glutamate induces de novo growth of functional spine in developing cortex

8/6/2019 Glutamate Induces de Novo Growth of Functional Spine in Developing Cortex

http://slidepdf.com/reader/full/glutamate-induces-de-novo-growth-of-functional-spine-in-developing-cortex 1/5

LETTERdoi:10.1038/nature09986

Glutamate induces de novo growth of functional spines in developing cortexHyung-Bae Kwon & Bernardo L. Sabatini

Mature cortical pyramidal neurons receive excitatory inputs ontosmall protrusions emanating from their dendrites called spines.Spines undergo activity-dependent remodelling, stabilizationand pruning during development, and similar structural changescan be triggered by learning and changes in sensory experiences1–4.However, the biochemical triggers and mechanisms of de novo

spine formation in the developing brain and the functional signifi-cance of new spines to neuronal connectivity are largely unknown.Here we develop an approach to induce and monitor de novo spineformation in real time using combined two-photon laser-scanning

microscopy and two-photon laser uncaging of glutamate. Our data demonstrate that, in mouse cortical layer 2/3 pyramidal neurons,glutamate is sufficient to trigger de novo spine growth from thedendrite shaft in a location-specificmanner. We find thatglutamate-induced spinogenesis requires opening of NMDARs (N -methyl-D-aspartate-type glutamate receptors) and activation of protein kinaseA (PKA) but is independent of calcium–calmodulin-dependentkinase II (CaMKII) and tyrosine kinase receptor B (TrkB) receptors.Furthermore, newly formed spines express glutamate receptors andare rapidly functional such that they transduce presynaptic activity into postsynaptic signals. Together, our data demonstrate that early neural connectivity is shaped by activity in a spatially precisemanner and that nascent dendrite spines are rapidly functionally incorpo-rated into cortical circuits.

During postnatal development, the formation and elimination of glutamatergic synapses are thought to be reflected in the growth andretraction of dendritic spines. In cortical pyramidal neurons, waves of new spine growth (spinogenesis) and synapse formation (synaptogen-esis) occur at specific developmental stages, followed by pruning as thebrain matures5. Many signals have been proposed to trigger and regu-late de novo spine growth in a developing circuit including neurotro-phins, neurotransmitters and cell-adhesion molecules6–9. To uncoverthe triggers for and mechanisms of spinogenesis, we imaged dendritesof enhancedgreen fluorescentprotein (EGFP)-expressing corticallayer2/3 pyramidal neurons while releasing glutamate at a specific dendriticlocation by two-photon laser-induced photolysis of (4-methoxy-7-nitroindolinyl)-glutamate (MNI-glutamate) (Fig. 1). Analysis was per-formed in acute corticalbrain slices from youngmice (postnatal day(P)8–12), a period in which spinogenesis occurs in vivo10.

Stimulation near theedge of a dendrite with 40 0.5-ms laser pulses at0.5Hz in a Mg 21-free extracellular solution induced growth of a new spine in approximately 14% of cases (Fig. 1a–d and Supplementary Fig.1),showing thepossibility of de novospinogenesis inducedby glutamateexposure11. Increasing stimulation frequency and laser pulse durationwhilemaintaining the total number of stimuli at 40 increased therateof spinogenesis such that, at 5 Hz with 4 ms duration, a maximal successrate of approximately 50% was achieved (Fig. 1c). Nascent spines arosefrom the dendrite where glutamate was released with high specificity (Fig. 1b) such that more than 70% of them grew within 1 mm of theuncaging spot (Fig.1e) and94%of themgrewon theside of the dendriteexposed to glutamate.

In 128 of 132 examples of glutamate-induced spinogenesis, thespine was seen to emerge without a filopodial stage (see Supplemen-tary Fig. 2a for an exception). Instead, spine growth occurred incre-mentally but explosively such that the spine head volume increasedfrom 10 to 90% of maximum within 11.861.5 pulses of glutamate(5.96 0.8 s at 2 Hz stimulation) (Fig. 2a–c and Supplementary Fig. 3).The final sizes and lengths of the newborn spines were heterogeneousbut not different from those of pre-existing neighbouring spines(Fig. 2d, e). The lifetime of newly formed spines was variable such thatapproximately 20% lasted less than 2min but those that lasted 5min

were stable and remained for at least 30 min (Supplementary Fig. 4).Thus, these newly formed spines either did not require continuedexposure to glutamate for maintenance or they received glutamatefrom an alternative source such as an axonal bouton.

Howard Hughes Medical Institute, Department of Neurobiology, Harvard Medical School, Boston, Massachusetts 02115, USA.

1 μm

1 μm

Uncaging

Uncaging

Example 1

Example 2

New spine

Layer 2, 3pyramidal neuron

–20

20

40

60

80

100

Distance ( μm)

Spatial specifcity

0.1

0

20

40

60

80

100

S u c c e s s ( % )

0.50

20

40Frequency Duration

Frequency (Hz) Laser pulse (ms)

(14)

(11)

(29)

(29)(33) (14)

(3) (2)(1)

a b

c d e

(3)

(22)

–1 10 240.5 2 5

Figure 1 | De novo spine generation is induced by glutamate uncaging.a , Dendrites of EGFP-expressingneurons in acute slicesfrom P8–12 mice were visualized with two-photon laser scanning microscopy, and glutamate wasreleased by photolysis of caged glutamate near a low-spine density section of dendrite. b, Examples of de novo spine formation induced by photolytic releaseof glutamate (40 pulses of MNI-glutamate uncaging at 2 Hz in Mg 21-freeartificial cerebrospinal fluid). Yellow circles, the uncaging spots; arrowheads,new spines. c–e, Most new spines grew near the uncaging spot and the successpercentage depended on the frequency (c, laser pulse duration5 4 ms) andduration (d, stimulation frequency 5 0.5 Hz) of glutamate uncaging.Experiment numbers for each bar are indicated in parentheses.

0 0 M O N T H 2 0 1 1 | V O L 0 0 0 | N A T U R E | 1

Macmillan Publishers Limited. All rights reserved ©2011

http://www.nature.com/doifinder/10.1038/nature09986

http://www.nature.com/doifinder/10.1038/nature09986



Glutamate-induced spinogenesis was restricted within postnataldevelopmental such that its efficiency diminished by P14–15 and itfailed to occur by P19–20 (Supplementary Fig. 5). This was not due todecreased glutamate receptor activation in older animals because theuncaging-evoked excitatory postsynaptic current (uEPSC) was largerat P19–20 than at P10–12 (Supplementary Fig. 6).

Previous ultrastructural studies have revealed a high frequency of dendrite shaft synapses in hippocampus in early postnatal life thatdecreases as spinogenesis occurs12, leading to a model of synapticdevelopment in which synapses are initially formed directly onto thedendriticshaft anda spine subsequentlygrows from this point with thesynapse attached. On the other hand, rapid movement of a physically connected spine head and axonal bouton together through a complex neuropile is difficult to reconcile with the high density of crossing axons and dendrites13. Our data demonstrate that glutamate uncaging-induced spinogenesis occurswith high spatial specificity and probability on the side of the dendrite exposed to glutamate. These findingsplace alower limit of approximately 1 mm21 for the density of dendritic shaftsynapses required to support this model of spinogenesis. To estimatethenumberof dendritic shaft synapses, Ca21 imaging wasperformed inconditions in which most synapses formed onto a stretch of dendritewere activated(approximately 90%, SupplementaryFig. 7). Under these

conditions,we observedhotspots of Ca21 influx in spineless stretches of dendrite at a density of 0.05mm21. This corresponds to a density of dendritic shaft synapses containing NMDARs that is approximately 20-fold less than necessary to explain the specificity and efficiency of glutamate-induced spinogenesis (see Supplementary Fig. 7 for furtherdiscussion).

Thehighsuccess rateof spinogenesis induced by glutamate uncaging allows identification of the signalling pathways that couple activity tospine growth (Fig. 3). Previous analyses of spine generation induced by electrical stimulation indicate a requirement for NMDARs in this pro-cess14,15, and at this age NMDARs are found throughout the dendrite(Supplementary Fig. 8). Preventing NMDAR activation with the ant-agonist 3-(2-carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP)nearly abolished spinogenesis whereas it was unaffected by inhibiting

a-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA)/

kainate glutamate receptors with 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione (NBQX) (Fig. 3a). The voltage-gatedsodium channel antagonist tetrodotoxin also did not affect spinogen-esis, discarding the possibility that postsynaptic action potentials arenecessary. Addition of extracellular Mg 21 significantly decreased thesuccess rate, suggesting that the degree of current flux throughNMDARs plays a crucial role in triggering spine formation. Blocking either mGluR1 or mGluR5 using 7-hydroxyiminocyclopropan[b]chromen-1a-carboxylic acid ethyl ester (CPCCOEt) or 2-methyl-6-(phenylethynyl)-pyridine (MPEP) or with the less selective group ImGluR antagonist1-aminoindan-1,5-dicarboxylic acid (AIDA) demon-strated that neither was strictly necessary for spinogenesis. Lastly,depleting Ca21 stores with cyclopiazonic acid or thapsigargin signifi-cantly inhibited spine formation. Thus ourdata indicatethat NMDARs,with additional contributions from intracellular stores provide, thecoupling between glutamate and activation of intracellular pathwaysresponsible for spinogenesis.

We further considered intracellular signalling pathways that mightbe activated in a spatially delimited fashion by glutamate and couldprovide the spatial information necessary for local spine growth.Previous studies of long-term potentiation and associated spineenlargement in older animals demonstrated a need for activation of

CaMKII or signalling by neurotrophin receptor tyrosine kinases suchas the BDNF receptor TrkB16–19. However, we found that activity-dependent spinogenesis is unaffected by the kinase inhibitors KN-62, KN-93 and K252a, indicating independence from CaMKII andTrkB signalling (Fig. 3a).

The cyclic AMP (cAMP)-activated kinase PKA is required for long-term potentiation induction in younger neurons20, and gradients in itsconcentration can be maintained over micrometre-length scales21.Raising cAMP concentration by applying theadenylatecyclaseactivatorforskolin was not sufficient to generate spines on its own (n53, datanot shown), but glutamate uncaging in its presence increased thespinogenesis success rate to approximately 80% (Fig. 3a). In the pres-ence of forskolin, multiple spines often grow with each inductionattempt (Fig. 3b, c), including at sites more distant from the uncaging

location (Fig. 3d). In addition, the PKA inhibitor H-89 prevented new

0 s

10 s

12 s

14 s

18 s

20 s

30% Maximum spineintensity to 100%

Line scan

a b c

1 μm

0 1 20.0

0.5

1.0

0.0 0.5 1.00.0

0.5

1.0

0.0 0.5 1.0 1.50.0

0.5

1.0

Spine length ( μm)

C u m u l a t i v e p r o b a b i l i t y

Spine head area ( μm2 ) C u m u l a t i v e p r o b a b i l l i t y

Spine length ( μm)

0.0

1.0

2.0

S p i n e l e n g t h

( μ m )

0.0

0.5

1.0

S p i n e w i d t h ( μ m )

E x i s t i n

g 0.0

0.3

0.6

S p i n e h e a d a r e a ( μ m 2 )

N a s

c e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

N a s

c e n t

C u m u l a t i v e p r o b a b i l l i t y

NS

NS NS

Existing ( n = 111)Nascent ( n = 95)

d

e

0 10 20 30 400

50

100

P e r c e n t a g e o m a x i m u m

Number o pulses

Figure 2 | New spines grow rapidly and acquire morphology similar to pre-existingspines. a , Left,time-lapse images of spineformationduring glutamateuncaging (40 pulses, 2Hz) showing the uncaging spot (yellow circle) andnascent spine (blue arrowhead). Right,fluorescence intensity profiles along theyellow line reveal that the spine head fluorescence increases gradually butrapidly (red arrowhead). b, Illustration of the measurement of spine headfluorescence during spinogenesis as a percentage of the maximum fluorescenceintensity reached. c, Time course of individual (black, 2 Hz; blue, 0.5Hz) and

average (red) fluorescence intensity increases during spinogenesis (n5 17).Errorbars, s.e.m. d, Averageof apparentspinelength,width andheadarea fromnascent (n5 95) and neighbouring existing (n5 111) spines (existing andnascent: length: 0.926 0.03mm, 0.8960.03mm, P . 0.1; width: 0.686 0.02mm, 0.706 0.02mm, P . 0.1; head area: 0.416 0.02 mm2, 0.386 0.03 mm2,P . 0.1). e, Cumulative distributions demonstrating that the morphologies of pre-existing and nascent spines are not different.

RESEARCH LETTER

2 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 1




spine growth (Fig. 3a), indicating that PKA activity is necessary but notsufficient for spinogenesis.

The small guanosine triphosphatase Ras is activated by Ca21 influx throughNMDARsandsignalsthrough mitogen-activated protein kinase(MAPK) to promote long-term potentiation17,22. We found that theMAPK pathway was necessary for spinogenesis because the success ratewas significantly reduced by blocking the upstream activator MAPKkinase 1/2 (MEK1/2) with U0126 or MEK1 with PD98059 (Fig. 3a).

In pyramidal neurons, activated Ras diffuses from active spines toneighbouring spines and heterosynaptically facilitates plasticity 17.To examine if a similar phenomenon potentiates spinogenesis, wedelivered a ‘priming’ stimulus to a pre-existing spine (40 pulses at2 Hz, Fig. 3e) and then a ‘test’ stimulus to the dendritic shaft within1–2min. In contrast to previous studies of older neurons in hip-pocampus16–18, wefound no consistent increasein volume ofthe exist-ing spine in response to this priming stimulus (Fig. 3f), supporting theidea that spinogenesis is not simply due to the growth of an undetect-able pre-existing spine. Nevertheless, the success rate of spine genera-tion was enhanced by the priming stimulus in a MEK1/2-dependentmanner such that the low level of spinogenesis evoked by a 0.5 Hz teststimulus was increased. In contrast, the priming stimulus did notincrease the success rate when the test stimulus was delivered at

2 Hz (Fig. 3g), indicating that the priming protocol shifted the induc-tion threshold for spinogenesis.Withthe exception of theNMDAR andPKA antagonists,none of the

pharmacological manipulations that altered spinogenesis rates affecteddendritic currents or Ca21 transients (Supplementary Fig. 8). Asexpected, CPP largely abolished the Ca21 transient and the prolongedphase of the currents. Consistent with a facilitation of Ca21 influx throughNMDARs by PKA23,24, H-89 lowered dendritic Ca21 transientsby approximately 20% (Supplementary Table 1). However, this effect isinsufficient to explain the near abolition of spinogenesis because therate of spinogenesis evoked by 0.5 ms uncaging pulses in control con-ditions was higher than that evoked by 4 ms pulses in H-89, despiteeliciting smaller dendritic Ca21 transients (Supplementary Fig. 8 andSupplementary Table 1).

We examined whether endogenous synaptic activity generates new spines in cortical tissue from young mice through similar pathways. Innormal extracellular Mg 21, high-frequency (100 pulses at 100 Hz,delivered twice, separated by 10 s) but not low-frequency (10Hz) elec-trical tetani rapidly triggerednew spine growth(Fig. 3h, i). Blockade of NMDARs or PKA prevented this synaptically evoked spine growth(Fig. 3i) as well as spontaneous new spine growth that occurred whenNMDAR activation was increased by removing extracellular Mg 21

and blocking inhibitory GABAA receptors (Supplementary Fig. 9).Hence, multiple experimental models demonstrate that activity-dependent spinogenesis in developing cortex requires NMDAR- andPKA-dependent signalling.

To determine if nascent spines detect synaptically released glutam-ate and are functionally incorporated into the circuit, we generatednew spines by high-frequency stimulation (HFS) and examined theirsynapticresponses using optical quantal analysisof synapticproperties.A whole-cell recording was obtained and the probability and amplitudeof synapticallyevoked NMDAR-mediated Ca21 transients in the spinehead were monitored (Fig. 4a). Using the Ca21-sensitive green fluor-ophore Fluo-5F, we detected stimulus-evoked Ca21 transients in theheadsof newly grown spines (Fig. 4b, c), demonstrating that they sensesynaptic activity within a neural circuit within 30 min after growth.Similar results were obtained in five of seven new spines and in 11 of 16 pre-existing spines. Analysis of the spines in which an evoked Ca21

transient could be detected indicated that nascent spines displayedsmaller and less frequent synaptically evoked Ca21 transients thanpre-existing spines (Fig. 4d, e). Similar results were obtained for new spines that grew in response to glutamate uncaging and were probedusing a glass-stimulating electrode placed near the spine (Supplemen-

tary Fig. 10).

e

C o n t r o l

P r i m i n g

S u c c e s s ( % )

0

50

100

db

N u m b e r o n e w s p i n e s

c

0 50 1000

50

100

S p e c i f c i t y ( % )

Success (%)

FSK

2–5 Hz

0.1–0.5 Hz

C

o n t r o l

F S K

0.0

0.5

1.0

1.5

2.0

*

1 μm

P r i m i n g

i n U 0 1

2 6

i

–8 –4 0 4 8

0

0.1

0.2

HFSHFS + CPPHFS + H-89LFS

N

u m b e r o n e w s p i n e s ( μ m – 1 )

Time (min)

* * *

New spine

1 μm

1 minater HFS

2 μm

–1 min

g

h

1 μm

H-89

FSK

0.5 Hz

*

2 Hz

Priming Test

Electrical

stimulation

a

C P P

M g 2 +

N B Q X

T T X

C P C C O

E t

M P E

P A I D A C P

A T G

K N - 6 2

K N - 9 3

K 2 5

2 a

F o r s k o

l i n H -

8 9

U 0 1

2 6

P D 9 8

0 5 9

0

20

40

60

80

100

(17)

(20)

(16)

(19)

(18)

(18)

*

*

*

* ** **

P e r c e n t a g e o s u c c e s s

–20

0

20

40

C h a n g e i n v o l u m e ( % )

Existingspine

f

(18)

(18)

(12)(12)

(8)

(27)

(16)(16)

(16)

(25)

Figure 3 | Molecular mechanisms of glutamate-induced spine formation.a , Spine formation using the 40 pulses at 2 Hz protocol was tested in thepresence of pharmacological agents. The dotted line indicates the successpercentage in control conditions with which statistical comparison was made.The number of induction attempts for each condition is given in parentheses,and the numbers of successes and totaltrials are summarized in Supplementary Table 2. TG, thapsigargin; CPA, cyclopiazonicacid. b, Examples of blockade of spinegeneration and of exuberant spinegrowth. Images takenbefore (left) andafter (right) glutamate uncaging in the presence of either H-89 or forskolin(FSK). Arrowheads, nascent spines. c, Summary of the average number of new spines with each induction attempt (control: 0.48 60.09, n5 33; FSK:1.126 0.21, n5 25, P , 0.005). d, Inverse relation between the locationspecificity and success rate. Data from three groups (0.1–0.5Hz, 2–5 Hz andFSK at 2Hz) are plotted. Specificity was measured as the percentage of cases in

which the spine arose within 1 mm of the uncaging spot. e, Representativeimages of priming experiments in which 40 pulses of 2 Hz glutamate uncaging were delivered to a pre-existing spine (yellow circle) followed by an additional40 pulses (0.5Hz or 2 Hz) delivered to the nearby dendrite (blue square).Releasing glutamate did not cause enlargement of the pre-existing spine head(left), but did trigger new spine growth from the dendrite (middle, right).f , Changes in the fluorescence of pre-existing spine heads exposed to thepriming stimulus (individual spines: circles; bar graph: average6 s.e.m.)(n5 31). g , Percentage of successful spine generation at the indicated testfrequencies with and without priming. U0126 prevented spinogenesisfacilitated by the priming protocol. h, The experiment (left) and images of adendrite 1 min before (middle) and after (right) HFS (23 100 pulses at100Hz). i, Average numbers of newspines generatedby HFSper micrometre of dendrite in control conditions (at 7 min: 0.1260.03, n5 11, P , 0.005), in thepresence of CPP (0, n5 8, P . 0.5), or H-89 (0.0046 0.001, n512; P . 0.5).The same number of pulses at a lower frequency (LFS, 10 Hz) generated fewer

new spines (0.0356 0.024, n5 7, P . 0.1). Error bars, s.e.m.

LETTER RESEARCH

0 0 M O N T H 2 0 1 1 | V O L 0 0 0 | N A T U R E | 3




The AMPA receptor (AMPAR) and NMDAR content of new spineswas characterized using glutamate uncaging after HFS-induced spino-genesis. In all cases, we detected fast uEPSCs at a holding potential of 260 mV and large prolonged uEPSCs at140 mV (Fig. 4g), consistentwith AMPAR- and NMDAR-mediatedcurrentspreviouslycharacterizedin these cells25. Similarly, uncaging-evoked Ca21 transients were clearly

visible at260mV (Fig. 4h) but were larger at140 mV (data notshown),consistent with the known properties of NMDAR-mediated Ca21 influx.Despite similar-sized AMPAR-uEPSCs in new and pre-existing spines,

NMDAR-uEPSCs and Ca21 influx were significantly lower in nascent

than pre-existing spines (Fig. 4g, h), similar to previous descriptions of spontaneously appearing new spines in hippocampal organotypicslices26. Therefore thesmallerCa21 transients measured in nascent spinesby synaptic activation probably reflect a smaller number of postsynapticNMDARs.

In this study, we established a protocol for the reliable and spatio-temporally precise induction of spinogenesis. These experimentsdemonstrate that glutamate is sufficient to trigger rapid spine forma-tion and suggest that neurons use glutamate release to establish circuit

wiring.Thus these data support the hypothesis that axonalgrowth and

1 μm

SpineCa2+

HFS

1 μm

Existing

Nascent D

S

D

S

100 ms

100 ms

100 ms

5 μm

Induction (HFS)

High-requencystimulation

Ca2+ imaging

Whole-cellrecording

Ca2+

indicator

Nascentspine

Whole-cell clamp

Linescan

0.0

0.2

0.4

0.6

0.8

1.0

S u c c e s s r a t e

*

*SpineCa2+

Dendritic Ca2+

Dendritic Ca2+

100 ms

D

S

100 ms

1 μm 5 μm

HFS

Spine Ca2+

Dendritic Ca2+

100 ms

2 pA

100 ms

3 pA

100 ms

5% ΔG / Gsat

A M P A R -

u E P S C

s ( p A )

0

2

4

N M D

A R -

u E P S C

s ( p A )

0

5

10

5

10

15

20

S p i n e C a 2 +

( % Δ G

/ G s a t )

0

5

10

15

20

+40 mV

–60 mV

gnitsixEtnecsaNgnitsixEtnecsaN

SpineCa2+

Dendrite Ca2+

5% ΔG / Gsat

* *

a b

edc

f

hg

D e n d r i t e C a 2 +

( % Δ G

/ G s a t )

10% ΔG / Gsat

10% ΔG / Gsat

0

20

40

60

S p i n e C a 2 +

( % Δ

G / G s a t )

N a s c

e n t

E x i s t i n

g

N a s c

e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

N a s

c e n t

E x i s t i n

g

Figure 4 | Functional characterization of new spines. a , Schematic of theexperimental procedure. HFS was delivered approximately 30 mm from thetarget dendritic region (red box). After nascent spines were identified, Ca21

indicator was loaded into the cell through a whole-cell recording pipette. Thenewly generated spine was examined at higher temporal and spatial resolutionto measure synaptically evoked Ca21 transients in the spine head and performoptical quantal analysis. b, Images before and after HFS showing the new spine(arrowhead) and the area subsequently analysed at higher resolution (yellow box). c, Images (left) of newlygenerated(top) and pre-existing (bottom) spinesfilled with the fluorophore Alexa 594 (red, 20 mM) and Fluo-5F (green,300mM). Fluorescence was collected (middle) and quantified (right) from aline-scan intersecting the spine (S) and dendrite (D) after electrical stimulation(arrowhead). The increases in green fluorescence indicate Ca21 entry. d, Greenfluorescence transients collected in consecutive trials (black) showing successesand failures. The average ‘success’ fluorescence transient is also shown (red).

e, Average amplitude (top, DG/Gsat) and rate (bottom) of success trials innascent and neighbouring existing spines for neurons heldat260 mV(nascentand existing: spine Ca21 DG/Gsat: 9.456 2.8%, n5 5, 25.56 4.4%, n5 11,P , 0.05; success rate: 0.266 0.09, n5 5, 0.586 0.09%, n5 11, P , 0.05).f , Similar experiments as those in a –c usingglutamateuncaging to characterizethe glutamate receptors on the nascent spine. g , Examples (top) and averageamplitudes (bottom) of AMPAR- and NMDAR-mediated uEPSCs at holding potentials of 260 and140 mV, respectively, using 1 ms uncaging pulses(nascent and existing: AMPAR-uEPSC: 21.961.4pA, n58, 22.560.5pA,n5 18, P . 0.05; NMDAR-uEPSC: 2.56 1.4pA, n5 8, 6.36 1.3pA, n5 18,P , 0.05). h, Average Ca21 transients measured in spine heads and dendritesfor neurons held at 260 mV (nascent and existing: spine Ca21 DG/Gsat:4.36 0.9%, n5 8, 12.36 1.8%, n5 18, P , 0.05; dendrite Ca21 DG/Gsat:1.16 0.1%, n5 8, 2.36 0.4%, n5 18, P . 0.05). Error bars, s.e.m.

RESEARCH LETTER

4 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 1




glutamate release may be the triggering event in synapse formationsuch that axonal bouton localization is an important early step forprecise neuronal circuit formation10,27. Given the involvement of NMDARs, Ca21 stores, cAMP, PKA and MAPK in activity-dependentspinogenesis, it is likely thatmany neuromodulators thatregulatethesemolecules may influence the capacity or threshold for new spineformation. For instance, activation of dopaminergic, serotonergic oradrenergic receptors that signal by Gas may facilitate spinogenesis,

whereas receptors that activate Gai-coupled signalling may functionas inhibitory signals.Lastly, we provide experimental evidence that spines that grow de

novo in developing cortical tissue becomerapidly functionally integratedinto the circuit such that they sense synaptically released glutamatethrough AMPARs and NMDARs. Whether these nascent spines arerapidly physically associated with a presynaptic bouton and display the ultrastructural correlates of a synapse is unknown26,28–30. Our resultsindicate that spines can grow de novo without the need for a filopodialintermediate and probably without a dendritic-shaft synapse stage. Intotal, this study demonstrates that synaptic activity can rapidly modify neuronal connectivity with high accuracy by generating new circuitelements.

METHODS SUMMARYAll procedures on animals followed protocols approved by the Harvard Standing Committee on Animal Care and National Institutes of Health guidelines. In utero

electroporation of EGFP was performed at embryonic day 15.5 in C57BL/6 mice.All studies were performed on layer 2/3 pyramidal neurons in acute coronal slicesidentified by their characteristic morphology, position in the slice and expres-sion of GFP. Two-photon glutamate uncaging and imaging was performed using custom microscopes. To induce spinegrowth,40 uncagingpulses weredeliveredat

varying frequencies to a spot approximately 0.5 mm from the edge of the dendrite.Synaptically induced spine growth was triggered with two high-frequency stimuli(100 pulsesat 100Hz) separatedby 10s delivered by a bipolar electrodepositionedapproximately 30mm from the target dendrite. For Ca21 imaging, neurons wereloaded through the whole-cell recording electrode with Alexa Fluor-594 (20 mM)and Fluo-5F (300 mM) and the amplitudes of fluorescence transients were quan-tified as a fraction of the maximal green fluorescence achieved in saturating (sat)Ca21 concentrations (Gsat). For optical quantal analysis, the synapse associated

with a visualized spine was stimulated using a closely positioned glass electrode.The position of the electrode and stimulus intensity were adjusted until (1) Ca21

transients were evokedin thespinehead that demonstratedstochastic failures andsuccesses, and (2) Ca21 transients in other spines and the dendritic shaft in thefield ofviewwerenot evoked. Fisher’s exacttest wasusedto compare theefficacy of spinogenesis across conditions. For each spinogenesis trial, an observer blind tothe experimental conditionwas asked to identify if (1)a newspinehad grown and(2), if so, how many spines had grown.

Received 25 October 2010; accepted 4 March 2011.

Published online 8 May 2011.

1. Grutzendler, J.,Kasthuri,N. & Gan, W. B. Long-termdendritic spine stability in theadult cortex. Nature 420, 812–816 (2002).

2. Hofer, S. B., Mrsic-Flogel, T. D., Bonhoeffer, T. & Hubener, M. Experience leaves alasting structural trace in cortical circuits. Nature 457, 313–317 (2009).

3. Trachtenberg, J. T. et al. Long-term in vivo imaging of experience-dependent

synaptic plasticity in adult cortex. Nature 420, 788–794 (2002).4. Zuo,Y., Yang, G., Kwon,E. & Gan, W. B. Long-term sensory deprivation prevents

dendritic spine loss in primary somatosensory cortex. Nature 436, 261–265(2005).

5. Rakic, P., Bourgeois, J. P., Eckenhoff, M. F., Zecevic, N. & Goldman-Rakic, P. S.Concurrent overproduction of synapses in diverse regions of theprimatecerebralcortex. Science 232, 232–235 (1986).

6. Dalva, M. B., McClelland, A. C. & Kayser, M. S. Cell adhesion molecules: signallingfunctions at the synapse. Nature Rev. Neurosci. 8, 206–220 (2007).

7. Scheiffele, P. Cell-cell signaling during synapse formation inthe CNS. Annu. Rev.Neurosci. 26, 485–508 (2003).

8. Sudhof, T.C. Neuroliginsand neurexinslink synapticfunctionto cognitive disease.Nature 455, 903–911 (2008).

9. Yuste,R. & Bonhoeffer,T. Genesisof dendritic spines:insights fromultrastructuraland imaging studies. Nature Rev. Neurosci. 5, 24–34 (2004).

10. Miller, M. & Peters, A. Maturation of rat visual cortex. II. A combinedGolgi-electronmicroscope study of pyramidal neurons. J. Comp. Neurol. 203, 555–573 (1981).

11. Richards, D. A. et al. Glutamate induces the rapid formation of spine headprotrusions in hippocampal slice cultures. Proc. Natl Acad. Sci. USA 102,6166–6171 (2005).

12. Fiala, J. C., Feinberg, M., Popov, V. & Harris, K. M. Synaptogenesis via dendriticfilopodia in developing hippocampal area CA1. J. Neurosci. 18, 8900–8911(1998).

13. Sorra, K. E., Fiala, J. C. & Harris, K. M. Critical assessment of the involvement ofperforations, spinules, andspine branching in hippocampal synapseformation. J.Comp. Neurol. 398, 225–240 (1998).

14. Engert, F. & Bonhoeffer, T. Dendritic spine changes associated with hippocampallong-term synaptic plasticity. Nature 399, 66–70 (1999).

15. Maletic-Savatic, M., Malinow, R. & Svoboda, K. Rapid dendritic morphogenesis inCA1 hippocampal dendrites induced by synaptic activity. Science 283,1923–1927 (1999).

16. Harvey, C. D. & Svoboda, K. Locally dynamic synaptic learning rules in pyramidalneuron dendrites. Nature 450, 1195–1200 (2007).

17. Harvey, C. D., Yasuda, R., Zhong, H. & Svoboda, K. The spread of Ras activitytriggered by activation of a single dendritic spine. Science 321, 136–140 (2008).

18. Matsuzaki, M., Honkura, N., Ellis-Davies, G. C. & Kasai, H. Structural basis of long-term potentiation in single dendritic spines. Nature 429, 761–766 (2004).

19. Tanaka, J. et al. Protein synthesis and neurotrophin-dependent structural

plasticity of single dendritic spines. Science 319, 1683–1687 (2008).20. Yasuda,H., Barth, A. L.,Stellwagen,D. & Malenka,R. C. A developmentalswitch in

the signaling cascades for LTP induction. Nature Neurosci. 6, 15–16 (2003).21. Zaccolo, M. & Pozzan, T. Discretemicrodomainswith highconcentration of cAMP

in stimulated rat neonatal cardiac myocytes. Science 295, 1711–1715 (2002).22. Di Cristo, G. et al. Requirement of ERK activation for visual cortical plasticity.

Science 292, 2337–2340 (2001).23. Chalifoux,J. R. & Carter, A. G.GABAB receptorsmodulate NMDAreceptor calcium

signals in dendritic spines. Neuron 66, 101–113 (2010).24. Skeberdis, V. A. et al. Protein kinase A regulates calcium permeability of NMDA

receptors. Nature Neurosci. 9, 501–510 (2006).25. Busetto, G., Higley, M. J. & Sabatini, B. L. Developmental presence and

disappearance of postsynapticallysilentsynapses on dendritic spines of rat layer2/3 pyramidal neurons. J. Physiol. (Lond.) 586, 1519–1527 (2008).

26. Zito,K., Scheuss, V., Knott, G.,Hill, T. & Svoboda, K. Rapidfunctional maturation ofnascent dendritic spines. Neuron 61, 247–258 (2009).

27. Hamori, J. The inductive role of presynaptic axons in the development ofpostsynaptic spines. Brain Res. 62, 337–344 (1973).

28. De Roo, M., Klauser, P., Mendez, P., Poglia, L. & Muller, D. Activity-dependent PSDformation and stabilizationof newly formed spines in hippocampal slice cultures.Cereb. Cortex 18, 151–161 (2008).

29. Nagerl, U. V.,Kostinger,G., Anderson, J. C.,Martin,K. A. & Bonhoeffer,T. Protractedsynaptogenesis after activity-dependentspinogenesis in hippocampal neurons. J.Neurosci. 27, 8149–8156 (2007).

30. Knott, G. W., Holtmaat, A., Wilbrecht, L., Welker, E. & Svoboda, K. Spine growthprecedes synapse formation in the adult neocortex in vivo. Nature Neurosci. 9,1117–1124 (2006).

Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.

AcknowledgementsWe thank membersof theSabatini laboratory fortheir commentson the manuscript and assistance with data analysis. We are grateful to S. Nazia fortechnical support and for acting as the blind evaluator. This work was supported by aSFARI grantfrom the Simons Foundation and the National Institute of NeurologicalDisorders and Stroke (NS046579).

Author ContributionsH.B.K. andB.L.S.designedthe experimentsand wrote thepaper.H.B.K.performed allthe experiments,analysedthe data(otherthanspinecountingby ablind, third-party observer) and prepared the figures.

Author Information Reprints and permissions information is available atwww.nature.com/reprints . The authors declare no competing financial interests.Readers are welcome to comment on the online version of this article atwww.nature.com/nature. Correspondence and requests for materials should beaddressed to B.L.S. ([email protected]) .

LETTER RESEARCH

0 0 M O N T H 2 0 1 1 | V O L 0 0 0 | N A T U R E | 5

M ill P bli h Li it d All i ht d©2011

http://www.nature.com/nature

http://www.nature.com/reprints


mailto:[email protected]

mailto:[email protected]


http://www.nature.com/reprints


glutamate induces de novo growth of functional spine in developing cortex

Documents