The main nonhormonal mechanism for controlling inorganic phosphate (Pi) homeostasis is renal adaptation of the proximal tubular Pi transport rate to changes in dietary phosphate content. Opossum kidney (OK) cell line is an in vitro renal model that maintains the ability of renal adaptation to the extracellular Pi concentration. We have studied how two competitive inhibitors of Pi transport, arsenate [As(V)] and phosphonoformate (PFA), affect adaptation to low and high Pi concentrations. OK cells show very high affinity for As(V) (inhibitory constant, Ki 0.12 mM) when compared with the rat kidney. As(V) very efficiently reversed the adaptation of OK cells to low Pi (0.1 mM), whereas PFA induced adaptation similar to 0.1 mM Pi. Adaptation with 2 mM Pi or As(V) was characterized by decreases in the maximal velociy (Vmax) of Pi transport and an abundance of the NaPi-IIa Pi transporter in the plasma membrane, shown by the protein biotinylation. Conversely, PFA and 0.1 mM Pi increased the Vmax and transporter abundance. Changes in the Vmax were limited to a 50% variation, which was not paralleled by changes in the concentration of Pi or of the inhibitor. OK cells are very sensitive to As(V), but the effects are reversible and noncytotoxic. These effects can be interpreted as As(V) being transported into the cell, thereby mimicking a high Pi concentration. PFA blocks the uptake of Pi but is not transported, and it therefore simulates a low Pi concentration inside the cell. To conclude, a mathematical definition of the adaptation process is reported, thereby explaining the limited changes in Pi transport Vmax.
- phosphate transport
- inorganic phosphate deprivation
- renal inorganic phosphate adaptation
- inorganic phosphate reabsorption
one of the main renal mechanisms for controlling inorganic phosphate (Pi) homeostasis is the adaptation of the Pi proximal tubular reabsorption (transport) rate to changes in dietary Pi content (20). This renal adaptation shows two distinct phases, an acute phase with a fast response, and a second, chronic phase of slow adaptation. The initial (acute) adaptation of the kidney is fast, it lasts for several hours, and it is independent of de novo protein synthesis in the cell. This acute adaptation involves the internalization and degradation of Pi transporters from the brush-border membrane in response to a high Pi diet, or the insertion of transporters in the brush-border membrane from intracellular, subapical stores as an adaptation to a low Pi diet (8). The slow or chronic response involves the synthesis of new transporters in the proximal tubular epithelial cells after a few hours of a low Pi diet (12, 14).
Previous studies have mainly been performed on type II Pi transporters (NaPi-IIa and -IIc; 25), but recently the involvement of type III transporters (Pit-2) has also been reported (23). However, we still do not know the precise mechanism(s) that are activated upon exposure to a diet with a different Pi content, according to which the renal reabsorption rate is consequently adapted and refined. These mechanisms seem to be independent of traditional factors that also control renal Pi transport (e.g., parathyroid and thyroid hormones, vitamin D3, phosphatonins, etc.; 1, 2, 8, 11). Recently, however, it has been shown that an unknown phosphaturic factor from the small intestine senses the Pi content of a diet and activates a gut-renal axis that modulates Pi reabsorption and the excretion rate before changes in the Pi plasma concentration are detected by the kidney (3). Independent of this new factor, the renal proximal tubular epithelial cells also contain an intrinsic mechanism that senses extracellular Pi changes and elicits adaptation of the Pi transport rate through the apical plasma membrane. Evidence of this autonomous system has been shown by using primary cultures of epithelial proximal tubular cells (5) and several established cell lines, including opossum kidney (OK) cells (15, 17), but the molecular nature of the intrinsic mechanism of adaptation is also unknown. Nevertheless, the failed attempts to identify this endogenous sensor have provided valuable information, such as the interaction between Pi transporters and PDZ domain-containing proteins (6).
The characteristics of renal Pi transport have also been studied using competitive inhibitors, such as arsenate [As(V); 9, 21] and phosphonoformic acid (PFA; 19). Both inhibitors show higher specificity for type II than for type III (Pit-1 and Pit-2) Pi transporters (21, 25). As(V) is readily transported through type II transporters when the concentration of Pi is low (21, 25), whereas PFA is not transported. Interestingly, as PFA blocks Pi transport, it also induces upregulation of Pi transport, similar to acute adaptation to a low Pi concentration (13).
In this study, we have extended to As(V) the seminal observations of Dousa's group on PFA (13), and we have characterized the effects of PFA and As(V) in OK cells. Our hypothesis is that whereas As(V) and PFA are competitive inhibitors of type II Pi transport, the cellular consequences of inhibition are different, and therefore both inhibitors can be used to elucidate the cellular mechanism of renal adaptation to Pi concentration. We have found that As(V) mimics accumulated Pi as it is transported into cells through Pi transporters, and consequently the cells do not adapt to a low Pi. The opposite is true for PFA, thereby confirming previous findings that have been reached through molecular characterization. We also report several differences between the adaptation to As(V) and to Pi, which have revealed certain characteristics of Pi transport in OK cells.
MATERIALS AND METHODS
Cell culture and uptake assays.
Wild-type OK cells were cultured as previously described (15) in passages 90–95. Culture media, including phosphate-free DMEM, were from Invitrogen (Carlsbad, CA). Radiotracer uptake assays were also performed, as reported previously, in cells grown on plastic 24-multiwell plates (21). All data are shown as total uptake in the presence of Na+, because Na-independent uptake of Pi in OK cells is minimal and only corresponds to diffusion and unspecific binding. In all experiments involving preincubation with Pi, As(V), or PFA, cells were washed with uptake solution before the radiotracer assays were performed.
Experiments with Xenopus laevis oocytes.
Uptake assays were performed as previously described (21) in all experiments related to in vitro transcription, the capping of RNA and polyadenylation, and the handling of Xenopus laevis and oocytes. Animals were maintained and handled according to APS guiding principles, and all procedures were approved by the Ethical Committee of the University of Zaragoza.
Polyclonal antibody preparation.
Polyclonal antibodies were prepared in rabbits against a synthetic peptide of the OK cell NaPi-IIa (formerly NaPi-4) amino terminus, MHSLKPLDQLITRATL (Davids Biotechnologie, Regensburg, Germany). The peptides were conjugated to keyhole limpet hemocyanin, mixed in Freund's complete adjuvant, and injected into rabbits. Two booster injections were given to the animals before sera were collected. A monospecific antibody was affinity purified from serum using the corresponding antigenic peptide. The specificity of antigen recognition was characterized by peptide protection and using preimmune serum, as described (23).
Membrane protein biotinylation.
OK cells were treated with As(V), PFA, or Pi for the indicated conditions, and membrane proteins were then biotinylated and detected by chemiluminescent immunoblot, as described (4). Densitometric signals were obtained with a Gel-Doc 1000 (Bio-Rad) after exposure to X-ray film. The specificity of signals was determined with preimmune serum and by blocking the affinity-purified antibodies with the corresponding antigenic peptide using the standard procedure and at a 50-fold molar excess of peptide per IgG. The antibody was used at a 1:1,000 dilution.
Cytotoxicity assays and microscopy.
OK cells were incubated with different concentrations of either As(V) or PFA for several time points. Cytotoxicity was assayed by quantifying the release of the intracellular lactate dehydrogenase. This was confirmed with a double staining using acridine orange-ethidium bromide, as reported (22). Actin filaments were visualized in paraformaldehyde-fixed cells using fluorescein isothiocyanate (FITC)-labeled phalloidin (Sigma). Cells were visualized using a Carl Zeiss Axiovert 200M inverted microscope (Jena, Germany) equipped with structured imaging (Apotome).
All experiments were performed three times, and the statistical significance of the comparison of means was determined by a t-test or an analysis of variance using Tukey's multicomparison posttest. In all cases, P < 0.05 was considered significant. The kinetic analysis of Pi transport has been described in detail in a previous work (21).
Pi transport and adaptation in OK cells.
We first checked whether Pi transport in our OK cells maintained the classical renal characteristics exhibited by these cells. Figure 1A shows total Pi uptake with increasing, saturating concentrations of substrate in the presence of Na+. The fit of the data to a Michaelian equation containing a saturable (transport) and a nonsaturable (diffusion plus unspecific binding) component (see legend of Fig. 1) provided a Pi Km value of 94 ± 12 μM. In addition, total uptake was also pH dependent, given that Na-coupled Pi uptake increased as the pH of the assay medium increased, up to a maximum of pH 7.5 (Fig. 1B). Both characteristics, affinity and pH dependence, coincide with the results of previous reports (8, 9, 17, 18, 21). A third characteristic of the renal proximal tubule that is also present in OK cells is the adaptation of the Pi transport rate to the extracellular concentration of Pi (Pi,out): the cell adjusts the capacity (maximal velocity, Vmax) or number of functional Pi transporters in the plasma membrane to maintain a constant transport rate of Pi molecules. Therefore, as Pi,out increases, fewer transporters are needed to handle a fixed amount of Pi; and conversely, as Pi,out decreases, more Pi transporters are needed to uptake all the necessary Pi molecules. If the maximal number of Pi transporters in the plasma membrane is obtained with a low Pi concentration such as 0.1 mM (i.e., 100%), then according to Michaelis-Menten [v = (Vmax·S)/(Km + S); where S is the Pi concentration], v should be 50 units, considering Vmax = 100 mol Pi·mg protein−1·min−1, Km = 0.1 mM, and S = 0.1 mM. If Pi transport through the cell (v) is kept constant as a consequence of adjusting the number of transporters in the plasma membrane to Pi,out, then only Vmax and Pi,out will change during adaptation of the Pi transport rate: then,
Table 1 shows the theoretical values of Vmax obtained with a different Pi,out according to Eq. 2. The real values obtained experimentally are shown next to them. The real values match perfectly with the theoretical data, thus validating our proposal. A representation of Vmax adaptation is also shown in Fig. 1C, which includes an Eadie-Hofstee linear transformation plot as an inset.
Inhibition of Pi transport with As(V) and phosphonoformate.
The inhibition of Pi transport by As(V) and PFA was also analyzed in our OK cells. A dose-response assay of As(V) showed a sigmoid inhibition of 50 μM Pi uptake, with a mean inhibitory concentration (IC50) of 0.29 mM (Fig. 2A). The inhibitory constant (Ki) was accurately determined with a global nonlinear regression fit to several saturating curves of Pi transport in the presence of multiple, fixed concentrations of As(V) (21). Figure 2B shows the different curve shapes of the global fit, which provides a Ki value of 120 ± 24 μM. The competitive behavior is also depicted by the two insets of Fig. 2B, namely a Dixon plot and the increase of Km values for Pi with increasing concentrations of As(V). With respect to PFA, the same inhibition kinetic assays were performed, and they are shown in Fig. 1, C and D. In this case, the affinity of PFA for the Pi transport system in OK cells was slightly lower, with a Ki value of 0.26 mM. The Ki values for As(V) and PFA did not change, regardless of whether the cells were previously adapted to a low or high Pi concentration (data not shown).
Pi transport by OK cell NaPi-IIa in X. laevis oocytes.
Pi transport in OK cells, as in the kidney, is mainly mediated through NaPi-IIa (formerly NaPi-4; 18). When expressed in Xenopus oocytes, NaPi-IIa exhibits an affinity for Pi similar to that of OK cells (89 ± 17 μM; Fig. 3A, Ref. 18). The pH dependence of NaPi-IIa transport is also identical to that of OK cells, in that it increases with increments of extracellular pH (data not shown; 18). Nevertheless, the inhibition of 50 μM Pi transport with increasing concentrations of As(V) or PFA provided kinetic constants that were different from the kinetic constants of OK cells and closer to those of rat NaPi-IIa orthologous (21): Ki values for As(V) and PFA, calculated using the corresponding IC50, were 0.56 mM (Fig. 3B) and 0.67 mM (Fig. 3C), respectively.
Effect of As(V) and PFA on Pi transport adaptation.
OK cells were adapted for 1 h (acute adaptation) to 0.1 (low), 1 (control), and 2 (high) mM Pi, and the uptake of 0.05 mM Pi was then determined to check for adaptation (Fig. 4A, open bars). The effects of As(V) and PFA were analyzed by preincubating the cells with 1 mM Pi, plus either As(V) (1 or 5 mM) or PFA (1 or 5 mM), and then measuring Pi uptake after washing the cells with uptake solution. As shown in Fig. 4A, the combination of 1 mM Pi plus As(V) induced changes in Pi transport that were similar to preincubation with 2 mM Pi. Conversely, when the cells were preincubated with 1 mM Pi plus PFA, the Pi transport rate increased to the level of 0.1 mM Pi. Therefore, whereas both As(V) and PFA are competitive inhibitors of Pi transport in OK cells, As(V) induces adaptation that is equivalent to high Pi (Vmax decrease), whereas PFA induces adaptation that is similar to low Pi (Vmax increase).
These findings on Pi transport were compared with the abundance in the plasma membrane of the only Na/Pi cotransporter cloned from OK cells (NaPi-IIa). A polyclonal antibody was prepared in rabbits, which recognized a specific band of ∼80 kDa (Fig. 4B). The membrane protein biotinylation of cells incubated for 1 h with 0.1 versus 2 mM PI, PFA, and As(V) revealed that the abundance of NaPi-IIa was proportional to the observed Pi transport changes (Fig. 4C). However, the total cell content of NaPi-IIa did not change, as expected of the acute adaptation mechanism (i.e., independent of de novo protein synthesis, Fig. 4D).
Kinetic characterization of the As(V) effect.
OK cells were adapted for 1 h to 0.1, 1, and 2 mM Pi, or 0.1 mM Pi plus increasing concentrations of As(V), and then they were assayed for 0.05 mM Pi uptake (Fig. 5A). As(V) progressively blunted the adaptive response of OK cells to 0.1 mM Pi in a dose-dependent way. The effect was already significant at 0.01 mM As(V), which showed the high sensitivity of the Pi transport system of OK cells to the metalloid salt. Given that the number of functional Pi transporters in the plasma membrane is equivalent to the Vmax of the Michaelian system, the effect of As(V) on Pi transport adaptation was more accurately analyzed using saturation kinetics (Fig. 5B). OK cells were incubated (adapted) for 1 h with 2 or 0.1 mM Pi and 0.1 mM Pi plus either 0.1 or 1 mM As(V). The data from the saturation kinetics were fit as in Fig. 1A, and the transport component is shown in Fig. 5B. The Vmax of the cells adapted to 0.1 mM Pi was 3.3 nmol Pi/mg of protein per minute, which was reduced to 1.5 nmol Pi/mg of protein per minute with 2 mM Pi. Preincubation of the cells with 0.1 mM Pi plus 0.1 mM As(V) dropped the Vmax by 27% (2.4 nmol Pi/mg of protein per min), whereas 1 mM As(V) completely blunted the adaptive response of the cells.
The time course of the As(V) effect was also assayed in cells adapted overnight to 0.1 mM Pi. The cells were incubated for different times with 2 mM Pi, 0.1 mM As(V), or 1 mM As(V), and then they were transport assayed with 0.05 mM Pi (Fig. 5C). All three conditions similarly diminished Pi transport, starting at 10 min of preincubation. Maximal inhibition of the adaptive response was obtained at 30 min of preincubation time, and longer preincubation times did not further inhibit Pi transport (Fig. 5D).
Reversibility of the As(V) effects.
To check that As(V) was not binding to the membrane or inhibiting Pi adaptation through mechanisms other than mimicking Pi uptake, cells that were adapted overnight to 0.1 mM Pi were then incubated for 4 or 60 min with 2 mM Pi or 10 mM As(V). The uptake of 0.05 mM Pi showed no effect after 4 min of incubation with Pi or As(V), but as expected, adaptation was suppressed after 60 min of incubation (Fig. 6A). We also studied the reversibility of the As(V) effect (Fig. 6B). Cells that were adapted overnight to 0.1 mM Pi (all bars) were incubated for 1 h with 2 mM Pi (left, solid and shaded bars) or 1 mM As(V) (right, solid and shaded bars). Some of the cells were then incubated for 4 h with 0.1 mM Pi (shaded bars), and the uptake assay showed that the cells had successfully readapted to 0.1 mM Pi, meaning that the effects of 2 mM Pi or 1 mM As(V) on cells adapted to 0.1 mM Pi are reversible.
Cytotoxicity of As(V).
The reversibility of the effects of Pi and As(V) on adaptation suggests that As(V) is not acting through cytotoxic mechanisms. Nevertheless, the possibility of cytotoxicity was discarded in several ways. First, total cell death was determined according to the activity of cytosolic lactate dehydrogenase in the cells treated with 1 mM As(V) for different times (Fig. 7A). When cells were incubated in a culture medium deprived of Pi, cell death was only significantly increased after 6 h of incubation. When cells were incubated with a culture medium containing a regular concentration of phosphate (0.92 mM phosphate in DMEM), cell death was only significant after 12 h of treatment with a As(V). This was confirmed by staining the cells with the fluorescent dyes ethidium bromide and acridine orange (Fig. 7B). Cells treated for 1 h with 10 mM As(V) did not show a significant increase in cell death, as shown by the absence of red cells from accumulated ethidium bromide (live cells only accept acridine orange, which appears green).
Given that As(V) could alter cell physiology or morphology at doses lower than those necessary for inducing cell death, the morphology of β-actin fibers was visualized using FITC-conjugated phalloidin after treatment with As(V) (Fig. 7B). Superficial actin was organized as patches (22), and As(V) did not induce any change, even when present at 10 mM.
In this paper we have studied the mechanism of functional adaptation to changes in Pi concentration in OK cells using As(V) and PFA as competitive inhibitors of Pi transport. Our work makes several contributions: 1) the affinity of OK cells for As(V) is 10 times higher (Ki 0.12 mM; Fig. 2, A and B) than the affinity in rat kidney (Ki 1.2 mM; Ref. 16); 2) As(V) is a competitive inhibitor of Pi transport and a substrate for Pi transporters, whereas PFA is also a competitive inhibitor, but it simply blocks the activity of Pi transporters because it is not transported (7, 25); 3) as a consequence of being transported into a cell, As(V) very efficiently prevents or reverses the adaptation of OK cells to a low Pi concentration (Fig. 5, A and C); 4) in PFA-treated cells, Pi uptake is decreased as a result of blocking the activity of type II Pi transporters, and consequently the cell activates the mechanism of adaptation to a low Pi concentration (Fig. 4A); and 5) whereas As(V) and PFA are competitive inhibitors of Pi transport (therefore they modify the affinity for Pi), the effects of As(V) and PFA on adaptation are manifested as changes in the capacity or the Vmax of the transport system, which is at least matched by changes in the abundance of the NaPi-IIa transporter in the plasma membrane of OK cells (Fig. 4C).
We have observed that the effects of As(V) in OK cells are not based on cytotoxicity. These effects can therefore be interpreted as a consequence of As(V) being carried by Pi transporters (21, 25). However, under the physiological concentrations of Pi in blood and ultrafiltrate (about 1 mM), the transport of As(V) by Pi transporters is nonsignificant, with the exception of the intestine (21). This conclusion is based on the low affinity of Pi transporters for As(V), compared with Pi, and it is based on the fact that the concentration of As(V) in blood is extremely low in usual exposures to the xenobiotic when compared with the millimolar range of Pi. The intestine is an exception because the Pi transporter NaPi-IIb shows high affinity for both Pi and As(V), and the concentration of Pi and As(V) mainly depends on ingested foods and liquids. Even in the presence of a high concentration of Pi, the absorption of As(V) could follow that of Pi. In OK cells, the As(V) Ki is also very low (0.12 mM), i.e., close to the Pi Km, which contrasts with the kidney as shown above, even though the OK cell is renally derived. The high affinity of As(V) transport in OK cells cannot be explained by the presence of any of the known type II and type III Pi transporters, because they all (except NaPi-IIb) show a high Ki when expressed in Xenopus laevis oocytes (Fig. 2; Ref. 21). It may be that NaPi-IIc, Pit-1, or Pit-2 from Didelphis virginiana (American opossum) are expressed in OK cells and exhibit a particularly high affinity for As(V) compared with the Ki values observed in the orthologous transporters of rodents or humans. Alternatively, OK cells may express a novel Pi transporter with an unusual affinity for As(V).
In addition to the transport activity, the OK cell mechanism of adaptation to the extracellular concentration of Pi also seems to be very sensitive to As(V), because at the low concentration of 0.01 mM, As(V) already reverses the adaptation induced by 0.1 mM Pi (Fig. 5A). This could be explained by the presence of a cellular sensor of Pi that has a greater affinity for As(V). This particular characteristic could be used in a strategy for the molecular identification of this sensor. In addition to high affinity, our data also suggest that the Pi sensor can only be intracellular, because PFA blocks Pi transport from the outside, and it induces adaptation to a low Pi even in the presence of a high concentration of Pi. The possibility that a Pi transporter is simultaneously a sensor for adaptation (i.e., depending on the transport rate) can also be excluded, because as little as 0.01 mM As(V) is able to reverse the adaptation induced with 0.1 mM Pi, and the change in the transport rate is negligible.
PFA mostly inhibits type II-mediated Pi transport, whereas As(V) is a poor substrate of type III Pi transporters (8, 21, 24, 25). Therefore, since Pi adaptation induced with PFA is mainly caused by the inhibition of Pi influx through NaPi-IIa and NaPi-IIc, this selectivity of effects could be used to analyze the involvement of the different transporters in this important physiological phenomenon. The reverse could be said about As(V), which is a poor substrate of Pit-1 and Pit-2. Downregulation of the adaptation to Pi deprivation with As(V) can be explained mainly by the influx through type-II Pi transporters. Using these tools [PFA and As(V)], the contribution by type II versus type III Pi transporters to acute adaptation mechanisms can be studied separately. Additionally, our results also provide evidence about the major contribution by type II transporters to Pi transport in OK cells, because treatment with PFA, which only blocks type II-mediated Pi transport, induces adaptation that is similar to a low Pi concentration. This occurs despite the total activity of type III Pi transporters, which are not inhibited by PFA at the concentrations used. This evidence agrees with another study under revision, in which we conclude that more than 90% of Pi transport is mediated through type II transporters.
It is noteworthy that all attempts to identify the intracellular sensor of extracellular Pi concentration in the proximal tubule have systematically failed. To successfully identify this sensor, it is critical to have detailed functional knowledge of the adaptation process. Therefore, this study also contributes by providing a mathematical definition of the changes in the Vmax that are observed during adaptation to the Pi concentration. To our knowledge, this is the first time that the generally accepted physiology of renal adaptation to Pi concentration has been described mathematically (Eq. 2). Our definition is based on the fact that the cell's aim is to maintain a constant transcellular transport of Pi molecules by adjusting the number of transporters to the extracellular concentration of Pi (Table 1 and Fig. 1C). Therefore, Vmax will be inversely proportional to the concentration of Pi, but proportional to the product of the actual velocity times the sum of Km plus the Pi concentration. Consequently, the interval for adjusting the Vmax of the transport system is very narrow, and independently of any extreme changes in the Pi concentration in a culture medium or in a diet, OK cells and proximal tubular cells only adapt the capacity or the Vmax of transport up to two times. In other words, cells that are adapted to 2 mM should show a 50% lower Vmax than cells adapted to a Pi concentration that is 20 times less (0.1 mM), as it has been demonstrated experimentally (Table 1). This limitation of Vmax adaptation has been reflected molecularly in many immunoblots of the Na/Pi cotransporters of rats and OK cells adapted to low and high Pi, because the Vmax is the consequence of the number of Pi transporters in the plasma membrane (e.g., 4, 12, 16, 23). This explains why the Vmax is very similar at 1 and 2 mM, and it also explains the difficulty of observing significant differences in NaPi-IIa immunoblots of OK cells or of rats adapted to a control (0.6–0.8% Pi diet) versus a high Pi diet (1.2% Pi; unpublished). The complexity of the renal control of Pi homeostasis is becoming more complicated (and interesting) than expected, considering that there are at least three different transporters that have the common denominator of adaptation to changes of the Pi content in a diet (23). It is therefore increasingly imperative to identify the common sensor of Pi concentration.
This work was supported by a grant from the Spanish Ministry of Education and Science (BFU2006-06284/BFI to V. Sorribas) and by a predoctoral fellowship from the Government of Aragón, Spain (B086/2007 to R. Villa-Bellosta).
- Copyright © 2009 the American Physiological Society