Fig 2: In vitro ubiquitination on protein microarrays identified WWP2 substrates.A, example of comparative zoomed-in images of representative WWP2 ubiquitination hits on the microarray chip. E3 ligase NEDD4L showed fluorescent signals on both control and active WWP2Y369E–treated protein microarrays. MED4 and OPTN only appeared on the active WWP2Y369E–treated protein microarrays. B, identified WWP2 E3 ligase hits using HuProt protein microarrays using linker phospho-mimetic form of WWP2. The hits are ranked by ubiquitination signal intensity/Z value. C, GO-cellular component enrichment analysis of WWP2 hits. The correlated components were ranked based on calculated enrichment scores using the GOrilla tool. D, GO-biological process enrichment analysis of WWP2 hits. The GO term (Log10P) was used for generating and visualizing the hit map using the REVIGO tool. Note: The axes in the plot have no intrinsic meaning. OPTN, optineurin.

Fig 3: Enzymatic analysis of NDP52 and OPTN as WWP2 substrates.A, in vitro ubiquitination assay of WWP2 with NDP52. Different forms of WWP2 E3 ligase were used for the ubiquitination of NDP52. The hyper autoubiquitination activity of WWP2?linker was highlighted by the red arrow. B, ubiquitination assay of NDP52 in the presence of allosteric activators. WT WWP2 with either Ndfip1 (activator with the proline-rich motifs that interacts with WW domains and activates WWP2) or UbV (allosteric activator that binds to the ubiquitin exosite in the catalytic HECT domain) showed an increased extent of ubiquitination on NDP52. C, in vitro ubiquitination assay of WWP2 with OPTN. Different forms of WWP2 E3 ligase (wt; ?linker; Y369E; Y392E) were used for the ubiquitination of OPTN. The hyper autoubiquitination activity of WWP2?linker was highlighted by the red arrow. D, ubiquitination assay of OPTN in the presence of allosteric activators. WT WWP2 with either Ndfip1 or UbV showed an increased extent of ubiquitination on OPTN. Gels are stained with Colloidal Blue. The depletion of the unmodified substrate proteins was determined by densitometry analysis and shown as a percentage compared to the samples at time zero. ?linker, hyperactive form with 2,3-linker removed; Y369E or Y392E, the active form of WWP2 with linker phosphorylation mimetics; OPTN, optineurin; UbV, ubiquitin variant.

Fig 4: Identification of WWP1 as the prime E3 ligase which mono-ubiquitinates AMOTL2 at Lys-347/408.(A) 293T cells were transfected with the indicated DNAs, then subjected to in vivo ubiquitination assays. Shown are the E3 ligases which were successfully expressed, and demonstrated a positive ubiquitination effect from the initial 160 E3 ligases from the ORF library (except for WWP2—a closely related family member of WWP1—which did not demonstrate a positive ubiquitination effect, and was included to serve as a negative control). (B) 293T cells were transfected with the indicated DNAs, then subjected to in vivo ubiquitination assays. (A) Shown are the E3 ligases which exhibited the greatest ubiquitination effect from (A). AMOTL2 2KR mutant was also included in this set, to assess whether the E3 ligase candidates cannot ubiquitinate this 2KR mutant. As can be seen, only WWP1 met this criteria.

Fig 5: WWP2 is a bona fide SOX2 E3 ubiquitin ligase in GSCs.A GSCs transduced with control or RNAi targeting CDH1, CUL4A, UBR5, and WWP2 were harvested 6 days later and subjected to immunoblotting. Data were representative of three independent experiments. B GSCs transduced with WWP2 RNAi or control were harvested 6 days later and subjected to immunoblotting. Data were representative of three independent experiments. C Cells isolated from patient tumors were subjected to scRNA-seq. Malignant cells were identified by expression-based inference of copy number variation as in Fig. 1I. Cells are colored by expression level for WWP2. D GSC lysates were subjected to anti-SOX2 immunoprecipitation (IP). Input lysates and IP samples were subjected to immunoblotting. Data were representative of three independent experiments. E GSCs transduced with lentiviruses to stably express 6X-Histidine epitope-tagged SOX2 and FLAG epitope-tagged ubiquitin were transduced with WWP2 RNAi or control. Lysates were subjected to IP under denaturing conditions using nickel-NTA beads. IP and input samples were subjected to immunoblotting. Data were representative of three independent experiments. F GSCs transduced with WWP2 RNAi or control were subjected to ELDA as in Fig. 3A. Data represent mean ± SEM (n = 3, unpaired t-test, **P < 0.01). See Source Data File for exact P values and statistical parameters.