Kavanagh, Deirdre M. and Smyth, Annya M. and Martin, Kirsty J. and Dun, Alison and Brown, Euan

15 Pages·2017·2.72 MB·English

ARTICLE
Received12Aug 2014|Accepted 6 Nov2014|Published 17 Dec 2014 DOI:10.1038/ncomms6774 OPEN
A molecular toggle after exocytosis sequesters
the presynaptic syntaxin1a molecules involved
in prior vesicle fusion
Deirdre M. Kavanagh1,2, Annya M. Smyth1,2,3,w, Kirsty J. Martin1,2,w, Alison Dun1,2, Euan R. Brown1,2,
Sarah Gordon3, Karen J. Smillie3, Luke H. Chamberlain4, Rhodri S. Wilson1,2, Lei Yang1,2,w, Weiping Lu1,2,
Michael A. Cousin3, Colin Rickman1,2 & Rory R. Duncan1,2
Neuronalsynapsesareamongthemostscrutinizedofcellularsystems,servingasamodelfor
all membrane trafficking studies. Despite this, synaptic biology has proven difficult to inter-
rogate directly in situ due to the small size and dynamic nature of central synapses and the
molecules within them. Here we determine the spatial and temporal interaction status of
presynaptic proteins, imaging large cohorts of single molecules inside active synapses.
Measuring rapid interaction dynamics during synaptic depolarization identified the small
number of syntaxin1a and munc18-1 protein molecules required to support synaptic vesicle
exocytosis. After vesicle fusion and subsequent SNARE complex disassembly, a prompt
switch in syntaxin1a and munc18-1-binding mode, regulated by charge alteration on the
syntaxin1a N-terminal, sequesters monomeric syntaxin1a from other disassembled fusion
complex components, preventing ectopic SNARE complex formation, readying the synapse
for subsequent rounds of neurotransmission.
1InstituteofBiologicalChemistry,BiophysicsandBioengineering,HeriotWattUniversity,EdinburghEH144AS,UK.2EdinburghSuper-ResolutionImaging
Consortium,www.esric.org.3CentreforIntegrativePhysiology,UniversityofEdinburgh,GeorgeSquare,EdinburghEH89XD,UK.4StrathclydeInstituteof
PharmacyandBiomedicalSciences,161CathedralStreet,GlasgowG40RE,UK.wPresentaddresses:ResearchGovernance&QAOffice,Universityof
Edinburgh,TheQueen’sMedicalResearchInstitute,47LittleFranceCrescent,EdinburghEH164TJ,UK(A.M.S.);BeatsonInstituteforCancerResearch,
SwitchbackRoad,Bearsden,GlasgowG611BD,UK(K.J.M.);OmniVisionTechnologies,Co.,Ltd,4275BurtonDrive,SantaClara,California95054,USA(L.Y.).
CorrespondenceandrequestsformaterialsshouldbeaddressedtoR.R.D.(email:[email protected]).
NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications 1
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
A
ll forms of membrane fusion rely on a core family of targets) accumulate at nerve terminals (Fig. 1c and Supple-
SNARE proteins1. The synchronized action of a number mentary Fig. 1).
ofaccessoryproteinsisalsorequiredtooverseethehighly Next we used PALM, comparing the distribution of hetero-
orderedandlocalizednatureofSNAREmediatedexocytosis(for logous syntaxin1a and munc18-1 with the spatial pattern of
review,seeref.2).Sec1/Munc18proteins(SMproteins)areaclass endogenous molecules ascertained using dSTORM. To achieve
ofsuchaccessoryfactorsthatarepresentatallSNARE-catalyzed themaximumresolution,wefirstexaminedchemicallyfixedcells
membrane fusion sites3. It is known that munc18-01 and co-expressing either PA-mCherry-Munc18-1/enhanced green
syntaxin1 (the principal SM protein and syntaxin involved in fluorescent protein (EGFP)-syntaxin1a (EGFP provided diffrac-
synaptic exocytosis) interact via at least two distinct modes; one tion-limited resolution data) or conversely, PA-mCherry-
with monomeric ‘closed’ syntaxin1a and the other involving its syntaxin1a/EGFP-Munc18-1. Positional information describing
highly conserved amino-terminal (N-terminal) peptide motif4–6. PA-mCherry-munc18-1 and PA-mCherry-syntaxin1a molecule
Current hypotheses incorporate these data into models where localization was rendered into maps (Fig. 1d), where munc18-1
munc18-01andsyntaxin1ainteractusingdistinctbindingmodes and syntaxin1a were seen to co-cluster with one another in
depending on intracellular location and function3,7. However, varicosities (Fig. 1d). To determine whether these areas
whether munc18-01 interacts with syntaxin1a and functions in represented nerve terminals, neurons were transfected with PA-
the late stages of synaptic vesicle fusion7–9 or whether it com- mCherry-syntaxin1a, fixed and co-immunolabelled against
pletely dissociates from syntaxin1a, or syntaxin1a-containing synapsin as before. PALM imaging confirmed that these
complexes, during exocytosis is still unspecified. Therefore, varicosities represented synapses; single syntaxin1a (and so also
despite a large amount of biochemical, electrophysiological and munc18-1) molecules clustered at synapsin-positive synapses
ultra-structuraldata,thespatiotemporalarrangementofmunc18- (Supplementary Fig. 1).By observingand measuring directlythe
01andsyntaxin1aatamolecularlevelinlivingneuronalcellsand molecular distributions of these endogenous and heterologous
particularly in central synapses remains undefined. Questions proteins in neurons for the first time, we reveal a molecular
surrounding the molecular interaction starting point of the distribution, with sparse, individual molecules in processes but
synaptic vesicle cycle have proven difficult to probe, principally denser accumulations in varicosities, suggesting that syntaxin1a
because of a dearth of single-molecule resolution approaches. andmunc18-1aretraffickedalongaxonsbeforeaccumulationin
Here we employed imaging and spectroscopic approaches to presynaptic areas22,23.
quantify the distributions, movements and interactions of To test this hypothesis directly, we quantified the mobility of
munc18-01 and syntaxin1a molecules in central synapses to individual munc18-1 molecules (as opposed to earlier bulk
identify directly the small number of molecules specifically studies) in living neurons using state-of-the-art single-particle
involved with synaptic vesicle exocytosis and to identify the tracking approaches combined with PALM (sptPALM12,13,24).
interactionpathwayinsynapsesbefore,duringandaftersynaptic Large cohorts of single munc18-1 molecules (6,584 single
vesicle exocytosis (Fig. 1a). molecules from n¼3 independent experiments) were tracked,
revealing kinetically and spatially distinct molecular populations
(Fig. 2a–d). Consistent with the notion that the molecules are
Results trafficked along processes before accumulation in synapses,
Munc18-1 and syntaxin1a single-molecule dynamics. We pre- munc18-1 molecules exhibited a restricted motion in puncta,
viously developed fluorescent syntaxin1a and munc18-1 probes whereas a separate population of munc18-1 molecules displayed
(Fig. 1b) that we showed target appropriately in neuroendocrine directed movement, travelling with long displacements between
cells5,10–14,aswellasbeingfunctional5,10,14(alsoshownbyothers synapses (Fig. 2a–d).
using similar constructs)15–17, engaging in protein–protein
interactions in a predictable way5,10,16,17. Importantly, the
single-molecule distributions, localizations and interactions of Munc18-1 trafficking depends on binding to syntaxin1a. To
these essential presynaptic proteins have never been elucidated determinewhethertheinteractionwithsyntaxin1aaffectedthese
in situ—that is, in central neurons. We believe that quantifying different behaviours,weintroduced adominant-negativemutant
protein–proteininteractionsatthemolecularlevel,asopposedto of the t-SNARE protein (Syx D6, L165A,E166A-mCerulean5) that,
the more common colocalization or bulk dynamic studies, is in neuroendocrine cells, disrupts both munc18-1-membrane
essentialtoallowprogresstobemadeinourunderstandingofnot localization (as this is dependent on syntaxin1a interaction25,26)
justsynapticbiologybutofthegamutofcellbiologicalquestions. and exocytosis10,14. This mutant, in contrast to the syntaxin1a
Toprobeandcomparethemolecularorganizationofendogenous ‘open’ mutant (L165A, E166A (ref. 27) that has an unaltered
andheterologousmunc18-1andsyntaxin1aonthenanoscale,we affinity for munc18-1, has a nearly 10-fold reduced affinity
employed both direct stochastic optical reconstruction invitro5asaresultofacombineddisruptionofboth‘N-peptide’
microscopy microscopy (dSTORM18,19) and photoacti- and ‘closed-form’ interaction. It is important to note that this
vatable (PA) localization microscopy (PALM20,21). dSTORM mutantstill interactswithmunc18-1atleastinvitro,and soany
involved the immunodetection of endogenous proteins with a effects we observe likely point to a partial disruption of
fluorophore-conjugatedantibody(Alexa-647)drivenintoalong- interaction as opposed to a complete abolition of binding. This
lived ‘dark-state’ using high-intensity laser illumination in the approachallowedustoquantifymunc18-1molecularbehaviours
presence of a reducing buffer19. Single Alexa-647 fluorophores, in detail in cells readily identifiable as containing mutant
conjugated to secondary antibodies, spontaneously re-emerge syntaxin1a by virtue of their cyan fluorescence. Comparing data
from this dark state, permitting the localization of individual from this system with that where we introduced wild-type (wt),
epitopes separated in a time stack. Munc18-1 and syntaxin1a full-length syntaxin1a as a control revealed that munc18-1
moleculeswerethusdetectedinchemicallyfixedcorticalneurons molecular trafficking was dependent on interaction with
(fixation was for 90min to ensure complete immobilization). syntaxin1a; disruption of binding with syntaxin1a resulted,
Subsequent co-staining against synapsin was performed to qualitatively, in a more uniform distribution of munc18-1
delineate presynaptic areas. dSTORM imaging revealed that tracks (Fig. 2b). Quantifying munc18-1 track lengths revealed
both endogenous munc18-1 and syntaxin1a molecules (with the significantly longer movements when SNARE interactions were
caveat that we detect immunolabelled complexes larger than the disrupted (B25% longer; 1.96±0.6nm (mean±s.e.) when co-
2 NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
Synaptic vesicle containing
neurotransmitter
Synapsin
Syntaxin1a Synaptobrevin
SNAP25
Munc18-1
(?)
(?) (?)
(?)
(i) (ii) (iii) (iv) (v)
Synaptic cleft
EGFP N Habc H3 TM Syntaxin1a
mCherry D1 D2 D3 D2 Munc18-1
Syntaxin1a Synapsin Merge PA-syntaxin
1
8-
1
c
n
u
M
P-
F
G
PA-Munc18-1
n
xi
a
nt
y
s
P-
F
G
Figure1|Munc18-1andsyntaxin1asingle-moleculedistributioninneurons.(a)Modeloftheproposedmunc18-1(red)andsyntaxin1a(green)
interactionsatasynapse.Munc18-1bindssyntaxin1ainclosedconfirmation,preventingsyntaxin1afromenteringtheSNAREcomplexandinhibiting
membranefusion(i)Thebindingmodeofmunc18-1boundtosyntaxin1aswitchesfromtheclosedtoopenmode,allowingtheformationofthebinary
t-SNAREcomplex(ii)SNAREbinarycomplexwitht-SNAREpartnerSNAP-25ingrey(iii).Ternarycomplexofopensyntaxin1a,SNAP-25andsynaptobrevin
requiredformembranefusion(iv).Questionmarksrepresentuncertainpointsofsyntaxin-munc18-1molecularinteractioninthesynapticvesiclecycle.
(b)Schematicillustratingthesyntaxin1aandmunc18-1constructsusedinthisstudy.(c)dSTORMmapofimmunodetectedsyntaxin1a(Alexa-647,upper
left)andsynapsin-EGFP(upperright)incorticalneurons.Amergedimage(grey,upperright)showsoverlap.Lowerpanel:adSTORMmolecularmapfrom
theboxedareainthemergeimageshowsthelocationsofsingleimmunodetectedsyntaxin1amoleculesconcentratedinsynapsin-positivesynapseswith
sparsedistributionelsewhereintheneuron.(d)PALMlocalizationmapsshowsinglemoleculesofPA-mCherry-syntaxin1aorPA-mCherry-munc18-1
co-clusteringwitheitherEGFP-munc18-1orEGFP-syntaxin1a,respectively.Theboxedregionsaredisplayedatahigherzoom(toppanels).Scalebars,
500nm.Thedistributionofheterologousmunc18-1andsyntaxin1afluorescentfusionproteinmoleculesissimilartotheendogenouspattern.
expressed with wt syntaxin1a, 2.45±0.5nm in the presence of with 36 bins of 10(cid:2) angles adding to a circle of 360(cid:2). The
SyxD6,L165A,E166A,n¼1,613and3,605tracks fromthree cells, size of each ‘pie piece’ reports the number of measurements in
respectively, Po0.05). each histogram bin, with single-molecule speed shown as a
A major advantage of sptPALM, particularly if the most coloured bar.
advancedtrackingalgorithmsareused12,13,28,isthedeliveryofa Using this approach, we wanted to determine whether
vast amount of content describing the movements of large munc18-1 molecule movements in living neurons were affected
numbers of single molecules. The large samples and accurate by the disruption of interaction with syntaxin1a. To attempt to
data in turn offer statistical certainty and resolution much measure this, we selected one parameter; the angle of every
greater than can be achieved when using bulk studies. Taking munc18-1 molecular movement relative to the direction of the
advantage of this high-content data, and to further understand preceding step in each track. This analysis reveals whether
the effect of syntaxin1a interaction on munc18-1 molecular molecules diffuse freely, apparent as a completely round
dynamics, we plotted every trajectory angle taken by every histogram (that is, with the same number of measurements in
munc18-1 molecule, plotting in circular histograms, or ‘Rose each 10(cid:2) bin), or if they are they are somehow directed or
Diagrams’12,13, to illustrate differences in trajectories and speeds constrained, evident as a skewing of the histogram in one
(Fig. 2c). The Rose diagram is simply a circular histogram preferred direction or another—so it becomes ovoid in shape.
showing the direction taken between steps of a molecular track, This detailed analysis found a decrease in the proportion of
NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications 3
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
Munc18-1 molecular
track displacement
a
1
n
axi
nt
WT sy S Ly1n6t5aAxi,n E1a1 6Δ66A
ΔSyntaxin1a 6L165A, E166A Time Log Munc18-1track density 05 WT syntaxin1a
wt syntaxin1a Syntaxin1a Δ6 L165A, E166A
interaction modes intact interaction modes disrupted
15% 15%
10% 10%
3’’
1 Θ’2’ Θ’
3’
2 12
Munc18-1
Velocity μm s–1
Molecular track reversals
a
n1 0.35 Syntaxin interaction modes disrupted
WT syntaxi d frequency 000...223050 Syntaxin interaction modes intact
e 0.15
Δntaxin1a 665A, E166A Normaliz 000...0010500 2 4 6 8 10 12
SyL1 Munc18-1 speed (μm s–1)
Figure2|Munc18-1moleculardistributions,speedsanddynamicsaresyntaxin1ainteractiondependent.(a)Theas-the-crow-fliesdisplacementsof
munc18-1single-moleculetracksarerepresentedinarrowform(thearrowconnectsthestartandendpointofeachmoleculartrajectory,withthe
arrowheadindicatingmoleculedirection).Moleculartracksareshownbycolouredlines,withthecolourindicatingthetimeintheimagesequence.
Munc18-1molecularaccumulationsinvaricositiesbecomedisruptedinthepresenceofsyntaxin1aD6,L165A,E166A).Scalebar,1mm.Colourbar:starttime,
black;endtime,white.Thediffraction-limitedimageisshowningrey.(b)Densityplotshowingthenumberofmoleculartrackscrossingeachpixelinthe
image,demonstratingmunc18-1accumulationinsynapsesdecreasedwheninteractionwithsyntaxin1aisdisruptedandthattheintroductionofthemutant
didnotdecreasemunc18-1molecularnumber.Colourcodingshowsthetracknumber.Scalebar,10mm.(c)Leftpanel:diagrammaticrepresentationofthe
Rosediagramcalculations.Onceamoleculardirectionisestablished,everyanglebetweenmoleculartrajectorysteps(2–30 or2–300)ismeasured.Each
munc18-1angle(Ø0orØ00)isthenincorporatedintocircularhistogram‘RoseDiagrams’(centreandrightpanels).Inthisanalysis,eachofthe36segments
representsabinof10(cid:2)angle,segmentmagnituderepresentsthenumberoftrajectoriesineachhistogrambinandcoloursindicatethemoleculartrack
speed.Molecularreversalsarethusshownasaleftwarddeflection,asmoleculesreturnfromwheretheyoriginated.Theseshowanincreaseinreversing
munc18-1moleculesandaspecificdecreaseinthenumberoftheslowest-movingmoleculeswithaconcomitantincreaseintheproportionofthefastest
munc18-1moleculeswheninteractionwithsyntaxin1aisdisrupted.(d)Reconstructinganimagetoillustratewheretrackreversalsoccurhighlightsalossof
munc18-1directionalitywheninteractionwithsyntaxin1aisdisrupted.Greenspotsindicatecompletetrackreversals;greybackgroundrepresentsthe
outlineofthecells.Scalebar,5mm.(e)Quantifyingmolecularspeedsshowsthatwhenbindingwithsyntaxin1aisdisrupted,thereisarelativedecreasein
theproportionofslowestmunc18-1moleculesandaconcomitantincreaseinfaster-movingmunc18-1.Greybars,munc18-1moleculespeedsinthe
presenceofsyntaxin1aD6,L165A,E166A);blackbars,controlmunc18-1molecularspeeds.Errorbars¼s.e.m.
slowest-moving munc18-1 molecules and an increase in the We also found a relative increase in munc18-1 trajectory
fastest (Fig. 2c,e), when both interaction modes with syntaxin1a reversals when syntaxin1a interaction was disrupted (5,602
were disrupted. This speed change is indicative of a loss of molecular track reversals when interaction was disrupted, versus
interaction,assolublemunc18-1moleculesarepredictedtomove 2,416wheninteractionwasnormal):thisisapparentasaskewin
morequicklythanthoseinamembrane-associatedcomplexwith theRosediagramtotheleft,asthemovementbetweentrajectory
syntaxin1a. steps is prone to be backwards.
4 NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
This analysis alone does not allow a full interpretation: to the microsecond timescale. This has the advantage over more
understandthesedata,spatialinformationisrequired.Therefore, commonly used FRET assays in that it can report molecular
to interpret this complex behavior, we extracted the molecules dissociations as well as interactions, essential when analysing a
that showed these predominant direction reversals and recon- dynamic molecular pathway (Fig. 3a). FCS was performed in
structed these into an image representing their original position electrically active cortical neuronal synapses in cells expressing
in the cell (Fig. 2d). This revealed that when both the available fluorescent reporter molecules (Fig. 3b). To gain specificity and
modesofinteractionwithsyntaxin1aweredisruptedbymutation examine interactions at different points in the synaptic vesicle
to Syx D6, L165A, E166A, munc18-1 molecules lost directionality pathway, we employed botulinum neurotoxin (BoNT)-resistant
throughout the entire cell, whereas in the presence of wt molecular probes, expressed at very low levels (10–100s of
syntaxin1a, such molecular behaviours were less spatially molecules per measurement) as tracers, and specifically removed
homogeneous (Figs 1 and 2 and Supplementary Fig. 1). Finally, endogenous SNARE proteins as required to reveal interaction
we plotted the same munc18-1 single-molecule velocities in a modes at defined points in the synaptic secretory pathway.
traditionalhistogram,reiteratingthatwhensyntaxin1ainteractions First,wetestedwhetherwecouldquantifymoleculardiffusion
were disrupted, molecular speeds increased (Fig. 2e). Together, rates and interactions (Fig. 3c–e) by expressing EGFP and
these experiments show that munc18-1 molecular behavior and mCherry fluorescent proteins, and performing FCS measure-
directionality, at least in part, depend on the interaction with ments. Protein expression was driven from a ‘crippled’ cytome-
syntaxin1ausing one of theavailablebinding modes. galovirus (CMV) promotor36 ensuring that tracer levels of
The molecular speeds we measure are faster than those heterologous proteins were present, with varicosities containing
previously reported for these and other neuronal proteins using the lowest detectable fluorescence selected for analysis. Both the
lower temporal resolution approaches23,29, possibly due to the EGFPandmCherryphotonfluctuationscouldbemeasuredwith
under-sampling in earlier studies. To clarify this, we performed a2-msacquisitionrateallowingproteinmolecularconcentrations
fluorescence recovery after photobleaching (FRAP) experiments to be measured in situ, ensuring similar expression levels in all
atdecreasingbleachingandsamplingrates.Comparingthiswith experiments.FCShastheadvantageofbeingapplicableequallyin
molecular tracking at the same rates, and with fluorescence cellularsystemsaswellasinvitro,sowecomparedthesedatawith
correlation spectroscopy (FCS) measurements (below) at micro- those acquired from defined concentrations of highly purified
secondrates,wefoundthatapparentmolecularspeedsvariedwith fluorescent proteins in vitro; these calibrations confirmed that we
acquisition rate down to B3Hz, at which point they plateaued, could accurately report molecular behaviours, diffusion rates and
confirmingthatoursamplingratesweresufficienttocapturerapid concentrationsinneurons(SupplementaryFigs3–5).Boxplotdata
molecular dynamicsfrom synapses (SupplementaryFig.2). are also simplified and presented in Supplementary Fig. 4.
Together this combination of single-molecule resolution Representative raw data and fitted autocorrelation functions are
imaging, high-resolution tracking and new informatics showninSupplementary Fig. 5.
approaches reveal that the behavior of munc18-1 in central Asoursingle-moleculeimagingdatasuggestedthatmunc18-1
neurons is controlled by an interaction with syntaxin1a. Which behaviour in synapses was syntaxin1a dependent, we measured
mode of interaction is utilized at the different neuronal sites the rates of diffusion of mCherry-munc18-1 and EGFP-
identified using these approaches, and how differences in mode syntaxin1a in the varicosities identified by synapsin co-staining
may be functionally important in neurons cannot be fully in earlier experiments, again acquiring data with an acquisition
addressedusingimagingalone;spectroscopyisrequiredtobegin rate of 2ms (Fig. 3b). Autocorrelation curves for these data
to probe molecular interactions. accumulated over 5–10s were generated, delivering similar
diffusion rates for munc18-1 and syntaxin1a molecules (Fig. 3f–i;
0.35±0.09mm2s(cid:2)1 and 0.41±0.07mm2s respectively; mean±
Syntaxin1a–munc18-1 interactions in central neurons. s.e.m., n¼at least 10 experiments). Importantly, munc18-1
Munc18-1hasbeenattributedarangeofessentialfunctions,asa diffused at a rate similar to that of syntaxin1a, suggestive of
chaperone26,30, asa ‘docking factor’31,acting withSNAP-25 and protein–proteininteractionintherestingsynapse(asmunc18-1is
synaptotagmin32 and as an essential modulator of the very last asolubleproteinandsyntaxin1aisanintegralmembraneprotein).
stages of synaptic vesicle fusion7, even shaping fusion pore Closer inspection of the diffusional behavior of these proteins
kinetics8.Thesedistinctfunctionssuggestamolecularinteraction revealed that both diffused with directed motion, indicative of
pathway between docking and fusion3, as it is accepted that if membraneassociation(Fig.3j).Toruleoutthepossibilitythatthis
munc18-1regulatesthelatestagesofvesicleexocytosis,itmustbe was simplya resultof the crowded synaptic microenvironment37,
viainteractionwiththeNterminusofsyntaxin1a(asopposedto we performed similar FCS experiments using unfused (that is,
‘closed’ syntaxin interaction) in the ternary SNARE complex6. soluble) EGFPincorticalvaricosities,findinga significantly faster
However, recent data using mutagenesis and electrophysiology diffusionrateof,(5.32±0.83mm2s(cid:2)1,mean±s.e.m.,n¼13,four
showed that whereas munc18-1 has a role in the docking and cells, Mann–Whitney U-test, Po0.001) that followed a model of
priming stages of exocytosis, its continued association with the Brownian freediffusion (Fig.3j).
N-peptide of syntaxin1a appears dispensable for normal Nexttodeterminetherateofdiffusionofmonomericmunc18-
secretion33. What is still not known, however, is whether 1 molecules in the cellular environment, we first chose HEK293
munc18-1 and syntaxin1a dissociate at all during synaptic cells, known not to express syntaxin1a or other munc18-1-
exocytosis or whether they remain bound, and when if tested bindingproteins.Munc18-1inthiscellularexpressionsystemwas
directly, the N-terminal mode interaction persists. Given that cytosolic, and FCS autocorrelation analysis yielded a diffusion
exocytosis is likely to be mediated by only a few SNARE/SM rate of 9.26±1.86mm2s(cid:2)1 (mean±s.e.m., n¼9 independent
molecules12,13,34,35 that we have shown here that localize experiments;SupplementaryFig.3).Thismoleculardiffusionwas
dynamically to synapses in an interaction-dependent way slowerthanforcytosolicEGFPalone(D¼19.66±1.06mm2s(cid:2)1,
(Figs 1 and 2), single-molecule resolution approaches with very mean±s.e.m., n¼9 independent experiments), consistent with
high temporal rates are required to dissect synaptic interactions. the molecular mass of munc18-1-EGFP being four times that of
Toexaminedynamicmolecularinteractions,asopposedtoco- unfused EGFP, resulting in a slower diffusion rate. As synaptic
localizations, in intact synapses, we used FCS that reports munc18-1 had a significantly slower rate of diffusion than
molecular concentrations, diffusion rates and interactions on cytoplasmic munc18-1 in a HEK293, where no syntaxin1a is
NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications 5
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
Z
0
(ii)
(i)
W
0
(iii)
A
p
5 500 ms
Autocorrelation0000011......0246800 DD == 54..98 μμmm22 Auto-ss––11correlation–0404.01tiLmoeg 1 (lm0ags)1,000 Residuals––4404040.01 0.1 1 10 100 1,000 μ2–1Diffusion rate (m s) 110246802 n=++15 n=+13
0.010.1 1 10 100 1,000 Log lag time (ms) mCherry EGFP
Log lag time (ms)
Autocorrelation00001.....24680 DD57 == % 00 ..521629%85 μμ mm22ss––11 Auto-correlation–4040.01 tiLmoeg (1lma0gs)1,000 Residuals–––440440220 μ2–1Diffusion rate (m s) 00112.....05050 + +
0.0 1 1 1 0 0 0 n=15 n=10
0.01 0.1Log l1ag tim1e0 (ms1)00 1,000 0.0 L0.og lag tim1e (m10s) 1,00 MmuCnhce1r8ry-1- syEnGtaFxiPn-1a
EGFP-syntaxin1a 2–1m s)1102 EmEGGCFFhPPer-rsyyntaxin1a
mCherry-Munc18-1 ** μe ( 8 mCherry-Munc18-1
at 6
EGFP on r 4
mCherry Diffusi 02
–1.0 –0.5 0.0 0.5 1.0 1.5 –0.8 –0.4 0.0 0.4 0.8 1.2 1.6
Deviation from brownian motion Deviation from brownian motion
Figure3|Syntaxin1aandMunc18-1areinanimmobilecomplexinrestingsynapses.(a)FCSandFCCSwereusedtoprobesyntaxin1a–munc18-1
interactioninsynapsesonarapidtimescale.Thedimensionsoftheexcitationspot(W andZ )aredeterminedexperimentallyusingpurifiedfluorescent
0 0
proteins(SupplementaryFig.3).FCSprovideshighspatiotemporalresolutionofproteindiffusionrateandmode(i)interaction(ii)andreactionkinetics(iii).
(b)Representativecorticalneurontransfectedwithmunc18-1-EGFP(leftpanel)showingvaricosities(whitecircles)wheremeasurementswereacquired.
Scalebar,5mm.ExamplespontaneousmEPSCtrace(rightpanel).(c)RepresentativeautocorrelationfitsofunfusedEGFPandmCherrymoleculesin
neuronalsynapses,autocorrelationtraceofthesamedata(insert).Nocross-correlationcouldbedetected.(d)Fitresidualsofthedatainc.(e)Boxplotof
EGFPandmCherrydiffusioninrestingsynapses.(f)Representativeautocorrelationandcross-correlationfit(blue)resultofsyntaxin1a(green)and
munc18-1(red),rawautocorrelationtraceofthesamedata(insert).(h)Fitresidualsofthedatainf.(i)Boxplotofsyntaxin1aandmunc18-1diffusiondata
inrestingsynapses(simplifiedbarchartsshowingmeansands.e.m.foreachtreatmentarepresentedinSupplementaryFig.3).Theratesofdiffusionwere
calculatedfromthederivedautocorrelationcurves;centrelinesrepresentthemedian;crossindicatesthemean;boxlimitsindicatethe25thand75th
percentilesandwhiskersextendtominimumandmaximumpoints.Notchesare95%confidenceintervalsthattwomediansdiffer.Therearenostatistical
differencesbetweengroups.(j)GraphicalrepresentationofthecalculateddeviationfromBrownianmotionofeachsamplediffusiondata.EGFPand
mCherryinsynapsesdiffusewithaBrownianmotion(indicatedbythebluedashedline),indicativeoffreediffusioninthesynapticcytosol,whereas
syntaxin1aandmunc18-1moleculesdeviatefromthisbehavior,indicativeofmembraneanchoringandthusinteraction(asmonomericmunc18-1isasoluble
protein).Inthisplot(left),thecentralblackbarsindicatethemedian,witherrorbarsindicatings.e.m.Rightpanel:plotting‘DeviationfromBrownian
motion’againstdiffusionraterevealstwopopulationsofmCherry-munc18-1behavioursinsynapses;bothpopulationshavelowdiffusionratesbutdifferin
diffusionalbehavior—onegroupofmoleculesdiffuseinadirectedmannerandtheotherappearscaged.
present, this provided further evidence of an interaction with synapses; bothsharedsimilardiffusionrateswithsyntaxin1abut
syntaxin1a. Correlating diffusion rate with ‘Deviation from differedindiffusionalbehaviour,suggestingthatasmallsubsetof
Brownian motion’—a parameter describing the molecular munc18-1moleculeswereassociatedwithacomplexinsynapses
motions as they pass through the excitation volume38—revealed distinct from a direct interaction with syntaxin1a. Notably here,
that two populations of munc18-1 molecules were detectable in no syntaxin1a was ever detected that did not have munc18-1
6 NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
molecules sharing identical dynamics (Fig. 3j). As a final, Determiningsyntaxin1a-munc18-1pre-fusioninteractionmode.
additional test of protein–protein interaction, we cross- Syntaxin1a and munc18-1 can adopt functionally and spatially
correlated the fluorescence fluctuation data (fluorescence distinct modes of interaction5,6. As we found that munc18-1
cross-correlation spectroscopy; FCCS) acquired from mCherry- interactswiththet-SNAREheterodimer,thissuggestsbindingto
munc18-1 and EGFP-syntaxin1a molecular signals in synapses, the so-called N-peptide motif in syntaxin1a. To address this, we
finding that B60% of each binding partner co-diffused in the transfected cells with a phosphomimetic BoNT/C-resistant
synapse (Fig. 3f). Despite this result, this latter measure, FCCS, Syntaxin1aS14ECR. Syntaxin1a has a consensus phosphorylation
proved unreliable in our hands as it was technically difficult to siteforCaseinkinaseIIatSerine14;40%oftotalbrainsyntaxin1
cross-correlate data from the short recordings we required in was shown to be phosphorylated at this site in axons but not in
small synapses. However, the multiple alternative parallel active zones23; furthermore, we showed recently that charge
measures confirm that we could quantify, robustly, protein– alteration at this site destabilized specifically the interaction
protein interactions inside synapses with high temporal betweentheN-terminalpeptidemotifofsyntaxin1aandmunc18-
resolution and that the diffusion rates measured, combined 1, whilst leaving the alternative modes of interaction between
withdatadescribingmolecularfreedom,actasanaccurate,rapid thesepartnersintact10,41.Asthismutanthasasignificantlylower
reporter for direct interactions. affinity for munc18-1 than wt-syntaxin1a10, a (toxin resistant)
A concern about these assays is that the background of construct was thus used in neuronal cells where endogenous
endogenous syntaxin1a could interact with our mCherry- syntaxin1a was removed using BoNT/C poisoning (see Fig. 4);
munc18-1 probe expressed at tracer levels, so skewing our data. EGFP-Syntaxin1aS14ECR and mCherry-munc18-1 co-localized
Tocontrolforthisandimprovethespecificityoftheseassays,we withanarrangementnotdissimilarfromtheendogenousprotein
removed endogenous syntaxin1a from synapses by treatment patterns,withenrichmentinsynapses(Fig.5a).FCSanalysiswas
with BoNT/C, which cleaves proteolytically syntaxin1a. As a test then employed to further understand the interaction between
for endogenous syntaxin1a cleavage and to further test our theseproteins(asopposedtotheco-localization,whichislimited
hypothesis that munc18-1 interacts with syntaxin1a in the bytheresolutionoftheimagingsystem(B250nminthiscase)).
synapse, we determined the diffusion rate of mCherry-munc18- Theseexperimentsfoundthattherateofdiffusionandrestricted,
1 molecules in these cells. mCherry-munc18-1 in these samples membrane-associated behaviour of syntaxin1aS14ECR was
had a significantly faster diffusion rate than in cells with intact indistinguishable from that of wt EGFP-syntaxin1a (Fig. 5b,d).
syntaxin1a (diffusion rate in BoNT/C-treated cells¼3.27±0.44 Importantly, in these synapses, mCherry-munc18-1 followed a
mm2s(cid:2)1, increasing from 0.35±0.09mm2s(cid:2)1, mean±s.e.m., freelydiffusingmodelwitharateofdiffusionapproximatelyfour
n¼10, Po0.001). This rate was also slower than that we times faster than syntaxin1aS14ECR (Fig. 5b,c) but in neuronal
determinedforcytosolicmonomericmunc18-1inHEK293cells, processes,behavedinamannerindistinguishablefromthatwhen
indicative of the crowded synaptic environment37. associated with wt syntaxin1a (Fig. 5b,d). Notably, when the
Next we introduced EGFP-syntaxin1a mutated to be BoNT C diffusionratewasplottedagainstdiffusionalbehaviourasbefore,
resistant39; (syntaxin1a-CR; Fig. 4a). Munc18-1 and syntaxin1a- complexes that behaved in a tightly caged manner, with a very
CRcolocalizedwithnogrossspatialreorganizationinthesecells, restricted diffusion rate, were absent compared with when wt
imaged at diffraction-limited resolution. FCS studies as before syntaxin1a was present (Fig. 5d).
confirmedthatthetwoproteinpartnersinthiscleanbackground TogetherthesedataindicatethatS14modificationdisruptsthe
co-diffused with statistically similar diffusion rates (Fig. 4b; interaction between syntaxin1a and munc18-1 and thus demon-
0.68±0.18mm2s(cid:2)1 and 0.72±0.2mm2s(cid:2)1, respectively)— strate that munc18-1 interacts predominantly with the N-term-
furthermore, these behaviours were identical to those found in inal motif of syntaxin1a in resting synapses. Furthermore, a
non-toxin-treated samples (see Fig. 3). specific subpopulation of complexes in synapses, identified by
Thus, these combined experiments demonstrate that we can virtueofsimultaneouslycaged,non-Brownianandslowdiffusion,
detect and quantify presynaptic protein interactions on the was absent when syntaxin1a–munc18-1 N-terminal interaction
molecularlevelatthehighspeedsnecessarytodeterminedynamics. was disrupted in synapses.
We hypothesized that this subpopulation of complexes,
representing on an average B10% of SNARE molecules,
Dissecting interaction modes during the vesicle fusion cycle. represented those ‘release-ready’ SNARE complexes with
Having confirmed that we could detect the interaction in munc18-1 engaged that may go onto support synaptic vesicle
synapses, we next set out to determine whether the munc18-1 fusion during an action potential. We tested this in two ways;
there interacted mainly with monomeric syntaxin1a, or with the first,astheassociationofthevesicularSNARE,synaptobrevin,is
t-SNARE heterodimer or the ternary SNARE complex. To knowntooccuraftert-SNAREheterodimerformation42,evenin
address the first question, we treated samples with BoNT A the presence of SNARE complex-associated munc18-1 (refs
(BoNT/A),whichcleavesthecarboxyl-terminalnineaminoacids 5,6,43), we treated neuronal preparations with tetanus
from SNAP-25, altering the t-SNARE heterodimer conforma- neurotoxin (TeNT) to cleave specifically synaptobrevin44. FCS
tion11,40. Further, it is also known that specifically BoNT/A after this treatment showed that in resting synapses, syntaxin1a
cleavage of SNAP-25 does not alter the affinity of the t-SNARE andmunc18-1interactedasbeforebutexhibitedratesofdiffusion
heterodimer partners40. Performing our FCS analyses as before, in complex significantly faster than found in untreated synapses
usingtracerprobesinsynapses,revealedthatbothsyntaxin1aand (SupplementaryFig.6),suggestingthatthemunc18-1–syntaxin1a
munc18-1 increased their diffusion rates by a similar extent complexes we detected could be associated subsequently with
comparedwithuntreatedsynapses,consistentwithbeingpartofa synaptobrevin. The hypothesized ‘release-ready’ molecular
complex with lower molecular mass (Fig. 4c,d). Furthermore, complex diffusional behaviours we had previously observed, were
analysisofthemodeofdiffusionfoundthatwhilstsyntaxin1aand absent afterthissynaptobrevincleavage (SupplementaryFig.7).
munc18-1remainedincomplex,theyco-diffuseddifferentlyafter
BoNT/Atreatment,adoptingalessrestrictedbehaviour(Fig.4e).
Together these data suggest that munc18-1 interacts with the Following synaptic protein interactions during exocytosis. We
t-SNARE heterodimer predominantly and not with monomeric next wanted to determine the synaptic protein interaction path-
syntaxin1a in resting synapses. way during depolarization and exocytosis. A recent study
NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications 7
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
Syx 1a 2–1m s) 3 + BoNT/C
Syx 1a/CR? μe ( 2
BoNT/CSNAP25 usion rat 1 ++ + + +
Diff 0
n number =15 10 15 8
Synaptic cleft mCherry- EGFP- mCherry- EGFP-
Munc18-1 syntaxin1a Munc18-1 syntaxin1a CR
+ BoNT/A
4
*
MuncS1y8x 1a 2–1m s) 3 *
SNAP25 BoNT/A μate ( 2
n r
o
usi 1 + +
Diff
+ +
Synaptic cleft 0
n number =15 10 8 8
mCherry- EGFP- mCherry- EGFP-
Munc18-1 syntaxin1a Munc18-1 syntaxin1a
A
T/ EGFP-syntaxin1a
N
o
B mCherry-Munc18-1
+
T/C EGFP-syntaxin1a CR
N
Bo mCherry-Munc18-1 *
+
EGFP-syntaxin1a
mCherry-Munc18-1
–1 0 1
Deviation from brownian motion
Figure4|Munc18-1interactsdirectlywithsyntaxin1ainthet-SNAREheterodimer.(a)CartoonillustratingBoNT/Ccleavageofendogenoussyntaxin1a
inasynapse.Thequestionmarkillustratestheexperimentalquestion:domunc18-1andsyntaxin1ainteractinthepresynapticarea.(b)Boxplotcomparing
thediffusionratesofWTMunc18pairedwithsyntaxin1aorsyntaxin1a/CR.LineindicatesBoNT/C-treatedneurons.Bothproteinpairsco-diffusedwith
similarratesandtoxincleavagedidnotresultinanysignificantchanges.(c)ModelillustratingBoNT/AcleavageofendogenousSNAP-25inasynapse.
(d)BoNT/AcleavageofSNAP-25resultsinanincreaseddiffusionrateofbothmunc18-1andsyntaxin1acomparedwithnon-treatedcontrols.Starsindicate
statisticalsignificance(Mann–WhitneyU-test,Po0.05).(e)Syntaxin1aandmunc18-1inBoNT/Atreatedneuronsco-diffuseatasimilarrateandexhibit
non-Browniandiffusion.
reported a large-scale, temporary dispersal of munc18-1 from that the presence of our probes at such low levels supported
synapses during electrical stimulation23, but it remains unclear normal fusion kinetics (Supplementary Fig. 8c). Exocytosis
whether all the munc18-1 redistributes from the synapse during continued for several seconds beyond voltage-gated Ca2þ
stimulationorwhethersomeremainsassociatedwithsyntaxin1a channel inactivation, determined by using transfected pHluorin
andonlysolublemunc18-1disperses.Ifthelatteristrue,thenitis reporters in combination with bafilomycin blockade of vesicle
credible that the munc18-1 molecules that remain bound to reacidification(SypHy)assays(SupplementaryFig.8d);thesedata
syntaxin1a are contributing functionally to the synapse, whereas indicated that Ca2þ concentration was sustained at levels
the soluble, mobile pool represents a reserve. Therefore, to sufficient to elicit vesicle fusion and that the expression of our
determine whether neuronal activity resulted in a change in fusion proteins had no effect on vesicular fusion kinetics, or the
functionalinteraction(eitherdissociationorinmodeofbinding) sizeoftheRPorRRP,ascomparedwithnon-transfectedcontrols.
between munc18-1 and syntaxin1a at the molecular level, we Maximal stimulation elicited a reported release of 50±10.2%
deliveredtrainsofelectricaldepolarizationsduringFCSrecording of synaptic vesicles across all synapses—equating to B100
to determine on the molecular level whether interactions synaptic vesicles per bouton. Thus, immediately on the
persisted. Parallel high-speed measurements (Supplementary initiation of depolarization for 30s, we performed FCS
Fig. 8a) of synaptic calcium levels and current-clamp electro- measurements, to probe specifically the period within synapses
physiology found that repetitive neuronal action potentials whenCa2þ wasmaximal.Nochangewasseeninthebehaviours
were maintained for the time course of the stimulation with a of unfused-mCherry or -EGFP that remained freely diffusing in
mean time of B5.4s in these nerve terminals (Supplementary synapses, confirming that the stimulation regime had no non-
Fig. 8b). We established that these stimulation regimes induced specific effect on the synaptic microenvironment or our assay
depletion of synaptic vesicle pools using FM-dye assays, and (Supplementary Fig. 8e).
8 NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
EGFP-syntaxin1a S14E CR Synapse
–1s) 8 *
2m 6
μ
mCherry-Munc18-1 ate ( 4 ++ Axon
n r
sio 2 + +
Diffu 0 +
Merge
n number = 6 5 8 8
mCherry- mCherry-
Munc18-1 Munc18-1
EGFP-syntaxin1a EGFP-syntaxin1a
S14E CR S14E CR
1.2 1.2
1.0 1.0
Autocorrelation 000...468 DD == 15..2150 μμmm22 ss––11 Autocorrelation 000...468 DD == 22..0383 μμmm22 ss––11
0.2 0.2
0.0 0.0
0.01 0.1 1 10 100 1,000 0.01 0.1 1 10 100 1,000
Log lag time (ms) Log lag time (ms)
?
Synaptic cleft
Figure5|Munc18-1interactswithsyntaxin1ainvaricositiesviathesyntaxin1aN-terminalpeptidemotifbutcanusealternativeinteractionmodesin
neuronalprocesses.(a)BoNT/C-treatedneuronstransfectedwithmCherry-munc18-1andEGFP-Syntaxin1aS14ECR;bothproteinshavealocalizationnot
dissimilarfromthenon-treatedmCherry-munc18-1-EGFP-syntaxin1a.Scalebar,1mm.Theboxedareaiszoomedintherightpanels(scalebar,1mm).
(b)Munc18-1andsyntaxin1adiffusionratesmeasuredinsynapticregions.Thediffusionrateofmunc18-1increasedsignificantlycomparedwith
syntaxin1aS14ECRindicatingthatthetwoproteinsnolongerinteract.Inaxons,theinteractionbetweenthetwoproteinsisstabilizedwithbothco-diffusing
withanidenticalrateandmode.(c)RepresentativeautocorrelationfitsofmCherry-munc18-1andEGFP-syntaxin1aS14ECRinsynapses.Thediffusioncurve
ofmunc18-1nolongerhasasharpdecayindicatingaswitchinbehaviourtofreediffusion.(d)RepresentativeautocorrelationfitsofmCherry-munc18-1and
EGFP-syntaxin1aS14ECRinneuronalprocesses.Bothproteinshavesimilarratesandmodesofdiffusion.(e)Cartoonsummarizingthisfinding:intheresting
synapse,munc18-1ispredominantlyassociatedwithopensyntaxin1aviatheN-peptideinteraction.
FCS experiments during depolarization showed identical 1 in resting neurons. Surprisingly, during depolarization, an
diffusion rates and behaviours for both syntaxin1a inducedinteractioncouldbedetected,withsyntaxin1aremaining
(1.18±0.37mm2s(cid:2)1) and munc18-1 (0.78±0.2mm2s(cid:2)1, n¼at membraneassociatedandthepreviouslyfast,Browniandiffusion
least 10 independent experiments, mean±s.e.m.), which did not of munc18-1 in syntaxin1aS14ECR synapses slowing to rates
alter from the resting state, with both showing a restricted identical to syntaxin1aS14ECR with a restricted motion (Fig. 6b).
diffusion pattern indicative of a persistent interaction during Thiseffect wasconfinedtosynapses;measurements fromaxonal
depolarization(Fig.6a).Importantly,therewasanenrichmentof or dendritic regions of the cells showed no interaction between
‘release-ready’ complexes during depolarization (to B20% of syntaxin1aS14ECRandmunc18-1underanycircumstance.These
detected molecules, compared with B10% in resting synapses), experiments indicate that whereas in resting synapses the
lendingweighttoourconclusionthatwecouldidentifythesubset N-peptide interaction predominates, during depolarization and
ofcomplexesactiveinsynapticvesicleexocytosisintheseboutons exocytosis, a molecular switch to munc18-1 interacting with
(Fig. 6a). closed-form syntaxin1a occurs. However, when we plotted dif-
fusion rate against ‘deviation from Brownian motion’ as before
(Fig. 6b),our measure thatwe postulated couldidentify ‘release-
Syntaxin1ainteractionswitchoccursafterSNAREdisassembly. ready’ syntaxin1a–munc18-1 complexes, we did not find the
Next we performed experiments using the syntaxin1aS14ECR slow-moving, caged complexes apparent previously (see Fig. 6a),
mutant,whichwefoundpreviouslynottointeractwithmunc18- presenting a conundrum.
NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications 9
&2014MacmillanPublishersLimited.Allrightsreserved.
ARTICLE
NATURECOMMUNICATIONS|DOI:10.1038/ncomms6774
Syntaxin1a Syx1 before stim
2–1m s) 23 20 Hz 21..06 MSMy11x881 b ddeuufrorininrgeg ssstttiimmim
μsion rate ( 01 + + + + Diffusion10..28
Diffu −1 0.4
n =15 10 15 6 0.0
Munc18-1Syntaxin1a Munc18-1 Syntaxin1a –0.4 0.0 0.4 0.8 1.2 1.6 2.0 2.4
Deviation from brownian motion
2–1m s) 68 * 20 Hz 2.0 EGFPS-s1y4nEtaCxRin1a Diffusion2468
μusion rate ( 24 + + + + Diffusion 110...628 –00.80.00.81.62.43.2
Diff 0 0.4
n = 6 5 8 7
0.0
mCherry- mCherry- –0.4 0.0 0.4 0.8 1.2 1.6 2.0 2.4
Munc18-1 Munc18
Deviation from brownian motion
EGFP-syntaxin1a EGFP-syntaxin1a
S14E CR S14E CR
Non-FRET
lifetime threshold
1.2
y
nc 1.0
EGFP-syntaxin1aS14E CR e
u
0.4 Munc18-1 -FCCS eq 0.8
Resting synapse d fr 0.6
Depolarized synapse e
0.2 Resting fit aliz 0.4
CCS 0.0 Depolarized fit Norm 00..20
F 0 1 2 3 4
–0.2
Fluorescence lifetime (ns)
20 Hz AP
–0.4 No STIM + NEM 20 Hz AP
0.01 0.1 1 10 100 1,000 a
Log lag time (ms) xin1nor
ao
yntR d
GFP-sS14E C
E
Closed-syntaxin1a
Munc18-1
Trafficking
complex
?
Closed-syntaxin1a
Recycling vesicle
Synaptic vesicle containing
neurotransmitter
? Open-syntaxin1a
SNAP25
Closed-syntaxin1a
Munc13
Munc18-1
?
Synaptic cleft
10 NATURECOMMUNICATIONS|5:5774|DOI:10.1038/ncomms6774|www.nature.com/naturecommunications
&2014MacmillanPublishersLimited.Allrightsreserved.