Primate-specific isoform of Nedd4-1 regulates substrate binding via Ser/Thr phosphorylation and 14-3-3 binding

Nedd4 (Nedd4-1) is an E3 ubiquitin ligase involved in crucial biological processes such as growth factor receptor signaling. While canonical Nedd4-1 comprises a C2-WW(4)-HECT domain architecture, alternative splicing produces non-canonical isoforms that are poorly characterized. Here we characterized Nedd4-1(NE), a primate-specific isoform of Nedd4-1 that contains a large N-terminal Extension (NE) that replaces most of the C2 domain. We show that Nedd4-1(NE) mRNA is ubiquitously expressed in human tissues and cell lines. Moreover, we found that Nedd4-1(NE) is more active than the canonical Nedd4-1 isoform, likely due to the absence of a C2 domain-mediated autoinhibitory mechanism. Additionally, we identified two Thr/Ser phosphoresidues in the NE region that act as binding sites for 14-3-3 proteins, and show that phosphorylation on these sites reduces substrate binding. Finally, we show that the NE region can act as a binding site for the RPB2 subunit of RNA polymerase II, a unique substrate of Nedd4-1(NE) but not the canonical Nedd4-1. Taken together, our results demonstrate that alternative splicing of the ubiquitin ligase Nedd4-1 can produce isoforms that differ in their catalytic activity, binding partners and substrates, and mechanisms of regulation.

characterized, including at least three isoforms that contain a common 516-amino acid NE region which replaces most of the C2 domain (Fig. 1A, Fig. S1A).The NE region does not contain any known domains or predicted structures and is not found in other Nedd4 family members.While Nedd4-1, especially the HECT domain, is well conserved throughout evolution, the NE region is only highly conserved amongst higher order mammals such as primates (Fig. S1B).Given the presence of this NE region, we have named the longest of these isoforms Nedd4-1(NE).
To confirm the expression of these alternatively spliced isoforms of human Nedd4-1, we used a cDNA panel derived from mRNA extracted from various human tissues.We confirmed mRNA expression of Nedd4-1(NE), but not Nedd4-1(NE) [Δ517-588] or Nedd4-1(NE) [Δ589-604] isoforms, across all tissues tested (Fig. 1B).Moreover, we found the same pattern of isoform expression across various human cell lines and colonic organoids (Fig. 1C,D).Together, these data suggest that alternative splicing produces ubiquitous expression of the Nedd4-1(NE) isoform in humans.Although Nedd4-1(NE) mRNA is widely expressed in human cells, its expression levels relative to the canonical Nedd4-1 are significantly lower (Fig. S2), and accordingly, we were unable to detect endogenous protein expression of this isoform using antibodies specific to the NE region that we generated.This could be due to the low abundance of the Nedd4-1(NE) protein in these cells, at least under steady-state conditions, low sensitivity of the antibodies, or both.

Nedd4-1(NE) exhibits enhanced catalytic activity compared to Nedd4-1
To better understand the role of the NE region in Nedd4-1(NE) function, we first compared the catalytic (ubiquitin ligase) activity of Nedd4-1(NE) with the canonical isoform of Nedd4-1.To do so, we transfected HEK293T cells with N-terminally Flag-tagged Nedd4-1 or Nedd4-1(NE) and assessed their ability to catalyze the transfer of ubiquitin onto themselves (i.e., autoubiquitination).Lysates from transfected cells were boiled in SDS to remove Nedd4-1-or Nedd4-1(NE)-interacting proteins, and autoubiquitination was determined by immunoprecipitating (IP) with anti-Flag antibodies and immunoblotting with anti-ubiquitin (Ub) antibodies.Interestingly, Nedd4-1(NE) was autoubiquitinated at higher levels than Nedd4-1 (Fig. 2A).To study catalytic activity towards a cellular substrate, we analyzed the ability of Nedd4-1 and Nedd4-1(NE) to catalyze the transfer of ubiquitin onto the transcription factor YY1, a known substrate of Nedd4-1, which is ubiquitinated by Nedd4-1 without affecting its (YY1) stability 19 .Consistent with the autoubiquitination data, we found that Nedd4-1(NE) was able to ubiquitinate YY1 to a greater extent than Nedd4-1 (Fig. 2B).Together, these data indicate that Nedd4-1(NE) displays enhanced ubiquitin ligase activity compared to canonical Nedd4-1.Next, we sought to understand the basis behind the increased catalytic activity of Nedd4-1(NE).In the case of Nedd4-1, an autoinhibitory mechanism exists whereby the C2 domain binds the HECT domain and inhibits its catalytic activity [9][10][11] .Given that only a small segment of the C2 domain is present in Nedd4-1(NE), it is uncertain whether this autoinhibitory mechanism is preserved.To address this, we utilized a HECT domain mutation (Y604A) that perturbs C2 domain binding, thus blocking the autoinhibitory mechanism and consequently favoring the active state of the HECT domain 20 .Our results show that unlike Nedd4-1, in Nedd4-1(NE), the analogous mutation (Y1023A) does not enhance catalytic activity, presumably due to the lack of a preserved autoinhibitory mechanism (Fig. 2C).Therefore, unlike the C2 domain of Nedd4-1, the NE region of Nedd4-1(NE) does not negatively regulate the HECT domain, a factor which likely contributes to the enhanced catalytic activity of Nedd4-1(NE) compared to Nedd4-1.
To explore the possibility that the interactome of Nedd4-1(NE) is altered by 14-3-3 binding, we performed quantitative IP-MS, comparing the interactomes of the WT and the 2A NE regions of Nedd4-1(NE) in HEK293T cells (Fig. 4B,C).Consistent with our earlier experiments described in Fig. 3 and Table S1, we identified six members of the 14-3-3 family as significant interactors of the WT NE region (Fig. 4B).Of all the interacting proteins that were significantly different between the WT and 2A samples, these six proteins were the only ones to exhibit reduced binding to the 2A mutant (Fig. 4C and Table S3).On the other hand, 50 proteins were identified with significantly enhanced binding to the 2A mutant than the WT NE region (Fig. 4C and Table S3), suggesting that these mutations stabilize or enhance various protein-protein interactions mediated by the NE region.To validate this hypothesis, we focused on the RPB2 subunit of RNA polymerase II, the most significant hit in the aforementioned cohort.Indeed, our results show that RPB2 bound Nedd4-1(NE), but not Nedd4-1 (Fig. 4D,E).In accord, we show that RPB2 is a substrate of Nedd4-1(NE), but not Nedd4-1 (Fig. 4F).Consistent with the IP-MS data, we found that RPB2 interacts with the 2A mutant of the NE region to a greater extent than the WT NE region (Fig. 4G,H).Moreover, we found that KPC1, an independently identified binding partner and substrate of Nedd4-1(NE), also exhibits increased binding to the 2A mutant compared to the WT NE region (Fig. 4I).Taken together, these data demonstrate that mutations in Thr 16 and Ser 458 have a global effect on stabilizing or enhancing substrate binding to the NE region of Nedd4-1(NE) by preventing either phosphorylation or 14-3-3 binding.

Discussion
Our studies here analyzed Nedd4-1(NE), a poorly characterized isoform of the ubiquitin ligase Nedd4-1, which is widely distributed in tissues and cells (although expressed at lower levels than the canonical Nedd4-1 under steady-state conditions).Specifically, our work distinguishes Nedd4-1(NE) from the canonical Nedd4-1 isoform with regards to its catalytic activity, substrates, and mechanisms of regulation.Our work represents the first direct comparison of alternatively spliced isoforms of Nedd4-1.
While several modes of regulation of catalytic activity of Nedd4-1 have been described (e.g. 28,29), the most pertinent to our studies here are those that implicate the C2 domain, which is largely absent in Nedd4-1(NE).Residues absent in Nedd4-1(NE) comprise calcium-binding sites, HECT domain-binding sites, and Src kinase phosphorylation sites, all of which are involved in regulating autoinhibition of enzymatic activity of Nedd4-1 10,30 .Accordingly, we found that Nedd4-1(NE) is more active than Nedd4-1, likely due to the absence of the autoinhibitory mechanism brought about by C2:HECT domains interaction.
Our quantitative IP-MS data not only identified novel binding partners of Nedd4-1(NE), which are not dependent on binding to its WW domains, but also identified binding partners and substrates that preferentially bind to the unphosphorylated NE region.Our results indicate that 14-3-3 binding exerts a more widespread effect on Nedd4-1(NE) than Nedd4-2 by regulating universal substrate binding, including binding to seemingly unrelated substrates such as RPB2 and KPC1.While we found that the mutation of 14-3-3 binding phosphosites in the NE region enhanced substrate binding, we do not know whether that is directly caused by the absence of 14-3-3 binding, or by the loss of phosphorylation at these residues.If the former is true, phosphorylation may recruit 14-3-3, thus blocking substrate binding, while if the latter is true, 14-3-3 binding may simply protect the phosphorylation of these residues, thereby limiting substrate binding.
Our above quantitative IP-MS experiments identified the RPB2 subunit of RNA polymerase II as an isoformspecific binding partner and substrate of Nedd4-1(NE).Previous reports had identified the canonical Nedd4-1, as well as its yeast ortholog Rsp5, as the ubiquitin ligase responsible for ubiquitination of RNA polymerase II 36,37 .However, these reports focused on the RPB1 subunit of the complex, while our proteomic screen identified the RPB2 subunit as a binding partner of Nedd4-1(NE).Thus, we propose that the different subunits of RNA polymerase II are ubiquitinated by different isoforms of Nedd4-1, with RPB1 targeted by the canonical Nedd4-1 through a WW domain-mediated interaction, and RPB2 targeted by Nedd4-1(NE) through a NE region-mediated interaction.
In summary, our study characterized the ubiquitin ligase Nedd4-1(NE), differentiating it from the canonical Nedd4-1 isoform with regards to its catalytic activity and binding partners.Moreover, we identified an isoformspecific mechanism by which substrate binding is regulated.

Cell culture and transfection
HEK293T (ATCC #CRL-1573) and 293 T-REx (Thermo Fisher Scientific #R78007) cells were cultured in Dulbecco's Modified Eagle Medium (DMEM), supplemented with 10% fetal bovine serum and antibiotic-antimycotic. 293 T-REx stable cell lines were generated by recombinase-mediated DNA recombination using the Flp-In system (Thermo Fisher Scientific) and maintained in DMEM containing 100 µg/ml hygromycin B. Recombinant

Figure 4 .
Figure 4. 14-3-3 binding site mutations in the NE region promote substrate binding.(A,F) HEK293T cells were transfected with the indicated constructs, with the Nedd4-1(NE) 2A mutant representing the T16A/S458A double mutant.Cell lysates were subjected to Flag IP following addition of 1% SDS and boiling to dissociate protein complexes.Ubiquitination was detected by immunoblotting for ubiquitin (Ub).(B,C) HEK293T cells were transfected with Flag-tagged NE region (WT or 2A mutant) or an empty vector control.Cell lysates were subjected to Flag IP and interacting proteins were identified by mass spectrometry.In (B), data points above the dashed line (p > 0.05) represent significant interactors of the WT NE region, with 14-3-3 isoforms highlighted in red.In (C), data points highlighted in red represent proteins that are significantly decreased in the 2A compared to the WT data set, while data points highlighted in green represent proteins that are significantly increased in the 2A compared to the WT data set.(D,E,G-I) HEK293T cells were transfected with the indicated constructs.Cell lysates were subjected to Flag IP and the indicated proteins were detected by immunoblotting.All data are means ± S.D. of 3 independent experiments.Statistical significance was determined using (A,F,G) oneway ANOVA, or (D,E,H,I) Student's t test (**p < 0.01; ***p < 0.001; ****p < 0.0001; N.S. not significant).