With the demonstration that up to 80% of Rett syndrome cases are caused by mutation in a methyl DNA-binding protein, MeCP2, the analysis of neurological effects and mental retardation in this disease must focus on associated changes in chromatin. The working hypothesis is that chromatin remodeling based on methylation must be modified to turn on or off critical proteins active in brain function. MeCP2 was suggested to function as a transcriptional repressor in this process. How MeCP2 represses transcription is controversial. Earlier work suggested that repression is through association with a histone deacetylase complex containing mSin3A and HDAC1. For example, independent studies by the labs of Drs. Bird and Wolffe have suggested that MeCP2 associates with HDAC1 and mSin3A in vitro, and co-immunoprecipitates with these proteins in vivo. However, a more recent publication by Bird's group showed MeCP2 purified by conventional chromatography from rat brain has no stably associated proteins. Moreover, they showed that MeCP2 in the crude extract fractionated with the same apparent molecular weight (400-500 kDa) as the purified protein, indicating that MeCP2 should have no stably associated proteins even in the crude extract. Our group has performed unbiased immunopurification of MeCP2 from human HeLa cells. We found that MeCP2 has no stably-associated proteins. Specifically, it neither co-copurifies nor co-fractionates with mSin3A and HDAC1. Our purification is one step, and the washing conditions are very mild (0.2 M salt concentration). Our data, together with the recent paper from A. Bird's lab, suggests that MeCP2 is not in stable complexes with other proteins. Very recently, Harikrishnan and colleagues reported that MeCP2 functions through association with brahma-containing SWI/SNF chromatin remodeling complex. Brahma (also called brm) and BRG1 are both ATPase subunits in mammalian SWI/SNF complexes. This latest paper contradicted data from several groups, including ours, so we decided to examine this issue carefully. We obtained antibodies from Harikrishnan and colleagues and repeated their experiments. We found that MeCP2 purified from mouse brain has no stably associated proteins. Specifically, MeCP2 does not co-purify, co-fractionate or co-localize with brahma-associated complex. Our data are consistent with previous findings that mice inactivated brm, the component in SWI/SNF that was suggested to link with MeCP2, develop and grow normally, whereas mice with various MeCP2 mutations develop severe neurological disorders that mimic RS. The absence of RS-like phenotypes in brm-null mice provides genetic evidence against involvement of brm in pathogenesis of RS as well as in the primary function of MeCP2. This work has been submitted for publication. Interestingly, we found that MeCP2 is a phosphoprotein by mass spectrometry. A previous study has shown that MeCP2 becomes hyperphosphorylated when neurons are induced to undergo depolarization. This hyperphosphorylation inhibits MeCP2 binding activity, which correlates with increased transcription of BDNF, an important regulator for neuronal function. We scaled up our purification to obtain enough phosphorylated MeCP2 for mass spectrometry identification of its phosphorylation sites. In collaboration with Dr. Yi Sun's lab at UCLA, we identified 5 phosphorylation sites in MeCP2 from normal mouse brain, and 2 additional sites in MeCP2 from mice induced to undergo seizure. We have also successfully generated one phosphorylation site-specific antibody for Ser80, and were able to use this antibody to confirm phosphorylation at Ser80 in vivo. In addition, we show that although the total level of MeCP2 phosphorylation is increased during depolarization of neurons, the level of phosphorylation at Ser80 is actually decreased. Furthermore, we find that mutating Ser80 to alanine (S80A) reduced induction of BDNF gene expression in cultured neurons derived from either normal mice or MeCP2-null mice. These data suggest an important role of phosphorylation at Ser80 in MeCP2 function. We are currently investigating the mechanism of how Ser80 phosphorylation affects MeCP2 function through collaboration with Bird and Sun's groups. By chromatin-IP, we found that S80A mutation increased the level of MeCP2 bound to the BDNF promoter in vivo, which correlates with the observation that the same mutation reduced induction of BDNF transcription. However, S80A (and S80D) mutation has no significant effect on the DNA binding activity of MeCP2 in vitro, arguing against a simple mechanism of direct stimulation of DNA binding by S80 phosphorylation.
| Hu, Keping; Nan, Xinsheng; Bird, Adrian et al. (2006) Testing for association between MeCP2 and the brahma-associated SWI/SNF chromatin-remodeling complex. Nat Genet 38:962-4; author reply 964-7 |
