The general phosphate need in mammalian cells is accommodated by members of the Pi transport (PiT) family (SLC20), which use either Na+ or H+ to mediate inorganic phosphate (Pi) symport. The mammalian PiT paralogs PiT1 and PiT2 are Na+-dependent Pi (NaPi) transporters and are exploited by a group of retroviruses for cell entry. Human PiT1 and PiT2 were characterized by expression in Xenopus laevis oocytes with 32Pi as a traceable Pi source. For PiT1, the Michaelis-Menten constant for Pi was determined as 322.5 ± 124.5 μM. PiT2 was analyzed for the first time and showed positive cooperativity in Pi uptake with a half-maximal activity constant for Pi of 163.5 ± 39.8 μM. PiT1- and PiT2-mediated Na+-dependent Pi uptake functions were not significantly affected by acidic and alkaline pH and displayed similar Na+ dependency patterns. However, only PiT2 was capable of Na+-independent Pi transport at acidic pH. Study of the impact of divalent cations Ca2+ and Mg2+ revealed that Ca2+ was important, but not critical, for NaPi transport function of PiT proteins. To gain insight into the NaPi cotransport function, we analyzed PiT2 and a PiT2 Pi transport knockout mutant using 22Na+ as a traceable Na+ source. Na+ was transported by PiT2 even without Pi in the uptake medium and also when Pi transport function was knocked out. This is the first time decoupling of Pi from Na+ transport has been demonstrated for a PiT family member. Moreover, the results imply that putative transmembrane amino acids E55 and E575 are responsible for linking Pi import to Na+ transport in PiT2.
- inorganic phosphate transport
- retroviral receptor
in every cell phosphate is needed for structural and metabolic purposes. Inorganic phosphate (Pi) has to be actively transported across the cell membrane against a chemical and electrical gradient. In mammals this task is managed by the type III Na+-dependent Pi (NaPi) symporters PiT1 (SLC20A1) and PiT2 (SLC20A2), which utilize the free energy provided by the Na+ concentration gradient as the driving force to mediate the uphill import of Pi (1, 16, 17, 22, 38, 44).
The mammalian type III transporters are part of the Pi transport (PiT) family [SLC20 (43); TC #2.A.20 (29)], but several members were originally identified as receptors for retroviruses belonging to the gammaretrovirus genus (18, 19, 21, 37, 42, 45); thus PiT1 and PiT2 are proteins that exhibit dual functions.
The PiT paralogs are highly related, and human PiT1 and PiT2 share overall 62% amino acid sequence identity with a higher degree of homology at the NH2- and COOH-terminal ends (21, 42). Hydropathy profiles show that the PiT paralogs are polytopic proteins and predict 10–12 transmembrane domains (4, 9, 14, 32) and a large hydrophilic domain near the center of each protein that has been experimentally assigned to the cytoplasmic space in PiT2 (6).
PiT1 and PiT2 show a broad tissue distribution, both being expressed in all investigated human tissues, albeit at different levels (40), and low extracellular Pi levels can result in upregulated PiT1 and PiT2 expression in mammalian cells (7, 17), which, together with the current knowledge, albeit limited, on their transport characteristics, suggests that they constitute the housekeeping Pi uptake system between cells and the extracellular fluid (7, 16, 17, 40). This entails that PiT1 and PiT2 have to possess the abilities to adjust to a wide range of extracellular surroundings and organ-specific milieus in the body. Moreover, the PiT proteins also were recently found to be implicated in chondro- and osteoblastic mineralization (3, 20, 24) as well as the phenomenon of transdifferentiation of vascular smooth muscle cells to osteoblast-like cells in the process of vascular calcification under hyperphosphatemic conditions (reviewed in Ref. 11).
Despite the essential role of Pi in life processes and the specific roles of PiT1 and PiT2 in supplying mammalian cells with their general Pi need, as well as the increasing evidence for the role of type III transporters in normal and pathological calcification, little is known about the basic transport functions of the individual proteins. Using the Xenopus laevis oocyte as a “host” in the present study allowed us to investigate the basal capacities of either of the human PiT proteins in surroundings without the influence of uncontrollable variables coupled to cell line-specific characteristics dependent on, for example, organ of origin. No previous study has addressed the basic NaPi transport characteristics of human PiT2, but some of these characteristics already have been analyzed for human PiT1 (16, 17) and rodent PiT proteins (rat PiT1 and mouse and rat PiT2) (1, 17, 40).
The functional characterization of human PiT1 and PiT2 presented in this study shows that although they differ in their transport kinetics with PiT2, showing positive cooperativity in Pi uptake, they also behave remarkably alike. For example, they both support Na+-dependent Pi uptake at acidic and alkaline pH values, and their Na+ requirements for Pi transport also are similar. However, only human PiT2 supports Na+-independent Pi uptake, albeit at low levels, at acidic pH, underscoring another difference in the transport capacities among these two quite similar proteins. Moreover, Ca2+ or Ca2+ and Mg2+ were found to increase PiT1- and PiT2-mediated NaPi import, but neither cation was critical for the transport functions in that, compared with transport in the presence of both divalent cations, omitting Ca2+ and Mg2+ only resulted in an ∼50% reduction in Pi uptake for both proteins. Interestingly, in the absence of Ca2+, Mg2+ could partially substitute Ca2+ for PiT1 but not for PiT2, implying another difference in their Pi transport mechanisms. We also addressed the Na+ transport function versus the Pi import function of PiT2 and included a mutant PiT2 protein lacking the ability to sustain Pi uptake (5). These studies showed that wild-type PiT2 transported Na+ regardless of whether Pi was available for import or not. Moreover, the mutant PiT2 protein with knocked out Pi uptake function supported wild-type levels of Na+ transport, allowing us to suggest two putative transmembrane glutamate residues as being responsible for the coupling of Pi import to Na+ transport in wild-type PiT2.
MATERIALS AND METHODS
The pcDNA1- and pcDNA1ARtkpA-derived expression plasmids pOJ9 and pOJ75, respectively (Wyeth-Ayerst Research, Pearl River, NY), encode human PiT1 and have been described elsewhere (15, 25). The pcDNA1ARtkpA-derived expression plasmid pOJ74 (Wyeth-Ayerst Research) encoding human PiT2 and the PiT2-derived NaPi transport knockout mutant PiT2 E55Q E91Q E575Q also have been described previously (5, 25).
X. laevis oocytes.
Female X. laevis frogs were obtained from Nasco (Modesto, CA) and were kept and handled according to guidelines approved by the Danish National Committee for Animal Studies. Oocytes were isolated from frogs anesthetized in a 0.1–0.2% MS.222 (3-aminobenzoic acid ethyl ester; Sigma, St. Louis, MO) solution for 10–30 min. A 1- to 1.5-cm incision was made in the abdomen, and several ovaries were removed surgically by authorized personnel.
The oocytes were manually dissected and subsequently treated with collagenase (Sigma) and maintained in modified Barth's solution [88 mM NaCl, 1 mM KCl, 0.82 mM MgSO4, 0.4 mM CaCl2, 0.33 mM Ca(NO3)2, 2.4 mM NaHCO3, 10 mM HEPES-KOH, pH 7.5, 100 IU/ml penicillin, and 100 μg/ml streptomycin] at 18°C as described previously (5). The following day, the oocytes were ready for cRNA or H2O injection.
32Pi transport assays in X. laevis oocytes.
The 32Pi transport assays in X. laevis oocytes were essentially done as described previously (5). Briefly, restriction enzyme Bln1 or Not1 was used to linearize the pOJ74 plasmid, Bln1 or Apa1 was used to linearize the pOJ75 plasmid, and Sfi1 was used to linearize the pOJ9 plasmid. Notice that Apa1, Bln1, and Sfi1 cut in the vector sequence downstream of the polylinker and Not1 cuts in the polylinker immediately after the 3′-untranslated region (UTR) of PiT2. Moreover, except for a few bases, Bln1 and Sfi1 cut pOJ75 and pOJ9, respectively, in corresponding positions. Thereafter, we applied the mMESSAGE mMACHINE kit (Ambion, Austin, TX) to produce cRNAs according to the manufacturer's instructions. The quality of the cRNA preparations was verified by agarose gel electrophoresis before injection as well as on remaining samples of the diluted cRNA preparations used for injection.
Stage V and VI oocytes were microinjected with 12.5 ng of cRNA (or H2O as negative control) 1 day after surgery and incubated at 18°C. After 2 or 3 days, we applied a standardized Na32Pi transport protocol in which the oocytes were washed in Pi-free uptake solution (100 mM NaCl, 2 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and 10 mM HEPES-KOH, pH 7.5), and uptake was allowed for 1 h at 18°C in 0.1 mM KH232PO4 (2 mCi/ml; New England Nuclear, Boston, MA) diluted in Pi-free uptake solution. The oocytes were subsequently washed in ice-cold Pi-free uptake solution containing 5 mM KH2PO4, and 32Pi uptake of each oocyte was measured in a liquid scintillation counter as described previously (5). This protocol was modified in individual experiments as indicated.
For determination of the affinity constants Km and K0.5 for Pi, oocytes were incubated for 5 min in uptake solution containing KH232PO4 in concentrations varying from 1,000 to 7.8 μM. For determination of the pH-dependent Na32Pi uptake pattern, oocytes were incubated for 1 h in uptake solutions modified to reach pH values ranging from 4.5 to 5.5 and 6.5 with HCl and to 8.5 with KOH. More acidic conditions could not be analyzed because the oocytes did not tolerate pH 3.5.
Na+-independent 32Pi transport was analyzed at pH values ranging from 5.5 to 7.5 with Tris·HCl substituting for NaCl; pH was adjusted to 6.5 and 7.5 with KOH and to 5.5 with HCl, otherwise the uptake solutions were comparable to the ones used for pH-dependent Na32Pi uptake studies. The dependency of Na32Pi transport on Mg2+ and/or Ca2+ was analyzed by incubating oocytes in 0.1 mM KH232PO4 in Pi-free uptake solution without Mg2+ or Ca2+ or without both Mg2+ and Ca2+.
For analysis of Na+ dependency of 32Pi transport function, 32Pi uptake was analyzed using Na+ concentrations ranging from 0.5 to 100 mM. The uptake solution was supplemented with Tris·HCl at pH 7.5 to reach a final concentration of 100 mM Tris/Na+.
22Na+ transport assay in X. laevis oocytes.
Plasmids encoding PiT2 (pOJ74) and PiT2 E55Q E91Q E575Q were linearized with Not1, and cRNAs were made as described above. H2O or 25 ng of cRNA was injected into each oocyte, and after 3 days, 22Na+ uptake studies were performed in Pi-free uptake solution (100 mM NaCl, 2 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and 10 mM HEPES-KOH, pH 7.5) containing 5.8 μM 22NaCl (11.6 mCi/ml: Amersham Biosciences, Uppsala, Sweden). Moreover, 22Na+ uptake in the presence of Pi was analyzed in the same way by supplementing the uptake solution with 5 mM KH2PO4. Uptake was allowed for 1 h, and the oocytes were washed in ice-cold uptake solution. Fifteen oocytes were used per setup, and 22Na+ uptake was measured as gamma decay on pools of five oocytes in a Packard Cobra II Auto gamma counter (Packard Instruments, Canberra, Australia).
Kinetics and statistical analyses.
Substrate kinetics calculations for PiT1 were determined by nonlinear regression to the Michaelis-Menten equation: , where ν represents the initial rate of uptake at a given substrate concentration, Vmax is the maximal rate of uptake at saturating substrate concentrations, [S] is the concentration of substrate, and KmPi is the Michaelis-Menten constant representing the substrate concentration that produces one-half Vmax. Substrate kinetics calculations for PiT2 were determined by nonlinear regression to a sigmoidal dose-response (variable slope) equation: , where the slope h denoting the Hill coefficient reveals positive cooperativity if >1, and K0.5 represents the substrate concentration that produces one-half Vmax. Both analyses were performed using GraphPad Prism version 4.01 for Windows (GraphPad Software, San Diego, CA; www.graphpad.com).
For ordinary transport function comparison of two sets of data, the hypothesis that two mean values were identical was tested using a two-tailed Student's t-test. Values were considered different at a 95% confidence level.
Calculation of chemical speciation.
Calculation of the chemical speciation in the uptake solutions was performed using MINEQL+ version 3.01b Chemical Equilibrium Management System, distributed by Environmental Research Software (33, 34). For most uptake solutions used at a pH >5.5, the program predicts the presence of precipitates. However, no precipitates were evident in any of the uptake solutions used except at pH 8.5, where slight precipitation was formed after 3 days. Therefore, in all calculations precipitates were neglected. HEPES was not included in the calculations because it does not form complexes with the cations present in the uptake solutions (12); however, its contribution to the ionic strength was taken into consideration.
Kinetic analysis of human PiT1 and PiT2 expressed in X. laevis oocytes.
To determine the fundamental kinetic parameters of human PiT1 and PiT2 for Pi, we injected cRNAs encoding PiT1 and PiT2 or H2O into oocytes and measured Na32Pi uptake as a function of the extracellular Pi concentration. 32Pi concentrations in the range of 62.5–1,000 μM at pH 7.5 resulted in a dose-dependent approximately linear uptake of 32Pi by PiT1-expressing oocytes when the background uptake was not subtracted (Fig. 1A). After correction for background, the Michaelis-Menten constant for PiT1, KmPi, was determined to be 322.5 ± 124.5 μM (Fig. 1B). For PiT2, the data in Fig. 1A suggested cooperative substrate binding and were instead analyzed using the variable slope curve fitting (Fig. 1C), which gave a PiT2 K0.5 value for Pi (K0.5Pi) of 163.5 ± 39.8 μM and a Hill coefficient, h, of 2.0 ± 0.7. This is the first time K0.5Pi for human PiT2 expressed in oocytes has been determined. Vmax for PiT1 and PiT2 were determined to be 452.8 ± 71.7 and 449.5 ± 57.3 pmol·oocyte−1·h−1, respectively (Fig. 1).
The physiological level of Pi in the blood of healthy adults is in the range of 0.9–1.4 mM. However, circadian rhythms, age group, and health are all elements that affect the level of Pi in the blood; e.g., in infants it is between 1.5 and 2.4 mM (26, 36). Thus the KmPi and K0.5Pi values obtained for human PiT1 and PiT2 seem to correlate well with physiologically relevant values and are in agreement with these proteins being high-affinity Pi transporters.
Expression of PiT1 protein in X. laevis oocytes is influenced by enzyme choice in the cRNA preparation process.
An important aspect of working with protein expression in oocytes is the step of making cRNAs from plasmid preparations. At first, we used the enzymes Apa1 and Not1 for linearization of PiT1- and PiT2-encoding plasmids, respectively. Apa1 linearizes 549 base pairs (bp) downstream of the 126-bp-long PiT1-specific 3′-UTR fragment in pOJ75, whereas Not1 linearizes 13 bp downstream of the 541-bp-long PiT2-specific 3′-UTR fragment in pOJ74. However, to make relative comparisons of PiT2 and PiT1 more feasible, we switched to an enzyme (Bln1) applicable for both pOJ74 and pOJ75; Bln1 linearizes 845 bp and 306 bp downstream of the Not1 site in pOJ74 and the Apa1 site in pOJ75, respectively.
Comparison of PiT2 cRNAs transcribed from Bln1- and Not1-linearized pOJ74 with our standardized Na32Pi transport protocol resulted in PiT2-expressing oocytes supporting equal (P = 0.171) levels of Na32Pi uptake (see Table 1, experiment 1), thus both Not1 and Bln1 are suitable enzymes when preparing PiT2 cRNAs from pOJ74. Moreover, we compared PiT1 cRNA transcribed from Apa1-linearized pOJ75 with PiT2 cRNA from Not1-linearized pOJ74 (see Table 1, experiment 2) as well as PiT1 cRNA transcribed from Bln1-linearized pOJ75 with PiT2 cRNA made from Bln1-linearized pOJ74 (see Table 1, experiments 1 and 3). PiT1 cRNA from Apa1-linearized pOJ75 gave rise to oocytes transporting significantly lower levels of 32Pi than oocytes injected with PiT2 cRNA from Not1-linearized pOJ74 (P < 0.001) (experiment 2). However, cRNAs made from Bln1-linearized pOJ74 and pOJ75 plasmids resulted in 32Pi uptake levels that were not statistically significantly different (P = 0.921 and P = 0.662, experiments 1 and 3, respectively). Thus, for PiT1 cRNA production from pOJ75, we suggest that Bln1 is used for linearization of the plasmid.
It was somewhat surprising that the change of restriction enzyme from Apa1 to Bln1, both of which cut within the pOJ75 vector sequence, led to increased NaPi uptake, especially given that the entire vector sequence was dispensable for high expression from pOJ74 (plasmid digested with Not1). The qualities of the cRNAs were equally good before injection in oocytes; however, a possible explanation could be that cRNA stability after injection into oocytes was impaired when the cRNA was made from pOJ75 linearized with Apa1. On the basis of our observations, we urge that caution in general be taken when selecting an enzyme for cRNA production.
Human PiT2 displays linear NaPi uptake from 1 to 4 h when expressed in X. laevis oocytes.
The time course of Pi uptake in the presence of Na+ showed that linear transport of 32Pi supported by PiT2-expressing and H2O-injected oocytes occurred in the range from 1 to 4 h (Fig. 2). We chose 1-h uptake for future experiments, because this time span resulted in data with lesser variation than longer transport periods did (see SE on Fig. 2). Allowing oocytes to transport for 1 h resulted in a PiT2 uptake of 114.4 ± 20.6 pmol 32Pi/oocyte (Fig. 2), in agreement with our former results obtained with PiT2-expressing oocytes (Table 1, experiments 1 and 2) (4, 5).
PiT1 and PiT2 sustain NaPi transport function at acidic and alkaline pH.
One of the main roles the PiT proteins have to fulfill is to accommodate the housekeeping Pi need in all cells and tissues in the human body, and we therefore investigated the abilities of PiT1 and PiT2 to sustain Pi uptake at pH varying from 4.5 to 8.5 (Fig. 3). The average 32Pi uptake for PiT2 and PiT1 at pH 7.5 was 117.6 ± 16.8 and 114.5 ± 24.7 pmol·oocyte−1·h−1, respectively (Fig. 3, A and B). At all pH values, PiT1- and PiT2-expressing oocytes supported 32Pi uptake at statistically significantly higher levels than H2O-injected oocytes (P ≤ 0.005). For both PiT1 and PiT2, there were tendencies to optimal 32Pi uptake at pH 6.5 and 7.5; however, for both proteins, comparison of 32Pi transport obtained at different pH values to that obtained at pH 7.5 revealed no data sets that were statistically significantly different (for P values, see Fig. 3 legend). Repetition of the experiments at pH values ranging from 5.5 to 8.5 gave similar results and statistics (not shown). Thus PiT1 and PiT2 sustain NaPi transport function at various pH values, in agreement with their ubiquitous expression in all human cells and tissues investigated (6, 28, 40).
Comparison of the phosphate species present at pH 4.5 to pH 8.5 (Tables 2 and 3) and the uptake levels at these pH values (Fig. 3) imply that both proteins can support uptake of the H2PO4− and HPO42− phosphate species. Indeed, not much is known about how Pi is translocated through the PiT proteins. One aspect is whether the Pi ion interacts with a divalent cation to generate a transport complex. PiT family members from Escherichia coli are believed to transport Pi in the form of a soluble, neutral metal Pi (MeHPO4 Aq) complex with Mg2+, Ca2+, Co2+, Mn2+, or Zn2+ (2, 41). Some of the Pi circulating in the human blood is in complex with Ca2+ or Mg2+ (26); however, in the experiment performed at pH 4.5, the 32Pi uptake solution contains <0.1% of the phosphate at neutral metal complex forms (Tables 2 and 3), suggesting that they are not the transported species. However, there was a tendency to lower uptake levels at pH 4.5 and 5.5 compared with those at pH 6.5 and 7.5 for both PiT1 and PiT2, although they were not statistically significantly impaired compared with uptake at pH 7.5 (Fig. 3). We therefore investigated the impact of Ca2+ and Mg2+, which are the sole divalent cations present in our and others' (1, 17, 38) standard 32Pi uptake solutions, on the Na32Pi transport function of human PiT1 and PiT2.
Ca2+ is involved in NaPi transport function of PiT proteins.
We found that PiT1 and PiT2 exhibited robust, albeit impaired, 32Pi uptake levels in uptake solutions without Ca2+ and Mg2+, sustaining Pi uptake of ∼54% (PiT1) and 58% (PiT2) compared with uptake in the presence of both divalent cations (P < 0.001 relative to H2O-injected oocytes in all cases) (Fig. 4). (For Pi species present in uptake solutions without Ca2+ and/or Mg2+, see Table 4 and 5). Thus, although neither protein was strictly dependent on Ca2+ and/or Mg2+ for Pi transport, the presence of at least one of the cations increased their Pi uptake. Mg2+ was dispensable for NaPi transport function when Ca2+ was present, whereas lack of Ca2+ resulted in impaired Pi uptake (Fig. 4). Thus the preferred divalent cation for NaPi transport function of the PiT proteins is Ca2+. However, in the absence of Ca2+, PiT1, but not PiT2, managed to use Mg2+ (P = 0.016 for PiT1 32Pi transport in the presence of Mg2+ compared with the absence of both divalent cations). The indifference of PiT2 transport function on the lack of Mg2+ was reexamined, and similar results were obtained (results not shown). These findings are interesting, because they indicate differences in the mechanism of Pi transport supported by PiT1 and PiT2. Moreover, the high levels of Pi uptake by PiT1 and PiT2 at pH 4.5 (Fig. 3), together with the observed abilities of both proteins to support Ca2+- and Mg2+-independent Pi uptake (Fig. 4), suggest that the effect of the divalent cations is to modulate the Pi transport capacities of the proteins, leading to increased uptake of H2PO4− and/or HPO42−. Thus the mammalian PiT proteins have preserved the ability to employ divalent cations for transport of Pi, although they are not strictly dependent on them.
pH affects Na+-independent Pi transport of PiT2.
An inwardly directed Na+ gradient is utilized by the human PiT proteins to make the transport of Pi against the prevailing electrochemical gradient possible; however, there are examples of PiT family members from more primitive species that utilize a gradient created by the translocation of H+ to import Pi (8, 13). Thus, with an eye on evolution as well as the aspect of characterizing the transport function of human PiT2, we analyzed for Na+-independent 32Pi transport function in X. laevis oocytes. At pH 7.5, PiT2-expressing and H2O-injected oocytes supported comparable Na+-independent 32Pi uptake levels of 2.5 ± 1.0 and 1.8 ± 0.2 pmol·oocyte−1·h−1, respectively (P = 0.477) (Fig. 5A). Interestingly, human PiT2 expressed in oocytes was capable of transporting 32Pi in a Na+-independent manner, significantly different from, and at levels approximately two to five times higher than, H2O-injected oocytes at acidic pH values (P < 0.05 compared with H2O-injected oocytes at pH values different from pH 7.5) (Fig. 5).
This is the first time that a Na+-independent Pi transport function of human PiT2 at acidic conditions has been reported, and it suggests that H+ is used to generate a gradient that drives Pi import. Three PiT family members from plant and bacteria sustain H+-dependent Pi transport (8, 13), and whether this ability of human PiT2 is reminiscent from the evolution of the PiT proteins or constitutes a physiological relevant function is an open question.
PiT1 lacks Na+-independent Pi transport function.
We also investigated whether human PiT1 could sustain Na+-independent 32Pi transport function. Oocytes were injected with cRNAs from pOJ75 (Fig. 6A) or, in an independent experiment, with cRNAs transcribed from pOJ9 (Fig. 6B). Surprisingly, PiT1-expressing oocytes were not able to import 32Pi in a Na+-independent manner significantly better than H2O-injected oocytes at pH 5.5 (Fig. 6). However, oocytes injected with PiT1 cRNA made from pOJ75 plasmid did import 32Pi at a level significantly different from that of H2O-injected oocytes at pH 6.5 (P = 0.010) (Fig. 6A), but the difference in transport was less than twofold and was not observed in the experiment shown in Fig. 6B; thus PiT1 is unable to import Pi at acidic pH in a Na+-independent manner. Our study confirms another study analyzing human PiT1 in MDTF cells at pH 6.5 (22).
Na+ levels required for Pi transport function of PiT1 and PiT2 are the same.
Na+ is transported by the type III NaPi transporters by utilizing an inwardly directed gradient, which then drives the cellular import of Pi. The observation that PiT1 and PiT2 exhibited different Na+-independent Pi uptake characteristics led us to investigate the relationship between Na+ and Pi, focusing on low levels of Na+. Analyzing the 32Pi uptake as a function of Na+ concentration (ranging from 0.5 to 100 mM) showed a clear relationship between the rate of 32Pi transport and the concentration of Na+ for both proteins, and they supported comparable levels of 32Pi uptake at the different Na+ concentrations (Fig. 7B). The experiment was repeated for PiT2 at 5 and 100 mM Na+ with similar results (not shown). This is the first time the Na+ dependencies of the Pi transport function have been analyzed for human PiT1 and PiT2; however, mouse PiT2 has been analyzed the same way, and a similar relationship was observed (1).
Interestingly, we found that 10-fold dilutions of Na+ only affected the 32Pi uptake supported by PiT1 and PiT2 to a slighter degree. Thus, at 100 mM Na+, both PiT1 and PiT2 sustained uptake of ∼120 pmol 32Pi·oocyte−1·h−1, whereas at 10 mM Na+, the 32Pi uptake was ∼4–6 times lower, and reducing the Na+ concentration 10-fold to 1 mM only resulted in 32Pi uptake ∼2–2.5 times lower (Fig. 7). Hence, even at low levels of Na+, the PiT proteins are able to sustain a remarkable level of Pi uptake. Because the physiological Na+ concentration in the human body is in the range of 130–145 mM, in vivo there always will be plenty to sustain Pi uptake via PiT1 and PiT2. Thus, whether the ability to sustain Pi import at low levels of Na+ is a relic from the past or does in fact constitute a physiologically relevant function of PiT1 and PiT2 remains to be investigated.
Na+ and Pi symport functions can be uncoupled in PiT2.
Na+ transport and Pi import are the two active events conducted by the PiT proteins, so to gain insight into the NaPi cotransport function, we analyzed PiT2 and a PiT2-derived Pi transport knockout mutant protein (PiT2 E55Q E91Q E575Q; mutant QQQ in Ref. 5) in oocytes by using 22Na+ as a traceable Na+ source. The mutant protein PiT2 E55Q E91Q E575Q lacks Pi transport function (Fig. 8B) (5), but the retroviral receptor function for two PiT2 cognate viruses is intact, implying that the mutant protein is correctly processed and that overall topology is preserved (5).
Interestingly, we found that 22Na+ was transported by wild-type PiT2 in both the absence and presence of Pi and at comparable levels (P = 0.363) (see Fig. 8A legend). This is in agreement with the observation that the PiT2 Pi transport knockout mutant supported 22Na+ transport at the same levels in the presence or absence of Pi (P = 0.143) (see Fig. 8A legend). These results thus demonstrate in two ways that the transport of Na+ by PiT2 can occur without subsequent import of Pi.
Moreover, comparison of the 22Na+ uptake levels obtained for wild-type PiT2 and PiT2 E55Q E91Q E575Q in the presence of Pi (P = 0.928) and in the absence of Pi (P = 0.732) revealed that both proteins supported comparable levels of 22Na+ uptake whether Pi were absent or present. Thus, in a Pi-rich environment, the Pi transport knockout PiT2 E55Q E91Q E575Q supported the same amount of 22Na+ transport as the wild-type PiT2 in an environment devoid of Pi (Fig. 8), showing that Na+ transport of human PiT2 can be uncoupled from Pi transport.
We have investigated a variety of physiological features of the human Na+-dependent Pi transporters PiT1 and PiT2 and used the X. laevis oocyte as expression system. Using the oocyte as “host” allowed us to investigate some basal capacities of the PiT proteins in surroundings without the influence of uncontrollable variables coupled to cell line-specific characteristics.
Whereas rat and mouse PiT2 have been analyzed previously in oocytes, this is the first report on human PiT2. Interestingly, our data reveal positive cooperativity (Hill coefficient of 2) in human PiT2 Pi transport function. This is in agreement with the observations that human PiT2 can be at dimeric form (5) and that Pi starvation promotes human PiT2 dimerization, suggesting that the transporting form is a dimer (30). We determined the K0.5Pi value of human PiT2 to be 163.5 ± 39.8 μM; however, rodent PiT2 were reported to obey Michaelis-Menten kinetics, and their KmPi values were determined to be 25.3 ± 6.0 μM (rat) (17) and 38 ± 12 μM (mouse) (1), respectively. The disparities regarding human and rodent PiT2 might be due to amino acid differences, because the overall homology between human and mouse PiT2 is 91% and that between mouse and rat PiT2 is 95% (1).
We determined the PiT1 Km value for Pi to be 322.5 ± 124.5 μM, and others have determined KmPi values for human (17) and rat PiT1 (38) in oocytes to be 24.1 ± 5.5 and 89 ± 13 μM, respectively. It cannot be excluded that the difference in KmPi values between human (present work) and rat PiT1 (38) reflects amino acid differences (human and rat PiT1 exhibit 90% overall sequence homology). Moreover, the KmPi value for human PiT1 determined in the present study and that determined by Kavanaugh et al. (17) differ by >13-fold. Kavanaugh et al. performed NaPi uptake in the presence of 1.8 mM CaCl2, whereas we applied 1.0 mM CaCl2, and because we found that human PiT1 can sustain Pi uptake without the presence of Ca2+ (or Mg2+) but that the presence of either at 1 mM increased human PiT1 Pi uptake (Fig. 4), we cannot exclude the possibility that the difference in Ca2+ levels contributes to the differences in KmPi values obtained for human PiT1.
The PiT1 KmPi and PiT2 K0.5Pi values we obtained in oocytes are in good agreement with former studies on endogenously expressed human PiT proteins, which revealed KmPi values of 454 ± 61 μM in human osteoblastic SaOS2 cells (23), 280 ± 30 μM in human primary erythrocytes (35), 360 μM in human erythroid precursor-like K562 cells (39), 110 ± 20 μM in human cervical epithelial-like HeLa cells (27), and 147 ± 14 μM in human embryonic kidney HEK-293 cells (10). Although Fernandes et al. (10) assigned their KmPi values in HEK-293 cells to PiT1, their cells most likely also expressed PiT2 mRNA, because another study (40) showed its presence in HEK-293 cells and kidney. These results are in agreement with the previous assignment of PiT1 as the major NaPi transporter in osteoblastic cells (20, 23) and type III transporters as the housekeeping NaPi transporters in human cells (16, 17, 40).
We found that human PiT1 and PiT2 sustained substantial Pi uptake in oocytes at pH values from 4.5 to 8.5 with a tendency to pH optima at 6.5 and 7.5 for both proteins. Human PiT1 NaPi uptake was previously investigated in oocytes at pH 7.5 and 8.5, and a higher reduction in uptake at pH 8.5 (55%) compared with pH 7.5 was observed (17). As found in the present study for human PiT2, mouse PiT2 was reported to sustain comparable levels of Pi transport in oocytes at pH 5.5–7.5. Moreover, mouse PiT2 also supported substantial levels of Pi uptake at pH 8.0 and 8.5, albeit significantly impaired compared with transport at pH 7.5 (1). The same tendency was observed for rat PiT2 at pH 7.5 and 8.5 (17); however, the Pi uptake levels in rat PiT2-expressing oocytes were in general surprisingly low (17) compared with those supported by human PiT2 (4, 5) (present work) and mouse PiT2 (1), hampering the full interpretation of these data. In mammalian cell lines originating from kidney, ovary, and connective tissue, the pH optima of the PiT proteins did not differ substantially from those found in oocytes; however, severely impaired uptake at pH 8.0 and 8.5 was observed (10, 22, 44). This underscores the difficulties in addressing general characteristics of proteins in mammalian cells, where a cell type-specific behavior might determine the outcome of the protein characterization.
Human PiT1 and PiT2 supported comparable levels of Pi uptake at Na+ concentrations varying from 0.5 to 100 mM when analyzed in oocytes. This Pi uptake was Na+ dose dependent and in agreement with results from similar studies of mouse PiT2 (1). Voltage-clamp studies of oocytes expressing rat PiT2 and human PiT1 proposed transport of 2 Na+:1 H2PO4− or 3 Na+:1HPO42− (16); our analyses of pH-dependent uptake indeed support the suggestion that both H2PO4− and HPO42− can be transported. Moreover, our 22Na+ uptake studies also are in agreement with uptake of a surplus Na+ to Pi. We also found that the PiT proteins are capable of sustaining striking levels of Pi uptake at low levels of Na+; however, in the human body, the physiological Na+ levels is in the range of 130–145 mM, and thus it is unlikely that this ability is of physiological relevance.
The cotransport of Na+ and Pi by the PiT proteins are the two actions responsible for Pi import into cells. Human PiT2 residues D28, D506, E55, and E575 (4, 5) as well as S113 and S593 (31) were found to be critical for Pi transport function of PiT2 (4, 5, 31). The present study is the first to address the Na+ transport function of a PiT protein. We found that wild-type PiT2 was capable of transporting Na+ without the presence of Pi, emphasizing that Pi and Na+ transport occur as two distinct functions. However, a full understanding of the implication and relevance of this finding, i.e., whether PiT2 exhibits substrate-independent Na+ uptake, a phenomenon referred to as slippage, in its physiological environment in the mammalian cell, demands more investigations. Nevertheless, the analyses of a Pi transport knockout PiT2 mutant protein, PiT2 E55Q E91Q E575Q, did provide knowledge on the mechanism of the Na+ transport, showing that it can occur even though the Pi uptake function is abolished. The glutamates in human PiT2 positions 55, 91, and 575 are believed to locate to transmembrane domains (4, 9, 14, 32). A former study (5) revealed that E91 is dispensable, whereas E55 and E575 are critical for Pi transport of PiT2. The putative location of E55 and E575 make them unlikely to fulfill structural roles in the PiT proteins, and instead we have previously suggested that they were involved in function-dependent conformational changes or being parts of a ligand site for the coupling cation (5). However, the data presented in the present study clearly show that they are not participants in the Na+ translocation pathway. Rather, we suggest that E55 and E575 are involved in the coupling of Na+ transport to Pi uptake. Noticeably, the conserved nature of the glutamates in PiT2 positions 55 and 575 among members of the PiT family, including in those dependent on H+ for Pi uptake, suggest that the residues have similar roles in other PiT family members.
The modulation of PiT1- and PiT2-mediated Pi uptake by Ca2+ within 1 h of exposure is interesting. The Ca2+ concentration used in the present experiments is slightly lower than the free serum Ca2+ level (1.1–1.3 mM) in adult healthy individuals (26). We found an increase in Pi uptake by adding 1 mM Ca2+ to the uptake solution; however, it is tempting to speculate that Ca2+ plays a general regulating function in Pi uptake via PiT1 and PiT2, because serum Ca2+ levels are tightly controlled and serum Pi levels are not. Whether the effects of Ca2+ and Mg2+ occur directly on the PiT proteins is presently not known.
In summary, we have in the present study characterized human PiT1 and PiT2 transport functions at different levels. We have determined kinetic constants of both proteins for Pi and found that they correspond well to KmPi values determined for endogenously expressed human PiT proteins. We also have found that both proteins supported NaPi uptake at pH 4.5–8.5, but only PiT2 was capable of HPi uptake at acidic pH. Moreover, Ca2+ increased both PiT1- and PiT2-mediated Pi uptake in a manner suggesting a regulatory function of Ca2+. Dissection of the transport function of human PiT2 revealed glutamate residues critical for the coupling of Pi import to Na+ transport.
This work was supported by Lundbeck Foundation Grants 14/02 and 68/04, the Novo Nordisk Foundation, Danish Medical Research Foundation Grants 9802349 and 22-03-0254, the Karen Elise Jensen Foundation, and Danish Cancer Society Grant DP00092.
We thank Bryan O'Hara for pOJ74, pOJ75, and pOJ9, and Jan Egebjerg Jensen for use of the X. laevis oocyte facilities.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society