Inorganic phosphate transporter type II (a, b, or c);  NaPi-IIa;  NaPi-IIb;  NaPi-IIc;  NPT2a;  NPT2b;  NPT2c;  Sodium/phosphate cotransporter type 2;  Solute carrier family 34 (sodium phosphate)

Phosphorus is one of the six essential elements for life. Molecules containing anionic phosphate (PO₄³⁻) are constituents of genetic material (DNA, RNA) and are an essential building block of phospholipid membranes, bones, and teeth; high-energy phosphate bonds drive cell energetics via ATP/ADP hydrolysis; phosphorylation/dephosphorylation reactions are key events in intracellular signaling, and phosphate acts as an intra- and extracellular pH buffer. In mammals, phosphate is obtained from the diet in the form of inorganic phosphate (P_i) that exists in solution as negatively charged mono- (H₂PO₄⁻) and divalent (HPO₄²⁻) ions in the physiological pH range. As such P_i must be actively transported “uphill” across the cell membrane from the external medium against its electrochemical gradient.

Products of two human gene families (SLC20 and SLC34) are known to mediate transmembrane P_i transport. These are secondary-active carrier proteins (cotransporters) that are driven by the Na⁺ concentration gradient and the transmembrane electric potential. Whereas those encoded by the SLC20 gene family (SLC20A1 (PiT1), SLC20A2 (PiT2)) are ubiquitously expressed and are considered to fulfill a “housekeeping” role (see Sorribas, this volume); those encoded by the SLC34 gene family (in humans: SLC34A1 (NaPi-IIa), SLC34A2 (NaPi-IIb), and SLC34A3 (NaPi-IIc)) play vital roles in P_i homeostasis and are expressed primarily at the luminal (or apical) side of epithelial or epithelial-like tissue.

NaPi-IIa and NaPi-IIc are found principally in the proximal tubule of the nephron, where they are responsible for P_i reabsorption from the glomerular filtrate. NaPi-IIa is also expressed in the basolateral membrane of osteoclasts, where it is thought to be involved in recovering P_i released from bone resorption (Khadeer et al. 2003). NaPi-IIb shows a more widespread expression pattern, being highly expressed in enterocytes of the small intestine (NaPi-IIb), where it mediates dietary P_i absorption and mRNA for NaPi-IIb is detected in the testes, lung, pancreas, mammary and salivary glands, liver, and uterus (Nishimura and Naito 2008). In the lung NaPi-IIb is thought to clear the alveolar space of P_i that accumulates as a result of surfactant turnover, whereas the specific roles of NaPi-IIb in other tissues have yet to be elucidated. A potential candidate for the molecular pathway mediating “downhill” P_i exit into the plasma across the epithelial basolateral membrane has been recently identified as the retroviral receptor XPR1 (Giovannini et al. 2013).

Studies in the pre-cloning era identified vectorial Na⁺-dependent P_i transport processes at the luminal membrane of proximal tubule and small intestine using the brush-border membrane preparation (for review of these earlier studies, see Murer et al. 2000). These seminal investigations established the existence of a Na⁺-dependent, secondary-active transport mechanism for P_i and underscored the importance of these transport pathways in the physiological context for P_i homeostasis. With the advent of expression cloning, beginning with NaPi-IIa cloned from rat and human tissue in 1993, followed by NaPi-IIb cloned from mouse in 1998 and finally NaPi-IIc from mouse tissue in 2002, the functional properties of each isoform could be studied independently using a variety of laboratory techniques (for review see Forster et al. 2002, 2012). These include tracer flux, electrophysiology, and fluorometry. At the whole animal level, elucidation of multiple regulatory pathways and identification of circulating factors involved have benefitted from having isoform- and species-specific SLC34 antibodies. More recently, studies using knockout animals (conditional and non-conditional) have allowed deeper insight into the isoform-specific roles in achieving P_i homeostasis in vivo (e.g., Lederer and Miyamoto 2012). Moreover the discovery of mutations in SLC34 protein in humans with P_i homeostasis dysfunction has led to further understanding of their physiological roles in a clinical context (Lederer and Miyamoto 2012; Wagner et al. 2014).

All three SLC34 isoforms preferentially transport divalent P_i (HPO₄²⁻) in a Na⁺-dependent manner; however a key functional difference between NaPi-IIa/b and NaPi-IIc is that the former are electrogenic and the latter is electroneutral. In both cases, there is a strict stoichiometric relationship between Na⁺ and P_i, namely, 3:1 for the electrogenic isoforms together with one net positive charge transported per transport cycle and 2:1 for the electroneutral isoform with zero net charge translocation. The stoichiometry of the three isoforms was determined by a combination of electrophysiology and simultaneous uptake (³²P and ²²Na) assays (Forster et al. 1999; Bacconi et al. 2005; Andrini et al. 2012). The P_i concentrating capacity of NaPi-IIa/b is approximately 100-fold greater than NaPi-IIc under physiological conditions (∼10:1 Na⁺ gradient and −60 mV transmembrane potential). The physiological basis for having both electrogenic and electroneutral isoforms expressed in the same organ has not been satisfactorily established, albeit that their regulation patterns are different. One possible rationale may relate to the reported higher NaPi-IIc expression during development that would place a lower energetic demand on renal epithelia.

The charge translocated by the electrogenic isoforms gives rise to a P_i-dependent transmembrane current that can be readily measured by means of the voltage-clamp technique (Forster et al. 2002, 2013). Under these experimental conditions, stoichiometric coupling implies that the P_i-dependent current is a direct measure of the transport velocity, assuming a fixed number of active transporters. This property has allowed detailed investigation of the transport kinetics in real time by means of electrophysiological assays, which have used the Xenopus laevis oocyte expression system almost exclusively.

The most important properties are now described in brief (see Table 1 and the following reviews: (Forster et al. 2002; Andrini et al. 2008; Forster et al. 2012)). The transport rate of SLC34 proteins is strongly pH dependent. This arises in part from the titration of P_i (at normal physiological pH ∼7.4 the ratio HPO₄²⁻/H₂PO₄⁻ is 4:1), and this ratio decreases significantly under acidosis. In addition, protons are known to interact directly with NaPi-IIa/b and thereby modulate the kinetics of partial reactions in the transport cycle (see Fig. 1). The apparent affinity for P_i (K_0.5^Pi) is <0.1 mM but varies somewhat among isoforms and species. For example, mammalian NaPi-IIb can display a K_0.5^Pi < 10 μM, which means that significant transport rates could still be achieved at acidic pH despite the reduced availability of the preferred substrate. The apparent Na⁺ affinities (K_0.5^Na), determined under saturating P_i (1 mM), are typically 40–50 mM, and, given the cooperative nature of Na⁺ interactions, this ensures high transport rates at physiological Na⁺ concentrations. The apparent substrate affinities are found to be similar for the electrogenic and electroneutral isoforms from the same species. This implies that for all isoforms the phenomenological constants K_0.5^PiK_0.5^Na are determined by the same molecular entities attributable to substrate interaction. Kinetic studies have shown that substrate interaction is ordered for inward P_i transport: two Na⁺ ions precede P_i binding, followed by a third Na⁺ ion before translocation can occur. This binding sequence is thought to hold even for NaPi-IIc, whereby the first Na⁺ is still able to interact but is prevented from translocation (Ghezzi et al. 2009) (Fig. 1).

SLC34, Table 1

Selected functional properties of SLC34 proteins expressed in Xenopus oocytes

Isoform	NaPi-IIa (SLC34A1)	NaPi-IIb (SLC34A2)	NaPi-IIc (SLC34A3)
Pi species preferred	HPO42−	HPO42−	HPO42−
Driving cations	Na+ (Li+)	Na+(Li+)	Na+
Electrogenicity (charge/cycle)	1	1	0
Stoichiometry (Na:Pi)	3:1	3:1	2:1
K0.5Pi (μM)a	~50	7; 29 (zf1); 31 (fl); 250(zf2)	70–80
K0.5Na (mM)b	40–50	25; 42(zf2); 46(fl); 67(zf1)	43–48
pH dependence (pH↓)	Strong	Strong, weak(zf2)	Strong
Transport rate (s−1) (−60 mV)	4	9; 13(fl)	n.d.
Concentrating capacity (10:1 Na gradient, −60 mV)	10,000:1	10,000:1	100:1
Arsenate (KiAs) (mM)	1.08	0.05	1.01
PFA (KiPFA) (mM)	1.05	0.16	0.9

^aDetermined with 100 mM Na⁺, pH 7.4

^bDetermined with 1 mM P_i, pH 7.4

All values relate to mammalian isoforms unless otherwise indicated. zf1 zebrafish NaPi-IIb1, zf2 zebrafish NaPi-IIb2, fl flounder NaPi-IIb, n.d. not determined, PFA phosphonoformic acid. A more complete table can be found in (Forster et al. 2012) with citations to original literature

Open image in new window

SLC34 proteins are highly specific for transporting P_i but can also transport arsenate at low rates, (Villa-Bellosta and Sorribas 2008). Li⁺ ions can substitute for the first of the three Na⁺ ions that bind and are subsequently translocated (Ghezzi et al. 2009; Andrini et al. 2012) although at a lower rate than Na⁺. However, if only Li⁺ ions are present, no P_i significant transport activity is detected. Currently no other transported substrates have been identified or characterized among commercially available phosphate analogs and compounds. A number of proprietary compounds have been reported that show strong inhibition in a noncompetitive manner (e.g., Weinstock 2004). The only commercially available reversible inhibitors are phosphonocarboxylates, with phosphonoformic acid (PFA) (foscarnet) being the most effective; however nephrotoxic and other secondary effects have limited their usage (e.g., Loghman-Adham 1996). Like arsenate, they act competitively at the P_i binding site with a typical inhibition constant (K_i^PFA) of ∼1 mM (NaPi-IIa) and 0.16 mM (NaPi-IIc) (Villa-Bellosta and Sorribas 2008). It is important to note that in its role as a P_i transport inhibitor, PFA is an effective inhibitor of SLC34 transporters, consistent with its specificity for divalent P_i but shows negligible inhibitory effect on members of the SLC20 family (PiT1 and PiT2) at similar concentrations ((Ravera et al. 2007); see also Sorribas, this volume).

Finally, like many solute-driven transporters, SLC34 proteins exhibit a cation (Na⁺) leak that is detectable in the absence of extracellular P_i and is blocked by PFA (Andrini et al. 2008). This would not normally be of physiological significance except at P_i concentrations well below the K_0.5^Pi, where the probability of P_i binding and translocation is reduced. Nevertheless, structure-function studies have revealed how single point mutations can result in significantly increased Na⁺ leak (Andrini et al. 2008) that could have important functional consequences in a physiological setting. It should be noted that Xenopus laevis oocyte assays are generally performed at ∼20 °C and that the temperature dependence of the kinetic parameters would have to be taken into account when extrapolating the data to mammalian systems at 37 °C.

Further understanding of the transport mechanism at the molecular level has come from so-called presteady-state assays applied to the electrogenic isoforms (Forster et al. 2002, 2012). When rapid changes in transmembrane voltage are imposed across the membrane of oocytes expressing either electrogenic isoform, current relaxations are observed that are superimposed on the normal capacitative transients associated with altering the charge on the cell membrane. Detailed analyses of these NaPi-IIa/b-associated relaxations show that they arise from displacement of charges intrinsic to the transporter as well as the movement of a single Na⁺ ion/protein across part of the transmembrane electric field (e.g., Forster et al. 2012). Thus the charge movements reflect voltage-dependent conformational changes that confer voltage dependence to transport at the macroscopic level. The charges involved are small, amounting to an effective charge of 1 e⁻/transport cycle/protein. As expected, for the electroneutral NaPi-IIc, no presteady-state relaxations are detectable, although the first Na⁺ ion in the binding sequence is thought to initiate the subsequent steps in the transport cycle but is not translocated and remains bound to the protein. Combining data from steady-state and presteady-state assays allows the number of active transporters to be estimated and the turnover (transport) rate can be predicted (e.g., Forster et al. 2012). Although there appear to be isoform- and species-specific differences, the transport rate determined using this approach is <100 s⁻¹, consistent with rates determined from other Na⁺-driven cotransporters.

Taken together, the steady-state and presteady-state functional assays have led to a kinetic scheme that accounts for most of the experimental data (Fig. 1). The transport cycle can be considered as a sequence of transitions between unique conformational states, as substrates are bound, translocated, and then released. The interconversions between states are determined by hypothetical rate constants (forward and backward) associated with the partial reactions between any two states, some of which are voltage dependent and/or substrate activity dependent and hence determine the probability of state occupancy in the steady state. Although highly simplified, this model has provided a useful basis for understanding the transport mechanism, and numerical solutions have allowed the assignment of structural entities to specific partial reactions (e.g., Patti and Forster 2014; Patti et al. 2016).

SLC34 proteins are functional monomers (Kohler et al. 2000) but most likely assemble in the membrane as dimers, based on evidence from biochemical assays, freeze fracture studies (Forster et al. 2006), and, more recently, the structural similarity to the dimeric VcINDY (Vergara-Jaque et al 2015b), a member of the dicarboxylate transporter family. The molecular weight of each monomer varies slightly among isoforms and lies in the range 75–90 kDa. The predicted secondary topology of SLC34 proteins has evolved from a basic eight transmembrane domain model originally proposed (Magagnin et al. 1993) to the current scheme that is characterized by an inverted repeat topology (Fig. 2) (Fenollar-Ferrer et al. 2014).

Open image in new window

Both N- and C-termini are intracellularly located, and a large extracellular loop contains two N-glycosylation sites and features a pair of critical cyteines that form an essential disulfide bridge linking the N-terminal and C-terminal halves (Kohl et al. 1998). A comparison of the hydropathy profile of the three slc34 isoforms with the bacterial homolog from Vibrio cholera suggested that a candidate transport pathway might be formed by regions within each half where the hydropathy predictions are less well defined and thus would allow substrate accessibility from either side of the membrane (e.g., Forster et al. 2002). Structure-function studies have combined bioinformatics and functional assays using the substituted cysteine accessibility (SCAM) technique, cross-linking, chimeras, and site-directed mutagenesis to identify functionally important regions (for review see Forster et al. 2002, 2012). Of particular note among these studies was the identification of a charged amino acid (D224 in human NaPi-IIa) whose presence is essential for electrogenic transport (Bacconi et al. 2005) (Fig. 2). When this aspartic acid together with two neighboring alanines is substituted at the equivalent sites in the electroneutral NaPi-IIc, electrogenic behavior is reestablished, including charge relaxations and uncoupled leak (Bacconi et al. 2005).

Currently no 3-D crystal structure of SLC34 proteins or their bacterial homologs is available; however recent homology modeling studies (Fenollar-Ferrer et al. 2014, 2015) suggest that they most likely display a 3-D architecture similar to a dicarboxylate cotransporter VcINDY (Vergara-Jaque et al. 2015b), a member of the divalent anion/Na⁺ symporter (DASS) family (TCDB 2.A.47). Homology modeling using the VcINDY template has allowed prediction of residues involved in substrate coordination, and these predictions were confirmed by analyzing the functional consequences of making single point mutations at these sites (Fenollar-Ferrer et al. 2014, 2015).

Interpretation of structural and functional data has led to the proposal of a so-called elevator mechanism to achieve substrate translocation. In this model the transport domain is predicted to move through both a rotation (∼37°) and vertical displacement (15 Å) relative to the oligomeric interface domain that acts as fixed scaffold (Patti et al. 2016) (Fig. 3). Substrates are first bound in the outward-facing conformation and have no direct access to the cytosol, thereby satisfying the alternating access criterion to which all solute-coupled transporters must adhere. When fully loaded, the transport domain can now translocate the substrates, and in the inward-facing conformation they are released to the cytosol. Membrane potential is predicted to accelerate the return to the outward-facing conformation (transition 0–1, Fig. 1), possibly through the movement of charges intrinsic to the protein, one of which is thought to be the previously identified D224, or its equivalent (Fig. 2). The relatively large movements predicted by this model have been detected by applying the voltage-clamp fluorometry (VCF) technique in which fluorophores report changes in their microenvironments in response to protein conformational changes induced by membrane potential and cation interaction (Patti et al. 2016).

Open image in new window

In mammals, P_i dietary intake and excretion (via urine and feces) are balanced, and the circulating plasma P_i concentration lies in the range 0.8–1.4 mM. The plasma P_i level is determined by intestinal P_i absorption, renal P_i reabsorption/secretion, and skeletal P_i during bone resorption/formation. Intestinal P_i absorption includes the active, NaPi-IIb-mediated component and a passive, Na⁺-independent, paracellular movement of P_i down its electrochemical gradient (e.g., Marks et al. 2013). The kidney plays the vital role as principal plasma P_i regulating organ. Along the proximal tubule 80–90% of filtered P_i is reabsorbed primarily via NaPi-IIa, NaPi-IIc (and possibly Pit2 (SLC20A2), and the remainder is excreted in the urine. The physiological roles for the three SLC34 isoforms as the targets for regulation have been confirmed by a number of knockout animal studies (see below).

Both P_i intestinal absorption and renal reabsorption/excretion are controlled by circulating metabolic and hormonal factors that ultimately determine the abundance of SLC34 proteins in the respective membranes. Currently, there are no reported physiological modulators of the transport kinetics per se at the molecular level (e.g., by altering substrate affinity or turnover rate). The regulatory factors include dietary P_i, hormones (parathyroid hormone (PTH), fibroblast growth factor 23 (FGF23) – a bone-derived hormone, together with its co-receptor Klotho, dopamine, and vitamin D₃), and glucocorticoids. Vitamin D₃ increases intestinal P_i absorption via NaPi-IIb, whereas PTH and FGF23 reduce plasma P_i by downregulating NaPi-IIa/c. One well-documented regulatory process at the posttranslational level leads to internalization of NaPi-IIa induced by PTH, FGF23, and dopamine. These hormones bind to their respective membrane receptors and trigger different intracellular signaling cascades that involve, among others, protein kinases, PKA and PKC. The regulatory signal converges on the Na⁺/H⁺ exchanger regulator NHERF1, whose presence stabilizes NaPi-IIa in the membrane by interaction of a PDZ domain with the C-terminal TRL motif of the transporter (Fig. 2). Phosphorylation of NHERF1 leads to removal of NaPi-IIa from the apical membrane (e.g., Forster et al. 2006). Klotho, in addition to its association in the membrane-bound form with FGF23 (Hu et al. 2013), in its soluble form has also been shown to act as a direct extracellular enzyme by deglycosylating NaPi-IIa, thereby leading to proteolytic degradation (Hu et al. 2010). For NaPi-IIb and NaPi-IIc the intracellular signaling pathways and molecular motifs involved in regulating the membrane abundance remain to be identified.

Dysfunction in P_i homeostasis results in pathological conditions that are linked to hyper- or hypophosphatemia (reviewed in Bergwitz and Juppner 2010; Lederer and Miyamoto 2012). Knockout studies have underscored the physiological role of SLC34 transporters and provide essential models for disentangling the complexities of whole-body P_i homeostasis. NaPi-IIa^−/− animals show severe P_i wasting and slow growth with an early bone phenotype (retarded secondary ossification) that reverses in adult mice. The NaPi-IIb knockout is embryonic lethal, which most likely reflects its expression in other organs. However, conditional deletion of NaPi-IIb in the murine intestine leads to compensatory upregulation of renal P_i reabsorption. During weaning, hypophosphatemia is observed in these mice, which reflects the P_i requirement for skeletal growth and underscores a physiological role for NaPi-IIb during development. In contrast, in older (20 weeks) mice, adequate dietary supply normal P_i levels can be maintained with presumably paracellular intestinal absorption and reduced renal P_i excretion. Disruption of the Slc34a3 gene (NaPi-IIc) in mice, has no apparent consequences, also in young animals. NaPi-IIc^−/− mice show hypercalcemia, hypercalciuria but no hypophosphatemia, even when targeted to the kidney. These findings point to species differences in renal P_i handling between mice and human as mutations in the human SLC34A3 gene are associated with hereditary hypophosphatemic rickets with hypercalciuria (HHRH). Whereas the link between HHRH and mutations in SLC34A3 was established within 4 years after discovery of the gene, it took almost two decades to identify patients with defects in NaPi-IIa. Interestingly, the loss of function mutations manifests in hypercalcemia and kidney stones rather than hypophosphatemia; severe cases present with Fanconi syndrome (Wagner et al. 2014; Dinour et al. 2016; Schlingmann et al. 2016). Mutations in the SLC34A2 gene (NaPi-IIb) are associated with pulmonary and testicular microlithiasis (reviewed in Lederer and Miyamoto 2012). NaPi-IIb has also recently come into focus as tumor marker and potential target for ovarian, lung, and bladder cancers. The monoclonal mouse MX35 antibody was generated against human ovarian ascites cells and later shown to specifically recognize NaPi-IIb (Lin et al. 2015). Moreover, NaPi-IIb itself is implicated in cancer progression as depletion of the transporter reduces cell viability in vitro and tumor growth in vivo (Ye et al. 2017). Finally, the ability of SLC34 proteins to transport Li⁺ in the presence of Na⁺ (Andrini et al. 2012) and the reported action of SLC34 proteins to reabsorb Li⁺ in rat kidney (Uwai et al. 2014) may have consequences for the clinical use of Li⁺ salts to treat psychiatric disorders.

Members of the SLC34 family of Na⁺-coupled P_i transporters have been the subject of intense research at the molecular, cellular, and whole organism level for over two decades since they were originally cloned, yet many questions remain unanswered. Of particular importance is the lack of a complete 3-D structure. This has hampered the development of SLC34-specific inhibitors that could find application in the treatment of chronic kidney disease (CKD), a consequence of which is hyperphosphatemia due to kidney dysfunction and increased calcium-phosphate deposits in blood vessels. In this respect, for such pharmacological intervention to be effective it will also be necessary to determine the contribution of NaPi-IIb to dietary P_i intake in humans. The development of isoform-specific drugs could also serve not only to inhibit transport but also stabilize membrane expression in the case of P_i-wasting diseases, where downregulation could be pharmacologically controlled. Identifying the elements in the regulatory cascades and their molecular interactions with SLC34 proteins may lead to the identification of useful therapeutic targets in addition to the transporters themselves. In this context, with mounting evidence of the side effects of a high P_i diet in Western society, it is to be expected that SLC34 transporters will therefore remain an important research focus in the future. Moreover, although the role of NaPi-IIa and NaPi-IIc as the main players in mammalian renal P_i reabsorption is well established, the role of NaPi-IIb in other organs is still unclear. Finally, it should also be noted that whereas much of our current understanding of the role of SLC34 proteins in P_i homeostasis has been derived from animal models, the translation of these data to humans should be made with caution given the isoform-specific differences observed among species.