Multiubiquitination of TRPV4 reduces channel activity independent of surface localization

Ubiquitin (Ub)-mediated regulation of plasmalemmal ion channel activity canonically occurs via stimulation of endocytosis. Whether ubiquitination can modulate channel activity by alternative mechanisms remains unknown. Here, we show that the transient receptor potential vanilloid 4 (TRPV4) cation channel is multiubiquitinated within its cytosolic N-terminal and C-terminal intrinsically disordered regions (IDRs). Mutagenizing select lysine residues to block ubiquitination of the N-terminal but not C-terminal IDR resulted in a marked elevation of TRPV4-mediated intracellular calcium influx, without increasing cell surface expression levels. Conversely, enhancing TRPV4 ubiquitination via expression of an E3 Ub ligase reduced TRPV4 channel activity but did not decrease plasma membrane abundance. These results demonstrate Ub-dependent regulation of TRPV4 channel function independent of effects on plasma membrane localization. Consistent with ubiquitination playing a key negative modulatory role of the channel, gain-of-function neuropathy-causing mutations in the TRPV4 gene led to reduced channel ubiquitination in both cellular and Drosophila models of TRPV4 neuropathy, whereas increasing mutant TRPV4 ubiquitination partially suppressed channel overactivity. Together, these data reveal a novel mechanism via which ubiquitination of an intracellular flexible IDR domain modulates ion channel function independently of endocytic trafficking and identify a contributory role for this pathway in the dysregulation of TRPV4 channel activity by neuropathy-causing mutations.

Ubiquitin (Ub)-mediated regulation of plasmalemmal ion channel activity canonically occurs via stimulation of endocytosis. Whether ubiquitination can modulate channel activity by alternative mechanisms remains unknown. Here, we show that the transient receptor potential vanilloid 4 (TRPV4) cation channel is multiubiquitinated within its cytosolic N-terminal and C-terminal intrinsically disordered regions (IDRs). Mutagenizing select lysine residues to block ubiquitination of the Nterminal but not C-terminal IDR resulted in a marked elevation of TRPV4-mediated intracellular calcium influx, without increasing cell surface expression levels. Conversely, enhancing TRPV4 ubiquitination via expression of an E3 Ub ligase reduced TRPV4 channel activity but did not decrease plasma membrane abundance. These results demonstrate Ubdependent regulation of TRPV4 channel function independent of effects on plasma membrane localization. Consistent with ubiquitination playing a key negative modulatory role of the channel, gain-of-function neuropathy-causing mutations in the TRPV4 gene led to reduced channel ubiquitination in both cellular and Drosophila models of TRPV4 neuropathy, whereas increasing mutant TRPV4 ubiquitination partially suppressed channel overactivity. Together, these data reveal a novel mechanism via which ubiquitination of an intracellular flexible IDR domain modulates ion channel function independently of endocytic trafficking and identify a contributory role for this pathway in the dysregulation of TRPV4 channel activity by neuropathy-causing mutations.
Plasma membrane-localized ion channels enable cells to rapidly respond to intracellular and extracellular stimuli but require tight regulatory control to maintain cellular homeostasis. Indeed, multiple disorders including neurological diseases are caused by genetic mutations that alter ion channel activity and cellular ion influx (1)(2)(3)(4). Dynamic regulatory control of plasmalemmal ion channels is ensured by modulation of channel gating as well as by changes in cell surface expression.
Ubiquitination is a rapidly reversible post-translational modification (PTM) shown to regulate the activity of multiple plasma membrane-localized ion channels (4). The covalent attachment of an 8 kDa ubiquitin (Ub) peptide to a lysine residue is catalyzed by the sequential action of a Ub-activating enzyme (E1), a Ub-conjugating enzyme (E2), and a Ub ligase (E3), the last of which provides specificity for the substrate protein. For a given substrate, Ub may be conjugated to a single lysine (monoubiquitination), multiple lysines (multiubiquitination), or be ubiquitinated itself to form Ub chains (polyubiquitination). It is well established that monoubiquitination/multiubiquitination of surface-expressed ion channels can stimulate removal of a channel from the plasma membrane, thereby negatively regulating channel activity (5)(6)(7)(8). Conversely, disrupting ubiquitination of such ion channels typically results in the accumulation of the channel at the plasma membrane, leading to a gain-of-channel function (9)(10)(11)(12)(13). In cases in which ion channels are internalized after ubiquitination, they may then be shuttled to proteolytic pathways (7,14) or trafficked to recycling compartments that are available for membrane reinsertion (15)(16)(17). In addition to regulating membrane protein trafficking and degradation, monoubiquitination or multiubiquitination events are increasingly appreciated to regulate protein function in other ways, including by altering protein-protein and protein-lipid interactions (18)(19)(20). Individual proteins may also undergo multiple distinct Ub modifications that together provide a diverse Ub architecture enabling complex and dynamic regulation of cellular processes (21)(22)(23). NEDD4 and ITCH, members of the NEDD4 family of E3 Ub ligases, acting alone or with adaptor/accessory proteins, have been shown to ubiquitinate several plasmalemmal proteins, including ion channels, resulting in their internalization (24). It has not been determined whether monoubiquitination/multiubiquitination of an ion channel can affect channel activity independent of internalization.
Transient receptor potential vanilloid 4 (TRPV4), a member of the vanilloid subfamily of TRP channels, is a broadly expressed calcium-permeable ion channel with diverse functions (25). TRPV4 functions as a homotetramer at the plasma membrane where it can be activated by mechanical stimulation, hypo-osmotic stress, arachidonic acid, and synthetic compounds (25). The channel is composed of six transmembrane domains flanked by intracellular N-terminal and C-terminal domains, the structures of which have been partly characterized by cryogenic electron microscopy (26), including the large cytosolic N-terminal ankyrin repeat domain (ARD). Dominant gain-of-function mutations of exposed arginine residues on one face of the ARD cause forms of peripheral neuropathy, including Charcot-Marie-Tooth disease type 2C and distal spinal muscular atrophy (27)(28)(29)(30). Though our understanding of how these missense mutations cause a gain of function is incomplete, they are not thought to increase channel surface expression (29,31) but may rather disrupt specific regulatory protein-protein and/or protein-lipid interactions (32). In contrast to the ARD, the intrinsically disordered regions (IDRs) of the N termi-and C termini have not been amenable to structural analysis (33). The function of IDRs, which have no stable ordered structure, is not fully understood, although they are increasingly recognized as important components of the cell signaling machinery as they contain multiple interaction motifs, dynamically bind diverse targets, and act as central hubs in signaling networks (34). The TRPV4 N-terminal IDR is located upstream of the ARD and contains a PI(4,5)P 2 -(hereafter PIP 2 ) binding domain (PBD) and a proline-rich region (PRR) (33,35,36). Intermolecular interactions at these domains modulate TRPV4 channel activity without altering TRPV4 surface density (36,37). IDRs are frequently post-translationally modified (33), but whether ubiquitination within the IDR affects TRPV4 function is unknown.
Here, we show that TRPV4 is multiubiquitinated within the intracellular flexible N-terminal IDR and that blocking ubiquitination at key residues within this domain results in increased channel activity without increasing TRPV4 surface abundance. Conversely, enhancing ubiquitination at these residues reduces channel activity without reducing plasma membrane abundance. We further demonstrate that neuropathy-causing mutations in TRPV4 reduce channel ubiquitination correlating with increased channel activity, but enhancing ubiquitination of neuropathy mutants dampens their responses to agonist stimulation. These data reveal a novel mechanism by which multiubiquitination of a disordered intracellular protein domain can regulate ion channel activity without altering plasma membrane abundance and provide insight into mechanisms contributing to TRPV4 gain-offunction mutations.

TRPV4 is monoubiquitinated/multiubiquitinated in multiple cell types
To determine whether endogenous TRPV4 protein is ubiquitinated in cells and whole tissue, we assessed an immortalized rat sensory neuron cell line (50B11 cells), the Madin-Darby canine kidney (MDCK) epithelial cell line, and mouse choroid plexus tissue; all of which express robust levels of TRPV4 (32,38,39). To immunopurify endogenous TRPV4, we used an anti-TRPV4 antibody that detects TRPV4 protein in cultured primary choroid plexus epithelial cells (40), kidney, and choroid plexus derived from WT, but not Trpv4 null mice (40)(41)(42) (Fig. S1, A and B). Using this antibody, we achieved robust TRPV4 immunoprecipitation (IP) with little background compared with whole cell lysate inputs (Fig. S1C). TRPV4 banding patterns can vary across tissues and cell lines likely because of variations in PTM including glycosylation (43). Our data showed a single TRPV4 band in 50B11 cells (Fig. 1A) and a doublet in MDCK cells and choroid plexus tissue (Fig. 1, B and C) consistent with previous studies (39,(44)(45)(46)(47).
To facilitate detection of ubiquitinated TRPV4, hemagglutinin (HA)-tagged Ub (HA-Ub) was expressed in 50B11 cells using lentiviral transduction, MDCK cells via transient transfection, and in C57BL/6J mice via gene targeting (48). Western blot (WB) analyses of TRPV4 immunoprecipitated from the two cell lines showed a single mass-shifted Ub band in 50B11 cells (Fig. 1A, arrow) and three mass-shifted bands in MDCK cells (Fig. 1B, arrows) following probing with anti-HA antibody. WB analyses of TRPV4 immunoprecipitated from choroid plexus tissues of HA-Ub-expressing mice, but not WT mice, demonstrated a mass-shifted Ub band (Fig. 1C). Together, these data indicate that endogenous TRPV4 protein is ubiquitinated in multiple cell types, both in vitro and in vivo.
To explore TRPV4 ubiquitination in depth, we established a system that allowed us to readily scale up cell-based Ub detection assays for downstream proteomics analyses. Cterminally FLAG-tagged TRPV4 (TRPV4-FLAG) was immunoprecipitated from transfected human embryonic kidney 293T (HEK293T) cells, separated on SDS-PAGE gels, and blotted for endogenous Ub. We detected multiple discrete Ub bands in the 100 to 140 kDa range, consistent with either polyubiquitination or multiubiquitination of TRPV4 (Fig. 1D, arrows). To assess the type of Ub linkage conjugated to TRPV4, we cotransfected cells with TRPV4-FLAG and either HA-Ub or a form of HA-Ub that cannot form poly-Ub chains (K0-HA-Ub) (49) (Fig. 1E). Comparable banding patterns of ubiquitinated TRPV4 were observed in cells expressing either HA-Ub or K0-HA-Ub (Fig. 1E, arrows), indicating that TRPV4 is predominantly multiubiquitinated under these conditions, consistent with the findings of a previous study (50). To determine whether the pattern of ubiquitination of TRPV4 is influenced by channel activity, transfected cells were cultured in the presence of the TRPV4-specific antagonist HC067047 (hereafter HC067) (51). The TRPV4 Ub banding pattern was not altered by the presence of HC067, indicating that basal TRPV4 channel activity does not play a significant role in the type of channel ubiquitination. However, the amount of ubiquitinated TRPV4 increased in the presence of antagonist suggesting a link between TRPV4 channel activity and its level of ubiquitination (Fig. 1F). To further confirm that TRPV4 was predominantly multiubiquitinated, WB analyses were performed on immunoprecipitated TRPV4 using antibodies specific for either K48-linked or K63-linked poly-Ub chains. Although diffuse poly-Ub signals were observed in the unprecipitated lysates (inputs), minimal signal was seen in immunoprecipitated TRPV4 eluates (Fig. 1G). Together, these data suggest that TRPV4 is predominantly multiubiquitinated.
TRPV4 is ubiquitinated at lysine residues in its cytoplasmic intrinsically disordered regions Mass spectrometry (MS) analysis of immunoprecipitated TRPV4 protein was performed to identify ubiquitinated lysine residues. We enriched for ubiquitinated TRPV4 by using a double IP approach ( Fig. 2A) utilizing HEK293T cells transfected with both TRPV4-FLAG and HA-Ub and sequential IP with anti-FLAG and anti-HA antibodies. Each TRPV4 monomer contains 45 lysines, most of which reside in the cytosolic intracellular domains: 10 in the N-terminal IDR, 13 in the ARD, and 5 in the C-terminal IDR (hereafter CTD) (Fig. 2B, black vertical lines). Seventeen lysines are distributed throughout the transmembrane regions and the TRP box (Fig. 2B, Figure 1. TRPV4 is monoubiquitinated/multiubiquitinated. A and B, ubiquitination of endogenous TRPV4 in HA-ubiquitin (Ub)-expressing 50B11 (A) and MDCK (B) cells, as detected by immunoprecipitation (IP) with anti-TRPV4 antibody and Western blot analysis with anti-HA antibody. C, Western blot analysis using anti-HA antibody of endogenous TRPV4 immunoprecipitated from the choroid plexus of HA-Ub mice; parallel IP from WT mice demonstrates the specificity of the anti-HA antibody. D, cell-based ubiquitination assay performed using HEK293T cells transiently transfected with TRPV4-FLAG. TRPV4 was immunoprecipitated from whole cell lysates with anti-FLAG antibody and Western blot analysis performed with anti-Ub antibody. E, cell-based ubiquitination assay utilizing HEK293T cells cotransfected with TRPV4-FLAG and either WT-HA-Ub or K0-HA-Ub, in the presence or the absence of the TRPV4-specific antagonist HC067 (0.5 μM). TRPV4 was immunoprecipitated from whole cell lysates with anti-FLAG antibody, and Western blot analysis was performed with anti-HA antibody. F, quantification of relative TRPV4 ubiquitination in the presence or the absence of HC067 from the experiment shown in E. n = 3 independent transfections. Data are presented as means ± SD. *p = 0.0422 (unpaired t test). G, cell-based ubiquitination assay performed using HEK293T cells transiently transfected with TRPV4-FLAG. TRPV4 was immunoprecipitated from whole cell lysates with anti-FLAG antibody, and Western blot analysis was performed with anti-Ub, anti-K48-linkage-specific polyubiquitin, or anti-K63-linkage-specific polyubiquitin antibodies. Arrows indicate ubiquitinated TRPV4 and arrowheads indicate polyubiquitinated TRPV4. HA, hemagglutinin; HC067, TRPV4-specific antagonist HC067047; HEK293T, human embryonic kidney 293T cell line; MDCK, Madin-Darby canine kidney; TRPV4, transient receptor potential vanilloid 4; TRPV4-FLAG, FLAG-tagged TRPV4.
TRPV4 ubiquitination reduces plasmalemmal TRPV4 activity ubiquitinated lysines were identified in the transmembrane region or extracellular domains. Interestingly, Lys 192 and 197 form part of an ATP-binding pocket within the TRPV4 ARD (52,53) (Fig. S1D). Ubiquitinated lysine residues were more frequently and consistently observed in the unstructured Nterminal and C-terminal IDRs than in the ARD (Table 1 and  Table S1). The ubiquitinated lysine residues in the N-terminal IDR are distributed throughout the domain including sites within or near the two known functional domains, the PBD and PRR (Fig. 2D).

Blocking ubiquitination in the N-terminal IDR increases TRPV4 channel activity
To determine the functional effects of ubiquitination of specific TRPV4 protein domains, we genetically blocked ubiquitination by mutating select lysine residues to chargepreserving arginines (K > R) by site-directed mutagenesis. We generated four TRPV4 expression constructs: (i) All-K-IDR, in which all 10 lysine residues in the N-terminal IDR were mutated to arginine (listed in legend to Fig. S2A); (ii) 5K-IDR, in which five lysine residues in the N-terminal IDR selected based on their identification by MS analyses and proximity to the PBD and PRR were mutated (Lys 77, 101, 130, 136, and 147); (iii) a more limited K130, 136R mutant; and (iv) 5K-CTD, in which all five lysine residues within the CTD were targeted for mutagenesis (listed in the legend to Fig. S2A). Following expression in HEK293T cells, ubiquitination was diminished in the 5K-IDR and All-K-IDR mutants relative to the WT channel, although as expected, ubiquitination was still observed at the remaining lysine residues ( Fig. S2A and Table 2). Culturing cells expressing the TRPV4 lysine mutants TRPV4 ubiquitination reduces plasmalemmal TRPV4 activity in the presence of the HC067 antagonist restored a normal level of ubiquitination of the 5K-IDR mutant but not of the All-K-IDR mutant (Fig. S2B).
To assess the subcellular distribution and channel activity of the TRPV4 K>R mutants, we generated constructs containing a V5 tag in the first extracellular loop and an enhanced GFP (EGFP) tag at the C terminus (TRPV4-V5-EGFP) (Fig. 3A). Surface TRPV4 was visualized by incubating nonpermeabilized cells with an anti-V5 antibody, and total TRPV4 was visualized using the EGFP tag. Fluorescence intensity was measured using confocal microscopy, and the ratio of surface to total TRPV4 was determined. Experiments were performed in the presence or the absence of the antagonist HC067, both to assess effects of channel activity on cell surface expression of the mutants and to determine whether observed changes in intracellular calcium resulted from TRPV4 activation. In the absence of HC067, the 5K-IDR mutant showed reduced surface expression compared with the WT channel (Fig. 3, B and C). Strikingly, however, 5K-IDR-expressing cells displayed elevated intracellular calcium levels, as measured by Fura-2 ratiometric calcium imaging (Fig. S2C), suggesting that this mutant exhibits increased channel activity. Culturing of 5K-IDR-transfected cells in the presence of HC067 rescued surface localization of the mutant channel and reduced intracellular calcium levels (Figs. 3C and S2C). Together, these data suggest that the 5K-IDR mutant exhibits increased channel activity at the cell surface. The more limited IDR mutant targeting lysines 130 and 136 (K130, 136R) behaved similarly to WT in its surface localization and channel activity in the absence or the presence of HC067 (Fig. S2, C-E). 5K-CTD-expressing cells also exhibited levels of TRPV4 surface expression and intracellular calcium comparable to cells expressing the WT channel (Figs. 3C and S2C).
To further explore the activity of the 5K-IDR TRPV4 mutant, we used an antagonist withdrawal paradigm in which transfected HEK293T cells were cultured in HC067 for 24 h, and calcium imaging was performed after removing the antagonist. While cells expressing WT, 5K-CTD, or K130, 136R TRPV4 proteins showed minimal changes in intracellular calcium levels over 60 min after HC067 removal (Figs. 3, D and E and S2F), cells transfected with 5K-IDR TRPV4 showed marked increases in intracellular calcium levels over time a Also identified in a screen of the Ub proteome of mouse kidney (40).  TRPV4 ubiquitination reduces plasmalemmal TRPV4 activity ( Fig. 3, E and F). Performing the antagonist withdrawal experiment in calcium-free media did not result in increased intracellular calcium levels of 5K-IDR TRPV4-expressing cells until extracellular calcium was provided (Fig. 3G). These data suggest that calcium flux through the TRPV4 channel at the cell surface is required for intracellular calcium accumulation. We also assessed the activity of the 5K-IDR TRPV4 mutant in cells treated with the TRPV4-specific agonist GSK1016790A (GSK101) following HC067 washout (Fig. 3H). 5K-IDRexpressing cells showed a 300% greater response to 10 nM GSK101 compared with WT-CTD-expressing cells or 5K-CTD-expressing cells (Fig. 3I). Examination of the surface abundance of the 5K-IDR mutant during the period of increased calcium influx following antagonist withdrawal ( Fig. 3E) revealed little change in 5K-IDR TRPV4 surface abundance and certainly not an accumulation of the 5K-IDR mutant at the cell surface relative to WT (Fig. 3, J and K).
Together, these data demonstrate that blocking ubiquitination at specific lysines in the N-terminal IDR, but not the CTD, results in increased TRPV4 channel activity with little change in surface abundance.

The TRPV4 IDR can be ubiquitinated by the NEDD4 family of E3 Ub ligases
Given our results demonstrating that blocking ubiquitination of the TRPV4 N-terminal IDR increased channel activity, we next sought to determine the functional consequences of enhancing ubiquitination. The NEDD4 family of E3 Ub ligases is well studied in the context of ion channel ubiquitination (24), and NEDD4 in particular is known to function at the plasma membrane because of its C2 lipid-binding domain (54,55). In HEK293T cells, we observed that TRPV4 coimmunoprecipitated with two NEDD4 family members, NEDD4 and ITCH, as well as their catalytically inactive mutants (NEDD4-CI and ITCH-CI) (56) (Fig. 4A). These data confirm a previous report demonstrating that ITCH interacts with TRPV4 (50). These interactions were specific for TRPV4, as neither NEDD4 nor ITCH modify the ubiquitination of TRPV1 (Fig. S3A). As expected, only coexpression of catalytically active NEDD4 or ITCH resulted in increased ubiquitination of TRPV4, with a 110% increase in TRPV4 ubiquitination observed with NEDD4 compared with ITCH ( Fig. 4, B and C). Neither NEDD4 nor ITCH coexpression reduced total TRPV4 levels at 24 h posttransfection, suggesting that ubiquitination is not a degradative signal for TRPV4 (Fig. 4D).
To determine which specific TRPV4 lysine residues become ubiquitinated in the presence of NEDD4, Myc-tagged NEDD4 was coexpressed with TRPV4-FLAG and HA-Ub in HEK293T cells, followed by double IP and MS analysis as outlined in Figure 2A. In the presence of only endogenous Ub ligases, five lysine residues were observed to be ubiquitinated in the Nterminal IDR and three in the CTD. In the presence of NEDD4, four of the five lysine residues in the N-terminal IDR and two of the three lysine residues in the CTD were ubiquitinated (Tables 1 and 3). However, one additional lysine residue (Lys 147) was found to be ubiquitinated in the Nterminal IDR solely in the presence of NEDD4 (Table 3).
NEDD4 and ITCH sometimes act directly on protein targets but in other cases require coadaptors to ubiquitinate a substrate (57)(58)(59)(60). To test whether their ubiquitination of TRPV4 is direct, we performed cell-free Ub assays combining recombinant NEDD4 or ITCH with either recombinant TRPV4 full N terminus (IDR + ARD) or ARD fragments (Fig. S3B). Autoubiquitination of each E3 ligase provided evidence of enzymatic activity (Fig. S3, C-E, red arrowheads). NEDD4 did not induce TRPV4 ubiquitination in this system, suggesting requirement of a coadaptor protein (Fig. S3C), whereas ITCH induced ubiquitination of the full TRPV4 N terminus but not of the ARD fragment or of two additional N-terminal domain fragments: IDR and ARD + PRR (Fig. S3, D and E). MS analysis indicated that four of the IDR residues identified from the whole cell analysis (Fig. 2) can be ubiquitinated by ITCH: Lys 77, 101, 130, and 136. The additional IDR lysine residue identified as a site of ubiquitination in cells expressing NEDD4 (Lys 147) was also ubiquitinated in the presence of ITCH in this cell-free assay (Fig. S3B, green circles, Table S2). Together, these data demonstrate that NEDD4 and ITCH primarily ubiquitinate TRPV4 at the same N-terminal lysine residues, and that these residues exhibit extensive overlap with those identified in cell-based assays performed in the absence of E3 ligase overexpression.

Enhancing TRPV4 IDR ubiquitination dampens stimulated channel responses
We next sought to determine the functional consequences of increased TRPV4 ubiquitination by NEDD4. HEK293T cells expressing WT TRPV4 alone, WT TRPV4 together with NEDD4, or WT TRPV4 together with NEDD4-CI for 24 h were treated with the TRPV4 agonist GSK101 (10 nM). Some cells expressing TRPV4 and NEDD4 showed an elevated baseline calcium prior to GSK101 application (Fig. S4A, left of dotted line). To ensure assessment of a healthy population of cells, we analyzed only those cells that did not exceed two standard deviations above WT TRPV4 baseline intracellular calcium levels. This analysis resulted in the exclusion of 11 (5.5%) cells from WT TRPV4 alone, 24 (13.6%) cells from WT TRPV4 + NEDD4, and 12 (6.2%) cells from WT TRPV4 + NEDD4-CI conditions (Fig. 5A). Coexpression of TRPV4 with NEDD4, but not NEDD4-CI, reduced the maximal intracellular calcium level observed following GSK101 application (Fig. 5B), delayed the time to reach this maximal response (Fig. 5C), and reduced the percent of cells responding to the TRPV4 agonist (Fig. 5D). NEDD4 coexpression also dampened and delayed calcium responses of cells treated with hypotonic saline (HTS, 37.5 mM NaCl), a well-characterized nonspecific TRPV4 activator (Figs. 5E, S4, B and C) (61). Increased calcium influx after HTS observed in cells expressing NEDD4-CI might result from NEDD4-CI binding and stabilizing other plasmalemma ion channels that respond to hypotonic solutions. Furthermore, when NEDD4 was overexpressed in 50B11 cells expressing endogenous TRPV4, we observed that NEDD4 but TRPV4 ubiquitination reduces plasmalemmal TRPV4 activity not NEDD4-CI reduced maximal intracellular calcium levels in response to GSK101 treatment (Fig. 5, F-H).
We next assessed whether the ability of NEDD4 to dampen GSK101-induced responses in HEK293T cells required the presence of specific lysine residues in the N-terminal IDR or CTD. We confirmed that the 5K-IDR and 5K-CTD lysine mutations did not interrupt binding of NEDD4 to TRPV4 (Fig. S4D) and then examined calcium responses of WT, 5K-IDR, and 5K-CTD TRPV4 to 10 nM GSK101 after antagonist washout. As expected, NEDD4 reduced WT TRPV4 responses to agonist stimulation (Fig. 5, I and J and S4E). Strikingly, NEDD4 did not suppress the response of 5K-IDR TRPV4 (Fig. 5, K and L and S4F) but did reduce 5K-CTD TRPV4 responses (Fig. 5, M and N and S4G) in a similar fashion to WT TRPV4. These findings indicate that NEDD4 suppressed WT TRPV4 channel activity through ubiquitination of the TRPV4 N-terminal IDR.
Increased TRPV4 IDR ubiquitination does not reduce TRPV4 cell surface abundance Canonically, ubiquitination of plasmalemmal ion channels reduces channel activity by stimulating channel endocytosis and removal from the plasma membrane (62). Conversely, blocking ubiquitination results in channel accumulation at the plasma membrane. However, our data from the 5K-IDR mutant suggest a gain-of-channel function without a change in cell surface expression (Fig. 3). We thus employed multiple TRPV4 ubiquitination reduces plasmalemmal TRPV4 activity methods to assess the effects of NEDD4 expression on TRPV4 cell surface localization. First, surface to total TRPV4 ratios were determined in HEK293T cells transfected with TRPV4-V5-EGFP alone or together with either NEDD4 or NEDD4-CI. Despite exhibiting increased ubiquitination with expression of NEDD4 (Fig. 4B), TRPV4 surface levels were not reduced by coexpression of NEDD4 or NEDD4-CI (Fig. 6, A and B). Comparable results were obtained using flow cytometry to assess TRPV4 surface localization in motor neuronlike 1 (MN-1) cells expressing TRPV4-V5-FLAG alone or with NEDD4 (Fig. S5A). We next used cell surface biotinylation of transfected HEK293T cells to assess TRPV4 expression at the plasma membrane at baseline and following stimulation with GSK101. In this assay, coexpression of NEDD4 did not alter surface TRPV4 abundance at baseline or in response to TRPV4 channel stimulation (Figs. 6, C and D, S5, B-E). Together, these approaches suggest that NEDD4mediated ubiquitination of TRPV4 does not significantly alter TRPV4 surface expression.
To assess cell surface localization in living cells we utilized total internal reflection fluorescence (TIRF) microscopy to measure TRPV4 surface fluorescence and intracellular calcium levels simultaneously using EGFP-tagged TRPV4 and the calcium indicator Cal590. Using TIRF microscopy, it has been previously reported that supramaximal doses of GSK101 (100 nM) cause rapid TRPV4 internalization from the plasma membrane in stably transfected HEK293 cells (63). We therefore first assessed TRPV4 dynamics and calcium levels in cells expressing WT TRPV4 treated with GSK101 at a dose of 10 nM (used throughout this study) or 100 nM. TRPV4 containing the M680K mutation that blocks the channel pore (64) was used as a negative control. Application of 10 nM GSK101 resulted in increased intracellular calcium concentrations but no loss of TRPV4 fluorescence from the TIRF evanescent field (Fig. 6, E-H, gray lines). In contrast, treatment with 100 nM GSK101 caused a large increase in intracellular calcium and a rapid loss of TRPV4 fluorescence from the TIRF evanescent field (Fig. 6, E-H, black lines). These data indicate that channel activation is not associated with internalization at low levels of stimulation, whereas supramaximal doses of GSK101 lead to rapid and robust TRPV4 internalization. We next assessed the effect of NEDD4 coexpression on TRPV4 TIRF fluorescence after treatment with 10 nM GSK101 (Fig. 6, I-L). NEDD4 delayed the onset of increased calcium levels following agonist application (Figs. 6, I and J, S5F), without an associated reduction in TRPV4 fluorescence from the TIRF evanescent field, indicating that the reduced calcium influx was not caused by decreased TRPV4 surface localization (Fig. 6, K and L).
Given that ubiquitination of TRPV4 did not alter cell surface expression, we investigated whether known N-terminal proteinprotein and protein-lipid interactions were affected by blocking ubiquitination. Interactions of the TRPV4 N-terminal domain with intracellular proteins such as RhoA and protein kinase C and casein kinase substrate in neurons 1-3 (PACSIN1-3) can regulate TRPV4 channel activity (Fig. S5G) without altering channel surface expression (29,32,35,65). Interactions with PIP 2 influence TRPV4 channel activity as well (36,65,66). Co-IP studies of PACSIN1 and TRPV4 showed reduced binding of PACSIN1 to 5K-IDR but not 5K-CTD TRPV4 (Fig. S5, H and I). However, mutation of TRPV4 residues K130 and K136 alone, which are part of the PACSIN SH3 domain-binding site, also reduced PACSIN1 binding (Fig. S5, H and I). As the K130, 136R mutations do not result in increased TRPV4 channel activity (Fig. S2F), these results indicate that the gain of function of the 5K-IDR mutant is not solely explained by disrupted PACSIN1 interaction. In addition, co-IP studies showed no differences in RhoA interaction between WT, 5K-IDR, and 5K-CTD TRPV4 (Fig. S5J). To assess whether altered lipid binding might contribute to increased channel activity of 5K-IDR TRPV4, we introduced neutralizing mutations within the positively charged residues of the PBD ( 121 KRWKR 125 to 121 AAWAA 125 ) in WT and 5K-IDR TRPV4 to disrupt interaction between the TRPV4 IDR and PIP 2 (36,65). The 5K-IDR-AAWAA TRPV4 mutant was expressed at the cell surface at similar levels to WT TRPV4 (Fig. 6, M and N) but partially rescued the gain of function of 5K-IDR TRPV4 in the antagonist withdrawal paradigm (Fig. 6, O and P). These data indicate that blocking PIP 2 binding can partially rescue the increased channel activity associated with decreased N-terminal ubiquitination and suggest that ubiquitination may regulate channel activity in part through modulation of lipid binding. Neuropathy-causing mutations exhibit reduced ubiquitination correlating with increased channel activity We and others have previously shown that neuropathycausing mutations in TRPV4 result in increased basal and stimulated channel activity (Fig. S6A) (29)(30)(31). While neuropathy-causing mutations are predominantly localized to one face of the ARD, interplay between the N-terminal IDR and the ARD has been proposed as a potential mechanism of TRPV4 channel modulation (31,33). In fact, it is still unknown how neuropathy-causing mutations result in increased channel activity. To determine whether neuropathy-causing mutations might impact ubiquitination of the TRPV4 N-terminal IDR, we were loaded with the calcium indicator Cal590 prior to imaging to allow for simultaneous assessments of TRPV4 surface localization and intracellular calcium levels. Representative TIRF images prior to and following GSK101 application are shown in E; warmer colors represent increased levels of performed cell-based Ub assays examining transiently transfected HEK293T cells, as well as HEK293T cell lines that inducibly express either WT (T-Rex-TRPV4 WT ) or R269C neuropathy mutant (T-Rex-TRPV4 R269C ) FLAG-tagged TRPV4 (32). Utilizing both approaches, we observed reduced ubiquitination of R269C TRPV4 compared with WT (Figs. 7, A  and B, S6B). The relative abundance of ubiquitinated peptides generated from WT and R269C TRPV4 was assayed by MS using isobaric labeling with tandem mass tags (TMTs) (Fig. 7C). Ubiquitination of R269C TRPV4 was reduced by approximately half at several lysine residues located in the Nterminal IDR, ARD, and CTD domains, indicating that R269C TRPV4 exhibits a broad reduction in ubiquitination, rather than a loss at specific sites (Fig. 7C).
To address whether neuropathy-causing TRPV4 mutants exhibit an intrinsic deficit in their capacity to be ubiquitinated, or whether reduced ubiquitination is a downstream consequence of elevations in intracellular calcium resulting from the gain-of-channel function, we assessed the effects of channel activity on TRPV4 ubiquitination. Culturing of T-Rex-TRPV4 cells in HC067 rescued T-Rex-TRPV4 R269C ubiquitination to T-Rex-TRPV4 WT ubiquitination levels (Fig. 7D) and equalized baseline intracellular calcium levels (Fig. 7E). Withdrawing the antagonist resulted in increased intracellular calcium levels in T-Rex-TRPV4 R269C cells compared with T-Rex-TRPV4 WT cells (Fig. 7E). Similar results were observed in transiently transfected HEK293T cells (Fig. 7F). In a Drosophila model of TRPV4 neuropathy (67), we also observed reduced ubiquitination of R269C TRPV4 as well as rescue of TRPV4 Ub levels in flies expressing R269C TRPV4 along with the poreinactivating M680K mutation (Fig. 7, G and H). These data suggest that the reduced ubiquitination of neuropathy-causing TRPV4 mutants is downstream of their increased channel activity and not because of a structural inability of the protein to be ubiquitinated.

Enhancing TRPV4 ubiquitination can partially rescue the gainof-channel function characteristic of neuropathy-causing mutants
Having observed that increasing WT TRPV4 ubiquitination reduces stimulated calcium responses, we postulated that increasing the ubiquitination of neuropathy mutant TRPV4 might ameliorate its gain-of-channel function. We found that NEDD4 coimmunoprecipitates with R269C TRPV4 (Fig. 7I), and that NEDD4 expression both enhanced R269C TRPV4 ubiquitination and inhibited mutant channel activity upon antagonist withdrawal (Fig. 7, I-L). Furthermore, NEDD4 expression delayed calcium influx in R269C TRPV4expressing cells stimulated with 10 nM GSK101 (Fig. 7, M-O). We also confirmed NEDD4-mediated suppression of neuropathy mutant channel activity by examining a second neuropathy-causing mutant (R237L TRPV4) in MN-1 cells. In these cells, NEDD4 coexpression reduced GSK101 responses in both WT and R237L neuropathy mutant-expressing cells (Fig. S6C). Together, these data further highlight the importance of N-terminal IDR ubiquitination in modulating TRPV4 channel activity and identify alterations of this pathway resulting from neuropathy-causing mutations that can be partially rescued by enhancing TRPV4 ubiquitination.

Discussion
As the plasmalemmal cation channel TRPV4 is increasingly implicated in both normal and pathologic states in numerous tissues (3,51,(68)(69)(70)(71)(72)(73)(74)(75)(76) and is activated by multiple physiological stimuli, it likely requires diverse regulatory mechanisms to be utilized in distinct cellular contexts (77). However, the molecular mechanisms regulating the TRPV4 channel under homeostatic and pathologic states remain poorly understood. Here, we investigated the role of ubiquitination, as this PTM is implicated in the regulation of numerous cell surfaceexpressed ion channels by triggering endocytosis (4). Using comprehensive unbiased MS profiling, we identified multiple monoubiquitinated lysine residues that cluster specifically in the IDRs of the intracellular N-terminal and C-terminal domains of the TRPV4 protein. We found that disrupting or enhancing ubiquitination of the N-terminal IDR modulates TRPV4-mediated calcium influx in opposing directions without altering TRPV4 cell surface abundance, revealing a novel mechanism for Ub-mediated tuning of plasmalemmal ion channel activity.
Although ubiquitination has not previously been shown to modulate ion channel activity in the absence of altered intracellular calcium (Cal590) or increased TRPV4 localization within the TIRF evanescent (TRPV4-EGFP). The scale bar represents 10 μm. Quantification of the fold changes (ΔF/F) in intracellular calcium and TRPV4 surface localization following GSK101 application are shown in F and G, respectively. Solid lines represent means, and shading represent SEM. n = 9-13 cells/condition from four independent experiments. H, quantification of surface TRPV4 fluorescence in the TIRF evanescent field 10 min after treatment with GSK101. Data are represented as means ± SD. n = 9 to 13 cells/condition from four independent experiments. ****p < 0.0001 (one-way ANOVA with Dunnett's multiple comparison test). I-L, TIRF imaging of HEK293T cells transfected with WT TRPV4-EGFP alone or with NEDD4 and stimulated by GSK101 (10 nM) application. Representative TIRF images prior to and following GSK101 application are shown in I. The scale bar represents 10 μm. Quantification of the fold changes (ΔF/F) in intracellular calcium and TRPV4 surface localization following GSK101 treatment are shown in J and K, respectively. Solid lines represent means, and shading represent SEM. n = 7 cells/condition from four independent experiments. L, quantification of the changes in TRPV4-EGFP TIRF signal for the experiments shown in K, examining the effects of NEDD4 coexpression on WT TRPV4 surface localization in response to 10 nM GSK101 treatment. Data are presented as means ± SD. n = 7-8 cells/condition from four independent experiments. Unpaired two-tailed t testK-. M, representative images of HEK293T cells expressing either WT or 5IDR TRPV4 and their AAWAA counterparts. Unpermeabilized cells were stained with V5 antibody to identify surface-localized TRPV4 (magenta), and total TRPV4 was visualized by the EGFP tag (green). The scale bar represents 10 μm.  trafficking, it has been suspected that this could occur (6,78). Importantly, the Ub-like PTM SUMO has been reported to modulate activity of plasmalemmal potassium channels without altering surface localization (79)(80)(81), and there is at least one example of a membrane protein, Gap1, that remains monoubiquitinated at the cell surface and is not internalized unless it becomes polyubiquitinated (21,82). Here, we employed several experimental methods including immunocytochemistry (ICC), cell surface biotinylation, and TIRF microscopy to demonstrate that little alteration of TRPV4 surface abundance was provoked by changes of TRPV4 N-terminal IDR ubiquitination. Ion channel IDRs are important allosteric modulators of channel activity (83)(84)(85) as well as sites for scaffolding protein interactions (86)(87)(88). Furthermore, protein IDRs are often hot spots for PTMs (89), which can affect the target protein's binding partners and structural conformation (90). Ubiquitination of protein IDRs is well studied, and evolutionarily conservation of Ub sites within IDRs is favored over ordered regions (91), although the effects of ion channel IDR ubiquitination on channel function are not well explored (33,87). However, there are numerous examples of monoubiquitination regulating protein function through alteration of protein-protein and protein-lipid interactions in other contexts (18-20, 92, 93), suggesting that monoubiquitination of the TRPV4 N-terminal IDR may similarly influence channel activity through effects on protein or membrane lipid interactions without affecting surface localization.
The TRPV4 N-terminal IDR is an extended 149 amino acid peptide region that, like other IDRs, is flexible, can undergo conformational changes, and is highly accessible for proteinprotein and protein-lipid interactions as well as PTMs (33). Previous studies have demonstrated multiple mechanisms by which TRPV4 IDRs can modulate channel activity: alteration of TRPV4 protein-protein interactions (37,94), disruption of PIP 2 interaction (31,36,66), and through phosphorylation of specific residues (95-97). We did not identify specific disruption of protein-protein interactions that accounted for the increased ion channel activity observed with disruption of N-terminal IDR ubiquitination. While changes in PACSIN1 interaction were observed with the 5K-IDR mutant that displays increased channel function, similar disruption of binding was seen with the more limited K130, 136R mutant which did not have increased channel function. These data indicate that the regulatory interplay between TRPV4 ubiquitination and channel activity is more complex than simply abrogating interactions with desensitizing binding partners. On the other hand, mutation of the PIP 2 -binding domain in the 5K-IDR mutant rescued the gain-of-ion channel function observed in the antagonist withdrawal paradigm, suggesting that ubiquitination of the N-terminal IDR may regulate channel activity at least in part through alterations in PIP 2 interactions. Our data argue that multiubiquitination of the TRPV4 N-terminal IDR modulates channel activity in the absence of altering cell surface abundance, thereby allowing TRPV4 to preserve its membrane scaffolding functions and ability to recruit binding partners such as RhoA, among others (32,35,40,59,65,98,99). Together, these results add to the emerging evidence of important modulatory roles of IDRs and provide a novel link between ubiquitination and IDR function in the regulation of ion channel activity.
Our finding of decreased ubiquitination of TRPV4 neuropathy mutants provides further evidence of the importance of Ub in the regulation of ion channel activity. We and others have shown that neuropathogenic mutations cause a gain-ofion channel function, both in vitro and in vivo (30,32,67). Here, we demonstrate that reduced ubiquitination of neuropathy-causing TRPV4 mutants is not a result of an intrinsic defect in their ability to be ubiquitinated, but is rather a consequence of the gain-of-ion channel function and resultant increased intracellular calcium levels. Restoring ubiquitination of neuropathy mutant TRPV4 through expression of NEDD4 was able to ameliorate increased channel activity. Given that the ubiquitination status of a protein is regulated by the balance of activity of both E3 ligases and deubiquitinating enzymes (DUBs) (100), one possibility is that excessive calcium influx through the mutant channels activates a DUB that deubiquitinates the mutant protein, further contributing to the pathological gain-of-channel function. Future studies will be needed to further define the ubiquitinating enzymes and DUBs that regulate TRPV4 in different cell types and the circumstances in which they are activated. This has particular relevance for emerging therapeutic strategies that aim to specifically modulate ubiquitination, such as designer DUBs and small-molecule inhibitors (101,102).
Increasingly, protein ubiquitination is appreciated as a complex code with a multiplicity of functional effects. While our studies focus on the functional consequences of monoubiquitination of specific lysines in the TRPV4 N-terminal IDR, there are likely multiple different types of ubiquitination events involving TRPV4 that contribute to its regulation. Specific ubiquitinated lysine residues were also very consistently observed in the C-terminal IDR of TRPV4, although their functional relevance is currently unclear. With the broad physiologic role of TRPV4 in different tissues and pathological states, and an increasing appreciation of the diverse functions of IDRs in regulating channel activity, understanding how TRPV4 function is modulated by its IDRs could serve as a Solid lines represent means, and shading represents SEM. K, quantification of the increase in intracellular calcium from baseline to 60 min after HC067 removal for the experiments shown in J. Data are presented as means ± SD. **p = 0.0069 (unpaired two-tailed t test). L, time to response quantified from J. n = 71 to 74 cells/condition. ****p < 0.0001 (unpaired two-tailed t test). M, calcium imaging of HEK293T cells expressing R269C TRPV4 without or with NEDD4 coexpression, showing responses to application of GSK101 (10 nM) immediately following HC067 removal (vertical dotted line). n = 51-57 cells/ condition from three independent transfections. Solid lines represent means, and shading represents SEM. N and O, quantification of the increase in intracellular calcium from baseline to 5 min after GSK101 treatment (N) and time to response (O) for the experiment shown in M. n = 46 to 52 cells/ condition. Data are presented as means ± SD. ****p < 0.0001 (unpaired two-tailed t test). GSK101, GSK1016790A; HC067, TRPV4-specific antagonist HC067047; HEK293T, human embryonic kidney 293T cell line; TRPV4, transient receptor potential vanilloid 4. gateway to studying the role of IDR regulation of other ion channels. Future studies dissecting what may be a complex Ub code targeting ion channel IDRs under different conditions and in different cell types will have important implications for understanding ion channel function in health and disease.

Mice
All animal procedures followed the National Institutes of Health guidelines and were approved by the Johns Hopkins University Animal Care and Use Committee. Conditions in the mouse facility were maintained at 42% humidity, 22.2 C, and a 14:10 h light-dark cycle. TRPV4-null mice (Fig. S1A) (41) were kindly provided by Michael Caterina, and HA-Ub mice (48) were a generous gift from Hidde Ploegh. A second TRPV4 KO mouse line (Fig. S1B) was generated by the Jackson Laboratory and previously characterized by our laboratory (42).
The following plasmids were obtained from Addgene:

Transient transfection
HEK293T, MN-1, and MDCK cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum and penicillin/streptomycin at 37 C with 5% CO 2 . For Ub assays, calcium imaging, and TIRF microscopy, HEK293T cells were plated at 4.5 × 10 5 cells/ml in 6-well plates 24 h prior to transfection. Cells were transfected using Lipofectamine LTX (Thermo Fisher Scientific) as per the manufacturer's instructions with a total of 2.5 μg DNA/well, comprising 0.5 μg of TRPV4 plasmid, 0.7 μg of E3 ligase plasmid (where required), and 1. MDCK cells were transfected with GenJet II transfection reagent (SignaGen) as per the manufacturer's instructions. Positively transfected cells were selected with 1000 μg/ml G418 sulfate (Thermo Fisher Scientific) 6 h after transfection and allowed to grow to confluency up to 72 h posttransfection.
50B11 cells were cultured in Neurobasal medium supplemented with 5% fetal bovine serum, 2% B-27 supplement, 1% GlutaMAX, and 0.2% glucose at 37 C with 5% CO 2 . For calcium imaging experiments, 50B11 cells were transfected with Fugene 6 (Promega) as per the manufacturer's instructions. Cells were plated at a density of 2 × 10 5 cells/ml in 6-well plates and transfected at 24 h postplating with a total of 2 μg DNA/well, comprising 0.7 μg of GFP plasmid (to identify transfected cells), 0.7 μg of E3 ligase plasmid, and 0.6 μg of empty pcDNA3.1 vector. All cells were assayed 24 h after transfection. For cell-based Ub assays, pLenti Puro HA-Ubiquitin lentivirus was added to 50B11 cells at 0.5 multiplicity of infection for 48 h. Transduced cells were selected with 2 μg/ml puromycin in media changed every 2 days until cells reached 80% confluency.

IP
At 24 h post-transfection, HEK293T, 50B11, and MDCK cells were washed three times in PBS containing 0.9 mM CaCl 2 and 0.5 mM MgCl 2 (PBS++) and then lysed in radioimmunoprecipitation assay (RIPA) buffer (MilliporeSigma) supplemented with Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific) and the DUB inhibitor PR-619 (50 μM; MilliporeSigma). Lysates were centrifuged at 17,000g for 15 min, and supernatants were then incubated for 1 h at 4 C with rotation in 25 μl of Dynabeads Protein G (Thermo Fisher Scientific) preconjugated with 2 μg of anti-FLAG M2 antibody. Beads were then washed three times with 0.2% Tween-20 in PBS and transferred to new tubes. Immunoprecipitated protein was eluted from the beads in sample buffer (RIPA buffer containing 1× Laemmli sample buffer and 10% 2-mercaptoethanol) at 70 C for 10 min. For IP of TRPV4 from mouse choroid plexus, mice were deeply anesthetized with isoflurane and euthanized by cervical dislocation. Choroid plexus tissues were dissected from the lateral and fourth ventricles, lysed in 50 μl of ice-cold RIPA buffer supplemented with protease inhibitor cocktail, and homogenized with a polypropylene pestle (MilliporeSigma). Samples were then processed as described previously, with incubation in Dynabeads Protein G preconjugated with anti-TRPV4 antibody (Abcam; catalog no.: ab39260). For Drosophila, 30 heads were collected from 3-day-old adult flies expressing TRPV4 variants selectively in neurons (C155-GAL4 pan-neuronal driver). Lysis and IP protocols were identical to those described previously for mouse choroid plexus tissue.

WB
Whole cell lysates (inputs) were reduced in Laemmli sample buffer with 10% 2-mercaptoethanol for 10 min at 70 C. Lysates and eluates were resolved on 4 to 15% or 7.5% TGX gels (Bio-Rad) and then transferred to polyvinylidene difluoride membranes (Thermo Fisher Scientific). Membranes were blocked with 5% bovine serum albumin (BSA) in Tris-buffered saline containing 0.1% Tween-20 (TBST) for 1 h at room temperature with rocking. For experiments requiring the Ubhorseradish peroxidase (HRP) primary antibody, blocking was performed in 1% BSA in TBST. Primary antibodies were applied overnight at 4 C in TBST containing 2.5% BSA, and secondary antibodies were applied for 1 h at room temperature in TBST. Membranes were developed with SuperSignal West Pico Substrate (Thermo Fisher Scientific) and imaged using an ImageQuant LAS 4000 system (GE Healthcare). Densitometric analyses were performed using open source FIJI ImageJ software.

ICC
At 24 h post-transfection, cells were washed briefly with PBS++ on ice and then blocked in PBS++ containing 0.1% BSA for 30 min at 4 C. Cells were then incubated in blocking solution containing mouse anti-V5 antibody (1:5000 dilution) for 1 h at 4 C, washed three times for 5 min in PBS++, and fixed in 2% paraformaldehyde for 30 min at room temperature. For experiments requiring immunolabeling of intracellular proteins (as in the experiments with NEDD4 coexpression, Fig. 6), cells were permeabilized with 0.3% Triton X-100 for 15 min, blocked in TBS containing 1% BSA and 0.2% milk for 1 h at room temperature, and then incubated in TBS blocking solution containing the required primary antibody for 1 h at room temperature. All cells were then washed and stained with secondary antibodies for 1 h at room temperature in TBS blocking solution. Coverslips were mounted on slides with Prolong Diamond Antifade Medium with 4 0 ,6-diamidino-2phenylindole (Thermo Fisher Scientific) and imaged using a Zeiss LSM 800 Airyscan confocal microscope. Images were quantified in FIJI by selecting individual cells and measuring the mean gray areas for both the surface stain (V5) and total TRPV4 (EGFP). Surface/total ratio was then calculated and reported.

Flow cytometry
Transfected MN-1 cells were gently harvested on ice using a cell scraper, pelleted, and washed three times with ice-cold PBS++. Immunocytochemical staining of cell surface TRPV4 was performed as described previously using a mouse DyLight 488-conjugated mouse anti-V5 antibody, with washing steps performed by pelleting cells at 156g for 5 min at 4 C. Immunolabeled cells were then analyzed using a FACS Lyric flow cytometer (BD Biosciences).

Cell surface biotinylation
HEK293T cells were cultured in 10 cm dishes and transfected at 80% confluency using polyethylenimine (Polysciences, Inc). Cell surface biotinylation was performed at 24 h post-transfection with the Pierce Cell Surface Protein Isolation Kit (Thermo Fisher Scientific). Briefly, cells were rinsed twice with PBS++ on ice and then incubated with 0.25 mg/ml EZ-Link Sulfo-NHS-SS-Biotin with gentle rotation (30 rpm) for 30 min at 4 C. Excess biotin was quenched with three successive washes of 50 mM glycine in PBS. Cells were subsequently lysed in RIPA buffer supplemented with protease inhibitor cocktail (MilliporeSigma) for 30 min on ice. Lysates were transferred to microcentrifuge tubes, sonicated at low power, and cleared by centrifugation at 17,000g in a refrigerated bench-top microcentrifuge. About 10% of the lysate volume was reserved as input, whereas the remainder was incubated with NeutrAvidin agarose slurry with end-over-end rotation for 2 h at 4 C. The slurry was then washed three times each by centrifugation with wash buffer and high-salt (0.5 M NaCl) wash buffer, respectively. Biotinylated proteins were eluted in Laemmli sample buffer supplemented with 50 mM DTT. Lysates were heated at 70 C for 10 min prior to WB analysis, as described previously. Surface/total TRPV4 ratios were quantified by densitometry using FIJI software. Normalized values for surface and total TRPV4 were calculated by dividing TRPV4 band intensity by the corresponding transferrin receptor band intensity. The ratio of surface to total TRPV4 was then calculated by dividing the normalized TRPV4 surface value by the normalized total value (as reported in Fig. S5, B and C).

Calcium imaging
Calcium imaging was performed using a Zeiss Axio Observer.Z1 inverted microscope equipped with a Lambda DG-4 wavelength switcher (Sutter Instrument Company). Cells were loaded with Fura-2 AM (8 μM; Thermo Fisher Scientific) for 60 min at 37 C in culture medium. For antagonist withdrawal assays, 0.5 μM HC067 was also included during Fura-2 loading. Coverslips were then transferred to the microscope and placed into calcium imaging buffer (150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 10 mM glucose, 10 mM Hepes, and pH 7.4). For antagonist withdrawal experiments, a gravity flow system was used to wash out HC067containing calcium imaging buffer from the coverslip chamber continuously over the course of an hour. Excess buffer was pumped out with a peristaltic pump (New Era Pump Systems). For calcium-free experiments, CaCl 2 was omitted from the buffer. Cells were imaged every 10 s for the first 6 min, with 1 min of baseline and 5 min following washout of antagonist. The imaging rate was then slowed to once per minute for the following 25 min and finally once every 5 min for the final 30 min of imaging. Calcium levels at each time point were computed by determining the ratio of Fura-2 AM emission at 340 nM divided by the emission at 380 nM. Data were expressed as either raw Fura ratio or change in final Fur-2 ratio from initial Fura-2 ratio. Images were processed using Axio-Vision 4 (Zeiss).
For HTS treatment, three volumes of NaCl-free calcium imaging buffer were added to one volume of standard calcium imaging buffer for a final NaCl concentration of 37.5 mM. For GSK101 treatment, GSK101-containing buffer (Milli-poreSigma) was added 1:1 (v:v) to the imaging chamber to achieve the appropriate final concentration. Cells were imaged every 10 s, first taking baseline measurements for 30 s, then imaging for 5 min after treatment.

TIRF imaging
Cells were plated on 35 mm glass bottom dishes (#1.5 coverglass; MatTek; catalog no.: P35G-1.5-14-C) and transfected as described previously. One hour before imaging, cells were loaded with Cal590 AM (1 μg/μl; AAT Bioquest) and subsequently imaged in calcium imaging buffer 24 h after transfection. The evanescent wave for TIRF was established using micromirrors positioned below a high numerical aperture apochromat objective (100×, 1.5 numerical aperture; Olympus) mounted on a Nikon Eclipse microscope running NIS-Elements software (Nikon). TRPV4-V5-EGFP was excited by a 488 nm Argon laser line, and Cal590 was excited by a 561 nm 0.5 W fiber laser. Emissions were collected through a GFP/red fluorescent protein/GFP-red fluorescent protein FRET filter set using an Evolve 512 EMCCD camera (Photometrics). Images were captured with a 50 ms exposure every 5 s. The surface expression level of TRPV4 was determined for one to three cells from a random field of view per dish. Mean fluorescence intensity was analyzed using FIJI. GSK101 treatment was performed as in the calcium imaging experiments described previously.

ITCH
Full-length ITCH (residues 1-903) was synthesized by GenScript in a pET11a vector with an N-terminal His6tobacco etch virus (TEV) cleavage site. Briefly, the construct was expressed in E. coli BL21-Gold(DE3) in LB medium supplemented with 100 μg/ml ampicillin. Cells were grown at 37 C to an absorbance of 0.8 at 600 nm, moved to 18 C, and were induced with 250 μM IPTG. Growth continued at 18  and sonicated on ice. The cell lysate was loaded onto a nickelnitrilotriacetic acid gravity flow column (Qiagen), and protein was ultimately eluted with 500 mM imidazole and dialyzed (50 mM Tris [pH 8], 20 mM imidazole, 200 mM NaCl, 1 mM DTT, and 1 mM EDTA) overnight at 4 C in the presence of TEV protease (molar ratio of 30:1). The His6-TEV tag was cleaved from the TRPV4 protein and further purified by sizeexclusion chromatography (HiLoad prep grade 16/60 Super-dex200; GE Healthcare) in 50 mM Tris (pH 8), 200 mM NaCl, 1.5% (v/v) glycerol, and 1 mM DTT. Protein purity was confirmed via SDS-PAGE and subsequent Coomassie blue staining.

NEDD4 and Ub pathway components
Recombinant NEDD4, UBE1, UbcH5c, and Ub were generous gifts of the Cole Laboratory (Harvard Medical School) and generated as previously described (104).

MS sample preparation: in-gel digestion
Proteins in gel bands (approximately 1 μg) were destained and reduced in 100 μl of 50 mM DTT in 50% MeOH at 54 C for 45 min while shaking. To confirm that the alkylating reagents did not affect the identification of modified lysines, proteins within gel bands were alkylated with various alkylating reagents independently, including 200 mM S-methyl thiomethanesulfonate, 200 mM iodoacetamide, and 200 mM chloroacetamide (Table S2) for 15 min in the dark. Gel pieces were processed as previously described (105) and digested with 40 μl trypsin/Lys-C (20 μg in 2 ml 10 mM triethylamine bicarbonate) overnight at 37 C. Peptides were extracted with three additions of 50 μl of 60% acetonitrile with 0.1% TFA and then dried. Acidified peptides were desalted on a house-made C18 stage tip. Samples were reconstituted in 20 μl of 2% acetonitrile/0.1% formic acid.
Peptides were analyzed by reverse-phase chromatography at 300 nl/min using a 2 to 90% acetonitrile gradient over 90 min on a nano-Orbitrap Fusion Lumos-ETD in FTFT (Thermo Fisher Scientific) interfaced with a nano-LC 1100 system (Thermo Fisher Scientific). Peptides were desalted on a C18 trap (S-10 μM, 120 Å, 75 μm × 2 cm; YMC) and subsequently separated on an in-house packed PicoFrit column (75 μm × 200 mm, 15u, +/−1 μm tip; New Objective) with C18 phase (ReproSil-Pur C18-AQ, 3 μm) 75 μm × 150 mm ProntoSIL-120-5-C18 H column (3 μm, 120 Å; BISCHOFF). Eluting peptides were sprayed into the Lumos-ETD mass spectrometer at 2.7 kV. Survey scans (full MSs) were acquired on Orbitrap within 400 to 1800 Da m/z using a data-dependent 3 s cycle time method to monitor the top 15 m/z with dynamic exclusion of 10 s. Precursor ions were individually isolated with 0.7 Da (no offset) and fragmented using higher-energy collisional dissociation with an activation collision energy of 30%. Precursor and the fragment ions were analyzed at resolution of 120,000 and 60,000 and an automatic gain control (AGC) of 400 k and 50 k, respectively.

MS data analysis
Mass spectra were processed by Proteome Discoverer, version 2.4 (Thermo Fisher Scientific) and searched against the RefSeq2021_204_Human protein database containing 71,622 sequences (including the human TRPV4 K>R constructs based on the National Center for Biotechnology Information accession number: NP_067638.3) using Mascot, version 2.7 (Matrix Science). The following were the search criteria: trypsin as enzyme, two missed cleavages, precursor mass tolerance of 5 ppm, and fragment mass of tolerance 0.01 Da. The following were the variable modifications: oxidation on M, Ub (GG) on K, deamidation on N and Q, and static carbamidomethyl on C. Low scoring spectra (less than 1% false discovery rate [FDR]) were filtered and redirected to BYONIC to search a custom database containing all human sequences plus the modified sequences for TRPV4 (e.g., K>R mutants). Peptide identifications were evaluated using Target Decoy Validator based on peptide Q values and with a strict target FDR for peptide-spectral matches and peptides at 0.01. A 1% FDR was calculated using a concatenated decoy database.

TMT 8-plex analysis: labeling
Peptide digests were reconstituted in 10 μl of 100 mM triethylamine bicarbonate and labeled using TMT 10-plex reagents (Thermo Fisher Scientific; lot no.: TJ68848) following the manufacturer's instructions. Labeling was quenched with 0.8 μl hydroxylamine, and samples were then combined and dried using a SpeedVac.

TMT 8-plex analysis: MS method
TMT-labeled peptides (25% of total) were injected and analyzed on a nano-Orbitrap Fusion Lumos-ETD as described previously with the following modifications: 120 min gradient, 360 to 1700 Da m/z survey scan, higher-energy collisional dissociation with an activation collision energy of 39%, precursor ions analyzed at resolution of 120,000, 50 ms injection time with an AGC target of 400,000 ions, fragment ions analyzed at 60,000 with a 118 ms injection time. Fragment ion scanning starting at 120 m/z with normalized AGC target set to 200% and 100,000 ions.
TMT 8-plex analysis: MS data analysis TMT 8-plex mass spectra were processed by Proteome Discoverer and Mascot software as described previously with the following modifications: variable modifications included TMT on K, oxidation on Met, phosphorylation STY, deamidation on N and Q, and GG on K; static modifications included carbamidomethyl on C and TMT on N terminus.
Reporter ions were quantified using S/N values for unique peptides at 5% FDR. Ratios were calculated using an ANOVA based on protein abundance, and either not normalized or normalized on the TRPV4 protein sequence.