Data resource: vasopressin-regulated protein phosphorylation sites in the collecting duct

Abstract

Vasopressin controls renal water excretion through actions to regulate aquaporin-2 (AQP2) trafficking, transcription, and degradation. These actions are in part dependent on vasopressin-induced phosphorylation changes in collecting duct cells. Although most efforts have focused on the phosphorylation of AQP2 itself, phosphoproteomic studies have identified many vasopressin-regulated phosphorylation sites in proteins other than AQP2. The goal of this bioinformatics-based review is to create a compendium of vasopressin-regulated phosphorylation sites with a focus on those that are seen in both native rat inner medullary collecting ducts and cultured collecting duct cells from the mouse (mpkCCD), arguing that these sites are the best candidates for roles in AQP2 regulation. This analysis identified 51 vasopressin-regulated phosphorylation sites in 45 proteins. We provide resource web pages at https://esbl.nhlbi.nih.gov/Databases/AVP-Phos/ and https://esbl.nhlbi.nih.gov/AVP-Network/, listing the phosphorylation sites and describing annotated functions of each of the vasopressin-targeted phosphoproteins. Among these sites are 23 consensus protein kinase A (PKA) sites that are increased in response to vasopressin, consistent with a central role for PKA in vasopressin signaling. The remaining sites are predicted to be phosphorylated by other kinases, most notably ERK1/2, which accounts for decreased phosphorylation at sites with a X-p(S/T)-P-X motif. Additional protein kinases that undergo vasopressin-induced changes in phosphorylation are Camkk2, Cdk18, Erbb3, Mink1, and Src, which also may be activated directly or indirectly by PKA. The regulated phosphoproteins are mapped to processes that hypothetically can account for vasopressin-mediated control of AQP2 trafficking, cytoskeletal alterations, and Aqp2 gene expression, providing grist for future studies.

NEW & NOTEWORTHY Vasopressin regulates renal water excretion through control of the aquaporin-2 water channel in collecting duct cells. Studies of vasopressin-induced protein phosphorylation have focused mainly on the phosphorylation of aquaporin-2. This study describes 44 phosphoproteins other than aquaporin-2 that undergo vasopressin-mediated phosphorylation changes and summarizes potential physiological roles of each.

INTRODUCTION

Vasopressin controls the osmotic water permeability of the renal collecting duct through actions in principal cells to regulate the water channel aquaporin-2 (AQP2) (1–10). AQP2 is regulated in at least three ways: 1) vasopressin controls trafficking of AQP2-containing vesicles to and from the plasma membrane (11–14), 2) vasopressin increases the half-life of the AQP2 protein (15, 16), and 3) vasopressin increases the rate of transcription of the Aqp2 gene (4, 17). In addition, extensive evidence indicates that the regulation of trafficking of AQP2 depends on vasopressin-induced rearrangements in the actin cytoskeleton (18–22). These actions depend on the binding of arginine vasopressin to V2 receptors in the basolateral plasma membrane, resulting in G_αs-mediated activation of adenylyl cyclase 6, increasing cAMP levels and protein kinase A (PKA) activity (8, 23). Ultimately, the effects of vasopressin in collecting duct principal cells depend almost entirely on PKA (24–26), either through direct PKA-mediated phosphorylation of proteins or indirectly through PKA-mediated regulation of other protein kinases (27). An example of indirect effects of PKA activation are mitogen-activated protein kinases (MAPKs), ERK1 and ERK2, which are decreased in activity in response to vasopressin in native collecting ducts in part via PKA-dependent Raf1 inhibition (28). In addition to cAMP, an important second messenger is Ca²⁺, which is increased in collecting duct principal cells in response to V2 receptor occupation (29–31). This response has been proposed to be secondary to PKA-dependent phosphorylation of inositol 1,4,5-trisphosphate receptors Itpr1 and Itpr3, sensitizing them to activation by fluctuations in basal Ca²⁺ levels (24). Regulation of water permeability in collecting ducts occurs in part through Ca²⁺/calmodulin-mediated activation of myosin light chain kinase (Mylk), a regulator of conventional nonmuscle myosins in collecting duct cells (20). Beyond its action to phosphorylate myosin regulatory light chain, several noncanonical targets of Mylk have been identified by CRISPR/Cas9-mediated gene deletion followed by phosphoproteomics (32). Beyond Mylk, there are many other kinases that are regulated by Ca²⁺/calmodulin, including Ca²⁺/calmodulin-dependent protein kinase type IIδ (Camk2d), which has been proposed to phosphorylate AQP2 at Ser²⁵⁶ (33, 34).

The intracellular actions of vasopressin in the renal collecting duct are shown in Table 1 (4, 11–22, 24, 29–31, 35–42). These actions all appear to be responses to vasopressin-dependent phosphorylation events catalyzed by PKA and PKA-regulated downstream protein kinases. However, the specific phosphorylation events that mediate the actions shown in Table 1 are largely unidentified. We know from phosphoproteomics and immunoblot analysis with phospho-specific antibodies that AQP2 itself undergoes changes in phosphorylation in a cluster of four COOH-terminal serines (Ser²⁵⁶, Ser²⁶¹, Ser²⁶⁴, and Ser²⁶⁹), which are known as pivotal phosphorylation sites for AQP2 trafficking to the plasma membrane (7, 43–55). Specifically, Ser²⁵⁶, Ser²⁶⁴, and Ser²⁶⁹ undergo increases in phosphorylation in response to vasopressin (45, 47, 48, 56–59), whereas Ser²⁶¹ undergoes a decrease (48, 57, 60). Recent studies have used powerful quantitative phosphoproteomics methods to identify additional vasopressin-dependent phosphorylation changes in native rat inner medullary collecting duct (IMCD) cells (61) and cultured mpkCCD cells derived from mouse collecting ducts (27). The former identified 156 vasopressin-regulated phosphopeptides, and the latter identified 429 vasopressin-regulated phosphopeptides. A long-term goal is to identify the roles of all of these sites in the vasopressin-regulated processes shown in Table 1. In this paper, to prioritize vasopressin-regulated phosphorylation sites for future study, we identified the vasopressin-regulated sites that are common to both rat native IMCDs and cultured mouse mpkCCD cells, arguing that phosphorylation sites that change in the same way in both models are the most likely to play physiological roles in the regulation of AQP2. Also, we except that the most consistent findings between rat and mouse collecting ducts will be representative of the species-independent behavior of the mammalian collecting duct. Using this approach, we identified 51 phosphorylation sites in 45 phosphoproteins that show the same changes in mpkCCD cells and IMCD cell suspensions in response to the V2 receptor-specific vasopressin analog desmopressin [1-deamino-8-d-arginine vasopressin (dDAVP)]. We mapped these phosphoproteins to known cellular actions of vasopressin (Table 1) and created a data resource to allow users to access the information. Although the major focus of this paper is on the regulation of AQP2, the information shown in Table 1 emphasizes that vasopressin regulates other processes in collecting duct cells (such as apoptosis, cell proliferation, and tight junction permeability) and some of the phosphorylation changes discussed here may be relevant to these other vasopressin-regulated processes.

Table 1.

Intracellular processes regulated by vasopressin in collecting duct cells

Vasopressin Actions	Reference(s)
Increased exocytosis of AQP2-containing vesicles	(11, 12, 14)
Decreased AQP2 endocytosis	(11, 13)
Increased AQP2 protein half life	(15, 16)
Increased Aqp2 gene transcription	(4, 17)
Intracellular Ca2+ mobilization	(29–31)
Reorganization of actin filaments	(18–20, 22)
Depolymerization of filamentous actin	(18, 19, 21)
Microtubule-dependent AQP2 redistribution	(35)
Increased tight junction permeability	(36)
Decreased rate of apoptosis	(37)
Decreased rate of proliferation	(38)
Increased principal cell size	(39–42)
Decreased ERK1/2 activity	(24, 28)

AQP2, aquaporin-2.

METHODS

Curation of Consensus Vasopressin-Regulated Phosphorylation Sites

Phosphorylation sites that significantly changed in abundance in response to 30-min treatment of the vasopressin analog dDAVP in both cultured mouse mpkCCD cells (27) and in native rat IMCD suspensions (61) were curated into a single data set. Because amino acid numbering in rat versus mouse proteins can differ, curation was done by mapping the four amino acid sequences surrounding the phosphorylation sites (X-X-p[S/T]-X), that is, by comparing two amino acids upstream from a given site and one amino acid downstream from the site. The mapping process is provided in Supplemental Spreadsheet S1 (see https://esbl.nhlbi.nih.gov/Databases/Supplemental-Phosphorylation/). The final list of consensus vasopressin-regulated phosphorylation sites is provided at https://esbl.nhlbi.nih.gov/Databases/AVP-Phos/. Automated Bioinformatic Extractor (ABE; https://esbl.nhlbi.nih.gov/ABE/) was used to create annotations for individual proteins based on official gene symbols. Protein functions were derived from the “[FUNCTION]” fields in UniProt records for individual proteins. Reported amino acid numbering is based on the UniProt Accession numbers provided at https://esbl.nhlbi.nih.gov/Databases/AVP-Phos/. Our laboratory carried out previous quantitative phosphoproteomic analysis in mouse mpkCCD cells (62) and rat IMCD cells (63), but these were not included in the present analysis owing to the fact that earlier LC-MS/MS methodologies were far less sensitive, leading to many fewer sites being quantified. In general, the sites found in these prior studies overlapped the sites reported here and showed the same direction of change.

RESULTS AND DISCUSSION

We curated a list of 51 phosphorylation sites in 45 proteins that significantly changed in abundance in response to the vasopressin analog dDAVP in both cultured mouse mpkCCD cells (27) and in native rat IMCD suspensions (61) (organized by “Molecular Function” in Fig. 1). The curation process is detailed in Supplemental Spreadsheet S1. Information about these sites is archived for browsing and download at https://esbl.nhlbi.nih.gov/Databases/AVP-Phos/. The vasopressin-regulated phosphoproteins are mapped to known effects of vasopressin at https://esbl.nhlbi.nih.gov/AVP-Network/. This webpage displays documentation from prior literature about each node by hovering over the node. Although the annotations provided focus mainly on regulation of AQP2-mediated water transport, we acknowledge that vasopressin regulates sodium transport in the cortical collecting duct and regulates urea transport in the IMCD and findings may be relevant to these regulatory events as well.

Figure 1.

Consensus vasopressin-regulated protein phosphorylation sites in collecting duct cells. Two separate datasets, one from the cultured collecting duct cells from the mouse (mpkCCD) cell line treated with 0.1 nM desmopressin [1-deamino-8-d-arginine vasopressin (dDAVP)] for 30 min and one from an inner medullary collecting duct (IMCD) suspension treated with 1 nM dDAVP for 30 min, were combined to identify the phosphorylation sites that changed significantly in response to vasopressin in both studies (27, 61). 51 consensus dDAVP-responsive phosphosites are identified.

Figure 2 shows a graphical display of the vasopressin-regulated phosphorylation sites mapped to molecular functions. Most of the vasopressin-regulated phosphorylation sites have not been studied in the context of the collecting duct, and their listing in Figs. 1 and 2 provide a convenient list of hypotheses for future investigations.

Figure 2.

Relational graph showing the vasopressin-regulated phosphoproteins mapped to molecular functions. An annotated version with popups showing descriptions of each node can be found online at https://esbl.nhlbi.nih.gov/AVP-Network/. The 51 phosphosites are categorized into 11 vasopressin-associated cellular processes (see Table 1). Each cellular process is known to play roles in the regulation of collecting duct water permeability.

The molecular functions highlighted in Figs. 1 and 2 include several categories related to known aspects of vasopressin-mediated signaling in collecting duct cells, viz. “cAMP-dependent signaling,” “calcium signaling,” “regulation of GTPase activity,” and “protein phosphorylation.” Notably, in the latter category, there were several protein kinases, namely misshapen-like kinase 1 (Mink1), Ca²⁺/calmodulin-dependent protein kinase kinase 2 (Camkk2), cyclin-dependent kinase 18 (Cdk18), Erbb3, and Src, which play likely roles downstream from PKA. The remaining molecular function categories shown in Figs. 1 and 2 also correspond closely to cellular level effects of vasopressin shown in Table 1, viz. “vesicle-mediated transport,” “regulation of transcription,” “actin cytoskeleton,” “microtubule cytoskeleton,” and “cell polarity.” An additional category, “RNA processing,” was not anticipated from prior studies.

Phosphorylation Changes in Response to Vasopressin Are Dependent on PKA Activity

We hypothesized that the demonstrated vasopressin-regulated phosphorylation sites shown in Fig. 1 are either PKA targets or are downstream from PKA actions, e.g., through regulation of non-PKA kinases secondary to PKA activity changes. If so, the phosphorylation changes seen with deletion of PKA should be opposite to the changes seen in response to vasopressin. The PKA catalytic subunit is coded by two separate genes expressing PKA-Cα and PKA-Cβ proteins. Isobe et al. (24) deleted both (PKA double knockout) in mouse mpkCCD cells and carried out quantitative phosphoproteomics. Figure 3A shows a plot of phosphorylation changes in response to vasopressin in mpkCCD cells versus changes in response to PKA deletion for all sites quantified in both studies. In general, there was a strong negative correlation, confirming the hypothesis. Specifically, this provides further support for the prior observation that almost all vasopressin-mediated signaling in collecting duct cells is dependent on PKA activation (25). One exception is the regulatory subunit of PKA, Prkar2a, which inhibits PKA activity by binding to the catalytic site when cAMP is absent. Its phosphorylation increases at Ser⁹⁶ when bound to the catalytic subunit, i.e., in the absence of cAMP (64). The complete set of phosphorylation changes in response to PKA deletion are shown as a volcano plot in Fig. 3B. The two catalytic subunit proteins, PKA-Cα and PKA-Cβ, have extremely similar amino acid sequences in their catalytic regions and would be expected to phosphorylate the same proteins if they were similarly located in the cell. As shown in Fig. 3C, these two PKA proteins have different effects on phosphorylation when deleted individually, implying that they may be compartmentalized to interact with different substrates (26). Specifically, it was concluded that PKA-Cα deletion chiefly affected phosphorylation of proteins associated with cell membranes and membrane vesicles, whereas target proteins in PKA-Cβ-null cells were largely associated with the actin cytoskeleton and cell junctions (26). Among the phosphorylation sites that stand out in Fig. 3C are Bin1, Sec22b, Slc9a3r1, Agfg1, and Ralgapa2, which change in opposite directions in response to deletion of the two PKA catalytic proteins. Overall, we conclude that vasopressin-dependent signaling is almost entirely PKA dependent, but the two PKA catalytic proteins perform different functions, presumably due to differences in subcellular localization.

Figure 3.

Role of protein kinase A (PKA) in vasopressin-mediated phosphorylation. A: correlation between the response to vasopressin for the sites shown in Fig. 1 and response to deletion of both PKA catalytic genes in cultured collecting duct cells from mouse (mpkCCD) cells. The plot shows a strong negative correlation (P < 0.01). B: volcano plot showing changes in phosphorylation for all sites in mpkCCD cells in response to deletion of both PKA catalytic genes, highlighting the phosphorylation sites shown in Fig. 1. C: desmopressin [1-deamino-8-d-arginine vasopressin (dDAVP)]-responsive phosphosites show differing responses to PKA-Cα single knockout (KO) versus PKA-Cβ single KO mpkCCD cells. Many of the dDAVP-responsive phosphosites showed opposite phosphorylation changes in PKA-Cα KO versus PKA-Cβ KO cells, suggesting that the two PKA catalytic subunits have different functions in the cell. Green indicates sites with increased phosphorylation in response to dDAVP; pink indicates sites with decreased phosphorylation in response to dDAVP. Ctrl, control.

Some Vasopressin-Regulated Phosphorylation Events Have Known Effects on Protein Function

Among the 51 phosphorylation sites regulated by vasopressin, eight sites have been investigated in prior studies in noncollecting duct tissues with regard to the effect of phosphorylation on the activities of the phosphoproteins. These are shown in Table 2 (65–78). The following paragraphs link these sites to possible roles in collecting duct principal cells.

Table 2.

Effect of phosphorylation on protein activity for vasopressin-regulated phosphorylation sites

Gene	Phosphosite	Effect of Phosphorylation	Reference(s)
Arhgef2	Ser885	PKA-mediated phosphorylation at Ser885 inhibits the GEF activity	(65, 66)
Exoc7	Ser250	ERK1/2-mediated phosphorylation stimulates exocyst complex assembly and exocytosis	(67)
Hdac4	Ser245	CaMK- and Mark kinase-mediated phosphorylation at Ser245 promote the interaction with 14-3-3, which induces nuclear export and cytoplasmic accumulation	(68, 69)
Itpr1	Ser1588, Ser1755	PKA-mediated phosphorylation at Ser1588 and Ser1755 enhances Ca2+ release	(70, 71)
Nsfl1c	Ser176	PKA-mediated phosphorylation at Ser176 induces Nsfl1c activity	(72)
Src	Ser17	PKA-mediated phosphorylation at Ser17 induces Src activity	(73)
Stim1	Ser575	ERK1/2-mediated phosphorylation at Ser575 activates Stim1-mediated store-operated Ca2+ entry	(74, 75)
Zfp36l1	Ser54	MAPK-activated protein kinase 2-mediated phosphorylation at Ser54 inhibits mRNA decay. Site is also phosphorylated by PKA	(76–78)

Rho guanine nucleotide exchange factor 2 (Arhgef2; also known as Lfc and GEF-H1) plays a role in the regulation of the state of actin polymerization (65, 66). Vasopressin increases phosphorylation at Ser⁸⁸⁵ (Fig. 1), likely mediated directly by PKA. It has been found that phosphorylation of Arhgef2 at Ser⁸⁸⁵ inhibits its GEF activity and its ability to activate RhoA (65, 66). Given the prior observations that vasopressin action is associated with the depolymerization of F-actin (18, 19, 21), thought to be critical to apical trafficking of AQP2, and that the depolymerization is in part dependent on RhoA (19, 79, 80), it appears plausible that phosphorylation of Arhgef2 at Ser⁸⁸⁵ plays an important causal role in F-actin depolymerization and AQP2 trafficking to the apical plasma membrane.

Exocyst complex component 7 (Exoc7) is an element of the exocyst complex that consists of Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, Exoc7, and Exoc8. Vasopressin decreases phosphorylation of Exoc7 at Ser²⁵⁰ (Fig. 1), a likely target of ERK1 or ERK2. The exocyst complex is associated with vesicle trafficking via post-Golgi vesicle targeting to the basolateral plasma membrane (81, 82). As a direct substrate of ERK1/2, phosphorylation of Exoc7 at Ser²⁵⁰ stimulates exocytosis to the basolateral plasma membrane by promoting assembly of the exocyst complex (67). Potentially, therefore, vasopressin’s action to reduce Ser²⁵⁰ phosphorylation could reduce exocytosis of AQP2 to the basolateral plasma membrane in lieu of the apical plasma membrane.

Histone deacetylase 4 (Hdac4) plays a role in transcriptional processes by regulating chromatin accessibility. Phosphorylation at Ser²⁴⁵ (along with two neighboring sites) is required for the binding of Hdac4 to 14-3-3 and sequestration in the cytoplasm (83). The sequestration results in the prevention of histone H3 deacetylation at Lys²⁷ (H3K27Ac), an activation chromatin mark (84). Vasopressin decreases phosphorylation of Hdac4 at Ser²⁴⁵ (Fig. 1), so it is predicted to translocate to the nucleus and deacetylate histone H3 at Lys²⁷ at the transcriptional start sites of target genes. Since histone H3K27Ac increases at the AQP2 promoter (85), this change is unlikely to be the consequence of decreased phosphorylation of Hdac4 at Ser²⁴⁵, which should cause the opposite effect. However, vasopressin action results in the repression of many genes (17, 86), which could potentially be the result of the observed phosphorylation change. The protein kinase responsible for phosphorylation of Ser²⁴⁵ of Hdac4 has been proposed to be Ca²⁺/calmodulin-dependent protein kinase (CaMK) (68, 87). In addition, polarity kinases [microtubule affinity regulating kinases (Mark2 and Mark3)] have also been shown to be capable of phosphorylating Ser²⁴⁵ in Hdac4 (69). Ser²⁴⁵ is found in a classical SNF1 family motif (L-X-(R/K)-X-X-p(S/T)-X-X-X-L), typical of microtubule affinity regulating kinases (88).

Itpr1 is an intracellular Ca²⁺ channel in the endoplasmic reticulum (ER) membrane, which mediates Ca²⁺ release from the ER (70). It has been reported that the Ca²⁺-releasing activity of Itpr1 is upregulated by PKA-mediated phosphorylation and substitution of two PKA consensus sequences (R-R-X-p[S/T]) at Ser¹⁵⁸⁸ and Ser¹⁷⁵⁵ to alanine inhibits Ca²⁺ mobilization (70, 71). As shown in Fig. 1, both of these sites display increases in phosphorylation in response to vasopressin, providing an explanation for the observation that vasopressin action through the V2 receptor increases the intracellular Ca²⁺ concentration (29–31). Working via calmodulin, vasopressin-dependent Ca²⁺ mobilization can activate certain protein kinases, such as Mylk (16, 20, 32) and calmodulin-dependent kinase IIδ, that are involved in the regulation of AQP2.

NSFL1 (p97) cofactor p47 (Nsfl1c) is a cofactor of valosin-containing protein (VCP; also known as p97), which has multiple functions in cellular processes, such as protein degradation and membrane fusion/trafficking (72, 89). Phosphorylation of Nsfl1c at Ser¹⁷⁶ is mediated by PKA and the phosphorylation enhances the activity of VCP/p47 in mouse neuronal cells (72). As shown in Fig. 1, vasopressin increases phosphorylation of Nsfl1c at Ser¹⁷⁶, possibly playing a role in vasopressin-mediated vesicle fusion and AQP2 trafficking to the plasma membrane.

The tyrosine-directed protein kinase Src (Src) is involved in diverse cellular signaling pathways. Vasopressin causes an increase in Src phosphorylation at Ser¹⁷ (Fig. 1), thought to be a PKA site (90). Src phosphorylation at Ser¹⁷ in the amino terminus facilitates its activity and activates the small G protein Rap1 (73, 90). In collecting duct cells, Rap1 activation inhibits AQP2 endocytosis, increasing retention of AQP2 in the apical plasma membrane (55). Thus, PKA-mediated phosphorylation of Src at Ser¹⁷ may cause AQP2 plasma membrane retention by activation of Rap1. Against this hypothesis is the observation that application of a tyrosine kinase inhibitor, dasatinib, is associated with increased accumulation of AQP2 in the apical region of IMCD cells (91). In addition, Rap1 plays an important role of inhibiting the ability of Ras to activate the ERK cascade (92), so vasopressin-mediated phosphorylation of Src at Ser¹⁷ may account for the ability of vasopressin to reduce ERK activity (24, 28). Overall, the conflicting observations described earlier indicate that more work will be required to fully understand the role of Src and other tyrosine kinases, including growth factor receptors, and should be a priority for future studies.

Stromal interaction molecule 1 (Stim1) is a key regulator of Ca²⁺ signaling, which activates store-operated Ca²⁺ entry by sensing Ca²⁺ (93). In response to Ca²⁺ depletion in the ER, Stim1 activates store-operated Ca²⁺ channels in the plasma membrane, such as Orai1, and promotes Ca²⁺ influx (94). Vasopressin decreases phosphorylation of Stim1 at Ser⁵⁷⁵ (Fig. 1), a likely target site for ERK1 or ERK2. Prior studies have shown that ERK1/2 phosphorylation of Stim1 at Ser⁵⁷⁵ facilitates Stim1 multimerization, which is a necessary prerequisite for its role in promoting Ca²⁺ entry (74, 75). Thus, the overall effect of vasopressin in the collecting duct is predicted to be inhibition of plasma membrane uptake of Ca²⁺.

mRNA decay activator protein ZFP36L1 (Zfp36l1) is an mRNA-destabilizing protein interacting with AU-rich elements (AREs) located in the 3′-untranslated regions of target mRNAs (95). Thus, Zfp36l1-dependent mRNA degradation is an important determinant of the abundances of target mRNAs that may exert control of gene expression in a transcription-independent manner. Vasopressin increases phosphorylation of Zfp36l1 at Ser⁵⁴, present in a motif compatible with phosphorylation by PKA (R-R-X-pS-V) (Fig. 1). Prior studies in other tissues have indeed shown that PKA (78) can phosphorylate Ser⁵⁴. Also, MAPK-activated protein kinase 2 (MAPKAPK2) is capable of phosphorylating Zpf36l1 at this site (76, 77). Phosphorylation at Ser⁵⁴ inhibits binding to the target mRNA and would be expected to increase target mRNA abundances. This action is unlikely to play a role in the vasopressin-mediated increase in Aqp2 mRNA (Table 1), given that the AQP2 transcript lacks an ARE motif in its 3′-untranslated region (17). However, a number of other vasopressin-regulated transcripts do contain ARE motifs (17) and are candidates for regulation by Zfp36l1-mediated mRNA degradation. Interestingly, Zfp36l1 is also regulated by vasopressin in another way. Cai et al. (96) found that the vasopressin analog dDAVP given for 7 days decreased mRNA abundance of Zfp36l1 protein in the mouse renal inner medulla to 66% of the control value. The relationship between the short-term phosphorylation response and the long-term protein abundance response is so far unexplored.

Time Course of Vasopressin-Induced Phosphorylation Changes

The data used to develop Fig. 1 were obtained after 30 min of vasopressin addition in both mpkCCD cells and the IMCD. To provide additional information about earlier time points, we cite a recent study that identified the time courses of phosphorylation changes in rat IMCD suspensions in response to the vasopressin analog dDAVP in the timeframe of 1–15 min (97). Figure 4 shows the time courses for phosphorylation sites identified in Fig. 1, categorized according to the molecular functions regulated by vasopressin. In general, the phosphorylation responses were complete within 2–5 min including the responses in the “vesicle-mediated transport” group of phosphoproteins. Typically, the water permeability response in isolated perfused collecting ducts displays an initial increase requiring ∼10 min followed by a slower secondary response lasting at least 30 additional min (11, 98). Thus, phosphorylation changes appear to be complete before water permeability changes. These data suggest that phosphorylation changes are not rate limiting for the water permeability response and that the dynamics may be due to secondary changes following the phosphorylation changes such as other posttranslational modifications.

Figure 4.

Dynamics of phosphorylation responses to vasopressin for different phosphorylation sites in native inner medullary collecting duct (IMCD) cells. Data are replotted from Leo et al. (97). A: sites associated with vesicle-mediated transport: Bin1 (Ser³⁰⁴), Exoc7 (Ser²⁵⁰), Nsfl1c (Ser¹⁷⁶), and Sec22b (Ser¹³⁷). B: sites associated with the regulation of GTPase activity/actin cytoskeleton/microtubule cytoskeleton: Agfg1 (Ser¹⁸¹), Arfgef1 (Ser¹⁵⁶⁶), Arhgef2 (Ser⁸⁸⁵), Specc1l (Ser³⁸⁵), and Rmdn3 (Ser⁴⁶). C: sites associated with the regulation of transcription: Ctnnb1 (Ser⁵⁵² and Thr⁵⁵¹), Lrrfip1 (Ser⁸⁸), and Tsc22d4 (Thr²²³). D: sites associated with protein phosphorylation: Camkk2 (Ser⁴⁹⁵ and Ser⁵¹¹), CDK18 (Ser⁶⁶), Erbb3 (Ser⁹⁸⁰), and Src (Ser¹⁷). E: sites associated with cAMP-dependent signaling/Ca²⁺ signaling: Prkar2a (Ser⁹⁶), Itpr1 (Ser¹⁵⁸⁸), Itpr3 (Ser¹⁸³²), and Stim1 (Ser⁵⁷⁵). F: sites associated with cell polarity: Igsf5 (Ser³³⁵), Luzp (Ser²⁶¹), and Slc9a3r1 (Ser²⁷⁵). Some sites shown in Fig. 1 did not yield time course data in Leo et al. (97). In general, phosphorylation responses were faster than observed water permeability responses, indicating that phosphorylation is not rate limiting for the overall response (see text).

Hypotheses about Roles of Vasopressin-Mediated Phosphorylation Changes in AQP2 Trafficking and Increases in Aqp2 Gene Expression

In Figs. 5, 6, and 7, we relate the phosphoproteins whose phosphorylation states are regulated by vasopressin to known physiological actions of vasopressin at a cellular level. These provide hypotheses for future studies of the mechanisms of AQP2 regulation in collecting duct principal cells.

Figure 5.

Vasopressin-regulated phosphoproteins with hypothetical roles in “vesicle-mediated transport.” Listed phosphoproteins are likely to be involved in three biological processes, “SNARE-mediated vesicle fusion,” “endosomal biogenesis,” and “interaction with dynamins,” which could regulate vasopressin-mediated aquaporin-2 (AQP2) trafficking. Cdk18, cyclin-dependent kinase 18; PKA, protein kinase A.

Figure 6.

Vasopressin-regulated phosphoproteins with hypothetical roles in the regulation of the “actin cytoskeleton,” “microtubule cytoskeleton,” and “GTPase activity.” The phosphorylation responses could potentially be involved in the regulation of aquaporin-2 (AQP2) trafficking. PKA, protein kinase A.

Figure 7.

Vasopressin-regulated phosphoproteins with hypothetical roles in the “regulation of transcription” and “RNA processing.” The phosphorylation responses could potentially be involved in the regulation of the abundance of aquaporin-2 (Aqp2) mRNA and secondarily AQP2 protein. PKA, protein kinase A.

Figure 5 shows the phosphoproteins that map to “vesicle-mediated transport” in Fig. 2. These can be categorized in three groups, viz. those that mediate “SNARE-mediated vesicle fusion,” “endosomal biogenesis,” and “interaction with dynamin.” Five of the phosphoproteins have recognized roles in SNARE-mediated vesicle fusion and therefore are candidates for roles in vasopressin-stimulated AQP2 exocytosis (11, 12, 14). Two of the phosphoproteins are known to interact with dynamin and therefore are candidates for roles in vasopressin’s action to inhibit endocytosis (11, 13). The other two phosphoproteins have roles in endosome biogenesis and therefore could be involved in AQP2 recycling or AQP2 degradation via late endosomes and lysosomes. Six of these proteins are likely direct targets of PKA, whereas two are likely targets of ERK1 or ERK2, both of which are downregulated through the action of PKA (24, 28). The remaining protein, amphiphysin-1, is a likely target of Cdk18 (or PCTAIRE-3). Cdk18 is activated by PKA-mediated phosphorylation at three sites, including Ser⁶⁶ (99). In addition, Cdk18 mRNA abundance is also upregulated in collecting duct mpkCCD cells in response to vasopressin (27, 86). In addition, Cdk18 protein may be regulated through control of its degradation, as it was found in a complex with the ubiquitin ligase Stub1 and regulatory subunits of PKA bound to AQP2 (100).

Figure 6 shows the phosphoproteins that map to three processes related to regulation of the cytoskeleton (Fig. 2), namely, “actin cytoskeleton,” “microtubule cytoskeleton,” and “regulation of GTPase activity.” All of these have relevance to AQP2 trafficking to and from the plasma membrane (Table 1). Five of these phosphoproteins are involved in the regulation of activities of Rho, Rac, and CDC42, small GTP-binding proteins that play central roles in the regulation of the actin cytoskeleton (101). RhoA in particular has been shown to play key roles in the regulation of AQP2 trafficking in collecting duct cells (19, 79, 80). Three phosphoproteins (with 4 phosphorylation sites) are directly involved in “actin cytoskeleton rearrangement,” which is a well-described cellular process in response to vasopressin (20, 22). Finally, two phosphoproteins (Clip1 and Rmdn3) are involved in “microtubule cytoskeleton organization.” The microtubule cytoskeleton has been shown to have roles in the determination of AQP2 localization within collecting duct cells (35, 102–106), and phosphorylation of Clip1 and Rmdn3 can be hypothesized to be involved in these processes.

Figure 7 shows the phosphoproteins that map to “regulation of transcription” and “RNA processing,” which together determine cellular levels of individual mRNAs (Fig. 2). One phosphoprotein, Zfp36l1, regulates mRNA degradation, as detailed earlier. Four phosphoproteins (5 phosphorylation sites) are transcription factors and transcriptional coregulators that are directly associated with control of RNA polymerase II-mediated transcription. Hypothetically, vasopressin-mediated phosphorylation changes in these proteins are involved in vasopressin-mediated regulation of AQP2 transcription (4, 17). Finally, an additional phosphoprotein, Hdac4, is involved in the regulation of DNA accessibility by control of histone H3 acetylation, as detailed earlier.

Perspectives and Significance

Vasopressin regulates renal water excretion through control of the AQP2 water channel in collecting duct cells. Studies of vasopressin-induced protein phosphorylation have focused mainly on the phosphorylation of AQP2. This study describes 44 phosphoproteins other than AQP2 that undergo vasopressin-mediated phosphorylation changes and summarizes potential physiological roles for each. Few of these sites have been studied in detail. We have laid out a bioinformatic analysis, highlighting potential roles for these sites, both in the regulation of AQP2 and in the mediation of other actions of vasopressin highlighted in Table 1. Although much has been gained through the systems/-omics approach, the full identification of physiological roles of each of the phosphorylation sites will likely require a reductionist/hypothesis-driven approach. New techniques of genome-editing (e.g., CRISPR/Cas9) have been developed, which will allow each of the 51 phosphorylation sites to be mutated, in either animals or cultured cells, to learn their roles in vasopressin-mediated regulation of collecting duct function.

DATA AVAILABILITY

Data access: https://esbl.nhlbi.nih.gov/Databases/AVP-Phos/ and https://esbl.nhlbi.nih.gov/AVP-Network/.

SUPPLEMENTAL DATA

Supplemental Spreadsheet S1: https://esbl.nhlbi.nih.gov/Databases/Supplemental-Phosphorylation/.

GRANTS

This work was funded by the Division of Intramural Research, National Heart, Lung, and Blood Institute Grants ZIAHL001285 and ZIAHL006129 (to M.A.K.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.A.K. conceived and designed research; E.P., C-R.Y., V.R., V.D., A.D., B.G.P., K.T.L., H.K., L.C., C-L.C., and M.A.K. analyzed data; E.P., C-R.Y., V.R., V.D., A.D., B.G.P., K.T.L., H.K., L.C., C-L.C., and M.A.K., interpreted results of experiments; E.P. and M.A.K. prepared figures; E.P., C-R.Y., V.R., L.C., C-L.C., and M.A.K. drafted manuscript; E.P., C-R.Y., V.R., V.D., A.D., B.G.P., K.T.L., H.K., L.C., C-L.C., and M.A.K. edited and revised manuscript; E.P., C-R.Y., V.R., V.D., A.D., B.G.P., K.T.L., H.K., L.C., C-L.C., and M.A.K. approved final version of manuscript.