The BAF (BRG1/BRM-Associated Factor) chromatin-remodeling complex exhibits ethanol sensitivity in fetal neural progenitor cells and regulates transcription at the miR-9-2 encoding gene locus

Abstract

Fetal alcohol spectrum disorders are a leading cause of intellectual disability worldwide. Previous studies have shown that developmental ethanol exposure results in loss of microRNAs (miRNAs), including miR-9, and loss of these miRNAs, in turn, mediates some of ethanol's teratogenic effects in the developing brain. We previously found that ethanol increased methylation at the miR-9-2 encoding gene locus in mouse fetal neural stem cells (NSC), advancing a mechanism for epigenetic silencing of this locus and consequently, miR-9 loss in NSCs. Therefore, we assessed the role of the BAF (BRG1/BRM-Associated Factor) complex, which disassembles nucleosomes to facilitate access to chromatin, as an epigenetic mediator of ethanol's effects on miR-9. Chromatin immunoprecipitation and DNAse I-hypersensitivity analyses showed that the BAF complex was associated with both transcriptionally accessible and heterochromatic regions of the miR-9-2 locus, and that disintegration of the BAF complex by combined knockdown of BAF170 and BAF155 resulted in a significant decrease in miR-9. We hypothesized that ethanol exposure would result in loss of BAF-complex function at the miR-9-2 locus. However, ethanol exposure significantly increased mRNA transcripts for maturation-associated BAF-complex members BAF170, SS18, ARID2, BAF60a, BRM/BAF190b, and BAF53b. Ethanol also significantly increased BAF-complex binding within an intron containing a CpG island and in the terminal exon encoding precursor (pre)-miR-9-2. These data suggest that the BAF complex may adaptively respond to ethanol exposure to protect against a complete loss of miR-9-2 in fetal NSCs. Chromatin remodeling factors may adapt to the presence of a teratogen, to maintain transcription of critical miRNA regulatory pathways.

Keywords: BRG1; DNAse hypersensitivity; Fetal alcohol spectrum disorders; MiR-9.

Figure 1. Specific BAF complex subunit transcripts are induced in NSCs following ethanol exposure

(a) mRNA levels of BAF subunits in neurospheres after 120mg/dl and 320mg/dl ethanol treatment calculated as fold change normalized to control (0mg/dl ethanol). Error bars shows standard error of the mean. (b) Schematic diagram shows the components of the BAF complex in embryonic stem cells (esBAF), neural progenitor cells (npBAF) and in neurons (nBAF). Orange color indicates the subunits which are maintained in different cell maturation stages from embryonic stem cells state through the differentiation and whose mRNA levels were induced by ethanol treatment. Green color indicates subunits which are exclusively found in the BAF complex in differentiated neuronal states, and whose mRNA levels were induced by ethanol treatment. White color indicates the subunits whose mRNA levels were not affected by ethanol.

Figure 2. Ethanol-sensitive BAF complex transcripts are transiently induced during neural differentiation

(a) BAF complex subunit expression over a three-day course of mitogen withdrawal, BDNF-induced neural differentiation calculated as fold change normalized to expression in NSCs (Day 0). Error bars show standard error of the mean. (b) Phase contrast microscopy of a neurosphere (left) and cells over three days (D1, D2 and D3) of neural differentiation. Scale bar, 100 µm.

Figure 3. BRG1 protein expression in NSCs

(a) Western immunoblot shows a single BRG1 band at ~238KD detected in the nuclear extract of neurosphere cultures but not in the cytoplasmic extract. (b) BRG1 protein levels in neurospheres after 0mg/dl, 120mg/dl and 320mg/dl ethanol treatment. Lower panel shows the corresponding immunoblot when probed for β-Actin as a loading control. (c) Bar graph, depicting the quantification of the western blot, shows that ethanol did not alter BRG1 protein expression, consistent with the lack of effect on BRG1 mRNA expression. The vertical axis shows the ratio of the band density of BRG1 to the ratio of the band density of β-Actin. Error bars indicates standard error of the mean. (d) Western blot analysis of BRG1 immunoprecipitation in cytoplasmic and nuclear cellular fractions probed with anti-BRG1 antibody. BRG1 is specifically precipitated with anti-BRG1 antibody (blue text), but not with an isotype-specific IgG control antibody (red text). Efficiency of anti-BRG1 immunoprecipitation is indicated by the relative depletion of BRG1 from the nuclear supernatant and enrichment in the immuno-precipitate (IP).

Figure 4. BRG1-containing complexes associate with both DNAse I-hypersensitive and insensitive sites on the miR-9-2 gene locus

(a) Positional representation of POU5F1/Oct4, c-myc and REST transcription regulatory factor binding sites along the human pre-miR-9-2 gene locus identified with the UCSC genome browser ENCODE analysis hub track. Colored circles indicate locations for primer pairs for pri-miR-9-2 regions 1,2,3,4 and 5 and the pre-miR-9-2 coding region. (b) qPCR results of BRG1-ChIP from neurospheres. Primers were used to amplify 6 distinct positions along the pri-miR-9-2 gene locus and a genetically sparse region on mouse chromosome 6. Vertical axis shows fold change relative to an IgG pulldown control. Error bars indicate standard error of the mean. (c) qPCR of neurosphere DNA following digestion with 20, 60 and 120U of DNAse I at the 6 positions along the pri-miR-9-2 gene locus. The vertical axis shows fold amplification relative to DNAse I-untreated neurosphere DNA.

Figure 5. ChIP analysis indicates that ethanol exposure increases BRG1 binding to DNAse I-resistant/CpG island containing region 3 and the pre-miR-9-2 coding exon

Bar graph shows the effect of ethanol on BRG1-association with regions 1 to 5 and the pre-miR-9-2 exon-coding region of the primary (pri)-miR-9-2 coding locus. Primers for a gene desert on chromosome 6 (not predicted to bind any transcription factors) shows the specificity of the immuno-precipitation. Vertical axis shows the fold change of the specific associated DNA region to BRG1 in ethanol treated group relative to control group. Error bars indicate standard error of the mean.

Figure 6. siRNA mediated BAF155 and BAF 170 knockdown results in specific downregulation of miR-9

(a) Levels of BAF155 and BAF170 mRNA after BAF155/BAF-170 knockdown, normalized to β-Actin mRNA expression levels and calculated as fold change relative to scrambled siRNA control group. (b) Cyclophilin-B mRNA level after either scrambled siRNA, cyclophilin-B siRNA or combined BAF155/BAF170 siRNA transfection, normalized to β-Actin mRNA expression levels and calculated as fold change relative to scrambled control. (c) mir-9 levels measured by qRT-PCR normalized to U6 snRNA expression levels and calculated as fold-change relative to control scrambled siRNA transfected group. BAF155/170 knockdown resulted in a significant decrease in miR-9. However, cyclophilin-D knockdown did not result in altered miR-9 expression. Error bars indicate standard error of the mean.