Imaging signal transduction in single dendritic spines during synaptic plasticity Ryohei Yasuda (HHMI, Duke)

Spine Spine: Biochemical compartment Small ~0.1 fL. Narrow neck (~100nmΦ) : Diffusional barrier Ca 2+ signaling in spines  Synaptic plasticity  Memory

Signaling networks R. Iyenger

Imaging signaling in single spines Measure FRET with 2-photon fluorescence lifetime imaging (2-photon FLIM) Develop and use FRET sensors optimized for 2-photon FLIM Image signal transduction, while inducing plasticity in single spine with 2-photon glutamate uncaging

FRET and fluorescence lifetime FRET decreases fluorescence lifetime. Use donor fluorescence only. –Independent of fluorophore concentrations. –Independent of wavelength-dependent light scattering Multiple populations can be deconvolved. Easy to combine with 2-photon microscopy Log (fluorescence) Time (ns) Laser pulse FRET Mixture Donor

2-photon fluorescence lifetime imaging microscopy High resolution and sensitivity in deep tissue. Quantitative measurements of FRET

Stimulate single spines using 2-photon glutamate uncaging Synapse-specific Ca 2+ elevation Matsuzaki, Ellis-Davies, Kasai

2p-uncaging to produce long lasting synaptic potentiation and spine growth times, Hz in Zero extracellular Mg 2+ Matsuzaki 2001, 2004

Imaging activity of CaMKII Ca 2+ /Calmodulin-dependent kinase II

Ca 2+ /Calmodulin-dependent kinase II: biochemical memory? Kinase domain Low Ca 2+ Ca 2+ Activation Memory? Ca 2+ /CaM CaM P Ca 2+ P Inactive GFP FRET d-YFP Takao et al, 2004 Lisman 2002 Autoinhibitory domain Kinase domain

Fluorescence lifetime change in lysates Time (min) CaM+, ATP+ CaM+, ATP+(T286A) CaM+, ATP– CaM–, ATP+ CaM–,ATP– Fluorescence lifetime change (ns) Ca 2+ EGTA Lee et al., Nature 2009

CaMKII activation during structural plasticity of spines Lee et al., Nature 2009

Transient and spine-specific activation of CaMKII Lee et al., Nature 2009

Transient and spine-specific activation of CaMKII ? CaMKII Activity Change Stimulated Spine Adjacent Spine Dendrite Uncaging 35 2 min CaMKII Uncaging Volume Lee et al., Nature 2009

Ca 2+ /Calmodulin-dependent kinase II: biochemical memory? Kinase domain Low Ca 2+ Ca 2+ Activation Memory? Ca 2+ /CaM CaM P Ca 2+ P Inactive Not for 1 hour What is the role of phosphorylation?

T286A Wild-type Fast fluorescence lifetime imaging Lee et al., Unpublished s 5.8 s Time (s) Fluorescence lifetime (ns) 1 μM Δ[Ca 2+ ] ~0.1 s Glutamate uncaging (4 ms)

Ca 2+ /Calmodulin-dependent kinase II: biochemical memory? Kinase domain Low Ca 2+ Ca 2+ Activation Memory? Ca 2+ /CaM CaM P Ca 2+ J. Lisman P Inactive Yes, but only for 6 s ~ 6 s

Ca 2+ CaMKII Long-term plasticity 0.1 s10 s10 min1 hour1 min Previous view of LTP

Ca 2+ CaMKII Long-term plasticity 0.1 s10 s10 min1 hour1 min Now ….

Imaging the activity of Ras superfamily proteins

Small GTPase signaling Several major subgroups: Ras, Rho, Rab, Rap, Arf, Ran etc… Acts as signaling switch. Regulate organization of actin cytoskeleton, membrane trafficking etc. Important for morphogenesis of dendritic spines and plasticity Mutations in the pathway are associated with mental retardation

Imaging binding between Ras and Ras binding domain (RBD) of Raf1 Yasuda et al., Nat.Neurosci Harvey et al., Science 2008 CaMKII Ras

Ca 2+ CaMKII Long-term plasticity 0.1 s10 s10 min1 hour1 min Ras  ERK AMPAR exocytosis

General approach to make sensors for Ras superfamily Cdc42 / RhoA: Important for regulation of actin cytoskeleton and dendritic spine morphology a.a. S79A, F89A mGFPCdc42 Pak3 mRFP P 8-89 a.a.. WT mGFP RhoA RTKN mRFP P Cdc42 sensor RhoA sensor

Making small GTPase sensors 1. Screen RXX Binding Domain in cuvette K d ~ 1 – 5 uM for GTP form (RBD inhibits Rho inactivation) K d > 50 uM for GDP form for low background 2. Test sensitivity & specificity in cell line 3. Test sensitivity & kinetics in neurons

Step1: Screen RBD and mutants FRET between GFP-CDC42 and PAK2-mCherry

GFP-RasCdc42 Rac1 RhoA mCherry-RBDGDPGMPPNPGDPGMPPNPGDPGMPPNP PAK PAK1 (F89A) PAK1 (F89A, F96A)913.4> PAK PAK PAK3 (F89A) PAK3 (S79A, F89A) WASP432.9 RTKN473.9

Step 2: Test sensitivity & specificity in HeLa cells

Step 3: Test sensitivity & reversibility in neurons

Cdc42/RhoA activation during LTP

Cdc42 activation is compartmentalized and sustained Hideji Murakoshi Stimulated spine Adjacent spine

RhoA activation spreads and sustained Hideji Murakoshi Stimulated spine Adjacent spine AP5 RhoA activation (%) Time (min) Uncaging

Cdc Binding fraction change (%) Cdc s1-5 min5-10 min10-20 min 0510 Distance (µm) Before 24 s 2.15 ns µm Spine Dendrite Spatial profile of Cdc42

Spatial spreading of RhoA RhoA 24 s Before 2.0 ns Distance (µm) RhoA µm s1-5 min5-10 min10-20 min Binding fraction change (%) Spine Dendrite

Effects of overexpression on length constant Length constant (µm) [mCherry-RBD-mCherry] (µM) [mEGFP-RhoA] (µM) RhoA A n = 20/18 r = 0.05 n = 20/18 r = 0.16

Diffusion coupling at the spine neck Lee, Harvey, Murakoshi Ras proteins: ~5 s CaMKII: ~60 s PA-GFP tagged Ras *Constitutively active mutants * * * *

Ca 2+ CaMKII Long-term plasticity 0.1 s10 s10 min1 hour Cdc42 1 min RhoA

RhoA activation is CaMKII dependent (NMDA receptor inhibitor) (CaMKII inhibitor) Late phase is CaMKII dependent Partial inhibition at early phase Time (min) RhoA activation (%) Ctrl (stim) KN62 AP5 Uncaging Hideji Murakoshi

Control Hideji Murakoshi AP5 (NMDA receptor inhibitor) KN62 (CaMKII inhibitor) Cdc42 activation (%) Time (min) Uncaging Cdc42 activation is CaMKII dependent

Ca 2+ CaMKII Long-term plasticity 0.1 s10 s10 min1 hour Cdc42 1 min RhoA

Cdc42 is required for long-term structural maintenance Hideji Murakoshi Cdc42Sustained growth

Cdc42 is required for long-term structural maintenance Hideji Murakoshi Cdc42Sustained growth Volume change (%) Time (min) Cdc42 binding domain (24 hours)

RhoA is required for transient phase Hideji Murakoshi RhoATransient growth

Stronger inhibition of RhoA inhibits both transient and sustained phases Volume change (%) Time (min) RhoATransient/Sustained growth

RhoA/Cdc42 does not alter Ca 2+ -CaMKII Cdc42 RhoA CaMKII Ca 2+ CaMKII activation Structural plasticity Control C3: Rho inhibitor WASP: Cdc42 inhibitor Time (min) Lifetime change (ns) Time (min) Volume change (%) Time (min) Spine growth

Regulation of spine volume by Rho GTPases Cdc42/RhoA Spine growth CaMKII Hideji Murakoshi Cdc42 Volume CaMKII 2 min Cdc42 Volume CaMKII 10 min Uncaging RhoA

Ca 2+ CaMKII Long-term plasticity 0.1 s10 s10 min1 hour Cdc42 Actin 1 min RhoA Actin Transient plasticity

Downstream of RhoA/Cdc42

Ca 2+ CaMKII Long-term plasticity 0.1 s10 s10 min1 hour Cdc42  PAK Actin 1 min RhoA  ROCK Actin Transient plasticity Ras  ERK AMPAR exocytosis

There are 100 more small GTPase proteins…

Rac1 sensor mGFPRac1 PAK1 mRFP P Rac1 activity Before Stimulation 30 s High Low

RhoA/Cdc42/Rac1 activation time courses RhoA Cdc42 Rac1 RhoA Cdc42 Rac1

Rap1 sensor 0 60 Binding fraction (%) -2 min 6 min Control Forskolin IBMX WTG12VS17N Control Forskolin IBMX 10 um Binding fraction (%) mGFPRap1A RalGDS mRFP P Neuroblastoma cell

Signaling networks R. Iyenger

Ca 2+ Hong Wang CaMKII Seok-Jin Lee Ras/ERK/Rap Ana Oliveira Erzsebet Szatmari Shenyu Zhai Rho Hideji Murakoshi Nathan Hedrick AMPAR Michael Patterson Technical assistance Airong Wan David Kloetzer Funding NIH/NIMH NIH/NINDS NIH/NIDA NSF HHMI Alzheimer’s Association Brandeis Univ. J. Lisman HHMI/Janelia Farm Christopher Harvey Karel Svoboda Yasuda lab Duke Sridhar Raghavachari Michael Ehlers Scott Soderling