CTCF Is Required for Neural Development and Stochastic - Cell Press

Cell Reports

Article CTCF Is Required for Neural Development and Stochastic Expression of Clustered Pcdh Genes in Neurons Teruyoshi Hirayama,1,2 Etsuko Tarusawa,2,3 Yumiko Yoshimura,3,4 Niels Galjart,5 and Takeshi Yagi1,2,* 1KOKORO-Biology

Group, Laboratories for Integrated Biology, Graduate School of Frontier Biosciences Science and Technology Agency-Core Research for Evolutional Science and Technology (CREST) Osaka University, Suita 565-0871, Japan 3Division of Developmental Neurophysiology, National Institute for Physiological Sciences, National Institutes of Natural Sciences, Okazaki, Aichi 444-8787, Japan 4Department of Physiological Sciences, The Graduate University for Advanced Studies, Okazaki 444-8585, Japan 5Department of Cell Biology and Genetics, Erasmus Medical Center, 3000 CA Rotterdam, The Netherlands *Correspondence: [email protected] http://dx.doi.org/10.1016/j.celrep.2012.06.014 2Japan

SUMMARY

The CCCTC-binding factor (CTCF) is a key molecule for chromatin conformational changes that promote cellular diversity, but nothing is known about its role in neurons. Here, we produced mice with a conditional knockout (cKO) of CTCF in postmitotic projection neurons, mostly in the dorsal telencephalon. The CTCF-cKO mice exhibited postnatal growth retardation and abnormal behavior and had defects in functional somatosensory mapping in the brain. In terms of gene expression, 390 transcripts were expressed at significantly different levels between CTCFdeficient and control cortex and hippocampus. In particular, the levels of 53 isoforms of the clustered protocadherin (Pcdh) genes, which are stochastically expressed in each neuron, declined markedly. Each CTCF-deficient neuron showed defects in dendritic arborization and spine density during brain development. Their excitatory postsynaptic currents showed normal amplitude but occurred with low frequency. Our results indicate that CTCF regulates functional neural development and neuronal diversity by controlling clustered Pcdh expression. INTRODUCTION CCCTC-binding factor (CTCF) is a highly conserved eukaryotic protein with multivalent functions. Depending on the genomic and epigenomic context, CTCF can act as a gene repressor, activator, or insulator, and with roles in genomic imprinting, X chromosome inactivation, and somatic recombination in lymphocytes (Ohlsson et al., 2010; Seitan et al., 2012). CTCF also appears to be essential for global organization of the chromatin architecture (Phillips and Corces, 2009). Because a complete CTCF deficiency disturbs early mouse development (Fedoriw et al., 2004), its in vivo function has been examined

using a conditional knockout (cKO) allele (Ctcf fl) that allows CTCF to be deleted in only specific cell types (Splinter et al., 2006; Heath et al., 2008; Soshnikova et al., 2010). Recent reports showed that CTCF regulates V(D)J recombination in B and T cells during lymphocyte diversification by mediating loop formation and conformational changes (Guo et al., 2011; Seitan et al., 2012). These studies revealed a critical role for CTCF in generating diversity in individual cells; however, nothing is known about its in vivo function in neurons. Neuronal diversity is an important feature of the complex neural networks in the brain (Sperry, 1963; Muotri and Gage, 2006). At the molecular level, enormous neuronal diversity is generated in the vertebrate brain by the stochastic and combinatorial expression of the clustered protocadherin (Pcdh) genes, which include the Pcdha, Pcdhb, and Pcdhg clusters and encode a diverse group of proteins belonging to the cadherin superfamily (Yagi, 2008; Zipursky and Sanes, 2010). These proteins are expressed in the brain and have important roles in axonal targeting, synapse formation, and dendritic arborization (Kohmura et al., 1998; Wu and Maniatis, 1999; Weiner et al., 2005; Hirayama and Yagi, 2006; Hasegawa et al., 2008; Katori et al., 2009; Garrett et al., 2012). Each clustered Pcdh isoform has its own promoter and is regulated by alternative promoter usage and cis-splicing (Tasic et al., 2002; Wang et al., 2002). Interestingly, each Pcdh gene cluster generates distinct sets of Pcdh isoforms in individual neurons, and these isoforms are combinatorially expressed by stochastic and monoallelic mechanisms (Esumi et al., 2005; Kaneko et al., 2006; Noguchi et al., 2009). In addition the different Pcdh proteins expressed in individual neurons form cis-tetrameric complexes with an enormous variety of combinations (Schreiner and Weiner, 2010; Yagi, 2012). Recent genome-wide analyses showed that the clustered Pcdh gene locus contains many CTCF-binding sites and engages in highly complex intrachromosomal interactions (Kim et al., 2007; Handoko et al., 2011). In addition CTCF regulates individual clustered Pcdh genes in SH-SY5Y, CAD, and N2a cells (Golan-Mashiach et al., 2012; Monahan et al., 2012). CTCF binds to each promoter of the clustered Pcdh genes, Cell Reports 2, 345–357, August 30, 2012 ª2012 The Authors 345

1

2

34 5 67 8 9101112

floxed

puro

neo

CTCF-cKO (fl/fl; +/Cre)

B

Control (fl/fl; +/+)

A lacZ SA

flox

Nex-Cre mediated recombination 1

del

2

CTCF-cKO

lacZ SA

C

Control ctx hip mb

ob

st

th cb

CTCF-cKO

ctx hip

ob

mb

st th

cb

and this binding is mediated by a transcriptional enhancer element of DNase I hypersensitive site (HS) 5–1 (Kehayova et al., 2011; Monahan et al., 2012). These results suggest that CTCF is a regulator for the stochastic expression of clustered Pcdh genes in individual neurons. However, CTCF’s function in individual differentiated neurons is largely unknown. Here, we show that the loss of CTCF in the postmitotic projection neurons of mice resulted in postnatal growth retardation, and abnormal behavior. These mutants also exhibited a somatosensory map defect in the brain. In the mutant cortex and hippocampus, the expression levels of 390 genes were significantly altered, and the levels of almost all the stochastically expressed clustered Pcdh genes were markedly reduced. Expression analyses of the clustered Pcdh genes revealed that CTCF is required for their stochastic expression in individual neurons. Each neuron showed decreased dendritic arborization and spine density but had mature synapses. Our data demonstrate that CTCF is important for building functional neural networks in the brain and for regulating neuronal diversity elicited by the stochastic expression of the clustered Pcdh genes. RESULTS

D

Control CTCF-cKO

F

Control

Percent Survival

100 80 60 40 20 0 0

5

10

15

20

25

30

Days Elapsed

E

CTCF-cKO Body Weights (g)

12

Control CTCF-cKO

10 8 6 4

*

*

14

21

*

2 0 1

7

Postnatal Days

Figure 1. Phenotypes of the Nex-Cre-Mediated CTCF cKO Mice (A) Strategy for the Cre-loxP-mediated gene targeting of Ctcf. (B) Southern blots for control (fl/fl;+/+) and CTCF-cKO (fl/fl;+/Cre) cortex and hippocampus at P1. About 70% of the genomic DNA was deleted in the CTCF-cKO. (C) Anti-CTCF immunostaining of the sagittal section of P0 brain. The CTCF deletion was found mostly in the dorsal telencephalon, including the cortex (ctx) and hippocampus (hip), but less in olfactory bulb (ob), striatum (st), thalamus (th), midbrain (md), and cerebellum (cb). High-magnification images of the cortex and hippocampus are shown in Figures 4A and 4B. Scale bar, 500 mm. (D) Survival curves of control (n = 42) and CTCF-cKO (n = 33) mice. (E) Body weight over time. The CTCF-cKO mice exhibited significant growth retardation by 7 days after birth. n > 10 for each group. *p < 0.01. Error bars represent SEM. (F) CTCF-cKO mice showed an abnormal limb-clasping reflex when suspended by the tail. See also Figures S1, S2, and S3 and Table S1.

346 Cell Reports 2, 345–357, August 30, 2012 ª2012 The Authors

CTCF-cKO Mice Show Growth Retardation and Abnormal Behavior and Die within 1 Month of Birth To investigate CTCF’s role in the brain, we produced CTCF-cKO mice by crossing Ctcf fl/fl mice with Nex-Cre+/Cre knockin mice, in which Cre recombinase, driven by the Nex promoter, is prominently expressed in postmitotic cortical and hippocampal projection neurons (Figure 1A). CTCF is ubiquitously expressed in the brain, including projection and inhibitory neurons and glia (Control in Figures 1C, 4A, and 4B). The Cre-mediated deletion in CTCF-cKO mice was confirmed by Southern blotting for the cortex and hippocampus on postnatal day 1 (P1) (Figure 1B) and by immunohistochemistry of the P0 brain (Figure 1C). The loss of CTCF expression was mostly observed in differentiated projection neurons of the dorsal telencephalon, consistent with previous Nex-Cre conditional targeting (Goebbels et al., 2006) (Figure 1C). In the cortex and hippocampus, the CTCF deficiency was specifically found in postmitotic projection neurons, and not in inhibitory neurons or glia (Figures 4A and 4B). The CTCF-cKO mice were born at a normal Mendelian ratio and exhibited no obvious differences from control littermates on P0 (Table S1). Nissl staining did not reveal any obvious differences in gross brain structure between the control and CTCFcKO mice at P0 (Figure S1A). Analysis of neural pathway and cortical layer structures by anti-neurofilament staining and in situ hybridization for cortical layer-specific markers showed no obvious differences in the patterns (Figures S1B and S2). In addition anti-ssDNA staining showed no increase in apoptosis (Figure S3). However, 30% of the CTCF-cKO mice died within the first day (Figure 1D), most with no milk in the stomach. The surviving CTCF-cKO mice showed obvious growth retardation by P7 (Figure 1E) and then exhibited abnormal behavior, including the limb-clasping reflex when suspended by the tail (Figure 1F). Most of these mice died between P23 and P27 (Figure 1D). These results suggested that CTCF in projection

neurons is essential for the postnatal acquisition of normal brain projection neuron function. Loss of Barrel Structure in the Somatosensory Cortex of CTCF-cKO Mice No obvious histological abnormalities were found in the CTCFcKO cortex (Figures S1, S2, and S3). However, no organized barrel structure was observed in the CTCF-cKO cortex (Figures 2A–2D). In the rodent somatosensory cortex, discrete cytoarchitectonic units called barrels, which correspond to individual whiskers, form during the first postnatal week as a result of whisker stimulation, and this process shapes the organization of the layer IV neurons in the somatosensory cortex and of the thalamocortical axon (TCA) terminals that synapse onto them (Inan and Crair, 2007). Here, we examined the cytoarchitectonic organization of the barrels in the CTCF-cKO mice by Nissl staining at P7 (Figure 2A). Compared with controls, the CTCF-cKO mice lacked the densely stained barrel structures normally seen in layer IV of the somatosensory cortex. Using cytochrome oxidase (CO) histochemistry, which strongly stains axon terminals and postsynaptic dendrites, we further investigated the barrel structure in layer IV of the CTCF-cKO mice. A continuous, high CO-intense band in layer IV was observed (Figure 2B), but no organized barrel structure. In contrast the controls showed clearly separated barrels. To observe the TCA terminals specifically, we examined the serotonin transporter (5-HTT) distribution in the somatosensory cortex (Figure 2C). In the CTCF-cKO mice, dense staining of the TCA terminals was found in layer IV, but not the segregated barrel pattern seen in controls. On the other hand, other 5-HTT-positive fibers, derived from the raphe nucleus of the brain stem, exhibited a normal distribution in the somatosensory cortex and hippocampus of the CTCF-cKO mice. Thus, not only the arrangement of postsynaptic neurons but also the segregation of TCA terminals were disrupted in the CTCF-cKO mice. Furthermore, tangential sections through layer IV of a flattened cortex preparation that the primary cortical areas (somatosensory, visual, auditory) established during perinatal development were maintained in the CTCF-cKO mice (Figure 2D). The somatosensory representations of the whisker pad, lower lip, forepaw, and hindpaw were appropriately separated as in the control; however, the barrel structure in the primary somatosensory area was specifically disrupted in the CTCF-cKO mice (Figure 2D). We next analyzed topographic projections from the thalamus to the somatosensory cortex. To label thalamic neurons from the cortex, we inserted two different lipophilic dyes, DiI and DiA, into different sites in the somatosensory cortex (Figure 2E) (Ince-Dunn et al., 2006). After 4 weeks, neuronal somata in the ventroposterior nucleus of the thalamus were labeled by dye transported retrogradely via TCAs. The DiI- and DiA-labeled neurons were located in different places within the ventroposterior nucleus in the controls (Figure 2F) and, to a large extent, in the CTCF-cKO mice as well. These results indicated that the TCAs projected to the somatosensory cortex with the correct topography, but the barrel structure could not be formed in layer IV in the CTCF-cKO mice, suggesting that CTCF affects the final refinement of the somatosensory map during postnatal brain development.

Structures of Barrelettes and Barreloids in the CTCFcKO Mice The somatosensory pathway from the facial vibrissae to the somatosensory cortex runs through the brain stem and thalamus, organized at each relay into topographic structures termed ‘‘barrelettes’’ and ‘‘barreloids.’’ In the Nex-Cre knockin mice, Cre is expressed in postmitotic projection neurons not only in the cortex and hippocampus but also in part of the brain stem (Goebbels et al., 2006). We therefore performed CO staining in these areas (Figure S4). In the brain stem the spinal nucleus of the trigeminal nerve (SpV) exhibited a clear barrelette pattern, even in the CTCF-cKO mice (Figure S4B). On the other hand the patterns in the principal nucleus of the trigeminal nerve (PrV) and in the ventral posteromedial (VPM) and posterolateral (VPL) nucleus in the CTCF-cKO mice were segmented but somewhat obscure compared to those of the normal mice (Figures S4A and S4C). In the somatosensory cortex the barrel segmentation pattern was completely lost in the CTCF-cKO mice, although there was intense CO staining (Figures 2B and S4D). These results showed that the CTCF cKO in the Nex-Cre knockin mice partially affected the barrelette and barreloid formation, and thus, these defects could have contributed to the loss of barrel formation in the CTCF-cKO cortex. In any case the CTCF deletion in postmitotic projection neurons disturbed the functional maps of neural networks during brain development. Gene Expression Profile in the CTCF-cKO Mouse Brain More than 30,000 CTCF-binding sites have been identified by genome-wide analyses in several human and mouse cell types (Kim et al., 2007; Handoko et al., 2011). Because most CTCFbinding sites are located near genes, we expected to find dramatic alterations in the mRNAs expressed in the CTCF-cKO mouse brain. Therefore, we examined the gene expression profile by microarray analysis. In the CTCF-cKO mice the levels of 390 transcripts were significantly altered compared with controls, in both the cortex and hippocampus (a false discovery rate [FDR] 1.2). Of these, 264 genes (67.7%) were downregulated, and 126 (32.3%) were upregulated (Tables S2 and S3). The tendency for more genes to be downregulated is consistent with previous results of CTCF depletion by RNAi in growing mouse oocytes (Wan et al., 2008). To assess the distribution of the affected genes, we plotted their transcriptional start sites on all the chromosomes. They were scattered among all the autosomes and the X chromosome, and the up- and downregulated genes were intermingled. We found the highest frequency of expression changes at the clustered Pcdh locus (Figure 3A). Next, we assessed the biological role of CTCF in neurons by gene ontology analysis and found that genes involved in cell adhesion, behavior, postembryonic development, and the cell morphogenesis associated with neuronal differentiation were significantly overrepresented (Table S4). Cell-adhesion genes were especially highly represented, and of these, 22 of 38 (58%) were clustered Pcdh genes (Table S5). Because the barrel structure in the somatosensory cortex of the CTCF-cKO mice was defective, we examined the expression of NR1 (NMDA receptor, NMDA1) and other barrel map formation-related genes in the microarray data but did not detect any significant changes (Table S6), indicating that the Cell Reports 2, 345–357, August 30, 2012 ª2012 The Authors 347

A

B Control

C

Control

Control

CTCF-cKO

CTCF-cKO

I II III IV

V

VI

CTCF-cKO I II III IV V

VI

D

Control

CTCF-cKO

v

v

hp

hp

a a sw

fp

lw

fp

lw

sw

ll

ll

E

F

Control

CTCF-cKO

DiA SC

4 weeks

DiI at 37°C

VP

Figure 2. Lack of Somatosensory Barrel Formation in CTCF-cKO Cortex (A–C) Staining to detect the barrel structures in coronal cortical sections at P7. Nissl staining (A), CO staining (B), and 5-HTT immunostaining (C). (D) Immunostaining for 5-HTT in tangential cortical sections at P7. A correct sensory body map was observed in the CTCF-cKO mice compared with control mice, but the barrel structures in the primary somatosensory map were deficient. v, visual cortex; a, auditory cortex; lw, large whisker; sw, small whisker; ll, lower lip; hp, hind paw; fp fore paw. (E) Schematic diagram for the analysis of topographic projections from the thalamus to somatosensory cortex. DiI and DiA were placed at different sites in the somatosensory cortex (SC) at P0. After 4 weeks, both dyes were retrogradely transported to the ventroposterior nucleus (VP) in the thalamus. (F) Neurons labeled with DiI (red) and DiA (green) were separated from each other in control and CTCF-cKO VP. Scale bars, 100 mm (A–C and F) and 500 mm (D). See also Figure S4.


loss of barrel structure observed in the CTCF-cKO mice was not due to barrel map formation-related genes but to other mechanisms. Alteration of Clustered Pcdh Gene Expression in the CTCF-cKO Mice The clustered Pcdha and Pcdhg genes have alternative promoters, and Pcdhb has 22 one-exon genes. The C-type isoforms, ac1–ac2 and gc3–gc5, are expressed in most neurons, whereas the other isoforms show stochastic and combinatorial expression in individual neurons. Microarray and quantitative reverse-transcription PCR (qRT-PCR) analyses showed that, except for b1, all the stochastically expressed clustered Pcdh isoforms, a1–a12, b2–b22, ga1–ga12, and gb1–gb8, were downregulated by the CTCF deletion. On the other hand the constitutively expressed C-type isoforms, ac2, gc4, and gc5, were upregulated, except for ac1, which was downregulated in both the cortex and hippocampus (Figures 3C and 3D). The other genes on the clustered Pcdh locus, for example Taf7 and Diap1, were unaltered in the cortex, although they were upregulated in the hippocampus. These results were also confirmed by qRT-PCR analysis of the total RNAs (Figure 3E). All the stochastically expressed isoforms, a4, a12, b3, b16, b22, ga3, and ga7, were downregulated, whereas the constitutively expressed isoforms, ac2 and gc4, were upregulated in both cortex and hippocampus (Figure 3E). These results showed that CTCF is required for the stochastic expression of clustered Pcdh genes in individual neurons, but not for their constitutive expression, even within the same gene cluster. Based on recent genome-wide CTCF-mediated chromatin interactome analysis data using mouse embryonic stem cells (Handoko et al., 2011), we mapped the CTCF-binding sites on the clustered Pcdh locus (Figure 3B). Interestingly, all the stochastically expressed clustered Pcdh isoforms except a8 and b1 had a CTCF-binding site in their promoter region. We also found that the enhancer regions, HS5-1 and Pcdhb cluster control region (CCR), had CTCF-binding sites. Of the C-type isoforms, ac1 had a CTCF-binding site in its promoter region, but ac2 and gc3–gc5 did not. Notably, down- or upregulation of the clustered Pcdh isoforms in CTCF-cKO mice corresponded well to whether their promoter region had a CTCF-binding site or not. These results were consistent with a recent report that CTCF specifically binds to alternative Pcdha promoters in two neuroblastoma cell lines, which express distinct Pcdha isoforms (Monahan et al., 2012). These findings supported the idea that CTCF regulates the generation of single-cell diversity elicited by the stochastic expression of the clustered Pcdh genes. To evaluate the expression patterns of clustered Pcdh isoforms in individual CTCF-deficient neurons, we next performed in situ hybridization for several of them. In the cortex and hippocampus, the loss of CTCF expression was specifically observed in differentiated projection neurons, and not in inhibitory neurons or glia (Figures 4A and 4B). The stochastically expressed isoforms, such as a4, a12, b16, b22, and ga7, were severely downregulated in the projection neurons of CTCF-cKO mice compared with controls (Figures 4C–4F, S5, and S6), whereas the inhibitory neurons and glia, which expressed CTCF in the same

animals, maintained the expression levels of these isoforms. In contrast the expression levels of the C-type isoforms, ac2 and gc3–gc5, were generally increased in the CTCF-cKO mice (Figures 4C–4F, S5, and S6). These results demonstrated that CTCF is a global regulator of the clustered Pcdh family that is especially required for the stochastically expressed isoforms and indicated that CTCF has an important role in the neuronal diversity elicited by the stochastic expression of the clustered Pcdh genes in the brain. Development of Dendritic Arbor Morphology and Dendritic Spine Density in Individual Projection Neurons To examine the morphology of individual projection neurons in the CTCF-cKO mice versus controls, we used Golgi-Cox staining and evaluated the dendritic growth (Figures 5A and 5B). At P7, the apical and basal dendrites showed normal orientation and total dendritic length. However, at P14, the average dendritic length was significantly reduced (Figures 5C and 5D). Sholl analysis (Figures 5E and 5F), which presents dendritic complexity as a function of the distance from the soma, showed that the number of dendritic intersections was significantly lower in the CTCF-cKO mice, especially near the proximal dendritic region, at P14 (Figure 5F). Although the number of dendritic branch points was reduced (Figure 5G), there was no difference in the number of primary dendrites per neuron (Figure 5H). These results indicated that the CTCF deficiency inhibits dendritic arborization during neuronal maturation. Dendritic arborization and spine formation are closely related (Vaughn, 1989; Niell et al., 2004). Therefore, we calculated the number of spines on the dendrites at P14 and found that their density was significantly decreased on the CTCF-deficient neurons (Figures 5I and 5J). These results indicated that the CTCF deficiency also inhibits dendritic spine formation during neuronal maturation. In the CTCF-cKO mice, CTCF expression was also lost in the pyramidal neurons of the hippocampus. Therefore, we examined the morphology of each pyramidal neuron in the CA1 hippocampal region at P14 by Golgi-Cox staining to see if these CTCF-deficient neurons also had dendrite-arborization and spine-formation defects (Figures 6A and 6B). The dendrite lengths were significantly reduced on both the basal and apical sides of the CTCF-deficient neurons (Figures 6C and 6D). Significant reductions in dendritic complexity and spine density of the projection neurons were also found in the CA1 hippocampal neurons at P14, similar to the observations in the cortex (Figures 6E–6G). These data demonstrated that CTCF is important for synapse formation and dendrite development in postmitotic projection neurons during postnatal development. Electrophysiological Properties of CTCF-Deficient Neurons To assess the electrophysiological properties of the CTCFdeficient neurons in P14–P15 mice, we performed whole-cell voltage-clamp recording from layer II/III pyramidal neurons after stimulating layer IV of the somatosensory cortex. AMPA- and NMDA-mediated excitatory postsynaptic currents (EPSCs) were similarly evoked in the CTCF-cKO mice compared with controls, indicating that the AMPA/NMDA ratio was normal (Figures 7A and 7B). There was also no difference in the NMDA Cell Reports 2, 345–357, August 30, 2012 ª2012 The Authors 349

A

B

C

D

E


decay time. These results showed that the CTCF-deficient neurons had physiologically mature synapses. We next recorded the miniature EPSCs (mEPSCs) and found that their frequency was significantly decreased in the CTCF neurons, with no changes in amplitude, rise time, or half-width (Figures 7C and 7D). These results were consistent with the decreased spine density in the CTCF-deleted neurons (Figures 5J and 6G). Thus, the CTCF-deficient neurons showed a defect in their ability to form appropriate numbers of synapses, but not in physiological synapse maturation during brain development. DISCUSSION Here, by studying mice with a conditional loss of CTCF function in postmitotic projection neurons, we demonstrated that CTCF plays critical roles in the cerebral cortex and hippocampus. Functional neuronal networks require appropriate dendritic arborization and synapse formation, which are largely postnatal events. The CTCF-cKO mice were born without apparent abnormalities but showed obvious growth retardation by P7. They also showed abnormal behavior and died within 1 month. These results indicated that the CTCF in postmitotic projection neurons is essential for normal brain development after birth. Using Golgi-Cox staining, we demonstrated that the dendritic length of the pyramidal neurons was decreased in the CTCFdeficient neurons at P14, owing to decreased dendritic arborization, although a significant difference from the control was not detected at P7. Our results indicated that CTCF is required for normal dendritic development. In addition we demonstrated that the spine density was reduced in the CTCF-deficient neurons at P14. Consistent with these histological observations, we observed a decreased frequency of mEPSCs in CTCFdeficient neurons, with normal mEPSC amplitude and AMPA/ NMDA ratio, even at P14. These results indicated that CTCF is required in the postmitotic projection neurons for the elaboration of dendritic arbors and the formation of sufficient synapse numbers during postnatal brain development, but not for individual synapse maturation. During postnatal development in the cortex, functional neural networks are completed in the context of environmental stimulus. Therefore, the abnormalities of individual CTCF-cKO neurons may be dependent on the neural activity induced by sensory input. We also observed that the barrel structure was disrupted in CTCF-cKO mice and that the barrelettes and barreloids of these mice were less clear than those of control mice. The formation of barrel and barreloid structures depends on sensory stimulation from the whiskers. The molecules involved in barrel map forma-

tion have been examined using knockout mice. Most are associated with neuronal activity and cell signaling. However, we did not observe any significant differences in the expression of these molecules in the CTCF-cKO mice (Table S6). NMDA receptors play a central role in excitatory synapse formation and in experience-dependent somatosensory map formation. NMDA-mediated EPSCs are important in the early postnatal period and then are superseded by AMPA-mediated EPSCs during refinement of the somatosensory map. Several previous studies showed that mice with a disrupted barrel map also exhibit an abnormal AMPA/NMDA ratio (Inan et al., 2006; Ince-Dunn et al., 2006; Ultanir et al., 2007; Iwasato et al., 2008). However, we did not detect significant differences in the AMPA/NMDA ratio between the CTCF-cKO and control mice (Figure 7). The CTCF-deficient neurons were defective in processes by which the dendritic arborization and number of spines increase during brain development. Therefore, dendritic arborization and spineformation processes independent of the NMDA pathway might be necessary for proper functional somatosensory mapping. In the CTCF-deficient cortex and hippocampus, we observed altered expression levels of 390 transcripts (Tables S2 and S3). These genes contribute to the phenotypes observed in the CTCF-cKO mice. The expression levels of the clustered Pcdh genes were particularly sensitive to the CTCF defect, especially the stochastically expressed isoforms. Similar phenotypes related to barrel formation and dendritic arborization in the cortex have been observed in mutants with loss of function in the clustered Pcdh genes. In Pcdhb CCR deletion mutant mice, the barrel structure of the somatosensory cortex is impaired (Yokota et al., 2011). In these mice all the Pcdhb isoforms except b1 are greatly downregulated, and the Pcdhg expression is disrupted and downregulated. At the same time, Diap1/mDia1, which is located downstream of the Pcdhg cluster and encodes a ubiquitously expressed Rho effector protein, is also disrupted. It is reported that mDia1 knockout mice develop without apparent abnormality (Sakata et al., 2007). On the other hand, mice with forebrain-restricted loss of the clustered Pcdhg show severely reduced dendritic arbors in their cortical pyramidal neurons (Garrett et al., 2012). These findings are consistent with the idea that the clustered Pcdh gene expressions contribute to the phenotypes observed in the CTCF-cKO mice. Collectively, these results lead to new questions about how the stochastic expression of clustered Pcdh genes in individual neurons is related to dendritic arborization, synapse formation, and functional map formation in the cortex. Here, we have demonstrated that the stochastically expressed isoforms of clustered Pcdh were markedly downregulated in

Figure 3. Severe Downregulation of the Stochastically Expressed Isoforms of the Clustered Pcdh Genes in the CTCF-cKO Cortex and Hippocampus (A) Chromosomal location of up- and downregulated genes and the frequency of expression changes identified by microarray analysis. The transcriptional start sites of altered genes are plotted on each chromosome with their frequency. The highest frequency was observed in the clustered Pcdh locus in chromosome 18 (asterisk). See also Tables S2, S3, S4, S5, and S6. (B) Genomic structure of the clustered Pcdh gene locus and binding sites for CTCF. Red triangles indicate CTCF-binding sites. Rhombuses indicate enhancer regions (HS5–HS1 enhances a3–a12 and ac1; CCR enhances b1–b22). (C and D) Relative gene expression of the clustered Pcdh gene locus in the CTCF-cKO mouse in the P7 cortex (C) and hippocampus (D) compared to control. Asterisks indicate that quantitative reverse-transcription PCR validation was performed. (E) Validation of microarray results by quantitative reverse-transcription PCR analysis. The stochastically expressed isoforms are shown as red characters. *p < 0.05. **p < 0.01. Error bars represent SEM. aCR and gCR represent constant region exons of the Pcdha and Pcdhg clusters, respectively.

Cell Reports 2, 345–357, August 30, 2012 ª2012 The Authors 351

A

Control

CTCF-cKO

B

CTCF-cKO

Control CA1

Anti-CTCF

Anti-CTCF

Control

C

CA1

Control

cKO

CTCF-cKO

CA2

CA3

CA3

Control

th

D

CA2

cKO

th

CTCF-cKO

Control CA1

Pcdh 4

Pcdh 4

CA1 CA2 CA2

CA3

CA3

CA1

Pcdh c2

Pcdh c2

CA1 CA2

CA2

CA3

CTCF-cKO

Pcdh 4

Control

Pcdh 16

Pcdh 16 Pcdh a7 Pcdh c2

F

Pcdh a7

CTCF-cKO

Pcdh c2

Control

Pcdh 4

E

CA3

Figure 4. In Situ Hybridization Analysis of Clustered Pcdh Expression in the Cortical and Hippocampal Regions in the CTCF-cKO Mouse (A and B) Anti-CTCF immunostaining patterns in the cerebral cortex (A) and hippocampus (B) at P0 were compared between CTCF-cKO and control mice. In CTCF-cKO mice, only subpopulation of cells, which are presumably inhibitory neurons and glial cells, were immunopositive. th, thalamus. (C–F) Expression patterns of the clustered Pcdh genes were examined by in situ hybridization at P7. (C and D) Example of the expression patterns of a stochastically expressed isoform, a4 (red), and a constitutively expressed C-type isoform, ac2, in the cortex (C) and hippocampus (D). In the CTCF-cKO mice the ac2 expression was constitutive and higher than in control mice. In contrast the a4 expression was dramatically downregulated in both the cortex and hippocampus of the CTCF-cKO mice. (E and F) High-magnification images of the boxed regions in (C) and (D), and other stochastically expressed isoforms, b16 and ga7. (See additional results in Figures S5 and S6.) The stochastically expressed isoforms are shown as red characters. Scale bars, 100 mm.

individual CTCF-cKO neurons. These isoforms have CTCFbinding sites in their promoter regions, except for a8 and b1, which are normally expressed at lower levels than other isoforms 352 Cell Reports 2, 345–357, August 30, 2012 ª2012 The Authors

in the brain. The C-type isoforms, ac2 and gc3–gc5, which are expressed by most neurons, were not downregulated by the CTCF deficiency. These C-type isoforms do not have

E

CTCF-cKO

Control

Basal dendritic length (mm) Apical dendritic length (mm)

D 1.5

Control CTCF-cKO

1.0

*

0.5

0 1.5

P7 Control CTCF-cKO

P14 *

1.0

0.5

0

P7

F

Number of intersections

P14


P7

C

CTCF-cKO

B

Control CTCF-cKO

10 5 0

0

100 200 300 Distance from soma (µm)

20

Control CTCF-cKO

15 10 5 0

P14

0

I H

Nodes per cell

40 30

Control CTCF-cKO

*

20 10

Primary dendrites

G

8

J

Control CTCF-cKO

6 4 2 0

0

P7

P14

P7

100 200 300 Distance from soma (µm)

P14

Control

CTCF-cKO

Density of spine and filopodia (µm)

Control

A

400

1.5

1.0

* 0.5

0.0

Control CTCF-cKO

Figure 5. Decrease in Dendritic Arborization and Spine Number in the Cortical Projection Neurons (A and B) Golgi staining of the cerebral cortex in control (A) and CTCF-cKO (B) at P14. Scale bars, 50 mm. (C) Example of traced pyramidal neurons of cortical layer II/III stained by the Golgi method. Scale bars, 50 mm. (D) Quantitative analysis of the apical and basal total dendritic length. n > 30 neurons from five mice for each group. (E) Sholl analysis for dendritic complexity at P7. (F) Sholl analysis for dendritic complexity at P14. (G) Total branch points of dendrites. (H) Number of primary dendrites. (I) Apical dendritic spines of cortical pyramidal neurons (layer II/III). Scale bar, 5 mm. (J) Quantitative analysis of spine density at P14. n > 30 neurons from five mice for each group. *p < 0.01. Error bars represent SEM.

CTCF-binding sites within their promoter region. Thus, our results indicate that CTCF-dependent chromatin interactions between promoters and enhancers regulate the stochastic and

combinatorial expression of clustered Pcdh isoforms in postmitotic projection neurons. Each neuron might therefore have a distinct chromatin architecture resulting at least in part from Cell Reports 2, 345–357, August 30, 2012 ª2012 The Authors 353

B

Control

Control

CTCF-cKO

D

2.0

2.0

Basal dendritic length (mm)

C

CTCF-cKO

Apical dendritic length (mm)

A

1.5

*

1.0

0.5

0

F


E

CTCF-cKO

G

Control CTCF-cKO

25 20 15 10 5 0 0

100

200

300

Distance from soma (µm)

400

Control

CTCF-cKO

Density of spine and filopodia (µm)

Control

1.5

1.0

*

0.5

0

Control

CTCF-cKO

1.5

* 1.0

0.5

0.0

Control

CTCF-cKO

Figure 6. Decrease in Dendritic Arborization and Spine Number in the Hippocampal Neurons (A and B) Golgi staining of the hippocampus in control (A) and CTCF-cKO (B) at P14. Scale bar, 50 mm. (C) Example of traced pyramidal neurons of the hippocampal CA1 at P14. Scale bar, 50 mm. (D) Quantitative analysis of dendrite length. (E) Sholl analysis for dendritic complexity. (F) Apical dendritic spines of a CA1 pyramidal neuron. Scale bar, 5 mm. (G) Quantitative analysis of the spine density. n > 30 neurons from five mice for each group. *p < 0.01. Error bars show SEM.

the CTCF-directed differential expression of clustered Pcdh genes. CTCF colocalizes with cohesin (Parelho et al., 2008; Wendt et al., 2008), and together, these proteins provide the basis for the cell-type-specific chromatin conformation (Hadjur et al., 2009; Hou et al., 2010) that enables specific transcriptions. Nipped B-like (NIPBL) facilitates cohesin’s loading onto DNA in chromosomal events, and mutations in NIPBL are responsible 354 Cell Reports 2, 345–357, August 30, 2012 ª2012 The Authors

for Cornelia de Lange syndrome (CdLS), which is characterized by severe mental and growth retardation (Krantz et al., 2004; Tonkin et al., 2004). Nipbl heterozygous mice exhibit characteristics of CdLS, and Pcdhb gene transcription is the most sensitive to Nipbl defects in the brain (Kawauchi et al., 2009). Cohesin-SA1 was found to regulate expression of the clustered Pcdh genes in mouse brain, using an embryonic lethal cohesinSA1 KO mutant (Remeseiro et al., 2012).

A

B

1.0

AMPA/NMDA

Control

0.5

0.0 20 pA 100 ms

CTCF-cKO

Control

CTCF-cKO

300 NMDA decay (ms)

CTCF-cKO

Control

200

100

0

C

D

EXPERIMENTAL PROCEDURES

3 2

*

1

Rise time (ms)

Frequency (Hz)

2

Control

0

Control CTCF-cKO

4

Half width (ms)

100 ms

Amplitude (pA)

CTCF-cKO

1

0

Control CTCF-cKO

10 pA

10

5

0

3 2 1 0

Control CTCF-cKO

isoforms (Monahan et al., 2012), which are regulated stochastically in individual neurons (Yokota et al., 2011). These recent results, the results of Remeseiro et al. (2012), and our findings indicate that the CTCF- and cohesin-dependent chromatin architecture contributes to the stochastic expression of the clustered Pcdh genes in individual neurons. In conclusion we found developmental defects, including the improper refinement of somatosensory maps and decreased dendritic arborization and synapse formation, in CTCF-cKO mice. In these mice, neuronal diversity, at least that involving the differential and stochastic expression of clustered Pcdh, was disrupted. Thus, our results suggest that CTCF is a key regulator of neuronal differentiation that is required to generate neuronal diversity and form functional neural networks. Furthermore, our results reveal that the significance of neuronal diversity in the vertebrate nervous system can be studied by manipulating CTCF in neurons.

Control CTCF-cKO

Figure 7. Comparison of Electrophysiological Properties between Control and CTCF-cKO Neurons (A) Current maps show representative superimposed traces of NMDAmediated EPSCs (positive) and AMPA-mediated EPSCs (negative) (average of 20 sweeps). Top, control; bottom, CTCF-cKO. (B) The amplitude ratio (mean ± SEM) of AMPA-EPSCs to NMDA-EPSCs (top) and decay time constant of NMDA-EPSCs (bottom). Number of cells is nine (control) and ten (CTCF-cKO). (C) Representative traces of mEPSCs. (D) Mean frequency, amplitude, rise time, and half width of mEPSCs. Number of cells is 9 (control) and 11 (cKO). *p < 0.01. Error bars represent SEM.

Monahan et al. (2012) recently demonstrated that both CTCF and the cohesin complex subunit Rad21 bind to transcriptionally active Pcdha promoters, in correlation with alternative isoform expression, and to the Pcdha HS5–HS1 enhancer in two mouse neuroblastoma cell lines: CAD and N2a. In CAD cells the downregulation of Rad21 decreases the expression of the a12, ac1, and ac2 isoforms, and CTCF knockdown reduces the expression of alternative Pcdha isoforms, but not of the constitutively expressed ac1 and ac2 isoforms (Monahan et al., 2012). Their result for ac1 expression is different from our findings in CTCFdeficient neurons (Figure 3). Nevertheless, the results of Monahan et al. (2012) strongly support the existence of a molecular mechanism in which CTCF regulates the stochastic expression of clustered Pcdh genes in individual neurons. In addition both CTCF and Rad21 can bind to HS17–HS20 in the CCR for Pcdhb

Animals All the experimental procedures were in accordance with the guide for the care and use of laboratory animals of the science council of Japan and were approved by the animal experiment committee of Osaka University. The Ctcf-floxed (Ctcfflox) mice, in which loxP sites flank exons 3–12, were described previously by Heath et al. (2008). To generate postmitotic neuronspecific Ctcf cKO mice, Ctcfflox/flox mice were bred to Ctcf+/flox; Nex+/Cre mice. Nex-Cre mice were generated using a knockin strategy and express Cre recombinase in differentiated neurons, particularly hippocampal and cortical projection neurons (Goebbels et al., 2006). Genotyping was carried out by PCR on DNA extracted from tail tissue. Cre-loxP-mediated recombination was confirmed by Southern blotting of DNA derived from the cortex and hippocampus at P1. PCR primers and Southern blotting probes were described previously by Heath et al. (2008). Homo- or heterozygous mice for the Ctcf-floxed allele that did not contain the Cre gene were used as controls. Histological Analysis Mice were anesthetized and perfused with 4% paraformaldehyde in phosphate buffer (pH 7.4). The brain was removed and postfixed in the same fixative overnight at 4 C, then cryoprotected in 20% sucrose in phosphate buffer. Frozen sections prepared on a microtome were cut 40 mm thick, and those prepared on a cryostat were 10 mm thick. Sections were stained with cresyl violet, anti-CTCF (Millipore), anti-ssDNA (IBL), and anti-neurofilament (clone 2H3; Developmental Studies Hybridoma Bank). For somatosensory map analysis, fixed tissues were cut into 50 mm sections. For Nissl staining the sections were stained with cresyl violet solution. For CO staining the sections were incubated in 4% sucrose, 0.05% cytochrome c, and 0.05% DAB in 0.1 M phosphate buffer (pH 7.4) at 37 C. For 5-HTT immunohistochemistry the sections were incubated in blocking buffer (20% Block Ace, 5% normal donkey serum, 0.1% Triton X-100 in PBS) for 1 hr at room temperature, an anti-5HTT antibody (1:500 dilution; Santa Cruz Biotechnology) was added in antibody dilution buffer (5% Block Ace, 5% normal donkey serum, 0.1% Triton X-100 in PBS), and the sections were incubated overnight at 4 C. The sections were rinsed in PBS and incubated for 1 hr at room temperature with a biotinylated anti-goat secondary antibody, and the signal was enhanced using the VECTASTAIN Elite ABC kit (Vector Laboratories) and visualized by DAB staining. Detailed protocols for DiI and DiA tracing are provided in the Extended Experimental Procedures. Golgi-Cox Staining Golgi-Cox staining was performed using the FD Rapid GolgiStain kit (FD NeuroTechnologies). Freshly dissected brain was immersed in the kit’s impregnation solution for 2 weeks. The frozen tissue was sectioned at 150 mm on


a microtome, followed by staining according to the manufacturer’s procedures. For the neuronal morphometric analysis, pyramidal neurons were randomly selected from somatosensory cortical layer II/III and hippocampal CA1. Photomicrographs were acquired (2 or 0.2 mm pitch) using a light microscope at 403 or 1003 magnification (BZ-9000; Keyence), and the dendrites and dendritic spines were traced using Neurolucida software (MicroBrightField). The total dendritic length, Sholl analysis, and spine density were also calculated by Neurolucida. The spinal density was obtained from the apical dendrites within the region 50–150 mm from the soma (total length 1,300– 1,800 mm). Statistical analyses were performed using the unpaired t test with GraphPad Prism 5 software (GraphPad). Microarray Analysis We used the Affymetrix Mouse Exon 1.0 ST Array to investigate gene expression. We used only male mice to eliminate potential confounds from gender differences. Total RNA was extracted from the cerebral cortex (n = 4) and hippocampus (n = 3) at P7, using TRIzol (Invitrogen), then purified with an RNeasy mini column (QIAGEN). Hybridization cocktail preparation and hybridization to the array were performed according to the Affymetrix GeneChip Whole Transcript (WT) Sense Target Labeling Assay Manual. After being washed and stained, the arrays were scanned with an Affymetrix GCS3000 7G, using the Affymetrix GeneChip Command Console (AGCC). The array data were analyzed using the Affymetrix GeneChip Expression Console and Partek Genomics Suite Version 6.4 (Partek). The probe set intensities were normalized using the RMA algorithm. Statistical analysis was performed by ANOVA using the core probe sets defined by Affymetrix. A list of genes showing a significant difference by Ctcf deletion was generated using a FDR 1.2 as a cutoff value. Gene ontology analysis was performed using the DAVID Bioinformatics Resources 6.7 (http://david.abcc.ncifcrf.gov/) (Huang et al., 2009). Quantitative Reverse-Transcription PCR Analysis Total RNA was extracted as described for the microarray analysis and treated with DNase I (Takara). cDNA was synthesized from 1 mg of the total RNA using SuperScript III reverse transcriptase (Invitrogen) and random hexamers. Quantitative reverse-transcription PCRs were performed on the ABI prism7900HT Sequence Detection System using SYBR-Green and gene-specific primers (Table S7). All RNA samples were analyzed in duplicate, and at least four independent analyses were performed. The expression of each gene was normalized to that of Gapdh. Statistical analyses were performed using the unpaired t test with GraphPad Prism 5 software. In Situ Hybridization The brain was removed and embedded in OCT compound as quickly as possible, then frozen in isopentane cooled with liquid nitrogen. The frozen tissue was cut into 10 mm sections on a cryostat. In situ hybridization was performed as described previously by Noguchi et al. (2009) and SchaerenWiemers and Gerfin-Moser (1993). Digoxigenin (DIG)-labeled RNA probes were synthesized from cDNA clones using the DIG RNA Labeling Mix (Roche). The RNA probes are listed in Table S8. Electrophysiology Detailed protocols for electrophysiology are provided in the Extended Experimental Procedures. ACCESSION NUMBERS Array data have been deposited in the GEO database (GSE38673).

LICENSING INFORMATION This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 Unported License (CC-BY; http://creativecommons. org/licenses/by/3.0/legalcode). ACKNOWLEDGMENTS We thank K.-A. Nave for the donation of the Nex-Cre mice and N. Yamamoto for the Tbr1 and RorB in situ hybridization probes. We also thank members of our laboratories, especially S. Yokota, for helpful discussions. This work was supported by a Grant-in-Aid for Scientific Research (S) (JSPS), Innovative Areas ‘‘Mesoscopic Neurocircuitry’’ (No. 23115513) and (Comprehensive Brain Science Network) from the Ministry of Education, Science, Sports, and Culture of Japan (MEXT), JST-CREST (to T.Y.), and by the Dutch Cancer Society (KWF) (to N.G.). Received: May 29, 2012 Revised: June 13, 2012 Accepted: June 15, 2012 Published online: July 26, 2012 REFERENCES Esumi, S., Kakazu, N., Taguchi, Y., Hirayama, T., Sasaki, A., Hirabayashi, T., Koide, T., Kitsukawa, T., Hamada, S., and Yagi, T. (2005). Monoallelic yet combinatorial expression of variable exons of the protocadherin-alpha gene cluster in single neurons. Nat. Genet. 37, 171–176. Fedoriw, A.M., Stein, P., Svoboda, P., Schultz, R.M., and Bartolomei, M.S. (2004). Transgenic RNAi reveals essential function for CTCF in H19 gene imprinting. Science 303, 238–240. Garrett, A.M., Schreiner, D., Lobas, M.A., and Weiner, J.A. (2012). g-protocadherins control cortical dendrite arborization by regulating the activity of a FAK/ PKC/MARCKS signaling pathway. Neuron 74, 269–276. Goebbels, S., Bormuth, I., Bode, U., Hermanson, O., Schwab, M.H., and Nave, K.A. (2006). Genetic targeting of principal neurons in neocortex and hippocampus of NEX-Cre mice. Genesis 44, 611–621. Golan-Mashiach, M., Grunspan, M., Emmanuel, R., Gibbs-Bar, L., Dikstein, R., and Shapiro, E. (2012). Identification of CTCF as a master regulator of the clustered protocadherin genes. Nucleic Acids Res. 40, 3378–3391. Guo, C., Yoon, H.S., Franklin, A., Jain, S., Ebert, A., Cheng, H.-L., Hansen, E., Despo, O., Bossen, C., Vettermann, C., et al. (2011). CTCF-binding elements mediate control of V(D)J recombination. Nature 477, 424–430. Hadjur, S., Williams, L.M., Ryan, N.K., Cobb, B.S., Sexton, T., Fraser, P., Fisher, A.G., and Merkenschlager, M. (2009). Cohesins form chromosomal cis-interactions at the developmentally regulated IFNG locus. Nature 460, 410–413. Handoko, L., Xu, H., Li, G., Ngan, C.Y., Chew, E., Schnapp, M., Lee, C.W., Ye, C., Ping, J.L., Mulawadi, F., et al. (2011). CTCF-mediated functional chromatin interactome in pluripotent cells. Nat. Genet. 43, 630–638. Hasegawa, S., Hamada, S., Kumode, Y., Esumi, S., Katori, S., Fukuda, E., Uchiyama, Y., Hirabayashi, T., Mombaerts, P., and Yagi, T. (2008). The protocadherin-alpha family is involved in axonal coalescence of olfactory sensory neurons into glomeruli of the olfactory bulb in mouse. Mol. Cell. Neurosci. 38, 66–79. Heath, H., Ribeiro de Almeida, C., Sleutels, F., Dingjan, G., van de Nobelen, S., Jonkers, I., Ling, K.W., Gribnau, J., Renkawitz, R., Grosveld, F., et al. (2008). CTCF regulates cell cycle progression of alphabeta T cells in the thymus. EMBO J. 27, 2839–2850.

SUPPLEMENTAL INFORMATION

Hirayama, T., and Yagi, T. (2006). The role and expression of the protocadherin-alpha clusters in the CNS. Curr. Opin. Neurobiol. 16, 336–342.

Supplemental Information includes Extended Experimental Procedures, six figures, and eight tables and can be found with this article online at http://dx.doi.org/10.1016/j.celrep.2012.06.014.

Hou, C., Dale, R., and Dean, A. (2010). Cell type specificity of chromatin organization mediated by CTCF and cohesin. Proc. Natl. Acad. Sci. USA 107, 3651–3656.


Huang da, W., Sherman, B.T., and Lempicki, R.A. (2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4, 44–57.

Remeseiro, S., Cuadrado, A., Go´mez-Lo´pez, G., Pisano, D.G., and Losada, A. (2012). A unique role of cohesin-SA1 in gene regulation and development. EMBO J. 31, 2090–2102.

Inan, M., and Crair, M.C. (2007). Development of cortical maps: perspectives from the barrel cortex. Neuroscientist 13, 49–61.

Sakata, D., Taniguchi, H., Yasuda, S., Adachi-Morishima, A., Hamazaki, Y., Nakayama, R., Miki, T., Minato, N., and Narumiya, S. (2007). Impaired T lymphocyte trafficking in mice deficient in an actin-nucleating protein, mDia1. J. Exp. Med. 204, 2031–2038.

Inan, M., Lu, H.C., Albright, M.J., She, W.C., and Crair, M.C. (2006). Barrel map development relies on protein kinase A regulatory subunit II beta-mediated cAMP signaling. J. Neurosci. 26, 4338–4349. Ince-Dunn, G., Hall, B.J., Hu, S.C., Ripley, B., Huganir, R.L., Olson, J.M., Tapscott, S.J., and Ghosh, A. (2006). Regulation of thalamocortical patterning and synaptic maturation by NeuroD2. Neuron 49, 683–695. Iwasato, T., Inan, M., Kanki, H., Erzurumlu, R.S., Itohara, S., and Crair, M.C. (2008). Cortical adenylyl cyclase 1 is required for thalamocortical synapse maturation and aspects of layer IV barrel development. J. Neurosci. 28, 5931–5943. Kaneko, R., Kato, H., Kawamura, Y., Esumi, S., Hirayama, T., Hirabayashi, T., and Yagi, T. (2006). Allelic gene regulation of Pcdh-a and Pcdh-g clusters involving both monoallelic and biallelic expression in single Purkinje cells. J. Biol. Chem. 281, 30551–30560. Katori, S., Hamada, S., Noguchi, Y., Fukuda, E., Yamamoto, T., Yamamoto, H., Hasegawa, S., and Yagi, T. (2009). Protocadherin-alpha family is required for serotonergic projections to appropriately innervate target brain areas. J. Neurosci. 29, 9137–9147. Kawauchi, S., Calof, A.L., Santos, R., Lopez-Burks, M.E., Young, C.M., Hoang, M.P., Chua, A., Lao, T., Lechner, M.S., Daniel, J.A., et al. (2009). Multiple organ system defects and transcriptional dysregulation in the Nipbl(+/-) mouse, a model of Cornelia de Lange Syndrome. PLoS Genet. 5, e1000650. Kehayova, P., Monahan, K., Chen, W., and Maniatis, T. (2011). Regulatory elements required for the activation and repression of the protocadherina gene cluster. Proc. Natl. Acad. Sci. USA 108, 17195–17200. Kim, T.H., Abdullaev, Z.K., Smith, A.D., Ching, K.A., Loukinov, D.I., Green, R.D., Zhang, M.Q., Lobanenkov, V.V., and Ren, B. (2007). Analysis of the vertebrate insulator protein CTCF-binding sites in the human genome. Cell 128, 1231–1245.

Schaeren-Wiemers, N., and Gerfin-Moser, A. (1993). A single protocol to detect transcripts of various types and expression levels in neural tissue and cultured cells: in situ hybridization using digoxigenin-labelled cRNA probes. Histochemistry 100, 431–440. Schreiner, D., and Weiner, J.A. (2010). Combinatorial homophilic interaction between gamma-protocadherin multimers greatly expands the molecular diversity of cell adhesion. Proc. Natl. Acad. Sci. USA 107, 14893–14898. Seitan, V.C., Krangel, M.S., and Merkenschlager, M. (2012). Cohesin, CTCF and lymphocyte antigen receptor locus rearrangement. Trends Immunol. 33, 153–159. Soshnikova, N., Montavon, T., Leleu, M., Galjart, N., and Duboule, D. (2010). Functional analysis of CTCF during mammalian limb development. Dev. Cell 19, 819–830. Sperry, R.W. (1963). Chemoaffinity in the orderly growth of nerve fiber patterns and connections. Proc. Natl. Acad. Sci. USA 50, 703–710. Splinter, E., Heath, H., Kooren, J., Palstra, R.J., Klous, P., Grosveld, F., Galjart, N., and de Laat, W. (2006). CTCF mediates long-range chromatin looping and local histone modification in the beta-globin locus. Genes Dev. 20, 2349–2354. Tasic, B., Nabholz, C.E., Baldwin, K.K., Kim, Y., Rueckert, E.H., Ribich, S.A., Cramer, P., Wu, Q., Axel, R., and Maniatis, T. (2002). Promoter choice determines splice site selection in protocadherin alpha and gamma pre-mRNA splicing. Mol. Cell 10, 21–33. Tonkin, E.T., Wang, T.J., Lisgo, S., Bamshad, M.J., and Strachan, T. (2004). NIPBL, encoding a homolog of fungal Scc2-type sister chromatid cohesion proteins and fly Nipped-B, is mutated in Cornelia de Lange syndrome. Nat. Genet. 36, 636–641. Ultanir, S.K., Kim, J.E., Hall, B.J., Deerinck, T., Ellisman, M., and Ghosh, A. (2007). Regulation of spine morphology and spine density by NMDA receptor signaling in vivo. Proc. Natl. Acad. Sci. USA 104, 19553–19558.

Kohmura, N., Senzaki, K., Hamada, S., Kai, N., Yasuda, R., Watanabe, M., Ishii, H., Yasuda, M., Mishina, M., and Yagi, T. (1998). Diversity revealed by a novel family of cadherins expressed in neurons at a synaptic complex. Neuron 20, 1137–1151.

Vaughn, J.E. (1989). Fine structure of synaptogenesis in the vertebrate central nervous system. Synapse 3, 255–285.

Krantz, I.D., McCallum, J., DeScipio, C., Kaur, M., Gillis, L.A., Yaeger, D., Jukofsky, L., Wasserman, N., Bottani, A., Morris, C.A., et al. (2004). Cornelia de Lange syndrome is caused by mutations in NIPBL, the human homolog of Drosophila melanogaster Nipped-B. Nat. Genet. 36, 631–635.

Wan, L.B., Pan, H., Hannenhalli, S., Cheng, Y., Ma, J., Fedoriw, A., Lobanenkov, V., Latham, K.E., Schultz, R.M., and Bartolomei, M.S. (2008). Maternal depletion of CTCF reveals multiple functions during oocyte and preimplantation embryo development. Development 135, 2729–2738.

Monahan, K., Rudnick, N.D., Kehayova, P.D., Pauli, F., Newberry, K.M., Myers, R.M., and Maniatis, T. (2012). Role of CCCTC binding factor (CTCF) and cohesin in the generation of single-cell diversity of Protocadherin-a gene expression. Proc. Natl. Acad. Sci. USA 109, 9125–9130.

Wang, X., Su, H., and Bradley, A. (2002). Molecular mechanisms governing Pcdh-gamma gene expression: evidence for a multiple promoter and cis-alternative splicing model. Genes Dev. 16, 1890–1905.

Muotri, A.R., and Gage, F.H. (2006). Generation of neuronal variability and complexity. Nature 441, 1087–1093. Niell, C.M., Meyer, M.P., and Smith, S.J. (2004). In vivo imaging of synapse formation on a growing dendritic arbor. Nat. Neurosci. 7, 254–260. Noguchi, Y., Hirabayashi, T., Katori, S., Kawamura, Y., Sanbo, M., Hirabayashi, M., Kiyonari, H., Nakao, K., Uchimura, A., and Yagi, T. (2009). Total expression and dual gene-regulatory mechanisms maintained in deletions and duplications of the Pcdha cluster. J. Biol. Chem. 284, 32002–32014. Ohlsson, R., Bartkuhn, M., and Renkawitz, R. (2010). CTCF shapes chromatin by multiple mechanisms: the impact of 20 years of CTCF research on understanding the workings of chromatin. Chromosoma 119, 351–360.

Weiner, J.A., Wang, X., Tapia, J.C., and Sanes, J.R. (2005). Gamma protocadherins are required for synaptic development in the spinal cord. Proc. Natl. Acad. Sci. USA 102, 8–14. Wendt, K.S., Yoshida, K., Itoh, T., Bando, M., Koch, B., Schirghuber, E., Tsutsumi, S., Nagae, G., Ishihara, K., Mishiro, T., et al. (2008). Cohesin mediates transcriptional insulation by CCCTC-binding factor. Nature 451, 796–801. Wu, Q., and Maniatis, T. (1999). A striking organization of a large family of human neural cadherin-like cell adhesion genes. Cell 97, 779–790. Yagi, T. (2008). Clustered protocadherin family. Dev. Growth Differ. 50 (Suppl 1), S131–S140. Yagi, T. (2012). Molecular codes for neuronal individuality and cell assembly in the brain. Front. Mol. Neurosci. 5, 45.

Parelho, V., Hadjur, S., Spivakov, M., Leleu, M., Sauer, S., Gregson, H.C., Jarmuz, A., Canzonetta, C., Webster, Z., Nesterova, T., et al. (2008). Cohesins functionally associate with CTCF on mammalian chromosome arms. Cell 132, 422–433.

Yokota, S., Hirayama, T., Hirano, K., Kaneko, R., Toyoda, S., Kawamura, Y., Hirabayashi, M., Hirabayashi, T., and Yagi, T. (2011). Identification of the cluster control region for the protocadherin-b genes located beyond the protocadherin-g cluster. J. Biol. Chem. 286, 31885–31895.

Phillips, J.E., and Corces, V.G. (2009). CTCF: master weaver of the genome. Cell 137, 1194–1211.

Zipursky, S.L., and Sanes, J.R. (2010). Chemoaffinity revisited: dscams, protocadherins, and neural circuit assembly. Cell 143, 343–353.
