LETTERS

The PtdIns(3,4)P2 phosphatase INPP4A is asuppressor of excitotoxic neuronal deathJunko Sasaki1,2, Satoshi Kofuji1,2, Reietsu Itoh1,2, Toshihiko Momiyama3, Kiyohiko Takayama2,4,Haruka Murakami1,2, Shinsuke Chida1,2, Yuko Tsuya1,2, Shunsuke Takasuga1,2, Satoshi Eguchi1,2, Ken Asanuma1,2,Yasuo Horie5, Kouichi Miura5, Elizabeth Michele Davies6, Christina Mitchell6, Masakazu Yamazaki2,Hirokazu Hirai2,4, Tadaomi Takenawa7,8, Akira Suzuki2,9 & Takehiko Sasaki1,2

Phosphorylated derivatives of phosphatidylinositol, collectivelyreferred to as phosphoinositides, occur in the cytoplasmic leafletof cellular membranes and regulate activities such as vesicle trans-port, cytoskeletal reorganization and signal transduction1,2. Recentstudies have indicated an important role for phosphoinositidemetabolism in the aetiology of diseases such as cancer, diabetes,myopathy and inflammation3–5. Although the biological func-tions of the phosphatases that regulate phosphatidylinositol-3,4,5-trisphosphate (PtdIns(3,4,5)P3) have been well characterized, littleis known about the functions of the phosphatases regulating theclosely related molecule phosphatidylinositol-3,4-bisphosphate(PtdIns(3,4)P2). Here we show that inositol polyphosphate phos-phatase 4A (INPP4A), a PtdIns(3,4)P2 phosphatase, is a suppressorof glutamate excitotoxicity in the central nervous system. Targeteddisruption of the Inpp4a gene in mice leads to neurodegenerationin the striatum, the input nucleus of the basal ganglia that has acentral role in motor and cognitive behaviours. Notably, Inpp4a2/2

mice show severe involuntary movement disorders. In vitro, Inpp4agene silencing via short hairpin RNA renders cultured primarystriatal neurons vulnerable to cell death mediated by N-methyl-D-aspartate-type glutamate receptors (NMDARs). Mechanistically,INPP4A is found at the postsynaptic density and regulates synapticNMDAR localization and NMDAR-mediated excitatory post-synaptic current. Thus, INPP4A protects neurons from excitotoxiccell death and thereby maintains the functional integrity of thebrain. Our study demonstrates that PtdIns(3,4)P2, PtdIns(3,4,5)P3

and the phosphatases acting on them can have distinct regulatoryroles, and provides insight into the unique aspects and physiologicalsignificance of PtdIns(3,4)P2 metabolism. INPP4A represents, to ourknowledge, the first signalling protein with a function in neuronsto suppress excitotoxic cell death. The discovery of a direct linkbetween PtdIns(3,4)P2 metabolism and the regulation of neuro-degeneration and involuntary movements may aid the developmentof new approaches for the treatment of neurodegenerative disorders.

PtdIns(3,4)P2 and PtdIns(3,4,5)P3 are phosphoinositides that arethought to be highly similar in terms of their scarce abundance andproposed biological functions6,7. Production of PtdIns(3,4)P2 andPtdIns(3,4,5)P3 depends on phosphoinositide 3-kinases (PI(3)Ks)that drive an extensive signalling network downstream of cell surfacereceptor activation8. Upon cellular stimulation, levels of PtdIns(3,4)P2

and PtdIns(3,4,5)P3 undergo an enormous, but transient, increase6.The subsequent degradation of these phosphoinositides is carried out

1Department of Medical Biology, Akita University Graduate School of Medicine, Akita 010-8543, Japan. 2Global COE Program, Gunma University and Akita University, Gunma 371-8511, Japan. 3Department of Pharmacology, Jikei University School of Medicine, Tokyo 105-8461, Japan. 4Department of Neurophysiology, Gunma University Graduate School ofMedicine, Gunma 371-8511, Japan. 5Department of Gastroenterology and Neurology, Akita University Graduate School of Medicine, Akita 010-8543, Japan. 6Department ofBiochemistry and Molecular Biology, Monash University, Clayton, Victoria 3800, Australia. 7Division of Lipid Biochemistry, Department of Biochemistry and Molecular Biology, KobeUniversity Graduate School of Medicine, Kobe 650-0017, Japan. 8Global COE Program, Kobe University Graduate School of Medicine, Kobe 650-0017, Japan. 9Division of Embryonicand Genetic Engineering, Medical Institute of Bioregulation, Kyushu University, Fukuoka 812-8582, Japan.

f

dc

e

ba

Time (days after birth)Time (days after birth)

0

2

4

6

8

10

7 14 21

Bod

y w

eigh

t (g

)

WTKO

0

20

40

60

80

100

0 10 20 30

Sur

viva

l (%

)

WTKO

0

10

20

30

40C

onvu

lsio

n (s

)

WT KOMK801 – +– +

****

0

1

2

3

4

5

6

7

Hin

dlim

b c

lasp

ing

(s)

WT KOMK801 – +– +

Figure 1 | Mortality and involuntary movements of Inpp4a2/2 mice. a,Reduced body weight. Wild-type and Inpp4a2/2 (KO) littermates wereweighed starting at day 7 after birth. Results shown are the mean bodyweight 6 s.e.m. (n 5 8 mice per genotype). b, Survival of wild-type andInpp4a2/2 mice (n 5 12 per genotype). c–f, Involuntary movements ofInpp4a2/2 mice involve NMDAR. All Inpp4a2/2 mice showed limbhyperkinesias, opisthotonos and dystonia starting at P14 and thereafter. (Seealso Supplementary Movies 1–5.) c, d, Inpp4a2/2 mice showing rigidextension of the limbs (c) and opisthotonos (d). e, f, Blockade of NMDAR byMK801 suppressed hindlimb clasping (e) and limb and trunk hyperkinesias(f) in Inpp4a2/2 mice. Results shown for e are the mean hindlimb claspingscores 6 s.e.m. (n 5 5 per genotype). Data shown for f are the mean durationof convulsion 6 s.e.m. observed during a 3-min period (n 5 5 per genotype).**P , 0.01 in e and f.

Vol 465 | 27 May 2010 | doi:10.1038/nature09023

497Macmillan Publishers Limited. All rights reserved©2010

by phosphatases, and seems to be important for suppressing thehyperactivation of molecular events downstream of PI(3)Ks. In con-trast to the wealth of knowledge about PTEN, the tumour suppressorgene product that specifically dephosphorylates PtdIns(3,4,5)P3

9, therole of the INPP4A enzyme that specifically dephosphorylatesPtdIns(3,4)P2

10 remains unclear.To investigate the physiological function of INPP4A, we generated

mice with a targeted disruption of Inpp4a (Supplementary Fig. 1a–c).Inpp4a2/2 mice were born at the expected Mendelian ratio and wereindistinguishable from wild-type and Inpp4a1/2 littermates at birth.However, by postnatal day 14 (P14), the growth rate of the homo-zygotes had slowed significantly and all mutants had died by thefourth week (Fig. 1a, b).

Whereas control mice initiated gait at about P12, Inpp4a2/2 micewere unable to walk throughout their short lives. In addition, themutants showed severe involuntary movements starting at aboutP14. Limb hyperkinesias such as ballism and chorea (Fig. 1c andSupplementary Movies 1 and 2), tonic contractions of the neckand trunk leading to upward deviation of the head and lumbar spineregion (opisthotonos; Fig. 1d), and hindlimb clasping (Fig. 1e andSupplementary Fig. 1d) all occurred in Inpp4a2/2 mice with 100%penetrance. Dystonic postures with forelimbs and hindlimbsextended were also typical (Supplementary Fig. 1e). It seemed thatthe involuntary movements frequently emerged when the Inpp4a2/2

mice attempted to perform voluntary movements (for example,walk), and the former prevented the latter. In addition, Inpp4a2/2

mice displayed convulsions every time they were placed in a newenvironment (Fig. 1f and Supplementary Movie 2). By 4 weeks ofage, most Inpp4a2/2 mice had progressed to rigidity and akinesia.Administration of MK801, an antagonist of the NMDA glutamatereceptor (NMDAR), ameliorated the hindlimb clasping and con-vulsions of Inpp4a2/2 mice (Fig. 1e, f and Supplementary Movies 3and 4) at a concentration where this agent had no significant effect on

wild-type mouse movements (not shown). Thus, INPP4A is essentialfor normal movement control, and loss of INPP4A leads to involu-ntary movements that are associated with an enhanced response ofneurons to glutamate via NMDAR. Indeed, the phenotype ofInpp4a2/2 mice includes several aspects that resemble the involu-ntary movements of human disorders such as Huntington’s diseaseand dystonia11.

Involuntary movement disorders are often related to the degenera-tion of basal ganglia11,12. Histological examination of Inpp4a2/2

brains revealed lesions specifically in the striatum, which receivesabundant glutamatergic projections from the cerebral cortex. Nisslstaining detected the presence of degenerating neurons that werereduced in size and hyperchromic (Fig. 2a, b), and TdT-mediateddUTP nick end labelling (TUNEL) staining revealed large numbersof dying cells in mutant (but not wild type) striatum (Fig. 2c, d).Importantly, neurodegeneration in Inpp4a2/2 striatum was rescuedby lentiviral-vector-mediated expression of recombinant INPP4A(Fig. 2e, f). Polymerase chain reaction with reverse transcription(RT–PCR) analyses and immunostaining of striatal tissues to detectMet-enkephalin showed that medium-sized spiny projection neurons(MSNs) were the cells most vulnerable to INPP4A deficiency(Supplementary Fig. 2). Neuronal damage frequently leads to reactivegliosis, a feature that was prominent specifically in Inpp4a2/2 striata(Fig. 2g) and markedly ameliorated by INPP4A reconstitution (Sup-plementary Fig. 3). These data demonstrate that INPP4A is essentialfor the structural and functional integrity of the striatum.

Phosphatase activity towards PtdIns(3,4)P2 was decreased inInpp4a2/2 striatum compared to the wild type (SupplementaryFig. 4a). Consistent with this observation, cellular PtdIns(3,4)P2 wassignificantly increased in mutant striatum (Supplementary Fig. 4b).To determine the importance of PtdIns(3,4)P2 phosphatase activity inmaintaining normal striatum, we took advantage of the natural mousemutant weeble13, which has a frameshift mutation (Inpp4a D744G)

a

KO

WT

c

KO

WT Sham

Virus

INPP4A/GFP TUNEL

e

g

3040506070

01020

WT KO WT KO WT KO

Cel

ls o

f red

uced

siz

e

**b

0

50

100

150

200 **

Inta

ct c

ells

0

10

20

30

40

**d f

Sham Virusinjection

**

0

5

10

15

20

25

TUN

EL+

cel

ls

TUN

EL+

cel

ls

KO

WT

Figure 2 | Neurodegeneration in Inpp4a2/2 striatum. a, Histology. Low(left) and high (right) magnification views of Nissl-stained sections fromwild-type (WT; top) and Inpp4a2/2 (KO; bottom) mice. The mutant showsdecreased striatal staining and dark cells of reduced size. Results shown arerepresentative of 5 mice examined per genotype. b, Striatal cell counts.Numbers of dark neurons of reduced size (left) and intact neurons (right)were counted in ten randomly selected microscope fields from mid-striatalsections of 20-day-old wild-type and Inpp4a2/2 littermates (n 5 2 pergenotype). Data are expressed as mean 6 s.e.m. per field. c, IncreasedTUNEL staining. The striata of wild-type and Inpp4a2/2 mice weresubjected to TUNEL staining to detect neuronal death. d, Quantification of

the TUNEL1 cells in c. Results shown are the mean 6 s.e.m. per field (n 5 3mice per genotype). e, Prevention of cell death by introduction of Inpp4acDNA. Inpp4a2/2 mice at P2 were injected in the striatum with (Virus; n 5 4mice) or without (Sham; n 5 2 mice) a lentiviral vector expressing GFP-P2A-INPP4A. Striata were isolated at P20 and cell death was detected by TUNELstaining. f, Quantification of the TUNEL1 cells in e. Results shown are themean 6 s.e.m. per field (5 fields per mouse). g, Striatal-selective gliosis. Mid-striatal coronal sections from wild-type and Inpp4a2/2 (KO) mice at P30were subjected to GFAP immunostaining. Scale bars: a, 100mm (left), 20 mm(right); c, e, 20 mm; g, 2 mm. **P , 0.01 in b, d, f.

LETTERS NATURE | Vol 465 | 27 May 2010

498Macmillan Publishers Limited. All rights reserved©2010

that expresses a messenger RNA encoding a truncated INPP4Aprotein (amino acids 1–263) lacking the phosphatase motif (Sup-plementary Fig. 4c, d). We created an Inpp4a2/D744G trans-compoundheterozygote that bears an Inpp4a null allele and a phosphatase-deficient Inpp4aD744G allele. Inpp4a2/D744G mice displayed the sameinvoluntary movements and neurodegeneration as the geneticallyengineered Inpp4a2/2 mice (Supplementary Movie 5 and Sup-plementary Fig. 4e), confirming the crucial nature of INPP4A’sPtdIns(3,4)P2 phosphatase activity. This neuroprotective role seemsto be unique to the INPP4A isozyme because loss of the otherPtdIns(3,4)P2 phosphatase INPP4B did not cause any neurologicaldefects (data not shown).

When we examined isolated striata by electron microscopy, a sig-nificant increase in electron-dense and vacuolated MSNs wasobserved in Inpp4a2/2 mice (Fig. 3a and Supplementary Fig. 5).This phenomenon is called ‘dark cell degeneration’, a process ofcellular demise linked to excitotoxicity induced by excessive activa-tion of glutamate receptors14,15. In humans, excitotoxicity is a majormechanism of neuronal loss in cases of brain ischaemia and chronicneurodegenerative disease16,17. Thus, understanding how neuronscan be protected from excitotoxic death might lead to significantclinical benefits. We speculated that the molecular mechanismunderlying INPP4A’s neuroprotective function in vivo might be thesuppression of excitotoxicity. To explore this hypothesis, we firstmeasured extracellular glutamate concentrations in the striatum ofour mutant mice by microdialysis. Extracellular glutamate levels werenot significantly different between wild-type and Inpp4a2/2 striata(Supplementary Fig. 6a). In addition, neither glutamate uptake norexpression levels of glutamate transporters were abnormal inInpp4a2/2 mice (Supplementary Fig. 6b, c), indicating that loss ofINPP4A does not affect glutamate homeostasis.

We next investigated whether INPP4A regulates neuronal survivalin a cell-autonomous manner. Cultured MSNs transfected with Inpp4ashort hairpin RNA (shRNA) were significantly more susceptible to

glutamate-induced excitotoxicity than were MSNs transfected withcontrol shRNA (Fig. 3b and Supplementary Fig. 7a, b). Notably, sup-pression of INPP4A markedly potentiated the toxic effect of glutamateat a concentration (50mM) that had no toxic effect on control cells(Fig. 3c). Furthermore, treatment of MSNs with PtdIns(3,4)P2 (but notPtdIns(3,4,5)P3) significantly potentiated glutamate-induced cell death(Fig. 3d, e). The glutamate excitotoxicity enhanced by Inpp4a shRNAand PtdIns(3,4)P2 was inhibited in both cases by MK801 (Fig. 3c, e).These data demonstrate that degradation of PtdIns(3,4)P2 by INPP4Ahas a key role in preventing NMDAR-mediated excitotoxicity.

The increased sensitivity of Inpp4a2/2 MSNs to NMDAR-mediated excitotoxicity prompted us to examine whether an absenceof INPP4A could lead to a change in the dynamics of NMDARdistribution. Purification of the postsynaptic density (PSD) fractionof wild-type striatal extract followed by immunoblotting showed thatINPP4A was indeed present at the PSD, together with NMDARs andthe synaptic scaffold protein PSD95 (Fig. 4a, left). Notably, althoughthere were no differences between wild-type and Inpp4a2/2 striata intotal expression of the NMDAR subunits NR1 and NR2B (data notshown), loss of INPP4A expression led to increases in NR1 and NR2Bat the PSD (Fig. 4a, right). It has previously been shown that, at thePSD, tyrosine phosphorylation of NR2B decreases its associationwith the clathrin adaptor AP2 and blocks internalization of the entireNMDAR; NMDAR surface expression is thereby promoted18.Consistent with the increased level of NR2B in the Inpp4a2/2 PSDfraction, NR2B tyrosine phosphorylation was enhanced in Inpp4a2/2

striatum (Fig. 4b). NR2C, an Akt substrate that promotes survival ofcerebellar granule cells19, was not expressed in either wild-type orInpp4a2/2 striatum (Supplementary Fig. 8).

The upregulation of synaptic NMDAR in Inpp4a2/2 striata wasalso evident in a series of electrophysiological recordings. Inpp4a2/2

synapses had significantly larger NMDAR-mediated excitatory post-synaptic current (EPSC)/a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR)-mediated EPSC ratios than did

0

20

40

60

80

0

20

40

60

80

100

120 **

**

Control shRNA

Inpp4a shRNA

**

Control

Glutam

ate

Glutam

ate+P

tdIns(3,4)P

2

Glutam

ate+P

tdIns(3,4,5)P

3

Glutam

ate+P

tdIns(4,5)P

2

Glutam

ate+P

tdIns(3,4)P

2+

MK

801

**Glutamate +PtdIns(3,4)P2

Glutamate +PtdIns(3,4,5)P3

Glutamate +PtdIns(4,5)P2

Glutamate

Control shRNA Inpp4a shRNA

GFP

MAP2MAP2

Control shRNA Inpp4a shRNA

GFP

WT

KO

a b

d

c

e

GFP

+M

AP

2+ (%

of c

ontr

ol)

MA

P2+

cel

ls p

er fi

eld

Glutamate: 0 50 µM 400 µM+ MK801

400 µM

Figure 3 | INPP4A suppresses glutamate excitotoxicity. a, Dark celldegeneration. Degenerating neurons in Inpp4a2/2 striata (bottom) can beidentified by their enhanced affinity for osmium and are rare in wild-typestriata (top). b, c, INPP4A downregulation enhances excitotoxicity. Pooledcultures of primary MSNs were co-transfected with a GFP expression vectorplus either control shRNA or Inpp4a shRNA. At 7 d in vitro, MSNs wereeither left untreated or treated for 16 h with 50 mM glutamate. b, MSNs werestained with anti-GFP (green) and anti-MAP2 (red). c, Quantification ofGFP1MAP21 neurons. The mean GFP1MAP21 neuron count in untreated(0 glutamate) cultures of control shRNA and Inpp4a shRNA transfectants

was arbitrarily assigned a value of 100. Data shown are GFP1MAP21 neuronnumbers expressed as the mean percentage 6 s.e.m. of the control for eachshRNA transfectant (triplicate samples). d, e, PtdIns(3,4)P2 potentiatesNMDAR-mediated excitotoxicity. MSNs (12 days in vitro) were incubatedwith the indicated phosphoinositides for 30 min and then were left untreatedor treated for 16 h with 10mM glutamate. d, MSN excitotoxicity was assessedby staining with anti-MAP2 (green). e, Quantification of MAP21 MSNs. Themean number 6 s.e.m. of MAP21 cells per field in 5–10 fields per treatmentis shown. Scale bars: a, 10 mm; b, d, 30 mm. **P , 0.01 in c and e.

NATURE | Vol 465 | 27 May 2010 LETTERS

499Macmillan Publishers Limited. All rights reserved©2010

wild-type synapses (Fig. 4c). In addition, current-clamp recordingsrevealed that the resting membrane potential of MSNs was signifi-cantly more depolarized in Inpp4a2/2 mice (265.2 6 0.86 mV;n 5 5; P , 0.05) than in wild-type mice (275.0 6 3.34 mV, n 5 4;data not shown). Such an effect has previously been attributed toalterations in the composition of synaptic glutamate receptors20.Taken together, these data indicate that Inpp4a2/2 synapses have ahigher density of NMDARs than do wild-type synapses. We proposethat, at least in MSNs, INPP4A suppresses excitotoxicity by down-regulating NMDARs and thus decreasing cellular sensitivity to glu-tamate stimulation. The molecular basis of this striatal-selectiveneurodegeneration in Inpp4a2/2 mice is unclear but may be due,in part, to differential composition or expression of NMDARs indifferent neuronal populations.

To our knowledge, INPP4A represents the first neuronal signallingprotein to suppress excitotoxic cell death and involuntary move-ments. The early onset and severity of the spontaneous neurologicaldisorders of Inpp4a2/2 mice are of particular interest because mostother genetic models of excitotoxicity depend on overexpression ofdisease-associated mutant proteins. In contrast, Inpp4a2/2 mice donot express any known disease-specific, gain-of-function mutantproteins; the sole defect in these animals is the loss of the phosphataserequired to control levels of the second messenger PtdIns(3,4)P2.

In view of the pro-apoptotic properties of PTEN21, the pro-survivalfunction of INPP4A is intriguing. PtdIns(3,4)P2 and PtdIns(3,4,5)P3

are both lipid products of PI(3)K and have long been thought to over-lap in function. However, in contrast to Inpp4a2/2 mice, mice withbrain-specific PTEN deficiency exhibit neither striatal neurodegenera-tion in vivo nor enhanced neuronal excitotoxicity in vitro22. Thus, ourconclusion is that PtdIns(3,4)P2, PtdIns(3,4,5)P3 and the phosphatasesacting on them can have distinct roles. Intervention in the PI(3)K/PtdIns(3,4,5)P3 signalling pathway is already under investigation as atreatment for cancers and inflammatory diseases23,24. Our results high-light the significance of PtdIns(3,4)P2 metabolism and point towardspotential novel therapies for acute and chronic neurological disordersin which excitotoxicity triggers neurodegeneration25–28.

METHODS SUMMARYInpp4a2/2 mice were generated by standard gene targeting methods in which the

Inpp4a gene was disrupted by homologous recombination in mouse embryonic

stem cells. All experimental protocols were reviewed and approved by the Akita

University Institutional Committee for Animal Studies. Dyskinesia assessments,

histological and immunohistochemical analyses, construction and intrastriatal

injection of lentiviral vector expressing INPP4A, PtdIns(3,4)P2 measurement,

electron microscopy, RT–PCR, in vivo microdialysis, glutamate uptake, striatal

cell culture, construction of short hairpin RNA-expressing plasmids, excitotoxi-

city in cultured striatal neurons, isolation of PSD fractions, immunoblotting,

electrophysiology, and occlusion of common carotid arteries were performed in

accordance with standard or published protocols. Statistical analyses were per-

formed using the Student’s t-test.

Full Methods and any associated references are available in the online version ofthe paper at www.nature.com/nature.

Received 2 July 2009; accepted 12 March 2010.Published online 12 May 2010.

1. Takenawa, T. & Itoh, T. Phosphoinositides, key molecules for regulation of actincytoskeletal organization and membrane traffic from the plasma membrane.Biochim. Biophys. Acta 1533, 190–206 (2001).

2. Di Paolo, G. & De Camilli, P. Phosphoinositides in cell regulation and membranedynamics. Nature 443, 651–657 (2006).

3. Sasaki, T. et al. Mammalian phosphoinositide kinases and phosphatases. J. LipidRes. 48, 307–343 (2009).

4. Kok, K., Geering, B. & Vanhaesebroeck, B. Regulation of phosphoinositide 3-kinaseexpression in health and disease. Trends Biochem. Sci. 34, 115–127 (2009).

5. Wymann, M. P. & Schneiter, R. Lipid signalling in disease. Nature Rev. Mol. Cell Biol.9, 162–176 (2008).

6. Stephens, L. R., Jackson, T. R. & Hawkins, P. T. Agonist-stimulated synthesis ofphosphatidylinositol(3,4,5)-trisphosphate: a new intracellular signalling system?Biochim. Biophys. Acta 1179, 27–75 (1993).

7. Vanhaesebroeck, B. et al. Synthesis and function of 3-phosphorylated inositollipids. Annu. Rev. Biochem. 70, 535–602 (2001).

8. Cantley, L. C. The phosphoinositide 3-kinase pathway. Science 296, 1655–1657(2002).

9. Suzuki, A., Nakano, T., Mak, T. W. & Sasaki, T. Portrait of PTEN: messages frommutant mice. Cancer Sci. 99, 209–213 (2008).

10. Norris, F. A., Auethavekiat, V. & Majerus, P. W. The isolation and characterizationof cDNA encoding human and rat brain inositol polyphosphate 4-phosphatase.J. Biol. Chem. 270, 16128–16133 (1995).

11. Walker, F. O. Huntington’s disease. Lancet 369, 218–228 (2007).12. Albin, R. L., Young, A. B. & Penney, J. B. The functional anatomy of basal ganglia

disorders. Trends Neurosci. 12, 366–375 (1989).13. Nystuen, A., Legare, M. E., Shultz, L. D. & Frankel, W. N. A null mutation in inositol

polyphosphate 4-phosphatase type I causes selective neuronal loss in weeblemutant mice. Neuron 32, 203–212 (2001).

14. Turmaine, M. et al. Nonapoptotic neurodegeneration in a transgenic mouse modelof Huntington’s disease. Proc. Natl Acad. Sci. USA 97, 8093–8097 (2000).

15. Portera-Cailliau, C., Price, D. L. & Martin, L. J. Non-NMDA and NMDA receptor-mediated excitotoxic neuronal deaths in adult brain are morphologically distinct:further evidence for an apoptosis-necrosis continuum. J. Comp. Neurol. 378,88–104 (1997).

16. Martin, L. J. et al. Neurodegeneration in excitotoxicity, global cerebral ischemia,and target deprivation: A perspective on the contributions of apoptosis andnecrosis. Brain Res. Bull. 46, 281–309 (1998).

17. Estrada Sanchez, A. M., Mejia-Toiber, J. & Massieu, L. Excitotoxic neuronal deathand the pathogenesis of Huntington’s disease. Arch. Med. Res. 39, 265–276(2008).

18. Zhang, S., Edelmann, L., Liu, J., Crandall, J. E. & Morabito, M. A. Cdk5 regulates thephosphorylation of tyrosine 1472 NR2B and the surface expression of NMDAreceptors. J. Neurosci. 28, 415–424 (2008).

PSD95 NR2B GluR2NR10

2

1

3

WT

KO

Pro

tein

leve

l in

PS

D fr

actio

n(a

rbitr

ary

unit)

a

**

*

WT WT KOKO

INPP4A

Synaptophysin

PSD95

NR2B

NR1

GluR2

PSDSup

b

0

NM

DA

/AM

PA r

atio

0.4

0.8

1.2

1.6

2.0

2.4 *c

NR2B

p-NR2B

WT KO

WT KO

WT

50 pA

WTx

x

y y50 pA

z

KO

KO

50 pA

z 50 ms50 pA

50 ms

50 ms50 ms

Figure 4 | Increased NMDARs at the synapses in the absence of INPP4A.a, Increased NR1 and NR2B at the PSD in the absence of INPP4A. PSD and non-PSD (Sup) fractions were prepared from striata of wild-type and Inpp4a2/2

(KO) mice. Left: western blot to detect the indicated proteins. Right:densitometric quantification of the indicated proteins in the PSD fraction (n 5 8mice per genotype). b, Increased tyrosine phosphorylation. Total striatal proteinlysates from 4 wild-type and 4 Inpp4a2/2 mice were successivelyimmunoblotted to detect phosphorylated and total NR2B. c, IncreasedNMDAR EPSC/AMPAR EPSC ratio. Synaptic currents were evoked by focalstimulation at 140 mV in the presence of c-aminobutyric acid subtype A(GABAA) and glycine receptor blockers. Left: for both wild-type and Inpp4a2/2

(KO) MSNs, traces labelled as ‘x’ are synaptic currents recorded under controlconditions. Traces labelled as ‘y’ are synaptic currents recorded in the presenceof the NMDAR antagonist D(2)-2-amino-5-phosphonopentanoate (D-AP5).Subtraction of y from x yields the traces labelled as ‘z’, the NMDAR-mediatedcomponent. Right: ratio of NMDAR-mediated/AMPAR-mediated currentscalculated as z/y. *P , 0.05 in a and c; **P , 0.01 in a.

LETTERS NATURE | Vol 465 | 27 May 2010

500Macmillan Publishers Limited. All rights reserved©2010

19. Chen, B. S. & Roche, K. W. Growth factor-dependent trafficking of cerebellarNMDA receptors via protein kinase B/Akt phosphorylation of NR2C. Neuron 62,471–478 (2009).

20. Wolf, J. A. et al. NMDA/AMPA ratio impacts state transitions and entrainment tooscillations in a computational model of the nucleus accumbens medium spinyprojection neuron. J. Neurosci. 25, 9080–9095 (2005).

21. Salmena, L., Carracedo, A. & Pandolfi, P. P. Tenets of PTEN tumor suppression.Cell 133, 403–414 (2008).

22. Groszer, M. et al. Negative regulation of neural stem/progenitor cell proliferationby the Pten tumor suppressor gene in vivo. Science 294, 2186–2189 (2001).

23. Knight, Z. A. et al. A pharmacological map of the PI3-K family defines a role forp110a in insulin signaling. Cell 125, 733–747 (2006).

24. Ruckle, T., Schwarz, M. K. & Rommel, C. PI3Kc inhibition: towards an ‘aspirin ofthe 21st century’? Nature Rev. Drug Discov. 5, 903–918 (2006).

25. Kemp, J. A. & McKernan, R. M. NMDA receptor pathways as drug targets. NatureNeurosci. 5 (Suppl.), 1039–1042 (2002).

26. Dirnagl, U., Iadecola, C. & Moskowitz, M. A. Pathobiology of ischaemic stroke: anintegrated view. Trends Neurosci. 22, 391–397 (1999).

27. Cowan, C. M. & Raymond, L. A. Selective neuronal degeneration in Huntington’sdisease. Curr. Top. Dev. Biol. 75, 25–71 (2006).

28. Lipton, S. A. The molecular basis of memantine action in Alzheimer’s disease andother neurologic disorders: low-affinity, uncompetitive antagonism. Curr.Alzheimer Res. 2, 155–165 (2005).

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

Acknowledgements We thank R. Shigemoto, T. Iwatsubo, I. Kanazawa, T. Shimizu,H. Ichijo, K. Yamada, Y. Imai, K. Kawamura, Y. Kanaho and T. Itoh for discussions;and H. Takahashi, K. Sasaki, Y. Sugihara, S. Kumagai, A. Kato, C. Horie, M. Nishio,S. Sato, T. Sugawara, R. Nakamura and K. Mizoi for technical support.Anti-Met-enkephalin antibody was provided by T. Kaneko and T. Furuta, andanti-NR2C antibody was provided by B. S. Chen and W. Roche. This work wassupported by research grants from the Ministry of Education, Culture, Sports andTechnology of Japan (MEXT); the Japan Society for the Promotion of Science(JSPS); the Japan Science and Technology Corporation (JST); the NaitoFoundation; and the Toray Science Foundation. H.H., T.T., A.S. and T.S. aresupported by the Global Centers of Excellence Program of MEXT. T.T. and T.S. aresupported by a Grant-in-Aid for Creative Scientific Research from MEXT.

Author Contributions J.S. and T.S. designed the research. J.S. generated theInpp4A-deficient mice. J.S., S.K., R.I., H.M. and Y.T. performed experiments andcollected and analysed data. T.M. carried out the electrophysiological studies. K.T.and H.H. performed the lentiviral reconstitution experiments. S.C. performed theultrastructural analyses. Y.H., S.T., S.E., K.M., K.A. and M.Y. contributed to thehistological analyses and provided technical assistance with the shRNAexperiments. M.D. and C.M. contributed to the analyses of weeble mice. T.T., A.S.and T.S. supervised the research. T.S. wrote the paper with the input of J.S., T.M.,H.H. and M.Y.

Author Information Reprints and permissions information is available atwww.nature.com/reprints. The authors declare no competing financial interests.Correspondence and requests for materials should be addressed to J.S.(sasakij@med.akita-u.ac.jp) or T.S. (tsasaki@med.akita-u.ac.jp).

NATURE | Vol 465 | 27 May 2010 LETTERS

501Macmillan Publishers Limited. All rights reserved©2010

METHODSGeneration of INPP4A-deficient mice. A knockout mutation of the mouse

Inpp4a gene was generated by replacing the first and second coding exons with

a LacZ-PGK-Neor cassette. The targeting vector contained a 7-kb genomic mouse

(129/Ola) Inpp4a fragment plus a LacZ-PGK-Neor cassette inserted in antisense

orientation to Inpp4a transcription. The linearized construct was electroporated

into 1 3 107 E14K mouse embryonic stem (ES) cells. ES cell colonies resistant to

G418 (0.3 mg ml21) were screened for homologous recombination by PCR.

Recombinant colonies were confirmed by hybridizing a Southern blot of

HindIII- and XbaI-digested genomic DNA to a 541-bp 39 flanking probe.

Targeted ES cells were injected into C57BL/6 blastocysts. Chimaeric male mice

were crossed with C57BL/6 females to achieve germline transmission. After

heterozygous matings, Inpp4a2/2 mice were distinguished from Inpp4a1/2

and Inpp4a1/1 mice by PCR. An oligo primer (59-TGAGGAATCAGGTGAG

GTGTGACC-39) common to the Inpp4a1 and Inpp4a2 alleles, a primer specific

for the Inpp4a1 allele (59-TCCCAGTTTCAGGAGCATTAGC-39), and a primer

specific for the Inpp4a2 allele (59-TTGAGGGGACGACGACAGTATC-39) were

combined in the same PCR reaction.

Dyskinesia assessments. Mice at postnatal day 15 (P15) to P19 received intra-

peritoneal administration of saline or 0.5 mg kg21 MK801 (Sigma). After 1 h,

hindlimb clasping in response to tail suspension was assessed as described29.

Briefly, P15–P21 mice were suspended by the tail and observed in 2-s time bins

for 14 s. Each mouse was allocated a score 5 1 for dystonic movement, and a

score 5 0 for an absence of abnormal movement (maximum score of 7).

Dystonic movement was defined as any movement where the hindlimbs were

pulled together into the central body axis. For convulsion assessment, mice were

removed from their home cages 1 h after MK801 administration, placed on an

observation table, and video-recorded for 60 s (3 trials per individual). The

duration of hyperkinetic movements was measured. All assessments were made

at room temperature (20–25 uC) during the day.

Histological analyses. Mice were fixed by cardiac perfusion with 10% formalin

neutral buffer solution (WAKO). Brain tissues were excised and post-fixed in the

same fixative overnight. Tissue blocks were embedded in paraffin and 4–10-mm

sections were cut and stained with Nissl (cresyl violet). Fluorescent TUNEL

staining was performed using a kit according to the manufacturer’s directions

(Roche). For immunohistochemistry, deparaffinized sections were incubated in

citrate buffer (pH 7.0), treated with 3% H2O2 followed by 2.5% normal horse

serum, and incubated overnight at 4 uC with primary antibodies against GFAP

(1:100). Immunostained sections were incubated with peroxidase-conjugated

secondary antibodies that were detected using DAB solution (WAKO).

Immunofluorescent detection of Met-enkephalin was conducted as described30

except that the vibratome sections were 50 mm thick. Fluorescent images were

acquired on a DM IRE2 microscope (Leica) fitted with a confocal imaging

system (Yokogawa) as described31.

Lentiviral INPP4A reconstitution. Human INPP4A cDNA was inserted into the

multi-cloning site (MCS) of the lentiviral plasmid pCL20c CMV GFP-P2A-

MCS32,33. This vector can be used to express a green fluorescent protein

(GFP)-labelled protein of interest under the control of the cytomegalovirus

(CMV) promoter34. P2A is a picornavirus ‘self-cleaving’ peptide sequence,

D-V/I-E-X-N-P-G, which is extremely rare and subject to cleavage after glycine.

The detailed procedures of viral vector production and assessment of viral titre

are described elsewhere35. The lentiviral preparation used for injection, which

had a mean titre 6 s.d. of 2.1 6 0.4 3 1010 transduction units per ml, was stored

at 4 uC and used within one week.

Striatal injection of the lentiviral vector. On P2, mouse pups were anaesthe-

tized using 2% isoflurane (flow speed: 1 l min21) inhalation and mounted on a

stereotactic frame. The blunt-ended tip (33 gauge) of a Hamilton syringe

attached to a micropump (UltramicroPump II; World Precision Instruments)

was placed in mouse striata at the following coordinates: 1.5 mm rostral to

bregma, 61.0 mm lateral to midline, and 1.5 mm ventral to the skull surface.

The anaesthetized animals were injected on one side of the striatum with a single

0.25-ml aliquot containing the lentiviral vector delivered at a rate of 100 nl min21.

Injected mice were housed in a safety rack (Type SCL; Oriental Giken Inc.) in

which clean air entered through a super-efficient filter and contaminated air was

exhausted through the same filter. On P20, vibratome sections of the injected

striata were subjected to histological analyses as described above.

PtdIns(3,4)P2 measurement. To determine endogenous PtdIns(3,4)P2 levels,

cellular lipids were extracted from isolated striata using the Bligh-Dyer method

and dried. In vitro lipid kinase reactions using phosphatidylinositol phosphate

5-kinase type I a (PIPKIa) were performed as described36. Briefly, dried cellular

lipids were re-suspended in a kinase reaction buffer (50 mM Tris-HCl pH 7.5, 1

mM EGTA) by vortexing and sonication for 3 min. The resulting micelles were

incubated with Myc-tagged PIPKIa and an ATP mixture (50mM [a-32P]ATP

(1mCi) plus 10 mM MgCl2) at 37 uC for 25 min. Reactions were quenched by the

addition of chloroform/methanol/8% HClO4 (5:10:4). Following vigorous vortex-

ing, chloroform/HClO4 (5:5) was added to recover the organic phase, which was

then washed twice in chloroform-saturated 1% HClO4 before drying. Dried lipids

were resolved in chloroform/methanol (6:1) and separated by thin-layer chro-

matography (Silica Gel 60; Merck) as described37. Radioactivity of PtdIns(4,5)P2

and PtdIns(3,4,5)P3 (synthesized from PtdIns(4)P and PtdIns(3,4)P2, respect-

ively) was quantified using a FLA-5000 bioimaging analyser (Fuji). To compare

PtdIns(3,4)P2 levels among striata, the raw radioactive counts determined for

PtdIns(3,4,5)P3 were normalized to the raw radioactive counts for PtdIns(4,5)P2

in each reaction.

Electron microscopy. Mice were perfused through the left cardiac ventricle with

35–50 ml 3% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4). Isolated striata

were diced and fixed overnight at 4 uC. After incubation for 2 h in 1% OsO4 in

0.1 M cacodylate buffer, the specimens were dehydrated in an ethanol series,

passed through propylene oxide, and embedded in epoxy resin. Ultrathin sec-

tions (60 nm) were collected on copper grids and stained for 10 min in 4% uranyl

acetate. The specimens were observed with a Hitachi H-7650 electron micro-scope at 80 kV.

RT–PCR. Total RNA was prepared using TRIzol reagent (Gibco BRL), and first-

strand cDNA was synthesized using 5 mg total RNA and M-MLV Reverse

Transcriptase (TOYOBO). Specific PCR primers were as follows: GLT1 forward,

59-GCTGGGGAAAAATCTCCTGC-39; GLT1 reverse, 59-TGACCGCCTTGG

TGGTATTG-39; EAAC1 forward, 59-TGAATGACATCAACAGGACGGG-39;

EAAC1 reverse, 59-AGGCAAAGCGGAAAGGGTTC-39; GLAST forward, 59-TC

ATTCACGCCGTCATCGTC-39; GLAST reverse, 59-CACCGAGTTCCCCATTT

CAAC-39; GAPDH forward, 59-CAACGACCCCTTCATTGACCTC-39; GAPDH

reverse, 59-ATCCACGACGGACACATTGG-39; Inpp4AD744G forward, 59-ATGA

CAGCAAGAGAGCACAG-39; Inpp4AD744G reverse, 59-TCACAAACTGCCGAG

GCAC -39.

Amplification conditions for detecting glutamate transporter gene expression

were: 94 uC for 0.5 min; 98 uC for 0.5 min for 20 cycles (GAPDH) or 25 cycles

(GLT1) or 28 cycles (EAAC1, GLAST, Inpp4AD744G); 94 uC for 0.5 min, 65 uC for

0.5 min, 72 uC for 0.75 min; 72 uC for 7 min; and 4 uC for 2 h (PerkinElmer). The

amplification of glutamate transporters showed a linear increase in detectable

levels with corresponding increases in cDNA, allowing a relative quantification

of glutamate transporter mRNA levels normalized to the expression of the

housekeeping gene GAPDH.

In vivo microdialysis. Extracellular concentrations of glutamate were measured

essentially as described38 except that the probes were stereotaxically implanted in

the left striatum (1 mm anterior and 1.5 mm lateral to the bregma) of 2–3-week-

old anaesthetized mice. At 26 h after implantation, dialysis probes were perfused

with Ringer’s solution at a rate of 2 ml min21. Microdialysate samples (20ml)

were collected every 10 min for 60 min and glutamate concentrations were mea-

sured by HPLC (Eicom Corp).

Glutamate uptake. Glia cells derived from newborn mouse forebrains were main-

tained in Dulbecco’s modified Eagle’s medium (DMEM; Sigma). The medium

was replaced by an incubation buffer (5 mM Tris-HCl, 10 mM HEPES-NaOH,

2.5 mM KCl, 1.2 mM CaCl2, 1.2 mM MgCl2, 1.2 mM K2HPO4, 10 mM glucose,

pH 7.5, supplemented with 140 mM NaCl or LiCl), and 40mM [3H]L-glutamate

(0.5mCi) was added at time zero. After incubation at 37 uC for 10 min, the assay

was terminated by washing twice with ice-cold incubation buffer and cells weresolubilized in 0.1 M NaOH.

To prepare synaptosomes, mouse striata were isolated and homogenized in

0.32 M sucrose. The homogenates were centrifuged at 800g for 10 min and the

resulting supernatant was further centrifuged at 13,000g for 10 min at 4 uC to

yield synaptosomal pellets. For the glutamate uptake assay, synaptosomal pellets

(200mg protein) re-suspended in incubation buffer were incubated with 1 mM

[3H]L-glutamate (0.5mCi) at 37 uC for 10 min. Unincorporated radioactivity was

removed by centrifugation (13,000g for 1 min) and the pellet was solubilized in

0.1 M NaOH containing 1% SDS. Incorporated glutamate was determined by

liquid scintillation counting (ALOKA). Sodium-dependent glutamate uptake

was defined to be the difference between the radioactivity detected in the pres-

ence of sodium- as opposed to lithium-containing buffer.

Striatal cell culture. Striata were dissected from 30–40 wild-type C57BL/6 mice

at embryonic day 18–19, pooled in ice-cold DMEM, diced, and incubated with

10 U ml21 papain at 36 uC for 20 min. Striatal cells were suspended in neurobasal

medium supplemented with 2% B-27, 0.5 mM glutamine, 100 U ml21 penicillin

and 100 mg ml21 streptomycin (all from Invitrogen), and plated at a density of

0.6–1.2 3 106 cells per cm2 in culture dishes (Greiner) coated with polyethyle-

neimine (Sigma). To enrich for neurons, 8mM cytosine arabinoside (Sigma) was

added for 2 days followed by a return to regular neurobasal medium.

Construction of shRNA-expressing plasmids. Sequences of the oligonucleo-

tides used for the construction of a shRNA plasmid directed against Inpp4a were

doi:10.1038/nature09023

Macmillan Publishers Limited. All rights reserved©2010

as follows: 59-GATCCGAATGTTGGAAGTCGCAAATTCAAGAGATTTGCGACTTCCAACATTCTTTTTTGGAAA-39 and 59-AGCTTTTCCAAAAAAGAATG

TTGGAAGTCGCAAATCTCTTGAATTTGCGACTTCCAACATTCG-39.

Complementary oligonucleotides were annealed and inserted into the

BamHI/HindIII sites of the pSilencer 2.1-U6 vector containing an RNA poly-

merase III promoter (Ambion). A scrambled shRNA (Ambion) was used as the

negative control. Freshly isolated striatal cells were electroporated (Amaxa) with

pEGFP-C1 plus either control shRNA or Inpp4a shRNA plasmids. Transfected

cells were plated as above and maintained in vitro for at least 7 days before

experimentation.

Excitotoxicity in cultured striatal neurons. To induce excitotoxicity, cultures

were incubated with glutamate (0, 50 or 400mM) for 16 h. Cells were fixed in

PBS(-) containing 5% formalin and 4% sucrose, permeabilized with methanol

on ice, extensively washed with PBS(-), and incubated with rabbit anti-GFP and

mouse anti-MAP2 in 1% normal goat serum in PBS(-) at 4 uC overnight. After

three washes in PBS(-), cells were incubated with Alexa488 goat anti-rabbit IgG

plus Alexa568 goat anti-mouse IgG. Images of fluorescently labelled cells in 5–10

randomly chosen fields per dish were captured on a DM IRE2 microscope

(Leica) fitted with a confocal imaging system (Yokogawa). Data were storedwithout enhancement and analysed offline. The mean number of GFP1 neurons

was determined from three independent dishes per treatment.

To compare the excitotoxic potency of glutamate on control shRNA- and

Inpp4a shRNA-expressing neurons, the GFP1 cell count in the control culture

(no glutamate) was assigned a value of 100 for each transfectant, and relative

numbers of viable cells were calculated for each treatment. For NMDAR inhibi-

tion studies, cells were incubated with MK801 (20 mM; Sigma) for 15 min before

glutamate treatment.

To study the effects of phosphoinositides on excitotoxicity, we used the

Shuttle PIP kit (according to the manufacturer’s instructions; Echelon) to com-

plex Di-C16 phosphoinositides to histone H1. (The H1 acted as a carrier to

facilitate entry into cells.) Neurons were incubated with histone H1 alone or

H1–phosphoinositide complexes for 30 min before challenge with 10 mM glu-

tamate for 16 h. Viable neurons were visualized by immunofluorescent staining

with anti-MAP2 and DAPI. The absolute number of MAP21 neurons was deter-

mined as for GFP1 neurons above.

Isolation of postsynaptic density fractions. Preparation of postsynaptic density

(PSD) fractions was performed as described39 with some modifications. Briefly,striata from P15–P20 mice were homogenized in buffer A (0.32 M sucrose,

50 mM Tris-HCl pH 7.4, 0.1 mM PMSF, 1 mM EGTA, plus protease inhibitors

(Roche)) and centrifuged at 1,000g for 20 min. The supernatant was collected

and centrifuged at 10,000g for 15 min, and the recovered pellet was subjected to

osmotic shock by dilution in double-distilled water, homogenization in buffer A,

and centrifugation at 25,000g for 20 min. The pellet (lysed synaptosomal mem-

brane fraction) was re-suspended in buffer A and loaded onto a discontinuous

sucrose gradient composed of 0.8 M and 1.5 M sucrose. After centrifugation at

100,000g for 2 h, the fraction recovered from the interface between 0.8 M and 1.5

M sucrose was subjected to centrifugation at 202,000g for 30 min. The resulting

pellet (synaptosomal membranes) was incubated for 30 min in buffer B (0.16 M

sucrose, 5 mM Tris-HCl pH 8.0, 1 mM EDTA, 75 mM KCl, 0.5% Triton X-100

plus protease inhibitors), and centrifuged at 202,000g for 20 min. The resulting

pellet was considered to be the purified PSD fraction whereas the supernatant

served as the non-PSD fraction. Protein concentrations were measured using a

protein assay kit (Bio-Rad). Each fraction (10mg protein) was analysed by standard

western blotting as described40.

Electrophysiology. Mice were killed by decapitation under deep halothaneanaesthesia, and coronal slices (300 mm) containing the striatum were cut using

a microslicer (PRO7, Dosaka) in ice-cold oxygenated Krebs cutting solution

(120 mM choline chloride, 2.5 mM KCl, 0.5 mM CaCl2, 7 mM MgCl2,

1.25 mM NaH2PO4, 26 mM NaHCO3, 15 mM D-glucose, 1.3 mM ascorbic acid).

Striatal slices were transferred to a holding chamber containing standard Krebs

solution (124 mM NaCl, 3 mM KCl, 2.4 mM CaCl2, 1.2 mM MgCl2, 1 mM

NaH2PO4, 26 mM NaHCO3, 10 mM D-glucose, pH 7.4 when bubbled with

95% O2-5% CO2) and incubated at room temperature (21–26 uC) for at least

1 h before recording.

For recording, a striatal slice was transferred to the recording chamber, held

submerged, and superfused with standard Krebs solution (bubbled with 95%

O2-5% CO2) at a rate of 3–4 ml min21. Neurons within the striatum were visually

identified with a 603 water immersion objective attached to an upright micro-

scope (BX50WI, Olympus Optics). Images were acquired using a cooled CCD

camera (CCD-300T-RC, Nippon Roper) and displayed on a video monitor (LC-

150M1, SHARP). Pipettes for whole-cell recordings were made from standard-

walled borosilicate glass capillaries (1.5 mm outer diameter; Harvard Apparatus).

Synaptic currents were evoked by focal stimulation in the presence of bicuculline

(10mM) and strychnine (0.5mM) to block GABAA and glycine receptor-mediated

current components, respectively. Holding potential was 140 mV. Synaptic cur-

rents were recorded in the presence or absence of the NMDAR antagonist D-AP5

(25mM) to obtain AMPAR-mediated component.

For analyses of membrane properties, current-clamp recordings were carried

out using a K-gluconate-based internal solution (120 mM K-gluconate, 6 mM

NaCl, 5 mM CaCl2, 2 mM MgCl2, 0.2 mM K-EGTA, 10 mM K-HEPES, 2 mM

Mg-ATP, 0.3 mM Na-GTP, pH 7.4 adjusted with 1 M KOH). For recording of

synaptic currents, patch pipettes were filled with a CsCl-based internal solution

(140 mM CsCl, 9 mM NaCl, 1 mM Cs-EGTA, 10 mM Cs-HEPES, 2 mM Mg-

ATP, pH 7.4 adjusted with 1 M CsOH). Whole-cell recordings were made from

striatal medium spiny neurons using a patch-clamp amplifier (Axon

Instruments). The series resistance was measured from the amplifier, and the

data were discarded if the series resistance changed by more than 10% of the

initial value. The access resistance was monitored by measuring capacitative

transients obtained in response to a hyperpolarizing voltage step (5 mV,

25 ms) from a holding potential of –65 mV. No correction was made for the

liquid junction potentials (calculated to be 5.0 mV by pCLAMP7 software, Axon

Instruments). Synaptic currents were evoked at a rate of 0.2 Hz (every 5 s) byextracellularly delivered voltage pulses (0.2–0.4 ms in duration) of suprathres-

hold intensity via a stimulating electrode filled with 1 M NaCl. The stimulating

electrode was placed within a 50–120mm radius of the recorded neuron. The

position of the stimulating electrode was varied until a stable response was evoked

in the recorded neuron. Experiments were carried out at room temperature

(21–26 uC).

Antibodies. Antiserum specific for Met-enkephalin was provided by T. Kaneko

and T. Furuta. The monoclonal antibody (mAb) against GFAP was from DAKO.

Anti-GFP polyclonal antibody (pAb) was from MBL. Anti-MAP2 mAb was from

Sigma. Antibodies against choline acetyltransferase, parvalbumin, somatostatin,

NR1, synaptophysin and GluR2 were from Chemicon. Anti-PSD95 and anti-

NR2B were from Upstate. Anti-NR2C (cross-reactive with NR2A and NR2B)

was from Enzo Life Sciences. Anti-INPP4A pAb and anti-phosphotyrosine mAb

(PY99) were from Santa Cruz. Alexa488 goat anti-rabbit IgG and Alexa568 goat

anti-mouse IgG were from Invitrogen, and other secondary antibodies were

from DAKO or Vector Laboratories.

Statistical analyses. All data are expressed as the mean 6 s.e.m. of at least trip-

licate samples. Data were analysed by two-tailed, unpaired Student t-tests. For all

experiments, P values ,0.05 were considered significant.

29. Gantois, I. et al. Ablation of D1 dopamine receptor-expressing cells generatesmice with seizures, dystonia, hyperactivity, and impaired oral behavior. Proc. NatlAcad. Sci. USA 104, 4182–4187 (2007).

30. Furuta, T., Zhou, L. & Kaneko, T. Preprodynorphin-, preproenkephalin-,preprotachykinin A- and preprotachykinin B-immunoreactive neurons in theaccumbens nucleus and olfactory tubercle: double-immunofluorescenceanalysis. Neuroscience 114, 611–627 (2002).

31. Nishio, M. et al. Control of cell polarity and motility by the PtdIns(3,4,5)P3

phosphatase SHIP1. Nature Cell Biol. 9, 36–44 (2007).32. Torashima, T. et al. In vivo transduction of murine cerebellar Purkinje cells by HIV-

derived lentiviral vectors. Brain Res. 1082, 11–22 (2006).33. Hanawa, H. et al. Efficient gene transfer into rhesus repopulating hematopoietic

stem cells using a simian immunodeficiency virus-based lentiviral vector system.Blood 103, 4062–4069 (2004).

34. Takayama, K. et al. Purkinje-cell-preferential transduction by lentiviral vectorswith the murine stem cell virus promoter. Neurosci. Lett. 443, 7–11 (2008).

35. Torashima, T. et al. Exposure of lentiviral vectors to subneutral pH shifts thetropism from Purkinje cell to Bergmann glia. Eur. J. Neurosci. 24, 371–380 (2006).

36. Yamazaki, M. et al. Phosphatidylinositol 4-phosphate 5-kinase is essential forROCK-mediated neurite remodeling. J. Biol. Chem. 277, 17226–17230 (2002).

37. Sasaki, T. et al. Function of PI3Kc in thymocyte development, T cell activation, andneutrophil migration. Science 287, 1040–1046 (2000).

38. Oda, M. et al. Direct correlation between ischemic injury and extracellular glycineconcentration in mice with genetically altered activities of the glycine cleavagemultienzyme system. Stroke 38, 2157–2164 (2007).

39. Inagaki, S. et al. Sema4c, a transmembrane semaphorin, interacts with a post-synaptic density protein, PSD-95. J. Biol. Chem. 276, 9174–9181 (2001).

40. Irie-Sasaki, J. et al. CD45 is a JAK phosphatase and negatively regulates cytokinereceptor signalling. Nature 409, 349–354 (2001).

doi:10.1038/nature09023

Macmillan Publishers Limited. All rights reserved©2010