By daylight, DNA methyltransferase 3L (DNMT3L) takes on the modest role of a catalytically inactive but stimulatory binding partner for its more popular methyltransferase siblings (DNMT3A and DNMT3B). But, when the sun goes down, DNMT3L gives new meaning to protein moonlighting.
DNMT3L is located on chromosome 21, which has made it a prime candidate for the Down syndrome (Trisomy 21) DNA methylation signature. Down syndrome is characterized by genome-wide differences in gene expression and DNA methylation across multiple tissues. The DNA methylation profile is skewed towards hypermethylation in all tissues except for blood, and a subset of genes are consistently hypermethylated across many tissues. While DNMT3L can create the Down syndrome DNA hypermethylation profile at select genes, it hadn’t yet been examined at the genome-wide level. Thankfully, two new studies bring forth the epigenomic landscape of DNMT3L overexpression in human neuronal cells and a mouse model.
DNMT3L Overexpression Across Neuronal Differentiation
Previously, the lab of Janine LaSalle (University of California, Davis) used whole genome bisulfite sequencing (WGBS) to confirm and expand upon DNA hypermethylation in Down syndrome brain and to also examine the DNA hypomethylation profile resulting from increased RUNX1 dosage in newborn dried blood spots. In their latest, the group stably overexpressed DNMT3L in a human neuronal cell line (SH-SY5Y) via piggyBac transgenesis and utilized WGBS to examine the genome-wide impact on DNA methylation across 3 differentiation timepoints.
They found that DNMT3L overexpression results in:
- Slight global and CpG island hypermethylation
- Thousands of differentially methylated regions (DMRs), which:
- Are skewed towards hypermethylation, but 1/3rd are hypomethylated
- Map to genes to involved in neurodevelopment, cellular signaling, and gene regulation
- Show a moderate scale of change that is similar to what it is seen in Down syndrome
- The hypermethylated DMRs are enriched for within regions marked by H3K4me3 and bivalent chromatin as well as regions known to show differential methylation across multiple tissues in Down syndrome, which includes the HOX gene clusters
- The hypomethylated DMRs are enriched for within heterochromatin and only the set from undifferentiated cells is enriched for within differentially methylated sites from fetal down syndrome brain
Overall, these findings suggest a mechanism where excess DNMT3L binds to regions of bivalent chromatin that lose H3K4me3 and results in hypermethylation of those loci, while also resulting in hypomethylation of heterochromatin by releasing DNMT3A from it and allowing it to target accessible regions. Furthermore, the findings demonstrate that DNMT3L only results in a facet of the Down syndrome DNA methylation profile.
DNMT3L Needs a Methylation Independent Mechanism … STAT
While this methylation mystery could easily take a new route investigating other components of epigenetic machinery on chromosome 21, the labs of Jie Lu (China Medical University) and Yu Wang (Shenyang Women & Children’s Hospital) persistence paid off with a clever combination of cell culture and mouse models.
First, they utilized early differentiated human neural precursor cells, where they found that DNMT3L overexpression:
- Results in a profile of ~350 differentially methylated CpGs on the EPIC array, which:
- Display moderate changes that are skewed towards hypermethylation, but 1/3rd are hypomethylated
- Map to genes involved in neuronal differentiation and synaptic formation
- Has a stronger impact on gene expression, as assayed by RNA-seq, where there is a slight skew towards down-regulation and an enrichment for genes involved in immune processes
Interestingly, the group didn’t observe a significant correlation between DNA methylation and gene expression, and thus sought to investigate whether mechanisms independent of DNA methylation were at play. They utilized immunostaining and western blots to show that DNMT3L decreases proliferation and increases differentiation in neural precursor cells. Mass spectrometry analysis of DNMT3L binding partners identified 1144 proteins, including the STAT1 and STAT3 transcription factors, where it increased their phosphorylation and translocation to the nucleus.
Then, by utilizing phosphorylation inhibitors for STAT1 and STAT3, they were able to:
- Abolish the pro-differentiation effects of DNMT3L overexpression in SH-SY5Y cells
- Rescue cell fate changes in a mouse model that was engineered via CRISPR/Cas9 to overexpress DNMT3L
- Treatment with decitabine, a DNA methylation inhibitor, did not produce the same effect
Moonlight’s Illumination of Methylation Machinery
Taken together, these two studies show that DNMT3L overexpression in the developing brain contributes to some—but not the majority—of the Down syndrome DNA methylation profile. Furthermore, DNMT3L is a moonlighting protein, as DNA hypermethylation isn’t the only result of its overexpression in developing neurons.