddedafter4daysofinvitroculture(DIV).508


http://dx.doi.org/10.1101/225847

Calciumphosphatetransfectionofculturedneurons509

Neurons were transfected using the calcium phosphate method 63. Briefly, the medium of primary510

hippocampal neurons cultured in a 24well platewas collected and replacedwith 400 µl serum-free511

MEMmedium(ThermoFisherscientific).30µltransfectionmix(2µgplasmidDNAand250µMCaCl2in512

HBSatpH7.05)wereaddedperwell.After1hincubationthecellswerewashed2timeswithMEMand513

themediumwaschangedbacktothecollectedoriginalmedium.Culturedneuronswereusedbetween514

14–17DIVforexperiments.515

Thefollowingplasmidswereusedinthisstudy:516

pAAV_hSyn1_GtACR2-eGFP_WPRE(basedonAddgene85463),pAAV_CamKIIα_mNeonGreen_WPRE,517

pAAV_CamKIIα(0.4kb)_mScarlet_WPRE,pCITF_KCC2-tdTomato(Addgene61404).518

Animals519

All experimental procedures were approved by the Institutional Animal Care and Use Committee520

(IACUC)attheWeizmannInstituteofScience.Six-week-oldC57BL/6mice(P35–45)wereobtainedfrom521

Envigo.Upto5maleorfemaleC57BL/6micewerehousedinacageinalight-dark(12h-12h)cyclewith522

foodandwateradlibitum.Micewerehousedfor6-12weeksfollowingsurgerytoallowforrecoveryand523

virusexpression.524

Stereotacticinjectionofviralvectors525

Six-week-oldC57BL/6mice(P35–45)wereinitiallyinducedwithketamine(80mgkg−1)andxylazine(10526

mg kg−1) and placed into a stereotaxic frame (David Kopf Instruments), before isoflurane anesthesia527

(~1% in O2, v/v). A craniotomy (∼1 mm in diameter) was made above the injection site. Virus528

suspensionswere slowly injected (100 nlmin–1) using a 34G beveled needle (Nanofil syringe,World529

Precision Instruments). After injection, the needlewas left in place for an additional 5min and then530

slowly withdrawn. The surgical procedure was either continued with optic fiber or optrode drive531

implantations (describedbelow),or the surgical incisionwasclosedwith tissueglueand0.05mgkg−1532

Buprenorphinewassubcutaneously injectedforpost-surgicalanalgesia. Injectionstargetingthemedial533

prefrontalcortex(mPFC)weremade1.8mmanterior,0.3mmlateraland2.53mmventraltobregma.534

Basolateralamygdala(BLA) injectioncoordinateswere1.15mmposterior,3.0mmlateraland5.0mm535

ventral to bregma. For mPFC injections, 1 µl of the indicated virus was injected. For fear extinction536


http://dx.doi.org/10.1101/225847

experiments mice were bilaterally injected with 500 nl AAV2/1.CamKIIα.stGtACR2-Fred.WPRE or537

AAV2/1.CamKIIα.eYFP.WPREwithagenomictiterintherangeof2-3x1011vpml-1.538

OpticfiberandOptrodedriveimplantation539

Forfiberopticimplantation,acraniotomy(∼1 mmindiameter)wasmadeabovetheimplantationsite540

and a ferrule-terminated optical fiber (ThorLabs) was placed at the desired coordinates using a541

stereotaxicframe(DavidKopfInstruments).ForbilateralBLAtargeting,thefibertipwasplaced1.15mm542

posterior, 3.0 mm lateral and 4.8 mm ventral to bregma. For nucleus accumbens, the fiber was543

implantedata45°anglewiththeferrulepointingposteriortoallowforoptrodedriveplacementabove544

themPFCinthesameanimals.Thefibertipwasaimedtoterminate1.42mmanterior,1mmlateraland545

5mmventraltobregma.TheopticalfiberwassecuredtotheskullusingMetabond(Parkell)anddental546

acrylic. In mice trained for fear extinction learning additional dental acrylic was applied in a second547

session under isoflurane anesthesia (~1% in O2, v/v) after fear learning (day 2). For optrode drive548

implantation,themovabledrivewasloweredtoaninitialrecordingpositionabovethePL(AP:1.8mm,549

ML:0.3mm,DV:–2.3mm).Priortothepermanentattachmentoftheoptrodetotheskull,theoptrode550

guidewasprotectedwithKwik-Kastsiliconeelastomer(WorldPrecisionInstruments)andsecuredusing551

dentalacrylic.Micewereallowedtorecoverforat least6weeksbeforeexperiments.Thelocationsof552

implantedopticalfibersandoptrodeswerevalidatedhistologicallyforallexperimentalmice.553

Acutebrainslicepreparation554

Mice were injected intraperitoneally with pentobarbital (130 mg kg−1, i.p.) and perfused with555

carbogenated (95% O2, 5% CO2) ice-cold slicing solution ([mM] 2.5 KCl, 11 glucose, 234 sucrose, 26556

NaHCO3, 1.25 NaH2PO4, 10MgSO4, 2 CaCl2; 340mOsm). After decapitation, 300 µm coronalmPFC557

sliceswereprepared in carbogenated ice-cold slicing solutionusinga vibratome (LeicaVT1200S)and558

allowed to recover for 20 min at 33°C in carbogenated high-osmolarity artificial cerebrospinal fluid559

(high-OsmACSF; [mM]3.2KCl, 11.8 glucose, 132NaCl, 27.9NaHCO3, 1.34NaH2PO4, 1.07MgCl2, 2.14560

CaCl2;320mOsm)followedby40minincubationat33°CincarbogenatedACSF([mM]3KCl,11glucose,561

123NaCl,26NaHCO3,1.25NaH2PO4,1MgCl2,2CaCl2;300mOsm).Subsequently,sliceswerekeptatRT562

incarbogenatedACSFuntiluse.TherecordingchamberwasperfusedwithcarbogenatedACSFatarate563

of2mlmin–1andmaintainedat32°C.564

Electrophysiologicalmethodsforcellcultureandacutebrainslicerecordings565


http://dx.doi.org/10.1101/225847

Whole-cellpatchclamprecordingswereperformedundervisualcontrolusingobliqueilluminationona566

two-photon laser scanningmicroscope (Ultima IV, Bruker) equipped with a 12 bit monochrome CCD567

camera (QImagingQIClick-R-F-M-12). Borosilicate glass pipettes (Sutter Instrument BF100-58-10)with568

resistancesrangingfrom3–7MΩwerepulledusingalasermicropipettepuller(SutterInstrumentModel569

P-2000).Forhippocampalneuroncultures,electrophysiologicalrecordingsfromneuronswereobtained570

inTyrode’smedium ([mM]150NaCl,4KCl,2MgCl2,2CaCl2,10D-glucose,10HEPES;320mOsm;pH571

adjustedto7.35withNaOH),AcOHTyrode’smedium([mM]125NaCl,25AcOH,4KCl,2MgCl2,2CaCl2,572

10 D-glucose, 10 HEPES; 320 mOsm; pH adjusted to 7.35 with NaOH). The recording chamber was573

perfused at 0.5 ml min–1 and maintained at 29°C. Pipettes were filled using standard intracellular574

solution([mM]135K-gluconate,4KCl,2NaCl,10HEPES,4EGTA,4MgATP,0.3NaGTP;280mOsmkg–1;575

pHadjustedto7.3withKOH)oranintracellularsolutionallowingforEPSCandIPSCrecording([mM]120576

Cs-gluconate,11CsCl,1MgCl2,1CaCl2,10HEPES,11EGTA,5QX-314;280mOsmkg–1;pHadjustedto577

7.3 with CsOH). Whole-cell voltage clamp recordings were performed using a MultiClamp 700B578

amplifier,filteredat8kHzanddigitizedat20kHzusingaDigidata1440Adigitizer(MolecularDevices).579

Invivoopticalsilencingandelectrophysiology580

Allelectrophysiologicalrecordingsinawake,freelymovingmicewereperformedusinganoptrodedrive581

consisting of an electrode bundle of 16 microwires (25 μm diameter straightened tungsten wires;582

Wiretronic Inc.) attached to an 18 pin electrical connector, concentrically arranged around an optical583

fiber inamechanicallyadjustabledrive (Anikeevaetal., 2011 64). Extracellularwaveformsignalswere584

collectedusingtheDigitalLynxintegratedhardwareandsoftwaresystem(NeuralynxInc.).Theelectrical585

signalwasfiltered(600–6,000Hz),amplifiedusingaHS-18-CNR-LEDunity-gainhead-stageamplifierand586

digitized at 32 kHz. The electrode–fiber assembly was lowered using themechanical drive to a new587

recording site at the end of each recording session, leaving at least 1.5 h before the next session to588

ensure stable recordings. Optical stimulation was applied through a ferrule-terminated optical fiber589

(ThorLabs)attachedtothepatch-chordbyazirconiasleeve(ThorLabs).ForopticalsilencingofmPFC,we590

used a blue diode laser (λ = 460 nm,OmicronNanoTechnology). Light transmission for each optrode591

drivewasmeasuredwithacalibratedpowermeter(ThorLabs)atthetipoftheopticalfiberattheendof592

theexperiment.Lightpowerwasmeasureddailybeforeexperimentsatthetipoftheopticalpatchcord.593

Neural data were sorted manually using Off-Line Spike Sorter 3.2.4 (OFSS, Plexon) and analyzed in594

Matlab(MathWorks).595

Invivooptogeneticsilencinginmiceduringextinctiontraining596


http://dx.doi.org/10.1101/225847

Mice in both the stGtACR2 and control group (eYFP expressing)were placed in the fear conditioning597

chamber(MedAssociates)incontextA.MicewerepresentedwithfivepairingsoftheCS(50mslong5598

kHz84.4dBtones,deliveredat10Hzfor30s)andUS(continuous0.5mAfootshockfor1s).EachCS599

coterminatedwiththeUS,witha60sintervalbetweenCS–USpairings.Onday3,micewereconnected600

totheopticalpatchchordandthenplacedinadifferentchamber(contextB).ContextBdifferedfrom601

contextAinthefollowingaspects:odor(A:1%Aceticacidvs.B:70%EtOH),lighting(A:IRvs.B:IR+602

white light), box size (A: small, B: large), floor texture (A: grid, B: plain),wall texture (A:metal vs. B:603

Plexiglas),andbackgroundnoise(A:nonevs.B:fan).Micewereallowed10minofhabituationandthen604

presentedwith20repetitionsoftheCS,separatedby60sintervals.TheCSwaspairedwith5mWblue605

light(447nm)administeredbilaterallyfromthefibertipinbothgroups.Totestextinctionlearning,mice606

were placed in context B on day 4 and presented with 20 repetitions of the CS, separated by 60 s607

intervals.Moviesrecordedat25framespersecondwereautomaticallyscoredforfreezingonday1and608

4 by EthoVision XT 11.5 (Noldus) and by a custom written OpenCV-Python script. The number of609

changedpixelscomparedtothelastframewasquantifiedandfilteredbyaGaussianfilterwith3frames610

standarddeviation.Whenmicewereconnectedtoopticalpatchcords,onlychangedpixelsaroundthe611

mousebodywereconsidered,todiscardpatchcordmotion.Amousewasconsideredtobefreezingif612

38consecutivevalues(1.5s)werebelow983pixels(0.5%ofallpixels,EthoVision)or100pixels(within613

theROIaroundthemouse,OpenCV-Pythonscript).614

Immunofluorescenceandmicroscopy615

Hippocampalneuronalcultureswerefixedfor15minwith4%paraformaldehydeinPBS.Coverslipswere616

washedthreetimesinPBS,incubatedinblockingsolutionfor45min(10%normaldonkeyserum(NDS)617

with0.1%Triton inPBS)and thenexposedovernightat4°C tomonoclonalmouseanti-KCC2primary618

antibody (diluted1:1500 in5%NDS,PBS; catalog#167594S1-12;USBiological)and rabbitanti-MAP2619

(diluted1:1000in5%NDS,PBS;catalog#4542S;CellSignalingTechnology).Following3washesinPBS,620

coverslipswere incubated for2hat room temperature (RT)witha Cy5DonkeyAnti-Rabbit IgG (H+L)621

(diluted1:500in5%NDS,PBS;catalog#711-175-152;JacksonImmunoResearch)andCy3DonkeyAnti-622

Mouse IgG (H+L) (diluted1:1000 in 5%NDS, PBS; catalog # 715-165-151; Jackson ImmunoResearch).623

Coverslipswerethenwashed2timeswithPBS,dippedbrieflyintodouble-distilledwaterandembedded624

inDABCOmountingmedium (Sigma). Immunostainedneuronswere imagedwith a confocal scanning625

microscope(LSM700,CarlZeiss)usinga20xobjectiveforoverviewimages(NA0.8;CarlZeiss)anda63626

xoil immersionobjective(NA1.40;CarlZeiss) forquantification.Miceweredeeplyanesthetizedusing627


http://dx.doi.org/10.1101/225847

pentobarbital (0.4mg g−1 bodyweight) and perfused transcardially with ice-cold phosphate buffered628

saline (PBS, pH 7.4) followed by a solution of 4% paraformaldehyde (PFA) in PBS. After overnight629

postfixation at 4 °C, brainswere removed from the skull and incubated overnight in 4% PFA in PBS.630

Brainswerestoredinto30%sucroseinPBSforatleast24horuntilsectioning.Coronalsections(30μm631

or50µm)werecutonamicrotome(LeicaMicrosystems)andcollectedincryoprotectantsolution(25%632

glycerol, 30% ethylene glycol in PBS pH 6.7). Free-floating sections weremounted on gelatin-coated633

slides,dehydratedandembeddedinDABCOmountingmedium(Sigma). Imageswereacquiredusinga634

virtual slide scanner V (Olympus). Acquisition settingswere kept constantwithin each experiment to635

allowforcomparisonbetweenmice.636

Invitroilluminationanddrugapplication637

Whole-fieldillumination invitrowasperformedusinga470nmlightemittingdiode(29nmbandwidth638

LED; M470L2-C2; Thorlabs) delivered through the microscope illumination path including a custom639

dichroicinordertoreflectthe470nmactivationwavelength.Lightpowerdensitieswerecalculatedby640

measuring the light transmitted through the objective using a power meter (Thorlabs PM100A with641

S146C sensor) and dividing by the illumination area, calculated from the microscope objective field642

number andmagnification 65. D-AP5 (25 µM; ab120003; Abcam) and CNQX (10 µM; C-141, Alomone)643

werebathappliedduringallcultureexperiments.Forspatially-restrictedilluminationofneuronalsoma644

orneurites,a473nmdiodelaser(Bruker)wasdirectedtotheimagingplanewithgalvanometricmirrors,645

yieldingadiffraction-limitedspotof light thatprovidedbrief lightpulses (1ms)ateach location,with646

500msinter-pulseintervalsbetweennonadjacentlocations.647

Dataanalysisandstatisticalmethods648

During whole-cell patch-clamp recordings, pClamp 10 software (Molecular Devices) was used for649

acquisition. Data was analyzed using custom scripts written in Matlab (Mathworks). To quantify650

postsynapticcurrentamplitudesinresponsetolightpulses,holdingcurrenttraceswerefilteredwitha651

Savitzky-Golay11point,secondorder,Welchwindowfunctionfilterandthemaximalchangeinholding652

currentwithin20ms (EPSCs) after lightdeliverywasdetermined. Fiji (basedon ImageJ2;USNational653

InstitutesofHealth)wasusedforimmunofluorescenceimageanalysis.IntheKCC2immunofluorescence654

experimentallnumbers(n)refertothenumberofimagedaxons,inthetargetinghistologynrefersto655

thenumberofmice,andinelectrophysiologicalrecordings,nreferstothenumberofrecordedneurons656

/ units. To detect significantlymodulated units during in vivo silencing experiments, a paired-sample657


http://dx.doi.org/10.1101/225847

student’s t-testwasperformed comparing thenumberofdetectedactionpotentials between the5 s658

pre-lightperiodand the5 s lightperiodduring10 trialsper lightpower.Todetectantidromic spiking659

units, a paired-sample student’s t-test was performed comparing the number of detected action660

potentialsbetweenthe20mspre-lightperiodand the20ms lightperiodstartingwith the5ms light661

pulse.Allvaluesareindicatedasmean±SEM.Significancewasdeterminedatasignificancelevelof0.05662

withTukey'shonestlysignificantdifference(HSD)posthoctestusedtocorrectformultiplecomparisons.663

Incaseofnon-normaldatadistributionnon-parametrictestswereused:Mann–WhitneyUtestwasused664

forasinglecomparisons,theKruskal-WallisHtestforone-wayanalysisofvariance,andtheScheirerRay665

Haretestfortwo-wayanalysisofvariance.Nostatisticaltestswereruntopredeterminesamplesize,but666

sample sizeswere similar to those commonly used in the field. Blinding and randomizationwere not667

performed;howeverautomatedanalysiswasusedwheneverpossible.668


http://dx.doi.org/10.1101/225847

Documents

High-efficiency optogenetic silencing with soma … · Anion-conducting channelrhodopsins could therefore relieve constrains 70 imposed by ion-pumping rhodopsins. The naturally-occurring