Abstract
This study was performed to identify genetic polymorphisms in multidrug and toxin extrusion 2-K (MATE2-K, SLC47A2), a proton/organic cation antiporter that plays a role in the transport of organic cations across the apical membrane in kidney epithelial cells into the urine, and to demonstrate their effects on MATE2-K functions in vitro. Four of the thirty single nucleotide polymorphisms (SNPs) we identified in three ethnic groups (Caucasian, African American, and Japanese) were novel [308C>G (P103R), c.487-8C>T, 818A>G (Y273C), and c.1018+14T>C]. The transport activities of the prototypical substrates, tetraethylammonium and metformin, for four nonsynonymous SNPs (P103R, P162L, G211V, and Y273C) were significantly different from those of the wild-type. In particular, transport activity was higher in P103R than in the wild-type, which is the first time elevated transport activity was demonstrated due to these coding SNPs. Kinetic analysis revealed that P103R had a higher Vmax value, whereas Y273C had a lower value than that in the wild-type. Cell surface protein expression levels were higher for P103R than for the wild-type, whereas Y273C expression was decreased. Immunofluorescence analysis revealed that the P103R protein was localized to the plasma membrane, whereas Y273C showed cytoplasmic localization. Therefore, the difference in transport activities between P103R and Y273C variants was suggested to be responsible for the different protein expression levels observed at the plasma membrane. Four nonsynonymous SNPs in this study showed relatively low allelic frequencies (0.5 to 2.1%), but these were associated with markedly reduced or increased MATE2-K function.
Introduction
The kidney is an essential organ that eliminates many endogenous compounds and drugs from the body. The proximal tubules play a key role in the renal elimination of drugs. Drug transporter proteins have been shown to contribute to drug absorption, distribution, and excretion. Organic cation drugs are secreted from the blood into urine by two distinct types of organic cation transporters. The first step is driven by a transmembrane electrical potential difference in the basolateral membranes, whereas the second step is driven by a transmembrane H+ gradient in the brush-border membranes (Tsuji, 2002; Mizuno and Sugiyama, 2002). The first step is mediated by organic cation transporter 2, which is highly expressed on the basolateral membrane, and the second step is mediated by multidrug and toxin extrusion 1 (MATE1, gene name SLC47A1) and MATE2-K (SLC47A2) transporters. MATE1 and MATE2-K transporters were identified as molecules that excrete organic cations in the human kidney and are characterized by the exchange of proton and organic cations via the apical membranes (Otsuka et al., 2005; Masuda et al., 2006).
The clinical significance of genetic variations has been demonstrated not only for drug-metabolizing enzymes but also for drug transporters. For example, genetic polymorphisms in the MATE1 transporter were recently identified and characterized. Several groups identified genetic variants in the promoter and exons of the SLC47A1 gene (Kajiwara et al., 2007, 2009; Chen et al., 2009; Ha Choi et al., 2009; Meyer zu Schwabedissen et al., 2010). Two single nucleotide polymorphisms (SNPs) in the 5′-untranslated promoter region were recognized as functional SNPs that regulate the transcriptional activity of SLC47A1, and some nonsynonymous SNPs in the coding region have been functionally characterized (Chen et al., 2009; Kajiwara et al., 2009; Meyer zu Schwabedissen et al., 2010). Furthermore, the genetic variant in the promoter significantly reduced the renal clearance of metformin-selected substrate of MATE1 (Christensen et al., 2013). These functional SNPs have shown either the complete loss of or decrease in the transport activity of MATE1.
Human MATE2 has also been cloned as a homolog of human MATE1 (Otsuka et al., 2005). The MATE2 family includes three isoforms, MATE2-K, MATE2, and MATE2-B, with MATE2-K being predominantly expressed in the human kidney and the active form of the SLC47A2 gene (Masuda et al., 2006). The physiologic roles of MATE2 and MATE2-B remain unclear. Two nonsynonymous SNPs, 192G>T (K64N) and 632_633GC>TT (G211V) in the SLC47A2 gene, showed decreased transport activity in the Japanese population (Kajiwara et al., 2009). Toyama et al. (2010) reported that G211V have little influence on the oral clearance of metformin in Japanese diabetic patients. Another study reported that the transport activities of two nonsynonymous variants, 485C>T (P162L) and 1177G>A (G393R), were significantly decreased, whereas the transcriptional activities of the promoter haplotypes containing g.−130G>A (rs12943590) were significantly increased (Choi et al., 2011). The carrier of g.−130G>A showed significantly increased renal and secretory clearance in healthy subjects (n = 57) (Stocker et al., 2013), but significant differences were not observed between the wild-type and the variant type in g.−130G>A in the retrospective data analysis (n = 96) (Yoon et al., 2013). Pharmacokinetic analysis of metformin based on the promoter haplotype demonstrated that the haplotype containing g.−130G>A significantly increased the renal and secretion clearance (Chung et al., 2013). In the present study, we identified two novel coding SNPs by screening all exons of the SLC47A2 gene from three ethnic groups (Japanese, Caucasian, and African American) and tested the functional analysis of coding SNPs by in vitro transient expression experiments.
Materials and Methods
[14C]Tetraethylammonium (TEA) (129.5 MBq/mmol) was purchased from PerkinElmer Life and Analytical Sciences (Boston, MA). [14C]Metformin (3.7 GBq/mmol) was purchased from American Radiolabeled Chemicals (St. Louis, MO). The mammalian expression vector pcDNA3.1(+) was obtained from Invitrogen (Carlsbad, CA). Human embryonic kidney (HEK) 293 cells were purchased from Dainippon Pharma (Osaka, Japan). Blood samples were obtained from unrelated Japanese, Caucasian, and African American subjects (96 of each; Tennessee Blood Services, Memphis, TN).
Screening of SLC47A2 Variants in the Three Ethnic Populations.
Genomic DNA was isolated from blood samples. SLC47A2 variants were screened as described previously (Sasaki et al., 2011). Briefly, genetic variations were examined in SLC47A2 sequences, including all 17 exons, their surrounding introns, and approximately 2,000 bp of the 5′-flanking region. PCR primers (Supplemental Table S1-a and -b) were designed on the basis of the reference sequences of the SLC47A2 gene (Genbank NC_000017.10 and NM_001099646.1). PCR was run for 30 cycles of 95°C for 40 seconds, 50–68°C for 45 seconds, and 72°C for 1 minute using AmpliTaq Gold DNA polymerase (Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions. The products were analyzed with the single-strand conformation polymorphism method for the screening of genetic variants followed by sequencing. Sequencing was performed directly on an ABI 3130xl DNA analyzer (Applied Biosystems) using a Big-Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems).
Construction of an Expression Plasmid Coding the FLAG-MATE2-K Wild-Type and Variants.
The MATE2-K coding plasmid was obtained from the Mammalian Gene Collection (MGC) (Strausberg et al., 2002). The MATE2-K coding region on pCMV-SPORT6/MATE2 (MGC clone ID: 5759988) was subcloned to the pcDNA3.1(+) expression plasmid. FLAG tag DNA was then inserted into the expression plasmid. Briefly, FLAG tag DNA was generated by annealing two oligonucleotides, containing the FLAG (amino acid: DYKDDDDK) sequence and restriction endonuclease site (N-ter: ClaI, C-ter: NotI) in the sequence (Supplemental Table S1-c). Linear DNA coding MATE2-K was amplified by inverse PCR using restriction endonuclease site conjugated primers and the MATE2-K expression plasmid. The MATE2-K wild-type and its variants were generated using the site-directed mutagenesis method (Supplemental Table S1-d). The sequences of these plasmids were confirmed by direct sequencing.
Expression of the MATE2-K Protein in HEK293 Cells.
HEK293 cells were maintained at 37°C and 5% CO2 in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. HEK293 cells were grown on poly-d-lysine-coated 6-well (TEA) or 24-well (metformin) plates and transfected with FLAG-MATE2-K cDNA plasmids. Cells were transfected with plasmid DNA using Lipofectamine 2000 reagent (Life Technologies, Grand Island, NY), following the manufacturer’s instructions. Transport activity was measured 48 hours after the transfection.
Uptake Assays.
Transfected HEK293 cells were preincubated with 0.5 ml incubation medium (145 mM NaCl, 3 mM KCl, 1 mM CaCl2, 0.5 mM MgCl2, 5 mM d-glucose, 5 mM HEPES, pH 7.4) for 20 minutes at 37°C. To manipulate the intracellular pH, intracellular acidification was performed by adding 30 mM ammonium chloride to the incubation medium. The medium was then removed, and 0.5 ml incubation medium (pH 8.5) containing the radiolabeled substrate was added. The medium was aspirated off at the end of the incubation, and the monolayers were rapidly rinsed two times with 1 ml ice-cold incubation medium. Cells were solubilized in 1 ml of 1 N NaOH, neutralized with 5 M HCl, and the radioactivity in aliquots was then determined using a liquid scintillation counter (LSC-6100, ALOKA, Tokyo, Japan). The uptake count was standardized by the amount of protein in each well. The protein concentrations of the cell lysates were measured by the BCA assay using bovine serum albumin (BSA) as a standard. Nonspecific uptake due to passive diffusion was assessed in parallel experiments using cells transfected with the pcDNA3.1 empty vector.
Determination of Kinetic Parameters.
These parameters were estimated by a nonlinear least-squares method (damping Gauss-Newton method) using S-PLUS for Windows V8.1 (TIBCO Software, Palo Alto, CA), and expressed as the mean ± computer-calculated S.D. The concentration dependence of the uptake of a substrate mediated by MATE2-K was analyzed using the following Michaelis-Menten equation:
where v, S, Km, Vmax, and Pdif represent the uptake velocity of the substrate (nmol/min/mg protein), substrate concentration in the incubation buffer (mM), Michaelis constant (mM), maximum uptake rate (nmol/min/mg protein), and nonsaturable transport clearance (ml/min/mg protein), respectively.
Western Blot Analysis and Quantification of Band Density.
HEK293 cells were harvested and homogenized with CelLytic M cell Lysis Reagent (Sigma-Aldrich, St. Louis, MO) and Protease Inhibitor Cocktail (Sigma-Aldrich) 48 hours after transfection. Lysate samples were separated on 10% SDS-polyacrylamide gels and transferred to PVDF membranes by a Trans-Blot SD Semi-Dry Electrophoretic Transfer Cell (Bio-Rad, Richmond, CA). Immunoblots were blocked in PBS with 0.1% Tween 20 and 5% nonfat dry milk for 1 hour at room temperature and incubated with the primary antibody against FLAG (Monoclonal ANTI-FLAG M2-Peroxidase Clone M2, Sigma-Aldrich) or Na+/K+-ATPase (Na+/K+-ATPase Antibody, CST) for 1 hour at room temperature. Primary antibody dilutions in Can Get Signal Immunoreaction Enhancer Solution 2 (Toyobo, Osaka, Japan) were as follows: anti-FLAG, 1:2,000 and anti- Na+/K+-ATPase, 1:1,500. After extensive washing and blocking for 30 minutes, the membranes were incubated for 1 hour at room temperature with the secondary antibody, anti-rabbit IgG-horseradish peroxidase (GE Healthcare, Waukesha, WI) diluted 1:10,000 in Tween-PBS. After washing the membranes, immunoreactive proteins were visualized by chemiluminescence (Amersham ECL, GE Healthcare, Piscataway, NJ), captured (Image Reader LAS-3000, Fujifilm, Tokyo, Japan), and analyzed (Multi Gauge, Fujifilm).
Immunofluorescence of MATE2-K-Transfected Cells.
HEK293 cells were seeded onto poly-d-lysine-coated cover glasses, and transfection was performed. Cultured cells were fixed with 4% paraformaldehyde in PBS. When staining for surface glycoproteins, 5 μg/ml Alexa Fluor 555-conjugated wheat-germ agglutinin (WGA; Invitrogen) was applied to the cells on ice for 5 minutes, followed by permeabilization with methanol for 5 minutes on ice. After blocking with 1% BSA/PBS, cells were incubated with the fluorescein isothiocyanate (FITC)-labeled monoclonal antibody, anti-FLAG M2-FITC (Sigma-Aldrich) diluted 1:1,100 in 1% BSA/PBS. Immunofluorescence images were visualized with an A1Rsi confocal laser microscope (Nikon, Tokyo, Japan).
In Silico Structural Analysis.
Structure models of MATE2-K (amino acid reference: NP_001093116.1) were generated using MODELER 9.10 (Sali, 1995) based on the X-ray structure of the MATE homolog from Vibrio cholerae, NorM [Protein Data Bank identified 3MKT (He et al., 2010)]. The image was created using PyMol (http://www.pymol.org).
Statistical Analysis.
The significance of differences between the wild-type and mutants was evaluated by a one-way analysis of variance followed by Dunnett’s test.
Results
Identification of Genetic Variants in the SLC47A2 Gene.
Thirty SNPs were detected in the SLC47A2 gene in the three ethnic groups by the PCR-single-strand conformation polymorphism analysis (Table 1 and Supplemental Table S2). Five SNPs were located in the 5′-flanking region and ten SNPs in the coding region. Four SNPs [308C>G (P103R), c.487-8C>T, 818A>G (Y273C), c.1018+14T>C] were novel. Five coding SNPs caused an amino acid change (nonsynonymous, including two novel SNPs), and their allele frequencies ranged from 0.5 to 2.1%.
TEA and Metformin Uptake Activities in MATE2-K-Expression HEK293 Cells.
To explore the impact of SNPs on transport activity, the uptakes of TEA and metformin were measured in transfected HEK293 cells (Fig. 1). Cells were incubated for 2 minutes in uptake medium containing 5 μM TEA or 20 μM metformin. The transport activity of 308C>G (P103R) was higher than that of the wild-type (TEA: 150%, metformin: 225%). In contrast, the transport activities of 485C>T (P162L), 632_633GC>TT (G211V) and 818A>G (Y273C) were lower than that of the wild-type. The uptake rate of TEA by P162L, G211V, and Y273C were 20.7, 33.9, and 35.3% of the wild-type rate. The transport activities of metformin in P162L, G211V, and Y273C were 35.8, 50.4, and 68.0% of the wild-type activity. No significant difference was observed in transport activities between the six other coding SNPs and the wild-type.
Kinetic Analysis of TEA Uptake in MATE2-K-Expressing HEK293 Cells.
We assessed the kinetic profile of MATE2-K P103R-, P162L-, G211V-, or Y273C-expressing cells. To estimate kinetic parameters for the uptake of TEA, concentration-dependent uptake was carried out (0.05, 0.1, 0.2, 0.5, 1.0, and 2.0 mM) (Fig. 2). The nonsaturable component was estimated by the linear regression analysis in the range of concentrations up to 10 mM TEA using mock-transfected cells (r2 > 0.97). The uptake of TEA by the MATE2-K variants was calculated by subtracting the nonsaturable component. P162L and G211V did not show saturable TEA uptake. The apparent Vmax and Km values for the wild-type, P103R, and Y273C are summarized in Table 2. Three separate experiments were performed to obtain the kinetic parameters, and the Vmax values of P103R and Y273C were significantly different from those of wild-type MATE2-K (1.207 ± 0.114 nmol/mg protein/min for the wild-type, 2.154 ± 0.294 nmol/mg protein/min for P103R and 0.387 ± 0.038 nmol/mg protein/min for Y273C, P < 0.05).
Western Blotting.
MATE2-K mRNA levels were determined to identify the cause of the variability in transport activity. No change was found between the MATE2-K wild-type and its four variants (Supplemental Fig. 1S). Western blotting was performed using the crude membrane from HEK293 transfectants, and the protein expression level of each MATE2-K variant was estimated by quantifying MATE2-K band densities (Fig. 3). Although the expression level of each variant differed, the molecular mass was approximately 40 kDa, which is consistent with the findings of a previous study (Tanihara et al., 2007). MATE2-K expression was higher in P103R than in the wild-type (230%); in contrast, the expression level of Y273C was decreased (40%). P162L and G211V showed markedly lower protein expression levels than those of the other MATE2-K-transfected cells.
Cellular Localization of the MATE2-K Wild-Type and Its Variants.
The localization of MATE2-K variants was investigated using immunocytochemical staining (Fig. 4). All cells were stained with fluorescent WGA (red) to visualize the plasma and nuclear membranes. The P103R MATE2-K protein (green) was localized at the plasma membrane, whereas the P162L and G211V proteins were not found at the plasma membrane or cytoplasm (Fig. 4). The Y273C protein was partially localized at the plasma and internal compartment membrane. No FLAG tag-derived staining was observed in mock-transfected HEK293 cells.
Protein Structure Prediction for Four Nonsynonymous SNPs in the Comparative Structure Model.
We predicted the structure of MATE2-K variants using comparable models (Fig. 5). Transmembrane helices were represented as ribbons using the rainbow gradient from the N terminus (blue) to the C terminus (red). A side view of the MATE2-K monomer showed that the P103 residue was placed in the cytoplasmic loop (Fig. 5A). P162, G211, and Y273 residues were predicted to be located in the 4th transmembrane helix, 6th helix, or 7th helix from the N terminus.
Discussion
Systematic screening was performed in the present study to determine genetic variations, and two novel coding SNPs [308C>G (P103R) and 818A>G (Y273C)] were identified. Among the five promoter variants identified, g.−130G>A promoter SNP frequency ranged from 26.0 to 35.4% and was reported to be associated with a significant increase in transcriptional activities, leading to a poor response to metformin (Tsuji, 2002). No significant interethnicity was observed in its frequency, which was similar to the findings of a previous study (Tsuji, 2002).
Ten nonsynonymous SNPs were characterized by in vitro experiments. A transporter assay showed that TEA or metformin uptake was altered in four coding SNPs (P103R, P162L, G211V, and Y273C) (Fig. 1). The transport activity of P103R, in particular, was increased. This is the first study to identify a coding SNP that increased the function of MATE-2K. The transport activity of another novel SNP Y273C was decreased. The defective transport activities of P162L and G211V were consistent with the findings of previous studies (Kajiwara et al., 2009; Choi et al., 2011). Kinetic analysis revealed that the Vmax of P103R was significantly higher than that of the wild-type, which supported the increase in MATE2-K activity shown in Fig. 1. In contrast, the Vmax of Y273C was lower than that of the wild-type, and the kinetic profiles of P162L and G211V were similar to the mock transfectant. These results also support Y273C, P162L, and G211V being associated with decreased transport activities.
We performed Western blotting and immunofluorescence analysis to examine whether the expression of MATE2-K proteins on the cell surface membrane was altered in HEK293 cells expressing variants. The markedly low protein expression levels on P162L- and G211V-generated cell surface membranes were consistent with the lack of transport activities. As shown in Figs. 3 and 4, the protein expression level of P103R was significantly higher than that of the wild-type, and localization was maintained at the plasma membrane. The increased transport activity and Vmax value may be responsible for higher expression of the MATE2-K protein on the cell surface membrane. The Y273C variant protein was localized at the internal compartment membrane, which suggested that reduced localization at the plasma membrane caused the lower Vmax value. Hoechst 33342 staining revealed that the Y273C variant did not localize at nuclei (data not shown); therefore, one possible reason for reduced Y273C variant protein is Golgi localization. The tyrosine-based sorting motif conforms to the consensus motifs YXXØ (Y is tyrosine, X is any amino acid, and Ø is an amino acid with a bulky hydrophobic side chain), and the lack of this motif has been suggested to cause its retention in the Golgi (Aoki et al., 1992; Nilsson et al., 1991). The MATE2-K wild-type contains the amino acid residue (273YEIG276), which has been predicted as the tyrosine-based sorting motif. However, the Y273C variant protein is a defective motif.
Comparison of the amino acid sequences of MATE2-K between organisms showed that Y273 residue is located in a highly conserved region of the amino acid sequence, but P103 is not conserved (Supplemental Fig. 2s). Topology analysis predicted that Y273C could change the secondary structure of MATE2-K (data not shown). Thus, the amino acid substitution at a highly conserved Y273 may play a crucial role in the altered formation of MATE2-K protein. Comparable structure models indicated that P103 exists in the intracellular loop region. An estimation of the ubiquitination site using BDM-PUB (prediction of ubiquitination sites with the Bayesian discriminant method) showed that the wild-type has ten sites; however, P103R has nine sites. This reduction in the number of ubiquitination sites may be responsible for the decrease in the protein level in P103R. A comparable model of Y273C indicated that Y273 is located in a putative proton-binding site. The findings of previous studies have indicated that mutations in the putative proton-binding site in MATE2-K caused reductions in protein expression and transport activity (Otsuka et al., 2005; Matsumoto et al., 2008; Choi et al., 2011).
In the present study, we demonstrated the screening of genetic variations in the SLC47A2 (MATE2-K) gene and functional analysis of coding SNPs. Four SNPs (P103R, P162L, G211V, and Y273C) affected protein expression and transport activities. Despite low allele frequencies, these alleles altered MATE2-K function. G393R was identified with a very low frequency in Caucasian (minor allele frequency = 0.01), and the significant effect on the trough steady-state metformin concentration was not observed. However, association analysis between the genotype and phenotype indicated that G393R could perhaps affect the long-term absolute decrease in hemoglobin A1c (Christensen et al., 2011). Additional studies are needed to analyze the contribution of the identified novel SNPs to the pharmacokinetics of MATE2-K substrates in humans.
Acknowledgments
The authors thank the Research Support Center, Graduate School of Medical Sciences, Kyushu University, for technical support.
Authorship Contributions
Participated in research design: Hirota and Ieiri.
Conducted experiments: Nishimura, Ide, Hirota, Kawazu, Kodama, and Takesue.
Performed data analysis: Nishimura, Ide, and Hirota.
Wrote or contributed to the writing of the manuscript: Nishimura, Ide, Hirota, and Ieiri.
Footnotes
- Received January 6, 2014.
- Accepted July 1, 2014.
This work was supported by JSPS KAKENHI Grant Number 26460199.
K.N. and R.I. contributed equally to this work.
Abbreviations
- BSA
- bovine serum albumin
- FITC
- fluorescein isothiocyanate
- HEK
- human embryonic kidney
- MATE2-K
- multidrug and toxin extrusion 2-K
- MGC
- Mammalian Gene Collection
- SLC47A2
- solute carrier family 47, member 2
- SNP
- single nucleotide polymorphism
- TEA
- tetraethylammonium
- WGA
- wheat germ agglutinin
- Copyright © 2014 by The American Society for Pharmacology and Experimental Therapeutics