SMARCA3, a Chromatin-RemodelingFactor, Is Required for p11-DependentAntidepressant ActionYong-Seok Oh,1 Pu Gao,3 Ko-Woon Lee,1 Ilaria Ceglia,1 Ji-Seon Seo,1 Xiaozhu Zhang,1 Jung-Hyuck Ahn,4 Brian T. Chait,2

Dinshaw J. Patel,3 Yong Kim,1,* and Paul Greengard1,*1Laboratory of Molecular and Cellular Neuroscience2Laboratory of Mass Spectrometry and Gaseous Ion Chemistry

The Rockefeller University, New York, NY 10065, USA3Structural Biology Program, Memorial Sloan-Kettering Cancer Center, New York, NY 10065, USA4Department of Biochemistry, Ewha Womans University School of Medicine, Yangcheon-ku, Seoul 158-710, Republic of Korea

*Correspondence: kimyo@rockefeller.edu (Y.K.), greengard@rockefeller.edu (P.G.)http://dx.doi.org/10.1016/j.cell.2013.01.014

SUMMARY

p11, through unknown mechanisms, is required forbehavioral and cellular responses to selective sero-tonin reuptake inhibitors (SSRIs). We show thatSMARCA3, a chromatin-remodeling factor, is a targetfor the p11/annexin A2 heterotetrameric complex.Determination of the crystal structure indicates thatSMARCA3 peptide binds to a hydrophobic pocketin the heterotetramer. Formation of this complexincreases the DNA-binding affinity of SMARCA3and its localization to the nuclear matrix fraction. Inthe dentate gyrus, both p11 and SMARCA3 are highlyenriched in hilar mossy cells and basket cells. TheSSRI fluoxetine induces expression of p11 in bothcell types and increases the amount of the ternarycomplex of p11/annexin A2/SMARCA3. SSRI-induced neurogenesis and behavioral responsesare abolished by constitutive knockout ofSMARCA3.Our studies indicate a central role for a chromatin-remodeling factor in the SSRI/p11 signaling pathwayand suggest an approach to the development ofimproved antidepressant therapies.

INTRODUCTION

Selective serotonin reuptake inhibitors (SSRIs) are currently the

most widely used class of antidepressants (Berton and Nestler,

2006). SSRI medications generally take several weeks to show

clinical efficacy, including mood elevation, despite their imme-

diate effect on serotonergic neurotransmission. This therapeutic

delay suggests the involvement of complicated downstream

mechanisms, including long-term changes in gene expression

and neuroplasticity. However, our knowledge of the molecular

mechanisms underlying the efficacy of long-term treatment

with SSRIs and of the pathophysiology of depression is still

rudimentary.

p11 (S100A10) is a pivotal regulator of depression-like behav-

iors and a mediator of antidepressant responses (Svenningsson

et al., 2006). Despite the importance of p11 in the actions of

SSRIs, our knowledge about its underlying molecular mecha-

nisms is limited (Svenningsson et al., 2006). Annexin A2

(AnxA2) is a well-characterized binding partner for p11. AnxA2,

together with p11, plays a role in trafficking of membranous/

cytoplasmic proteins to plasma membrane or in providing

themwith firm anchorage at the plasmamembrane and the cyto-

skeletal structure (Rescher and Gerke, 2008). p11 and AnxA2

were also found to localize in the nucleus and interact with

nuclear proteins (Das et al., 2010; Liu and Vishwanatha, 2007),

although the precise roles of p11 and AnxA2 in the nucleus

have not been clearly defined. In this study, we have observed

that chronic treatment with an SSRI, fluoxetine (FLX), increases

the levels of the p11/AnxA2 complex. We have identified

SMARCA3, a chromatin-remodeling factor, as a downstream

target of the p11/AnxA2 complex. Our data indicate that the

p11/AnxA2/SMARCA3 pathway mediates both neurogenic and

behavioral responses to SSRIs.

RESULTS

Identification of SMARCA3 as a Specific Binding Partnerof p11/AnxA2 Complexp11, together with AnxA2, forms a heterotetramer in cells.

However, it is not yet established that p11 exists in a protein

complex with AnxA2 in brain tissue. Here, we show that the

protein level of AnxA2 is drastically downregulated in the frontal

cortex and hippocampus of p11 knockout (KO) mice (Figure 1A),

despite unchanged levels of AnxA2 transcript (Figure 1B). In

contrast, the protein level of AnxA1, another annexin family

member, and of S100B, another S100 family member, was not

altered in the brain of p11 KO mice, indicating the specificity of

the physiological interaction between p11 and AnxA2 in the brain

(Figure 1A). We have also observed that p11 protein level in the

hippocampal lysates from AnxA2 KO mice is reduced (data not

shown). Given that components of a protein complex often stabi-

lize each other, the data strongly support the existence of

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. 831

Figure 1. p11/AnxA2 as an Antidepressant-

Regulated Protein Complex

(A) Brain lysates from frontal cortex and hippo-

campus of WT (+/+) and p11 KO (�/�) mice were

immunoblotted for p11, AnxA2, S100B, and

AnxA1.

(B) mRNA levels of p11 and AnxA2 in the

hippocampus were measured using qPCR in WT

(+/+) and p11 KO (�/�) mice. Data represent

mean ±SEM.

(C) Coimmunoprecipitation of p11 and AnxA2 from

lysates of N2a neuroblastoma cells and mouse

hippocampus using anti-p11 antibody. IP, immu-

noprecipitate.

(D) WT (+/+) or p11 KO (�/�) mice were adminis-

tered VEH or FLX for 2 weeks. Hippocampal

lysates were immunoblotted for p11, AnxA2, and

b-actin.

(E) Quantitation of the immunoblot shown in (D)

using infrared imaging system (Odyssey; LI-COR).

Data represent mean ±SEM. *p < 0.05 and **p <

0.01, t test. ns, nonsignificant; n.d., not detectable.

a protein complex of p11 and AnxA2 in the brain. In addition, we

were able to coimmunoprecipitate AnxA2 with p11 from lysates

of the hippocampus as well as from lysates of N2a neuroblas-

toma cells (Figure 1C).

Previous studies showed that p11 was induced in the frontal

cortex (Svenningsson et al., 2006) and hippocampus (Warner-

Schmidt et al., 2010) by chronic administration of antidepres-

sants. In the present study, we observed concomitant upre-

gulation of p11 (170.4% ± 7.3% of p11(+/+)-vehicle (VEH) group;

p = 0.004) and of AnxA2 (167.1% ± 20.8%; p = 0.042) (Figures 1D

and 1E). This AnxA2 increase was not observed in p11 KO mice

(Figure 1E). Collectively, these results suggest that p11 and

AnxA2 exist as a protein complex, which can be induced by

antidepressant administration.

Next, we undertook a search for binding partners for p11/

AnxA2. To ensure the specificity of the interaction with the

p11/AnxA2 heterotetramer, we compared wild-type (WT) versus

C83S and C83Q mutants of p11, which prevent the interaction

between p11 and AnxA2 (Kube et al., 1992). C83S and C83Q

mutations in p11 significantly decreased the interaction with

AnxA2, without altering the interaction with endogenous p11 to

form a p11 dimer, suggesting that C83 mutations selectively

interfere with the heterotetramer formation, but not the homodi-

merization of p11 molecules (Figure S1A available online). After

transfection of p11 WT and C83 mutant plasmids into HEK293

cells, we isolated the protein complex of p11 using immunopre-

cipitation (Figure 2A). Four proteins with relative molecular mass

of 700, 260, 125, and 36 kDa were coprecipitated with WT p11

and were identified by mass spectrometry as AHNAK1 (AHNAK

nucleoprotein), SPT6 (suppressor of Ty 6 homolog S. cerevisiae),

832 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

SMARCA3 (SWI/SNF-related, matrix-

associated, actin-dependent regulator of

chromatin, subfamily A, member 3), and

AnxA2, respectively (Figure 2A). The iden-

tity of each protein was further confirmed

by immunoblotting with specific anti-

bodies (Figure 2B). AHNAK1, SPT6, and SMARCA3 were copre-

cipitatedwithWT p11, and the interactionwas greatly reduced or

abolished by either C83S or C83Q mutation of p11, indicating

that the interaction likely needs AnxA2 binding to p11. AHNAK1

has been reported as a binding protein of p11/AnxA2 (Benaud

et al., 2004). We focused on SMARCA3 in the following studies

because of the potential physiological importance of this chro-

matin-remodeling factor. To evaluate the role of AnxA2 in the

intermolecular interaction, an in vitro pull-down assay using

GST-p11, AnxA2, and 35S-labeled SMARCA3 was used. The

SMARCA3 interaction was significantly increased by the addi-

tion of AnxA2 to the pull-down assay mixture (Figure S1B). The

interaction of SMARCA3 with p11/AnxA2 was further confirmed

by the inverse immunoprecipitation using anti-SMARCA3

antibodies (Figure 2C). Collectively, these results identified

SMARCA3 as a binding partner of p11/AnxA2.

SMARCA3 belongs to the family of SWI/SNF proteins that use

the energy of ATP hydrolysis to remodel chromatin in a variety of

nuclear processes, such as transcriptional regulation, and DNA

replication and repair (Debauve et al., 2008). SMARCA3 contains

multiple domain structures, including DNA-binding, helicase

ATP-binding, RING-type Zinc finger, and helicase C-terminal

domains (Figure 2D). We next performed in vitro pull-down assay

with a series of deletion constructs of SMARCA3 to determine

the binding region for p11/AnxA2 (Figures 2D, S1C, and S1D).

Although the serial deletion from the SMARCA3 C terminus

had no effect on the interaction with p11 (Figure S1C), the dele-

tion of the N terminus of SMARCA3 abolished the interaction

(Figure S1D), localizing a binding region close to the N terminus

of SMARCA3 (Figure 2D). Through sequence alignment between

Figure 2. Identification of the Binding

Proteins of p11/AnxA2 Complex

(A) Control vector (Vec.) and vectors containingWT

or interaction-defective mutants (C83S and C83Q)

of p11 were transfected into HEK293 cells. Pro-

teins that coprecipitated with WT are marked with

arrowheads (red). A nonspecific band is indicated

by a black arrowhead. The proteins were identified

by tandem mass spectrometry (black arrows).

(B) Immunoblots for AHNAK1, SPT6, SMARCA3,

AnxA2, and Flag-p11, as indicated.

(C) Coimmunoprecipitation of p11/AnxA2/

SMARCA3 complex from brain lysates by two

different anti-SMARCA3 antibodies.

(D) WT, deletion mutants, and a peptide were

tested for their interaction with p11/AnxA2.

(E) Putative p11-binding sequences of AHNAK1

(aa 5654–5671), AHNAK2 (aa 5382–5399), and

SMARCA3 (aa 26–42).

(F) Interaction of the AHNAK1 or SMARCA3

peptides with p11/AnxA2 was confirmed by the

pull-down assay of biotinylated peptides. Scram-

bled peptide was used as control.

See also Figure S1.

AHNAK family proteins and the N terminus of SMARCA3, we

found a highly conserved putative binding motif, represented

byV-P-#-F-X-F (V, hydrophobic amino acid; P, proline; #, basic

amino acid; F, phenylalanine; X, any amino acid) (Figure 2E).

Next, we performed a peptide pull-down assay to validate p11/

AnxA2 binding to the putative binding motif. Indeed, the short

synthetic peptides from AHNAK1, AHNAK2, and SMARCA3

covering the putative motif were sufficient to bind to p11/

AnxA2 (Figure 2F; data not shown for AHNAK2 peptide).

Crystal Structure of p11/AnxA2 Bound to Its TernaryTarget, SMARCA3To understand the molecular interaction of SMARCA3 and

AHNAK1 with p11/AnxA2 in detail, we conducted structural

studies of p11 and AnxA2 complexed with SMARCA3 or

AHNAK1 peptides. We have used a p11-AnxA2 fusion protein

(henceforth named p11-AnxA2 peptide cassette) in which

p11 (1-92) is connected to AnxA2 peptide with a nine amino

Cell 152, 831–843,

acid linker (QENLYFQGD) (Rezvanpour

et al., 2009) to generate complexes

with added SMARCA3 (P26-F39) and

AHNAK1 (G5654-F5668) peptides. Crys-

tals of the complexes with SMARCA3

peptide (Figure 3) and AHNAK1 peptide

(Figure S2) diffracted to 3.0 and 2.0 A

resolution, respectively. The structures

of the two complexes exhibited similar

recognition principles; although the pep-

tide sequences are different, the back-

bones of bound SMARCA3 and AHNAK1

peptides superimpose well with an rmsd

of 0.4 A. The higher-resolution structure

of the AHNAK1 peptide complex (Figures

S2A–S2F) proved useful in resolving ambiguities in the lower-

resolution structure of the SMARCA3 peptide complex (Figures

3A–3F).

We have solved the 3.0 A crystal structure of the complex

between SMARCA3 (P26-F39) peptide and the p11-AnxA2

peptide cassette (Figures 3A and 3B; crystallization statistics in

Table S1). The fusion linker does not affect the binding between

p11 and AnxA2 peptide because we observe the same intermo-

lecular contacts in this complex as those reported previously

for nonlinked components (Rety et al., 1999). The p11-AnxA2

peptide cassette forms a symmetrical dimer, with individual

SMARCA3 peptides (electron density for bound peptide shown

in Figure 3C), binding eachmonomer in the dimer, while retaining

2-fold symmetry (Figures 3A and 3B). Complex formation

between SMARCA3 and p11-AnxA2 peptide cassette is medi-

ated by van derWaals contacts and hydrogen bond interactions,

whereby the SMARCA3 peptide interacts with elements of both

p11 and AnxA2 peptide (Figure 3D). Two highly conserved amino

February 14, 2013 ª2013 Elsevier Inc. 833

Figure 3. Crystal Structure of SMARCA3Peptide Bound to p11-AnxA2PeptideCassette asBinary Complex and to p11 and Full-Length AnxA2

as Ternary Complex

(A) Ribbon view of crystal structure of p11-AnxA2 peptide cassette in complex with SMARCA3 peptide. p11 (blue and green) and AnxA2 peptide (salmon) are

shown in ribbon, with the bound SMARCA3 peptides (yellow) in stick representations.

(B) Space-filling (peptide) and electrostatic surface (p11-AnxA2 peptide cassette) view of the complex illustrating the positioning of the pair of SMARCA3peptides

within the binding groove of the dimeric complex.

(C) Omit 2Fo�Fc electron densitymaps contoured at 1s level of bound SMARCA3 peptide in onemonomer of the complex. AnxA2 peptide (salmon) and p11 (blue)

are represented in ribbon views, SMARCA3 peptide (yellow) is represented in a stick view, and electron density contours are in green.

(D) Intermolecular contacts within one monomer of the complex between the SMARCA3 peptide (yellow) and p11 (blue)-AnxA2 peptide (salmon) cassette, with

hydrogen bonds depicted as red dashed lines.

(legend continued on next page)

834 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

acids (Pro35 and Phe37) in the bound SMARCA3 peptide are

directed inward and anchored within a hydrophobic pocket

formed by residues from both p11 (Phe42, Ile55, Leu59, Leu75,

Leu79, and Ala82) and AnxA2 peptide (Leu8, Leu11, and

Leu13) in the complex (Figure 3E), with similar interactions

observed for Pro5664 and Phe5666 in the AHNAK1 complex

(Figure S2E). The complex is stabilized by additional intermolec-

ular hydrogen bonding interactions involving Phe33(SMARCA3)-

Asp60(p11) and Arg36/Glu38(SMARCA3)-Ser12/Glu14(AnxA2

peptide) pairs (Figure 3D), with similar intermolecular contacts

observed in the AHNAK1 complex (Figure S2D). In addition,

the two bound SMARCA3 peptides are aligned in a head-

to-head arrangement that is stabilized through intermolecular

hydrogen bond formation between their N-terminal residues

(Figure 3A). As expected, two point mutations (Pro35A and

Phe37Y) in the binding consensusmotif of SMARCA3 diminished

the interaction with p11/AnxA2 (Figure S3A). In addition, Leu13A

mutation of AnxA2 abolished the interaction with SMARCA3 with

no effect on p11 interaction, whereas Leu8A and Leu11A

mutation blunted the interaction to p11 as well as to SMARCA3

(Figure S3B), showing a unique role of the Leu13 residue in

creating a binding pocket for the ternary targets. These structural

and mutagenesis studies account for the molecular basis of the

preferential binding of SMARCA3 to p11/AnxA2.

We compared the structures of p11-AnxA2 peptide cassette in

the free (Protein Data Bank [PDB] ID code 1BT6) and SMARCA3

peptide-bound states following structural superimposition (Fig-

ure 3F). We observe that Asp60 in p11 is looped out from the

peptide-binding groove in the free state but flips inward by

4.8 A on complex formation with the bound SMARCA3 peptide,

in the process forming additional intermolecular hydrogen bonds

(Figure 3D). In addition to this change, the Asp58-Asp64 loop

segment also undergoes a conformational change by shifting

toward and completes the peptide-binding groove on complex

formation (Figure 3F). A similar conformational transition is

observed on addition of AHNAK1 peptide to the p11-AnxA2

peptide cassette (Figure S2F). Indeed, the mutation on the p11

Asp60 residue abolished the interaction with SMARCA3 while

maintaining homodimerization of p11, as well as heterotetramer

formation with AnxA2, although the D60A mutant of p11 dis-

played a decrease in protein stability (Figure S3C). In our struc-

tures of complexes with bound SMARCA3 and AHNAK1 pep-

tides, the disulfide bridges observed for both p11 alone and

the p11-AnxA2 peptide cassette (Rety et al., 1999) are disrupted

due to the movement of the Asp58-Asp64 loop on complex

formation. As to the conformational change in AnxA2, we

observe an inward flip of Ser12 toward the peptide-binding site

by 4.6 A for the SMARCA3-bound complex, thereby forming

intermolecular hydrogen bonds with bound peptide in the

complex (Figures 3D and 3F). In addition, due to this movement

(E) Positioning of the conserved residues Pro35 and Phe37 of the bound SMARCA

(salmon) and p11 (blue).

(F) A view of the superimposed structures of the p11-AnxA2 peptide cassette in th

blue and AnxA2 peptide in salmon) states. Black arrows highlight the conformat

(G) Two views of 2.8 A structure of SMARCA3 peptide (yellow) in stick represen

representation shown for two different angles.

See also Figures S2 and S3, and Tables S1 and S2.

of Ser12, adjacent Leu13 in AnxA2 contributes to the closing of

an additional face of the hydrophobic pocket, thereby anchoring

the conserved residues Pro35 and Phe37 of bound SMARCA3

within the binding channel.

We have also successfully grown 2.8 A crystals of SMARCA3

peptide bound to p11 and full-length AnxA2, with two views of

the structure of the ternary complex shown in Figure 3G (X-ray

statistics summarized in Table S2). Importantly, the crystal struc-

ture of p11/full-length AnxA2/SMARCA3 peptide illuminated the

three-dimensional organization of each component within the

ternary complex (Figure 3G). Full-length AnxA2 is composed of

an N-terminal p11-binding region and a C-terminal annexin

repeat region with opposing convex and concave sides (Gerke

et al., 2005). The convex side of AnxA2 faces the cellular

membrane to mediate PIP2 binding, and the concave side faces

away from the membrane. Although two N-terminal peptides of

AnxA2 contact the lateral/bottom side of the inverted p11 homo-

dimer, two N-terminal peptides of SMARCA3 are anchored on

the top position of the inverted p11 dimer while crossing each

other. On the other hand, C-terminal annexin repeat regions of

full-length AnxA2 are not involved in intermolecular recognition

in the ternary complex. Furthermore, the C-terminal ends of

the SMARCA3 peptides point downward, suggesting that the

rest of the C-terminal parts of SMARCA3 including DNA-binding,

ATP-dependent helicase, and Zinc finger domains are likely

positioned opposite to the convex side of the annexin repeat

region of AnxA2. This mode of ternary complex assembly may

leave C-terminal annexin repeat regions of AnxA2 free to execute

the binding to PIP2 and actin without steric hindrance. Taken

together, our structural data on the complex with full-length

AnxA2 implicate additional regulatory mechanisms through

extended molecular interactions via C-terminal annexin repeat

region.

Regulation of SMARCA3 Activity through DirectInteraction with p11/AnxA2SMARCA3was initially cloned as a regulatory factor that binds to

the target DNA motifs of several gene promoters and enhancers

and was shown to regulate the transcription of tissue-specific

target genes such as PAI-1, b-globin, and prolactin (Debauve

et al., 2008). It is noteworthy that the p11-binding region

(aa 34–39) of SMARCA3 is located adjacent to the DNA-binding

domain (aa 38–287) and seems to be partially overlapped. This

observation prompted us to examine whether p11/AnxA2 inter-

action may regulate the DNA-binding affinity of SMARCA3.

SMARCA3 was known to interact with the B box element in the

promoter of the PAI-1 gene (Ding et al., 1996). We carried out

an in vitro reconstitution assay using purified recombinant pro-

teins and the B box oligonucleotide immobilized on beads. The

oligonucleotide pull-down assay showed that SMARCA3 can

3 peptide within a hydrophobic pocket formed by residues from AnxA2 peptide

e free (p11 in yellow and AnxA2 peptide in cyan) and SMARCA3-bound (p11 in

ional changes in p11 and AnxA2 peptide upon complex formation.

tation bound to full-length AnxA2 (salmon) and p11 (blue and green) in ribbon

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. 835

Figure 4. SMARCA3 Regulation by p11/

AnxA2 Complex

(A) The B box oligonucleotide was incubated with

the N-terminal domain of SMARCA3 (aa 1–350)

and p11 and/or AnxA2. Bound proteins were

immunoblotted.

(B) Quantitation of SMARCA3 bound to the B box

oligonucleotide.

(C) Whole cells (WC) and the nuclear matrix (NM)

were prepared from control (Control) or p11-

knockdown (KD) COS-7 cells and immunostained

with the indicated antibodies. Arrows indicate

knockdown cells, and open arrowheads show

nonknockdown cells. Scale bars, 20 mm.

(D) Whole-cell lysates and NM were prepared from

COS-7 cells transfected as indicated and im-

munoblotted for SMARCA3-V5 (a-V5), p11, AnxA2,

and Lamin-B (nuclear matrix marker).

(E) Transcriptional activity of SMARCA3 in N2a

cells after cotransfection of luciferase reporter

gene conjugated to PAI-1 promoter, together with

indicated plasmids. Immunoblots of cell lysates

and luciferase activity are shown.

(F) Transcriptional activity of SMARCA3 in Control

or p11-knockdown (KD) COS-7 cells transfected

with indicated SMARCA3 plasmids.

Mean ±SEM. *p < 0.05, **p < 0.01, and ***p < 0.001,

t test.

form a quaternary complex together with p11/AnxA2 and the

B box oligonucleotide (Figure 4A). Furthermore, p11/AnxA2

interaction increases the DNA-binding affinity of SMARCA3 by

up to 2.5-fold, whereas the equivalent amount of AnxA2 alone

did not show any effect (Figure 4B).

The distinct subnuclear localizations of nuclear factors are

essential to conduct chromatin remodeling, transcription, repli-

cation, and mRNA processing (Zaidi et al., 2007). In fact, among

SWI/SNF family chromatin remodelers, the hBAF (Brg1-associ-

ated factors) complex is known to be targeted to the nuclear

matrix/chromatin through direct interaction with nuclear PIP2

836 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

and actin/actin-related protein (Rando

et al., 2002; Zhao et al., 1998). It is well

established that p11/AnxA2 binds to

PIP2 as well as actin (Rescher and Gerke,

2008). Thus, we examined the possibility

that p11/AnxA2 may regulate the subnu-

clear localization of SMARCA3. In the

whole-cell preparation, p11 is predomi-

nantly cytoplasmic, and a much less but

significant level of p11 is in the nucleus,

whereas SMARCA3 is mainly localized

inside the nucleus. Our unpublished stud-

ies with primary cultured neurons indicate

that p11, together with AnxA2, shuttles

between cytoplasm and nucleus (data

not shown). In the nuclear matrix, which

is prepared after cell permeabilization fol-

lowed by chromatin digestion, the pres-

ence and the colocalization of p11 and

SMARCA3 are evident (Figure 4C). Impor-

tantly, the retention of SMARCA3 in the nuclear matrix is dramat-

ically reduced by the silencing of p11 expression with siRNA.

Consistent with that, biochemical fractionation of the nuclear

matrix revealed that WT SMARCA3 is in the nuclear matrix prep-

aration, whereas the P35A mutant, defective in p11/AnxA2 inter-

action, is not (Figure 4D), confirming the importance of the p11/

AnxA2 complex in the subnuclear localization of SMARCA3.

Taken together, p11/AnxA2 not only increases the DNA-

binding affinity of SMARCA3 but also anchors SMARCA3 to

the nuclear matrix presumably via the interaction of AnxA2 with

actin and PIP2.

Figure 5. SSRI Regulates p11 Expression in

Mossy Cells and Basket Cells in the Dentate

Gyrus

(A) Cell types expressing [p11]-EGFP in the dentate

gyrus. CRT (hilar mossy cells [MC]), parvalbumin

(PV, subpopulation of basket cells [BC(PV+)]),

calbindin (CBD, mature granule cells [GC]) neuro-

peptide Y (NPY, hilar interneurons perforant path

[HIPP]), glial fibrillary acidic protein (GFAP, astro-

cytes), and nestin (neural stem cells [SC]) were

used to identify cell types. Solid arrowheads indi-

cate representative doubly labeled cells. Open

arrowheads show cells labeled only with markers.

Scale bars, 20 mm.

(B) Distinct laminar projections of mossy cells

and parvalbumin-positive basket cells in the den-

tate gyrus. GCL, granule cell layer; OML, outer

molecular layer. Scale bar, 100 mm.

(C–E) Induction of [p11]-EGFP in dentate gyrus by

chronic SSRI. The dentate gyrus slices were

costained with anti-EGFP antibody (C and D) and

either anti-CRT (C) or anti-PV (D) antibodies. Scale

bars, 100 mm. EGFP intensity was quantitated in

the indicated subregions (E). Values were nor-

malized to fluorescence intensity in OML. Data

represent mean ±SEM (n = 5–6 mice per group).

*p < 0.05 and **p < 0.01, t test. Rel. Fluorescence,

relative fluorescence.

See also Figure S4.

To assess the functional consequence of p11-AnxA2 interac-

tion in the transcriptional activity of SMARCA3, we utilized a

PAI-1 promoter-driven luciferase reporter assay. First, we exam-

ined the effect of overexpression of the p11/AnxA2 complex in

N2a cells, which express low endogenous levels of p11 and

AnxA2. Cotransfection of SMARCA3 with p11 and AnxA2 poten-

tiated luciferase activity (Figure 4E). In addition, we examined the

effect of p11 gene silencing in COS-7 cells, which express rela-

tively higher endogenous levels of p11 and AnxA2. The silencing

of the p11 gene leads to the downregulation of SMARCA3-

mediated luciferase activity (Figure 4F). In the same experimental

set, we also compared WT SMARCA3 with its mutants (P35A,

F37Y) defective in the interaction with the p11/AnxA2 complex.

As expected, SMARCA3-mediated luciferase activity was abol-

ished in the two mutants (Figure 4F). Collectively, these results

suggest that p11/AnxA2 regulates transcriptional activity of

SMARCA3 by controlling the DNA-binding affinity of SMARCA3,

as well as its localization.

Regulation of the p11/AnxA2/SMARCA3 Complex inHilar Mossy Cells and Basket Cells in the Dentate Gyrusp11 mediates the actions of antidepressants (Svenningsson

et al., 2006). Chronic antidepressant administration increased

the level of p11 in the hippocampus (Figures 1D and 1E). Antide-

pressant actions including neurogenesis require several weeks

Cell 152, 831–843,

to show therapeutic effects. SMARCA3-

mediated regulation of transcription may

be associated with the therapeutic delay.

Our previous study showed that p11 is

expressed in GABAergic basket cells in

the dentate gyrus and might play a critical role in antidepres-

sant-induced hippocampal neurogenesis (Egeland et al., 2010).

We thus examined the neuronal types expressing p11 and

SMARCA3 in the dentate gyrus.

We took advantage of BAC (bacterial artificial chromosome)

transgenic mice (Heintz, 2001), in which the expression of

EGFP reporter is driven by p11 promoter activity ([p11]-EGFP).

The excellence of immunostaining analysis using anti-EGFP anti-

body enabled us to visualize the entire neuronal processes of

the neurons expressing p11. Because EGFP-positive neurons

in the BAC-[p11]-EGFPmice are doubly positive for the immuno-

staining of endogenous p11, EGFP-positive neurons represent

p11-expressing neurons (Figure S4A). p11-expressing cells as-

sessed with [p11]-EGFP signal localize in the hilus region of

the dentate gyrus (Figure 5A). To identify neuronal types for the

p11-expressing neurons, hippocampal sections from BAC-

[p11]-EGFP transgenic mice were doubly stained with EGFP

and neuronal-type markers. Notably, [p11]-EGFP reporter signal

was found to be enriched both in calretinin (CRT)-positive mossy

cells and in PV-positive basket cells, whereas the signal was

negligible in granule cells and is rarely (<5%) observed in HIPP

cells (Figure 5A). Furthermore, p11 expression is not observed

in astrocytes or neural stem cells (Figure 5A).

We next examined whether p11 expression in hippocampal

neuronal subpopulations is altered in response to antidepressant

February 14, 2013 ª2013 Elsevier Inc. 837

Figure 6. SMARCA3 Expression in Mossy

Cells and Basket Cells in the Dentate Gyrus

(A) Dentate gyrus slices from WT (top) or

SMARCA3 KO mice (bottom) stained with anti-

SMARCA3 antibody (red) and nuclear dye DraQ5

(blue). Scale bars, 100 mm.

(B) Neuronal cell types expressing SMARCA3

in the dentate gyrus. Scale bars, 20 mm. Repre-

sentative cells doubly labeled are indicated by

arrowheads.

(C) Coexpression of [p11]-EGFP and SMARCA3.

BAC-[p11]-EGFP mice were stained with anti-

EGFP and anti-SMARCA3 antibodies. Represen-

tative cells doubly labeled are indicated by

arrowheads. Scale bar, 20 mm.

(D and E) Level of SMARCA3 is not altered by FLX

administration for 2 weeks. Hippocampal lysates

were immunoblotted for SMARCA3 and b-actin

(D). mRNA level of SMARCA3 in the hippocampus

was measured using qPCR (E). Data repre-

sent mean ±SEM (n = 6–7 mice per group). ns,

nonsignificant.

(F) Formation of p11/AnxA2/SMARCA3 complex is

increased after treatment with FLX for 2 weeks.

SMARCA3 was immunoprecipitated from hippo-

campal lysates. Total lysates and immunoprecip-

itates (IPs) were immunoblotted for SMARCA3,

p11, and AnxA2.

See also Figure S5.

administration. We took advantage of the distinct laminar

segregation of the two types of p11-expressing neurons. The

mossy cells assessed by CRT expression locate their soma

and dendrites within the deep hilus and project their axonal

arbor to the inner molecular layer (IML), whereas a subpopula-

tion of basket cells visualized by PV staining locates its somas

in the hilar border and projects its axons to the granule cells

within the granular cell layer (GCL) (Figure 5B). Because p11

is mainly enriched in mossy cells and basket cells, the [p11]-

EGFP signal in the hilus and IML represents p11 promoter

activity in mossy cells, and the signal of EGFP in GCL does

for basket cell activity. BAC-[p11]-EGFP mice were treated

with FLX for 2 weeks, and [p11]-EGFP signal was examined

in the areas of hilus, IML, and GCL (Figures 5C–5E). Because

a high density of serotonergic innervation prevails in the caudal

part of the hippocampal formation, whereas the rostral part

receives only a moderate-to-weak serotonergic innervation

(Bjarkam et al., 2003; Gage and Thompson, 1980), the caudal

part (bregma �2.5 to �4.0) of the dentate gyrus was used for

the analysis. FLX treatment increased [p11]-EGFP in the den-

tate gyrus (Figures 5C and 5D). Quantitative analysis revealed

that FLX increased [p11]-EGFP intensity in hilus and IML as

well as in GCL, suggesting p11 induction in mossy cells and

basket cells (Figure 5E). p11 is also induced in the rostral part

of the hippocampus but with less potency (Figure S4B).

AnxA2, together with p11, is induced by FLX and requires

p11 for its protein stability (Figures 1D and 1E). Conversely,

p11 protein requires AnxA2 for its protein stability (data not

838 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

shown). Thus, AnxA2 is also likely to be induced in the same

neurons.

We examined the expression of SMARCA3 in the dentate

gyrus by immunohistochemistry. The specificity of the immuno-

staining was tested using SMARCA3 KO mice. SMARCA3 KO

mice were generated by targeting exons 11–13 of the SMARCA3

gene (Figure S5A). Immunoblotting (Figures S5B and S5C) and

immunohistochemistry (Figure 6A) confirmed the absence of

SMARCA3. SMARCA3 is expressed primarily in the hilar area

of the dentate gyrus (Figure 6A), where it is enriched in mossy

cells and parvalbumin-positive basket cells (Figure 6B), and is

expressed in [p11]-EGFP-positive cells (Figure 6C).

In contrast to p11 and AnxA2 (Figures 1D and 1E), the levels of

SMARCA3 protein (Figure 6D) and mRNA (Figure 6E) were not

altered after treatment with FLX. Analysis of hippocampal

lysates, following immunoprecipitation of SMARCA3, revealed

that chronic FLX administration increased the ternary complex

of p11/AnxA2/SMARCA3 by about 2.3-fold (Figure 6F). Thus,

p11 and AnxA2 induction facilitates the assembly of the p11/

AnxA2/SMARCA3 complex. Taken together, these results iden-

tify the mossy cells and the basket cells in the dentate gyrus as

primary neuronal types for SSRI/p11/SMARCA3 signaling.

SMARCA3 Is Required for Neurogenic and BehavioralResponse to Chronic SSRI Administrationp11 KO results in the loss of enhanced hippocampal neurogen-

esis and behavioral change in response to chronic antidepres-

sant treatment (Egeland et al., 2010), suggesting a crucial role

Figure 7. SMARCA3 Is Required for SSRI-

Induced Neurogenesis and Behavioral

Changes

(A and B) FLX-induced cell proliferation in WT and

SMARCA3 KO mice. WT (+/+) and SMARCA3 KO

(�/�) mice were administered VEH or FLX for

14 days and labeledwith BrdU for the last 2 hr prior

to perfusion. (A) Immunostaining with anti-BrdU.

Scale bars, 100 mm. (B) Quantitation of BrdU-

positive cells in the subgranular zone (n = 6–8 mice

per group).

(C and D) FLX-induced increase of DCX-positive

cells in WT and SMARCA3 KO mice. (C) Immu-

nostaining with anti-DCX. Scale bars, 20 mm. (D)

Quantitation of DCX-positive cells in the sub-

granular and granular zone (n = 6–8 per group).

(E and F) Survival of newborn cells in WT and

SMARCA3KOmice treatedwith VEH or FLX. BrdU

was injected for three consecutive days prior to

FLX administration for 28 days. (E) Immunostaining

with anti-BrdU and anti-NeuN. Scale bars, 100 mm.

(F) Quantitation of BrdU-positive cells.

(G and H) Behavior was assayed using (G) the NSF

paradigm after chronic administration of VEH or

FLX (4 weeks, n = 14–16 per group), or (H) the SPT

in NS or RS mice after chronic administration of

VEH or eCIT (4 weeks, n = 8–11 per group).

All data represent mean ±SEM. *p < 0.05, **p <

0.01, and ***p < 0.001, two-way ANOVA followed

by the post hoc Bonferroni test. ns, nonsignificant.

See also Figure S6 and Table S3.

for p11 and its downstream signaling pathways in those

antidepressant actions. We examined here whether SMARCA3,

as a downstream signaling molecule of p11, might mediate

chronic antidepressant-induced hippocampal neurogenesis

and behaviors.

Adult neurogenesis is controlled by multistep processes

including the proliferation of neural progenitors, differentiation,

and maturation into functional granule neurons (Ming and

Song, 2011). We examined proliferation using in vivo BrdU

labeling of the neural progenitors in the S phase in WT and

SMARCA3 KO mice. A significant increase was observed in

the number of BrdU-labeled neural progenitors in WT mice

Cell 152, 831–843,

treated with FLX (160.6% ± 13.7% of

VEH group; p < 0.01), but not in

SMARCA3 KO mice (Figures 7A and

7B). We also carried out an alternative

assay measuring an endogenous mitotic

marker Ki-67. Ki-67-positive cells were

significantly increased in response to

FLX in WT mice (143.8% ± 8.6%; p <

0.01), but not in SMARCA3 KO mice

(Figures S6A and S6B), confirming the

results of the BrdU assay.

We next analyzed the expression level

of doublecortin (DCX), a marker for

immature neurons, which represents a

snapshot of newborn cells undergoing

neuronal maturation and differentiation

(Couillard-Despres et al., 2005). DCX immunofluorescence

signal was increased by chronic FLX administration in WT mice

(218.4% ± 18.0%; p < 0.001), but not in SMARCA3 KO mice

(Figures 7C and 7D). Chronic FLX administration promotes the

newborn cells to survive at postmitotic stages and also to

becomemature neurons (Wang et al., 2008). Chronic FLX admin-

istration greatly increased the survival of the BrdU-labeled

newborn cells in WT mice (288.2% ± 21.1%; p < 0.01), but the

effect of FLX was attenuated in SMARCA3 KO mice (196.0% ±

20.3%; p < 0.05) (Figures 7E and 7F). Taken together, these

results indicate that SMARCA3 contributes to multiple stages

of antidepressant-stimulated neurogenesis, and this phenotype

February 14, 2013 ª2013 Elsevier Inc. 839

of SMARCA3 KOmice is reminiscent of that observed in p11 KO

mice (Egeland et al., 2010).

Next, we investigated the functional significance of the p11/

AnxA2/SMARCA3 complex for SSRI-induced behavioral

changes. WT and SMARCA3 KOmice did not display a baseline

difference in locomotor activity (open field test, Figure S6C),

depressive behaviors (tail suspension test [TST] and sucrose

preference test [SPT], Figures S6D and S6E), or anxiety behav-

iors (light/dark and elevated plus maze tests, Figures S6F and

S6G). Novelty-suppressed feeding (NSF) is believed to represent

depression-like and anxiety behavior, and this test is commonly

used to measure the chronic effect of antidepressants (David

et al., 2009; Surget et al., 2011). p11 KO mice are refractory to

behavioral change in response to chronic FLX administration in

the NSF test (Egeland et al., 2010; Schmidt et al., 2012). In the

same model, we analyzed the behavior of SMARCA3 KO mice.

Chronic treatment with FLX shortened latency to feed compared

to VEH treatment in WT mice (VEH versus FLX, 279 ± 13 versus

187 ± 20; p < 0.01), but not in SMARCA3 KO mice (Figure 7G),

as with p11 KO mice (Egeland et al., 2010). Neither FLX nor

SMARCA3 KO caused any significant effect on home cage

feeding (Figure S6H) or body weight (data not shown).

Anhedonia is a core symptom of human depression (Berton

and Nestler, 2006), and a chronic stress-induced decrease in

sucrose preference in rodents is regarded as a sign of a hedonic

deficit (Katz, 1982), which can be treated with chronic SSRI. We

examined the effect of chronic administration of escitalopram

(eCIT, another SSRI) on the behaviors of nonstressed (NS)

or restraint-stressed (RS) mice in the SPT. Daily restraint for

14 days induced a reduction of sucrose consumption in both

WT (NS plus VEH versus RS plus VEH, 100% ± 12% versus

68% ± 8%, respectively) and SMARCA3 KO mice (NS plus

VEH versus RS plus VEH, 97% ± 6% versus 61% ± 8%, respec-

tively). Poststress treatment with eCIT recovered sucrose

consumption to the normal level in WTmice (RS plus VEH versus

RS plus eCIT, 68% ± 8% versus 99% ± 13%, respectively), but

not in SMARCA3 KO mice (Figure 7H). No differences were

observed between WT and SMARCA3 KO mice as well as

among the experimental groups used for the SPT with regard

to their locomotor activities (Figure S6I). Notably, the behavioral

despair in the TST is known to be treated by acute SSRI, whereas

the therapeutic effect of SSRI in the NSF and SPT requires

chronic treatment. SMARCA3 KO did not alter the effect of acute

administration of eCIT in the TST (Figure S6J). Taken together,

our results suggest a crucial role for SMARCA3 in p11-depen-

dent neurogenic and behavioral response to chronic antidepres-

sant administration.

DISCUSSION

Molecular Interaction of p11/AnxA2 with SMARCA3p11 and AnxA2 cooperate to create a unique binding pocket, but

the optimal binding condition is not achieved without conforma-

tional changes associated with target binding. Upon interaction

with the SMARCA3 peptide, both Asp60 in p11 and Ser12

in AnxA2 flip inward toward the peptide-binding groove,

forming additional intermolecular hydrogen bonds (Figures

3D–3F). Similar intermolecular contacts were also observed in

840 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

the AHNAK1 complex (Figures S2D–S2F). These findings

support an induced-fit model for the assembly of the p11/

AnxA2/SMARCA3 complex. Among the key residues found in

the binding pocket, the role of Ser12 (AnxA2) is of particular

interest because this residue is regulated by phosphorylation

(Jost and Gerke, 1996), which may interfere with high-affinity

binding of ternary targets.

The current study identifies a binding motif, which is repre-

sented as f-P-#-F-X-F, and can be used for in silico analysis

to identify binding targets of p11/AnxA2. Our binding motif is

different from the p11-binding sequences previously reported

for TRPV5/6 (VATTV) (van de Graaf et al., 2003) and TASK1

(RRSSV) (Girard et al., 2002). It is possible that the p11/AnxA2

heterotetramer has additional binding pocket(s) on the surface.

Regulation of SMARCA3 Function by Interaction withthe p11/AnxA2 ComplexMost chromatin remodelers, including four well-characterized

subfamily members (SWI/SNF, INO80/SWR1, ISWI, and CHD),

form a large multisubunit complex with a core ATPase motor

subunit and unique accessory/regulatory subunits (Hargreaves

and Crabtree, 2011). The core subunit displays DNA- and nucle-

osome-dependent ATPase activity, and the accessory/regula-

tory subunits are essential for the function of the core ATPase

subunit, by facilitating the interaction with the transcriptional

regulatory factors, mediating the indirect binding to DNA and/

or modified histones, and targeting the complex to subnuclear

locations (Mohrmann andVerrijzer, 2005). SMARCA3, a relatively

uncharacterized member of the SWI/SNF protein family, is

composed of multiple functional domains, including a DNA-

binding domain, a SWI2/SNF2 ATPase domain, a RING-type

zinc finger domain for the binding to RFBP, a Type IV P-type

ATPase, and a C-terminal domain for the binding of transcription

factors such as Sp1, Sp3, Egr1, and cRel (Debauve et al., 2008)

(Figure 2D). Although p11 and AnxA2 stabilize each other as

structural components of a protein complex (Figure 1), the levels

of p11 and AnxA2 are not altered in SMARCA3 KOmice (Figures

S5B and S5C), and the level of SMARCA3 is not altered in p11

KO mice (data not shown). Thus, p11 and AnxA2 act as regula-

tory proteins but are not likely structural core components of

SMARCA3. Notably, the DNA-binding domain at the N-terminal

side of the ATPase domain mediates the sequence-dependent

binding of SMARCA3 to the target DNA (Debauve et al., 2008).

p11/AnxA2 facilitates the DNA-binding affinity of SMARCA3

(Figures 4A and 4B). Most chromatin-remodeling complexes,

including SMARCA3, display DNA- and/or nucleosome-depen-

dent ATPase activity (Hargreaves and Crabtree, 2011). There-

fore, it is conceivable that the enhanced DNA binding of

SMARCA3, upon the interaction with p11/AnxA2, may lead to

the activation of SMARCA3 to initiate ATP-dependent chromatin

remodeling of the target genes. Thus, p11/AnxA2 binding to

SMARCA3 would open the chromatin structure and recruit

specific transcription factors bound to the C-terminal domain

of SMARCA3 to the specific locus of the genomic DNA.

Critical nuclear events such as gene expression, replica-

tion, and repair processing occur at a distinct subnuclear

region, the nuclear matrix, which is composed of nuclear

lamins, nuclear actin/actin-related proteins, and phospholipids

(Barlow et al., 2010; Zaidi et al., 2007). Indeed, key gene regula-

tory machineries such as transcription factors, chromatin-

remodeling complexes, RNA polymerase II, and processing

factor SC35 are associated with nuclear matrix structures (Zaidi

et al., 2007). p11/AnxA2 mediates the subnuclear targeting of

SMARCA3 (Figures 4C and 4D). The subnuclear targeting of

SMARCA3 is likely controlled by intrinsic phospholipid- and

actin-binding properties of AnxA2 (Gerke et al., 2005). Our

crystal structure of p11/full-length AnxA2/SMARCA3 peptide

visualized the spatial organization of each component in the

ternary complex, in which the p11 dimer is ideally positioned in

the core of the complex to link SMARCA3 to AnxA2 (Figure 3G).

Thus, our current model is reminiscent of the role of p11/AnxA2

in mediating the membrane translocation of AHNAK1 (Benaud

et al., 2004). Both SMARCA3 and AHNAK1 use the interaction

with p11/AnxA2 to localize properly to the distinct subcellular

sites where they become functionally active.

p11/AnxA2/SMARCA3 Complex in AntidepressantActionp11/AnxA2/SMARCA3 complex-mediated hippocampal neuro-

genesis may contribute to the behavioral response to SSRIs.

However, the involvement of hippocampal neurogenesis in

depression andantidepressant-inducedbehavioral change is still

controversial (Hanson et al., 2011). Such neurogenesis has been

associated with mood, stress responses, and antidepressant

effects in some studies (Santarelli et al., 2003; Snyder et al.,

2011; Surget et al., 2011), but not in others (Bessa et al., 2009).

p11 and SMARCA3 are enriched in mossy cells and basket

cells but not detectable in neural progenitor cells. Thus, the

regulation of SSRI-induced neurogenesis by the p11/AnxA2/

SMARCA3 signaling pathway is non-cell autonomous. In fact,

GABAergic interneurons play an important role in the differentia-

tion, development, and integration of newborn neurons (Ge et al.,

2006; Tozuka et al., 2005). Specifically, interneuronal GABA

release is thought to act directly on neural progenitor cells and,

due to an age-specific increase in intracellular chloride levels in

these young cells, causes an atypical depolarization that subse-

quently affects neurogenic processes (Tozuka et al., 2005). The

synaptic regulation of granule cells by basket and mossy cells

may contribute to the neurogenic and behavioral responses to

SSRIs. It has been suggested that, upon activation by local

inputs from the granule cells in the dentate gyrus, glutamatergic

mossy cells primarily provide excitatory feedback to the granule

cells through their axonal projections to the IML and also provide

excitatory drive to local GABAergic HIPPs in the hilus (Henze and

Buzsaki, 2007; Scharfman, 1995). The basket cells primarily

provide feedforward inhibition to the granule cells in response

to excitatory inputs from the entorhinal cortex and feedback inhi-

bition to the granule cells in response to excitatory inputs from

the granule cells (Houser, 2007). Thus, despite their relatively

low abundance (�33 104mossy cells and 0.53 104 basket cells

versus 1 3 106 granule cells in rat), the mossy cells and basket

cells regulate the flow of extrinsic input to the dentate gyrus

through modulation of the activity of the granule cells and/or

the HIPPs, and such regulation is important to generate the

distinct oscillation patterns of neuronal excitation in the dentate

gyrus (Amaral et al., 2007; Henze and Buzsaki, 2007).

In this study, we have identified and characterized a nuclear

protein complex mediating chronic actions of SSRIs. Many inter-

esting questions have been raised as a result of this study: What

are the roles of p11 and SMARCA3 in the excitability and

synaptic transmission of basket cells and mossy cells? Which

genes are regulated by the p11/AnxA2/SMARCA3 complex in

basket cells and mossy cells? Which of these genes contribute

to the neurogenic and behavioral responses to antidepressants?

Current and future studies of the p11/AnxA2/SMARCA3

pathway should contribute not only to our understanding of

SSRI actions but also provide molecular and cellular targets for

the development of advanced therapeutics for mood and anxiety

disorders.

EXPERIMENTAL PROCEDURES

Generation of Transgenic Mice

All procedures involving animals were approved by the Rockefeller University

Institutional Animal Care and Use Committee and were in accordance with the

National Institutes of Health guidelines. The constitutive SMARCA3 KO mice

were generated in Taconic-Artemis (Germany) and maintained at The Rock-

efeller University. The BAC-[p11]-EGFP transgenic line (GENSAT; Clone No.

HC85) was provided by GENSAT (Gong et al., 2002). The mouse breeding

and drug treatment methods are in the Extended Experimental Procedures.

Crystallization and Structure Determination of p11/AnxA2/

SMARCA3 and p11/AnxA2/AHNAK1 Complexes

Details of protein preparations, protein expression, purification, crystallization

conditions, data collection, and refinement are included in Extended Experi-

mental Procedures.

Preparation of Nuclear Matrix Fraction

Cells were incubated with cell-permeable crosslinker, DSP (1 mM), and

extracted with CSK buffer containing 0.5% NP40. After chromatin digestion

with DNase I and then elution with 0.25 M (NH4)2SO4, the nuclear matrix

fraction was scraped for the biochemical assay or fixed with 2% parafor-

maldehyde for immunocytochemistry (detailed description is in Extended

Experimental Procedures).

Histological Methods

Immunostaining was carried out using the standard free-floating method.

Detailed description of antibody preparation, antigen retrieval, image acquisi-

tion, and quantification is in Extended Experimental Procedures.

Behavioral Analysis

Mood and anxiety-related behaviors (NSF, SPT, TST, light-dark box test, and

elevated plus maze) and locomotor activity (open field test) were tested as

described in Extended Experimental Procedures.

Data Analysis and Statistics

All data are presented as mean ±SEM. Two group comparisons were done by

two-tailed, unpaired Student’s t test. Multiple group comparisons were as-

sessed using a one-way or two-way ANOVA, followed by the post hoc

Newman-Keuls test or Bonferroni test, respectively, when significant main

effects or interactions were detected. Statistical significance was set at p <

0.05 level. Summary of statistical analysis for animal experiments is included

as Table S3.

ACCESSION NUMBERS

Coordinates and structure factors for p11-AnxA2 peptide cassette in complex

with AHNAK1 peptide (PDB 4HRG) and SMARCA3 peptide (PDB 4HRH) and

full-length p11/AnxA2 heterotetramer in complex with SMARCA3 peptide

(PDB 4HRE) have been deposited in the RCSB Protein Data Bank.

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. 841

SUPPLEMENTAL INFORMATION

Supplemental Information includes Extended Experimental Procedures, six

figures, and three tables and can be found with this article online at http://

dx.doi.org/10.1016/j.cell.2013.01.014.

ACKNOWLEDGMENTS

This work was supported by DOD/USAMRAA Grant W81XWH-09-1-0392 to

Y.K.; DOD/USAMRAA Grant W81XWH-09-1-0402 to P. Greengard; the JPB

Foundation to P. Greengard; the Fisher Center Foundation to P. Greengard;

NIH grants (MH090963, DA10044, and AG09464) to Y.K. and P. Greengard;

and the Maloris Foundation and the Abby Rockefeller Mauze Trust to D.J.P.

We thank the staff at beamline 24ID-C of the Advanced Photon Source at

the Argonne National Laboratory and beamline X29 of the National Synchro-

tron Light Source at the Brookhaven National Laboratory for assistance with

data collection. We thank Daesoo Kim, Eric Schmidt, Jennifer Wagner-

Schmidt, and YotamSagi for their helpful advice and discussion, and Elisabeth

Griggs for technical assistance.Wewould like to thank Ji-Eun Kim for themass

spectrometry analysis. Finally, we would like to acknowledge Rada Norinsky

and the Rockefeller University Transgenics Services Laboratory for their excel-

lent IVF services; and Henry Zebroski III and Nagarajan Chandramouli from

The Rockefeller University Proteomics Resource Center.

Received: May 16, 2012

Revised: September 14, 2012

Accepted: January 8, 2013

Published: February 14, 2013

REFERENCES

Amaral, D.G., Scharfman, H.E., and Lavenex, P. (2007). The dentate gyrus:

fundamental neuroanatomical organization (dentate gyrus for dummies).

Prog. Brain Res. 163, 3–22.

Barlow, C.A., Laishram, R.S., and Anderson, R.A. (2010). Nuclear phosphoino-

sitides: a signaling enigma wrapped in a compartmental conundrum. Trends

Cell Biol. 20, 25–35.

Benaud, C., Gentil, B.J., Assard, N., Court, M., Garin, J., Delphin, C., and

Baudier, J. (2004). AHNAK interaction with the annexin 2/S100A10 complex

regulates cell membrane cytoarchitecture. J. Cell Biol. 164, 133–144.

Berton, O., and Nestler, E.J. (2006). New approaches to antidepressant drug

discovery: beyond monoamines. Nat. Rev. Neurosci. 7, 137–151.

Bessa, J.M., Ferreira, D., Melo, I., Marques, F., Cerqueira, J.J., Palha, J.A.,

Almeida, O.F., and Sousa, N. (2009). The mood-improving actions of antide-

pressants do not depend on neurogenesis but are associated with neuronal

remodeling. Mol. Psychiatry 14, 764–773.

Bjarkam, C.R., Sørensen, J.C., and Geneser, F.A. (2003). Distribution and

morphology of serotonin-immunoreactive axons in the hippocampal region

of the New Zealand white rabbit. I. Area dentata and hippocampus. Hippo-

campus 13, 21–37.

Couillard-Despres, S., Winner, B., Schaubeck, S., Aigner, R., Vroemen, M.,

Weidner, N., Bogdahn, U., Winkler, J., Kuhn, H.G., and Aigner, L. (2005).

Doublecortin expression levels in adult brain reflect neurogenesis. Eur. J.

Neurosci. 21, 1–14.

Das, S., Shetty, P., Valapala, M., Dasgupta, S., Gryczynski, Z., and

Vishwanatha, J.K. (2010). Signal transducer and activator of transcription 6

(STAT6) is a novel interactor of annexin A2 in prostate cancer cells. Biochem-

istry 49, 2216–2226.

David, D.J., Samuels, B.A., Rainer, Q., Wang, J.W., Marsteller, D., Mendez, I.,

Drew, M., Craig, D.A., Guiard, B.P., Guilloux, J.P., et al. (2009). Neurogenesis-

dependent and -independent effects of fluoxetine in an animal model of

anxiety/depression. Neuron 62, 479–493.

Debauve, G., Capouillez, A., Belayew, A., and Saussez, S. (2008). The

helicase-like transcription factor and its implication in cancer progression.

Cell. Mol. Life Sci. 65, 591–604.

842 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

Ding, H., Descheemaeker, K., Marynen, P., Nelles, L., Carvalho, T.,

Carmo-Fonseca, M., Collen, D., and Belayew, A. (1996). Characterization of

a helicase-like transcription factor involved in the expression of the human

plasminogen activator inhibitor-1 gene. DNA Cell Biol. 15, 429–442.

Egeland, M., Warner-Schmidt, J., Greengard, P., and Svenningsson, P. (2010).

Neurogenic effects of fluoxetine are attenuated in p11 (S100A10) knockout

mice. Biol. Psychiatry 67, 1048–1056.

Gage, F.H., and Thompson, R.G. (1980). Differential distribution of norepi-

nephrine and serotonin along the dorsal-ventral axis of the hippocampal

formation. Brain Res. Bull. 5, 771–773.

Ge, S., Goh, E.L., Sailor, K.A., Kitabatake, Y., Ming, G.L., and Song, H. (2006).

GABA regulates synaptic integration of newly generated neurons in the adult

brain. Nature 439, 589–593.

Gerke, V., Creutz, C.E., and Moss, S.E. (2005). Annexins: linking Ca2+ signal-

ling to membrane dynamics. Nat. Rev. Mol. Cell Biol. 6, 449–461.

Girard, C., Tinel, N., Terrenoire, C., Romey, G., Lazdunski, M., and Borsotto,

M. (2002). p11, an annexin II subunit, an auxiliary protein associated with the

background K+ channel, TASK-1. EMBO J. 21, 4439–4448.

Gong, S., Yang, X.W., Li, C., and Heintz, N. (2002). Highly efficient modification

of bacterial artificial chromosomes (BACs) using novel shuttle vectors contain-

ing the R6Kgamma origin of replication. Genome Res. 12, 1992–1998.

Hanson, N.D., Owens, M.J., and Nemeroff, C.B. (2011). Depression, antide-

pressants, and neurogenesis: a critical reappraisal. Neuropsychopharmacol-

ogy 36, 2589–2602.

Hargreaves, D.C., and Crabtree, G.R. (2011). ATP-dependent chromatin

remodeling: genetics, genomics and mechanisms. Cell Res. 21, 396–420.

Heintz, N. (2001). BAC to the future: the use of bac transgenic mice for neuro-

science research. Nat. Rev. Neurosci. 2, 861–870.

Henze, D.A., and Buzsaki, G. (2007). Hilar mossy cells: functional identification

and activity in vivo. Prog. Brain Res. 163, 199–216.

Houser, C.R. (2007). Interneurons of the dentate gyrus: an overview of

cell types, terminal fields and neurochemical identity. Prog. Brain Res. 163,

217–232.

Jost, M., and Gerke, V. (1996). Mapping of a regulatory important site for

protein kinase C phosphorylation in the N-terminal domain of annexin II.

Biochim. Biophys. Acta 1313, 283–289.

Katz, R.J. (1982). Animal model of depression: pharmacological sensitivity of

a hedonic deficit. Pharmacol. Biochem. Behav. 16, 965–968.

Kube, E., Becker, T., Weber, K., and Gerke, V. (1992). Protein-protein interac-

tion studied by site-directed mutagenesis. Characterization of the annexin

II-binding site on p11, a member of the S100 protein family. J. Biol. Chem.

267, 14175–14182.

Liu, J., and Vishwanatha, J.K. (2007). Regulation of nucleo-cytoplasmic

shuttling of human annexin A2: a proposed mechanism. Mol. Cell. Biochem.

303, 211–220.

Ming, G.L., and Song, H. (2011). Adult neurogenesis in the mammalian brain:

significant answers and significant questions. Neuron 70, 687–702.

Mohrmann, L., and Verrijzer, C.P. (2005). Composition and functional

specificity of SWI2/SNF2 class chromatin remodeling complexes. Biochim.

Biophys. Acta 1681, 59–73.

Rando, O.J., Zhao, K., Janmey, P., and Crabtree, G.R. (2002). Phosphatidyli-

nositol-dependent actin filament binding by the SWI/SNF-like BAF chromatin

remodeling complex. Proc. Natl. Acad. Sci. USA 99, 2824–2829.

Rescher, U., and Gerke, V. (2008). S100A10/p11: family, friends and functions.

Pflugers Arch. 455, 575–582.

Rety, S., Sopkova, J., Renouard, M., Osterloh, D., Gerke, V., Tabaries, S.,

Russo-Marie, F., and Lewit-Bentley, A. (1999). The crystal structure of a

complex of p11 with the annexin II N-terminal peptide. Nat. Struct. Biol. 6,

89–95.

Rezvanpour, A., Phillips, J.M., and Shaw, G.S. (2009). Design of high-affinity

S100-target hybrid proteins. Protein Sci. 18, 2528–2536.

Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F., Dulawa, S.,

Weisstaub, N., Lee, J., Duman, R., Arancio, O., et al. (2003). Requirement of

hippocampal neurogenesis for the behavioral effects of antidepressants.

Science 301, 805–809.

Scharfman, H.E. (1995). Electrophysiological evidence that dentate hilar

mossy cells are excitatory and innervate both granule cells and interneurons.

J. Neurophysiol. 74, 179–194.

Schmidt, E.F., Warner-Schmidt, J.L., Otopalik, B.G., Pickett, S.B., Greengard,

P., and Heintz, N. (2012). Identification of the cortical neurons that mediate

antidepressant responses. Cell 149, 1152–1163.

Snyder, J.S., Soumier, A., Brewer, M., Pickel, J., and Cameron, H.A. (2011).

Adult hippocampal neurogenesis buffers stress responses and depressive

behaviour. Nature 476, 458–461.

Surget, A., Tanti, A., Leonardo, E.D., Laugeray, A., Rainer, Q., Touma, C.,

Palme, R., Griebel, G., Ibarguen-Vargas, Y., Hen, R., and Belzung, C. (2011).

Antidepressants recruit new neurons to improve stress response regulation.

Mol. Psychiatry 16, 1177–1188.

Svenningsson, P., Chergui, K., Rachleff, I., Flajolet, M., Zhang, X., El Yacoubi,

M., Vaugeois, J.M., Nomikos, G.G., and Greengard, P. (2006). Alterations in

5-HT1B receptor function by p11 in depression-like states. Science 311,

77–80.

Tozuka, Y., Fukuda, S., Namba, T., Seki, T., and Hisatsune, T. (2005).

GABAergic excitation promotes neuronal differentiation in adult hippocampal

progenitor cells. Neuron 47, 803–815.

van de Graaf, S.F., Hoenderop, J.G., Gkika, D., Lamers, D., Prenen, J.,

Rescher, U., Gerke, V., Staub, O., Nilius, B., and Bindels, R.J. (2003).

Functional expression of the epithelial Ca(2+) channels (TRPV5 and TRPV6)

requires association of the S100A10-annexin 2 complex. EMBO J. 22, 1478–

1487.

Wang, J.W., David, D.J., Monckton, J.E., Battaglia, F., and Hen, R. (2008).

Chronic fluoxetine stimulates maturation and synaptic plasticity of adult-

born hippocampal granule cells. J. Neurosci. 28, 1374–1384.

Warner-Schmidt, J.L., Chen, E.Y., Zhang, X., Marshall, J.J., Morozov, A.,

Svenningsson, P., and Greengard, P. (2010). A role for p11 in the anti-

depressant action of brain-derived neurotrophic factor. Biol. Psychiatry 68,

528–535.

Zaidi, S.K., Young, D.W., Javed, A., Pratap, J., Montecino, M., van Wijnen, A.,

Lian, J.B., Stein, J.L., and Stein, G.S. (2007). Nuclear microenvironments in

biological control and cancer. Nat. Rev. Cancer 7, 454–463.

Zhao, K., Wang, W., Rando, O.J., Xue, Y., Swiderek, K., Kuo, A., and Crabtree,

G.R. (1998). Rapid and phosphoinositol-dependent binding of the SWI/SNF-

like BAF complex to chromatin after T lymphocyte receptor signaling. Cell

95, 625–636.

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. 843

Supplemental Information

EXTENDED EXPERIMENTAL PROCEDURES

Animal Breeding and Drug TreatmentsWe produced the progeny for each line using in vitro fertilization (IVF) and embryo transfer (ET) techniques, to produce a number of

animals sufficient for the behavioral tests and other animal experiments. SMARCA3KOmice were bred using heterozygousmale and

female. We carried out all the animal experiments using age (10-15 weeks)- and gender (male)-matched littermates. The BAC-[p11]-

eGFP transgenic mice were bred against C57BL/6 mice (Taconics) to obtain hemizygote and used at an adult age (12-18 weeks).

Food and water were provided ad libitum. For chronic drug treatments, mice were housed 1-2 per cage and SSRIs (fluoxetine or es-

citalopram) were administered either by the subcutaneous implantation of time-release fluoxetine pellets (Innovative Research of

America; Cat. No. X-999; 10 mg/kg/day) for 2 weeks, or by the drinking water in a mixture of fluoxetine (0.160 mg/ml) and b-cyclo-

dextrine (0.45%) up to 4 weeks, or daily I.P. injection (Escitalopram, 10 mg/kg/day) up to 4 weeks. Control groups were administered

placebo pellet or vehicle only (Covington et al., 2009; Vialou et al., 2010). Drug solution in the amber water bottles was replaced with

fresh drug solution every 3 days.

Plasmid Construction and MutagenesisTo generate the plasmid encoding p11 (pIRESneo-p11-Flag-HA), the p11 CDS with Flag-HA epitopes in the C terminus was subcl-

oned into pIRES-neo3 vector (Clontech) at the two restriction enzyme cleavage sites (NheI, NotI) by PCR tagging. The WT p11

construct was used as a template to generate the mutants including C83Q, C83S and D60A using the mutagenesis kit (QuikChange

II, Agilent). The plasmid encodingAnxA2 (pEGFP-N3-AnxA2-myc-His) was generated by cloning theAnxA2CDSwithmyc-6xHis plus

stop codon into pEGFP-N3 vector using PCR tagging with enzyme cleavage (BamHI, SmaI/EcoRV). The WT construct was used as

a template to generate the mutants including L8A, L11A, and L13A. The plasmids encoding SMARCA3 (pAAV-CBA-WPRE-

SMARCA3-V5) were generated by cloning the SMARCA3 CDS into pAAV-CBA-WPRE vector (In lab) with insertion of V5 epitope

in the C terminus using PCR tagging with enzyme cleavage (NheI, NotI). The WT SMARCA3-V5 construct was used as a template

to generate the mutants including P35A and F37Y. For the PAI1-promoter-luciferase construct, PAI-1 promoter region ranging

from �791 to +38 was amplified using primer pairs (50-ggt acc agg ctg ctg tac tgg ttc ttg-30/50-gat atc gtc ctc ggg gct ctg ct-30)from the mouse genomic DNA, cloned into pCR2.1 TOPO vector. The sequence-verified insert was cloned into pGL4.10 luciferase

reporter vector (Promega) with enzyme cleavage (KpnI, EcoRV).

For recombinant protein production in the bacterial system, p11 CDS was amplified using the primer pair (50-gcc cat atg cca tcc

caa atg gag cac gcc-30/50-cgg gaattc cta ttt ctt ccc ctt ctg ctt cat-30), cloned into both pGEX-6P1 vector (GE healthcare) and pET29b+

vector (Novagen) for GST-tag (pGEX-6P1-p11) and 6xHis tag (pET29b-p11), respectively by using the restriction enzyme cut (NdeI,

EcoRI). For recombinant AnxA2 production, AnxA2CDSwas amplified using the primer pair (50-gct cat atg tct act gtc cac gaa atc ctg

tg-30/50-cgt gcg gcc gct cag tca tcc cca cca cac ag-30), cloned into pET29b+ vector (Novagen) to generate the pET29b+AnxA2

construct (NdeI, NotI). For expression of recombinant SMARCA3 protein, the coding region for the N-terminal fragment of SMARCA3

(Met1-350) was amplified using the primer pair (50-ccctctagaaataatgtcctggatgttcaagagggat-30/50-ggcaagctttgccttttcactggtattgtttcc-30) and inserted into pET29b(+) vector (XbaI, HindIII).

SMARCA3 deletion constructs were generated using PCR with the primer pairs shown below, followed by cloning into pCR2.1

TOPO vector (Invitrogen), and used for in vitro translation (Figures S1B–S1D). Primer pairs (forward/reverse) for PCR: SMARCA3

full length (cacc agg cgc tcc tct tgt catc/ taa gtc aat taa tgt tct gat ttc at), D854-C (cacc agg cgc tcc tct tgt catc/ttt tat gtt ggg att

ctt ctt tct), D744-C (cacc agg cgc tcc tct tgt catc/ctt ctt tct cag ttc ttc agg tgt), D631-C (cacc agg cgc tcc tct tgt catc/aat tcc tcc

ttc atc tcc cat t), D231-C (cacc agg cgc tcc tct tgt catc/cat ttc atg ggt ttt atc atc ttc), DN1-122 (cac caa agt aaa caa tgt gaa tgg

aaa tc/taa gtc aat taa tgt tct gat ttc at), DN1-155 (cac cat gaa tag aaa agc ggt ttc aga tca gt/taa gtc aat taa tgt tct gat ttc at),

DN1-230 (cac cat ggt tct agc ttg gat ggt gtc/taa gtc aat taa tgt tct gat ttc at).

Quantitative RT-PCRTotal RNA was extracted with standard method using Trixol-LS reagent (Invitrogen), precipitated with sodium acetate, treated with

DNase I (QIAGEN), and purified with RNeasy kit (QIAGEN). RNA quantity was measured with a Nanodrop 1000 spectrophotometer

and quality was assayed with RNA nanochip kit (Agilent 2100 Bioanalyzer). cDNA was synthesized from 1 mg of total RNA using

ImProm-II Reverse Transcription System (Promega) with Oligo dT primer according to manufacturer’s protocol. 10-50 ng of cDNA

was used for each qPCR reaction and all samples were run in triplicate. Q-PCR was carried out in Applied Biosystems 7900HT

system following standard cycling conditions (50�C for 2 min, 95�C for 10 min, then 40 cycle of 95�C for 15 s and 60�C for

1 min). All Taqman assays were purchased from Applied Biosystems as follows. Tagman assay ID for each gene; p11

(Mm00501457_m1), AnxA2 (Mm00500307_m1), SMARCA3 (Mm00447302_m1). All data were normalized to TaqMan Rodent

GAPDH Control, and relative expression levels between conditions were calculated by the comparative CT (2-DDCT) method, and

all statistical analyses were done using Student’s t test.

Western Blot AnalysisCell pellets were lysed in lysis buffer (30 mM HEPES, pH7.2, 140 mM NaCl, 1 mM EDTA, 2 mM MgCl2 and 0.2% Triton X-100) sup-

plemented with a protease inhibitor cocktail (Complete-EDTAfree; Roche) and a phosphatase inhibitor cocktail (PhosStop, Roche).

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. S1

The lysates were incubated on ice for 30 min with intermittent vortex mixing every 5 min. The hippocampal tissues were added in 5

times excess volume of the lysis buffer, sonicated with probe-type sonicator (Branson) for 10 s, twice. The cell or tissue lysate was

centrifuged and the protein level in the supernatant was measured by BCA. The protein samples were boiled in standard protein

sample buffer, and subjected to SDS-PAGE followed by protein transfer onto a nitrocellulose membrane. Immunoblotting was per-

formed with a standard protocol using the following antibodies: anti-p11 (mouse monoclonal, 1:1000, BD bioscience), anti-p11 (goat

polyclonal, 1:200, R&D systems), anti-SMARCA3 (goat polyclonal, 1:200, NOVUS), anti-SMARCA3 (rabbit polyclonal, 1:500), anti-

AnxA2 (mouse monoclonal, Santa Cruz), anti-myc (mouse monoclonal, 9E10, Sigma-Aldrich), anti-V5 (rabbit polyclonal, Chemicon),

anti-GAPDH (mouse monoclonal, Chemicon), anti-AnxA1 (rabbit polyclonal, Zymed), anti-S100B (mouse monoclonal, BD transduc-

tion lab), anti-SPT6 (rabbit polyclonal, Bethyl lab), anti-AHNAK1 (mouse monoclonal, NOVUS), anti-lamin-B (rabbit polyclonal,

Abcam), Anti-HA (rat monoclonal, Roche), and anti-actin (rabbit polyclonal, Cytoskeleton).

ImmunoprecipitationLysates from cultured cells or brain tissue were prepared as described above. After centrifugation at 12,000 x g for 10min, the super-

natant was precleared with protein A/G agarose beads (Pierce), and incubated with antibody-coupled agarose beads for 3 hr at 4�C,with constant rotation. After washing three timeswith the lysis buffer and one timewith the lysis buffer without Triton X-100, the bound

proteins were eluted by boiling in the protein sample buffer for 5 min, and subjected to SDS-PAGE. Antibody-coupled beads were as

follows: anti-p11 antibody-coupled protein A/G agarose gel (mouse polyclonal anti-p11 antibody), anti-SMARCA3 antibody-coupled

protein A/G agarose gel (rabbit polyclonal anti-SMARCA3 antibody I or II), anti-Flag affinity agarose gel (mouse monoclonal anti-Flag

M2 antibody; Sigma-Aldrich), anti-Myc affinity agarose gel (goat polyclonal anti-Myc antibody; Bethyl laboratory), and anti-V5 affinity

agarose gel (mouse monoclonal anti-V5 antibody, Sigma-Aldrich).

Antibody GenerationMouse polyclonal antisera against mouse p11were obtained by immunizing recombinantmouse p11 into p11 homozygous KOmice.

p11 homozygous KO mice at 6-7 weeks old were immunized subcutaneously with recombinant p11 (50 mg) protein emulsified with

TiterMax adjuvant (CytRx Co.) according to themanufacturer’s instruction. Themice were boostedmore than 6 times with emulsified

antigen at 3 week intervals prior to the final boosting with the intravenous injection of soluble antigen (100 mg). A week later, sera were

collected with retro-orbital bleeding under deep anesthesia, and tested by immunoblotting and immunoprecipitation. Rabbit poly-

clonal antibody was purified as reported (Czernik et al., 1991) with minor modification. The 14-mer peptides derived from SMARCA3

(rabbit polyclonal-I; DIIPPDDFLTSDEE) and (rabbit polyclonal-II; EERKIYQSVKNEGK) were synthesized with C-terminal cysteine

residue, coupled to the KLH protein using Sulfo-MBS cross-linker (Pierce Biotech). The peptide-KLH conjugates were desalted using

a PD-10 column (GE healthcare), emulsified with TiterMax adjuvant (CytRx Co.), and immunized into rabbits (Cocalico Inc.). The IgG

fraction was purified using protein A/G affinity chromatography from the sera of the immunized rabbit and the specific antibody frac-

tion was further purified using peptide-coupled resins (SulfoLink Coupling Resins and Immobilization Kits, Pierce Biotech).

In Vitro Binding AnalysisGST-p11 Pull-Down Assay

GST or GST-p11 (5 mg) was immobilized on GSH-agarose (GE healthcare), and preincubated with AnxA2 (15 mg or as indicated) for

1 hr at 4�C to produce p11/AnxA2 complex in the binding buffer (30mMHEPES, pH7.2, 140mMNaCl, 1 mMEDTA, 2mMMgCl2 and

0.2% Triton X-100) supplemented with the protease inhibitor cocktail (Roche). After washing out the unbound AnxA2with the binding

buffer, GST, GST-p11 or GST-p11/AnxA2 complex were incubated with either full-length SMARCA3 or its fragments which were

generated by in vitro transcription/translation with [35S-Methionine] labeling (TnT Quick Coupled Transcription/Translation System,

Promega). After 3 hr incubation at 4�C with constant rotation, the protein complexes on the beads were eluted using the sample

buffer, and were subjected to SDS-PAGE. The 35S-labeled SMARCA3 proteins on the dried gels were visualized using PhosphoIm-

ager (Figures S1B–S1D).

Peptide Pull-Down Assay

Short peptides derived from SMARCA3 (aa26-42) or AHNAK1 (aa5654-5671) were synthesized with a biotin-tag at the C terminus,

and immobilized onto streptavidin-coupled magnetic beads (DynabeadM-280 Streptavidin, Invitrogen). The peptide-coupled beads

were incubated with p11/AnxA2 complex (1 mg of the complex) in the binding buffer. After three timeswashing, the bound protein was

eluted by boiling in the sample buffer, and subjected to SDS-PAGE followed by western blot analysis using the antibodies indicated

(Figure 2F).

Oligonucleotide Pull-Down Assay

Oligonucleotide pull-down assay was performed as described (Herlt et al., 1988) with minor modification. Briefly, the streptavidin

magnetic beads (Dynabeads MyOne Streptavidin T1, Invitrogen) were preabsorbed with the blocking/coupling buffer for 30 min

at room temperature with constant rotation. The preabsorbed beads were dispensed into individual reaction tubes, coupled with

1 mg of the biotinylated oligonucleotide duplex (tandem B-box from mouse PAI-1 promoter: 50-TTA GAA AGT GGG TGG GGC

TGG AAC ATG TTA GAA AGT GGG TGG GGC TGG AAC ATG-30/50- CAT GTT CCA GCC CCA CCC ACT TTC TAA CAT GTT CCA

GCCCCACCCACT TTC TAA-30-Biotin) (Ding et al., 1996), andwashed three timeswith 13 oligonucleotide binding buffer to remove

the unbound oligonucleotide duplex. The washed beads were incubated with the protein solution prediluted in the reaction buffer, for

S2 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

3 hr at 4�C with constant rotation. The protein complex bound to the magnetic beads was washed three times with 1 3 oligonucle-

otide binding buffer, eluted by boiling in the protein sample buffer, and subjected to SDS-PAGE followed by quantitative western blot

analysis (Odyssey Infrared System, LI-COR).

Oligonucleotide binding buffer (3 3 ): 36 mM HEPES, pH 7.9, 12 mM TRIS, 450 mM KCl, 36% glycerol, 3 mM EDTA, 3 mM DTT.

Blocking/coupling buffer: 13 oligonucleotide binding buffer supplementedwith 2mg/ml BSA (DNase- and lipid-free; Calbiochem),

20 mg/ml poly dI-dC (poly deoxyinosinic-deoxycytidylic acid; Sigma-Aldrich), 200 mg/ml sheared salmon spermDNA (Invitrogen), and

0.2% NP-40 (Sigma-Aldrich).

Reaction buffer: 1 3 oligonucleotide binding buffer supplemented with 2 mg/ml BSA, 20 mg/ml poly dI-dC, and 0.2% NP-40.

Protein Expression and PurificationThe p11-AnxA2 peptide cassette [p11(1-92) - QENLYFQGD linker - AnxA2 peptide (2-15)] was designed according to a published

protocol (Rety et al., 1999; Rezvanpour et al., 2009). The gene fragment encoding p11-AnxA2 peptide cassette was inserted into

a sumo-pRSFDuet-1 vector (modified pRSFDuet-1 vector with 6xHis plus yeast Sumo as the N-terminal fusion tag). The protein

was expressed in BL21 (DE3) RILE. coli cells in LBmedium. Cells were grown at 37�C till OD600 reached around 0.8. Then themedium

was cooled and 0.4 mM IPTGwas added to the culture to induce protein expression overnight at 25�C. The cell extract was prepared

using a French Press in buffer A (50 mM TRIS, pH 7.5, 400 mM NaCl, 20 mM imidazole, 5 mM b-mercaptoethanol and 5% glycerol).

After centrifugation, the supernatant was loaded onto a Ni-NTA affinity column. The eluted target protein was exchanged in buffer A

by dialysis overnight, with Ulp1 protease added for removal of the Sumo-tag. Digested protein was loaded onto a Ni-NTA column

again to remove the 6xHis-Sumo tag and the His-tagged protease. The flow through containing target protein was further purified

by using a Hiload 16/60 G75 column in buffer B (10 mM HEPES, pH 7.5, 150 mM NaCl, 1 mM DTT, 0.1 mM EGTA and 5% glycerol).

The p11-AnxA2 fusion protein was concentrated to about 20 mg/ml for crystallization trials.

We used the coexpression strategy to prepare the complex of full-length p11 and full-length AnxA2. The gene fragments encoding

p11 and AnxA2were inserted into the firstMCS and the secondMCSof themodified pRSFDuet-1 vector, respectively. All the expres-

sionandpurificationconditionswere the sameas for thep11-AnxA2peptide cassette except for usingaheparin column to remove free

p11 protein from the complex. The complex of full-length p11 and full-length AnxA2 was concentrated to about 20 mg/ml in buffer B.

CrystallizationCrystals of p11-AnxA2 peptide cassette in complex with AHNAK1 peptide were obtained by mixing at a cassette:peptide molar ratio

of 1:2 on ice for 30 min, followed by mixing with an equal volume of reservoir buffer containing 100 mM Tris, pH 7.4, and 30%

PEG6000 at 20�C. The complex of the p11-AnxA2 peptide cassette with SMARCA3 peptide was also prepared by mixing the

cassette with peptide at a 1:2 molar ratio. When adding the SMARCA3 peptide, precipitation was found in the mixture. By incubating

the sample at 37�C for 10 min, the precipitation almost disappeared. Crystals were obtained by mixing 1 ml complex with 1 ml crys-

tallization buffer containing 100 mM MES, pH 5.0 and 1.6 M (NH4)2SO4.

The crystals of full-length p11 and full-length AnxA2 in complex with SMARCA3 peptide were obtained by mixing at a full-length

complex:peptide molar ratio of 1:2 on ice for 30 min, followed by mixing with an equal volume of crystallization buffer containing

100 mM citric acid, pH 3.7 and 25% PEG3350.

Structure DeterminationX-ray diffraction data sets for complexes of p11-AnxA2 peptide cassette with AHNAK1 and SMARCA3 peptides were collected at the

Advanced Photon Source NE-CAT 24ID-E beam line. The data sets were indexed, integrated and scaled using HKL2000. The struc-

ture of p11-AnxA2 peptide cassette - AHNAK1 peptide complex was solved by the molecular replacement method in PHASER

(McCoy et al., 2007) using the structure of p11 and AnxA2 complex (PDB 1BT6) as the search model. Further modeling of AHNAK1

peptide was carried out using COOT (Emsley and Cowtan, 2004), and model refinement was performed in PHENIX (Adams et al.,

2002). The final model of the complex was refined to 2.0 A.

The structure of p11-AnxA2 peptide cassette - SMARCA3 peptide complex was solved by molecular replacement method in

PHASER using the 2.0 A structure of the p11-AnxA2 peptide cassette - AHNAK1 peptide complex as the search model. The

SMARCA3 peptide was then modeled in COOT and the whole structure of the complex was refined using PHENIX. The final model

of the complex was refined to 3.0 A resolution. Data collection statistics and structural refinement parameters for complexes of p11-

AnxA2 peptide cassette with bound AHNAK1 and bound SMARCA3 peptides are summarized in Table S1.

The structure of the ternary complex of full-length p11/full-length AnxA2/SMARCA3 peptide was solved by the molecular replace-

ment method in PHASER using the 3.0 A structure of the p11-AnxA2 peptide cassette bound to SMARCA3 peptide and also free

AnxA2 (PDB 1W7B) as search models. The missing parts were then modeled in COOT and the entire structure of the complex

was refined using PHENIX. Data collection statistics and structural refinement parameters for the p11/AnxA2/SMARCA3 peptide

ternary complex are summarized in Table S2.

Nuclear Matrix PreparationNuclear matrix was prepared as reported with minor modification (van Steensel et al., 1995). After washing two times with ice-cold

PBS, cells were incubated with a cell-permeable cross-linker, 1 mM DSP (Pierce), in 13 PBS for 1 hr on ice. The cells were washed

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. S3

twice in CSK buffer, incubated with CSK buffer supplemented with 0.5% NP-40, and 0.5 mM sodium tetrathionate (Sigma-Aldrich),

and incubated 3min with CSK buffer supplemented with 0.5mM sodium tetrathionate. After washing twice with CSK buffer and once

with digestion buffer, the chromatin was digested with 200 mg/ml RNase-free DNase I (Sigma-Aldrich) in the digestion buffer for

30 min, extracted with 0.25 M (NH4)2SO4 in digestion buffer for 10 min, and washed twice with the CSK buffer. For western blot anal-

ysis, the nuclearmatrix preparationwas solubilizedwith 13RIPA buffer and scraped. For the in situ immunocytochemistry, thewhole

cells and the nuclear matrices on coverslips (BD BioCoat) were fixed for 10 min with 2% PFA solution. The cells were incubated with

PBS containing 0.5%NP-40, followed by 5 min incubation in PBS containing 100 mM glycine. The washed whole cells or the nuclear

matrices were subjected to the standard immunocytochemistry procedures using anti-p11 antibody (mouse monoclonal, 1:500, BD

bioscience) and anti-SMARCA3 antibody (goat polyclonal, 1:200, Novus biological).

CSK buffer: 10 mM PIPES, pH 6.8, 100 mM NaCl, 300 mM sucrose, 3 mM MgCl2, 1 mM EGTA, protease inhibitor cocktail

(Complete-EDTAfree; Roche).

Digestion buffer: 10 mM PIPES, pH 6.8, 50 mM NaCl, 300 mM sucrose, 3 mM MgCl2, 1 mM EGTA, protease inhibitor cocktail.

Luciferase Reporter AssayPAI-1 promoter-driven luciferase activity (pGL4.10-mPAI1 promoter) was measured using the Dual-Luciferase Reporter Assay

System (Promega) with Renilla (pGL4.73, Promega) as a reference. The reporter plasmid mixture of Firefly luciferase/Renilla lucif-

erase vectors (50 ng/0.5 ng) was cotransfected with SMARCA3 (0.2 mg), p11 and AnxA2 expression plasmids (0.4 mg each) into

N2a cells. At 24 hr posttransfection, the luciferase activity in the transfected cells was assayed according to the manufacturer’s

manual. For p11 silencing, COS-7 cells were transfected with either p11 siRNA duplex (5’-UUG CCA UCU CUA CAC UGG UCC

AGGU-3’/5’-ACC UGG ACC AGU GUA GAG AUG GCAA-3’) or negative control duplex (20 pmol; medium GC; Invitrogen). At

24 hr posttransfection, the cells were detached and replated onto 12-well plates. Next day, the cells were cotransfected with

SMARCA3 plasmids (0.2 mg) along with the reporter plasmid mixture (50 ng/0.5 ng), and the luciferase activity was measured

24 hr later. All analyses were performed in triplicate, with each experiment repeated at least twice.

ImmunohistochemistryAnimals were deeply anesthetized and perfused transcardially with PBS, followed by 4% paraformaldehyde (PFA) in PBS. Brains

were postfixed in 4% PFA 12 hr at 4�C, and then saturated with 30% sucrose, followed by freezing in the OCT medium on the

dry ice block. A cryostat was used to collect coronal sections of 40 mm thickness along the rostro-caudal axis of the hippocampus.

Immunofluorescent staining of the sections was carried out by the free floatingmethod. For immunostaining for p11, SMARCA3, nes-

tin, and Ki-67, the sections were pretreated with citrate buffer (10 mM sodium citrate, pH 6.0, 0.05% Triton X-100) for 30 min at 80�Cfor antigen retrieval, and then rinsed three times with PBS. For BrdU immunostaining, the sections were pretreated with 1 N HCl for

1 hr at 45�C to denature DNA. After being neutralized, the sectionswere rinsed twicewith PBS, and incubatedwith the blocking buffer

(13 PBS, 0.2%BSA, 2% normal donkey serum, 0.3% Triton X-100) at room temperature for 1 hr. After blocking, sections were incu-

bated with the primary antibodies diluted in the blocking buffer. The immunohistochemistry was done using the following antibodies:

anti-p11 (goat polyclonal, 1:100, R&D systems), anti-eGFP (chicken polyclonal, 1:200, Abcam), anti-eGFP (rabbit polyclonal, 1:200,

Abcam), anti-SMARCA3 (rabbit polyclonal, 1:500, Thermo Scientific), anti-calretinin (mousemonoclonal, 1:4000, Swant), anti-parval-

bumin (rabbit polyclonal, 1:1000, Swant), anti-parvalbumin (mouse monoclonal, 1:1000, Swant), anti-calbindin-D28K (mouse mono-

clonal, 1:1000, Swant), anti-neuropeptide Y (rabbit polyclonal, 1:1000, Phoenix Pharmaceutical), anti-GFAP (rabbit polyclonal,

1:1000, Sigma-Aldrich), anti-nestin (mouse monoclonal, 1:200, Abcam), anti-BrdU (rat monoclonal, 1:200, Abcam), anti-Ki-67 (rabbit

polyclonal, 1:200, Abcam), anti-doublecortin (goat polyclonal, 1:200, Santa-Cruz Biotech), and anti-NeuN (rabbit polyclonal, 1:1000,

Millipore). After incubation for 48-72 hr, sections were washed, incubated with Alexa-fluor-conjugated secondary antibodies (Invitro-

gen), and counterstained with DraQ5. All the sections were examined under a Zeiss LSM710 confocal microscope or wide-field fluo-

rescence microscope (Zeiss).

Immunohistochemistry of BAC-[p11]-eGFP Mice and Densitometric AnalysisBrain tissue (n = 5-6/group) was prepared for immunohistochemistry using a standard protocol. Brains were sectioned at 40 mm, and

every twelfth coronal section, along the rostro-caudal axis of hippocampus (�2.5 to �4.0 mm AP) was stained immunohistochemi-

cally, as described above. Sections were scannedwith a Zeiss LSM710 confocal microscopewith a 20 x objective lens. Densitometry

on digitized images was carried out using ImageJ software (NIH). Hippocampal subregions (Amaral et al., 2007) were defined by

referencing the well-characterized laminar distribution pattern, as well as well-defined anatomical landmarks visualized with counter-

staining with DraQ5, a nuclei-staining dye. The average fluorescent intensity is measured by outlining the region of interest (ROI) for

each laminar subregion including the hilus, the inner molecular layer (IML), the granular cell layer (GCL), and the outer molecular layer

(OML). Relative fluorescence intensity of each subregion was given after normalization within the same section: the value of individual

subregion was divided by the value of the outer molecular layer (OML), which shows background fluorescence with negligible immu-

nostaining and also with no change upon fluoxetine treatment. All imaging and analyses were performed blind to the experimental

conditions. For quantification comparisons, we used a two-way ANOVA followed by Bonferroni’s t test for pairwise comparisons

using Prism (GraphPad Software).

S4 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

BrdU Labeling and Neurogenesis AssayThe effect of chronic fluoxetine treatment on cell proliferation or cell survival in the WT and KO mice was assessed as described

(Malberg et al., 2000). For the proliferation assay, the mice were labeled with BrdU solution (200 mg/kg) for 2 hr prior to sacrifice

by perfusion. For the survival/neurogenesis assay, the mice were injected with BrdU (75 mg/kg) two times per day for three consec-

utive days, and were perfused 4 weeks later. Preparation of brain sections and immunostaining were done as described above, and

stereological quantitation of BrdU-positive cells were performed as described (Malberg et al., 2000). Briefly, serial coronal sections

(40 mm) were cut through the entire hippocampus in its rostro-caudal extension on a cryostat and stored in PBS with 0.1% NaN3.

Every sixth section throughout the hippocampus was processed for BrdU immunohistochemistry as described. After the staining

procedure, all the slides were coded. An experimenter blinded to the slide code counted all BrdU-labeled cells in the granule cell

layer (GCL) and the subgranular zone (SGZ) of the dentate gyrus (DG) in the total 12 sections from the individual mouse. The total

number of BrdU-labeled cells per section was determined and multiplied by 6 to obtain the total number of cells per dentate gyrus.

Two-way ANOVA and post hoc Bonferroni tests were performed on these totals.

Behavioral TestingAll the behavioral tests were done during the light cycle in a dedicated sound-proof behavioral facility by experimenters blind to treat-

ment- and genotype information.

Novelty suppressed feeding was performed as described (Samuels and Hen, 2011). The behavior testing room was soundproof

and the illumination level was maintained at 100 lux. After food-deprivation for 20-24 hr, mice were habituated to the testing room

in their home cage for 1 hr, and then singly housed in a fresh plastic cage with no food, or water for an additional 30 min. During

the test sessions, three pellets of standard mouse chow were placed on a white filter paper platform in the center of the box

(50x50x20 cm, 2 cm of wooden bedding) and the latency to feed (defined as the mouse sitting on its haunches and biting the pellet

with the use of forepaws) wasmeasured. Eachmouse was weighed before food deprivation to assess the percentage of body weight

loss during the period of drug treatment.

The restraint treatment used in this study was performed as previously described (Seo et al., 2012). In brief, mice were individually

placed head-first into well ventilated 50 ml polypropylene conical tubes, which were then plugged with a 4.5-cm-long middle tube,

and finally tied with a cap of the 50ml tube. After each session of restraint, themicewere returned to their home environment, in which

theywere housed in pairs in normal plastic cageswith free access to food andwater. From the next day after the last restraint session,

mice were intraperitoneally injected with escitalopram daily for 4 weeks) in the posttreatment paradigm.

Tail suspension test was performed as described (Svenningsson et al., 2006).

For sucrose preference test, a bottle with tap water and another bottle with 1% (Figure S6E) or 2% (Figure 7H) sucrose solution

were given to mice. After habituation for 3 days, mice were given a free choice between two bottles. To prevent a possible effect

of drinking behavior, the left/right location of the bottles was switched every day. The consumption of water and sucrose solution

was measured daily for 3 days by weighing the bottles. The sucrose preference was calculated as the ratio of consumed sucrose

solution to consumed water.

For open-field test, mice were habituated in the testing room in their home cages for 30 min and the open-field behavior was

assayed in a square arena (50x50x22.5 cm) for 60 min (Figure S6C) or 30 min (Figure S6I). The measures were automatized using

two rows of infrared photocells placed 20 and 50 mm above the floor, spaced 31 mm apart. Photocell beam interruptions were re-

corded on a computer using the superflex software (Accuscan Instruments). Horizontal activity was measured by counting all inter-

ruptions of photobeams in the lower rows. Total distance traveled was measured using the automated superflex software.

For elevated plus maze test, the test apparatus comprised two open arms (25 3 5 x 0.5 cm) across from each other and perpen-

dicular to two closed arms (253 5 x 16 cm) with a center platform (53 5 x 0.5 cm) and the entire apparatus was 50 cmabove the floor.

Amouse was placed in the center area of themaze and wasmonitored using a video camera connected to a computer. Time spent in

the open arms and the number of entries into each arm were quantified as described (Lister, 1987).

For light/dark box, a dark insert was placed into the open field arena to divide the chamber into light and dark compartments. Mice

were placed in the dark side of the box. The time spent and distance traveled in the light side of the arena were recorded for 10 min.

For the elevated plus maze test, mice were placed in the middle of the elevated plus maze (50 cm from the floor) and allowed to

explore the maze for 10 min. Time spent and the number of entries into open arms were measured using a tracking software

(Ethovision).

SUPPLEMENTAL REFERENCES

Adams, P.D., Grosse-Kunstleve, R.W., Hung, L.W., Ioerger, T.R., McCoy, A.J., Moriarty, N.W., Read, R.J., Sacchettini, J.C., Sauter, N.K., and Terwilliger, T.C.

(2002). PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr. D Biol. Crystallogr. 58, 1948–1954.

Covington, H.E., 3rd,Maze, I., LaPlant, Q.C., Vialou, V.F., Ohnishi, Y.N., Berton, O., Fass, D.M., Renthal,W., Rush, A.J., 3rd,Wu, E.Y., et al. (2009). Antidepressant

actions of histone deacetylase inhibitors. J. Neurosci. 29, 11451–11460.

Czernik, A.J., Girault, J.A., Nairn, A.C., Chen, J., Snyder, G., Kebabian, J., and Greengard, P. (1991). Production of phosphorylation state-specific antibodies.

Methods Enzymol. 201, 264–283.

Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. S5

Herlt, M., Schwarz, H.P., Neumann, R., and Eggers, H.J. (1988). Detection of DNA-protein binding in western blots by phosphorus-labeled and biotinylated DNA

probes. Biotechniques 6, 324–331.

Lister, R.G. (1987). The use of a plus-maze to measure anxiety in the mouse. Psychopharmacology (Berl.) 92, 180–185.

Malberg, J.E., Eisch, A.J., Nestler, E.J., andDuman, R.S. (2000). Chronic antidepressant treatment increases neurogenesis in adult rat hippocampus. J. Neurosci.

20, 9104–9110.

McCoy, A.J., Grosse-Kunstleve, R.W., Adams, P.D., Winn, M.D., Storoni, L.C., and Read, R.J. (2007). Phaser crystallographic software. J. Appl. Cryst. 40,

658–674.

Samuels, B.A., and Hen, R. (2011). Novelty-suppressed feeding in the mouse. In Mood and Anxiety Related Phenotypes in Mice: Characterization Using Behav-

ioral Tests, Volume II, T.D. Gould, ed. (New York: Springer), pp. 107–121.

Seo, J.S., Park, J.Y., Choi, J., Kim, T.K., Shin, J.H., Lee, J.K., and Han, P.L. (2012). NADPH oxidase mediates depressive behavior induced by chronic stress in

mice. J. Neurosci. 32, 9690–9699.

van Steensel, B., Brink, M., van der Meulen, K., van Binnendijk, E.P., Wansink, D.G., de Jong, L., de Kloet, E.R., and van Driel, R. (1995). Localization of the gluco-

corticoid receptor in discrete clusters in the cell nucleus. J. Cell Sci. 108, 3003–3011.

Vialou, V., Robison, A.J., Laplant, Q.C., Covington, H.E., 3rd, Dietz, D.M., Ohnishi, Y.N., Mouzon, E., Rush, A.J., 3rd, Watts, E.L., Wallace, D.L., et al. (2010).

DeltaFosB in brain reward circuits mediates resilience to stress and antidepressant responses. Nat. Neurosci. 13, 745–752.

S6 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

Figure S1. Characterization of SMARCA3 as a Binding Partner of p11/AnxA2 Complex, Related to Figure 2

(A) Effect of p11mutation in C83 residue on the interactionwith AnxA2 and endogenous p11.WT, C83S, and C83Qmutants of p11were immunoprecipitated from

transfected COS-7 cells, and immunoblotted for AnxA2 and p11 as indicated.

(B) Potentiation of the interaction between p11 and SMARCA3 by AnxA2. GST-p11 was immobilized on GSH-agarose beads, preincubated with various

concentrations of AnxA2, and incubated with 35S-labeled SMARCA3. The coprecipitated protein complex was visualized by autoradiography (top) and Coo-

massie staining (bottom). Intensities of autoradiographic SMARCA3 were quantified and compared with the value in the absence of AnxA2.

(C and D) GST-p11 pull-down assay using a series of C-terminal (C) and N-terminal (D) truncations of SMARCA3.

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. S7

Figure S2. Crystal Structure of AHNAK1 Peptide Bound to p11-AnxA2 Peptide Cassette as a Binary Complex, Related to Figure 3

(A) A ribbon view of the crystal structure of the p11-AnxA2 peptide cassette in complex with AHNAK1 peptide solved at 2.0 A resolution. In this dimeric complex,

p11 (blue and green) and AnxA2 peptide (salmon) are shown in ribbon, with the bound AHNAK1 peptides (yellow) in stick representations.

(B) Space-filling (peptide) and electrostatic surface (p11-AnxA2 peptide cassette) view of the complex illustrating the positioning of the pair of AHNAK1 peptides

within the binding groove of the dimeric complex.

(C) Omit 2Fo-Fc electron density maps contoured at 1s level of bound AHNAK1 peptide in one monomer of the complex. AnxA2 peptide (salmon) and p11 (blue)

are represented in ribbon views, SMARCA3 peptide (yellow) is represented in a stick view, and electron density contours are in green.

(D) Details of intermolecular contacts within one monomer of the complex between the AHNAK1 peptide (yellow) and p11 (blue) - AnxA2 peptide (salmon)

cassette, with hydrogen bonds depicted as red dashed lines.

(E) Positioning of the conserved residues Pro5664 and Phe5666 of the bound AHNAK1 peptide within a hydrophobic pocket formed by residues of AnxA2 peptide

(salmon) and p11 (blue).

(F) View of the superimposed structures of the p11-AnxA2 peptide cassette in the free (p11 in yellow and AnxA2 peptide in cyan) and AHNAK1-bound (p11 in blue

and AnxA2 peptide in red) states. The black arrows highlight the conformational changes in p11 and AnxA2 peptide upon complex formation.

S8 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

Figure S3. Point Mutations in the Key Residues Found in the Binding Interface of p11, AnxA2, and SMARCA3 Abolish Ternary Complex

Formation, Related to Figure 3

(A) Effect of SMARCA3 mutation (P35A and F37Y) on the interaction with p11/AnxA2 complex. WT, P35A, and F37Y mutants of SMARCA3 were immunopre-

cipitated using C-terminal V5 epitope from transfected COS-7 cells, and immunoblotted for SMARCA3-V5, AnxA2, and p11 as indicated.

(B) Effect of AnxA2mutation (L8A, L11A, L13A) on the interaction with p11 and SMARCA3. WT, L8A, L11A, and L13Amutants of AnxA2were immunoprecipitated

using C-terminal myc epitope from transfected COS-7 cells, and immunoblotted for SMARCA3, AnxA2, and p11 as indicated.

(C) Effect of p11 mutation (D60A) on the interaction with AnxA2 and SMARCA3. WT and D60A mutant of p11 were immunoprecipitated using C-terminal Flag-

epitope from transfected COS-7 cells, and immunoblotted for Flag-p11 (overexpressed), p11 (endogenous), SMARCA3, and AnxA2 as indicated. D60A mutant

appears to be in lower amount in total lysate (not shown) and in the immune complex, presumably due to instability. Although comparable amounts of

endogenous p11 and AnxA2were coprecipitated, SMARCA3wasmarkedly reduced in the immune complex containing the D60Amutant compared to that of WT

p11, supporting a critical role for the D60 residue in the interaction with SMARCA3.

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. S9

Figure S4. SSRI Regulates p11 Expression in the Dentate Gyrus, Related to Figure 5

(A) Expression of p11 in BAC-[p11]-eGFP reporter mice. P11 expression was assessed in dentate gyrus slices from BAC-[p11]-eGFP reporter mice after staining

with anti-eGFP or anti-p11 antibody. The layered dentate gyrus structure was visualized with nuclear dye staining (Blue; DraQ5) in the merged image. IML, inner

molecular layer; GCL, granule cell layer. Scale Bar, 20 mm. Representative cells doubly labeled are marked with arrow heads.

(B) Induction of [p11]-eGFP in dentate gyrus by chronic SSRI. BAC-[p11]-eGFP mice were treated with either VEH or FLX for 2 weeks. The dentate gyrus slices

were costained with anti-eGFP antibody (Green) and nuclear dye (Blue; DraQ5). Representative images are from the rostral (upper panels) or the caudal (lower

panels) side of the hippocampus. Scale Bars, 100 mm.

S10 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.

Figure S5. Generation and Characterization of SMARCA3 KO Mouse Line, Related to Figure 6

(A) Schematic for SMARCA3 gene KO strategy. In the targeting vector, exons 11-13 were flanked by loxP sites. Selection marker (Puro) was flanked by F3 sites,

inserted into intron 10. Constitutive KO line was established after in vivo Cre recombinase-mediated deletion of exons 11-13 and selectionmarker (Puro). Deletion

of exons 11-13 creates a frame shift in downstream exons and generates a premature STOP codon.

(B and C) Western blot analysis of SMARCA3, p11, AnxA2 and actin proteins in the hippocampus of SMARCA3 WT (+/+), heterozygote (+/�), homozygote KO

(�/�) mice. Representative images (B) and quantitation (C) are shown. The levels of p11 and AnxA2 are not altered. Data represent mean ± SEM (n = 5 mice per

group). ns (non-significant).

Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc. S11

Figure S6. Fluoxetine-Induced Increase in Number of Ki-67-Positive Cells; Baseline Locomotor Activity and Depression-like and Anxiety

Behaviors in SMARCA3 WT and KO Mice; Food Consumption and Locomotor Activity as Control Behaviors in Chronic Antidepressant-

Treated Mice; the Effect of SMARCA3 KO in Acute Antidepressant Action, Related to Figure 7

(A and B) Fluoxetine-induced increase of Ki-67-positive cells in WT (+/+) and SMARCA3 KO (�/�) mice. (A) Immunostaining with anti-Ki-67. Scale Bars, 100 mm.

(B) Quantitation of Ki-67-positive cells in the subgranular zone (n = 7-8 mice per group). Data represent mean ± SEM. *p < 0.05, **p < 0.01, ns (non-significant),

two-way ANOVA followed by the post-hoc Bonferroni test. See also Table S3.

(C–G) Naive WT and SMARCA3 KO mice were tested for the following behaviors. Total distance traveled was measured in an open field (C). Depression-like

behavior was measured by immobility in the tail suspension test (D) and anhedonia was assayed by sucrose preference (E). Anxiety was measured by time spent

in the light compartment of the light/dark box (F) or by time spent in the open arm of the elevated plusmaze (G). All data are presented asmean ±SEMn= 7-8mice

per group. ns (nonsignificant), t test.

(H) WT and SMARCA3 KO mice were treated with either VEH or FLX for 4 weeks. Hunger level was assessed by measuring food consumption over a period of

5 min. Data represent mean ± SEM (n = 14-16 per group).

(I) WT and SMARCA3 KO mice were exposed to control condition (nonstressed, NS) or restraint stress (RS) for 2 weeks, followed by either VEH or eCIT

(escitalopram) for 4 weeks. The total distance traveled was assessed using open field test. Data represent mean ± SEM (n = 8-11 per group).

(J) WT and SMARCA3 KO mice were tested for behavioral despair after acute escitalopram, for 30 min, in the tail suspension test. Data represent mean ± SEM

(n = 8 per group). *p < 0.05, ns (nonsignificant), two-way ANOVA followed by the post hoc Bonferroni test.

See also Table S3.

S12 Cell 152, 831–843, February 14, 2013 ª2013 Elsevier Inc.