Log in / Register
Home arrow Health arrow DNA Modifications in the Brain. Neuroepigenetic Regulation of Gene Expression


The three TET proteins, TET1, TET2, and TET3, are members of the Fe(II)/a-KG— dependent dioxygenase family of enzymes. It has been proposed that the enzymes use molecular oxygen to catalyze oxidative decarboxylation of a-KG, creating a highly

reactive intermediate that converts 5mC to 5hmC (Fig. 4.2). This proposed mechanism is based off of other Fe(II)/a-KG-dependent dioxygenase family proteins, since a structure of mammalian TET enzymes has not been solved. Each of the TETs contain a conserved catalytic domain (double-stranded 3-helix, DSBH, fold) that contains the metal-binding residues required for the oxidation reaction (Fig. 4.3) (Kohli & Zhang, 2013). Additionally, a cysteine-rich (Cys-rich) domain is found in all TET proteins upstream of the catalytic domain and is thought to be required for activity (Iyer, Tahil- iani, Rao, & Aravind, 2009; Tahiliani et al., 2009). TET and TET3 contain a CXXC domain near the N-terminal end of the protein, which is known to bind to CpG sites (Kohli & Zhang, 2013). Although each of the human TET enzymes only share ~18-24% sequence identity (UniProt Consortium, 2015), the catalytic and Cys-rich domains are highly conserved between the three enzymes (Fig. 4.4). It is hypothesized that the remaining nonconserved portions of the protein may serve as regulatory domains and convey different functionality between the three TETs.

Tet1, Tet2, and Tet3 are all expressed in the brain, with Tet3 having the highest expression, followed by Tet2; Tet1 has much lower expression levels than the other two family

methylcytosine to 5-hydroxymethylcytosine conversion by Ten-eleven translocation enzymes

Figure 4.2 5-methylcytosine to 5-hydroxymethylcytosine conversion by Ten-eleven translocation enzymes. Ten-eleven translocation (Tet) enzymes oxidize the 5-methyl group of 5-methylcytosine (5mC). With cofactors a-ketoglutarate (a-KG) and molecular oxygen, Tet oxidizes the 5-methyl carbon, adding a hydroxyl group, thereby yielding 5-hydroxymethylcytosine (5hmC). Other by-products of the enzymatic reaction include CO2 and succinate.

Schematic of Ten-eleven translocation enzymes

Figure 4.3 Schematic of Ten-eleven translocation enzymes. Ten-eleven translocation (Tet)1, Tet2, and Tet3 all share a conserved catalytic double-stranded [3-helix fold (DSpH) domain at the C-terminal end of the protein. Additionally, a conserved cysteine-rich (Cys-rich) domain is found at the N-terminal portion of the catalytic domain. Tet1 and Tet3 contain an additional CpG-binding CXXC domain near the N-terminal end of the protein.

Sequence conservation of Ten-eleven translocation domains

Figure 4.4 Sequence conservation of Ten-eleven translocation domains. (A) The cysteine-rich (Cys- rich) domain is almost fully conserved between the Ten-eleven translocation (Tet) enzymes. (B) The catalytic double-stranded [3-helix fold (DS|3H) domain is also highly conserved between the three Tet enzymes, even though the proteins only share 18-24% sequence identity. The remaining nonconserved domains are hypothesized to serve as regulatory domains.

members (Szwagierczak et al., 2010). Knockout, loss of function, and overexpression studies have revealed diverse functions of these enzymes, and the importance of 5hmC, in neuronal function.


Currently, Tet1 is the most well-studied Tet family member in the brain, most likely because TET1 was the first enzyme discovered to convert 5mC to 5hmC (Tahiliani et al., 2009). Although Tet1 expression is markedly lower than Tet2 and Tet3 in the brain, various studies have demonstrated the importance of Tet1 in neuronal function.

Tet1 whole-body knockout (KO) mice are viable and fertile without apparent health deficits, albeit a smaller body weight and litter size than WT animals (Dawlaty et al., 2011; Rudenko et al., 2013; Zhang et al., 2013). Additionally, there are no obvious morphological or developmental brain abnormalities (Rudenko et al., 2013; Zhang et al., 2013). In agreement with the importance ofTet1 in the generation of 5hmC, there is a small, but significant, reduction of 5hmC in the brains of Tet1 KO mice, but no change in 5mC. The fact that the change is small is likely due to presence of Tet2 and Tet3, the other members of the Tet family that are endogenously expressed at much higher levels than Tet1 in the brain. There are also no apparent deficits in synaptic connectivity as measured by Synapsin I, a marker of synaptic abundance (Rudenko et al., 2013). This compensation effect is supported by the abnormalities observed in Tet1 and Tet2 double knockout (DKO) mice. The majority of DKO die perinatally, although a small percentage survive without gross abnormalities. Compared to WT mice, DKO adult mice (2.5 months) have reduced 5hmC levels (34%) and increased 5mC levels (~5%) in the cerebrum and cerebellum. Although these are appreciable changes in methylation, a large portion of 5hmC remains intact, suggesting that Tet3 plays a critical role in its maintenance (Dawlaty et al., 2013).

Behaviorally, adult Tetl single KO mice (4 months) do not show deficits in locomotion, anxiety, fear memory acquisition, or depression-related behaviors. Multiple groups have observed memory deficits; however, there is not a consensus as to the specific type of memory deficit. According to one group, Tetl KO mice have impairments in shortterm memory and spatial learning, but normal long-term memory, as assessed by Morris water maze (MWM) (Zhang et al., 2013). Another group reported normal short-term memory and spatial learning, but impaired spatial memory extinction in the MWM and classical Pavlovian fear conditioning (Rudenko et al., 2013).When Tetl is overexpressed in the CA1 region of the hippocampus, long-term memory was affected (fear conditioning), but not locomotion, anxiety, or short-term memory. This deficit in long-term memory formation was observed for both catalytically active and inactive forms of TET1, suggesting that TET1’s role in memory formation is independent of its catalytic activity. Tet1 expression, but not that of Tet2, Tet3, or other proteins involved in the demethylation pathway, is significantly downregulated in the dorsal CA1 of mice after fear learning (Kaas et al., 2013). These findings support that Tet1 contributes to basal neuronal 5hmC levels that are potentially important for neuronal function. The behavioral effects of Tet1 in the brain still warrant further investigation considering the confounding results.

At the cellular and molecular level, evidence suggests that Tet1 is important in neurogenesis and hippocampal function. When Tet1 KO mice were bred with Nestin-GFP transgenic mice, the number of GFP-positive cells in the subgranular zone in adult mice was dramatically reduced by 45% compared to WT animals (Zhang et al., 2013). This is different from two other non-neurogenic brain regions examined, the cingulate cortex and hippocampus CA1 (Rudenko et al., 2013). The reduction in proliferation potential of NPCs is likely to underlie this deficit as evidenced by a reduction in neurospheres isolated from Tet1 KO mice, the decrease in bromodeoxyuridine (BrdU, which marks dividing cells)-positive neurons in Tet1-dentate gyrus (DG)-knockdown in adult mice, and the 35% decrease in BrdU-positive neurons in animals in which Tet1 is specifically deleted in neural progenitors at 2 months of age. Examination of the gene expression and methylation changes in Tet1 KO mice revealed that the decreased expression of a cohort of genes involved in neurogenesis was associated with an increase in 5mC at their promoters, suggesting that TET1 positively regulates adult neurogenesis through the oxidation of 5mC to 5hmC at these genes (Zhang et al., 2013).

Tet1 overexpression in the DG or CA1 region of the hippocampus of mice results in a dramatic increase in 5hmC and decrease in 5mC, providing evidence that TET1 in vivo oxidizes 5mC to 5hmC (Guo, Su, Zhong, Ming, & Song, 2011b; Kaas et al., 2013). The overexpression ofTet1 in the DG led to a significant decrease in methylation at promoter IX of Bdnf (brain-derived neurotrophic factor IX (Bdnf IX)) and the brain specific promoter of Fgfl (FgflB), and a concomitant increase in the expression of these two genes, supporting the role of Tet1 in the demethylation pathway, and subsequent gene activation (Guo et al., 2011b). Tet1 overexpression in area CA1 or DG of the hippocampus leads to the increased expression of various activity-dependent genes (Fos, Arc, Egrl, Homerl, and Nf4a2), as well as genes downstream of the Tet-mediated oxidation (Tdg, Apobecl, Smugl, and Mbd4) (Kaas et al., 2013). In the DG, the increased expression of these genes is dependent upon the catalytic domain TET1, as evidence by the fact that only the expression of human TET1 catalytic domain, but not expression of the catalyti- cally inactive version ofTET1; however, in the CA1 region, either the catalytic active or inactive TET1 leads to increase in expression of majority of these genes. This implies that TET1 acts via region-dependent mechanisms (Guo et al., 2011b; Kaas et al., 2013). Furthermore, Tet1 is required for neuronal activity-induced active DNA demethylation and gene expression since short hairpin-mediated knockdown of endogenous Tet1 in the DG abolished electroconvulsive stimulation (ECS)-induced demethylation of Bdnf IX and FgflB promoters. These in vivo findings are in agreement with in vitro work showing that Tet1 knockdown in hippocampal neurons leads to the hypermethylation of promoter IV of Bdnf and subsequent decreased expression from this promoter (Yu et al., 2015). Given that demethylation at these promoters is similarly abolished after ECS with knockdown ofApobec1, this suggests that TET1 and APOBEC1 work together through oxidative deamination to achieve active demethylation in the adult mouse brain (Guo et al., 2011b).

Loss of Tet1 also causes electrophysiological deficits in the hippocampus. Tet1 KO mice have normal basal synaptic transmission and intrinsic neuronal properties, as measured by paired-pulse facilitation and presynaptic excitability, respectively. However, long-term potentiation, assessed in the Schaffer collateral-CA1 pathway, is attenuated, and long-term depression (LTD) is amplified. These in vivo electrophysiological findings confirm what is found in vitro. Overexpression of the catalytically active form of TET1 prevents tetrodotoxin (TTX)-induced scaling-up, and knockdown of Tet1 leads to synaptic scaling-down that is unaltered by bicuculline treatment (Yu et al., 2015). Further analysis in vivo has demonstrated that alterations in the metabotropic glutamate receptor-dependent form of LTD is not affected, suggesting a deficit in N-methyl-D-aspartate receptor (NMDAR)-dependent LTD. Neuronal activity-regulated genes, including c-Fos, Egr2, Egr4, Arc, and Npas4, are affected in Tet1 KO mice. Analysis of the Npas4 promoter-exon 1 region confirmed a decrease of 5hmC and an increase in 5mC, which could explain the downregulation of this group of genes. After memory extinction in Tet1 KO mice (but not after fear memory acquisition), the Npas4 and c-Fos genes exhibit a decrease in 5hmC and an increase in 5mC, concomitant with a decrease in mRNA and protein expression levels in both brain regions assessed, the cortex and hippocampus. Since Tet1/Tet2/Tet3 expression does not increase during either fear memory extinction or acquisition, the activity of these proteins change, rather than absolute levels (Rudenko et al., 2013).

This body of work on Tet1 function in the brain suggests that Tet1, although expressed at much lower levels in the mammalian brain than the other Tet family members, plays an important role in maintaining 5hmC levels, and subsequent gene expression levels, at basal and activity-induced conditions.


Despite its high level of expression, Tet2 is presently the least well-studied Tet family member in the brain, but the limited studies conducted thus far have demonstrated the importance of Tet2 in brain function. There are no reported brain abnormalities or dysfunction in Tet2 KO mice (Ko et al., 2011; Li et al., 2011). However, when Tet2 is knocked down in hippocampal neurons in vitro, there is an increase in miniature excitatory postsynaptic current (mEPSC) amplitudes compared to controls (Yu et al., 2015). This implies that neuronal function may be impaired in the absence ofTet2.

Tet2 is also thought to play a role in the demethylation of developmentally dependent genomic loci. With the use of Tet2 KO mice, it was found that this member of the Tet family is responsible for the oxidation of large fraction (19.7%) of CpG genomic regions that gain hydroxymethylation status over development. In contrast, CpG regions with higher 5hmC in the adult than fetal stage are largely unaffected in Tet2 KO mice. Across development and aging (6 week, 10 week, and 22 mo) in Tet2 KO mice, there are greater than four-fold more hypermethylated CpG regions (14,000 CpG regions in total) than hypomethylated regions, suggesting that Tet2 plays a role in the demethyl- ation over development and aging (Lister et al., 2013). Tet2 may also play a role in neurogenesis since the knockdown ofTet2 and Tet3 via electroporation of shRNAs into the cortex lead to defects in the progression of differentiation from the subventricular zone (Hahn et al., 2013). These Tet2 findings suggest that Tet2 plays an important role in regulating developmentally dependent, differentially hydroxymethylated regions.


Various studies on Tet3 function in the brain have confirmed that this most highly expressed Tet family member in the brain is essential for regulating neuronal activity. When mice undergo extinction training, there is a significant increase in Tet3 mRNA in the cortex. Tet3 knockdown via lentiviral plasmids in the infralimbic prefrontal cortex results in normal fear memory acquisition, but impaired fear memory extinction. Furthermore, inhibiting NMDAR activity blocks the increase in Tet3 expression associated with fear memory extinction, suggesting that the rise in Tet3 occurs via an NMDAR- mediated pathway. Fear extinction causes in genome-wide changes in 5hmC at locations that contain CpA or CpT dinucleotide repeats, but not at CpGs. Additionally, there is a reduction in 5hmC at intronic and intergenic sites and an increase in 5hmC enrichment at distal promoters, 5' untranslated region (UTR), 3' UTR, exonic sequences, and DHS regions. Gene ontology analysis revealed that 16% genes enriched for 5hmC after extinction learning are involved in synaptic signaling. When one of these genes, gephyrin (Gphn), was evaluated, it was found that there was enrichment for 5hmC, co-occurring with decrease in 5mC, within one intron. Moreover, in response to extinction, there was an increase Tet3 occupancy at Gphn, as well as an increase in specificity protein 1 (Sp1), a transcription factor that activates gene expression by preventing the active loci from becoming methylated. Additionally, the observed reduction in transient H3K9me3 and increase in H3K27ac, p300, H3K4me1, and di-methyl arginine of histone H3 (H3R2me2), which support a euchromatic state, support the role ofTet3 in extinction- induced gene expression changes. All of these changes at the Gphn appear to be specifically regulated by Tet3 function, since the changes are blocked with the use of a Tet3 shRNA (Li et al., 2014).

Tet3 expression levels correlate with neuronal activity in vitro as well; an increase in synaptic transmission correlates with an increase in Tet3, but not Tetl or Tet2. When Tet3 is knocked down from hippocampal neurons in culture, mEPSC amplitudes are significantly larger than controls, and the reciprocal effect occurs when Tet3 is overexpressed. Notably, knockdown of either Tetl or Tet2 also increases mEPSC amplitudes, but not as drastically as Tet3 knockdown. Tet3 is also essential for the maintenance of homeostatic synaptic plasticity since knockdown of Tet3 leads to synaptic scaling-up that is unaltered by TTX or retinoic acid treatment; knockdown of Tet3 leads to synaptic scaling-down that occludes further alterations with bicuculline treatment; and Tet3 overexpression prevents TTX-induced synaptic scaling-up or bicuculline- induced scaling-down. Given that a similar effect on mEPSC amplitudes and synaptic scaling occurs when poly(ADP-ribose) polymerase or apurinic/apyrimidinic endonuclease, the two major components of the BER pathway, is inhibited this suggests that excitatory synaptic transmission is regulated by the oxidation of DNA via TET, followed by BER (Yu et al., 2015).

The molecular mechanism through which Tet3 elicits these effects is likely through the regulation of surface glutamate receptor 1 (GluR1). Knockdown of Tet3 leads to an increase in surface GluR1 receptors that is resistant to a further increase or decrease in surface GluR1 receptors. When gene expression changes were assessed in Tet3 knockdown neurons, gene ontology term enrichment revealed expression changes of genes involved in the synapse and synaptic transmission. Genes with differential expression due to TTX or bicuculline treatment in control neurons lost responsiveness in Tet3 knockdown neurons. In Tet3 knockdown neurons, promoter IV of the Bdnf is hypermethylated, and there is a decrease in expression from this promoter. The bicuculline-induced hypomethylation, as well as the TTX-induced hypermethylation, of Bdnf promoter IV is occluded in Tet3 knockdown neurons. Chromatin immunoprecipitation (ChIP)-PCR revealed that TET3 binds to the Bdnf promoter IV (Yu et al., 2015). These findings together suggest that neuronal activity can regulate TET3 function, which in turn controls the expression of target genes via altering methylation levels.

The importance ofTet3 in neural function is conserved across vertebrates, as knockdown of Tet3 in Xenopus by morpholino antisense oligonucleotide leads to marked developmental abnormalities, including malformation of the eye, small head, and early death. At the molecular level, Tet3 depletion causes a reduction in expression of master eye developmental genes (pax6, rx, and six3), primary neuronal markers (ngn2 and tubb2b), neural crest markers (sox9 and snail), and major sonic hedgehog (shh) signaling components (shh and ptc-1). Additionally, TET3-ChIP assays confirm the binding ofTet3 to the promoters of pax6, rx, six3, ptc-1, ptc-2, sox9, and ngn2. Furthermore, at the promoters of some of these target genes, there was found to be a developmental increase in 5hmC and decrease in 5mC from stage 10 to 19 in Xenopus development, which is perturbed when Tet3 is knocked down. These findings suggest that TET3 acts as an upstream activator of key neural developmental genes (Xu et al., 2012). Furthermore, these studies suggest that TET3 plays an important role in brain function that is conserved across animals.

Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
Business & Finance
Computer Science
Language & Literature
Political science