Stability and DNA Methyltransferase Activity of DNMT3A are Maintained by Ubiquitin-Specific Peptidase 11 (USP11) and Sumoylation Countering Degradation

Abstract

DNA methyltransferase 3A (DNMT3A) plays crucial roles in mammalian development and hematopoiesis. DNMT3A protein instability is associated with blood diseases such as myelodysplastic syndrome (MDS), acute myeloid leukemia (AML), as well as Tatton-Brown-Rahman syndrome, an overgrowth disorder. We found that certain unstable DNMT3A mutations cause DNMT3A localization changes, resulting in loss of function. This mislocalization is partially rescued by E1 enzyme inhibition or stable USP11 expression, as DNMT3A stability is maintained by deubiquitinating enzyme USP11 countering degradation by CUL4-DCAF8 E3 ligase. We also found that USP11 enhances DNMT3A SUMOylation by promoting the interaction between DNMT3A and SUMO E3 ligases. DNMT3A SUMOylation also is essential to maintain DNMT3A protein stability. Furthermore, we found USP11 enhances the binding of DNMT3A to the polycomb complex and maintains DNMT3A DNA methyltransferase (MTase) activity. This study uncovers the mechanism for DNMT3A protein turnover through USP11, which is essential to DNMT3A function, and may be a therapeutic approach for diseases caused by DNMT3A protein instability.

Figure 1. Targeted CRISPR screening identifies E3 ligases and deubiquitinating enzymes essential for DNMT3A protein turnover.

A, The image depicts DNMT3A and DCAF8 protein expression followed by western blotting using DNMT3A, DCAF8 and GAPDH antibodies. B, The image depicts the scheme of two major isoforms of DNMT3A exist: DNMT3A1 (full-length) and DNMT3A2 (short) C, The image depicts the scheme of ubiquitin-proteasome inhibitor assay for confirming DNMT3A protein turnover changes. D, The graph depicts the protein expression of DNMT3A WT and unstable mutants (W297del and G685R) with ubiquitin-proteasome inhibitors. Alterations in DNMT3A protein stability after administration of proteasome inhibitor (MG132), a CRL inhibitor (MLN4924), an E1 and ubiquitin enzyme inhibitor (TAK-243) Stability ratio of MFI of DNMT3A-GFP versus MFI of mCherry before and after treatment with the ubiquitin-proteasome inhibitors as measured by flow cytometry 48 hours after transfection. E, Schematic of targeted CRISPR screening to identify ubiquitin modulators essential for DNMT3A protein turnover. Kelly cells were engineered to constitutively overexpress Cas9 and the indicated bicistronic DNMT3AW297Del reporter and then infected with sgRNA libraries targeting ubiquitin ligases. Nine days after infection, we sorted both the top and bottom 5% of cells for DNMT3A-GFP expression. F, the graph depicts the gene enrichment score and P value in targeted CRISPR screening for ubiquitin modulators. DCAF8, USP20, and USP28 genes (red) and USP10, USP11 and USP41 (blue) were enriched and statistically significant.

Figure 2. USP11 regulated DNMT3A protein turnover countering protein degradation by FCAF8

A, Schematic of immunoprecipitation (IP) mass spectrometry to identify deubiquitinating enzymes essential to interact with DNMT3A. Kelly cells were engineered to constitutively overexpress the indicated bicistronic GFP-DNMT3AW297Del reporter and treated with TAK-243. To subjects cell lysate was performed I P and mass spectrometry analysis. B, The graph depicts the enriched score of the interaction between DNMT3AW297del and deubiquitinating enzymes with or without 1nM TAK-243 for 6 hours. C. 293T cells were transfected with GFP-DNMT3A1W297Del. Cell lysates were immunoprecipitated with anti-GFP antibody, following the detection of USP11-bound DNMT3A1W297Del with anti-USP11 antibody. D. 293T cells were transfected with GFP-DNMT3A1W297Del. Cell lysates were immunoprecipitated with anti-GFP antibody, following the detection of DNMT3A1W297Del ubiquitination with anti-ubiquitin antibody. E, F. In the presence or absence of DCAF8, 293T cells were transfected with GFP-DNMT3A1W297Del and/or Myc-USP11. Cell lysates were immunoblotted following the detection of DNMT3A1W297Del, USP11 and DCAF8 protein expression with anti-GFP -Myc and -DCAF8 antibodies. F: The graph depicts the protein expression in Figure 2E (data are shown as the mean ± S.E.M, *, p = 0.0310) The W297del samples were set as 1.

Figure 3. DNMTA deubiquitination restored the mislocalization of DNMT3AW297del

A, B 293T cells were transfected with GFP-DNMT3A WT or -W297del and were stained anti-USP11 (rabbit) antibodies followed by anti-rabbit Alexa 647 (yellow). DNMT3A expression detected GFP (Green), nuclei and cytoplasm detached, DsRed (Red). Nuclei were visualized with DAPI (Blue). Laser scanning microscopy (Keyence) was used for the imaging. C, The graph depicts the subcellular localization pattern of GFP-DNMT3A WT or -W297del. D, 293T cells were transfected with GFP-DNMT3A WT or -W297del. nuclear, membrane/organelle and cytoplasmic fractions were isolated, following the detection of DNMT3A with anti-GFP antibody. GAPDH was used as a loading control for cytoplasm, AIF for membrane/organelle, and LAMIN-B for nuclear proteins. E, 293T cells were transfected with GFP-DNMT3A1W297del and were stained anti-USP11 (rabbit) antibodies followed by anti-rabbit Alexa 647 (yellow) with or without 1nM of TAK-243 for 6 hours. DNMT3A expression detected GFP (Green), nuclei and cytoplasm detached, DsRed (Red). Nuclei were visualized with DAPI (Blue). Laser scanning microscopy (Keyence) was used for the imaging. The graph depicts the subcellular localization pattern of GFP-W297del with or without TAK-243. F, 293T cells were transfected with GFP-W297del and USP11 were stained anti-USP11 (rabbit) antibodies followed by anti-rabbit Alexa 647 (yellow), DNMT3A expression detected GFP (Green), nuclei and cytoplasm detected DsRed (Red), nuclei were visualized with DAPI (Blue). Laser scanning microscopy (Keyence) was used for the imaging. The graph depicts the subcellular localization pattern of GFP-W297del with or without USP11 overexpression.

Figure 4. USP11 induced DNMT3A SUMOylation and its essential for DNMT3A protein stability

A. Schematic of immunoprecipitation (IP) mass spectrometry to identify comparing DNMT3A stable (DNMT3A WT or W297Del in the absence of DCAF8) or an unstable protein (W297Del) interactome. 293T cells were engineered to constitutively overexpress the indicated bicistronic GFP-DNMT3A WT or W297Del reporter in the presence or absence of DCAF8. To subjects cell lysate was performed I P and mass spectrometry analysis. B, C. The graph depicts the enriched score of the interaction proteins with stable DNMT3A or unstable DNMT3A. D. 293T cells were transfected with FLAG-DNMT3A1, Myc-USP11 and HA-SUMO3. Cell lysates were immunoprecipitated with anti-FLAG (DNMT3A) antibody, following the detection of DNMT3A1 SUMOylation with anti-HA antibody. E. 293T cells were transfected with FLAG-DNMT3A1, Myc-USP11. Cell lysates were immunoprecipitated with anti-FLAG (DNMT3A) antibody, following the detection of DNMT3A1 endogenous SUMOylation with anti-SUMO2/3 antibody. F. The scheme depicts TAK-981, a first-in-class Inhibitor of SUMO-Activating enzymes. G, H. Kelly or THP- cells were cultured with 1uM ofTAK-981 for 24 hours. Cell lysates were detected with anti-DNMT3A, -SUMO2/3 and -B-actin antibodies.

Figure 5. USP11 induced DNMT3A SUMOylation and its essential for DNMT3A protein stability

A, 293T cells were transfected with FLAG-DNMT3A1 or 2 and Myc-USP11. Cell lysates were immunoprecipitated with anti-FLAG (DNMT3A) antibody, following the detection of DNMT3A interacted USP11 with anti-myc antibody for western blotting. B, The graph depicts the enriched score of the interaction between DNMT3AW297del and deubiquitinating enzymes with or without 1nM TAK-243 for 6 hours. C, 293T cells were transfected with FLAG-DNMT3A1 and Myc-USP11. Cell lysates were immunoprecipitated with anti-FLAG (DNMT3A) antibody, following the detection of DNMT3A interacted USP11 with anti-myc, CBX6 with anti-CBX6, RING1B with anti-RING1B antibodies for western blotting. D, DNMT3A DNA Methyltransferase activity analysis using the HOXA5-Snrpn-BFP methylation. Knocking in (KI) Snrpn promoters to the HOXA5 locus in reverse orientation in 293T cells linked to BFP and bovine polyA signals. E, F, Methyltransferase activity assay. The figure depicts blue fluorescence intensity in the FLAG-DNMT3A and/or HA-USP11 transduced cells as measured by flow cytometry on day 10. The graphs depicted the BFP positive cells. The negative control frequency of BFP+ cells was set as 1. One-way ANOVA with Tukey’s multiple comparisons was completed. *p < 0.05, **p < 0.01 *** p < 0.001 ****p < 0.0001.