Abstract
Gastrointestinal (GI) absorption of certain therapeutic agents is thought to be mediated by solute carrier (SLC) transporters, although minimal in vivo evidence has been reported. Here, we show key roles of postsynaptic density 95/disk-large/ZO-1 (PDZ) domain-containing protein, PDZK1, as a regulatory mechanism of two solute carriers, Slc15a1 (oligopeptide transporter PEPT1) and Slc22a5 (carnitine/organic cation transporter OCTN2) in mouse small intestine by using pdzk1 gene knockout (pdzk1–/–) mice. GI absorption of cephalexin, a substrate of PEPT1, after p.o. administration was delayed in pdzk1–/– mice compared with wild-type mice. Absorption of carnitine, a substrate of OCTN2, was also decreased in pdzk1–/– mice. Immunohistochemical analysis revealed the localization of both PEPT1 and OCTN2 at apical membrane of small intestinal epithelial cells in wild-type mice, whereas such apical localization was reduced in pdzk1–/– mice, with a concomitant decrease in their protein levels assessed by Western blotting in intestinal brush-border membranes. Electron microscopy revealed localization of PEPT1 in intracellular vesicular structures in pdzk1–/– mice. In addition, we first identified interaction between PEPT1 and PDZK1 in mouse small intestine and found that PDZK1 stimulates transport activity of PEPT1 by increasing its expression level in human embryonic kidney 293 cells. Taken together, the present findings provide direct evidence that PDZK1 regulates two intestinal SLC transporters in vivo as an adaptor protein for these transporters and affects oral absorption of their substrates. These findings also raise the possibility that intestinal absorption of the substrate drugs for PEPT1 and OCTN2 is governed by the protein network of these transporters and their adaptor PDZK1.
The membrane of intestinal epithelial cells is the first barrier to the absorption of therapeutic agents into the circulation, and permeation across this membrane is one of the critical factors determining oral bioavailability. It had been considered that intestinal absorption of lipophilic drugs is governed by simple diffusion according to the pH-partition hypothesis, but the intestinal absorption of several hydrophilic drugs, including amino β-lactam antibiotics, could not be simply explained by passive diffusion (Tsuji et al., 1979); therefore, carrier-mediated system(s) were suggested to be involved in the intestinal absorption (Tsuji et al., 1981). Since then, many studies have established the involvement of oligopeptide-specific transporters in the absorption of certain types of β-lactam antibiotics (Tsuji and Tamai, 1996).
The H+/oligopeptide transporter PEPT1 (SLC15A1) is localized on apical membranes of small intestinal epithelial cells and recognizes not only dipeptides and tripeptides but also peptide-like drugs, including oral β-lactam antibiotics, angiotensin-converting enzyme inhibitor, rennin inhibitor, and bestatin (Hu and Amidon, 1988; Frieman and Amidon, 1990; Kramer et al., 1990; Saito et al., 1993; Nakanishi et al., 1997; Tamai et al., 1997). Moreover, intestinal uptake and permeability of some β-lactam antibiotics are well correlated with the expression level of PEPT1 in the small intestine (Chu et al., 2001; Naruhashi et al., 2002). These data suggest that at least certain types of β-lactam antibiotics may be absorbed mainly via PEPT1 in the small intestine.
Transport of substrates by PEPT1 is known to be electrogenic, using the proton gradient across the plasma membrane as a driving force (Ganapathy and Leibach, 1985). This proton gradient is produced by another transporter, Na+/H+ exchanger (NHE), which pumps out protons using the Na+ gradient as a driving force. Thus, NHE is thought not only to supply the driving force for PEPT1 but also to export protons transported by PEPT1 together with its substrates to maintain the proton balance across the plasma membrane. Thus, it is hypothesized that transport function of PEPT1 is also affected by NHE activity. This idea has been supported by the observation in gastrointestinal (GI) epithelial cells, Caco-2, and human embryonic kidney (HEK) 293 cells transiently transfected PEPT1 and NHE3 (Kennedy et al., 2002; Thwaites et al., 2002; Watanabe et al., 2005). Thus, it is possible that NHE is localized in proximity to PEPT1 to maintain its transport function, and such cooperation with other transporter(s) may be a key feature of PEPT1 function. However, little information is available on protein-protein interactions involving PEPT1 in the small intestine.
On the other hand, carnitine/organic cation transporter OCTN2 (SLC22A5) is also localized on apical membrane of small intestinal and renal epithelial cells and is physiologically important for intestinal absorption and renal tubular reabsorption of its endogenous substrate, carnitine, which plays an important role for the β-oxidation of fatty acids in mitochondria (Nezu et al., 1999; Kato et al., 2006). Recently we have revealed that OCTN2 colocalized with postsynaptic density 95/disk-large/ZO-1 (PDZ) domain-containing protein PDZK1 on apical membrane of absorptive epithelial cells, and OCTN2 interacts with PDZK1 in mouse small intestine (Kato et al., 2006). PDZK1 has four PDZ domains in its structure, and each PDZ domain alone can directly bind to the C terminus of various transporters. PDZK1 has been suggested to be a functional regulator of various transporters and may be a scaffold protein that links those transporters at certain microdomains on the epithelial cell surface (Gisler et al., 2003; Anzai et al., 2004; Kato et al., 2004, 2005; Miyazaki et al., 2005; Sugiura et al., 2006), raising the possibility that PDZK1 interacts with various intestinal transporters, including OCTN2 and PEPT1. However, most of the data suggesting such functional regulation by PDZK1 were obtained in in vitro double-transfection studies with cultured cell lines. Although Wang et al. (2005) recently showed an essential role of PDZK1 in basolateral expression of organic anion transporting peptide 1a1 in the liver, the in vivo evidence for a fundamental role of PDZK1 as a regulatory mechanism for xenobiotic transporters is quite limited in small intestine.
In the present study, we focused on PDZK1 as a candidate regulator of PEPT1 and OCTN2 and examined its pharmacological relevance in vivo. A pharmacokinetic study of cephalexin (a substrate of PEPT1) and carnitine (a substrate of OCTN2) using pdzk1 gene knockout mice (pdzk1–/–) was performed to clarify the role of PDZK1 in intestinal absorption of substrates of intestinal transporters. In addition, in vitro studies were also performed to support the regulatory role of PDZK1 for PEPT1. Our findings show that both PEPT1 and OCTN2 are regulated in vivo by an interaction with PDZK1 in the small intestine. The present findings also provided further support for a fundamental role of PEPT1 in the intestinal absorption of β-lactam antibiotics in vivo.
Materials and Methods
Materials. Rabbit polyclonal antibody to PEPT1, rabbit polyclonal antibody to OCTN2, and rat polyclonal antibody to PDZK1 were raised as described previously (Sai et al., 1996; Tamai et al., 2000; Kato et al., 2005). Monoclonal antibodies against c-myc epitope tag (9E10), green fluorescent protein (GFP), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and β-actin (AC-15) were obtained from Covance Inc. (Princeton, NJ), Roche Diagnostics (Basel, Switzerland), Chemicon International, Inc. (Temecula, CA), and Sigma-Aldrich, Inc. (St. Louis, MO), respectively. [3H]Glycylsarcosine (GlySar, 0.5 Ci/mmol) and [14C]mannitol (53 mCi/mmol) were purchased from Moravek Biochemicals Inc. (Mercury Lane, Brea, CA); l-[3H]carnitine (84 Ci/mmol) was purchased from GE Healthcare (Piscataway, NJ); and [14C]inulin (1–3 mCi/g) was purchased from PerkinElmer Life Sciences (Boston, MA). Cephalexin hydrate was purchased from Sigma-Aldrich Japan K.K. (Tokyo, Japan). Antipyrine was provided from Wako Pure Chemical Industries (Osaka, Japan). The protein assay kit was purchased from Bio-Rad Laboratories, Inc. (Hercules, CA). cDNA encoding full-length human PEPT1 was subcloned into the pEYFP-C1 vector (Clontech, Mountain View, CA). Other reagents were obtained from Sigma Chemical Co. (St. Louis, MO), Wako Pure Chemical Industries, Funakoshi Co. (Tokyo, Japan), and Nacalai Tesque, Inc. (Kyoto, Japan) and were used without further purification.
Animals. Male mice were used for all the experiments at 6 to 12 weeks of age. The pdzk1–/– mice had been previously produced (Lan and Silver, 2005), and pdzk1–/– and littermates were of a mixed genetic background (C57BL/6J and 129Sv/Ev) produced by intercrossing pdzk1+/– mice. They had free access to food and water. This study was carried out in accordance with the Guide for the Care and Use of Laboratory Animals at the Takara-machi Campus of Kanazawa University.
Pharmacokinetic Studies. Mice were fasted overnight with free access to water and anesthetized with diethyl ether during drug administration and blood sampling. Cephalexin (2.0 mg/kg b.wt.), [3H]carnitine (250 ng/kg b.wt.), and antipyrine (20 mg/kg b.wt.) were administered p.o. by gavage. At various intervals after administration, aliquots of blood samples were collected through the caudal vein. All the blood samples were immediately centrifuged to obtain plasma, which was further mixed with two volumes of acetonitrile, then centrifuged, and the resultant supernatant was used for quantitation.
Amount of [3H]Carnitine Retained in GI Tract after p.o. Administration. [3H]Carnitine and [14C]inulin were dissolved in saline and administered p.o. by gavage. The initial concentration of carnitine was 125 ng/ml, and the volume of the solution administered was 2 ml/kg. Mice then were sacrificed at 30 min after administration, and their abdomen was opened immediately to take the sample of residual [3H]carnitine and [14C]inulin from each segment of the GI tract. The mouse alimentary tract was divided as follows: stomach, upper small intestine, middle small intestine, lower small intestine, cecum, and large intestine.
Uptake Studies in Everted Intestinal Sac. An everted intestinal sac was prepared from the upper, middle, and lower part of the small intestine of wild-type and pdzk1–/– mice. After 5 min of preincubation in the transport buffer (125 mM NaCl, 4.8 mM KCl, 1.2 mM KH2PO4, 5.6 mM d-glucose, 1.2 mM CaCl2, 1.2 mM MgSO4, and 25 mM MES, pH 6.0), the sac was incubated with [3H]carnitine and [14C]mannitol (extracellular marker) at 37°C for the designated times. The sac was then washed with ice-cold buffer, weighed, and solubilized in Soluene-350 (Packard Co., Canberra, NSW, Australia) by incubation at room temperature for 24 h. The solubilized samples were decolorized with 0.2 ml of H2O2, neutralized with 0.1 ml of 5 N HCl, left at room temperature for 4 h, and mixed with 3 ml of a scintillation mixture. Radioactivity was then counted using a liquid scintillation counter.
Analytical Procedures. Cephalexin and antipyrine were measured with a liquid chromatography/tandem mass spectrometry (LC/MS/MS) system equipped with a constant flow pump (Agilent 1200 series G1312A, Agilent Technologies, Tokyo, Japan), an automatic sample injector (G1367B, Agilent Technologies), a column oven (G1316A, Agilent Technologies), and a mass spectrometer (API 3200, Applied Biosystems). For cephalexin, the analytical column was COSMOSIL AR-II (2.0 × 150 mm, Nacalai Tesque, Inc.). Mobile phase A was 0.01 M ammonium formate, and mobile phase B was 100% methanol. The gradient elution time program was set as follows: 0 to 4 min, B 5 to 35%; 4 to 16 min, B 35%; and 16.1 to 25 min, B 5%. The flow rate was 0.2 ml/min. For antipyrine, the column was COSMOSIL MS-II (2.0 × 50 mm, Nacalai Tesque, Inc.), and the gradient elution time program was set as follows: 0 to 1 min, B 7 to 70%; 1 to 9.5 min, B 70%; and 9.6 to 19.6 min, B 7%. The radioactivity of [3H]GlySar, [3H]carnitine, and [14C]mannitol was measured with a liquid scintillation counter, LSC-5100 (Aloka, Tokyo, Japan), with Clearsol I (Nacalai Tesque, Inc.) as a scintillation fluid.
Immunohistochemical Analysis. Frozen sections of mouse small intestine were prepared as described previously (Kato et al., 2006). Following successive pretreatments with 0.3% Tween 20 in phosphate-buffered saline (PBS), 0.3% H2O2 in methanol, and 5% bovine serum albumin in PBS, the sections were incubated with anti-PEPT1 or anti-OCTN2 antibody. They were washed with PBS, and the immunoreaction product was visualized by incubating the section successively with biotinylated anti-rabbit IgG (Vector Laboratories, Burlingame, CA) for 2 h, horseradish peroxidase-conjugated streptavidin (DakoCytomation, Kyoto, Japan) for 1 h, and 3′,3′-diaminobenzidine tetrahydrochloride containing H2O2 for a few minutes. Finally, they were mounted in Vectashield mounting medium with 4,6-diamidino-2-phenylindole (Vector Laboratories) to fix the sample. The specimens were examined with an Axiovert S 100 microscope (Carl Zeiss, Jena, Germany).
For electron microscopic immunocytochemistry, the pre-embedding immunoreaction method was used (Wakayama et al., 2004). Cryosections immunostained with anti-PEPT1 antibody were postfixed in 0.5% OsO4 for 20 min, stained with 1% uranyl acetate for 20 min, dehydrated in a graded ethanol series, and embedded in Glicidether 100 (Selva Fenbiochemica, Heidelberg, Germany). Ultrathin sections were cut with an ultramicrotome and observed with a JEM-1210 electron microscope (JEOL, Tokyo, Japan).
Immunoprecipitation and Western Blot Analysis of Intestinal Brush-Border Membrane Vesicles. Brush-border membrane vesicles (BBMVs) were prepared from mouse small intestine as described by Tsuji et al. (1987) with some modifications. Briefly, the mucosa was scraped off and homogenized in ice-cold 2 mM Tris-HCl buffer, pH 7.1, containing 50 mM mannitol, in a volume (milliliters) equal to 30 times the weight (grams) of the mucosa scrapings. Homogenization was carried out with an ULTRA-TURRAX T25 (IKA-Werke GmbH & Co. KG, Staufen, Germany) for 5 min at the maximum speed. CaCl2 solution (1 M) was added to the homogenate to give a final concentration of 10 mM, and the mixture was stirred in a cold room for 15 min. It was then centrifuged at 3000g for 15 min, and the supernatant was centrifuged at 27,000g for 30 min. The resultant pellet was resuspended in a washing medium through a 25-gauge needle. The washing medium used was 10 mM Tris-HEPES buffer, pH 7.5, containing 270 mM mannitol.
Thus, intestinal BBMVs obtained were solubilized in radioimmunoprecipitation assay-Y buffer containing 1% Nonidet P40, 75 mM NaCl, 50 mM Tris-HCl, pH 7.5, and protease inhibitors. Anti-PDZK1 antibody prebound to Protein L Sepharose (Pierce Chemical, Rockford, IL) or anti-PEPT1 antibody prebound to Protein G Sepharose (GE Healthcare, Little Chalfont, Buckinghamshire, UK) was added to the obtained lysate, and the mixture was incubated at 4°C for 8 h, followed by centrifugation and washing three times with PBS. Samples were analyzed by SDS-polyacrylamide gel electrophoresis, followed by immunoblotting with anti-PEPT1 or anti-OCTN2 antibody. Western blot analysis was performed as described previously (Kato et al., 2005).
Determination of mRNA Expression by Real-Time Polymerase Chain Reaction. Total RNAs were extracted from small intestine of wild-type and pdzk1–/– mice (male, 8 weeks), and cDNA was synthesized as described previously (Naruhashi et al., 2002). Quantification of the mRNA coding PEPT1 and β-actin was performed using LightCycler technology (Roche Diagnostics). The PEPT1 forward and reverse primers were 5′-AGGGTCACCGGCACAACCTT-3′ and 5′-GACGGAGCCTGGGAATAAGAG-3′, respectively, which produce a 469-base pair amplicon. The β-actin forward and reverse primers were 5′-AGAGGGAAATCGTGCGTGACA-3′ and 5′-CAATAGTGATGACCTGGCCGT-3′, respectively, which produce a 138-base pair amplicon. The polymerase chain reaction (PCR) conditions for PEPT1 were as follows: initial denaturation at 95°C for 10 min, followed by 45 cycles of denaturation at 95°C for 0 s (followed by an immediate change to the next temperature) and combined annealing extension at 60°C for 0 s and 72°C for 19 s, followed by fluorescence emission reading at 78°C for 10 s. In the case of the PCR for β-actin, the annealing temperature was 58°C for 0 s, and the extension temperature was 72°C for 6 s. The quantification value was calculated as the average of three determinations of each of three different dilutions.
Transport Studies in Cell Culture. HEK293 cells were routinely grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, penicillin, and streptomycin in a humidified incubator at 37°C and 5% CO2. HEK293 cells stably expressing myc-tagged PDZK1 (HEK293/PDZK1 cells) were previously obtained (Kato et al., 2004) and were grown under the same conditions, except that the medium also contained 0.1% G-418. cDNA encoding human PEPT1 was subcloned into the pEYFP-C1 vector (Clontech). cDNA encoding PEPT1Δ4 (with deletion of the last four amino acids) was amplified from pEYFP-C1/PEPT1 by PCR using 5′- and 3′-specific primers (5′-AGATCTGCCGCCATGGGAATGTCCAAATC-3′ and 5′-CAATTCACAGAAACAGATGTGAGTCGAC-3′, respectively). The PCR product was subcloned into the pEYFP-C1 vector using BglII/SalI sites, and sequence for the insert was verified. After 24-h cultivation, the cDNA construct was transiently transfected into HEK293 or HEK293/PDZK1 cells by adding 20 μg/15-cm dish of the plasmid DNA according to the calcium phosphate precipitation method (Tamai et al., 2000). At 48 h after transfection, cells were harvested and suspended in transport medium (125 mM NaCl, 4.8 mM KCl, 5.6 mM d-glucose, 1.2 mM CaCl2, 1.2 mM KH2PO4, 1.2 mM MgSO4, and 25 mM HEPES, pH 7.4). A [3H]GlySar uptake experiment was then performed according to the silicone oil layer method (Tamai et al., 2000). According to our preliminary analysis, uptake of [3H]GlySar was linear until 3 min; therefore, initial uptake was examined at 3 min as reported previously (Watanabe et al., 2005). To analyze transport of cephalexin, the cell suspension and the transport medium containing cephalexin (100 μM) were mixed to initiate the transport reaction. At the designated times, 200-μl aliquots of the mixture were withdrawn, and the cells were separated from the transport medium by centrifugal filtration through a layer of a mixture of silicon oil (SH550, Toray Dow Corning, Tokyo, Japan) and liquid paraffin (Wako Pure Chemicals) on top of 3 M KCl solution. After sonication, cells was deproteinized with acetonitrile, then centrifuged for 5 min at 15,000g, and the resulting supernatant was measured by LC/MS/MS.
Subcellular Localization and Immunoprecipitation Analysis in HEK293 Cells. HEK293 cells were grown on cover glass (15 × 15 mm; thickness, 0.12–0.17 mm) (Matsunami Glass Inc., Osaka, Japan) and transiently transfected with cDNA encoding yellow fluorescent protein (YFP)-PEPT1. At 48 h after the transfection, cells were fixed with 3% formaldehyde in PBS. Attached cells were sealed onto the slides using Vectrashield mounting medium with 4′,6-dianidino-2-phenylindole (Vector Laboratories), and the fluorescence was detected with a confocal laser scanning fluorescence microscope (LSM 510, Carl Zeiss). At 48 h after transfection of YFP-PEPT1 in HEK293/PDZK1, the cells were washed twice with PBS, harvested, and solubilized in radioimmunoprecipitation assay-Y buffer. Western blot analysis was then performed using anti-GFP or anti-GAPDH antibodies as described previously (Kato et al., 2005). For immunoprecipitation, cell lysate was added to anti-myc antibody prebound to Protein L Sepharose (Pierce Chemical), and the mixture was incubated at 4°C for 8 h, followed by centrifugation, washing three times with PBS, and Western blot analysis.
Data Analysis. The results are expressed as mean ± S.E.M., and statistical analysis was performed by using Student's t test. A difference between means was considered to be significant when p < 0.05.
Results
Characterization of the Interaction of PEPT1 and PDZK1 in HEK293 Cells. Coexpression with PDZK1 increased the apparent uptakes by several solute carrier (SLC) transporters of their substrates in HEK293 cells, although the effect of PDZK1 on PEPT1 has not yet been examined. To examine whether the function of PEPT1 is modulated by PDZK1, uptake study was first performed in HEK293/PDZK1 cells transiently expressing PEPT1 in the present study. Time-dependent uptake of [3H]GlySar by PEPT1 was observed in both HEK293 and HEK293/PDZK1 cells, with the initial slope of the uptake being higher in HEK293/PDZK1 cells than in HEK293 cells (Fig. 1A). To examine the structural requirements at the C terminus of PEPT1 for the regulation of PEPT1, a mutant construct, PEPT1Δ4, lacking the last four amino acids of PEPT1 (–QKQM) was constructed. [3H]GlySar uptake by PEPT1Δ4 was also higher in HEK293/PDZK1 than HEK293 cells (Fig. 1D). As in the case of [3H]GlySar uptake, uptake of cephalexin mediated by PEPT1 was stimulated in the presence of PDZK1 (Fig. 1C). Saturable uptake of GlySar was observed during transient expression of PEPT1 in both HEK293/PDZK1 and HEK293 cells. Eadie-Hofstee plots revealed the existence of a single component, with a similar slope for the two cell lines, for the saturable uptake. The obtained values for Km and Vmax were 0.516 ± 0.015 and 0.549 ± 0.128 mM and 2.74 ± 0.12 and 5.61 ± 1.26 nmol/mg protein/3 min in HEK293 and HEK293/PDZK1 cells, respectively (Fig. 1B).
To examine the physical interaction of PEPT1 and PDZK1 in HEK293 cells, we next performed immunoprecipitation using anti-myc antibody (to myc-PDZK1) in lysates of HEK293/PDZK1 cells transiently transfected with PEPT1, followed by Western blot analysis using anti-GFP antibody, because HEK293/PDZK1 cells stably express myc-tagged PDZK1, whereas PEPT1 was transiently transfected as a YFP fusion protein. Immunoprecipitation with c-myc antibody resulted in a predicted band of approximately 100 kDa with immunoreactivity to GFP antibody (Fig. 2A, left, lane 3). This YFP-PEPT1 band was also observed when PEPT1Δ4 was transfected (Fig. 2A, left, lane 4) but not when PDZK1 was not stably expressed or when PEPT1 was not transiently transfected (lanes 1 and 2, respectively). Immunoprecipitation also yielded a 75-kDa protein that was immunoreactive, with c-myc antibody (Fig. 2A, right) having the predicted molecular mass for myc-PDZK1.
The subcellular localization of PEPT1 and its possible colocalization with PDZK1 were further examined. In both HEK293 and HEK293/PDZK1 cells, PEPT1 was mainly localized in plasma membranes with some minor localization in intracellular space (Fig. 2C). To compare expression levels of PEPT1, Western blot analysis was then performed for cellular lysates from HEK293 or HEK293/PDZK1 transiently transfected with PEPT1. Expression of PEPT1 was slightly higher in HEK293/PDZK1 cells compared with HEK293 cells (Fig. 2B, top). Expression of endogenous GAPDH, examined in a control experiment, was comparable in the two lines (Fig. 2B, bottom). The -fold increase in the PEPT1 expression was 1.7 when normalized to GAPDH expression, and this value approximated the difference in the uptake of [3H]GlySar (Fig. 1A).
Physical Interaction between PEPT1 and PDZK1 in Mouse Small Intestine. To examine the possible interaction between PDZK1 and PEPT1 under physiological conditions, immunoprecipitation of lysates of intestinal BBMVs was performed using anti-PDZK1 antibody, followed by Western blot analysis with anti-PEPT1 antibody (Fig. 3A). PEPT1 was detected in anti-PDZK1 immunoprecipitates together with PDZK1, but not when preimmune control serum was used for immunoprecipitation (Fig. 3A). These results revealed that PEPT1 interacts with PDZK1 in mouse small intestine, as in the case between OCTN2 and PDZK1 (Kato et al., 2006).
Reduced Expression of PEPT1 and OCTN2 on the Apical Membranes of the Small Intestine inpdzk1–/–Mice. To further examine the interaction of PDZK1 with PEPT1 and OCTN2, pdzk1–/– mice were used to analyze the change in expression and localization of these transporters in the small intestine of pdzk1–/– mice. Immunoreactive bands of PEPT1 and OCTN2 assessed in Western blot analysis were found to be expressed in the small intestine from wild-type mice, whereas these transporters in the small intestine from pdzk1–/– mice were less than wild-type mice (Fig. 3B). We confirmed by Western blot analysis that PDZK1 protein was not expressed in pdzk1–/– mice, in contrast with wild-type mice (Fig. 3B). Expression of β-actin was observed in all the samples of small intestine (Fig. 3B).
We also undertook an immunohistochemistry analysis of PEPT1 and OCTN2 in the small intestine from wild-type and pdzk1–/– mice (Fig. 4). PEPT1 and OCTN2 were localized on the apical brush-border membrane in wild-type mice (Fig. 4, A and C). However, in pdzk1–/– mice, PEPT1 and OCTN2 stainings at the apical surface were noticeably weaker in the small intestine (Fig. 4, B and D). These results were compatible with the results of Western blot analysis (Fig. 3B).
We also examined the mRNA copy number of PEPT1 in the upper intestine from both wild-type and pdzk1–/– mice by real-time quantitative PCR analysis. PEPT1 mRNA expression normalized to that of β-actin was not significantly different in pdzk1–/– compared with wild-type mice (Fig. 3C), indicating that changes in PEPT1 levels occurred on a post-transcriptional level.
Subcellular Localization of PEPT1 in Absorptive Epithelial Cells. The subcellular localization of PEPT1 in mouse small intestine was further analyzed by immunoelectron microscopy (Fig. 5, A and B). Immunoreactivity of PEPT1 in wild-type mice was mainly found in microvilli in the apical region of absorptive epithelial cells (Fig. 5A). However, in pdzk1–/– mice, immunoreactivity of PEPT1 on microvilli was quite weak, but slight, diffuse immunoreactivity was seen in the cytoplasmic region of the absorptive epithelial cells (Fig. 5B). In wild-type mice, the localization of PEPT1 was similar to that of the terminal web of actin filaments, as in the cases of OCTN2 and PDZK1 (Kato et al., 2006). The ultrastructural abnormality of the intestinal epithelial cells was observed in pdzk1–/– mice (Fig. 5D). In pdzk1–/– mice, vesicular structures were more obviously found in the cytoplasm of absorptive epithelial cells compared with wild-type mice (Fig. 5, C and D).
Intestinal Absorption of Cephalexin, Carnitine, and Antipyrine in Wild-Type andpdzk1–/–Mice. To investigate the role of PDZK1 in intestinal drug absorption in vivo, a mixture of several β-lactam antibiotics (cephalexin, cephradine, cefaclor, cefatrizine) was administered i.v. or p.o. to both wild-type and pdzk1–/– mice, followed by LC/MS/MS measurement of the plasma concentration-time profile of each compound. Plasma concentrations of the i.v. infected compounds exhibited only minimal differences between wild-type and pdzk1–/– mice, whereas those after p.o. administration exhibited slight but consistent differences between the two strains for all four compounds examined (data not shown). Based on these initial findings, we next investigated in detail the pharmacokinetic profile of cephalexin, a known substrate of PEPT1. The i.v. administration of cephalexin resulted in similar monophasic plasma concentration-time profiles in wild-mice and pdzk1–/– mice (data not shown). In contrast, plasma concentration of cephalexin after p.o. administration was lower in pdzk1–/– mice than that in wild-type mice (Fig. 6A). Because GI absorption of cephalexin is thought to be mediated by PEPT1 expressed on the apical membrane of the small intestine, it is possible that the delay of cephalexin absorption in pdzk1–/– mice (Fig. 6A) is caused by a defect of PEPT1 in plasma membrane (Figs. 4 and 5).
Plasma concentration of radioactivity after p.o. administration of [3H]carnitine, a substrate of OCTN2, was lower in pdzk1–/– mice compared with wild-type mice (Fig. 6B). Our previous report suggested that OCTN2 plays a predominant role for the uptake of carnitine from apical surface of mouse small intestine (Kato et al., 2006). Decreased absorption of carnitine (Fig. 6B) was presumably caused by decreased expression of OCTN2 at apical membrane of intestinal epithelial cells in pdzk1–/– mice (Fig. 4). In contrast, the plasma concentration of antipyrine after p.o. administration was similar between wild-type and pdzk1–/– mice (Fig. 6C).
Because l-carnitine is known to be converted to acetylcarnitine and other acylcarnitines in tissues (Rebouche and Paulson, 1986), we measured the amount of [3H]carnitine retained in the GI tract at 30 min after p.o. administration (Fig. 7A). An extracellular marker, [14C]inulin, retained in the GI tract was also measured (Fig. 7B). [14C]Inulin retained in each segment of the GI tract was similar, and total recovery of [14C]inulin was approximately 100% in both wild-type and pdzk1–/– mice (Fig. 7B). On the other hand, [3H]carnitine retained in the upper and middle parts of the small intestine was higher in pdzk1–/– mice than that in wild-type mice (Fig. 7A), indicating that absorption of [3H]carnitine was smaller in pdzk1–/– mice compared with wild-type mice. Moreover, the uptake of carnitine from the apical surface into the everted intestinal tissue was investigated (Fig. 8). The uptake of [3H]carnitine from apical surface by the intestinal tissue was much higher in wild-type mice than that in pdzk1–/– mice and decreased in the presence of 20 mM unlabeled carnitine (Fig. 8). The extracellular space assessed as the apparent uptake of [14C]mannitol was almost similar between wild-type and pdzk1–/– mice (0.11 ± 0.01 and 0.14 ± 0.02 μl/mg tissue, respectively).
Discussion
The H+/oligopeptide transporter PEPT1 (SLC15A1) is thought to be involved in absorption of dipeptides and tripeptides, as well as various types of therapeutic agents, including β-lactam antibiotics. In a previous report, we showed that intestinal transport of another β-lactam antibiotic, cefadroxil, is correlated with the PEPT1 expression level in the small intestine of both fed and starved rats (Naruhashi et al., 2002). Thus, at least some β-lactam antibiotics may be absorbed mainly via PEPT1 in the small intestine. Despite the transport activity of PEPT1, the regulatory mechanisms of this transporter in vivo are still poorly understood. Our findings here indicate that PEPT1 is functionally regulated by PDZK1 in mouse small intestine and is likely to be involved in GI absorption of cephalexin in vivo. The following four lines of evidence support this conclusion: first, after p.o. administration of cephalexin in vivo, plasma concentration was lower in pdzk1–/– mice compared with wild-type mice (Fig. 6A), suggesting intestinal absorption of cephalexin was decreased in pdzk1–/– mice. Second, in pdzk1–/– mice PEPT1 levels on apical membranes of small intestine were reduced compared with wild-type mice (Figs. 4 and 5), with a concomitant decrease in its protein expression level in BBMVs of the small intestine (Fig. 3B). Third, PEPT1 is immunoprecipitated by antibody to PDZK1 (Fig. 3A), suggesting a physical interaction of the two proteins that was confirmed in vitro in HEK293 cells transfected with these proteins (Fig. 2A). Fourth, coexpression with PDZK1 affected the PEPT1 expression level in HEK293 cells (Fig. 2B), thereby increasing PEPT1-mediated transport of substrates, GlySar and cephalexin (Fig. 1, A and C), with an increase in transport capacity (Vmax; Fig. 1D). This in vitro result seems consistent with the reduction of PEPT1 on the apical membrane of pdzk1–/– mice in vivo (Figs. 4 and 5). Together, these data indicate that PDZK1 is a regulatory factor of PEPT1 levels and function in mouse small intestine, providing the first demonstration of the potential pharmacological role of PEPT1-PDZK1 complex in small intestine in vivo.
Similarly, plasma concentration of [3H]carnitine after p.o. administration was decreased in pdzk1–/– mice compared with wild-type mice (Fig. 6B), suggesting intestinal absorption of carnitine was decreased in pdzk1–/– mice, as in the case of juvenile visceral steatosis (jvs) mice, which have a hereditary deficiency of the octn2 gene (Yokogawa et al., 1999; Kato et al., 2006). The reduced oral absorption of carnitine in pdzk1–/– mice was also supported by the elevated [3H]carnitine retained in the GI tract after p.o. administration (Fig. 7) and decreased uptake by the intestinal tissue from apical surface (Fig. 8) in pdzk1–/– mice. We have recently reported that OCTN2 is localized with PDZK1 in microvilli of absorptive epithelial cells, and uptake of carnitine was quite small and unsaturable in enterocytes obtained from jvs mice (Kato et al., 2006), suggesting carnitine is mainly transported by intestinal OCTN2. In pdzk1–/–mice, reduced expression of intestinal OCTN2 (Fig. 4, C and D) may lead to the decreased intestinal absorption of carnitine. On the other hand, plasma concentration of antipyrine, a marker compound for passive diffusion, was similar between wild-type and pdzk1–/– mice (Fig. 6C), suggesting that PDZK1 might affect only the substrates of intestinal transporters that interact with PDZK1 in the small intestine. From in vitro investigation, OCTN2 could recognize other organic cation drugs including verapamil and quinidine (Ohashi et al., 2001); therefore, OCTN2 could be a possible target for oral delivery of therapeutic agents.
Although the transport activity of oligopeptide and carnitine was reduced in small intestine of pdzk1–/– mice, no obvious phenotypic change except hyperlipemia has been reported in pdzk1–/– mice (Kocher et al., 2003; Lan and Silver, 2005). In pdzk1–/– mice, weight gain was similar to wild-type mice. Cardiac hypertrophy and fatty liver, both of which are observed in systemic carnitine-deficient jvs mice with octn2 gene deficiency (Nezu et al., 1999), were not observed in pdzk1–/– mice. One of the possible reasons would be a compensatory hyperfunction of amino acid transporters because of the decrease in the oligopeptide transport activity in the small intestine of pdzk1–/– mice. It is also possible that metabolism and/or excretion of carnitine may be changed in pdzk1–/– mice, thereby carnitine deficiency was not observed. In addition, other adaptor protein(s) than PDZK1 may function as a compensatory regulator of PEPT1 and/or OCTN2 in pdzk1–/– mice. Therefore, additional studies are required to clarify the compensatory mechanisms caused by the loss of pdzk1 gene in the small intestine.
Considering that there is little difference in PEPT1 mRNA level between wild-type and pdzk1–/– mice (Fig. 3C), PDZK1 appears to be involved in the post-transcriptional regulation of PEPT1. This may be explained by internalization of PEPT1 in intestinal epithelial cells in pdzk1–/– mice. In fact, electron microscopy revealed a remarkable loss of PEPT1 protein in microvilli, whereas immunoreactivity was observed in the cytoplasmic region in intestinal epithelial cells from pdzk1–/– mice (Fig. 5B). It is noteworthy that in pdzk1–/– mice vesicular structures were more obviously seen in the cytoplasmic region (Fig. 5, C and D). We have not yet been able to identify these intracellular structures exhibiting immunoreactivity against PEPT1 antibody (Fig. 6, B and D). Considering the essential role of PDZK1 in the apical localization of PEPT1 in vivo (Figs. 5 and 6), the appearance of such vesicular structures may reflect defective sorting of PEPT1 to apical membranes and/or unstable localization of PEPT1 on the apical membrane surface. Similar loss of PEPT1 expression on the apical membrane was recently found in small GTP-binding protein rab8 gene knockout (rab8–/–) mice (Sato et al., 2007), indicating a pivotal role of Rab8 in sorting of PEPT1 to the apical membranes. However, the regulatory mechanism for PEPT1 is clearly different between Rab8 and PDZK1 because the length of microvilli was reduced in rab8–/– mice, and the mutation was lethal just after the weaning period, whereas the reduction in the length of microvilli was not so obvious in pdzk1–/– mice, which have normal fertility (Lan and Silver, 2005). However, in the present study, we first revealed ultrastructural change in cytoplasm of absorptive epithelial cells in pdzk1–/– mice (Fig. 5D).
There has been no report regarding proteins that physically interact with PEPT1; this is the first demonstration of the interaction between PEPT1 and PDZK1 in small intestinal lysates and HEK293 cells transfected with both proteins by immunoprecipitation analysis (Figs. 2A and 3A). This physical interaction between PEPT1 and PDZK1 was unexpected because, according to our previous yeast two-hybrid analysis, no interaction was observed between the C terminus of PEPT1 and PDZK1 (Kato et al., 2004), implying that some structural domain(s) other than the C terminus could be involved in the interaction with PDZK1. This was also confirmed by the present finding that PDZK1 not only stimulated [3H]GlySar uptake by PEPT1 but also that by PEPT1Δ4 (Fig. 1, A and D). This is in contrast to the regulation by PDZK1 of other SLC transporters (Anzai et al., 2004; Kato et al., 2004, 2005; Miyazaki et al., 2005; Sugiura et al., 2006), in which the PDZ binding motif at the extreme C terminus is essential for a PDZK1-mediated increase in expression and/or transport activity. Although the PDZ binding motif is localized at the C terminus of most interacting transporters and receptors, several PDZ domains also recognize a specific internal peptide motif (Hillier et al., 1999). Another possibility would be an indirect interaction of PEPT1 and PDZK1 via another protein. For example, the small PDZK1-associated protein DD96/MAP17 has been shown to interact directly with PDZK1 as a regulatory mechanism (Pribanic et al., 2003; Silver et al., 2003) but also interacts with the N terminus of Na+/π -IIa (Pribanic et al., 2003).
In summary, our present findings show that PDZK1 regulates apical localization of PEPT1 and OCTN2 in the small intestine and is presumably involved in the GI absorption of their substrates in vivo. The regulation of PEPT1 localization may be closely related to our recent observation that the small GTP-binding protein Rab8 plays pivotal roles in apical localization of PEPT1 in vivo (Sato et al., 2007). Thus, PEPT1 does not work alone under physiological conditions but forms a complex with the other proteins. PDZK1 may also interact with the transporters other than PEPT1 and OCTN2 expressed on apical membranes of small intestine (Wang et al., 2000; Rossmann et al., 2005; Kato et al., 2006; Hillesheim et al., 2007), and such protein-protein interactions may also provide functional coupling of these transporters, as proposed in the case of PEPT1 and NHE3 in vitro (Watanabe et al., 2005). Therefore, an understanding of protein-protein interactions involving xenobiotic transporters is likely to be important to predict the permeability and GI absorption of drugs and drug candidates.
Acknowledgments
We thank Lica Ishida for technical assistance. We also thank Tomoko Yamada for performing Western blot analysis.
Footnotes
-
This study was supported in part by a Grant-in-Aid for Scientific Research provided by the Ministry of Education, Science and Culture of Japan and a grant from the Mochida Memorial Foundation (Tokyo, Japan) for Medical and Pharmaceutical Research.
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.107.020321.
-
ABBREVIATIONS: PEPT1, oligopeptide transporter 1; NHE, Na+/H+ exchanger; GI, gastrointestinal; HEK, human embryonic kidney; OCTN2, carnitine/organic cation transporter 2; PDZ, postsynaptic density 95/disk-large/ZO-1; GFP, green fluorescent protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GlySar, [3H]glycylsarcosine; LC/MS/MS, liquid chromatography/tandem mass spectrometry; PBS, phosphate-buffered saline; BBMV, brush-border membrane vesicle; PCR, polymerase chain reaction; YFP, yellow fluorescent protein; SLC, solute carrier; jvs, juvenile visceral steatosis.
- Received December 27, 2007.
- Accepted February 26, 2008.
- The American Society for Pharmacology and Experimental Therapeutics