Na⁺/H⁺ Exchanger Regulatory Factor 3 Is Critical for Multidrug Resistance Protein 4–Mediated Drug Efflux in the Kidney

Abstract

Na⁺/H⁺ exchanger regulatory factor 3 (NHERF3) is a PSD-95/discs large/ZO-1 (PDZ)-based adaptor protein that regulates several membrane-transporting proteins in epithelia. However, the in vivo physiologic role of NHERF3 in transepithelial transport remains poorly understood. Multidrug resistance protein 4 (MRP4) is an ATP binding cassette transporter that mediates the efflux of organic molecules, such as nucleoside analogs, in the gastrointestinal and renal epithelia. Here, we report that Nherf3 knockout (Nherf3^−/−) mice exhibit profound reductions in Mrp4 expression and Mrp4-mediated drug transport in the kidney. A search for the binding partners of the COOH-terminal PDZ binding motif of MRP4 among several epithelial PDZ proteins indicated that MRP4 associated most strongly with NHERF3. When expressed in HEK293 cells, NHERF3 increased membrane expression of MRP4 by reducing internalization of cell surface MRP4 and consequently, augmented MRP4-mediated efflux of adefovir, a nucleoside-based antiviral agent and well known substrate of MRP4. Examination of wild-type and Nherf3^−/− mice revealed that Nherf3 is most abundantly expressed in the kidney and has a prominent role in modulating Mrp4 levels. Deletion of Nherf3 in mice caused a profound reduction in Mrp4 expression at the apical membrane of renal proximal tubules and evoked a significant increase in the plasma and kidney concentrations of adefovir, with a corresponding decrease in the systemic clearance of this drug. These results suggest that NHERF3 is a key regulator of organic transport in the kidney, particularly MRP4-mediated clearance of drug molecules.

Assembly of protein complexes by adaptor proteins with PSD-95/discs large/ZO-1 (PDZ) domains plays an important role in the regulation of many membrane proteins, especially in epithelial and neuronal tissues.^1,2 For example, PDZ-based adapter proteins in epithelia, such as Shank2, synaptic scaffolding molecule (S-SCAM), and Na⁺/H⁺ exchanger regulatory factors (NHERFs), are known to be involved in the regulation of cell surface expression and the activity of many membrane transporters and receptors in the respiratory, digestive, urinary, and reproductive organs.^3–6 The four members of the NHERF proteins (NHERF1 to 4) contain either two or four PDZ domains and were the first family of PDZ-containing proteins shown to be involved in epithelial transport. NHERF3, also known as PDZK1 or CAP70, has four PDZ domains and is normally localized to the apical microdomains of gastrointestinal and kidney epithelia. Each PDZ domain of NHERF3 has been proposed to bind independently to the PDZ binding motif of various membrane proteins, such as the cystic fibrosis transmembrane conductance regulator (CFTR), the Na⁺/H⁺ exchanger 3, the urate transporter, and the organic anion transporter 4.^7–10 However, Nherf3 knockout (Nherf3^−/−) mice do not have a discernible phenotype except mild hypercholesterolemia,¹¹ and, consequently, the in vivo role of NHERF3 in transepithelial transport remains poorly understood.

Transporters belonging to either the ATP binding cassette (ABC) or the solute-linked carrier superfamilies of membrane proteins play diverse roles in the pharmacokinetic and pharmacodynamic pathways of drugs and their metabolites.¹² Multidrug resistance protein 4 (MRP4/ABCC4) is a member of the C subfamily of ABC transporters and mediates the efflux of a group of organic molecules using energy generated from the binding and hydrolysis of ATP.¹³ Cumulative evidence suggests that MRP4 pumps out purine compounds and plays an important role in the renal elimination of purine-based antiviral and antineoplastic agents.^14,15 However, the regulatory mechanisms for MRP4 expression and function have not been extensively studied.

The COOH terminus of MRP4 contains a class 1 PDZ interaction motif (-S/T-X-Ф, where Ф is a hydrophobic amino acid),¹⁶ indicating that MRP4 may interact with PDZ-based adaptors in epithelia. In preliminary studies using a yeast two-hybrid assay, MRP4 was found to strongly interact with NHERF3 among several PDZ adaptor proteins expressed in epithelial tissues. The aim of the present study was to identify the physiologic and pharmacologic roles of the interaction of NHERF3 with MRP4 using a variety of in vitro and in vivo experimental approaches. The results obtained provide strong evidence that NHERF3 is a key regulator of organic transport in the kidney, particularly the MRP4-mediated clearance of purine-based drug molecules.

Results

NHERF3 Interacts with MRP4

Interactions between the COOH termini of ABC transporters that have a PDZ binding motif and the PDZ domains of adaptor proteins expressed in epithelia were initially identified by a yeast two-hybrid screening assay. The results indicated that the PDZ domains of S-SCAM, NHERF1, and NHERF3 might interact with MRP4. NHERF3 showed the strongest interaction when examined by reporter gene activities (Supplemental Table 1). The interaction between MRP4 and NHERF3 found in the yeast two-hybrid system was confirmed in mammalian human embryonic kidney (HEK) 293 cells by a coimmunoprecipitation assay (Figure 1A). Experiments with an expression vector, which lacks the last four amino acids of the COOH terminus of MRP4 comprising the PDZ binding motif (MRP4-ΔETAL), indicated that a PDZ-based interaction mediates the association between MRP4 and NHERF3 in HEK293 cells.

Figure 1.

MRP4 directly binds to NHERF3 through a PDZ-based interaction. (A) Immunoprecipitation assays using transfected HEK293 cells confirmed the interaction between MRP4 and NHERF3 in mammalian cells. The interaction was eliminated by the deletion of the PDZ binding motif at the COOH terminus of MRP4 (MRP4-ΔETAL). (B–E) Pull-down assay using recombinant proteins. (B) Schematic diagram of the four NHERF3 PDZ domain constructs linked to a 6X His tag. (C) Polypeptides corresponding to each of the four PDZ domains in NHERF3 were purified using a nickel-nitrilotriacetic acid protein purification system. (D) Peptides containing the GST-conjugated COOH-terminal 40 amino acids of MRP4 (arrow) and a GST control protein were purified using glutathione-Sepharose beads. (E) GST pull-down showing the direct association between the COOH terminus of MRP4 and the first PDZ domain (PDZ1) of NHERF3.

To determine whether MRP4 directly associates with NHERF3 and, if so, which PDZ domain of NHERF3 is responsible for the interaction, pull-down assays were performed with recombinant proteins. Thus, four His₆-tagged recombinant proteins encoding individual PDZ domains of NHERF3 (Figure 1, B and C) and a glutathione S-transferase (GST) fusion protein containing the last 40 amino acids of the COOH terminus of MRP4 (GST-MRP4-C40) (Figure 1D) were expressed in Escherichia coli and purified. Pull-down assays with the recombinant proteins and glutathione agarose beads indicated that the COOH terminus of MRP4 directly associates with PDZ1 of NHERF3 (Figure 1E).

NHERF3 Increases the Expression and Function of MRP4 in HEK293 Cells

Next, we examined the effects of NHERF3 on the expression levels and function of MRP4 in HEK293 cells. First, cell surface and total MRP4 levels were measured in the presence and absence of transfected NHERF3. As shown in Figure 2A, NHERF3 greatly augmented the total and cell surface expression of MRP4. Quantitative analyses of multiple experiments are summarized in Figure 2, B–D, and show that coexpression of NHERF3 increased the amount of MRP4 in cell lysates and the cell surface by 4.0- and 5.9-fold, respectively (Figure 2, B and C). As a consequence, cell surface expression of MRP4 relative to the total amount in cell lysates was also increased by 38% (Figure 2D).

Figure 2.

NHERF3 upregulates MRP4 expression and function in HEK293 cells. (A–D) The effects of NHERF3 coexpression on the total and cell surface expression of MRP4 were examined in HEK293 cells. (A) A representative immunoblot of cell surface and lysate proteins in the presence and absence of NHERF3. (B) Densitometric analyses of total MRP4 protein levels in the lysate immunoblots (n=8). MRP4 levels were corrected relative to aldolase levels. (C) Densitometric analyses of biotinylated surface MRP4 protein levels (n=7). The amount of surface MRP4 was corrected for the amount of aldolase in the corresponding lysate blots. (D) Relative abundance of MRP4 at the cell surface (corrected for the total MRP4 in the lysate blots; n=7). (E and F) Functional activity of MRP4 was measured by loading HEK293 cells with [³H]adefovir and measuring remaining tritium in the cells after 2 hours. (E) Chemical structure of [³H]adefovir. (F) Efflux of [³H]adefovir, which represents MRP4 function, was augmented in the cells coexpressing NHERF3. *P<0.05; **P<0.01.

Second, the effect of NHERF3 coexpression on the functional activity of MRP4 in HEK293 cells was measured. The cells were loaded with the dipivoxil ester form of [³H]adefovir (Figure 2E), an antiviral agent and well known substrate of MRP4,¹⁷ and the efflux activity of MRP4 was indirectly estimated by measuring the radioactivity remaining in the cells after 2 hours at 37°C. As shown in Figure 2F, expression of MRP4 strongly decreased the amount of radioactivity (and hence, adefovir) that remained in the cells. Notably, coexpression of NHERF3 further decreased the [³H]adefovir radioactivity from 4.08±0.79×10⁵ to 2.03±0.52×10⁵ counts per minute/mg protein (P<0.05). These results indicate that NHERF3 can increase the functional activity of MRP4 in HEK293 cells, which correlates well with the results obtained from the cell surface labeling experiments.

Molecular Mechanism of MRP4 Upregulation by NHERF3

Enhanced levels of MRP4 could be explained by transcriptional upregulation or increased stability of MRP4 mRNA. However, results from quantitative RT-PCR revealed that the amount of MRP4 mRNA in HEK293 cells was not affected by NHERF3 coexpression (Supplemental Figure 1). Several PDZ-based adaptors have shown that they increase the cell surface stability of associated membrane proteins.^6,18 Therefore, the effects of NHERF3 on the cell surface stability of MRP4 were examined. HEK293 cells transfected with Myc-tagged MRP4 were treated with cycloheximide (0.1 mg/ml) for up to 36 hours to inhibit new protein synthesis; proteins on the cell surface were biotinylated, and cell surface MRP4 was detected with an anti-Myc antibody. As shown in Figure 3, A and B, only 8.6±2.1% of MRP4 remained at the cell surface 36 hours after the cycloheximide treatment. Interestingly, coexpression of NHERF3 increased this value to 35.7±5.3%. Coexpression of NHERF3 also enhanced the stability of MRP4 in total cell lysate, as well as the surface protein. Thus, the amount of total MRP4 remaining 36 hours after treatment with cycloheximide was increased from 37.4±4.7% to 67.9±6.1% by coexpression of NHERF3 (Figure 3, C and D), indicating that degradation of MRP4 was attenuated by NHERF3 coexpression.

Figure 3.

NHERF3 increases the stability of MRP4 protein in HEK293 cells. (A and B) The stability of MRP4 protein at the cell surface was examined using a biotinylation assay. After treatment with cycloheximide (0.1 mg/ml), protein samples were harvested at the indicated times. Immunoblotting for aldolase confirmed the absence of cytosolic proteins in the biotinylated samples. Results obtained in multiple experiments (n=4) are summarized in B. (C and D) The stability of MRP4 protein in total cell lysates was examined up to 36 hours. Results of multiple experiments (n=4) are summarized in D. Stability of MRP4 both on the cell surface and in the cytosol was increased by coexpression of NHERF3. For more accurate densitometric quantification of MRP4 levels, a 5-fold larger amount of protein sample was loaded in lanes without NHERF3 (aldolase immunoblot). *P<0.05; **P<0.01. CHX Tx, cycloheximide.

To explore mechanisms responsible for the NHERF3-induced increase in MRP4 expression, internalization and recycling of cell surface MRP4 were examined. As shown in Figure 4, NHERF3 significantly reduced the internalization of surface MRP4; however, surface recycling of MRP4 was not affected by NHERF3. To further explore the mechanisms responsible, protein degradation assays were performed with inhibitors of lysosomal (bafilomycin A) and proteasomal (MG132) degradation. As shown in Supplemental Figure 2, treatment with bafilomycin A, but not with MG132, greatly inhibited the degradation of MRP4. The effect of bafilomycin A was significantly reduced in cells where NHERF3 was coexpressed (Supplemental Figure 2). Collectively, the above results indicate that NHERF3 reduces internalization of surface MRP4, an effect associated with a decrease in the lysosome-mediated degradation of MRP4.

Figure 4.

NHERF3 reduces internalization of surface MRP4. (A and B) The internalization of cell surface MRP4 protein was assayed in HEK293 cells with or without NHERF3 coexpression. Cell surface proteins were biotinylated and then internalized for indicated times at 37°C. Proteins remaining at the cell surface were stripped of biotin with the MESNA stripping buffer. Internalized, biotinylated proteins were immunoblotted. (A) A representative MRP4 immunoblot of internalization assays is shown. Lane 1, total cell surface MRP4; lane 2, MESNA stripping control; lanes 3–5, internalized MRP4 that remained in the cells after the indicated times. (B) Shown are the quantitative analyses of internalized MRP4 in cells with control or NHERF3 coexpression (n=4). MRP4 internalization was significantly reduced by NHERF3 coexpression. (C and D) The recycling of cell surface MRP4 protein was assayed in HEK293 cells with or without NHERF3 coexpression. Cell surface proteins were biotinylated and then internalized for 1 hour at 37°C. Proteins remaining at the cell surface were stripped of biotin with the MESNA stripping buffer. Nonrecycled, biotinylated proteins were collected at the indicated times after another round of MESNA treatment. (C) A representative MRP4 immunoblot of recycling assays in HEK293 cells is shown. Lane 1, total cell surface MRP4; lane 2, MESNA stripping control; lane 3, internalized MRP4 after 1 hour at 37°C; lanes 4–6, internalized MRP4 that remained in the cells after 1, 2, and 4 hours, respectively. The difference between lane 3 and lanes 4–6 represents recycled MRP4. (D) Shown are the quantitative analyses of recycled MRP4 in cells with control or NHERF3 coexpression (n=4). MRP4 recycling was not significantly altered by NHERF3 coexpression. *P<0.05; **P<0.01.

NHERF3 Upregulates MRP4 in the Mouse Kidney

We next investigated whether the profound in vitro effect of NHERF3 on MRP4 expression could be reproduced in vivo in a mouse model. NHERF3 has been reported to regulate membrane proteins in the digestive organs and the kidney,^7,8,10,19 and therefore, we first examined expression patterns of NHERF3 using tissues from wild-type (Nherf3^+/+) and Nherf3^−/− mice. Immunoblotting experiments confirmed the absence of Nherf3 protein in the Nherf3^−/− mice (Supplemental Figure 3). The amount of Nherf3 protein varied among the four tissues examined, with the highest levels of protein in the kidney, comparatively lower levels in the ileum and liver, and very low levels in the colon (Supplemental Figure 3). Because Nherf3 was most highly expressed in the kidney, subsequent morphologic and biochemical studies were focused mainly on that organ. Immunofluorescence imaging of wild-type mice revealed that Nherf3 is highly expressed in the apical regions of most tubules of the kidney (Figure 5A). Similarly, Mrp4 was observed in the apical membrane of some tubules, consistent with previous reports that MRP4 is principally expressed in the proximal tubules.²⁰ Thus, Nherf3 and Mrp4 colocalize in the apical regions of kidney tubules (Figure 5A).

Figure 5.

Mrp4 levels are decreased in the kidney tubules of Nherf3^−/− mice. (A) Immunofluorescence image of Mrp4 (green) and Nherf3 (red) in the kidney of Nherf3^+/+ (wild-type) mice. Nherf3 colocalizes with Mrp4 at the apical membrane in the kidney tubules. DAPI, 4′,6-diamidino-2-phenylindole. (B) Immunohistochemical detection of Mrp4 in the kidney tissues of Nherf3^+/+ and Nherf3^−/− mice. Mrp4 stained with 3,3′-diaminobenzidine is localized at the apical membrane of proximal tubules of Nherf3^+/+ mice (arrow). Note that Mrp4 levels are reduced in the Nherf3^−/− mice (arrowheads). (C–F) The effects of Nherf3 deletion on the total and apical surface (BBM) expression of MRP4 in kidney tubules were examined. (C) A representative immunoblot of total (lysate) and apical surface (BBM) MRP4 in the kidneys of Nherf3^+/+ and Nherf3^−/− mice. (D) Densitometric analyses of total MRP4 protein levels in the lysate immunoblots (n=6). (E) Densitometric analyses of MRP4 protein levels in the BBM preparations (n=6). (F) Relative abundance of MRP4 at the BBM (corrected for the total MRP4 in the lysate blot; n=6). **P<0.01.

We next examined the effect of Nherf3 ablation. Both immunohistochemistry and immunoblotting experiments showed that the absence of Nherf3 evoked a profound decrease in Mrp4 expression in the mouse kidney. Immunohistochemical images showed that Mrp4 is densely localized at the apical membrane of proximal tubules in wild-type mice in contrast to kidneys from Nherf3^−/− mice, where the staining intensity of Mrp4 is reduced (Figure 5B). Quantification of the immunoblotting results further indicated that the total and apical surface (brush border membrane [BBM]) Mrp4 protein are greatly reduced in Nherf3^−/− mice (Figure 5, C–F). NHERF3 ablation decreased total and BBM MRP4 by 62.1±7.5% and 94.1±1.8%, respectively, in the Nherf3^−/− kidney. As a consequence, BBM expression of MRP4 relative to the total amount in the kidney tubules was also decreased by 85.4±3.2% (Figure 5F).

Effect of Nherf3 Ablation on Mrp4-Mediated Adefovir Transport in Mice

MRP4 has been reported to transport a broad spectrum of endogenous and exogenous substrates, particularly nucleoside-based molecules.^14,21 We chose adefovir for evaluating the in vivo role of NHERF3-mediated MRP4 regulation for following reasons. (1) Renal excretion is a major route for systemic clearance of adefovir.²² (2) MRP4 plays a significant role in the renal elimination of adefovir.¹⁴ (3) Metabolic enzymes do not play a major role in the systemic clearance of adefovir.²³ (4) As a hydrophilic molecule, adefovir is barely permeable to cell membrane by itself,²¹ and (5) adefovir causes a dose-dependent nephrotoxicity.²⁴

[³H]Adefovir was intravenously injected into wild-type and Nherf3^−/− littermates, and pharmacokinetic parameters were analyzed using blood samples. As shown in Figure 6 and Table 1, plasma concentrations of adefovir and some pharmacokinetic parameters were significantly altered in the Nherf3^−/− mice. For example, the plasma concentration of adefovir in Nherf3^−/− mice was 14.3±3.8 μg/ml compared with 7.8±2.0 μg/ml in wild-type mice at 90 minutes after injection (Figure 6A). As a consequence, the area under the plasma time concentration curve was increased from 64.6±8.4 to 84.7±12.7 h⋅μg/ml (P<0.05) (Figure 6B), and systemic clearance of adefovir was decreased from 20.8±2.8 to 15.9±1.8 ml/h (Figure 6C). In addition, the tissue concentration of adefovir 3 hours after injection was 2.0-fold greater in the kidney of Nherf3^−/− mice (Figure 6D). Taken together, these results indicate that the Mrp4-mediated adefovir efflux in renal tubules is significantly reduced in Nherf3^−/− mice.

Figure 6.

Plasma clearance of adefovir is reduced in the Nherf3^−/− mice. (A) [³H]Adefovir was administered intravenously to Nherf3^+/+ and Nherf3^−/− littermates, and plasma concentrations of adefovir were measured at the indicated times. (B) Shown are the values of area under the plasma time concentration curve (AUC) from five littermate pairs. (C) Shown are the values of plasma clearance of adefovir from five littermate pairs. (D) The concentration of adefovir in kidney tissue was measured 3 hours after administration of the drug using a different set of mice (n=5). *P<0.05; **P<0.01.

Table 1.

Pharmacokinetic parameters of adefovir in Nherf3^+/+ and Nherf3^−/− mice

Parameter	Nherf3+/+ (n=5)	Nherf3−/− (n=5)	P Value
ke (1/h)	1.70±0.24	1.37±0.12	0.13
t1/2 (h)	0.44±0.05	0.52±0.04	0.09
AUCinf (h⋅μg/ml)	64.62±8.37	84.66±12.72	0.03a
Vd (ml)	12.46±0.67	11.58±0.93	0.19
Cl (ml/h)	20.81±2.75	15.95±1.84	0.02a

Data are mean±SEM. k_e, elimination rate constant; t_1/2, apparent terminal half-life; AUC_inf, area under the concentration time curve with the last concentration extrapolated based on the elimination rate constant; Vd, volume of distribution; Cl, clearance.

^aP<0.05.

Discussion

Although several reports suggest that NHERF3 is involved in the regulation of many epithelial transporters, its precise physiologic roles are still poorly understood. In the normal circumstances, the only discernible phenotype of Nherf3^−/− mice is mild hypercholesterolemia that is possibly because of a dysregulation of the hepatic HDL receptor, scavenger receptor class B type 1.^11,25 In the present study, we provide evidence that NHERF3 is crucial for the regulation of MRP4 expression and activity both in cell culture model as well as in vivo in the mouse kidney. Physical interactions between NHERF3 and MRP4 were shown using three different experimental approaches (Figure 1, Supplemental Table 1). Augmented expression of NHERF3 was associated with elevated MRP4 levels in both human HEK293 cells and the mouse kidney (Figures 2 and 5). Finally, the Mrp4-mediated systemic clearance of adefovir was significantly reduced in Nherf3^−/− mice (Figure 6, Table 1).

The most notable finding of this study is that NHERF3 enhanced the cell surface expression of MRP4. This finding was more clearly evident in kidney epithelia that natively express MRP4 on the BBM (Figure 5). Results from the MRP4 internalization and recycling assays indicated that NHERF3 stabilizes cell surface MRP4 primarily by reducing its internalization rather than augmenting its membrane insertion (Figure 4). Thus, the decrease in internalization-associated cellular degradation of MRP4 (Supplemental Figure 2) seems to be responsible for the increase in total MRP4.

Interestingly, unlike NHERF3, several PDZ-containing proteins have been reported to reduce the expression of MRP4 by facilitating internalization of cell surface MRP4. For example, NHERF1/EBP50, a member of the NHERF family that also binds to MRP4 through its COOH-terminal PDZ motif, decreases the cellular and surface expression of MRP4.¹⁶ In addition, sorting nexin 27, a member of the sorting nexin family of proteins known to be involved in the cellular trafficking and sorting of cargo proteins, also binds to cell surface MRP4 and increases its internalization.²⁶ Therefore, it is possible that NHERF3 expression may inhibit the associations between these internalization-promoting PDZ proteins and MRP4. Especially in cells that express both NHERF1 and NHERF3, such as kidney tubules, ablation of NHERF3 may play a particularly important role in disrupting the balance between the MRP4-NHERF1 and MRP4-NHERF3 associations, increasing the internalization and degradation of MRP4.

As mentioned earlier, NHERF3 has four PDZ domains, but no other distinct domain structures. The specific roles of each of the four PDZ domains in cell physiology are not currently well understood. The two sodium-dependent phosphate (Na/P_i) transporters (NaPi-2a and NaPi-2c) are known to bind to NHERF3 in kidney tubules.¹⁹ However, the patterns of NaPi-2a and NaPi-2c regulation by NHERF3 differ. Unlike Mrp4, ablation of Nherf3 did not affect NaPi-2a and NaPi-2c levels in mice on normal diets.¹⁹ One of the unique characteristics of the NaPi protein is that its levels are regulated by the amount of P_i in the diet, where diets low in P_i result in an upregulation of NaPi. When Nherf3^−/− mice were fed a low P_i-containing diet, however, the upregulation of NaPi was attenuated for NaPi-2c but not NaPi-2a.¹⁹ Interestingly, NaPi-2c interacts with the PDZ2 domain of NHERF3, whereas NaPi-2a interacts with PDZ3. In this study, we have shown that MRP4 predominantly binds to the PDZ1 domain of NHERF3. As suggested in the case of NaPi,¹⁹ the association of different transporters with different NHERF3 PDZ domains may determine the differences in the regulation of transporters by this adaptor protein.

As a nucleotide transporter, MRP4 has been shown to function as an efflux pump for cAMP and at least in some circumstances, be involved in the termination of cellular cAMP signals.^27–29 For example, the cAMP-induced activation of CFTR is augmented by the MRP4 inhibitor MK571 in HT29-CL19A colonic epithelial cells and the intestines of Mrp4^−/− mice.²⁹ Based on results from in vitro studies, Li et al.²⁹ suggested that NHERF3 may mediate the association between CFTR and MRP4 in the gut. As shown in Supplemental Figure 3, however, levels of NHERF3 expression in the gut, especially in the colon, are very low compared with levels in the kidney. In addition, ablation of Nherf3 does not significantly affect the cAMP-induced activation of Cftr in the mouse intestine and colon.¹⁰ These observations suggest that PDZ adaptor proteins other than NHERF3 may mediate the CFTR-MRP4 association in the gut and that the MRP4-NHERF3 interaction has the highest physiologic relevance in the kidney tubules.

In the present study, ablation of Nherf3 caused a >80% increase in the plasma concentration of adefovir and a 24% reduction in its systemic clearance from mice. These results are highly consistent with results obtained from Mrp4^−/− mice, in which the systemic clearance of adefovir was decreased by 24%.¹⁴ In Mrp4^−/− mice, the reduction in the systemic clearance of adefovir was mainly caused by a 62% decrease in the kidney tubular clearance.¹⁴ Although MRP4-mediated renal clearance is the major route for systemic elimination of adefovir, it should be remembered that there are other routes for adefovir clearance, such as glomerular filtration and non-MRP4–mediated tubular secretions. Therefore, a partial decrease in the systemic clearance of adefovir may not be clinically meaningful in itself. However, the reduced plasma and renal clearances of adefovir in Mrp4^−/− mice were associated with accumulation of adefovir and its phosphorylated toxic metabolites in the kidney,¹⁴ an observation that was reproduced in the current study of Nherf3^−/− mice (Figure 6D). Nephrotoxicity is a major adverse reaction of adefovir, which is often observed with the use of this drug.³⁰ Thus, it is reasonable to conclude that NHERF3-mediated upregulation of MRP4 could play a critical role in preventing toxicities associated with adefovir and possibly, other related MRP4 substrates. Together, the observations that NHERF3 is most highly expressed in the kidney and that Nherf3^−/− mice exhibit a mild phenotype are consistent with the conclusion that a major physiologic/pharmacologic role of NHERF3 is the regulation of organic transporters in the kidney epithelia, especially those transporters involved in the disposition and/or elimination of xenobiotic substrates, such as MRP4.

Concise Methods

Plasmids, Cell Culture, and Animals

HEK293 cells were maintained in DMEM-HG (Invitrogen, Carlsbad, CA) supplemented with 10% FBS. Plasmids were transiently transfected into HEK293 cells using Lipofectamine Plus Reagent (Invitrogen). The pcDNA3.1/Hygro-MRP4 plasmid¹⁶ was subcloned into pCMV-myc-N using the EcoRI/KpnI sites to generate pCMV-N-myc-wt-MRP4. The MRP4-ΔETAL plasmid was generated by a PCR-based site-directed mutagenesis. GST-MRP4-C40 containing the DNA sequence for the COOH-terminal 40 amino acids (amino acids 1286–1325) of MRP4 was generated by inserting PCR fragments amplified from pCMV-N-myc-wt-MRP4 into the bacterial expression vector pGEX-4T-1 (GE Healthcare) using the BamHI/EcoRI sites. Plasmids encoding for the four PDZ domains of NHERF3 fused with His₆ were generated by inserting PCR fragments amplified from each PDZ domain (amino acids 1–105, 119–230, 231–348, and 365–474) into the bacterial expression vector pET-28c(+) (Novagen, Darmstadt, Germany) using the BamHI/KpnI sites. The open reading frame of human NHERF3 was PCR-amplified from pOTB7-hNHERF3 (Thermo Scientific, Waltham, MA) and subcloned into pcDNA3.1(+)-N-hemagglutinin (HA) using the BamHI/XbaI sites to generate pcDNA3.1(+)-N-HA-hNHERF3. Nherf3^−/− mice were handled and genotyped as previously described.^10,11 Animal use and welfare adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and the protocol was reviewed and approved by the Committee on Animal Research at Yonsei Medical Center, Seoul, Korea (no. 2010–0217).

Immunoblotting, Immunoprecipitation, and Surface Biotinylation

For immunoblotting, cell and tissue lysates were suspended in an SDS buffer and separated by SDS-PAGE. The separated proteins were transferred and immobilized onto a nitrocellulose membrane and blotted with anti-MRP4 (Alexis Biochemical, San Diego, CA), anti-aldolase (Santa Cruz Biotechnology, Santa Cruz, CA), and anti-HA (Santa Cruz Biotechnology) monoclonal antibodies. Protein bands were detected by enhanced chemiluminescence (Amersham Biosciences).

For immunoprecipitation, cell lysates were mixed with the anti-HA polyclonal antibody (Santa Cruz Biotechnology) and incubated overnight at 4°C in a lysis buffer containing 50 mM Tris⋅HCl (pH 7.4), 150 mM NaCl, 1% (vol/vol) Nonidet P-40, 0.25% (vol/vol) sodium deoxycholate, 1 mM EDTA, and a complete protease inhibitor mixture (Roche Applied Science). Immune complexes were collected by binding to mixed protein G/A beads and washed three times with lysis buffer before electrophoresis.

Surface biotinylation was performed as previously described.³¹ Briefly, transfected cells were washed with ice-cold PBS containing 0.1 mM CaCl₂ and 1 mM MgCl₂, and the plasma membrane proteins were biotinylated by sulfo-NHS-SS-biotin (Pierce) in PBS for 30 minutes at 4°C. After biotinylation, the cells were washed extensively with PBS to remove excess biotin. The cells were lysed, and an avidin-containing solution (Streptavidin Agarose Resins; Pierce) was added to the supernatant; the mixture was incubated at 4°C overnight. Avidin-bound complexes were washed three times, eluted in SDS sample buffer, resolved by SDS-PAGE, electrotransferred, and immunoblotted with appropriate antibodies.

To determine the stability of MRP4 on the cell surface and in total cell lysates, cells were treated with cycloheximide (0.1 mg/ml; Sigma-Aldrich) to prevent additional protein synthesis 72 hours after transfection with pCMV-N-myc-wt-MRP4. Surface proteins were biotinylated, and cells were harvested at 0, 12, 24, and 36 hours after the cycloheximide treatment. MRP4 on the cell surface and in lysates was detected by immunoblotting with anti-Myc antibody (Santa Cruz Biotechnology). The density of signals on the immunoblots was analyzed using an imaging software package (Multi Gauge Version 3.0; FujiFilm, Valhalla, NY).

Surface Protein Internalization and Recycling Assay

The internalization assay of cell surface MRP4 was performed as described previously with a slight modification.³² Briefly, transfected cells were biotinylated using EZ-Link sulfo-NHS-SS-biotin (Pierce) for 30 minutes in the dark at 4°C. The cells were then warmed to 37°C for the indicated times to induce internalization, and the remaining disulphide bonds on sulfo-NHS-SS-biotinylated proteins were stripped of biotin with four 15-minute washes in the sodium 2-mercaptoethanesulfonate (MESNA) stripping buffer of 50 mM MESNA, 150 mM NaCl, 1 mM EDTA, 0.2% BSA, and 20 mM Tris (pH 8.6). Subsequently, the cells were lysed with lysis buffer. The lysates were centrifuged for 10 minutes (13,000×g), and the pellet was discarded. Avidin solution (20 μl, Streptavidin beads; Pierce) was added to the supernatant (200 μg protein per 200 μl lysis buffer), and the mixture was incubated for 3 hours with gentle agitation. Avidin-bound complexes were pelleted by centrifugation, washed three times with lysis buffer, and immunoblotted.

Cell surface recycling of internalized MRP4 was assayed in HEK293 cells as described previously.³³ Because it is difficult to positively visualize the recycled MRP4, the amount of recycled MRP4 was estimated by subtracting nonrecycled MRP4 from total internalized MRP4. Briefly, the plasma membrane proteins were biotinylated and then internalized for 1 hour at 37°C. Proteins remaining at the cell surface were stripped of biotin with the MESNA buffer as described earlier. To detect the recycling of internalized, biotinylated MRP4, cells were incubated at 37°C for indicated times and then quickly cooled to 4°C. Biotinylated MRP4 that is recycled to the plasma membrane was again stripped of biotin by the MESNA buffer washes. The remaining nonrecycled biotinylated MRP4 up to 4 hours was measured, and the recycled MRP4 was calculated as described in Figure 4. Because the cellular degradation rate of MRP4 was very slow (Figure 3D), intracellular degradation of biotinylated MRP4 was not considered.

Immunohistochemistry

For immunofluorescence detection, kidney tissues from Nherf3^+/+ (wild type) and Nherf3^−/− mice were first embedded in tissue-freezing medium (TissueTec OCT; Sakura), frozen in liquid nitrogen, and then cut into 4-µm sections. The sections were fixed and permeabilized by incubation in ethanol-acetone solution (1:1; vol/vol) for 10 minutes at −20°C. Nonspecific binding sites were blocked by incubation for 1 hour at room temperature with 0.1 ml PBS containing 5% goat serum, 1% BSA, and 0.1% gelatin (blocking medium). After blocking, the sections were stained by incubating them with anti-MRP4 (Alexis Biochemical) and anti-NHERF3 (Novous Biologic) antibodies followed by fluorophore-tagged secondary antibodies. DNA was stained with 4′,6-diamidino-2-phenylindole. Images were obtained with a Zeiss LSM 710 confocal microscope.

For detection of Mrp4 by 3,3′-diaminobenzidine staining, kidney tissue was fixed in formalin, and paraffin blocks were prepared. Sections were incubated with the anti-MRP4 antibody (Alexis Biochemical) and then followed by anti-rat IgG (H+L) conjugated with horseradish peroxidase (Thermo Scientific); 3,3′-diaminobenzidine signals were developed with a peroxidase substrate kit (Vector Laboratories, Burlingame, CA) according to the manufacturer’s instructions.

Pull-Down Assays

All recombinant fusion proteins were produced in BL-21 (DE3) E. coli strain. The synthesis of GST and His₆ fusion proteins was induced by isopropyl β-D-1-thiogalactopyranoside (1 mM) at 37°C. Recombinant proteins were purified with glutathione-Sepharose beads (Amersham Bioscience) or a nickel-nitrilotriacetic acid protein purification system (Qiagen) according to the manufacturer’s instructions. Eluted His₆ fusion proteins were mixed with the GST fusion recombinant proteins bound to glutathione-Sepharose. After overnight incubation at 4°C, the bead-bound complexes were washed, eluted in SDS sample buffer, and immunoblotted.

Adefovir Cell Accumulation Assay

HEK293 cells were seeded in triplicate in 12-well plates and transfected with pCMV-N-myc-wt-MRP4. Cells were exposed to the dipivoxil ester form of [³H]adefovir ([³H]bis(POM)PMEA; specific activity 5.5 Ci/mmol; Moravek Biochemicals, Brea, CA) mixed with a 100-fold greater amount of nonradiolabeled adforvir dipivoxil (total concentration of 1 μM; Sigma-Aldrich) for 2 hours at 37°C in HBSS, and the transport reaction was terminated by washing cells with ice-cold PBS and solubilizing in 1 M NaOH. Radioactivity was measured in a liquid scintillation counter (Beckman Coulter, Inc., Fullerton, CA) after the addition of scintillation cocktail (PerkinElmer Life and Analytical Sciences). Data were expressed relative to the amount of cell protein (counts per minute per milligram protein) in each sample, which was determined by the Bradford colorimetric method.

Mouse In Vivo Pharmacokinetic Study

Sibling pairs of female mice (weight ∼ 30 g) were anesthetized with ketamine. The right jugular vein and left carotid artery were surgically exposed and cannulated with polyethylene tubing (PE-10; Clay Adams) to facilitate [³H]adefovir injections and blood sampling, respectively. The cannula connected to the artery was exteriorized by a three-way stopcock, and blood samples (40 µl) were collected at 0, 30, 60, and 90 minutes after administration of the drug. [³H]adefovir (specific activity 10 Ci/mmol; Moravek Biochemicals) mixed with a 30,000-fold larger amount of nonradiolabeled adefovir (Santa Cruz Biotechnology) was injected at a total dose of 50 mg/kg using a Hamilton syringe. A heparinized 0.9% NaCl solution was flushed into the cannula right after each blood sampling. Blood samples were centrifuged immediately, and 20 µl plasma was collected. The radioactivity of the plasma samples at each time point was measured by liquid scintillation counting (Beckman Coulter) after the addition of scintillation cocktail (PerkinElmer Life and Analytical Sciences). The concentration of adefovir in kidney tissue was measured as described previously.¹⁴ Briefly, 3 hours after drug injection, the mice were euthanized, and the whole kidney was excised immediately. The kidney was homogenized in lysis buffer using Precellys 24 (Bertin Techonologies), and the radioactivity of homogenate was measured by liquid scintillation counting after addition of the scintillation cocktail.

Statistical Analyses

The results of multiple experiments are presented as the mean±SEM. Statistical analyses of in vitro data were performed with t test or ANOVA followed by Tukey’s multiple comparison test as appropriate. Pharmacokinetic parameters were determined using WinNonlin (Pharsight), and testing for statistical significance was performed using the paired t test. P values<0.05 were considered statistically significant.

Disclosures

None.