The voltage sensing phosphatase (VSP) localizes to the apical membrane of kidney tubule epithelial cells

Voltage-sensing phosphatases (VSPs) are transmembrane proteins that couple changes in membrane potential to hydrolysis of inositol signaling lipids. VSPs catalyze the dephosphorylation of phosphatidylinositol phosphates (PIPs) that regulate diverse aspects of cell membrane physiology including cell division, growth and migration. VSPs are highly conserved among chordates, and their RNA transcripts have been detected in the adult and embryonic stages of frogs, fish, chickens, mice and humans. However, the subcellular localization and biological function of VSP remains unknown. Using reverse transcriptase-PCR (RT-PCR), we show that both Xenopus laevis VSPs (Xl-VSP1 and Xl-VSP2) mRNAs are expressed in early embryos, suggesting that both Xl-VSPs are involved in early tadpole development. To understand which embryonic tissues express Xl-VSP mRNA, we used in situ hybridization (ISH) and found Xl-VSP mRNA in both the brain and kidney of NF stage 32–36 embryos. By Western blot analysis with a VSP antibody, we show increasing levels of Xl-VSP protein in the developing embryo, and by immunohistochemistry (IHC), we demonstrate that Xl-VSP protein is specifically localized to the apical membrane of both embryonic and adult kidney tubules. We further characterized the catalytic activity of both Xl-VSP homologs and found that while Xl-VSP1 catalyzes 3- and 5-phosphate removal, Xl-VSP2 is a less efficient 3-phosphatase with different substrate specificity. Our results suggest that Xl-VSP1 and Xl-VSP2 serve different functional roles and that VSPs are an integral component of voltage-dependent PIP signaling pathways during vertebrate kidney tubule development and function.


Introduction
Phosphatidylinositol phosphates (PIPs) are lipid second messengers involved in almost all facets of cell biology, including differentiation, proliferation, migration, and polarity [1,2]. Many human diseases are linked to mutations in PIP-modifying enzymes, including cancer, peripheral neuropathy, stroke, bipolar disorder, autism and developmental disorders [3][4][5][6]. As a result, PIP kinases and phosphatases have been extensively studied to understand their roles in a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 these diverse cellular processes and are increasingly viewed as potential therapeutic targets [7][8][9]. Here, we focus on an exceptional member of the PIP-modifying family of enzymes, the voltage-sensing phosphatase (VSP).
VSP is a unique PIP-modifying enzyme whose activity is regulated through a voltage sensing domain (VSD) [10]. VSDs are composed of four transmembrane helices with the fourth helix (called S4) containing arginines responsible for sensing changes in the electrical membrane potential. In response to membrane depolarizations, the S4 helix changes conformation leading to activation of the cytosolic phosphatase domain (PD), which is homologous to PTEN (phosphatase and tensin homolog deleted on chromosome 10) [10], a well characterized lipid phosphatase. Once VSP is activated, its PD dephosphorylates both the 5-and 3-phosphates from PIPs [10][11][12][13][14]. As a result, VSPs provide a direct connection between the electrical signaling and PIP signaling pathways. While electrical signaling pathways are most often discussed in terms of neurons, all cells maintain an ionic gradient that creates a membrane potential. This electrochemical force is utilized to initiate essential cell-signaling functions. For example, pancreatic beta cells use their membrane potential to respond to increasing glucose concentrations by activating ATP-sensitive potassium channels and voltage-gated calcium channels to release insulin [15]. Renal tubules also use membrane potentials, specifically activating different potassium channels for regulating cell volume, potassium secretion and tubuloglomerular feedback [16]. Since many ion channels are regulated by PIPs [17][18][19], VSPs are likely to play a role in modulating ion channel permeability and their subsequent electrical signaling when channels and VSP are co-expressed.
In line with the idea that VSP serves a fundamental physiological role, VSP orthologs have been identified in several tissues across diverse phyla, including ascidians, newts, salamanders, zebrafish, chicken, amphibians, mice, and humans [10,11,[20][21][22][23][24][25][26]. In particular, VSP mRNA transcripts have been found in the adult nervous systems of sea squirts and the brains of frogs, mice and humans [20,22,27,28] as well as in kidney, stomach, heart, testis and ovary [20,[22][23][24][25][27][28][29][30]. In addition to adult tissues, VSP transcripts have been detected in embryonic tissues such as the kidney and eye of zebrafish [31], the kidney, brain and stomach of chicks [21,30], and the brain, spinal cord and eye of mice [25,27,32]. The majority of studies investigating VSP localization focused on measuring mRNA expression, with a few studies reporting cellular localization of VSP protein. The first report of the sea squirt VSP (Ciona intestinalis VSP, Ci-VSP) found protein expression in sperm tails [10]. In chicks, Gallus gallus VSP (Gg-VSP) protein was found in Purkinje neurons throughout the cell, which the authors suggested means Gg-VSP is expressed on the plasma membrane as well as intracellular membranes [30]. Lastly, mouse VSP (Mm-VSP) was found in the brains of adult and neonatal mice, specifically in dissociated cortical neurons [32]. However, to our knowledge, a connection between VSP activity and PIP-mediated biological processes in native tissues remains unknown.
To better understand VSP's biological role, we tested for Xl-VSP localization in Xenopus laevis embryos. X. laevis are an allotetraploid species and have two highly similar VSP proteins, termed Xl-VSP1 and Xl-VSP2 [22,33]. We found the RNA transcripts for both in multiple stages of embryonic development. The transcript expression patterns differed between the two homologs, suggesting that they fulfill different biological roles. We validated a new VSP antibody, N432/21 (NeuroMab), which recognizes both Xl-VSPs and shows that Xl-VSP protein is expressed in the brain and kidney of adult and embryonic X. laevis. Furthermore, we found that Xl-VSP protein is specifically located on the apical membrane of both embryonic and adult kidney tubules. To probe the function of Xl-VSP in these tubules, we tested the catalytic activity of both Xl-VSP homologs. We found that they dephosphorylate the 5-phosphate from PIPs, but unlike other VSPs, Xl-VSP2 is a significantly weaker 3-phosphatase. From our results, we suggest that by dephosphorylating PIPs in a voltage-dependent manner, Xl-VSPs play a fundamental role in kidney development and function.

Xl-VSP mRNA transcripts in X. laevis embryos
The location and function of a biological molecule are often tightly coupled characteristics. Previous investigations showed adult X. laevis transcript expression levels in several tissues including testes, kidney, ovary, liver, and brain [22]. Although previous studies found VSP transcripts in the embryos of ascidians [28], fish [31], chicken [21,30] and mice [27], expression in X. laevis embryos has not been reported. To better understand Xl-VSP's physiological role, we determined the RNA transcript expression of Xl-VSP1 and 2 during X. laevis development. We made polyadenylated cDNAs from total RNA from Nieuwkoop and Faber (NF) [34] stage 12-40 embryos and conducted semi-quantitative reverse transcriptase PCR (sqRT-PCR) using Xl-VSP1 and Xl-VSP2 specific PCR primers. We found that both Xl-VSP1 and 2 transcripts are expressed in the developing embryo and accumulate by swimming tadpole stages (NF stages >36; Fig 1A and 1B). Furthermore, we observed a high level of Xl-VSP1 transcripts in early embryos, which are likely remnants of maternal transcripts in agreement with previous finding of high level expression of Xl-VSP1 in the ovary (Fig 1A) [22]. Our results indicate that both Xl-VSP1 and Xl-VSP2 transcripts are present at early stages of X. laevis development and that the two show different patterns of embryonic RNA expression. These results agree with previous findings that the transcripts are differentially expressed in adult tissues [22] and further suggest discrete functional roles for each in developing embryos.
We next localized Xl-VSP mRNA in embryos by in situ hybridization (ISH). We tested NF stage 32-36 X. laevis embryos by ISH with an antisense-strand probe against full-length Xl-VSP1 mRNA since our sqRT-PCR analysis showed accumulation of mRNA expression at those stages. In agreement with our sqRT-PCR results, we found RNA transcripts in the developing pronephros, brain and brachial arches (Fig 2A), whereas no staining was seen when using a sense-strand control probe ( Fig 2B). Since Xl-VSP1 and Xl-VSP2 mRNAs share 93% identity on the nucleotide level, it is possible this staining is due to transcripts from either Xl-VSP1, Xl-VSP2, or both.   using PCR primers specific for Xl-VSP1 (A) and Xl-VSP2 (B). Both Xl-VSP1 and Xl-VSP2 transcripts appear to accumulate by stage 36. No bands were seen without reverse transcriptase (bottom). All cDNAs were made with equal amounts of total RNA as determined by spectrophotometry and confirmed by agarose gel electrophoresis to visualize ribosomal RNA bands (S1 Fig). sqRT-PCR was repeated at least two times with at least two different embryonic cohorts. The expected PCR amplicon sizes are 389 bp for Xl-VSP1 and 478 bp for Xl-VSP2. Shown are representative gels.

Validating a VSP antibody
To determine Xl-VSP protein expression in tissues, we validated a mouse monoclonal VSP antibody, N432/21, developed by the NeuroMab facility at the University of California, Davis. We used X. laevis oocytes as a heterologous expression system to determine the specificity of the anti-VSP. Oocytes were injected with 20 ng of VSP RNAs from different species, incubated for 36 hours before being lysed, and the lysates were run on SDS-PAGE gels for Western blotting. The anti-VSP N432/21 recognized several species of VSP, including Xl-VSP1 (58 kDa), Xl-VSP2 (58 kDa), Xenopus tropicalis VSP (Xt-VSP, 58 kDa), FLAG-tagged Ciona intestinalis VSP (Ci-VSP, 66 kDa), and Danio rerio VSP (Dr-VSP, 58 kDa) ( Fig 3A). Xl-VSPs heterologously-expressed in X. laevis oocytes often displayed a double band at their predicted molecular weight (MW) of 58 kDa, with the Xl-VSP2 migrating slightly slower through the gel than Xl-VSP1 and Xt-VSP (Fig 3A-3C). The nature of these doublets and the slight difference in electrophoretic mobility between Xl-VSP1, Xl-VSP2, Xt-VSP, and Dr-VSP (despite their nearly identical predicted MWs) is not presently understood. To further validate the VSP antibody, we injected either a control morpholino or a translation-blocking morpholino targeting the Xl-VSP2 transcript into X. laevis zygotes to knockdown Xl-VSP2 expression. At NF stage 42 the embryos were lysed and processed for Western blotting. The un-injected and control morpholino-injected embryos show a strong band at the expected MW for Xl-VSPs while the Xl2 morpholino-injected embryos displayed no band, further demonstrating that the N432/21 antibody is specific for VSP ( Fig 3B). Despite several attempts, a morpholino against Xl-VSP1 could not be validated. These results demonstrate that anti-VSP N432/21 recognizes both Xl-VSP1 and 2 proteins, as well as the most commonly-studied VSP (Ci-VSP) and VSPs from D. rerio and X. tropicalis, which are two other common vertebrate models of development.

Western blots of Xl-VSP protein in embryos and adult tissues
After validating the VSP antibody, we turned to tissue samples from adult and embryonic X. laevis. Lysates made from adult kidney, testis, and brain show a single band of the correct approximate MW (58 kDa) by Western blot, indicating that these tissues express Xl-VSPs ( Fig  3C). The anti-VSP N432/21 recognizes both Xl-VSP1 and 2 so we cannot definitively distinguish whether Xl-VSP1, Xl-VSP2, or both are present. However, two pieces of evidence suggest that, at least for the kidney, Xl-VSP2 may predominate. First, our previous work showed a larger Xl-VSP2 mRNA concentration in the adult kidney compared to Xl-VSP1 mRNA [22]. In addition, our Western blots of over-expressed Xl-VSP1 and 2 show a slight difference in their electrophoretic mobility, and the band from the kidney sample aligns with the slightly and Xl-VSP2 (as seen in panels A and C) has not been determined. These results show that anti-VSP N432/21 is specific for VSP and cross-reacts with VSPs from multiple species. (B) X. laevis zygotes were injected with either a control morpholino, an Xl-VSP2 morpholino or left un-injected. (top) Western blot analysis of equal amounts of lysates from embryos at stage 42 shows a band at the predicted MW for Xl-VSPs (58 kDa) for un-injected and control embryos and no band for the Xl2 morpholino-injected embryos. (bottom) Blots were stripped and re-probed with anti-actin for a loading control. (C) Western blot analysis of X. laevis tissues. Lysates from adult kidney (K, 3 μg), testis (T, 30 μg), and brain (B, 10 μg) were run against lysates from oocytes injected with RNAs for Xl-VSP1 (Xl1) and Xl-VSP2 (Xl2) and analyzed by Western blot with anti-VSP. A single band of approximately the correct MW (58 kDa) was observed in the tissue lysates, indicating the presence of Xl-VSP protein in all tissues tested. (D) Western blot analysis of X. laevis embryos. Lysates from NF stage 12-40 embryos (30 μg each) were run against a lysate from adult kidney (K, 5 μg). Blots were probed either with anti-VSP (top) or anti-actin (bottom) as a loading control (predicted MW 42 kDa). A weak band (potentially corresponding to Xl-VSP1) is present only at early embryonic stages 12-20 (red arrowheads), whereas a slightly slower-migrating band (potentially corresponding to Xl-VSP2) accumulates at later embryonic stages 36-40. The difference between the kidney and embryo lysate MWs for both VSP and actin may reflect a different degree of post-translational modification in embryos versus adult X. laevis. Lysates, gels and blots were repeated three times with either three different adults or three different embryonic cohorts. Shown are representative gels for each. slower mobility of Xl-VSP2 ( Fig 3C). Combining these two lines of evidence, we suggest that Xl-VSP2 expression may predominate in the kidney while Xl-VSP1 expression may predominate in the testes, further supporting the likelihood of different roles for the two Xl-VSP proteins.
Next, we tested for Xl-VSP protein expression in X. laevis embryos, NF stages 12-40. Western blots show a faint band at the early embryonic stages, NF 12-20, and darker bands at the later swimming tadpole stages, NF 36-40 ( Fig 3D) corresponding to the correct approximate MW of 58 kDa. Based on the slight electrophoretic differences observed in the heterologous system, the slightly lower apparent MW faint bands may represent Xl-VSP1. The darker bands have a slightly higher apparent MW, suggesting they represent Xl-VSP2. This designation is further supported by our morpholino data ( Fig 3B) where a morpholino designed against Xl-VSP2 dramatically reduced the visible protein band from stage 42 embryos. Interestingly, these results suggest that Xl-VSP2 is up-regulated as the tadpole develops. This pattern of Xl-VSP protein levels agrees with the high level of Xl-VSP1 maternal transcripts in the ovary [22] and up-regulation of Xl-VSP2 RNA during later developmental stages ( Fig 1B). Furthermore, from these blots, we suggest post-transcriptional regulation of Xl-VSP protein expression in the developing embryo since Xl-VSP1 transcripts also are present in later embryonic stages yet are not detectable by Western blot (Fig 3D). Overall, Xl-VSP protein is present in both adult and embryonic X. laevis. The differential protein expression between Xl-VSP1 and Xl-VSP2 is consistent with the differences observed at the RNA level and further suggests the two homologs serve different functional roles.

Immunohistochemistry of Xl-VSP proteins in kidney and brain
Since protein accumulation was not detected by immunoblotting until later embryonic stages, we next used whole-mount indirect IHC on NF stage 42 embryos. Because we observed high RNA and protein levels in the kidneys, we initially focused on kidney expression for the IHC experiments. To facilitate kidney identification, we used two transgenic lines that express GFP in the developing kidney, either Pax8:GFP [35] or Cdh17:GFP [36]. Whole-mount immunostained Pax8:GFP embryos showed distinct Xl-VSP staining in the proximal pronephros, indicating Xl-VSP protein is present during kidney development (S2A-S2C and S2A'-S2C' Fig). This staining was not seen in sibling embryos stained with the secondary antibody alone (S2A"-S2C" Fig). Focusing on the pronephritic tubules, we fixed, embedded, and sectioned NF stage 42 Cdh17:GFP embryos for IHC to determine subcellular localization of Xl-VSPs ( Fig 4A-4F). Interestingly, we observed Xl-VSP staining in the developing pronephritic tubules, specifically localized to the lumenal surface or apical membrane of the tubules ( Fig  4A). This is in contrast to the soluble GFP marker that is visible throughout the tubules ( Fig  4B). Consecutive sections probed with the secondary antibody alone show the GFP marker without any Xl-VSP staining ( Fig 4A'-4F'). Our findings demonstrate embryonic expression of Xl-VSP protein in pronephritic tubules with subcellular localization on their apical membranes, thus, suggesting Xl-VSPs play an important role in the development and function of the embryonic X. laevis kidney.
In addition to embryos, we also tested adult kidney tissue using IHC. Adult male X. laevis kidneys were removed, fixed, embedded, sectioned, and stained with anti-VSP. The sections also show Xl-VSP staining on the lumenal surface of the tubules (Fig 5A), which is consistent with the subcellular staining observed in embryonic kidney tubules ( Fig 4A). No staining was seen in adult kidney tubules when consecutive sections were probed with secondary antibody alone ( Fig 5B). In contrast to our IHC of embryonic kidney sections, we only observed staining in adult kidney sections by anti-VSP after post-fixation antigen retrieval, which suggests a These results suggest a long-term functional role for Xl-VSPs in the kidney since they are present from the earliest stages of kidney development and continue into adulthood.

Functional activity of Xl-VSPs
While VSPs from Ciona intestinalis and Danio rerio function as both 5-and 3-phosphatases [10][11][12]14,37], the Xl-VSPs were previously characterized as only a 5-phosphatase against phosphatidylinositol-4,5-bisphosphate [PI(4,5)P 2 ] [22]. To understand how Xl-VSPs may contribute to kidney function, we examined their phosphatase activity to determine whether Xl-VSP1 and 2 are able to dephosphorylate both the 3-phosphate and the 5-phosphate from PIPs. Because Xl-VSPs are regulated by voltage, we used two electrode voltage clamp (TEVC) to control Xl-VSP activation. To monitor all four possible dephosphorylation reactions ( Fig  6A), we turned to pleckstrin homology (PH) domain-based optical biosensors. PH domains are well known protein domains that bind specifically to certain PIPs [38][39][40]. We chose PH domains from tandem PH domain containing protein 1 (TAPP) and phospholipase C (PLC) because they bind to phosphatidylinositol-3,4-bisphosphate [PI(3,4)P 2 ] and PI(4,5)P 2 , respectively, allowing us to monitor both 5-and 3-phosphatase activities [12,14,41,42]. We monitored the reactions that create and deplete PI(3,4)P 2 using the TAPP-PH domain (upper reaction scheme, Fig 6A) and the reactions that create and deplete PI(4,5)P 2 using the PLC-PH domain (lower reaction scheme, Fig 6A). However, all four reactions occur during each voltage step. For example, to follow PI(3,4)P 2 , we used the Förster resonance energy transfer (FRET)based biosensor fTAPP that utilizes the TAPP-PH domain flanked by an N-terminal CFP and C-terminal YFP. The whole sensor is anchored to the membrane through a prenylation site at the C-terminus ( Fig 6B) [41][42][43]. When VSP dephosphorylates the 5-phosphate from phosphatidylinositol-3,4,5-trisphosphate [PI(3,4,5)P 3 ], the TAPP-PH domain of the fTAPP biosensor will bind the PI(3,4)P 2 and the resulting conformational change between the CFP and YFP will change the FRET between them, leading to an increase in the FRET ratio ( Fig 6B). As PI (3,4)P 2 is depleted when VSP dephosphorylates the 3-phosphate, the TAPP-PH unbinds and the FRET ratio decreases (Fig 6B). Similarly, biosensors using the PLC-PH domain will bind to PI(4,5)P 2, allowing us to monitor the PI(3,4,5)P 3 to PI(4,5)P 2 and PI(4,5)P 2 to PI(4)P reactions. Thus, we combine the optical measurements of the biosensors with TEVC to precisely control and monitor catalytic activity of Xl-VSPs.
When oocytes co-expressed Xl-VSP1 and fTAPP, we observed Xl-VSP1 inducing an initial increase and subsequent decrease in FRET over time, indicating that it functions as both a 5and a 3-phosphatase (Fig 6C, left). To determine the voltage dependence of activity, we tested several voltages and plotted the FRET versus voltage relationships. To better understand each reaction, we separated the FRET increase from the FRET decrease during our analysis (Fig 6C, right, blue). Note the voltage dependence of the Xl-VSP1 5-phosphatase reaction (FRET increase, V 1/2 = 55 mV) is lower than the voltage dependence of the 3-phosphatase (FRET decrease, V 1/2 = 97 mV) (S1 Table). As a control, we also tested a catalytically inactive Xl-VSP2 C301S (Xl2-CS) [22] with fTAPP. As expected, a slight, continuous FRET increase was observed. This small signal is expected because it has been previously shown that a low level of endogenous Xl-VSP activity is present in X. laevis oocytes [14,22,41,42,44]. Interestingly, the same fTAPP experiment with Xl-VSP2 only showed a FRET increase (Fig 6C, green). Our previous investigations into Ci-VSP showed a change in apparent activity when expression was low which could be explained by the contribution of competing endogenous enzymes [42,45].
Using Western blots, we tested individual cells used for the fTAPP activity assays and found similar expression levels between the Xl-VSP1 and Xl-VSP2 proteins when staining with anti-VSP. The similar expression levels between Xl-VSP 1 and 2 indicate that the observed activity change is an inherent property of Xl-VSP2 and not due to lower protein levels (S3 Fig). Additionally, as was previously reported [22], the voltage dependence of Xl-VSP2 5-phosphatase activity against PI(3,4,5)P 3 is shifted to higher voltages (V 1/2 = 92 mV) compared to that of Xl-VSP1 (V 1/2 = 55 mV) (S1 Table). These results indicate that, although the Xl-VSP homologs are 93% identical, they do not share identical voltage-dependent functional roles in the cell.
To follow PI(4,5)P 2 , we used two different biosensors based on the PLC-PH domain, one FRET sensor which utilized the PH domain from PLC instead of TAPP to create fPLC, and a fluorescence sensor where GFP is attached to the N-terminus of PLC, called gPLC. We used fPLC to monitor the 5-phosphatase activity during the PI(4,5)P 2 to PI(4)P reaction. When oocytes co-expressed either Xl-VSP1 or Xl-VSP2 with fPLC, we observed robust FRET ratio decreases over time (Fig 6D, left), indicating 5-phosphatase activity for both homologs, as was previously published [22]. Interestingly, the kinetics of Xl-VSP2 are significantly slower than the kinetics of Xl-VSP1, as seen by the voltage steps to 160 mV for a 2 second voltage step for Xl-VSP1 and a 10 second voltage step for Xl-VSP2 (Fig 6D, left). The FRET ratio versus voltage relationship shows a similar voltage-dependence for both Xl-VSP1 and 2 (Fig 6D, right; S1 Table). The voltage-dependence of the PI(4,5)P 2 to PI(4)P reaction we report here is shifted compared to the previous published results ( [22]). Two key differences with those previous results are that we pre-treated the oocytes with insulin and used a different biosensor, fPLC. The shift in voltage dependence is likely due to the biosensor because fPLC has very different response characteristics, including a faster response time and shifted voltage dependence [41]. We used gPLC to monitor the PI(3,4,5)P 3 to PI(4,5)P 2 reaction, because the gPLC biosensor is more sensitive for this reaction than fPLC. Instead of a change in FRET, the gPLC translocates to the membrane when PI(4,5)P 2 concentrations increase, resulting in an increase in fluorescence. When oocytes co-expressed Xl-VSP1 with gPLC, we observed a strong increase in fluorescence over time, indicating the production of PI(4,5)P 2 from the Xl-VSP1 3-phosphatase reaction (Fig 6E, left, blue). Unexpectedly, the same experiment with Xl-VSP2 also gave a slow and small, but reproducible fluorescence increase (Fig 6E, left, green), indicating Xl-VSP2 is also able to function as a 3-phosphatase. However, we did not observe any 3-phosphatase activity when using the fTAPP sensor. When plotting the FRET ratio vs voltage relationship for the PI(3,4,5)P 3 to PI(4,5)P 2 reaction, the Xl-VSP2 is significantly shifted to higher voltages compared to Xl-VSP1 (Fig 6E, right; S1 Table). These results show that Xl-VSP2 can remove the 3-phosphate from PI(3,4,5)P 3 but not from PI(3,4)P 2 . Further investigations are needed to better understand this striking difference in substrate specificity. Overall, our results show that Xl-VSP1 can catalyze the same four reactions as other VSPs while Xl-VSP2 is restricted by substrate. Since the reactions and kinetics appear to be different between Xl-VSP1 and 2, they likely fulfill different functional roles in tissues which correlates with their differential expression patterns.

Discussion
VSPs provide a direct link between the electrical state of a cell and its PIP signaling pathways. By dephosphorylating PIPs in a voltage dependent manner, VSPs could regulate PIP-dependent processes such as cortical cytoskeleton remodeling, ion channel function, G protein-coupled receptor signaling (GPCR) and intracellular calcium signaling [18,19,46]. Though the biophysical properties of VSPs have been the subject of intense study since the discovery of Ci-VSP [10], the physiological relevance of this direct link remains unclear.
To elucidate Xl-VSPs' biological role, we determined the cellular expression pattern of Xl-VSP RNA and protein in Xenopus laevis. We extended our previous findings in adult X. laevis tissues [22] by showing both Xl-VSP1 and 2 RNA transcripts in developing embryos. Though VSP is seen in the embryonic stages of other species, this is the first study showing Xl-VSP RNA and protein in X. laevis embryos. Our findings indicate that Xl-VSP RNAs are present at activation are significantly faster for Xl-VSP1 than for Xl-VSP2. (right) GFP fluorescence measurements were tested at several voltages and the fluorescence voltage relationship plotted showing a fluorescence increase (net 3-phosphatase reaction, dashed line, open symbols). While Xl-VSP1 shows significant levels of 3-phosphatase activity against PI(3,4,5) P 3 , Xl-VSP2 is a much less efficient 3-phosphatase. All error bars are ± SEM., n � 9. Data fit with single Boltzmann equations. For each experiment, the oocyte was incubated in 8 μM insulin to increase PI(3,4,5)P 3 concentrations. early stages of development, NF stages 12-40, with both Xl-VSP1 and 2 RNA levels increasing from NF stage 20 to 40. At the same stages, we observed an increase in protein levels by Western blot, consistent with our RNA detection, indicating an upregulation of the protein during development. Using ISH, we observed Xl-VSP RNA transcripts in both the brain and pronephros of NF stage 32-36 embryos. Using IHC, we found Xl-VSP protein in the pronephroi of NF stage 42 embryos. Upon sectioning the embryos, we observed anti-VSP staining on the lumenal surface of the pronephritic tubules, corresponding to the apical membrane of the epithelial cells lining the tubule lumen. Additionally, we sectioned adult kidneys and found the same apical membrane localization in the tubules. To directly test Xl-VSP catalytic activation, we combined electrophysiology with optical biosensors to show that both Xl-VSPs dephosphorylate the 3-and 5-phosphate from PIPs though with different substrate specificities and efficiencies. Combining our localization results with our electrophysiological activity data, we suggest that Xl-VSPs have specific functional roles in both the brain and kidney.
Concentrating on the kidney, we found Xl-VSPs spanning the earliest stages of the developing pronephros into the full adult kidney, suggesting Xl-VSPs modulate PIP concentrations on the apical membrane of the kidney epithelial cells in a voltage dependent manner. Kidney tubules are lined with specialized epithelial cells that regulate vertebrate solute homeostasis [47,48]. The transepithelial potential difference across these cells contributes to tubular reabsorption of solutes and water [49,50]. The lumenal surface of these epithelial cells are decorated with microvilli and primary cilia, both of which function in tubular reabsorption by sensing fluid flow [51][52][53][54][55]. Microvilli provide a large absorptive membrane surface for kidney epithelial cells and are actin-dependent dynamic structures with turnover rates in the tens of minutes [56]. Heterologously expressed Gg-VSP influences actin-based cytoskeleton rearrangements in cultured chick fibroblasts leading to morphological changes in cortical cell structure [57], suggesting Xl-VSPs could modify the actin-based microvilli in kidney epithelial cells. In addition, GPCR signaling originating in microvilli at the apical membrane of proximal kidney tubules is directly influenced by local PI(4,5)P 2 concentrations [58], which is a substrate for both Xl-VSPs. Likewise, primary cilia are dependent on local PI(4,5)P 2 concentrations [59], suggesting Xl-VSPs could be involved in ciliary functions during renal tubule development and adult renal function.
Our results are consistent with VSP homologs having a broad role in multiple tissues. Specifically, our study agrees with previous studies that also showed embryonic kidney tubule expression of VSP in zebrafish [31], and chicks [21]. In particular, Gg-VSP RNA was found on the proximal part of the nephrogenic tubules of HH stage 26 chicks and not the distal or collecting tubules, consistent with our IHC results showing proximal tubule localization in NF stage 32 embryos (S2 Fig). Our findings in X. laevis are also consistent with the findings in chicks in that not all tubules appear to immunostain with our VSP antibody, indicating that the Xl-VSPs may be restricted to a functional subset of the developing and adult kidney. Our results go further because we observe clear Xl-VSP protein immunostaining on the apical membrane of the tubules where reabsorption of nutrients occurs and where PIP-mediated signaling originates.
Not all VSP studies analyzed kidney expression patterns. The human VSP, Hs-VSP1 (previously called TPIP and TPTE2), was not examined in kidneys, making human VSP kidney expression unknown [20]. In addition to kidney expression, VSP is also expressed in the brain and nervous system of multiple species [20,22,28,30,32]. Indeed, we observed Xl-VSP RNA transcripts and protein in the brain of adult and embryonic X. laevis. Some reports of VSP expression conflict, particularly regarding expression of mouse VSP, Mm-VSP (previously called PTEN2). Early studies identified Mm-VSP as a testis specific protein [25] while subsequent investigations report expression in the brain [32]. While more research is needed, a pattern is emerging with the kidney and brain being the most consistent tissues expressing VSP. Our results support this pattern and a functional role for VSP in both the brain and kidney of multiple vertebrate species.
The subcellular localization of VSP is also debated in the field. Prior studies of native VSP protein have found VSP on the plasma membrane of ascidian sperm [10] and of Purkinje neurons in chick cerebellum sections [30]. Along with plasma membrane localization, however, the cerebellum sections also showed intracellular staining with a Gg-VSP antibody. A more recent study in mice used dissociated cortical neurons from P0-1 mice and the staining appears throughout the cells [32]. In addition, heterologous expression of mouse and human VSP consistently show intracellular membrane localizations [20,25,29]. In contrast, our subcellular localization of endogenous Xl-VSP protein is on the apical membrane of kidney tubule epithelia. The differences in our results and those from prior studies may lie in the tissues tested, the antigen retrieval discussed below, or in the antibodies used. Further experiments are needed to determine whether VSPs function on both intracellular and plasma membranes. While not tested against either chicken or mouse VSPs, our antibody does cross react with VSPs from D. rerio, X. tropicalis and C. intestinalis. As a result, N432/21 may prove to be an invaluable tool in future studies of VSP's physiological role in vertebrate kidney development and neural function.
It is interesting to note that unlike in fixed embryonic kidney tissue, we only observed staining of adult kidney tissue after post-fixation antigen retrieval. This difference in VSP epitope availability could be due to different roles for VSP in adult versus embryonic kidney tubules, and differential VSP binding partners in these tissues. We recently demonstrated that Ci-VSP forms dimers [42]. Since the two Xl-VSP homologs shown here display different expression in the embryo and adult frog, it is also possible that the difference we observed in VSP epitope availability is due to the presence of Xl-VSP homodimers and heterodimers.
In addition to our localization experiments, we also tested the electrophysiological characteristics of both Xl-VSP homologs. We controlled their activation using TEVC and monitored the production and depletion of different PIPs using optical biosensors. Our previous report indicated that both homologs catalyzed the dephosphorylation of the 5-phosphate from PI (4,5)P 2 . VSPs from other species have been shown to utilize PI(3,4,5)P 3 and PI(3,4)P 2 as substrates as well. Here, we tested all three substrates by using PIP sensors that are sensitive to either PI(3,4)P 2 or PI(4,5)P 2 concentrations (TAPP-PH or PLC-PH respectively), allowing us to monitor the 3-phosphate removal from PI(3,4,5)P 3 and PI(3,4)P 2 as well as the 5-phosphate removal from PI(3,4,5)P 3 and PI(4,5)P 2 . We found that Xl-VSP1 behaves much like other VSPs in that it dephosphorylates both 5-and 3-phosphates. Xl-VSP2, on the other hand, is a slower 5-phosphatase and a weaker 3-phosphatase. After modifying our protocols for longer depolarizations to account for the slower kinetics, we observed clear 5-phosphatase activity for Xl-VSP2 using both fTAPP and fPLC sensors. The voltage-dependence of activation was shifted to higher voltages for Xl-VSP2 compared to Xl-VSP1. We also observed 3-phosphatase activity of Xl-VSP2 against PI(3,4,5)P 3 , but not against PI(3,4)P 2 , even after a one minute depolarization. This difference between the Xl-VSP homologs further indicates that the two are not interchangeable and serve different physiological roles in the cell.
In conclusion, we observe Xl-VSP expression in the brain and kidney of X. laevis embryos and adults. We localize Xl-VSP protein on the apical membrane of kidney tubule epithelial cells, and we characterize both Xl-VSPs as 3-and 5-phosphatases with different enzymatic activities. Our results indicate a role for voltage-dependent dephosphorylation of both 5-and 3-phosphates from PIPs on the apical membrane during kidney tubule development and renal function.

Materials
Ciona intestinalis VSP (Ci-VSP, NM_001033826) and Danio rerio VSP (Dr-VSP, NM_001025458) plasmids were kindly provided by Y. Okamura (Osaka University, Osaka, Japan). X. laevis VSP1 (Xl-VSP1, NM_001096603), X. laevis VSP2 (Xl-VSP2, NM_001280607), and X. tropicalis VSP (Xt-VSP, NM_001015951) expression plasmids were made in Laurinda Jaffe's laboratory and are available through Addgene (Cambridge, MA). While the original human gene was named transmembrane phosphatase with tensin homology (TPTE), the majority of the literature refer to this protein as the voltage sensing phosphatase or VSP. Here, we follow this more common nomenclature to avoid confusion. fPLC and fTAPP were both kindly provided by E.Y. Isacoff (University of California, Berkeley). GFP-PLC-PH was kindly provided by T. Meyer (Stanford University). Mutations and epitope tags were created by site-directed mutagenesis with Pfu Turbo DNA polymerase (Agilent, Santa Clara, CA) by standard protocols. All DNA constructs were confirmed by sequencing. cRNA was transcribed using SP6 or T7 mMessage mMachine kits (Thermo-Fisher, Waltham, MA). X. laevis oocytes for Xl-VSP activity assays were purchased from Xenopus 1 (Dexter, MI). Mouse monoclonal anti-VSP (N432/21; RRID:AB_2716253) was made by the UC Davis/NIH Neu-roMab Facility and is available through Antibodies Incorporated (Davis, CA). Anti-beta actin (clone C4, catalog # sc-47778) was purchased from Santa Cruz Biotechnology (Dallas, TX). 3G8.2C11 antibody was obtained from the European Xenopus Resource Centre [60,61]. Goat antimouse IgG, light chain specific, conjugated to horseradish peroxidase (HRP, catalog # 115-035-174) was purchased from Jackson ImmunoResearch (West Grove, PA). Goat anti-mouse IgG conjugated to Alexa Fluor 488 or Alexa Fluor 594 was purchased from Thermo Fisher. Anti-digoxygenin (DIG) conjugated to alkaline phosphatase (AP) was purchased from Sigma-Aldrich (St. Louis, MO). Cdh17:GFP and Pax8:GFP transgenic Xenopus laevis frogs were obtained through the National Xenopus Resource (Woods Hole, MA, RRID:SCR_013731) [35,36,61]. All animals in this study were used in accordance with protocols approved by the Montana State University IACUC.

Reverse-transcriptase PCR
Total RNA was isolated from embryos with TRIzol (Thermo Fisher), quantitated by Nanodrop spectrophotometry and analyzed by ethidium bromide agarose gel electrophoresis (S1 Fig). Two μg of embryo RNA was treated with DNAse I (DNA-free, Thermo Fisher) to remove genomic DNA and cDNAs were reverse-transcribed by priming with oligo-dT using Super-Script IV reverse transcriptase (Thermo Fisher). A parallel control reaction was performed without reverse-transcriptase for each DNAse I-treated RNA. Semi-quantitative PCR was performed on 1/20 th of the resulting cDNA using primers for Xl-VSP1 (forward: GATGCTGGAA ACAACTCCATAGTCC; reverse: GCTGTGTATGTGGTCAGAACAC) or Xl-VSP2 (forward: GATGCTGGGAACAATTCCGTAGTCA; reverse: GGGTAATAGTACGTTAAGAAGTG) with Taq polymerase in Standard Taq PCR buffer (New England Biolabs, Ipswich, MA) with the thermocycling parameters: 95˚C for 30 sec, 37 cycles of 62˚C for 20 seconds followed by 68˚C for 30 seconds. For each primer pair the forward primer is located upstream of an intron and the reverse primer is located in the 3' UTR to ensure alloallele specificity and amplification of only cDNA. PCR products were analyzed by ethidium bromide agarose gel electrophoresis.

Western blotting
Adult Xenopus laevis tissue, oocyte, and embryo lysates were made by homogenization and brief sonication in lysis buffer (150 mM NaCl, 150 mM dithiothreitol, 0.1% CHAPS, 50 mM Tris, pH 7.4) with protease inhibitors (Roche, Switzerland). X. laevis oocytes were injected with 20 ng VSP RNA and cultured at 18˚C for 36 hours before lysis. X. laevis zygotes were injected with either 20 pg of the following antisense morpholino (CTGCTCGTATCTTACACT TGACATG) that targets the N-terminal coding region of the Xl-VSP2 transcript, or 20 pg of the Standard Control morpholino (Genetools, Philomath, OR). The corresponding N-terminal region of the Xl-VSP1 transcript has two internal base pair mismatches making the Xl-VSP2 morpholino unlikely to block Xl-VSP1 translation. Embryos were raised at 18˚C in 0.1X Marc's Modified Ringers solution (10 mM NaCl, 0.2 mM KCl, 0.1 mM MgSO 4 , 0.2 mM CaCl 2 , 0.5 mM HEPES, pH 7.4) with 50 mg/ml gentamicin until the desired stage. Protein concentrations were determined by a Quick Start Bradford assay (Bio-Rad, Hercules, CA). Lysates of un-injected, control morpholino, or Xl-VSP2 morpholino embryos were made by homogenizing eight embryos from each cohort into 200 μL lysis buffer. Lysates were run on 10% polyacrylamide gels at 200 volts for one hour, transferred to ProTran nitrocellulose membranes (GE Healthcare, Chicago, IL) in sodium borate transfer buffer for 70 minutes at 350 mA and stained with Ponceau-S to evaluate equal loading of protein between the wells. Membranes were blocked in TBS-T block buffer (5% non-fat dry milk, 150 mM NaCl, 10 mM, 0.1% Tween-80, 10 mM Tris, pH 7.4) and probed with anti-VSP supernatant at 1:25 dilution or anti-beta actin at 1:500 dilution. Blots were then washed in TBS-T, probed with anti-mouse IgG-HRP and developed with ECL reagent (Sycamore Life Sciences, Houston, TX).

Immunohistochemistry (IHC)
Adult Xenopus laevis kidneys and NF stage 40-42 embryos [34] were fixed in phosphate-buffered 4% paraformaldehyde (pH 7.4) for 90 minutes followed by sequential equilibration for two hours each in 30% sucrose, 30% sucrose/50% O.C.T (Optimal Cutting Temperature) embedding media (Fisher, Hampton, NH), and 100% OCT. Kidney tissues were then snap-frozen in OCT using a dry ice/ethanol slurry and sectioned at 14 μm onto glass slides coated with 0.5% porcine gelatin, type A (Sigma-Aldrich, catalog # G2500) and 0.05% chromium potassium sulfate. Sections were air-dried overnight at room-temperature and stored at -80˚C. Sections were rehydrated with phosphate-buffered saline (PBS; 137 mM NaCl, 27 mM KCl, 100 mM Na 2 HPO 4 , 18 mM KH 2 PO 4 , pH 7.4) and subjected to antigen retrieval in 10 mM sodium citrate, 0.05% Tween-80, pH 6.0 for 20 minutes at 95˚C followed by quick cooling on ice. Sections were re-equilibrated with PBS-T (0.1% Triton X-100) and blocked in PBS-T with 10% normal goat serum (Fisher, catalog # 16-210-064). Sections were probed with anti-VSP supernatant at 1:25 dilution, washed in PBS-T, and probed with anti-mouse IgG-Alexa Fluor 488 or anti-mouse IgG-Alexa Fluor 594 at 1:1000 dilution with 1 μg/ml Hoechst 33342 (Thermo-Fisher) to stain nuclei. Immuno-stained sections were mounted in Vectashield (Vector Labs, UK) under glass coverslips and imaged with a 63x/1.4 NA objective on a Leica SP8 confocal microscope (Leica, Buffalo Grove, IL). Embryo sections were similarly fixed, sectioned, stained and imaged without the antigen retrieval process. Whole-mount embryos were similarly fixed and stained without the antigen retrieval process and with four hours to overnight incubations to allow for antibody penetration. Whole-mount IHC images were captured using Progres Mac CapturePro software on a ProgRes C14plus camera through a Zeiss Planneofluar 5x/0.16 NA objective on an Axioscope.A1 microscope. GFP or Alexa Fluor 594 fluorescence in whole-mount immuno-stained embryos was excited by an LED and collected through Zeiss FITC (EX BP 475/40, BS FT 500, EM BP 530/50) or Texas Red (EX BP 560/40, BS FT 585, EM BP 630/75) filter sets, respectively. Anti-VSP, anti-3G8 and matching secondary-alone control samples were imaged using the same excitation intensity and digital gain. IHC was repeated a total of three times with tissue from different animals or clutches of embryos.

Electrophysiology and fluorescence measurement of activity
Two electrode voltage clamp (TEVC) was performed as previously described [14]. FRETbased PIP sensors [41,42] were used to measure depletion of PI(4,5)P 2 or the production and depletion of PI(3,4)P 2 . The PIP sensors were designed by adding an N-terminal CFP and a Cterminal YFP to pleckstrin homology (PH) domains. The PH domains were originally derived from phospholipase C (PLC) for PI(4,5)P 2 and the tandem PH domain containing protein 1 (TAPP) for PI(3,4)P 2 , and hence were called fPLC and fTAPP respectively. To measure production of PI(4,5)P 2 , the diffusion based sensor GFP-PLC (gPLC) was used which has a GFP was attached to the N-terminus of the PLC-PH domain. All Xl-VSP cRNAs (0.8 μg/μl) were mixed with either fPLC cRNA (0.4 μg/μl), fTAPP cRNA (0.4 μg/μl) or gPLC cRNA (0.06 μg/μl) and injected into X. laevis oocytes. In all experiments, 50 nl of the cRNA mixtures were injected into oocytes and incubated in ND-96 for 30-40 hours. On the day of the experiments, all cells were transferred from ND-96 to ND-96' (ND-96 without gentamicin and sodium pyruvate) containing 8 μM insulin to promote PI3 kinase activity and up-regulate PI(3,4,5)P 3 levels.
A Leica DM IRBE inverted microscope with a Leica HC Pl APO 20×/0.7 fluorescence objective was used with a Dagan CA-1B amplifier and illuminated with a Lumen Dynamics X-Cite XLED1 light source. Fluorescence was measured with a ThorLabs photomultiplier tube (PMT). The amplifier and light-emitting diode were controlled by a Digidata-1440A board and Axon™ pClamp™ 10.7 software package (Molecular Devices). For the FRET experiments, light was filtered through a HQ436/20 excitation filter and directed to the objective with a 455LP dichroic (Chroma). The microscope cube did not contain an emission filter, because the ThorLabs PMT module contains its own cube. Thus, the emitted light was filtered before the PMTs with a 510-nm dichroic, an HQ480/40 emission filter for CFP, and an HQ535/30 emission filter for YFP (Chroma). For the gPLC experiments, light was filtered through a HQ470/40 excitation filter, an HQ525/50 emission filter and a Q496LP dichroic (Chroma). The voltage protocol consisted of steps from −100 to 180 mV in irregular increments. The length of the voltage pulse was 1.5-2 s for Xl-VSP1 and varied for Xl-VSP2, 2-10 s for fTAPP, and fPLC, and 3 s for gPLC. Rest periods of 1-3 min between voltage steps were used to allow the cell to recover depleted PIP concentrations before the next voltage step. The resulting FRET or fluorescence was then plotted versus the voltage to generate the fluorescence versus voltage relationship.

Data analysis
For the FRET experiments, Axon™ Clampfit™ 10.7 (Molecular Devices) was used to calculate the FRET ratio. For the FRET or fluorescence increases (fTAPP and gPLC), the ΔF/F, ðF after À F before Þ =F before Þ , was calculated from the pre-pulse baseline to the max signal increase.
For the fPLC FRET decrease, the ΔF/F was calculated from the pre-pulse baseline to the max FRET decrease. For the fTAPP FRET decrease, the ΔF/F was calculated from the max FRET increase to the max FRET decrease. Calculations were done in Excel (Microsoft) or IGOR Pro (WaveMetrics) and data was plotted in IGOR Pro. When bleaching was observed, it was corrected by fitting to a line and dividing the values by the fit to obtain the final, bleaching-corrected signal. Final voltage dependent curves were plotted with change in FRET or fluorescence on the Y-axis and voltage on the X-axis. The data were fit to single Boltzmann equations. Activity assays were repeated on oocytes extracted on different days from at least three different frogs (biological replicates) until data from a minimum of 8 oocytes were acquired and analyzed. Error bars indicate standard error of the mean (SEM). IHC images were processed with ImageJ [64]. Fluorescence signals from anti-VSP and matching secondary-alone control images were processed in parallel while applying the same post-image processing parameters in each case. Embryos stained with 3G8 antibody (A-C), anti-VSP (A'-C') or secondary antibody alone control (A''-C''). The Pax8:GFP embryos show strong GFP expression in the proximal pronephros (B-B'') as confirmed by staining with the proximal pronephros marker 3G8 antibody (C). Xl-VSP staining of tubules is clearly visible (C') and co-localizes with the GFP, indicating Xl-VSP in the proximal pronephros. Positive staining is marked with arrowheads, whereas no signal above autofluorescence was seen in control animals stained with secondary antibody alone (C''). (TIF)

S3 Fig. Western blots of individual oocytes used for activity assays.
Individual X. laevis oocytes used for the activity assay in Fig 6 were probed for protein expression by Western blot with anti-VSP N432/21. Oocytes were injected with cRNA for wild type Xl-VSP1 (Xl1), wild type Xl-VSP2 (Xl2) and catalytically dead Xl-VSP2 C301S (Xl1-CS). (TIF)