Drug Metab. Pharmacokinet. 21 (5): 375–383 (2006).

Regular Article Mutation in an Adaptor Protein PDZK1 AŠects Transport Activity of Organic Cation Transporter OCTNs and Oligopeptide Transporter PEPT2

Tomoko SUGIURA,YukioKATO, Yoshiyuki KUBO and Akira TSUJI* Division ofPharmaceutical Sciences, Graduate School ofNatural Science and Technology, Kanazawa University, Kanazawa, Japan

Full text of this paper is available at http://www.jstage.jst.go.jp/browse/dmpk

Summary: Genetic polymorphisms in xenobiotic transporters have recently been clariˆed to be associ­ ated with change in drug distribution and disposition. To expand on recent identiˆcation of direct interaction and functional regulation of several transporters by a PDZ (PSD95, Dlg and ZO1) domain containing protein PDZK1, the eŠect of mutation in PDZK1 on transport activity and subcellular localiza­ tion of organic cationWcarnitine transporters OCTN1 and OCTN2, and oligopeptide transporter PEPT2 was examined in the present study. HEK293 cells stably expressing a mutant transcript PDZK1­E195K (HEK293WPDZK1­E195K) were constructed, followed by transient transfection of cDNA for each trans­ porter. Uptake of tetraethylammonium by OCTN1 was much higher in HEK293WPDZK1 cells, com­ pared with that in the parent HEK293 cells, the uptake in HEK293WPDZK1­E195K cells showing middle range between the two values. Such diŠerence in transport activity was accounted for the diŠerence in transport capacity, with minimal change in a‹nity of OCTN1 to the substrate or other compounds. The similar diŠerence among HEK293WPDZK1, HEK293WPDZK1­E195K and HEK293 cells was also observed in transport property of OCTN2 and PEPT2, whereas the diŠerence was not so remarkable in each transporter with the last four amino acids deleted, that has much lower interaction potential with PDZK1. Immunohistochemical analysis indicated that OCTN1 was colocalized with PDZK1 on cell­ surface, whereas colocalization with PDZK1­E195K was partially observed in cytoplasmic region. These results suggest a novel hypothesis that mutation in PDZK1 potentially changes transport property of various types of xenobiotic transporters by aŠecting their subcellular localization, possibly leading to change in disposition of various types of substrate drugs.

Key words: transporters; OCT; organic cations; PDZ domain; protein interaction

cell­surface stability of, at least, certain types of Introduction xenobiotic transporters including organic cationW Recent molecular biological researches have speciˆed carnitine transporters OCTN1 and OCTN2, oligopep­ PDZ (PSD95, Dlg and ZO1) domain containing pro­ tide transporter PEPT2, urate­anion exchanger teins which may play a role as scaŠolds for transmem­ URAT1, organic anion transporter OAT4 and organic brane proteins such as transporters, thereby aŠecting anion transporting polypeptide (Oatp) 1a1.1–6) PDZ subcellular localization, transporting activity andWor domains generally consist of 80–90 amino acids, and can recognize internal andWor extreme C­terminal pep­ tide sequences (PDZ binding motif) usually composed This study was supported in part by a Grant­in­Aid for Scientiˆc of 3 to 8 amino acids.7) PDZ domain containing proteins Research provided by the Ministry of Education, Culture, Sports, are considered to play important roles in the targeting of Science and Technology of Japan, in part by Uehara Memorial Foundation Research Grant provided by Uehara Memorial Founda­ proteins to speciˆc cell membranes, as well as assem­ tion and in part by Research Grant provided by the Research Founda­ bling proteins for e‹cient signal transduction.8,9) Our tion for Pharmaceutical Sciences. recent screening using yeast two­hybrid revealed physi­

Received; April 13, 2006, Accepted; June 6, 2006 *To whom correspondence should be addressed: Prof. Akira TSUJI,Ph.D.,Division ofPharmaceutical Sciences, Graduate School ofNatural Science and Technology, Kanazawa University, Kakuma­machi, Kanazawa 920­1192, Japan. Tel. {81­76­234­4479, Fax. {81­76­264­6284, E­mail: tsuji—kenroku.kanazawa­u.ac.jp

375 376 Tomoko SUGIURA, et al. cal interaction between the PDZ proteins andseveral (3108 GbqWmmol), [14C]TEA (2.035 GbqWmmol) and xenobiotic transporters including OCTN, PEPT, OAT [3H]Glycylsarcosine (GlySar, 148 GbqWmmol) were andOATP family members, almost all of which have purchasedfrom GE Healthcare (Buckinghamshire, UK) class I PDZ domain interacting motif at their C­termi­ andMoravek Biochemicals Inc (Mercury Lane Brea, nus (­SWT­X­F, F is a hydrophobic amino acid).3) In­ CA). The c­myc epitope taggedmutant PDZK1 terestingly, most of such transporters interacting with (PDZK1­E195K) was subclonedin pcDNA3 (Invitro­ PDZ proteins are expressedon apical membranes of gen). cDNAs encoding full­length human OCTN1 and epithelial cells in the kidney and small intestine, im­ human PEPT2 were subclonedinto pcDNA3, whereas plying possible importance of the protein interaction in cDNA encoding full­length human OCTN2 was sub­ subcellular localization of xenobiotic transporters.3,10) clonedinto pEYFP­C1 vector (Invitrogen). Among the PDZ proteins interacting with the trans­ Transport studies in HEK293 cells stably expressing porters, regulatory function of PDZK1 has been best PDZK1: HEK293 cells were transfectedwith the characterized. PDZK1 was ˆrst identiˆed as a protein c­myc­taggedmutant PDZK1, andcells were selectedby that is up­regulatedin rats on a low­phosphate diet 11) or adding G418 (Invitrogen) to the culture medium to in several carcinoma cell lines.12) PDZK1 has four PDZ obtain HEK293WPDZK1­E195K cells. The parent cell domains in its structure. In in vitro studies, PDZK1 has line HEK293 andHEK293 cells stably expressing wild­ been demonstrated to increase substrate transporting type PDZK1 (HEK293WPDZK1 cells) were usedas a activityofURAT1,OCTN2andOAT4bydiŠerent control.4) The cDNA encoding OCTN1, OCTN2 and mechanisms, ie., increase in cell­surface expression of PEPT2 was then transiently transfectedaccordingto the URAT1 andOAT4, anddirectregulation without calcium phosphate precipitation method, and, at 48 hr obvious change in cell­surface expression of after transfection, cells were harvestedandsuspended OCTN2.2,4,5) Transport activity of PEPT2 was also in a transport medium (125 mM NaCl, 4.8 mM KCl, 3) slightly increasedin the presence of PDZK1. Wang 5.6 mM D­glucose, 1.2 mM CaCl2,1.2mMKH2 PO4, 6) et al. have recently clariˆeddown­regulation of 1.2 mM MgSO4, and25 mM HEPES, pH 7.4). Uptake Oatp1a1 in hepatocytes of pdzk1(|W|) mouse, demon­ experiment was then performedby the silicon­oil layer strating fundamental role of PDZK1 in cell­surface methodas describedpreviously. 4) Because we compared expression of the transporter. transport activity between three cell lines (HEK293, It might be noteworthy that PDZK1 can thus bindto HEK293WPDZK1 andHEK293 WPDZK1­E195K) in the various types of xenobiotic transporters andmay aŠect present study, we routinely observed ‰uorescence their transport properties, leading to a novel hypothesis intensity in each cell line by ‰uorescence microscope that genetic variation in PDZK1 may aŠect function (Axiovert S­100; Carl Zeiss, Jena, Germany) after andWor cell­surface expression of those transporters, transient transfection of YFP (as a cytosolic protein) or thereby aŠecting membrane transport of many types of a chimeric protein YFP­OCTN2 (as a membranous substrate drugs. Therefore, it appears important to protein) with an aim to check transfection e‹ciency. clarify whether the genetic variation of PDZK1 aŠects There was no clear diŠerence in the ‰uorescence among the property of interacting transporters. Understanding each cell line. of the potential eŠect of genetic variation in transport­ Immunocytochemical analysis: HEK293 cells were ers and their regulators on drug disposition may help grown on cover glass (15~15 mm; thickness, 0.12–0.17 development of personalized medicine. However, up to mm; micro cover glass, Matsunami Glass Inc., Osaka, now, no information has been available on how the Japan) andtransiently transfectedwith cDNA encoding mutation in PDZK1 aŠects transport properties of the each transporter as described above. At 48 hr after the interacting transporters. transfection, cells were ˆxedwith 3 z formaldehyde in In the present study we attempted to examine the phosphate­buŠeredsaline (PBS), permeabilizedwith eŠect of a mutation in PDZK1 on its regulatory eŠect on methanol for 5 min, andincubatedwith PBS containing several xenobiotic transporters including OCTN1, 3z blocking agent (Amersham Biosciences) for 30 min OCTN2 andPEPT2. at room temperature. Cells were then incubatedwith anti­human OCTN1 andanti­c­myc antibodies for 1 hr Materials and Methods at room temperature at a dilution of 1:20 and 1:100, Materials: Rabbit polyclonal antibody for OCTN1 respectively, washedwith PBS, andthen incubatedwith was raisedas describedpreviously. 13) Monoclonal Alexa Fluor 488 goat anti­rabbit IgG andAlexa Fluor { { antibodies against Na WK ATPase, His5,c­myc 594 goat anti­mouse IgG (Molecular Probes Inc., epitope tag (9E10) and g­adaptin were obtained from Eugene, OR) at a dilution of 1:200 in PBS containing Upstate Biotechnology (Lake Placid, NY), Invitrogen 3z Blocking Agent. Attachedcells were sealedonto the (San Diego, CA), Covance Inc (Princeton, NJ) and slides using Vectrashield mounting medium with 4?,6­ Sigma (St. Louise, MO), respectively. L­[3H]Carnitine dianidino­2­phenylindole (Vector Laboratories, Burlin­ Mutation in PDZK1 AŠects Transport Activity of Various Interacting Transporters 377 game, CA)and examined under a confocal laser scan­ PDZK1. According to crystal structure,14) PDZ domain ning ‰uorescence microscope (LSM 510; Carl Zeiss). is formed from six­stranded antiparallel b­barrels Western blot analysis: At 48 hr after the transfec­ capped by two a­helices, and E195K is presumed to be tion of OCTN1, the cells were washed twice with PBS, located in the second a­helix that could be important to harvested and solubilized in RIPA­Y buŠer containing interact with C­terminus of membrane proteins. Since 1z NONIDET P­40, 75 mM NaCl, 50 mM Tris­HCl the second PDZ domain in PDZK1 has been proposed (pH 7.5)and protease inhibitors. Western­blotting was to be involved in interaction with C­terminus of various then performed using anti­OCTN1, anti­c­myc and anti­ transporters including OCTN1, OCTN2 and PEPT2,3,4) g­adaptin antibodies as described previously.4) we then attempted to clarify whether such diŠerence in Yeast two­hybrid analysis: Yeast two­hybrid amino acid sequence aŠects regulatory eŠect of PDZK1 analysis was performed as described previously.3) on these three transporters in the present study. Brie‰y, yeast cells were co­transformed with Although Kocher et al.12,15) have reported interaction of pGBKT7(TRP)encoding GAL4bd fused to the C­termi­ PDZK1­E195K with the C­terminus of multidrug nus of each transporter, and pGADT7(LEU2)vector resistance associated protein 2 and another associated encoding GAL4ad fused to wild­type and mutant protein MAP17, no information is available on its PDZK1 construct. Co­transformed cells were further interaction with other transporters. In addition, up to cultured on plates lacking leucine and tryptophan, with now, there is no information on regulatory function and or without histidine. association with any diseases of PDZK1­E195K. Plate assays: Glutathione S­transferase (GST) To examine the eŠect of mutant PDZK1 on activity fusion proteins and His6­tagged PDZK1 were obtained of these transporters, we obtained HEK293WPDZK1­ from E. coli (BL­21 strain)transformed with pGEX6P­ E195K cells and performed uptake study in HEK293W 1 and pET30 constructs, respectively, according to the PDZK1, HEK293WPDZK1­E195K and HEK293 cells, manufacturer's instructions. The protein content of those were further transiently transfected with cDNA each His6­tagged and GST­fusion proteins was deter­ encoding each transporter. Time­dependent uptake of mined both by CBB­staining after SDS­PAGE and [14C]TEA, [3H]carnitine and [3H]Gly­Sar was observed Bradford method with a protein assay kit (Bio­Rad, in HEK293 cells transfected with OCTN1, OCTN2 and Hercules, CA), in both of which bovine serum albumin PEPT2, respectively (Fig. 1A, C and E),whereassuch were used as a standard. His­Select HS Nickel­Coated uptake was relatively small in HEK293 cells transfected 96­well plate (Sigma)was ˆrst coated with 0.02 mMof with vector alone (Fig. 1B, D and F).Allofthese

His6­taggedPDZK1dissolvedinPBSwith1z BSA for uptakes were higher in HEK293WPDZK1 cells, com­ 4hrat49C. After washing the plate twice with PBS paredwiththoseinHEK293cells(Fig. 1A, C and E). containing 0.1z Tween20, GST C­terminus fusion Theuptakeof[14C]TEA, [3H]carnitine and [3H]Gly­Sar proteins dissolved in PBS with 0.05z Triton X­100 was by OCTN1, OCTN2 and PEPT2, respectively, in added and incubated at 49C for 2 hr. After three HEK293WPDZK1­E195K cells, was higher than that in washings, the plate was incubated at room temperature HEK293 cells. The uptake by OCTN2 in HEK293W for 1 hr with HRP conjugated anti­GST antibody PDZK1­E195K cells was lower than that in HEK293W (Amersham Biosciencces), followed by three­times PDZK1 cells (Fig. 1C). On the other hand, the diŠer­ washings. Reaction mixture (TMB peroxydase substrate ence in the uptake by OCTN1 and PEPT2 between kit, Bio­Rad)was added, and absorption at 655 nm was HEK293WPDZK1 and HEK293WPDZK1­E195K cells then monitored. was not so obvious in the initial phase, but more obvious at longer time periods (Fig. 1A and E). Student Results t­test was performed for the uptake by each transporter EŠect of PDZK1­E195K on transport property of between HEK293WPDZK1 and HEK293 cells, and that OCTN1, OCTN2 and PEPT2: PDZK1 with NCBI between HEK293WPDZK1­E195 and HEK293 cells. All reference sequence (accession# NMä002614)was used as the uptake values for OCTN1, OCTN2 and PEPT2 a wild­type in the present study. This PDZK1 has the were signiˆcantly higher in HEK293WPDZK1 or sameaminoacidsequencewiththatinotherPDZK1 HEK293WPDZK1­E195K cells than those in HEK293 clones (accession# BC006518 and BC006496)reported cells (pº0.05). in NCBI database. On the other hand, PDZK1 (acces­ To examine speciˆcity of the stimulatory eŠect of sion# AF012281)originally isolated by Kocher et al.12) PDZK1 and PDZK1­E195K, each cell line was tran­ has a diŠerence in the sequence (PDZK1­681GÀA, siently transfected with OCTN1D4, OCTN2D4and encoding E195K). Because this PDZK1­E195K clone PEPT2D4, these lacking a PDZ binding motif consist­ was isolated from normal human kidney cDNA libra­ ing of 4 amino acids at their extreme C­terminus, since ry,12) this clone should also be present in human ge­ interaction potential of these constructs with PDZK1 nomes. E195K is located in the second PDZ domain of was almost abolished.3,4) The stimulatory eŠect of 378 Tomoko SUGIURA, et al.

Fig. 2. EŠect of PDZK1 and its mutant PDZK1­E195K on concen­ tration dependence of [14C]TEA uptake by OCTN1. HEK293WPDZK1 (å), HEK293WPDZK1­E195K ($)andHEK293 (ç) cells were transiently transfected with OCTN1 or vector alone, and uptake of [14C]TEA was then measured for 5 min at 379C. OCTN1­mediated uptake was then calculated by subtraction of the background uptake obtained in the vector­transfected cells from the uptake obtained in OCTN1­transfected cells. The results are shown as Eadie­Hofstee plot, and the data represent mean}S.E.M. of three de­ terminations. When error bars are not shown, they are smaller than the symbols.

398}54, 441}67 and 401}95 mMinHEK293, HEK293WPDZK1 and HEK293WPDZK1­E195K cells, Fig. 1. EŠect of PDZK1 and its mutant PDZK1­E195K on transport respectively (Fig. 2). Thus, the diŠerence in K and V activity of OCTN1, OCTN2and PEPT2. m max values for initial uptake of TEA by OCTN1 was not so HEK293WPDZK1 (å), HEK293WPDZK1­E195K ($)andHEK293 (ç) cells were transiently transfected with OCTN1 (A), OCTN1D4 obvious between HEK293WPDZK1 and HEK293W (B), YFP­OCTN2(C), YFP­OCTN2 D4(D),PEPT2(E)and PDZK1­E195K cells. PEPT2D4 (F), and uptake of [14C]TEA (A, B), L­[3H]carnitine (C, D) To further characterize the stimulatory eŠect of 3 and [ H]GlySar (E, F) was measured. Control experiment was also PDZK1 and PDZK1­E195K, we next compared performed in HEK293WPDZK1 (ä), HEK293WPDZK1­E195K (#) and HEK293 (æ) cells transfected with vector alone, and results were inhibitory eŠect of pyrilamine and verapamil, both 16) shown in panels B, D and F. The data represent mean}S.E.M. of being substrates of OCTN1, on OCTN1­mediate three determinations. uptake of TEA between HEK293, HEK293 WPDZK1 and HEK293WPDZK1­E195K cells (Fig. 3). Both compounds inhibited the TEA uptake in a concentra­ PDZK1 and PDZK1­E195K was not so obvious on tion dependent manner, showing almost comparable uptake of each substrate by OCTN1D4, OCTN2D4and inhibition proˆle between each cell line (Fig. 3), suggest­ PEPT2D4(Fig. 1B, D and F). This result, especially ing that substrate recognition speciˆcity of OCTN1 was showing minimal diŠerence in transport activity of minimally aŠected by PDZK1 or PDZK1­E195K. OCTN2D4 between HEK293, HEK293WPDZK1 and EŠect of PDZK1 and PDZK1­E195K on subcellular HEK293WPDZK1­E195K cells, supported minimal localization and expression of OCTN1: To investigate diŠerence in transfection e‹ciency between each cell possible reasons for the present ˆnding that stimulatory line. eŠect on transport activity of OCTN1 of PDZK1­E195K To characterize the stimulatory eŠect of PDZK1 and was weaker than that of wild­type PDZK1 (Figs.1,2), PDZK1­E195K, then kinetic analysis of TEA uptake by subcellular localization of OCTN1 and its possible OCTN1 was performed. Eadie­Hofstee plots revealed colocalization with PDZK1 were further examined by single component in OCTN1­mediated uptake in immunohistochemical analysis (Fig. 4). In HEK293W HEK293, HEK293 WPDZK1 and HEK293WPDZK1­ PDZK1 cells, OCTN1 was colocalized with PDZK1, E195K cells, showing similar slope (Fig. 2). The and such colocalization was observed on cell­surface obtained values for Vmax and Km were 5.05}0.76, 7.10 (Fig. 4A, B and C). In HEK293 WPDZK1­E195K cells, }1.19 and 5.63}1.43 nmolWmg proteinW5min, and on the other hand, OCTN1 was colocalized with the Mutation in PDZK1 AŠects Transport Activity of Various Interacting Transporters 379

Fig. 3. Inhibitory eŠect of pyrilamine (A) and verapamil (B) on OCTN1­mediated uptake of [14C]TEA in HEK293WPDZK1, HEK293WPDZK1­ E195K and HEK293 cells. HEK293WPDZK1 (å), HEK293WPDZK1­E195K ($) and HEK293 (ç) cells were transiently transfected with OCTN1 or vector alone, and up­ take of [14C]TEA for 5 min was measured in the presence or absence of each compound. OCTN1­mediated uptake was then calculated by subtrac­ tion of the background uptake obtained in the vector­transfected cells from the uptake obtained in OCTN1­transfected cells. The data represent mean}S.E.M. of six or nine determinations.

Fig. 4. Subcellular localization of OCTN1 and PDZK1. HEK293WPDZK1 (A, B, C) and HEK293WPDZK1­E195K (D, E, F) cells were transiently transfected with OCTN1, and subcellular localization of OCTN1 and PDZK1 was then immunocytochemically analyzed with an anti­human OCTN1 (A, D) and anti­c­myc (B, E) antibodies. Overlay image (C, F) showed that both OCTN1 and PDZK1 are co­localized, but such colocalization was observed in diŠerent subcellular compartments between PDZK1 and PDZK1­E195K. In panels G, H and I, OCTN1 (green) and Na{ WK{ ATPase (red) were double stained in HEK293WPDZK1, HEK293WPDZK1­E195K and HEK293 cells, respectively, after transient transfection with OCTN1. Arrowheads indicated cytoplasmic localization of OCTN1.

PDZK1 mutant, but both proteins were partially change in localization of OCTN1. To further character­ localized in cytoplasmic regions (Fig. 4D, E and F). ize the localization of OCTN1, colocalization with Thus, mutation in PDZK1 may aŠect subcellular Na{ WK{ ATPase, a plasma membrane marker, was localization of PDZK1 itself, thereby causing the then examined. In HEK293WPDZK1 cells, OCTN1 was 380 Tomoko SUGIURA, et al.

Fig. 5. Expression of OCTN1 and PDZK1. HEK293WPDZK1 and HEK293WPDZK1­E195K cells were transient­ ly transfected with OCTN1, followed by Western blot analysis for cel­ lular lysates. Band intensity assessed by a densitometric analysis was Fig. 6. Physical interaction of PDZK1 and PDZK1­E195K with C­ normalized to that in HEK293 WPDZK1 cells, and shown. terminus of each transporter. In panel A, yeast cells were cotransformed with plasmids encoding PDZK1 or its mutants (subcloned as a fusion protein with GAL4AD into pGADT7 vector) and the C­terminus of each transporter (sub­ expressed on plasma membranes, colocalizing with clonedasafusionproteinwithGAL4BDinpGBKT7vector).P53was Na{ WK{ ATPase (Fig. 4G), whereas, in HEK293W used in the control experiment. Interaction was indicated by growth PDZK1­E195K cells, a partof OCTN1 was notcolocal­ on agar plates made with medium without histidine (­His). In panel B, nickel­coated 96­well plate coated with His ­tagged ized with Na{ WK{ ATPase, and again found in 6 PDZK1 (black bars), PDZK1­E195K (grey bars) and PDZK1­H193A cytoplasmic region (Fig. 4H). Such intracellular locali­ (white bars) was incubated with GST­C­terminus fusion proteins for 2 zation of OCTN1 was also observed in HEK293 cells hr. After washings, the plate was further incubated with HRP conju­ (Fig. 4I). gated anti­GST antibody, followed by the detection of the antibody To compare expression level of PDZK1, Western blot with TMB peroxydase substrate kit. Data were normalized by the absorbance obtained for His ­tagged PDZK1. analysis was then performed for cellular lysates. Expres­ 6 sion of OCTN1 was much higher in HEK293WPDZK1 cells than HEK293WPDZK1­E195 cells, whereas almost comparable expression of PDZK1 was observed Another PDZK1 mutant PDZK1­H193A was also between each cell line (Fig. 5). Expression of an en­ constructed by replacement of a histidine residue that is dogenous protein g­adaptin was also examined in a located in the second a­helix of the PDZ domain and control experiment, showing comparable level (Fig. 5). involved in hydrogen bond with hydroxy group of serine Physical interaction between PDZK1­E195K and or threonine residue in PDZ domain interacting motif, C­terminus of the transporters: To assess the eŠect of with alanine.14) The interaction of PDZK1­H193A with E195K mutation in PDZK1 on its interaction potential the C­terminus of OCTN1 was slightly lower in both with the C­terminus of each transporter, we examined experiments than that of wild­type PDZK1, but there their direct interaction by two methods, yeast two­ was no obvious diŠerence from wild­type PDZK1 in hybrid analysis (Fig. 6A) and plate assay (Fig. 6B). interaction potential of PDZK1­H193A (Fig. 6). Obvious diŠerence in interaction potential with the Discussion C­terminus of OCTN1, OCTN2 and PEPT2 was not observed between PDZK1 and PDZK1­E195K, OCTN1, OCTN2 and PEPT2 are localized on apical although the interaction of PDZK1­E195K with the membranes of renal proximal tubules. Both OCTN1 C­terminus of OCTN1 was slightly lower in both and OCTN2 acceptcarnitineand various organic experiments than that of wild­type PDZK1 (Fig. 6). cations as substrates, whereas PEPT2 transports di­ and Mutation in PDZK1 AŠects Transport Activity of Various Interacting Transporters 381 tripeptidesaswellasvarioustypesoftherapeuticagents was then analyzed (Fig. 4). Immunocytochemical including b­lactam antibiotics.16–21) Genetic deˆciency in analysis revealed that OCTN1 was colocalized with octn2 gene hinders renal reabsorption of carnitine and PDZK1 and PDZK1­E195K, although OCTN1 and secretion of tetraethylammonium (TEA) in mice, sug­ PDZK1­E195K were partially localized in intracellular gesting that OCTN2 plays an important role in the renal compartments (Fig. 4). Thus, mutation of PDZK1 may handling of its substrates.22,23) Essential role of renal aŠect subcellular localization of PDZK1 itself, thereby reabsorption of peptidic compounds has recently been causing the change in localization and activity of demonstrated in pept2(| W|) mice.24) Although the phar­ transporters that interact with PDZK1. This may macokinetic roles of OCTN1 has not yet been demon­ explain the present ˆnding that the E195K mutation strated in vivo, recent studies revealed the association of aŠectsregulatoryactivityofPDZK1forvarioustrans­ the risk for Crohn's disease or rheumatoid arthritis with porters (Fig. 1). The partial localization of OCTN1 in OCTN1 single polymorphism nucleotide (SNPs).25,26) intracellular compartments of HEK293WPDZK1­E195K Thus, these three transporters could play a major role in was compatible withthelower OCTN1­mediated TEA drug disposition or pathophysiological status, leading uptake especially at steady­state in HEK293W to necessity of the present studies regarding potential PDZK1­E195K, compared withHEK293 WPDZK1 cells eŠect of the mutation in their adaptor protein PDZK1 (Fig. 1A). In addition, expression level of OCTN1 in on activity of eachtransporter. In thepresent study, we HEK293WPDZK1­E195K was also lower than that in demonstrated that stimulatory eŠect of PDZK1 on HEK293WPDZK1 cells (Fig. 5), and this can also uptake of substrates by OCTN1, OCTN2 and PEPT2 explain the diŠerence in TEA transport (Fig. 1A). especially at longer time periods was decreased by Although the exact molecular mechanism(s) for the mutational change in PDZK1, whereas such eŠect was decreased expression of OCTN1 was currently not so obvious on the uptake by OCTN1D4, OCTN2D4 unknown, one of the hypotheses may include that such and PEPT2D4(Fig. 1), suggesting that mutational intracellular OCTN1 could be unstable, compared with change in PDZK1 may aŠect transport properties of that on cell­surface, presumably due to the intracellular various transporters. Considering the wide spectrum in degradation. It was reported that PDZK1 stabilizes substrate recognition speciˆcity of these transporters, URAT1 expression on cell­surface,2) and similar stabili­ suchmutation in PDZK1 may changemembrane zation by PDZK1 may occur for OCTN1. However, transport of numerous substrate drugs of the transport­ expression of OCTN1 was assessed in whole cell lysates ers. Although the present studies were restricted only to in the present study (Fig. 5). Further studies are needed the in vitro ones, the present ˆndings have proposed a to quantitatively determine OCTN1 level on cell­surface novel hypothesis that genetic variation in such adaptor of eachcell line. proteins may change membrane transport of various To date, it has been demonstrated that deˆciency of substrates of various transporters that directly interact pdzk1 gene leads to decrease in the expression level of withtheadaptors. scavenger receptor BI27) and translocation of Oatp1a1 to Our previous studies have suggested that PDZK1 is a intracellular compartments6) in liver. Therefore, PDZK1 functional regulator of OCTN2 and PEPT2,3,4) whereas could play critical roles in expression or subcellular eŠect of PDZK1 on OCTN1 has not yet been reported. localization of several membrane proteins including In the present study, we ˆrst identiˆed similar stimulato­ OCTN1. On the other hand, deˆciency of pdzk1 gene in ry eŠect of PDZK1 on OCTN1­mediated transport mice causes hypercholesterolemia, due to the loss in (Fig. 1A). The stimulatory eŠect of PDZK1­E195K on expression of the scavenger receptor.27) Mutaions in substrate uptake at longer time periods by OCTN1, OCTN2 are associated withprimary systemic carnitine OCTN2 and PEPT2 was weaker than that of wild­type deˆciency,23) whereas more recent studies have suggest­ PDZK1 (Fig. 1A, C and E). Thus, eŠect of the E195K­ ed the association of the risk for Crohn disease25) or mutation on the stimulatory activity of PDZK1 was rheumatoid arthritis26) withOCTN1 SNPs, whichmeans observed for various transporters. However, it should variation of transporters may aŠect not only phar­ be noted that eŠect of the mutation in PDZK1 on initial macokinetics but also several diseases. Considering phase (¿5 min) of TEA uptake by OCTN1 was not so these ˆndings, several diseases might be attributed to remarkable (Fig. 1A, Fig. 2), whereas that on the mutations andWor SNPs in PDZK1, that change the uptake at 30 min was more obvious (Fig. 1A). Thus, the expression, localizatoin andWor activity of transporters change by the PDZK1 mutation is more evident at that interact with PDZK1. steady­state than initial phase, implying that there may According to sequence comparison of class I PDZ be diŠerent molecular mechanisms in stimulatory eŠect domains,14,28) the position of E195K is located in a2 helix of PDZK1 on OCTN1­mediated uptake. within second PDZ domain in PDZK1 that is involved To understand how the mutation in PDZK1 aŠects its in binding to C­terminus of several membrane proteins. regulatory activity, subcellular localization of OCTN1 The E195K­mutation causes electric change and then 382 Tomoko SUGIURA, et al. may vary structure of the hydrophobic pocket in the 3) Kato, Y., Yoshida, K., Watanabe, C., Sai, Y. and Tsuji, PDZ domain, altering its interaction potential to other A.: Screening of the interaction between xenobiotic proteins. However, it is noted that PDZK1­E195K transporters and PDZ proteins. Pharm. Res., 21: exhibits almost comparable interaction potential with 1886–1894 (2004). wild­type PDKZ1 for the C­terminus of each transport­ 4) Kato, Y., Sai, Y., Yoshida, K., Watanabe, C., Hirata, er (Fig. 6). This may explain that activity of each trans­ T. and Tsuji A.: PDZK1 directly regulates the function of organic cationWcarnitine transporter OCTN2. Mol. porter in HEK293 PDZK1­E195K cells was higher, as W Pharmacol., 67: 734–743 (2005). in case of HEK293WPDZK1cells,thanthatinHEK293 5) Miyazaki, H., Anzai, N., Ekaratanawong, S., Sakata, cells. PDZ proteins may interact not only with trans­ T., Shin, H. J., Jutabha, P., Hirata, T., He, X., porters, but also with other PDZ proteins and form a Nonoguchi, H., Tomita, K. and Kanai, Y.: Endou H. large protein complex consisting of several proteins.1,8,9) Modulation of renal apical organic anion transporter 4 Therefore, the E195K mutation may aŠect the forma­ function by two PDZ domain­containing proteins. J. tion of such large protein complexes, and this cannot be Am.Soc.Nephrol., 16: 3498–3506 (2005). detected in experimental systems that investigate 6)Wang,P.,Wang,J.J.,Xiao,Y.,Murray,J.W., interaction between just two proteins. It should also be NovikoŠ, P. M., Angeletti, R. H., Orr, G. A., Lan, D., noted that interaction studies shown in Fig. 6 were Silver,D.L.andWolkoŠ,A.W.:Interactionwith performed just for the C­terminal domain of each trans­ PDZK1 is required for expression of organic anion trans­ porting protein 1A1 on the hepatocyte surface. J. Biol. porter. Therefore, such interaction may not precisely Chem., 280: 30143–30149 (2005). represent the one of full­length transporters. Actually, 7) Skelton, N. J., Koehler, M. F., Zobel, K., Wong, W. L., stimulatory eŠect of PDZK1 on OCTN1­mediated TEA Yeh, S., Pisabarro, M. T., Yin, J. P., Lasky, L. A. and uptake was still observed, when the last four amino Sidhu, S. S.: Origins of PDZ domain ligand speciˆcity. acidsofOCTN1weredeleted(Fig. 1B), implying that Structure determination and mutagenesis of the Erbin functional modulation by PDZK1 may occur via other PDZ domain. J. Biol. Chem., 278: 7645–7654 (2003). domains than the C­terminal region of OCTN1. This 8) Biber, J., Gisler, S. M., Hernando, N. and Murer, H.: phenomenon was in contrast to that observed for ProteinWprotein interactions (PDZ) in proximal tubules. OCTN2, stimulatory eŠect of PDZK1 being minimal J. Membr. Biol., 203: 111–118 (2005). for OCTN2D4(Fig. 1D). Therefore, in physiological 9) Brone, B. and Eggermont, J.: PDZ proteins retain and conditions, E195K mutation may cause more obvious regulate membrane transporters in polarized epithelial cell membranes. Am.J.Physiol., 288: C20–C29 (2005). reduction in interaction potential between OCTN1 and 10) Russel, F. G., Masereeuw, R. and van Aubel, R. A.: PDZK1. Molecular aspects of renal anionic drug transport. In conclusion, the present research suggested that Annu. Rev. Physiol., 64: 563–594 (2002). variation of PDZK1 may aŠect subcellular localization 11) Custer, M., Spindler, B., Verrey, F., Murer, H. and and activity of several transporters, which interact with Biber, J.: Identiˆcation of a new gene product (diphor­1) PDZK1, altering membrane transport of many types of regulated by dietary phosphate. Am.J.Physiol., 273: substrate compounds. Further investigations in this F801–F806 (1997). adaptor protein are needed to understanding of its 12) Kocher, O., Comella, N., Tognazzi, K. and Brown, L. pharmacological roles. F.: Identiˆcation and partial characterization of PDZK1: a novel protein containing PDZ interaction Acknowledgement: We thank Ms Lica Ishida for domains. Lab. Invest., 78: 117–125 (1998). 13) Tamai, I., Yabuuchi, H., Nezu, J., Sai, Y., Oku, A., technical assistance. We also acknowledge Mr. Kei­ichi Shimane, M. and Tsuji, A.: Cloning and characteriza­ Sakamoto and Mr. Yasuhito Nakadera for their help in tion of a novel human pH­dependent organic cation in vitro studies. transporter, OCTN1. FEBS Lett., 419: 107–111 (1997). References 14) Karthikeyan, S., Leung, T. and Ladias, J. A.: Structural basisoftheNa{ WH{ exchanger regulatory factor PDZ1 1) Gisler, S. M., Pribanic, S., Bacic, D., Forrer, P., interaction with the carboxyl­terminal region of the Gantenbein, A., Sabourin, L. A., Tsuji, A., Zhao, Z., cystic ˆbrosis transmembrane conductance regulator. J. Manser, E., Biber, J. and Murer, H.: PDZK1. I. A Biol. Chem., 276: 19683–19686 (2001). major scaŠolder in brush borders of proximal tubular 15) Kocher, O., Comella, N., Gilchrist, A., Pal, R., cells. Kidney Int., 64: 1733–1745 (2003). Tognazzi,K.,Brown,L.F.andKnoll,J.H.:PDZK1,a 2) Anzai, N., Miyazaki, H., Noshiro, R., Khamdang, S., novel PDZ domain­containing protein up­regulated in Chairoungdua, A., Shin, H. J., Enomoto, A., carcinomas and mapped to chromosome 1q21, interacts Sakamoto, S., Hirata, T., Tomita, K., Kanai, Y. and with cMOAT (MRP2), the multidrug resistance­associ­ Endou, H.: The multivalent PDZ domain­containing ated protein. Lab. Invest., 79: 1161–1170 (1999). protein PDZK1 regulates transport activity of renal 16) Yabuuchi, H., Tamai, I., Nezu, J., Sakamoto, K., Oku, urate­anion exchanger URAT1 via its C terminus. J. A., Shimane, M., Sai, Y. and Tsuji, A.: Novel mem­ Biol. Chem., 279: 45942–45950 (2004). brane transporter OCTN1 mediates multispeciˆc, bi­ Mutation in PDZK1 AŠects Transport Activity of Various Interacting Transporters 383

directional, and pH­dependent transport of organic deˆciency is caused by mutations in a gene encoding cations. J. Pharmacol. Exp. Ther., 289: 768–773 (1999). sodium ion­dependent carnitine transporter. Nat. 17) Tamai, I., Ohashi, R., Nezu, J., Yabuuchi, H., Oku, A., Genet., 21: 91–94 (1999). Shimane, M., Sai, Y. and Tsuji, A.: Molecular and func­ 24) Rubio­Aliaga,I.,Frey,I.,Boll,M.,Groneberg,D.A., tional identiˆcation of sodium ion­dependent, high a‹n­ Eichinger, H. M., Balling, R. and Daniel, H.: Targeted ity human carnitine transporter OCTN2. J. Biol. Chem., disruption of the peptide transporter Pept2 gene in mice 273: 20378–20382 (1998). deˆnes its physiological role in the kidney. Mol. Cell. 18) Shen, H., Smith, D. E., Yang, T., Huang, Y. G., Biol., 23: 3247–3252 (2003). Schnermann, J. B. and Brosius, F. C. 3rd.: Localization 25) Peltekova,V.D.,Wintle,R.F.,Rubin,L.A.,Amos,C. of PEPT1 and PEPT2 proton­coupled oligopeptide I.,Huang,Q.,Gu,X.,Newman,B.,VanOene,M., transporter mRNA and protein in rat kidney. Am. J. Cescon, D., Greenberg, G., Gri‹ths, A. M., St George­ Physiol., 276: F658–F665 (1999). Hyslop, P. H. and Siminovitch, K. A.: Functional vari­ 19) Tamai, I., Nakanishi, T., Kobayashi, D., China, K., ants of OCTN cation transporter genes are associated Kosugi, Y., Nezu, J., Sai, Y. and Tsuji, A.: Involvement with Crohn disease. Nat. Genet., 36: 471–475 (2004). of OCTN1 (SLC22A4) in pH­dependent transport of or­ 26) Tokuhiro, S., Yamada, R., Chang, X., Suzuki, A., ganic cations. Mol. Pharm., 1: 57–66 (2004). Kochi, Y., Sawada, T., Suzuki, M., Nagasaki, M., 20) Daniel, H. and Rubio­Aliaga, I.: An update on renal Ohtsuki, M., Ono, M., Furukawa, H., Nagashima, M., peptide transporters. Am.J.Physiol., 284: F885–F892 Yoshino, S., Mabuchi, A., Sekine, A., Saito, S., (2003). Takahashi,A.,Tsunoda,T.,Nakamura,Y.and 21) Tamai,I.,Ohashi,R.,Nezu,J.I.,Sai,Y.,Kobayashi, Yamamoto, K.: An intronic SNP in a RUNX1 binding D.,Oku,A.,Shimane,M.andTsuji,A.:Molecularand site of SLC22A4, encoding an organic cation transport­ functional characterization of organic cationWcarnitine er, is associated with rheumatoid arthritis. Nat. Genet., transporter family in mice. J. Biol. Chem., 275: 35: 341–348 (2003). 40064–40072 (2000). 27) Kocher, O., Yesilaltay, A., Cirovic, C., Pal, R., Rigotti, 22)Ohashi,R.,Tamai,I.,Nezu,J.,Nikaido,H., A. and Krieger, M.: Targeted disruption of the PDZK1 Hashimoto, N., Oku, A., Sai, Y., Shimane, M. and gene in mice causes tissue­speciˆc depletion of the high Tsuji, A.: Molecular and physiological evidence for density lipoprotein receptor scavenger receptor class B multifunctionality of carnitineWorganic cation transport­ type I and altered lipoprotein metabolism. J. Biol. er OCTN2. Mol. Pharmacol., 59: 358–366 (2001). Chem., 278: 52820–52825 (2003). 23) Nezu, J., Tamai, I., Oku, A., Ohashi, R., Yabuuchi, H., 28) Scott, R. O., Thelin, W. R. and Milgram, S. L.: A novel Hashimoto, N., Nikaido, H., Sai, Y., Koizumi, A., PDZ protein regulates the activity of guanylyl cyclase C, Shoji, Y., Takada, G., Matsuishi, T., Yoshino, M., the heat­stable enterotoxin receptor. J. Biol. Chem., Kato, H., Ohura, T., Tsujimoto, G., Hayakawa, J., 277: 22934–22941 (2002). Shimane, M. and Tsuji, A.: Primary systemic carnitine