Fig 2: Pharmacological manipulation of gDNA methylation in F11 cells alters 5-mC levels at CpG sites in Cacna1b e37a locus and e37a inclusion.(A) 5-Azacytidine (5-Aza) inhibits the activity of DNA methyltransferases (DNMTs), and 2-hydroxyglutarate (2-HG) inhibits the activity of ten-eleven translocation proteins (TETs). (B) Global 5-mC levels in F11 cells after 3 day exposure to different concentrations of 5-Aza (left) (ANOVA P value = 0.056, Dunnett’s multi-comparison P values for 0 vs 0.5 µM = 0.5371, 0 vs 1 µM = 0.1333, 0 vs 5 µM = 0.0688, 0 vs 10 µM = 0.0421, and 0 vs 10 µM = 0.0280) and associated qRT-PCR of Cacna1b e37a relative to e36 levels (right) (ANOVA P value = <0.0001, Dunnett’s multi-comparison P values for 0 vs 0.5 µM = 0.9874, 0 vs 1 µM = 0.9998, 0 vs 5 µM = 0.0015, 0 vs 10 µM = <0.0001) (three biological replicates per condition). (C) Location of 5 CpG sites in Cacna1b e37a locus (open circles) and CTCF binding motif (underlined). (D) 5-mC levels at CpG sites in Cacna1b e37a locus in control F11 cells (DMSO, left) and after 3 day treatment with 5 µM 5-Aza (right). Methylated (5-mCpG, filled circle) and unmethylated (CpG, open circle) sites are shown for independent clones ordered from most to least methylated. 5-mC levels for both conditions (two-way ANOVA P value = 0.0260 for treatment factor) (three biological replicates per condition). (E) qRT-PCR of Cacna1b e37a relative to e36 in F11 cells in the absence and after 3 day treatment with 5 µM 5-Aza for cells expressing control siRNA (Con) or Ctcf siRNA (t-test P value = 0.0219 for untreated vs 5-Aza with Con siRNA; and = 0.9303 for untreated vs 5-Aza with Ctcf siRNA) (three biological replicates per condition). (F) Global 5-mC levels in F11 cells following 1 day exposure to different concentrations of 2-HG (left) (ANOVA P value = 0.0022, Dunnett’s multi-comparison P values for 0 vs 0.01 mM = >0.9999, 0 vs 0.1 mM = 0.9878, 0 vs 0.5 mM = 0.6062, 0 vs 1 mM = 0.4755, and 0 vs 10 mM = 0.0010), and associated qRT-PCR of Cacna1b e37a relative to e36 levels (right) (ANOVA P value = 0.0001, Dunnett’s multi-comparison P values for 0 vs 0.01 mM = 0.9997, 0 vs 0.1 mM = 0.6026, 0 vs 0.5 mM = 0.1799, 0 vs 1 mM = 0.0487, and 0 vs 10 mM = 0.0003) (three biological replicates per condition). (G) 5-mC levels at CpG sites in Cacna1b e37a locus in control F11 cells (H2O, left) and after 1 day treatment with 10 mM 2-HG (right). Methylated (5-mCpG, filled circle) and unmethylated (CpG, open circle) sites are shown for independent clones ordered from most to least methylated. 5-mC levels for both conditions (two-way ANOVA P value = 0.0745 for treatment factor) (three biological replicates per condition). Biological replicates represent independent cell cultures, treatment and transfections.

Fig 3: Proposed model for cell-specific epigenetic modifications in a synaptic calcium ion channel gene controls cell-specific splicing.In naïve animals, in most neurons, Cacna1b e37a locus is hypermethylated (5-mC) and CTCF is presumed not to bind this locus based on our results in the DRG-derived F11 cell line. During splicing, e37a is skipped and Cacna1b mRNAs include e37b. In contrast, in Trpv1-lineage neurons, Cacna1b e37a locus is hypomethylated and is presumed to be permissive for CTCF binding. Hypomethylation promotes e37a inclusion and both Cacna1b e37a and e37b mRNAs are expressed. E37a confers strong sensitivity to the CaV2.2 channel to inhibition by µ-opioid receptors (µOR; [Andrade et al., 2010; Raingo et al., 2007]). Morphine is more effective at inhibiting e37a-containing CaV2.2 channels. After peripheral nerve injury that results in pathological pain, methylation level of Cacna1b e37a locus is increased, CTCF binding is presumed to be impaired, and Cacna1b e37a mRNA levels are decreased. This disrupted splicing pattern is associated with reduced efficacy of morphine in vivo (Jiang et al., 2013).

Fig 4: The DNA binding protein CTCF binds CACNA1B e37a but not e37b in vivo.(A) The 100 vertebrate basewise conservation track for ~30 kb region of CACNA1B (Hg19; chr9:104,970,785–141,003,093). Five conserved elements align to e35, e36, e37a, e37b, and e38. (B) ChIP-seq signals for CTCF binding in nine different human cell lines are aligned to CACNA1B region in A. Y-axes for ChIP-seq tracks are scaled to the maximum signal within the selected region. In total, there is CTCF signal at CACNA1B e37a in 27 of 50 human cell lines. Link to the UCSC genome output (https://genome.ucsc.edu/s/ejlopezsoto/Cacna1b%20e35%20to%20e38%20conservation%20track) (ENCODE Project Consortium, 2012).

Fig 5: CTCF binds mouse Cacna1b e37a in vitro.(A) Mouse Cacna1b e37a and e37b sequences are ~60% identical. Nucleotides conserved across exons are shown as dashes in the e37b sequence. A 15 nucleotide CTCF binding motif in e37a (underlined) is not present in e37b. (B) Gel shift assay for CTCF and a Cacna1b e37a DNA probe (139 bp) that also contains 5’ and 3’ flanking intron sequences alone (0 ng) or with increasing concentrations of recombinant CTCF (10–200 ng). Quantification of CTCF bound to e37a probe relative to total e37a probe at different CTCF concentrations (red line represents fitted Hill equations r2 = 0.9338; Kd = 0.0461 ± 0.0055 ng) (at least two biological replicates per CTCF concentration). (C) Gel shift assay for CTCF and DNA probes (139 bp) containing either Cacna1b e37a or e37b alone (lanes 1 and 5) or preincubated with 50 ng recombinant CTCF (lanes 2 and 6), CTCF plus 0.5 µg anti-CTCF (lanes 3 and 7), or CTCF plus 1000-fold excess unlabeled e37a or e37b probes (lanes 4 and 8). (D) Quantification of CTCF bound to e37 probes relative to total e37a or e37b probe (ANOVA P value = 0.0006 for e37a probe, Dunnett’s multi-comparison P values for lane 1 vs 2 = 0.0067, lane 1 vs 3 = 0.0009, lane 1 vs 4 = 0.9999) (three biological replicates per condition). Biological replicates represent independent gel shift assays.