Distinct functional properties of two electrogenic isoforms of the SLC34 Na‐Pi cotransporter

Abstract Inorganic phosphate (Pi) is crucial for proper cellular function in all organisms. In mammals, type II Na‐Pi cotransporters encoded by members of the Slc34 gene family play major roles in the maintenance of Pi homeostasis. However, the molecular mechanisms regulating Na‐Pi cotransporter activity within the plasma membrane are largely unknown. In the present study, we used two approaches to examine the effect of changing plasma membrane phosphatidylinositol 4,5‐bisphosphate (PI(4,5)P2) levels on the activities of two electrogenic Na‐Pi cotransporters, NaPi‐IIa and NaPi‐IIb. To deplete plasma membrane PI(4,5)P2 in Xenopus oocytes, we utilized Ciona intestinalis voltage‐sensing phosphatase (Ci‐VSP), which dephosphorylates PI(4,5)P2 to phosphatidylinositol 4‐phosphate (PI(4)P). Upon activation of Ci‐VSP, NaPi‐IIb currents were significantly decreased, whereas NaPi‐IIa currents were unaffected. We also used the rapamycin‐inducible Pseudojanin (PJ) system to deplete both PI(4,5)P2 and PI(4)P from the plasma membrane of cultured Neuro 2a cells. Depletion of PI(4,5)P2 and PI(4)P using PJ significantly reduced NaPi‐IIb activity, but NaPi‐IIa activity was unaffected, which excluded the possibility that NaPi‐IIa is equally sensitive to PI(4,5)P2 and PI(4)P. These results indicate that NaPi‐IIb activity is regulated by PI(4,5)P2, whereas NaPi‐IIa is not sensitive to either PI(4,5)P2 or PI(4)P. In addition, patch clamp recording of NaPi‐IIa and NaPi‐IIb currents in cultured mammalian cells enabled kinetic analysis with higher temporal resolution, revealing their distinct kinetic properties.


Introduction
Inorganic phosphate (P i ) is essential for a variety of cell activities, including bioenergetics and cell signaling. Phosphate is also an essential constituent of biological membranes and bone matrices, which are composed mainly of calcium phosphates. Dietary P i is absorbed in the small intestine and circulates in the blood, where its concentration is controlled mainly through reabsorption in the kidney. Absorption of P i in both the intestine and kidney is mediated primarily by type II Na-Pi cotransporters (or SLC34 proteins), which consist of three members, NaPi-IIa (SLC34A1), NaPi-IIb (SLC34A2), and NaPi-IIc (SLC34A3), showing different patterns of expression (Wagner et al. 2014): NaPi-IIa and NaPi-IIc are localized specifically in the apical brush border membrane of renal proximal tubule cells, whereas NaPi-IIb is found in many tissues, including the luminal brush border membrane of the small intestine and alveolar type II epithelial cells Wagner et al. 2014). All three forms transport one divalent P i (HPO 4 2À ) coupled with multiple Na + per transport cycle, but with different P i :Na + stoichiometries: NaPi-IIa and NaPi-IIb show a 1:3 stoichiometry (electrogenic), whereas NaPi-IIc exhibits a 1:2 stoichiometry (electroneutral) (Forster 2019). Studies of knockout mice and human disease have shown the importance of these transporters at the whole body level. Mice lacking NaPi-IIa exhibit hypophosphatemia and hyperphosphaturia (Beck et al. 1998), while those lacking NaPi-IIb die in utero (Shibasaki et al. 2009). Human NaPi-IIc mutations cause hereditary hypophosphatemic rickets with hypercalciuria (Bergwitz et al. 2006). P i absorption mediated by Na-Pi cotransporters is known to be regulated by hormones such as parathyroid hormone (PTH), fibroblast growth factor 23 (FGF23), and 1,25-(OH) 2 -Vitamin D 3 , which change the number of these transporters at the cell surface. PTH facilitates endocytosis of NaPi-IIa and NaPi-IIc in kidney, reducing reabsorption of P i (Segawa et al. 2007;Picard et al. 2010). 1,25-(OH) 2 -Vitamin D 3 , which stimulates calcium uptake in the small intestine, increases the number of NaPi-IIb in the brush border membrane of this tissue . These cotransporter activities are also regulated by membrane potential Hilfiker et al. 1998) and extracellular pH (de la Horra et al. 2000). However, other regulation mechanisms at the plasma membrane remain to be explored.
The activities of many ion channels and some transporters are reportedly regulated by phosphatidylinositol 4,5-bisphosphate (PI(4,5)P 2 ) (Suh and Hille 2008), which is the most common phosphoinositide in the plasma membrane and accounts for 1% of phospholipid (Mclaughlin et al. 2002). PI(4,5)P 2 can be hydrolyzed by receptor-activated phospholipase C (PLC) to produce inositol trisphosphate (IP 3 ) and diacylglycerol (DAG) (Rhee and Choi 1992). PI(4,5)P 2 -sensitive ion channel activities are regulated through stimulation of G q -coupled receptors, which in turn mediate PLC-catalyzed PI(4,5)P 2 hydrolysis (Hille et al. 2015). Such regulation is well characterized in cardiac cells and neurons (Cho et al. 2005;Morita et al. 2006). In neurons, for example, KCNQ2/3 activity is regulated by G q -coupled receptors that inhibit channel activity, thereby enhancing neurotransmitter-induced excitation (Hughes et al. 2007).

Ethics approval
Experiments using Xenopus laevis were performed in accordance with the guidelines of the Animal Care and Use Committee of the Osaka University Graduate School of Medicine.

Two-electrode voltage clamp technique
Electrical recordings were made using two-electrode voltage clamp (TEVC) carried out with an OC-725C amplifier (Warner Instruments, USA). Output signals filtered at 1 kHz were digitized with an AD/DA converter (Instru-TECH LIH 8 + 8; HEKA Elektronik, Germany) and analyzed using PatchMaster software (HEKA Elektronik). The resistance of the glass pipettes was 0.1-1.0 MΩ after filling with 2 mol/L KOAc and 1 mol/L KCl solution. The oocytes were placed in a bath chamber (volume, 1.0 ml) connected to an original gravity perfusion system. The oocytes were initially voltage clamped at À60 mV while being perfused with Ca 2+ -free ND100 solution (100 mmol/L NaCl, 2 mmol/L KCl, 5 mmol/L MgCl 2 , 5 mmol/L HEPES, pH 7.6 [adjusted with NaOH]) at the rate of 1.02 ml/min. 100 mmol/L NaH 2 PO 4 /Na 2 HPO 4 solution (pH 7.4) was diluted with Ca 2+ -free ND100 solution to make following 1 and 10 mmol/L P i solution.
Once the current reached a steady state, the solution was changed to 1 mmol/L P i solution (99 mmol/L NaCl, 1 mmol/L NaH 2 PO 4 /Na 2 HPO 4 , 2 mmol/L KCl, 5 mmol/L MgCl 2 , 5 mmol/L HEPES, pH 7.6). In the experiments summarized in Figure 1, 10 mmol/L P i solution (90 mmol/L NaCl, 10 mmol/L NaH 2 PO 4 /Na 2 HPO 4 , 1.8 mmol/L KCl, 4.5 mmol/L MgCl 2 , 4.5 mmol/L HEPES, pH 7.6) was administered to the oocytes at a rate of 25.0 ml/h using the puff application system described in Figure 1A. To prevent P i solution flowing out from the tip of the puff pipette before application, the tip was positioned such that it was not in contact with the surface of the bath solution.
In the experiments using Ci-VSP, a depolarization pulse to +50 mV was applied for 10 sec to activate Ci-VSP and to reduce PI(4,5)P 2 levels in the plasma membrane. To check the plasma membrane PI(4,5)P 2 level, rKCNQ2/3 currents were measured by following protocol: a step pulse to À10 mV was applied for 200 msec, 5 sec before and 5, 25, and 55 sec after Ci-VSP activation. Holding potential was set to À60 mV which activates little rKCNQ2/3 currents (Zhang et al. 2003).

Confocal fluorescence imaging
Confocal fluorescence imaging was performed using an inverted LSM 710 confocal microscope (Carl Zeiss, Germany) at room temperature (22-25°C). Transfected cells were reseeded onto poly-L-lysine-coated coverslips (Matsunami, Japan). Each coverslip was moved into the chamber for imaging analysis, which was filled with the 140 NaCl solution. Cells were scanned every 2 sec for 60 cycles. Between the 10th and 11th scan, 10 lmol/L rapamycin in the 140 NaCl solution was added to the chamber; the final rapamycin concentration was estimated to be 1 lmol/L. Acquired data were analyzed by using Ima-geJ software (NIH, USA).

Data analysis
Data are presented as the mean AE standard error of the mean (SEM). All data were analyzed using PatchMaster and Igor Pro (WaveMetrics Inc., USA) software. Statistical comparisons were made using Microsoft Excel 2016 (Microsoft, USA). Values of P < 0.05 were considered statistically significant.
Results P i -induced currents produced by two types of electrogenic Na-Pi cotransporters in a Xenopus oocyte expression system As in earlier reports (Hartmann et al. 1995;Hilfiker et al. 1998), we first measured the electrogenic transport activities of mNaPi-IIa and mNaPi-IIb using TEVC with a Xenopus oocyte expression system. To apply P i solution more rapidly than a gravity perfusion system, we utilized a puff application system with a syringe pump (Fig. 1A). In oocytes expressing mNaPi-IIa, an increase in the inward current was observed after application of 10 mmol/L P i solution, as was previously reported (Hartmann et al. 1995) (Fig. 1B). After washing out the P i , the current gradually returned to the initial basal level. A similar inward current induced by application of external P i was also observed with mNaPi-IIb ( Fig. 1C), as was previously reported (Hilfiker et al. 1998).

NaPi-IIa activity is unaffected by PI(4,5)P 2 depletion by Ci-VSP
We tested the sensitivity of mNaPi-IIa to PI(4,5)P 2 by depleting the phosphoinositide using Ci-VSP, which dephosphorylates PI(4,5)P 2 upon depolarization of the membrane potential (Murata and Okamura 2007;Okamura et al. 2018). The extent of the reduction in PI(4,5)P 2 was monitored electrophysiologically by coexpressing rKCNQ2/3 ion channels. mNaPi-IIa and rKCNQ2/3 currents were recorded from single oocytes to evaluate the relationship between the PI(4,5)P 2 level and mNaPi-IIa current ( Fig. 2A and B). When the cell membrane was depolarized to +50 mV for 10 sec after the mNaPi-IIa current reached a steady state ( Fig. 2A 2C and D), indicating PI(4,5)P 2 was depleted during that period. We also confirmed that within 55 sec after terminating the 10-sec depolarization (shown in upper panel of Fig. 2A), the rKCNQ2/3 current had recovered to the predepolarization level ( Fig. 2A, trace 1-4 and 5-8 in the inset), indicating that PI(4,5)P 2 levels are able to completely recover following repolarization of the membrane. These results show that mNaPi-IIa is insensitive to PI(4,5)P 2 , in the range of PI (4,5)P 2 reductions attained through activation of Ci-VSP.
PI(4,5)P 2 depletion by Ci-VSP inhibits NaPi-IIb activity We examined the sensitivity of mNaPi-IIb to PI(4,5)P 2 using the same protocol used for mNaPi-IIa (Fig. 3A). In contrast to mNaPi-IIa currents, the 10-sec depolarization led to reductions in the mNaPi-IIb current (Fig. 3B). Immediately after the depolarization, mNaPi-IIb currents were significantly smaller than before it (Pre: 69 AE 20 nA, Post: 43 AE 17 nA, n = 3, paired t-test: P = 0.025, ratio: 0.57 AE 0.07) (Fig. 3D). Within 55 sec after membrane repolarization, the currents had recovered to the predepolarization amplitude (Fig. 3B). Recovery of the rKCNQ2/3 currents was also observed (Fig. 3A, trace 5-8 in the inset). To verify that it was PI(4,5)P 2 depletion by Ci-VSP that inhibited the mNaPi-IIb activity, we assessed mNaPi-IIb activity using the same protocol with a Ci-VSP C363S mutant in which a cysteine in the phosphatase active center is substituted with a serine, eliminating its enzymatic activity (Murata et al. 2005). Using the mutated Ci-VSP, neither the mNaPi-IIb nor the rKCNQ2/3 current was affected by the 10 sec depolarization (for mNaPi-IIb, Pre: 26 AE 2 nA, Post: 28 AE 1 nA, n = 3, paired t-test: P = 0.28, ratio: 1.08 AE 0.05) ( Fig. 3C and D). Following the depolarization, mNaPi-IIb current amplitudes were significantly smaller with WT Ci-VSP than with the C363S mutant (Fig. 3E). These results indicate that mNaPi-IIb is sensitive to the membrane PI(4,5)P 2 level, which is diminished by Ci-VSP activity. Interestingly, even in the absence of P i , holding currents decreased after the 10-sec depolarization (Fig. 3B arrow) and then gradually returned to the level before the depolarization as found in rKCNQ2/3 currents (Fig. 3A, trace 1-4 in the inset). No changes in the holding current were observed when the Ci-VSP C363S mutant was expressed (Fig. 3C). Previous studies reported that in the absence of P i , holding current contains a baseline current carried by Na + in oocytes heterologously expressing mouse NaPi-IIa, rat NaPi-IIa, and flounder NaPi-IIb (Forster et al. 2002;Andrini et al. 2008). These led us to speculate that the holding currents in our experiments also contained such baseline current derived from P i -independent basal transporter activity (hereafter, thus called the "P i -independent basal inward current"). We conclude that both the P i -independent basal inward current and currents linked with P i transport by mNaPi-IIb are sensitive to PI(4,5)P 2 depletion.

NaPi-IIa and NaPi-IIb currents heterologously expressed in Neuro 2a cells
Ci-VSP-mediated dephosphorylation of PI(4,5)P 2 leads to production of PI(4)P (Iwasaki et al. 2008). Therefore, a possible explanation for the lack of change in the amplitude of mNaPi-IIa currents with activation of Ci-VSP may be that mNaPi-IIa is sensitive to both PI(4,5)P 2 and PI(4)P.
To test this possibility, we made use of the engineered chimeric lipid phosphatase PJ, which has both 5-phosphatase and 4-phosphatase activities (Hammond et al. 2012). In this system, recruitment of PJ to the plasma membrane can be readily induced by adding rapamycin to the bath solution. Upon binding of rapamycin to PJ, the enzyme is recruited to the plasma membrane through linkage to a scaffold protein (Lyn 11 -FRB) localized at the membrane, which in turn leads to depletion of both PI(4,5)P 2 and PI(4)P (Hammond et al. 2012). However, in our pilot study using PJ in Xenopus oocytes, we were unable to induce robust phosphatase activity. We therefore decided to use a mammalian cell expression system. 2019 | Vol. 7 | Iss. 14 | e14156 Page 5 The whole-cell patch clamp recording of type II Na-Pi cotransporters from cultured mammalian cells has not been reported. We initially chose HEK293T cells as a heterologous expression system, but an inward current induced by P i was occasionally recorded even in untransfected cells possibly due to some endogeneous P i -sensitive channel or transporter. We next tried Neuro 2a cells. Unlike with HEK293T cells, no inward currents were recorded from untransfected Neuro 2a cells (n = 9) (Fig. 4A). When we transfected the cells with mNaPi-IIa-pIRES2-EGFP plasmid . Ci-VSP-induced PI(4,5)P 2 depletion has no effect on NaPi-IIa but inhibits KCNQ2/3 activity. (A) pulse protocol (upper) and membrane currents (lower) in an oocyte coexpressing mNaPi-IIa, rKCNQ2, rKCNQ3, and Ci-VSP. rKCNQ2 and rKCNQ3 form heterotetrameric rKCNQ2/3 channels, which carry delayed-rectifier voltage-gated potassium currents sensitive to PI(4,5)P 2 . To activate Ci-VSP, oocytes were depolarized to +50 mV for 10 sec twice. The first 10-sec depolarization was applied before application of P i in the bath solution and the second after application of P i . To measure rKCNQ2/3 currents, depolarization steps to À10 mV were applied for 200 msec, 5 sec before and 5, 25, and 55 sec after each 10-sec depolarization. The inset shows rKCNQ2/3 currents recorded to monitor the PI(4,5)P 2 levels on an enlarged timescale at times 1-4 and 5-8 of the pulse protocol. Membrane potential was maintained at À60 mV throughout the recording.  and measured the currents from GFP-positive cells, we were able to record inward transporter currents after 5 mmol/L P i solution was perfused using an ALA Scientific rapid perfusion system (Fig. 4C). The current recovery to baseline after washing out the external P i was slow. We also observed a P i -independent basal inward current (Fig. 4C), as was previously reported in oocyte experiments (Forster et al. 2002;Andrini et al. 2008). This basal inward current decreased in Na + -free solution (Fig. 4B), consistent with the idea that it is carried by Na + . By measuring currents through SCN5A (hH1), a human cardiac voltage-gated Na + channel, with distinct external Na + concentrations, we estimated the time needed for replacement of the external solutions to be within 5 sec (Fig. 4D). This suggests that the slow current recovery reflects an innate property of mNaPi-IIa, not a slow exchange rate of external P i in our recording system. Of note, the kinetics of P i -induced currents of mNaPi-IIb differed from those of mNaPi-IIa. In the presence of P i , mNaPi-IIb exhibited marked current decay after the current reached its maximum (Fig. 4E arrow). No such current decay was seen in studies with Xenopus oocytes (Hilfiker et al. 1998). The relative mNaPi-IIb current amplitude 19 sec after reaching the peak was significantly smaller than the mNaPi-IIa current (mNaPi-IIb: 0.51 AE 0.03, n = 8 vs. mNaPi-IIa: 0.93 AE 0.06, n = 7) (Fig. 4F). The decay phase of the mNaPi-IIb current was fitted by a single-exponential function, and its time constant (s) was 6.3 AE 1.1 sec (Fig. 4G). Like mNaPi-IIa (Fig. 4C), mNaPi-IIb exhibited a slow recovery after P i washout (Fig. 4E). A P i -independent basal inward current decreased in Na + -free solution as with mNaPi-IIa, consistent with the idea that it is also carried by Na + .
Depletion of PI(4,5)P 2 and PI(4)P using PJ does not affect NaPi-IIa activity We coexpressed PJ in Neuro 2a cells with PH PLCd1 -mCherry, a PI(4,5)P 2 sensor, or with GFP-P4M-SidMx1, a PI(4)P-sensitive fluorescent probe (Hammond et al. 2014), and used confocal microscopy to assess the phosphatase activities of PJ (Fig. 5A). PH PLCd1 -mCherry and GFP-P4M-SidMx1 moved from the plasma membrane to the cytoplasm after addition of rapamycin ( Fig. 5B and C), indicating that both PI(4,5)P 2 and PI(4)P were depleted in the plasma membrane. This experiment was performed in two more cells for each probe and similar results were obtained.
With this system, we next examined the effects of PI(4,5)P 2 and PI(4)P depletion on mNaPi-IIa activity using whole-cell patch clamp (Fig. 6A). Rapamycin was added after inward currents were induced by P i . mNaPi-IIa currents were unaffected by the addition of rapamycin (Pre: 5.8 AE 1.4 pA, Post: 5.7 AE 1.3 pA, n = 6, paired ttest: P = 0.80, ratio: 1.01 AE 0.06) (Fig. 6B, D, and F). By contrast, mNaPi-IIb currents were significantly decreased after perfusion of rapamycin (Pre: 5.9 AE 1.8 pA, Post: 2.6 AE 0.9 pA, n = 7, paired t-test: P = 0.037, ratio: 0.43 AE 0.09) (Fig. 6C upper panel, E, and F), but were unaffected by perfusion of vehicle (DMSO) (Pre: 5.4 AE 0.9 pA, Post: 5.2 AE 1.0 pA, n = 3, paired t-test: P = 0.43, ratio: 0.95 AE 0.03) (Fig. 6C lower panel, E, and  F). These results are consistent with the results from Xenopus oocytes and indicate that whereas mNaPi-IIa activity is insensitive to the membrane PI(4,5)P 2 level, mNaPi-IIb activity is reduced upon depletion of PI(4,5)P 2 . We sometimes observed a decrease in the basal current after rapamycin application (Fig. 6C upper panel); that is, after washout of P i , mNaPi-IIb currents decreased to a level lower than the initial current level before perfusing P i -containing solution ( Fig. 6C upper panel, dashed line). This phenomenon was not observed with vehicle application nor with mNaPi-IIa ( Fig. 6B and C lower panel). This suggests that both P i -coupled current and the P i -independent basal inward current of mNaPi-IIb depend on PI(4,5)P 2 , as was suggested by experiments with oocytes (Fig. 3B).

Discussion
In this study, we used voltage-induced depletion of PI(4,5) P 2 with Ci-VSP and rapamycin-induced depletion of both PI(4,5)P 2 and PI(4)P with a PJ system to examine the sensitivity to phosphoinositides of two mouse electrogenic Na-Pi cotransporters, mNaPi-IIa and mNaPi-IIb. Our results indicate that mNaPi-IIa is insensitive to both PI(4,5)P 2 and PI(4)P, whereas mNaPi-IIb is sensitive to depletion of PI (4,5)P 2 . Our study also provided the first case of characterization of NaPi-IIa and NaPi-IIb by whole-cell patch clamp recording. We found that these transporters exhibit a P i -independent basal inward current and that a slow recovery after washout of P i could be an innate property of both mNaPi-IIa and mNaPi-IIb. We also found that these transporters have distinct kinetics in response to application of P i . Collectively, our findings indicate that mNaPi-IIa and mNaPi-IIb have distinct functional properties including different sensitivities to PI(4,5)P 2 .
Depletion of PI(4,5)P 2 by VSP or Pseudojanin VSP's nature of rapid and reversible PI(4,5)P 2 depletion is effective for examining the PI(4,5)P 2 sensitivity of ion channels and transporters, as it minimizes cell damage and enables repeated recordings from the same cell. With the PJ system, mNaPi-IIb currents were also decreased during application of rapamycin (Fig. 6C 2019 | Vol. 7 | Iss. 14 | e14156 Page 8 upper panel). However, the rate of PI(4,5)P 2 depletion was slower than with Ci-VSP, and the depletion was irreversible because after combining with the scaffold protein (Lyn 11 -FRB) through rapamycin, the phosphatase remains irreversibly anchored to the plasma membrane. representative time course of SCN5A (hH1) current decline upon rapid perfusion of Na + -free solution by using an ALA Scientific perfusion system. Right: current traces recorded from a HEK293T cell expressing SCN5A (hH1) (upper) and the pulse protocol (lower). Eight current traces recorded at different times during the solution exchange are superimposed. Currents were induced by 10-msec step pulses to -20 mV every 1 sec after starting perfusion at 0 sec; the pulse was repeated eight times. The holding potential was À80 mV. (E) Representative current trace recorded from a mNaPi-IIb-expressing Neuro 2a cell. 5 mmol/L P i solution was applied during the period indicated by the thick bar. An arrow points to the current decay. (F) Relative amplitudes of currents recorded 19 sec after reaching its maximal amplitude. Currents were normalized to the maximal amplitude. Normalized amplitudes were compared using Student's t-test (means AE SEM, ***P < 0.001). (G) The trace in E shown on an enlarged timescale (gray trace). The current decay was fitted by a single-exponential equation (red line). The time constant is the mean AE SEM (n = 8). In A, B, C, and E, membrane potential was clamped to À60 mV. Insensitivity of any given protein to depletion of PI(4,5)P 2 by VSP does not necessarily mean that the protein is totally insensitive to phosphoinositide, since dephosphorylation of PI(4,5)P 2 by VSP produces PI(4)P (Iwasaki et al. 2008). For example, the activity of mammalian transient receptor potential vanilloid 1 (TRPV1) is unaffected by VSP-induced depletion of PI(4,5)P 2 (Lukacs et al. 2013), but its activity is suppressed by depletion of both PI(4,5)P 2 and PI(4)P (Hammond et al. 2012;Lukacs et al. 2013), suggesting that mammalian TRPV1 is sensitive both to PI(4,5)P 2 and PI(4)P. In our study, mNaPi-IIa activity was not affected by Ci-VSP-induced PI(4,5)P 2 depletion (Fig. 2B) or by depletion caused by the PJ system (Fig. 6B), indicating that mNaPi-IIa is insensitive to both PI(4,5)P 2 and PI(4)P. Thus, the combination of VSP and PJ enables precise evaluation of ion channel and transporter sensitivity to PI(4,5)P 2 and PI(4)P.
How does PI(4,5)P 2 bind to mNaPi-IIb to regulate its transport activity? PI(4,5)P 2 is known to bind to ion channels electrostatically (Hille et al. 2015;Dickson and Hille 2019). For example, X-ray crystallographic analysis of the PI(4,5)P 2 -binding sites on the inwardly rectifying potassium channel 2.2 (Kir2.2) showed that positively charged amino acids are important for PI(4,5)P 2 binding (Hansen et al. 2011). By analogy, positively charged amino acids on the intracellular side of mNaPi-IIb may be important for PI(4,5)P 2 binding. A topological model based on epitope labeling, cysteine scanning mutagenesis, and in vitro glycosylation assays (Radanovic et al. 2006) showed that the N-and C-terminal regions of type II Na-Pi cotransporters are situated on the intracellular side of the protein and that there are five intracellular linkers. In mNaPi-IIb, these intracellular regions contain many positively charged amino acids, which could be candidate sites for PI(4,5)P 2 binding. Amino acid sequence alignment among type II Na-Pi cotransporters shows that as compared to mNaPi-IIb, a large section has been deleted from the C-terminal region of mNaPi-IIa and mNaPi-IIc. This suggests that the C-terminal region of mNaPi-IIb may be important for its PI(4,5)P 2 sensitivity. Several positively charged amino acids within the C-terminal region of NaPi-IIb conserved among mammals may be important for PI(4,5)P 2 binding.
Both basal inward currents in the absence of extracellular P i and P i -coupled currents were reduced upon depletion of PI(4,5)P 2 (Figs. 3B and 6C upper panel). Basal inward currents reflect a Na + -dependent P i transport mode intrinsic to electrogenic Na-Pi cotransporters (Andrini et al. 2008;Forster 2019). Our finding that these two transport modes are sensitive to PI(4,5)P 2 in mNaPi-IIb suggests that they involve common structural changes, consistent with a previously proposed idea (Forster 2019).
Type II Na-Pi cotransporters are conserved among species from bacteria to humans (Werner and Kinne 2001). With the exception of chickens and mammals, all tested species express only a single type II Na-Pi cotransporter classified as NaPi-IIb (Forster et al. 2011). NaPi-IIa and NaPi-IIc are reported only in chicken, mouse, rat, and human (Forster et al. 2011). Interestingly, the NaPi-IIb homolog in winter flounder is found in tissues throughout the animal, including the kidney (Werner et al. 1994), but mammalian NaPi-IIb is not expressed in kidney, where NaPi-IIa and NaPi-IIc mediate P i reabsorption (Forster et al. 2011). These observations suggest that NaPi-IIb is more closely related to the ancestor of type II Na-Pi cotransporters than the other two isoforms are. It will be intriguing to know whether the teleost NaPi-IIb homolog is PI(4,5)P 2 -dependent. It is possible that Figure 5. Changes in plasma membrane PI(4,5)P 2 and PI(4)P upon activation of PJ in Neuro 2a cells. (A) schematic representation of PI(4,5)P 2 and PI(4)P depletion upon PJ recruitment to the plasma membrane. PI(4,5)P 2 and PI(4)P were monitored using PH PLCd1 -mCherry and GFP-P4M-SidMx1, respectively. After PI(4,5)P 2 or PI(4)P depletion by PJ, these fluorescent probes translocate to the cytoplasmic region. Sac 1 and INPP5E are 4-phosphatase and 5-phosphatase, respectively. INPP5E, inositol polyphosphate-5-phosphatase E; PI, phosphatidylinositol. (B) Fluorescence images of PH PLCd1 -mCherry before (left) and after (right) application of rapamycin (upper). Representative time course of normalized PH PLCd1 -mCherry fluorescence intensity in the plasma membrane and cytoplasmic region (lower). Images were captured at 0 sec (left) and 118 sec (right) from a Neuro 2a cell expressing PH PLCd1 -mCherry, PJ, and Lyn 11 -FRB. Scale bar, 5.0 lm. Rapamycin (10 lmol/L) was applied to the recording chamber at the time indicated by the arrow in the lower panel. (C) Fluorescence images of GFP-P4M-SidMx1 (upper) and representative time course of normalized GFP-P4M-SidMx1 fluorescence intensity in each region (lower). Neuro 2a cells expressing GFP-P4M-SidMx1, PJ, and Lyn 11 -FRB were studied. Images were captured at 0 sec (left) and 118 sec (right). Scale bar, 5.0 lm. Rapamycin was applied at the time indicted by the arrow in the lower panel. PI(4,5)P 2 -independence of mNaPi-IIa is a more recent innovation along vertebrate evolution.
Insights into physiological significance of different PI(4,5)P 2 dependence between NaPi-IIa and NaPi-IIb Our finding that mNaPi-IIa is insensitive to PI(4,5)P 2 suggests that it does not require PI(4,5)P 2 for its transport activity in renal proximal tubules (Fig. 7). The activities of many ion channels sensitive to PI(4,5)P 2 are known to be altered by G q -mediated activation of PLC, which hydrolyzes PI(4,5)P 2 (Hille et al. 2015). In the kidney, G q -coupled receptors such as PTH receptors are present on the basolateral membrane of epithelial cells in the renal cortical tubules (Amizuka et al. 1997). However, our results exclude the possibility that NaPi-IIa activity is altered through changes of PI(4,5)P 2 concentration upon activation of G q -coupled receptors. It is possible that PI(4,5)P 2 -independence of NaPi-IIa activity is important for P i reabsorption activity in the plasma membrane of renal proximal tubule cells. We also showed that mNaPi-IIa is still active upon depletion of both PI(4,5)P 2 and PI (4)P (Fig. 6B), raising a possibility that P i transport activity could be maintained by NaPi-IIa which is located in endomembranes such as endosomes or lysosomes. The transport ability of NaPi-IIa within endomembranes may be inhibited by the acidic luminal environment (Demaurex 2002) for its pH-sensitivity (de la Horra et al. 2000).
In contrast to NaPi-IIa, the activity of NaPi-IIb in endomembranes, which have little PI(4,5)P 2 , may be limited due to its PI(4,5)P 2 dependence. Because G q -coupled receptor signaling is the main pathway to PI(4,5)P 2 hydrolysis, the same pathway likely regulates the activity of PI(4,5)P 2 -sensitive NaPi-IIb in cells that natively express it. NaPi-IIb is known to be expressed at the apical pole of alveolar type II (AT2) epithelial cells , which produce pulmonary surfactant. P i is an essential component of surfactant, and NaPi-IIb may bind P i in the liquid covering the surface of the alveolar epithelium and provide it to AT2 cells. GPR116, a G q -coupled receptor, reportedly localizes at the apical side of AT2 cells (Brown et al. 2017) where it negatively regulates surfactant secretion (Brown et al. 2017). This suggests that NaPi-IIb's transport activity is inhibited by activation of GPR116. Thus, simultaneous restriction of P i transport and surfactant synthesis through inhibition of NaPi-IIb activity via activation of GPR116 may take place under some physiological condition.
NaPi-IIb is also known to be expressed in intestine (Hilfiker et al. 1998). Our finding that NaPi-IIb is sensitive to PI(4,5)P 2 (Fig. 7) raises the possibility that an as yet unidentified G q -coupled receptor signaling pathway may be important for P i homeostasis in the intestine. The composition of ingested food is known to stimulate G q -coupled receptors expressed by some types of secretory cells of the intestine (Reimann et al. 2012). This three Na + . NaPi-IIb activity, for example in the intestine, is regulated by PI(4,5)P 2 in the plasma membrane. NaPi-IIa activity in the kidney is unaffected by PI(4,5)P 2 . Figure 6. Effect of PJ-induced depletion of PI(4,5)P 2 and PI(4)P on NaPi-IIa and NaPi-IIb activities in Neuro 2a cells. (A) Schematic representation of the measurement of Na-Pi cotransporter activity with PJ-induced depletion of PI(4,5)P 2 and PI(4)P. Rapamycin links PJ to Lyn 11 -FRB present in the plasma membrane. (B) Representative current trace recorded from a Neuro 2a cell coexpressing mNaPi-IIa with PJ and Lyn 11 -FRB. Rapamycin (1 lmol/L) was applied after P i -induced transporter currents emerged. (C) Representative mNaPi-IIb currents recorded from Neuro 2a cells coexpressing PJ and Lyn 11 -FRB in the presence (upper) or absence (lower) of 1 lmol/L rapamycin. In B and C, membrane potential was clamped at À60 mV and dashed lines indicate the initial current level before perfusing P i -containing solution. (D) Comparison of the amplitudes of mNaPi-IIa currents recorded before (pre) and after (post) perfusion of rapamycin. (E) Changes in mNaPi-IIb current amplitude elicited by perfusion of rapamycin (open squares and solid line) or vehicle (filled circles and dashed line). (F) Summary of the effect of PJ-induced depletion of PI(4,5)P 2 and PI(4)P on mNaPi-IIa and mNaPi-IIb currents. Current amplitudes after rapamycin application were normalized to the amplitudes before application. Normalized amplitudes were compared using Student's t-test (n = 6 for mNaPi-IIa, n = 7 for rapamycin-treated mNaPi-IIb, n = 3 for vehicle-treated mNaPi-IIb, means AE SEM, **P < 0.01). suggests the possibility that NaPi-IIb may be involved in mediating mucus secretion, though it is not known whether NaPi-IIb is expressed in mucus-secreting cells.
The physiological significance of different PI(4,5)P 2 sensitivity between NaPi-IIa and NaPi-IIb will be an interesting topic for future study.
Kinetic properties of the two electrogenic Na-Pi cotransporters, NaPi-IIa and NaPi-IIb The slow recovery of mNaPi-IIa and mNaPi-IIb after P i washout has previously been shown (Forster et al. 2002;Andrini et al. 2008). In these studies using Xenopus oocytes, the possibility of inefficient washout of external P i could not be excluded. In the present study, mNaPi-IIa and mNaPi-IIb currents from cultured Neuro 2a cells under whole-cell patch clamp using a rapid perfusion system also revealed the slow recovery after P i washout ( Fig. 4C and E). To test if the slow recovery observed in this study could be an innate property of these transporters, more rigorous experiments are necessary.
In this study, we succeeded in measuring electrogenic type II Na-Pi cotransporter currents by whole-cell patch recording for the first time. In the presence of P i , mNaPi-IIb exhibited marked current decay after the current reached its maximum (Fig. 4E arrow), whereas mNaPi-IIa did not show such current decay. Further experiments will be necessary to determine molecular basis underlying distinct kinetic properties between mNaPi-IIa and mNaPi-IIb.