XB-ART-4580Mol Cell August 1, 2003; 12 (2): 425-35.
A mutant form of MeCP2 protein associated with human Rett syndrome cannot be displaced from methylated DNA by notch in Xenopus embryos.
MeCP2 is a DNA binding protein that represses transcription of methylated genes in vitro, but the endogenous function of MeCP2 in vivo is unclear. Here, we demonstrate that in Xenopus laevis embryos MeCP2 is a partner of the SMRT corepressor complex that regulates the expression of a neuronal repressor xHairy2a in differentiating neuroectoderm. The MeCP2/SMRT complex is bound to the promoter of the silenced xHairy2a gene and is displaced upon activation by the Notch intracellular domain (NICD). A truncated form of MeCP2 (R168X) found in patients with Rett syndrome cannot interact with the SMRT complex or fully activate xHairy2a during primary neurogenesis. This disruption of MeCP2 activity results in abnormal patterning of primary neurons during neuronal differentiation. Our results support a model whereby the dynamic association of MeCP2 with methylated DNA and the SMRT complex regulates a gene involved in cell fate decisions during primary neurogenesis in Xenopus.
PubMed ID: 14536082
Article link: Mol Cell
Species referenced: Xenopus laevis
Genes referenced: acta4 actc1 actl6a dll1 dnmt1 ecd esr-5 gast hdac1 hes4 hes5.1 hes7.2 isyna1 mecp2 ncam1 ncor2 neurod1 neurog1 neurog2 notch1 odc1 pcna rbpj sin3a smo tub tubb2b
Morpholinos: mecp2 MO1 sin3a MO1
Disease Ontology terms: drug-induced mental disorder
OMIMs: RETT SYNDROME; RTT
Article Images: [+] show captions
|Figure 1. Expression of Neuronal Differentiation Markers Is Altered in MeCP2-Deficient and R168X Embryos (A) Western blot with α-xMeCP2 of staged extracts from control morpholino (CMO)- and xMeCP2 morpholino (MMO)-injected embryos. An α-PCNA blot serves as a loading control for the MMO extracts. (B) Western blot with α-xMeCP2 of staged extracts from embryos coinjected with MMO and R168X sense RNA. Blast, blastula; Gast, gastrula; Neur, neurula; Tailb, tailbud extracts; and Tadp, tadpole. (C) Semiquantitative RT-PCR analysis of normal (WT), R168X, MeCP2 morpholino (MMO), and embryos coinjected with MMO and GlXMeCP2 (MMO/Resc) neurula stage 15 embryos. The expression levels of neural markers: Neurogenin (Ngnr), NeuroD, xNotch, xDelta, and additional control genes: muscle actin (mAct), ornithine decarboxylase (ODC), and Enhancer of split related genes (Esr1,4,5, 6) were equal in each case. xHairy2a expression was upregulated in the MMO neurula and downregulated in the R168X neurula. Neural β-tubulin (N-Tub) and neural cell adhesion molecule (NCAM) expression in the MMO and R168X neurula are inverse to that of xHairy2a. Expression levels were restored to normal in MMO/rescue neurula (MMO/Resc). −RT is a control reaction in which reverse transcriptase is omitted. (D and E) Whole-mount in situ for xHairy2a on wild-type (WT) and embryos half-injected with MMO at neurula stage. (F–L) Whole-mount in situ for neural β-tubulin of embryos half-injected with Hairy2a (F) and wild-type (G). The three stripes of neural β-tubulin expression represent primary neurons; “I,” “l,” and “m” are the intermediate, lateral, and medial stripes, respectively. (H) Embryos injected with MMO. The intermediate and lateral stripes of neural β-tubulin expression are absent. (I) As in (H) but the embryos were injected into one cell at two-cell stage with MMO (1/2 MMO) and a β-galactosidase mRNA (β-gal) as a lineage tracer. The intermediate and lateral stripes of neural β-tubulin expression are absent (white arrows) in the injected (inj.) half of the embryo that stains light blue for β-gal. (J) The patterns of neural β-tubulin expression were normal in MMO embryos rescued by GlXMeCP2. (K) Embryos injected with R168X RNA. Notice the appearance of ectopic patches (black arrows) of neural β-tubulin expression. (L) Whole-mount in situ as in (D) of embryos half coinjected with the R168X mRNA (1/2 R168X) and a β-galactosidase mRNA (β-gal). The ectopic patches (white arrows) of neural β-tubulin expression were present only in the injected (inj.) half of the embryo, light blue for β-gal, but not in the uninjected half of the embryo.|
|Figure 2. Mutant Forms of MeCP2 Have Reduced Capacity to Interact with the SMRT Complex (A) Protein gel of the fusion proteins used in pull-down experiments. Size markers are on the left. (B) Western blot of GST-xMeCP2 pulled down fraction from staged embryo extracts with antibodies to Sin3A (α-Sin3A), SMRT (α-SMRT), and HDAC1 (α-HDAC1). Blast, blastula; Gast, gastrula; Neur, neurula; and Tail, tailbud extracts. (C) Western blot of GST-SMRT(N) and SMRT(C) pull-down fractions from egg and neurula extracts with α-Sin3A, and α-MeCP2. GST-SMRT(N) pulls down Sin3A and MeCP2. (D) Western blot of GST-CBF1-bound fractions from staged extracts with α-SMRT, α-HDAC1, α-MeCP2, and α-Sin3A. (E) Western blots of staged extracts from embryos injected with Sin3A (SMO) and control (CMO) morpholinos. Sin3A is undetectable in the extracts of postgastrula stage Sin3A MO embryos. The stability of SMRT protein is not affected by the loss of Sin3A. (F) Western blot showing that GST-MeCP2 (human or Xenopus) can pull down SMRT (α-SMRT) from CMO-injected neurulae extracts but not from SMO extracts. (G) Cartoon (not to scale) of GST proteins used in pull-down experiments; GST-xMeCP2 equals full-length Xenopus MeCP2; GST-R306C carries a point mutation in the TRD, indicated by an asterix (*). GST-R168X corresponds to a human truncation mutant after the MBD. GST-TRD contains only the TRD and the C-terminal part of human MeCP2. (H) Protein gel of the fusion proteins used. Size markers are shown on the left. (I) Western blot of pulled down fractions from neurula extracts with GST, GST-xMeCP2, GST-R306C, GST-R168X, and GST-TRD, probed with α-Sin3A, α-SMRT, and α-HDAC1. GST-R168X is unable to bind these proteins, and the GST-R306C pull-down of all three proteins was reduced compared to GST-xMeCP2. (J) Neurula extracts immunoprecipitated (IP) with α-Sin3A and protein G sepharose-bound fraction tested for the presence of MeCP2 (α-MeCP2), Sin3A (α-Sin3A), HDAC1 (α-HDAC1), and SMRT (α-SMRT). IPs were done from wild-type (WT), MeCP2 morpholino (MMO), and R168X extracts. There is no MeCP2 in the MMO and R168X extract-derived Sin3A-bound complex.|
|Figure 3. Changes in DNA Methylation and Histone Acetylation Alters xHairy2a and Neural β-Tubulin Expression in Neuralized Animal Caps (A) Diagram illustrating how ectodermal explants were generated. Two-cell stage embryos were injected with neurogenin RNA (Ngnr) or neurogenin plus a plasmid that expresses Dnmt1 antisense RNA (asDnmt1). Embryos were cultured until stage 7 and ectodermal explants isolated. These were incubated in medium with or without a histone deacetylase inhibitor trichostatin (TSA) and caspase inhibitors z-DEVD-fmk. Similar explants were isolated from two-cell stage embryos injected with MeCP2 morpholino (MMO) without or with R168 mRNA (MMO/R168X). (B) The expression of specific transcripts, xHairy2a, neural β-tubulin (N-tub), and ornithine decarboxylase (ODC) was monitored by RT-PCR of untreated and TSA-treated explants from wild-type (WT) and MeCP2 morpholino (MMO)-injected embryos. XHairy2a expression is absent in untreated explants but is induced to a low level by neurogenin (Ngnr). TSA treatment or loss of methylation (by asDnmt1 injection) further induces xHairy2a expression, and has an additive effect. In the MMO explants, TSA alone strongly induced xHairy2a expression, whereas the effect of asDnmt1 was negligible. The expression of N-tub varies inversely to xHairy2a expression, whereas ODC is equal in all cases. (C) The same experiment as in (B) except that the RT-PCR analysis was of WT explants and explants from embryos injected with MMO and R168X (MMO/R168X). In the MMO/R168X explants, asDnmt1 alone strongly induced xHairy2a expression, whereas TSA had little effect. The expression of N-tub inversely mirrored xHairy2a levels, whereas ODC served as a ubiquitously expressed control. On the right side of the panel, the same experiment was repeated with explants from embryos injected with a methylation-dependent activator MBD-VP16 (MBD-VP16) mRNA. xHairy2a expression was strongly induced in untreated MBD-VP16 explants. Ngnr treatments lead to low expression of N-tub without effecting xHairy2a levels. Loss of DNA methylation by inhibition of Dnmt1 (asDnmt1) abolished MBD-VP16 binding, and xHairy2a expression was reduced as a consequence. ODC expression was similar in all cases.|
|Figure 4. MeCP2 and SMRT Are Displaced from xHairy2a Promoter by NICD (A) Schematic of a 3 kb PstI fragment of the xHairy2a promoter. CpGs are denoted by horizontal bars. The region around the transcription start site (arrow) is a CpG island which includes a TATA box (black bar) and two Su(H)/CBF1 binding sites (hatched bars). The upstream CpG-rich region corresponds to Vi repeats that are not CpG islands. Gray bars denote the regions analyzed by PCR after chromatin immunoprecipitation (ChIP) with α-MeCP2 and α-SMRT. The striped bar denotes the region that was bisulphite sequenced to determine the methylation status of the CpGs. (B and C) ChIP analysis of wild-type (WT) and Notch-intracellular domain (WT + NICD)-injected neurula stage embryos with α-MeCP2 (ChIP-MeCP2) and α-SMRT (ChIP-SMRT) antibodies. The numbers above the lanes represent the gray fragments in (A) that were PCR amplified after ChIP. Note that fragment 4 (lane 4) was enriched after ChIP for MeCP2 in WT but not in NICD embryos, indicating reduced binding of MeCP2 to activated xHairy2a promoter. ChIP with α-SMRT show strong binding of SMRT to fragments 4–6 and 2 in WT but not in NICD embryos. (D and E) Similar ChIP analysis was done on R168X (R168X) and R168X and NICD mRNA (R168X + NICD)-coinjected neurula stage embryos with α-MeCP2 (ChIP-MeCP2) and α-SMRT (ChIP-SMRT). Fragment 4 (lane 4) was enriched after ChIP for MeCP2 both in R168X embryos and in R168X + NICD embryos, indicating persistent binding of MeCP2 independent of the presence of NICD. SMRT was enriched at fragments 4–6 and 2 in R168X but not in R168X + NICD embryos, similar to its binding pattern in WT embryos. (F and G) ChIP analysis for MeCP2 and SMRT in Sin3A morpholino SMO and SMO + NICD-injected embryos. Notice that as in R168X embryos, MeCP2 remains bound to fragment 4 upon removal of SMRT by NICD. (H and I) Input controls for (B)–(E). (J) Bisulphite sequencing of the xHairy2a promoter (striped bar in [A]) in neural stage 15 WT and Notch intracellular domain (WT + NICD)-injected embryos. In each case ten independent sequencing reactions are shown. Methylated CpGs are denoted by a filled circle and nonmethylated CpGs are denoted by an open circle. The overall pattern is similar between both types of DNA.|
|Figure 5. Model for the Regulation of xHairy2a by MeCP2 (Wild-Type and Mutants) (A) In proneural cluster cells (left panel), MeCP2 binds to a methylated region upstream of the xHairy2a promoter and is tethered to the SMRT-complex bound to Su(H)/CBF1 via interaction of its TRD with Sin3A. The promoter of xHairy2a is in a stably repressed state (indicated by the bold red X) which maintains the neuronal potential of the cell. Upon activation of Notch by Delta, the Notch-intracellular domain (NICD) (right panel) binds to CBF1 and displaces the complex from the promoter region resulting in normal activation of the xHairy2a gene. This leads to suppression of neuronal differentiation genes such as N-β-tubulin in xHairy2a-expressing cells, and ensures a balanced selection of cells for neuronal differentiation. (B) When MeCP2 is absent (left panel) this results in unstable (leaky) transcriptional repression of xHairy2a by SMRT complex (indicated by hatched red X) and reduced number of cells that remain competent for neuronal differentiation. NICD may more easily displace SMRT, resulting in enhanced expression of xHairy2a (right panel). (C) The truncated R168X mutant MeCP2, lacking the TRD domain (indicated by the interrupted red disc), does not interact with the SMRT complex, but still binds the upstream methylated regions. This possibly also results in stable repression of xHairy2a as in (B) (indicated by the red X in the left panel). However, when NICD binds toCBF1, only a partial derepression occurs, because R168X remains at the upstream methylated regions after removal of SMRT. This may restrict the access of additional transcriptional activators (?) to the promoter region occupied by R168X and hence less potent xHairy2a activation (right panel). The overall result is an increased number of cells that are competent to express early neuronal differentiation genes, resulting in an increased number of primary lateral neurons. (Only the relevant partners of the SMRT complex are shown for clarity.)|
|Supp. Fig 1. Spatial Expression of xMeCP2 [mecp2] during Xenopus Development Whole-mount in situ hybridization was performed with xMeCP2 antisense RNA on staged Xenopus embryos. The embryos were cleared with Murray’s clear solution after staining to allow three-dimensional observation of all tissues. (A) Egg (side view); (B) blastula (side view), stage 7; (C) gastrula (anterior view) stage10.5; (D) neurula (dorsal view) stage 15; (E) tailbud (side and dorsal view) stage 25; and (F) tadpole (side view) stage 35. Note that MeCP2 is not expressed in the intermediate neurons ( IN ) at stage 15 (D). An, animal pole; Bp, blastopore; DM, dorsal mesoderm; ECD, ectoderm; Veg, animal pole; and VM, ventral mesoderm.|
|mecp2 (methyl CpG binding protein 2) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF 1 stage lateral view, anterior left, dorsal up.|
|mecp2 (methyl CpG binding protein 2) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF 7 stage lateral view, anterior left, dorsal up.|
|mecp2 (methyl CpG binding protein 2) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF 10.5 stage anterior view.|
|mecp2 (methyl CpG binding protein 2) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF 15 stage lateral view, anterior left, dorsal up.|
|mecp2 (methyl CpG binding protein 2) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF 25 stage (top) lateral view, anterior left, dorsal up, and (bottom) dorsal view, anterior left.|
|mecp2 (methyl CpG binding protein 2) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF 35 & 36 stage lateral view, anterior left, dorsal up.|
|Figure S2. Phenotypes of Microinjeceted Embryos (A) Control embryos injected with 500 pg β-galactosidase mRNA were phenotypically normal (see Supplemental Table S1). (B) Injection of 10 ng of a control morpholino (CMO) into a 2-cell embryo did not lead to any phenotypic abnormalities. (C) Injection of a MeCP2 morpholino (MMO) induced microcephaly and axial defects. (D) The phenotype of MMO-injected embryos could be rescued by coinjection of Xenopus MeCP2 in which the 5’UTR was replaced with a globin 5′UTR (MMO+GlxMeCP2). (E) The phenotype of MMO-injected embryos was partially rescued by coinjection of mutant R168X MeCP2 (MMO + R168X). The embryos were bent and displayed abnormal hyperactive swimming motion. (F) The short dorsal axis phenotype of MMO injected embryos was to a great extent rescued by coinjection with mutant R306C MeCP2 (MMO + R306C). (G) Coinjection of 50 pg, 300 and 750 pg human MeCP2 (MMO + hMeCP2) was able to rescue the MMO phenotype in a dose-dependent manner. All the embryos shown in (A)-(G) are at stage 25 or the equivalent, lateral view, except (E) - dorsal view.|