|Home | About | Journals | Submit | Contact Us | Français|
The genetic basis of cognition and behavioral adaptation to the environment remains poorly understood. Here we demonstrate that the histone methyltransferase complex GLP/G9a controls cognition and adaptive responses in a region-specific fashion in the adult brain. Using conditional mutagenesis in mice, we show that postnatal, neuron-specific deficiency of GLP/G9a leads to de-repression of numerous non-neuronal and neuron progenitor genes in adult neurons. This transcriptional alteration is associated with complex behavioral abnormalities, including defects in learning, motivation and environmental adaptation. The behavioral changes triggered by GLP/G9a deficiency are similar to key symptoms of the human 9q34 mental retardation syndrome that is associated with structural alterations of the GLP gene. The likely causal role of GLP/G9a in mental retardation in mice and humans suggests a key role for the GLP/G9a controlled histone H3K9 di-methylation in regulation of brain function through maintenance of the transcriptional homeostasis in adult neurons.
A possible clue to the genetic mechanisms underlying cognition and environmental adaptation could be obtained through identification of genes responsible for mental retardation disorders. Recent studies show that many of the proteins associated with mental retardation syndromes in mice and humans are involved in epigenetic control of gene expression (Chelly et al., 2006; Inlow and Restifo, 2004; Kramer and van Bokhoven, 2009). It has been found that altered expression of histone acetyltransferases and deacetylases as well as of proteins involved in histone or DNA methylation are associated with mental retardation syndromes (Kramer and van Bokhoven, 2009). These epigenetic regulators govern expression of large numbers of potentially unrelated genes. Yet, despite differences in specific functions of the distinct epigenetic regulators and their targets, impairment of chromatin function leads to symptomatically similar mental retardation syndromes (Chelly et al., 2006; Inlow and Restifo, 2004; Kramer and van Bokhoven, 2009). Therefore, it is conceivable that mental retardation is triggered not by changes in specific target gene(s) but by the inability of neurons to respond adequately to environmental signals under conditions of greatly distorted transcriptional homeostasis (Ramocki and Zoghbi, 2008).
Transcriptional homeostasis relies largely on the balance between positive and negative regulation of gene transcription. In mammals, di- or tri-methylation of histone H3 on lysine 9 (H3K9me2 and H3K9me3) is commonly associated with gene silencing in eu- or hetero-chromatin respectively (Jenuwein and Allis, 2001; Kouzarides, 2007; Zhang and Reinberg, 2001). The H3K9me2 is catalyzed by an enzymatic complex comprised of the histone methyltransferases G9a and G9a-like protein GLP (Ogawa et al., 2002; Shi et al., 2003; Peters et al., 2003; Rice et al., 2003; Tachibana et al., 2001; Tachibana et al., 2002; Tachibana et al., 2005). Genetic ablation of either of these proteins results in loss of euchromatic H3K9me2 (Peters et al., 2003; Rice et al., 2003; Sampath et al., 2007; Tachibana et al., 2001; Tachibana et al., 2002; Tachibana et al., 2005). Despite the genome-wide distribution of H3K9me2, the scope and pattern of gene expression changes in the absence of this mark is rather limited and depends on the affected cell type (Kondo et al., 2008; Sampath et al., 2007; Tachibana et al., 2002; Wagschal et al., 2008). It has been demonstrated recently that aberrant function of G9a in non-neuronal cells leads to de-repression of neuronal genes outside of the nervous system (Ding et al., 2008; Roopra et al., 2004; Tahiliani et al., 2007). This finding raises a question about the potential ability of the GLP/G9a complex to function as a master regulator of lineage-specific gene expression in the brain. In support of this model we show that postnatal neuron-specific ablation of the histone methyltransferases GLP/G9a leads to de-repression of non-neuronal and early neuron progenitor genes in various brain areas. We also demonstrate that these changes in gene expression are associated with cognitive and behavioral defects in adult mice. The impaired brain function manifests itself either by a mental retardation-like syndrome in mice with forebrain-specific ablation of GLP/G9a, or by more selective behavioral defects that follow GLP/G9a deficiency in specialized neuronal subpopulations. Collectively, our findings have revealed GLP/G9a as a key regulator of cognition, motivation and adaptive behavior in adult mice.
To address the role of GLP and G9a in adult brain function, we employed conditional GLP and G9a gene inactivation in postnatal mouse neurons. Gene exons encoding the SET-domain of GLP (Fig. 1A) or G9a (Sampath et al., 2007) were flanked with loxP sites in vivo. The SET-domain is essential for the histone lysine methyltransferase activity of GLP and G9a (Tachibana et al., 2002; Tachibana et al., 2005). The functional inactivation of GLP and G9a genes in postnatal neurons was accomplished by using Cre recombinase driven by the Ca2+/Calmodulin-dependent protein kinase II alpha gene promoter (Camk2a-Cre). This promoter is specifically active in postnatal forebrain neurons and Camk2a-Cre mice have been successfully employed for conditional modification of numerous genes in the forebrain (Casanova et al., 2001; Zhang et al., 2004). In addition to Camk2a-Cre mice, we also used Cre expressing mice that enable conditional ablation of GLP or G9a specifically in dopamine 2 receptor (Drd2) or dopamine 1 receptor (Drd1) expressing medium spiny neurons (MSNs) in the striatum (Gong et al., 2007).
The immunohistochemical analysis of brain tissues from Camk2a-Cre; GLPfl/fl mice showed GLP deficiency in neurons of the cortex, hippocampus, striatum and hypothalamus (Fig. 1B). Postnatal, forebrain-specific deficiency of GLP led to drastic reduction in neuronal euchromatic H3K9me2 levels (Fig. 1C and Supplementary Fig. S1). Similar changes in euchromatic H3K9me2 levels occurred in neurons of Camk2a-Cre mice homozygous for the mutant G9a allele (Camk2a-Cre; G9afl/fl) (Fig. 1C), confirming the previously reported essential role of the GLP/G9a complex in euchromatic H3K9me2 (Sampath et al., 2007; Tachibana et al., 2005). Similarly, the Cre-mediated alteration of G9a specifically in Drd1- or Drd2-expressing neurons in the striatum resulted in loss of euchromatic H3K9me2 (Supplementary Fig. S1). As previously shown for G9a deficient ES-cells and fibroblasts (Peters et al., 2003; Sampath et al., 2007), the GLP/G9a deficient neurons still show heterochromatic H3K9me2 (Fig. 1C, Supplementary Fig. S1, S2), which may derive from enzymatic demethylation of heterochromatic H3K9me3. In agreement with the previously reported selective role of GLP/G9a in formation of euchromatic H3K9me2 in cells of various types (Peters et al., 2003; Rice et al., 2003; Sampath et al., 2007; Tachibana et al., 2002; Tachibana et al., 2005), the postnatal, neuron-specific deficiency of GLP and G9a had no impact on neuronal heterochromatic H3K9me3 (Supplementary Fig. S2).
Conditional inactivation of GLP or G9a in striatum, hippocampus, hypothalamus, or cortex led to up-regulation of a large number of genes (Fig. 2A, Supplementary Table S1-9). Comparison of gene expression within the functionally distinct, GLP or G9a deficient brain regions, as well as the cross-area comparison, revealed about 60 genes that collectively comprise a genetic “signature” of GLP or G9a deficiency in the adult forebrain (Fig. 2B-D). In addition to the commonly expressed genes, each of the individual G9a or GLP deficient brain areas shows up-regulation of additional neuronal genes depending on the specific brain region (Supplementary Table S2-S9). Combined GLP and G9a deficiencies (Camk2a-Cre; GLPfl/fl; G9afl/fl) had no additional impact on gene expression changes as compared to mice with a single methyltransferase deficiency (data not shown). The vast majority of the “signature” genes encode for proteins that are normally not expressed in adult neurons (Fig. 2C, D). A significant fraction of these genes is involved in development and function of the skeletal, muscular, cardiovascular, haematological, and immune systems (Fig. 2C, D). The degree of GLP and G9a impact on repression of non-neuronal genes is highlighted by neuronal expression of Afamin (Afm) or α-fetoprotein (Afp) in the striatum and cortex of GPL or G9a deficient mice (Supplementary Tables S1, S4, S5, S8, S9). Afamin and α-fetoprotein belong to the family of serum albumins and are highly expressed in the embryonic liver, but are transcriptionally silenced in mature neurons (Lichenstein et al., 1994; Nguyen et al., 2005). In addition to suppression of genes of non-neuronal lineage, GLP and G9a appear to control development stage-dependent neuronal gene expression. In particular, deficiency in GLP or G9a in adult neurons reactivates expression of the transcriptional regulator Dach2 (Supplementary Table S1-S11), which is expressed during embryonic brain, eye and limb development (Davis et al., 2001). Up-regulation of Dach2 is particularly interesting in view of the human DACH2 gene localization on the X chromosome in a region that is affected in a number of human mental retardation syndromes (Davis et al., 2001).
The neuron-specificity of the observed gene expression changes was validated by using the bacTRAP approach (Doyle et al., 2008; Heiman et al., 2008) that enables analysis of cell-type specific mRNA expression in tissues, which are composed of multiple different cell types. Expression of an eGFP tagged ribosomal protein selectively in Drd1 or Drd2 MSNs (Drd1-Cre; Drd1-bacTRAP; G9afl/fl or Drd2-Cre; Drd2-bacTRAP; G9afl/fl) allows Drd1 or Drd2 cell-type-specific polyribosome associated mRNA purification. Comparative analysis of polyribosome associated mRNA of G9a deficient and wild-type Drd1 or Drd2 MSNs showed that G9a deficiency in Drd1 or Drd2 MSNs leads to neuron-intrinsic expression of most of the previously identified non-neuronal/neuron-progenitor genes including Afm, Afp, Anxa10, Bank1, Ces7, Corin, Dach2, Defb1, Expi, Gbe1, Gkn1, Lama2, Lamb3, Musk, Myh1, Nxf3, Olfm4, Serpinb1b, Serpinb5, Ttn, and Tnnt2 (Supplementary Tables S1, S10, S11). This result indicates that the de-repression of non-neuronal genes in GLP or G9a deficient postnatal neurons is likely to be neuron-intrinsic.
Deficiency in GLP or G9a and ensuing changes in gene expression had no apparent impact on the structural organization of the affected brain areas (Fig. 3A). Moreover, the specific neuronal cell architecture in the striatum and the hippocampus was not altered by forebrain-specific postnatal GLP or G9a deficiency (Fig. 3B-E). Similarly, the lack of GLP/G9a had no major impact on regulation of neuronal morphology and basal electrophysiological features of Drd2-expressing MSNs in the striatum of Drd2-Cre; G9afl/fl mice (Fig. 4). Preservation of neuronal cell survival, morphology and cell type specific electrophysiological features in GLP or G9a deficient neurons rules out an essential role of GLP/G9a and H3K9me2 in regulation of vital cellular processes. Similar conclusions have been achieved by studies that evaluated the impact of G9a deficiency on fibroblasts, peripheral lymphocytes and cardiomyocytes (Sampath et al., 2007; Shirato et al., 2009; Thomas et al., 2008).
In humans, genetic alterations of the GLP gene are associated with a severe mental retardation syndrome that is further characterized by craniofacial abnormalities, hypotonia, obesity and a gradual decline in goal directed cognition and behavior including a loss of interest in the environment, reduced motor activity and emotional responsiveness (Cormier-Daire et al., 2003; Kleefstra et al., 2006; Kleefstra et al., 2009; Kramer and van Bokhoven, 2009; Verhoeven et al., 2009). Mental retardation in humans is defined by impaired cognitive function and deficits in adaptive behaviors in response to the environment (APA, 2000). Similarly, postnatal forebrain-specific ablation of GLP or G9a (Camk2a-Cre; GLPfl/fl or Camk2a-Cre; G9afl/fl) in the mouse brain leads to complex behavioral abnormalities. Starting around 6-8 weeks of age Camk2a-Cre; GLPfl/fl and Camk2a-Cre; G9afl/fl mice show a diminished exploratory behavior in response to a new environment. When challenged to explore a novel space in the open field, the GLP and G9a mutant mice travelled significantly less than control mice (Fig. 5A, B) and failed to explore the space as judged by the significantly reduced vertical episodes and lower center/total distance ratio (Fig. 5C, D). Notably, a similar level of reduced locomotor activity and exploration can be observed in mice with a haploinsufficiency in GLP (GLP+/−) (Fig. 5E-F), a genetic condition similar to the one in patients with altered function of a single GLP allele in all cells. Combined postnatal forebrain-specific deficiency of GLP and G9a (Camk2a-Cre; GLPfl/fl; G9afl/fl) does not potentiate the abnormal behavioral symptoms observed in mice with single enzyme deficiency (Fig. 5C, D), confirming the essential role of a functional GLP/G9a complex in regulation of motor activity and explorative behavior. Importantly, none of the observed behavioral aberrations is found in mice with a Camk2a-Cre mediated ablation of the histone lysine methyltransferase Ezh2 (CamK2a-Cre; Ezh2fl/fl mice), which is essential for H3K27 tri-methylation (Cao et al., 2002; Su et al., 2003) (Fig. 5C, D) and has been shown to play an important role in lineage-specific gene expression and differentiation of various types of cells including neurons and astrocytes (Bracken et al., 2006; Cao and Zhang, 2004; Hirabayashi et al., 2009).
Abnormal exploratory behavior in the absence of GLP/G9a could conceivably reflect a defect in motor function or balance. However, the Camk2a-Cre; GLPfl/fl mice did not display any reduction in motor activity or balance during the standard accelerated rotarod analysis. In fact, the mutant mice seemed to perform even slightly better than their littermate controls (Fig. 6A). Likewise, the elevated plus maze analysis, a test used to diagnose anxiety-like behavior in rodents, did not reveal any anxiety-like phenotype that could explain the observed reduction in exploratory behavior (Fig. 6B). On the contrary, the Camk2a-Cre; GLPfl/fl and the Camk2a-Cre; G9afl/fl mice showed no evidence that they were aware of the potential danger of open, unprotected areas. While wild-type mice avoided the bright and open arms of the elevated plus maze, the mutant mice showed a significant increase in the percentage of entries into the open arms and time spent in the open arms as compared to their littermate controls (Fig. 6B). GLP/G9a deficient mice also showed a significant decrease in total arm entries (Fig. 6B), which is consistent with the previously described reduction in explorative behavior in the open field (Fig. 4).
The reduction in locomotion and exploration in postnatal, neuron-specific GLP/G9a deficient mice could as well represent a general loss of curiosity and interest in the environment, implying potential deficits in the motivation or reward mechanism. To address the ability of these mice to seek/respond to rewards independently from their impaired locomotor activity and exploration, GLP/G9a deficient mice were tested for their preference for sucrose, a sweet-tasting natural reward. Wild type mice, when given a choice to drink regular water versus 1% sucrose, clearly prefer sucrose. Mice with a postnatal, neuron-specific deficiency in GLP or G9a almost completely lacked the preference for sucrose (Fig. 6C). Decreased sucrose preference may indicate an underlying dysfunction in the motivation/reward mechanism in these mice. These data are further supported by our recent findings describing an essential role of G9a in the regulation of addictive behavior in mice (Maze at al., in press, Science).
Contextual and cued fear conditioning is a behavioral test that evaluates the ability of mice to learn and remember associations between environmental cues and an aversive experience. The mice learn to associate a neutral context (test chamber) and a neutral stimulus (audio cue) with an aversive stimulus (mild foot shock). Learned association of the previous neutral cues with the aversive stimulus leads to the exhibition of a freezing response, which is characterized by an immobile and tense posture. Analysis of the Camk2a-Cre; GLPfl/fl mice using the standard fear conditioning paradigm demonstrated that while initial pain sensitivity and freezing in response to mild foot shock were not affected by GLP deficiency, the mutant mice did not remember their negative experience (Fig. 6D). The freezing response to the contextual conditioned fear, as well as to the audio cued conditioned fear was significantly impaired in the GLP mutant mice as compared to their littermate controls (Fig. 6D). Thus, the deficiency of GLP in postnatal neurons causes severe defects in learning and memory.
Many of the mental retardation syndromes in humans and mice are frequently associated with the development of obesity (Cormier-Daire et al., 2003; Fyffe et al., 2008; Goldstone and Beales, 2008). The same is true for the Camk2a-Cre; GLPfl/fl and the Camk2a-Cre; G9afl/fl mice, which became obese and almost doubled their body weight as compared to their controls by 5-6 months of age (Supplementary Fig. S3). Importantly, the described increase in body weight appears several weeks after the manifestation of reduced locomotion and exploration in these mice and therefore could be a consequence of, but not a reason for, the behavioral abnormalities described above (Supplementary Fig. S3).
Mice lacking G9a specifically in Drd1- or Drd2-expressing MSNs only (Drd1-Cre; G9afl/fl, Drd2-Cre; G9afl/fl), did not display the spontaneous locomotor or exploratory abnormalities observed in the Camk2a-Cre expressing mice (Fig. 7A-F). However, challenge with a Drd1 or Drd2 cell-specific stimulus revealed an important role of GLP/G9a in regulation of cell-type-specific behavioral responses to environmental signals. Locomotor activity is regulated by the Drd1 and Drd2 MSNs in opposite directions; either increased Drd1 or decreased Drd2 neuronal activity leads to an increase of locomotion in mice. Activation of Drd1-expressing MSNs via the dopamine 1 receptor leads to an increase of locomotor activity and exploration in wild-type mice. Loss of G9a in Drd1 MSNs led to a reduction of Drd1-agonist-induced locomotor activity and explorative behavior (Fig. 7A, C, E). On the other hand, decreased activity of Drd2 neurons is achieved either by inhibiting the adenosine-2- alpha (A2a) receptor or by activating the Drd2 receptor. Injection of caffeine, an A2a-receptor antagonist, which specifically reduces the activity of Drd2 MSNs, induced an increase in locomotor and exploratory activity in wild-type mice. Drd2 cell-specific deletion of G9a led to a hyper-responsiveness to caffeine, which was marked by a significantly higher increase in locomotor and exploratory response upon Drd2 MSN inhibition in vivo (Fig. 7B, D, F). In summary, deficiency of GLP/G9a in either of the two distinct neuronal populations in the striatum is associated with an alteration of the responsiveness of these cells to specific stimuli/inhibitors in vivo and reveals a significant reduction of cell-type-specific activity in both neuronal populations. These data demonstrate neuron/brain region specific functions of GLP/G9a in regulation of specific behavioral responses.
We have identified GLP/G9a as key regulators of cognition and adaptive behavior in adult mice. The mechanism of GLP/G9a involvement in brain function is likely to reflect the ability of these enzymes to maintain neuron-specific transcriptional homeostasis and to protect adult neurons from expression of numerous non-neuronal and neuronal progenitor genes. It is conceivable that ectopic expression of regulatory proteins that are involved in calcium and cAMP signaling (Bank1, Annexin-10, ArhGAP15, CaBP5, MuSK, Plce1), and cytoskeletal function (Myosin-1, Myosin-7, Myomesin-2, Titin, Tnnt2) may interfere significantly with the function of the tightly regulated intracellular signaling networks. Additionally, proteins of the laminin family like Lama2, Lama3 and Lamb3 that become upregulated in the absence of GLP/G9a are known to control cell-to-cell communication (Scheele et al., 2007) and may, therefore, interfere with neuronal connectivity in the brain. Finally, several of the upregulated “signature” proteins including Alpha-fetoprotein, Afamin, Carboxylesterase 7, Beta-defensin 1, Gastrokine-1, Olfactomedin-4, Serpin B1b, and Serpin B5 have been described as potent secreted regulators of various cell functions (Bailey et al., 2006; Benarafa et al., 2002; Holmes et al., 2008; Jensen et al., 2009; Niyonsaba et al., 2005; Oien et al., 2004). Production of these proteins by neurons could induce secondary changes in neighboring neurons or non-neuronal cells. Overall, it is likely that multiple gene defects rather than a single gene alteration are responsible for the complex behavioral changes in the absence of GLP/G9a.
Our data show that deficiency in GLP/G9a and ensuing changes in gene expression have a differential impact on specialized brain areas. Inactivation of GLP/G9a in large numbers of functionally diverse neurons in the postnatal forebrain leads to a mental retardation-like phenotype in adult mice. Contrary to forebrain-specific GLP/G9a deficiency in the Camk2a-Cre mice, ablation of these enzymes in highly specialized Drd1 or Drd2 neurons in the striatum causes behavioral changes only after pharmacological perturbation of Drd1 or Drd2 cell function. It is conceivable that the observed differences in the severity of the basal behavioral phenotypes are simply proportional to the size and functional diversity of the affected neuronal populations. Accordingly, the potential impact of GLP/G9a deficiency in specialized brain areas or neuronal subpopulations, such as Drd1- or Drd2-expressing MSNs, could be revealed only after perturbation of these neurons.
Comparison of the complex behavioral abnormalities in mice with conditional postnatal GLP/G9a deficiency revealed a strong resemblance to the human mental retardation syndrome associated with the subtelomeric deletion of the human chromosome 9 (9q34), which includes the histone lysine methyltransferase GLP/EHMT1 gene. In addition to major defects in cognition and learning, the 9q34 syndrome is further characterized by obesity, childhood hypotonia and a gradual, age-dependent development of severe apathetic behavior, reduced motor activity and loss of goal directed activities (Cormier-Daire et al., 2003; Kleefstra et al., 2006; Kleefstra et al., 2009; Kramer and van Bokhoven, 2009; Verhoeven et al., 2009). The potential causal role of the GLP/EHMT1 gene alterations (OMIM 607001) in the human 9q34 mental retardation syndrome has been underscored by the identification of various intragenic GLP/EHMT1 mutations in patients with a mental retardation syndrome clinically indistinguishable from 9q34 deletion syndrome (Kleefstra et al., 2009; Kramer and van Bokhoven, 2009). Five of the identified mutations in the GLP/EHMT1 gene predict a premature termination codon resulting in nonsense mediated mRNA decay, whereas one of the mutations leads to an amino acid substitution that is predicted to affect the conformation and hence the activity of the highly conserved, histone methyltransferase encoding SET-domain of the GLP/EHMT1 protein (Kleefstra et al., 2009).
The presence of key symptoms of the human 9q34 mental retardation syndrome in mice with a postnatal, neuron-specific GLP/G9a deficiency supports a causal role of GLP/G9a in impaired neuron-specific function in the human disease. In humans, deficiency in GLP is likely to affect both the development and the function of the adult neurons. Our findings suggest that deficiency in GLP/G9a in the postnatal mouse brain is sufficient to cause complex behavioral changes. The latter finding indicates a possible role of GLP/G9a regulation of adult brain function, including cognition, in normal individuals. It has been demonstrated recently that the protein level and histone methyltransferase activity of G9a can be regulated by oxygen concentration (Chen et al., 2006). Therefore, it is possible that changes in local oxygen concentrations in the brain may affect changes in brain function due to alteration of GLP/G9a mediated gene regulation. Thus, it seems plausible that genetically predetermined or environmentally induced changes of enzymes controlling H3K9me2 methylation may determine individual differences in cognition and environmental adaptation in mice and humans.
To create the GLP targeting vector, a region of the GLP genomic locus containing the long and short arms of homology was recombinogenically subcloned from a BAC clone (RPCI24-156K12; BL/6; CHORI, Oakland CA, USA) into pBlueScriptIIKS+ essentially as described (Lee et al., 2001; Sampath et al., 2007; Yu et al., 2000), using primers which inserted a 5′ AscI site and a 3′ FseI site. The properly recombined plasmid was digested with NheI, and an annealed double-stranded oligonucleotide containing the upstream loxP site and a BsrGI site was inserted. The vector was then digested with BamHI, and a NsiI fragment from pZeroloxP-FRT-neoR-FRT(-) containing the downstream loxP site and FRT-flanked neo gene was inserted and screened for proper orientation. The resulting plasmid was digested with AscI and FseI to release the targeted locus, which was inserted into XhoI-digested pDTA-TK to produce the final targeting construct. Induced deletion of exon 23 is predicted to lead to out-of-frame splicing and Nonsense-Mediated Decay of the transcript.
E14.1 (129/Ola) embryonic stem cells were transfected, selected, and used to produce chimeric mice as described (Torres and Kuhn, 1997). Chimeras were crossed to C57/BL6 and germline transmission was assessed by coat color, PCR and Southern Blot analysis (GLP-probe-B: CCTGTTAAACATGGCTGCTTG). Deletion of the FRT-flanked neo gene was accomplished by crossing of mice carrying the GLPTarg allele in the germline to FLPe-transgenic mice (Rodriguez et al., 2000). Routine genotyping was performed by tail biopsy and PCR using following primers:
GLPfl/fl, G9afl/fl (Sampath et al., 2007) and Ezh2fl/fl (Su et al., 2003) were bred to Camk2a-Cre (Casanova et al., 2001) mice to generate Camk2a-Cre; GLPfl/fl and Camk2a-Cre; G9afl/fl and Camk2a-Cre; GLP fl/fl; G9a fl/fl and Camk2a-Cre; Ezh2 fl/fl mice, respectively. GLP+/− mice were generated by crossing GLPfl/fl with mice that express Cre-recombinase under the control of the ubiquitously expressed CMV promotor (CMV-Cre mice, Jackson Laboratory). Drd2-bacTRAP (Heiman et al., 2008) and Drd2-eGFP (Heintz, 2004) mice were bred to Drd2-Cre (ER44 line from Gensat (Gong et al., 2007)) and G9afl/fl mice to generate Drd2-Cre; Drd2-bacTRAP; G9afl/fl and Drd2-bacTRAP; G9afl/fl control mice and Drd2-Cre; Drd2-eGFP; G9afl/fl and Drd2-eGFP; G9afl/fl control mice. Drd1-bacTRAP (Heiman et al., 2008) mice were bred to Drd1-Cre (EY262 line from Gensat (Gong et al., 2007)) and G9afl/fl mice to generate Drd1-Cre; Drd1-bacTRAP; G9afl/fl and Drd1-bacTRAP; G9afl/fl control mice and genotyping was performed as previously described. All mice were backcrossed to the C57Bl/6 background for ≥4 generations. Mice were housed under standard laboratory conditions at the Rockefeller University Comparative Bioscience Center (CBC). Protocols were approved by the Rockefeller University Institutional Animal Care and Use Committee.
Eight- to twelve-week old sex- and age-matched Camk2a-Cre; GLPfl/fl, Camk2a-Cre; G9afl/fl, Drd2-Cre; Drd2-eGFP; G9afl/fl, and their respective littermate controls were anesthetized with pentobarbitol and transcardially perfused with 10ml of phosphate buffered saline (PBS) followed by 40ml of 4% paraformaldehyde (PFA) in PBS. Brains were processed as previously described (Schaefer et al., 2007). For GLP immunohistochemistry, sections underwent antigen retrieval by heating in citrate solution (pH=6.0), were incubated with anti-GLP antibody (1:1000, R&D Systems, Minneapolis, MN) and visualized by the avidin–biotin–peroxidase complex method (Vector Laboratories, Burlingame, CA) as recommended by the manufacture. DAB enhanced liquid substrate system (Sigma-Aldrich) was used to detect peroxidase activity and sections were mounted using Crystal Mount solution (Biomedia, Foster City, CA). GFP (1:5000, ab6556), H3K9me2 (1:500, ab1220) and H3K9me3 (1:1000, ab8898, all Abcam, Cambridge, MA) were visualized using indirect immunofluorescense analysis (AlexaFluor 546/488 labeled goat anti-mouse/anti-rabbit IgGs (H+L) dilution 1:500, Invitrogen Corporation, Carlsbad, CA). Draq5 (Biostatus Limited, Leicestershire, UK) was used to counterstain the nucleus and sections were mounted using Prolong Gold antifade (Invitrogen Corporation, Carlsbad, CA).
Sagittal brain sections of twelve week-old Camk2a-Cre; GLPfl/fl, Camk2a-Cre; G9afl/fl and their respective littermate controls were visualized using standard Nissl-stain (FD NeuroTechnologies, Inc., Ellicott City, MD). Coronal sections of 12-16 week-old Camk2a-Cre; GLPfl/fl, Camk2a-Cre; G9afl/fl and their respective littermate controls (n=3/genotype) were stained using the FD Rapid GolgiStain™ Kit (FD NeuroTechnologies, Inc., Ellicott City, MD) as recommended. All sections were visualized on a Zeiss LSM510 confocal miroscope (Carl Zeiss, Thornwood, NY).
Three Drd2-Cre; Drd2-eGFP; G9afl/fl and Drd2-eGFP; G9afl/fl control mice between postnatal days 35-40 were anesthetized with ketamine/xylazine and perfused transcardially with ice cold artificial CSF (aCSF) containing in mM: 125 NaCl, 2.5 KCl, 1.25 NaH2PO4, 2.0 CaCl2, 1.0 MgCl2, 25 NaHCO3, and 12.5 glucose, bubbled continuously with carbogen (95% O2 and 5% CO2). The brains were rapidly removed, glued to the stage of a VT1000S slicer (Leica, Nussloch, Germany), and immersed in ice-cold aCSF. Striatal slices were cut at a thickness of 240 μm and transferred to a holding chamber, where they were submerged in aCSF, incubated at 35°C for 30 minutes, and returned to room temperature before recording. Recordings were made at room temperature (20-22°C) with patch electrodes fabricated from filamented, thick-wall borosilicate-glass (Sutter Instruments, Novato, CA) pulled on a Flaming-Brown puller (P-97; Sutter). Pipette resistance was typically ~3-5 MΩ when filled with internal solution consisting of (in mM): 135 KMeSO4, 5 KCl, 10 Na-phosphocreatine, 5.0 EGTA, 0.5 CaCl2, 2.0 Mg-ATP, 0.5 Na3-GTP, 5 HEPES, 0.2% w/v biocytin, pH 7.25-7.30, 300 mOsm. The liquid junction potential in recordings was ~7 mV and not corrected for. MSNs within the dorsal striatum were identified by their somatic morphological characteristics under IR-DIC optics. Somatic eGFP expression was verified routinely in cell-attach mode using epifluorescence microscopy to confirm cell identity before breaking into whole-cell mode. Somatic whole-cell patch-clamp recordings were obtained with a MultiClamp 700B amplifier (Molecular Devices, Union City, CA) interfaced to a Pentium-based PC running pClamp9 (Molecular Devices). For current-clamp recordings, the amplifier bridge circuit was adjusted to compensate for electrode resistance and subsequently monitored. For anatomical reconstruction, slices in which eGFP positive Drd2 cells were recorded in whole-cell mode for at least 30 minutes were flat-mounted on nitrocellulose filters (Millipore; Billerica, MA), fixed in 4% paraformaldehye for 72 hours, and reacted with 2 μg/ml streptavidin-AlexaFluor 594 (Invitrogen; Carlsbad, CA) in 2% Triton-X, 1% NGS prepared in PBS overnight. Sections were then washed, coverslipped, and imaged as detailed previously (Gertler et al., 2008).
Total RNA from cortex, hippocampus, hypothalamus and striatum of twelve week-old Camk2a-Cre; GLPfl/fl, Camk2a-Cre; G9afl/fl, and Camk2a-Cre; GLPfl/fl; G9afl/fl mice and age and sex matched littermate controls (n=4/genotype) was extracted using the Trizol/Chloroform technique as recommended by the manufacturer (Invitrogen Corporation, Carlsbad, CA). After extraction, RNA was precipitated overnight at −80°C in isopropanol with sodium acetate and Glycoblue (Ambion, Austin, TX), washed twice with 70% ethanol, re-suspended in water, and further purified using an Rneasy Micro Kit (Qiagen, Valencia, CA) with in-column DNase digestion. Purified samples were analyzed using a Nanodrop 1000 spectrophotometer (Thermo Scientific, Wilmington, DE) and a Bioanalyzer (Agilent Technologies, Santa Clara, CA) in order to assess mRNA quantity and quality, as reflected by rRNA levels and integrity.
Polyribosome associated mRNAs from age and sex matched Drd2-Cre; Drd2-bacTRAP; G9afl/fl and Drd2-bacTRAP; G9afl/fl, or Drd1-Cre; Drd1-bacTRAP; G9afl/fl and Drd1-bacTRAP; G9afl/fl mice (n=3×5 mice/genotype) were obtained as previously described using a mix of two monoclonal anti-GFP antibodies (19C8, 19F7) for the immunoprecipitation reaction (Doyle et al., 2008; Heiman et al., 2008).
Purified mRNA or total RNA was amplified and processed for using the Affymetrix two cycle cDNA Synthesis kit (Affymetrix, Santa Clara, CA) as previously described (Doyle et al., 2008; Heiman et al., 2008). Affymetrix Mouse Genome 430 2.0 arrays were used in all experiments.
Information regarding the array design and features can be found at www.affymetrix.com. Mouse Genome 430 2.0 arrays were scanned using the GeneChip Scanner 3000 (Affymetrix, Santa Clara, CA) and globally scaled to 150 using the Affymetrix GeneChip Operating Software (GCOS v1.4). Three/four biological replicates were performed for each experiment. GeneChip CEL files were imported together into Genespring GX 7.3.1 (Agilent Technologies, Santa Clara, CA), processed with the GC-RMA algorithm, and expression values on each chip were normalized to that chip’s 50th percentile. Statistical analysis was carried out to determine which genes are differentially expressed in the GLP and G9a deficient cortex, striatum, hippocampus and hypothalamus as compared to their respective controls, and G9a deficient Drd1 and Drd2 MSNs as compared to their respective controls. The Minimum Information about a microarray experiment (MIAMI) compliant summary of the gene expression analysis can be found in the Supplementary Material and Method section.
All behavioral tests have been performed on 6-16 week old Camk2a-Cre; GLPfl/fl, Camk2a-Cre; GLP+/fl, GLPfl/−, Camk2a-Cre; G9afl/fl, Camk2a-Cre; GLPfl/fl; G9afl/fl, Camk2a-Cre; Ezh2fl/fl, Drd2-Cre; G9afl/fl, Drd1-Cre; G9afl/fl mice and their respective age and sex matched littermate controls. Mice were housed with a 12 hr light/dark cycle (lights on at 7 A.M. and off at 7 P.M.) with access to food and water ad libitum. All behavioral tests were performed between 9 A.M. and 5 P.M. Experiments were conducted by an experimenter blind to the genotypes of the mice. GraphPad Prism version 4.03 for Windows (GraphPad Software, San Diego, CA) was used for statistical analysis of the data. Statistical analysis was performed using Student’s t-test, one-way, two-way, or two-way repeated measures ANOVA with Bonferoni posttest.
Locomotion and exploratory behavior was measured using the open field analysis in a new environment (clear plexiglas 40 × 40 × 30 cm open-field arena). Activity in the open-field was quantified by a computer-operated Photobeam activity system (AccuScan Instruments, Columbus, OH). Mice were recorded for the total distance moved (cm), number of vertical episodes (rearing), and distance moved in the center of the arena (cm). The distance moved in the center (cm) was divided by the total distance moved (cm) to obtain center/total distance ratio values. Data were collected at 5-10-min intervals over 20-60 min test sessions. Saline, Drd1-agonist SKF 81297 (5mg/kg in saline, Tocris, Ellisville, MO), and caffeine (10mg/kg in saline, Sigma, St. Louis, MO) have been used for intraperitoneal injections.
The motor function and balance of mice was analyzed using the standard accelerated rotarod test (4-40 rpm/5 min, Med Associates, St. Albans, VT). The time taken for the mice to fall from the rod was measured in seconds. If a mouse, clinging on to the rod, completed two full passive rotations, the mouse was removed from the trial and the time was recorded as fallen from the rod. If a mouse stayed on the rod until the end of the 5-minute trial, a time of 300 seconds was recorded. After one training trial, mice were subjected to three consecutive trials/day with 5 min inter-trial intervals for three consecutive days and measurements were taken from each trial.
The elevated plus maze test was used to determine the unconditioned response to a potentially dangerous environment. Anxiety-related behavior was measured by the degree to which the rodent avoided the open arms of the maze. The mice were placed at the junction of the elevated 4-arm maze in which 2 arms are open and 2 are enclosed. The number of times the animal entered each of the arms and the time spent in each arm was recorded for 5 min by the EthoVision video-tracking system (Noldus Information Technology Inc., Leesburg, VA). Total arm entries, percentage of open arm entries, and percentage of time spend in the open arms were calculated.
Sucrose preference of Camk2a-Cre; GLPfl/fl, Camk2a-Cre; G9afl/fl and their sex and age matched littermate controls (n=5/genotype) was measured as an index for motivation and reward. All animals were single housed 3-4 days prior to the beginning of the sucrose preference testing. Two 50ml bottles fitted with two-balled sipper tubes (Ancare, Bellmore, NY) were positioned on the food rack. The experiment started with 2 days of water in both bottles, followed by 2 days with 1% sucrose in both bottles, and 4 days of a water-sucrose choice. The fluid levels in both bottles were measured every evening at 6pm, and the bottle position was switched daily to avoid side biases. The total amount of sucrose was divided by the total amount of water consumed during the 4 days of water-sucrose choice to calculate the sucrose/water preference ratio.
Memory and learning abilities of Camk2a-Cre; GLPfl/fl and their sex and age matched littermate controls (n=13/genotype) have been analyzed using a standard fear conditioning paradigm (Med Associates, St. Albans, VT). The test chamber (neutral context) was made of clear Plexiglas. The bottom of the test chamber was a grid floor used to deliver a mild electric foot shock. The test chamber was placed inside a sound-attenuated chamber. Mice were observed through a window in the front of the sound-attenuated chamber. One mouse was placed in the test chamber (house lights on) and allowed to explore freely for 2 min. A white noise (80 dB), which served as the conditioned stimulus, was then presented for 30 sec, followed by a mild (2 sec, 0.5 mA) foot shock, which served as the unconditioned stimulus. Two minutes later the same sequence of auditory cue-shock pairing was repeated. The mouse was removed from the chamber 30 sec later and returned to its home cage. Freezing behavior was recorded every 10 sec during the time spend in the test chamber. Responses (run, jump, and vocalize) to the foot shock were also recorded. Twenty-four hours later, the mouse was placed back into the test chamber for 5 min, and the presence of freezing behavior was recorded (context test). Two hours later, the mouse was tested for its freezing respond to the auditory cue. Environmental and contextual cues were changed for the auditory cue test: a black Plexiglas triangular insert was placed in the chamber to alter its shape, spatial cues and lightening; the wire grid floor was covered with white Plexiglas; and orange extract was placed in the chamber to alter the smell. The auditory cue test was divided into two phases. During the first phase, freezing behavior was recorded for 3 min in the absence of the auditory cue. In the second phase, the auditory cue was turned on, and freezing was recorded for another 3 min. The number of freezing intervals for each test was converted to a percentage of freezing value.
Body weight of Camk2a-Cre; GLP fl/fl, Camk2a-Cre; G9afl/fl, and Pomc1-Cre; G9afl/fl mice and their age and sex matched littermate controls was recorded monthly starting at 4 weeks of age (n=10-15).
We would like to thank Myriam Heiman and Nathaniel Heintz for providing the Drd1-bacTRAP and Drd2-bacTRAP mice, Nathaniel Heintz and Gensat for providing the Drd2-eGFP and Drd1- and Drd2-Cre mice, Guenther Schuetz for providing the Camk2a-Cre mice, Donal O’Carroll for providing vectors for conditional gene targeting of GLP, Ayse Begum Tekinay for her help with the behavioral analysis, Selina Riddick and Sharon Kuca for animal work, Fekrije Selimi and Lars Brichta for critical reading of the manuscript, and Elisabeth Griggs and Debra Poulter for her help in the preparation of the manuscript. This work was supported in part by NIH grants AG09464 (P.G.), AI068058 NS13742 (A.T.), DA025962 (A.S.), GM07739 (S.C.S.) and MH074866 (D.J.S.). A. S. was a recipient of a DFG postdoctoral fellowship (SCHA 1482/1-1). T.G. was recipient of NRSA award MH082522. S.C.S. was supported by the Rudin Foundation.
Author Information The authors declare that they have no conflicting financial interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Supplementary Material is linked to an online version of the paper.
Supplementary Figures S1-3
Supplementary Tables S1-11
Supplementary Materials and Methods