Logo of biochemjBJ Latest papers and much more!
Biochem J. Nov 1, 2007; 407(Pt 3): 363–372.
Published online Oct 12, 2007. Prepublished online Jul 24, 2007. doi:  10.1042/BJ20070716
PMCID: PMC2275063

Heterodimerization, trafficking and membrane topology of the two proteins, Ostα and Ostβ, that constitute the organic solute and steroid transporter


Co-immunoprecipitation studies using mouse ileal proteins and transfected HEK-293 (human embryonic kidney-293) cells revealed that the two proteins, Ostα and Ostβ, which generate the organic-solute transporter are able to immunoprecipitate each other, indicating a heteromeric complex. Mouse ileal Ostα protein appeared on Western blots largely as bands of 40 and 80 kDa, the latter band consistent with an Ostα homodimer, and both of these bands were sensitive to digestion by the glycosidase PNGase F (peptide:N-glycosidase F). Ostβ appeared as bands of 17 and 19 kDa, and these bands were not sensitive to PNGase F. Both the 40 and 80 kDa forms of Ostα, and only the 19 kDa form of Ostβ, were detected among the immunoprecipitated proteins, indicating that the interaction between Ostα and Ostβ is associated with specific post-translational processing. Additional evidence for homodimerization of Ostα and for a direct interaction between Ostα and Ostβ was provided by BiFC (bimolecular fluorescence complementation) analysis of HEK-293 cells transfected with Ostα and Ostβ tagged with yellow-fluorescent-protein fragments. BiFC analysis and surface immunolabelling of transfected HEK-293 cells also indicated that the C-termini of both Ostα and Ostβ are facing the intracellular space. The interaction between Ostα and Ostβ was required not only for delivery of the proteins to the plasma membrane, but it increased their stability, as noted in transfected HEK-293 cells and in tissues from Ostα-deficient (Ostα−/−) mice. In Ostα−/− mice, Ostβ mRNA levels were maintained, yet Ostβ protein was not detectable, indicating that Ostβ protein is not stable in the absence of Ostα. Overall, these findings identify the membrane topology of Ostα and Ostβ, demonstrate that these proteins are present as heterodimers and/or heteromultimers, and indicate that the interaction between Ostα and Ostβ increases the stability of the proteins and is required for delivery of the heteromeric complex to the plasma membrane.

Keywords: bimolecular fluorescence complementation (BiFC), co-immunoprecipitation (co-IP), heterodimer, membrane topology, organic-solute transporter (Ost), Ostα−/− mouse
Abbreviations: BiFC, bimolecular fluorescence complementation; Endo H, endoglycosidase H; ES, embryonic stem; FXR, farnesoid X receptor; GPCR, G-protein-coupled receptor; IP, immunoprecipitation; MRP1/Mrp1, multidrug-resistance-associated protein 1; Neo, neomycin; Ost, organic-solute transporter; PNGase F, peptide:N-glycosidase F; RAMP, receptor activity-modifying protein; TM, transmembrane; YFP, yellow fluorescent protein


Regulation of bile acids and other steroid-derived molecules is important for cellular homoeostasis. The recently discovered Ostα–Ostβ, a heteromeric organic solute and steroid transporter, appears to be a major basolateral transporter of bile acids and conjugated steroids in a variety of tissues, including kidney, intestine, adrenal gland and liver [16]. This transporter comprises a predicted seven-transmembrane (TM)-domain protein, Ostα, and a single TM-domain ancillary polypeptide, Ostβ, but the roles of the two proteins in generating the functional transporter is unknown. Ostα and Ostβ are encoded by genes on different chromosomes, they exhibit no sequence identity with each other and appear to lack paralogues in the mouse or human genomes [2,5]. Ostα–Ostβ mediates transport of bile acid and conjugated steroids, including taurocholate, oestrone 3-sulfate and dihydroepinadrosterone sulfate by a facilitated diffusion mechanism [4] and thus can mediate either efflux or uptake, depending on the electrochemical gradient of a given substrate. Expression of both Ost genes is positively regulated by bile acids through the bile-acid-activated FXR (farnesoid X receptor) [610].

Although the mechanism by which OSTα and OSTβ interact is unknown, some insight has been provided by the studies of Dawson et al. [3] and the very recent studies of Sun et al. [11]. Dawson and co-workers [3] demonstrated that, when mouse Ostα or Ostβ are expressed individually, they display largely an endoplasmic-reticulum-staining pattern, with little detectable plasma-membrane staining. In contrast, when they are co-expressed, they are found largely at the plasma membrane. Dawson et al. [3] also demonstrated that co-expression of Ostα and Ostβ is required to convert the Ostα subunit into a mature N-glycosylated, Endo H (endoglycosidase H)-resistant form, indicating that co-expression facilitates the movement of Ostα through the Golgi apparatus. Thus stable association of both subunits may be required for transporter function, or the Ostβ subunit may function as a chaperone to promote the egress of Ostα from the endoplasmic reticulum. This question was also recently examined by Sun et al. [11], who also showed that membrane targeting of the human OST proteins requires their co-expression, but the mechanism was not identified.

Additional insight into how OSTα and OSTβ may interact has been provided by a comparison of the predicted membrane topologies of OSTα and OSTβ with those of other heteromeric membrane proteins. In particular, the overall predicted membrane architecture of the OSTα–OSTβ transporter is similar to that of GPCRs (G-protein-coupled receptors) that interact with the single-TM domain RAMPs (receptor activity-modifying proteins)-1, -2, or -3 [1214]. A growing body of evidence supports the existence of GPCR–GPCR homodimers and heterodimers, and of GPCR–RAMP heterodimers and hetero-oligomers [15,16]. The similarity in membrane topology between OSTα and the GPCRs, and between OSTβ and the RAMP proteins, raises the possibility that OSTα may also be found in cells as a homodimer (i.e., OSTα–OSTα), and as hetero-oligomers with OSTβ (i.e., OSTα–OSTβ or OSTα–OSTα–OSTβ).

The goal of the present study was to test this hypothesis and to probe for possible interactions between mouse Ostα and Ostβ using IP (immunoprecipitation), BiFC (bimolecular fluorescence complementation) and by analysis of the fate of Ostβ in Ostα-deficient mice.



HEK-293 cells were obtained from the A.T.T.C. (CRL-1573) and grown as monolayers at 37 °C under an atmosphere of 5% (v/v) CO2. Cells were maintained in medium consisting of Dulbecco's modified Eagle's medium (GIBCO; 10-013-CV) with 10% (v/v) fetal-calf serum, and antibiotics. Antibodies to mouse Ostα (mA315) and Ostβ (mB91) have been described previously [3,4]. The anti-FLAG M2 monoclonal (F 3165) and anti-c-Myc antibodies clone 9E10 (M 5546) were purchased from Sigma–Aldrich, the anti-MRP1 [anti-(multidrug-resistance-associated protein 1)] antibody (ALX-210-841, which is known to cross-react with mouse Mrp1) was from Alexis Biochemicals, and the anti-β-Actin (JLA20) antibody was purchased from Calbiochem. PNGase F (peptide:N-glycosidase F; P0704L) was purchased from New England Biolabs. [3H(G)]Taurocholic acid (1 Ci/mmol) was purchased from NEN. pDsRed2-ER (6982-1) plasmid was purchased from Clontech. All other chemicals and reagents were purchased from Ambion, Amersham Biosciences, Bio-Rad, Fermentas, Integrated DNA Technologies, Invitrogen, Jackson ImmunoResearch Laboratoties, J.T. Baker Inc., New England Biolab, PerkinElmer, Qiagen Roche, Sigma–Aldrich or Stratagene. The Ostα–IRES–Ostβ–c-Myc construct was kindly provided by Dr Paul Dawson (Wake Forest University School of Medicine, Winston-Salem, NC, U.S.A.). C57Bl/6 mice were purchased from Charles River Laboratories (Kingston, NY, U.S.A.) and were kept under a 12h-light/12 h-dark cycle at room temperature (20–22 °C). All experimental protocols were approved by the local Animal Care and Use Committee, according to criteria outlined in [16a].

Generation of Ostα−/− mice

A targeting vector was designed to replace exons 3–9 of Ostα (cDNA nucleotides 490–1545 or a deletion of ~4.3 kb) with a Neo (neomycin)-containing cassette (Caliper Life Sciences/Xenogen, Cranbury, NJ, U.S.A.). The 5′ homologous arm (2.4 kb) was generated by PCR using BAC (bacterial artificial chromosome) clone rp23-123l9 and proofreading LA Taq DNA polymerase (TaKaRa) and cloned into the pCRXL-TOPO vector. The 3′ homologous arm (7.5 kb) was generated by RED cloning/gap repair and cloned into the FtNwCD vector. Both plasmids were confirmed by restriction digestion and end sequencing. The final FtNwCD vector contained Neo and diphtheria toxin fragment A expression cassettes for positive and negative selection in ES (embryonic stem) cells respectively, and the sequence was confirmed by both restriction digestion and end-sequencing analysis. The vector was linearized with NotI before electroporation into C57BL/6 ES cells (Caliper Life Sciences/Xenogen, Cranbury, NJ). G418-resistant ES clones were analysed by Southern-blot analysis, the selected clones injected into FVB blastocysts, and the blastocysts implanted into the uteri of pseudo-pregnant females for the generation of chimaeras. Male chimaeric mice were bred, and the resulting progeny were genotyped until heterozygous germline transmission was achieved. Heterozygous animals (Ostα+/−) were bred to generate Ostα−/− mice, and all animals were maintained on a standard laboratory diet (LabDiet, PMI Nutrition International, Saint Louis, MO, U.S.A.) at the University of Rochester School of Medicine and Dentistry Vivarium. Genotyping was performed by PCR analysis of DNA isolated from tail biopsies. Genomic DNA was isolated from 0.5 cm of tail using DirectPCR-Tail lysis reagent (Viagen Biotech, Inc., Los Angeles, CA, U.S.A.). A 1 μl portion of crude lysate was used in a one-step genotyping PCR using primers for amplification of a ~600 bp segment of the Neo cassette (forward: 5′-CTGTGCTCGACGTTGTCACTG-3′; reverse: 5′-GATCCCCTCAGAAGAACTCGT-3′), and a ~800 bp region of Ostα genomic DNA containing exons 4 and 5 (forward: 5′- ATTTGTGGTGTCAGTTCTCCTGTCT-3′; reverse: 5′-TATTATTGGCTTTGCCCTACACAAG-3′) designed with Primer Express 1.5.

Protein and total mRNA isolation from various mouse tissues

Wild-type and Ostα−/− mice were anesthetized with pentobarbital sodium (50 mg/kg, intraperitoneally). For protein isolation, liver, kidney and ileum were homogenized in buffer A [10 mM Tris/HCl, pH 7.4, 10 mM KCl, 1.5 mM MgCl2, 200 μg/ml EDTA, 0.25 M sucrose, 0.625% protease-inhibitor cocktail (Sigma–Aldrich P8340) and 2 mM PMSF], and then centrifuged at 800 g for 20 min at 4 °C. The supernatant was the protein lysate for each tissue. For RNA isolation, liver, kidney, duodenum, jejunum and ileum were collected. Tissue samples not immediately processed for RNA isolation were stored in RNAlater (Ambion). Total mRNA was isolated using Rneasy Midi Kit (Qiagen).

Isolation of a membrane fraction from mouse ileum

Ileum from C57BL/6 mice was homogenized in buffer A and then centrifuged at 800 g for 20 min at 4 °C. The supernatant was centrifuged at 32 000 g for 20 min at 4 °C. The mouse ileum membrane fraction was obtained by resuspending the pellet of the second centrifugation in buffer B (10 mM Tris/HCl, pH 7.4, 200 μg/ml EDTA, 0.125 M sucrose, 0.625% protease-inhibitor cocktail and 2 mM PMSF).

Co-IP and immunoblotting

IPs in mouse ileum membrane fractions and in transfected HEK-293 cells were performed using the Protein G IP kit (IP-50) from Sigma–Aldrich. After pre-clearing with Protein G–agarose, mouse ileum membrane fraction or proteins extracted from transfected HEK-293 cells (~300 μg) were solubilized in 250 μl of RIPA buffer (50 mM Tris/HCl, pH 7.4, 1% Nonidet P40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF and 0.625% protease-inhibitor cocktail). Proteins were incubated with 2.5 μg of anti-Ostα (mA315), anti-Ostβ (mB91), anti-Mrp1, anti-FLAG or 2 μl of anti-c-Myc antibodies overnight at 4 °C. Immunocomplexes were isolated by incubation with 50 μl of Protein G–agarose for 2 h at room temperature. The agarose beads were washed four times with PBS, and the product of the IP was eluted by adding 90 μl of 1×SDS sample buffer (40 mM Tris/HCl, 4% glycerol and 2% SDS) with 2-mercaptoethanol, incubating for 30 min at 37 °C. The protein in the eluate was separated on SDS/4–20%-(w/v)-PAGE gels (Bio-Rad). The separated polypeptides were electrotransferred on to a PVDF membrane (Bio-Rad) for 90 min at 95 V using a wet-transfer apparatus (Bio-Rad). The membranes were blocked overnight in 5% (w/v) dried milk in TBST (20 mM Tris, 140 mM NaCl, 0.05% Tween-20, pH 7.4) at 4 °C. Antibodies were used at 1:1000 [anti-Ostα (mA315) and anti-Ostβ (mB91)] and at 1:3000 [horseradish-peroxidase-conjugated anti-rabbit secondary antibody (Amersham Biosciences)]. Antibody binding was then detected using an enhanced-chemiluminescence technique (Perkin–Elmer).

Immunofluorescence staining

HEK-293 cells were grown on lysine-coated four-well chambered coverslips (Lab-Tek II) and transiently transfected with 0.3 μg each of the following plasmids: Ostα-FLAG or Ostβ–c-Myc (negative controls), both Ostα-FLAG and Ostβ–c-Myc, or both Ostα–FLAG and c-Myc–Ostβ. For surface antigen immunofluorescence, the cells were not permeabilized. Cells were fixed for 15 min in 4% (w/v) formaldehyde in PBS2+ (PBS containing 1 mM MgCl2 and 0.1 mM CaCl2), blocked for 30 min with 5% (v/v) normal goat serum in PBS2+, and then incubated at room temperature for 1 h with 1:50-diluted anti-c-Myc (M 5546) or 1:20-diluted anti-FLAG (F 3165) antibody. Subsequently, the cells were incubated at room temperature for 1 h with rhodamine-conjugated anti-mouse secondary antibodies (Jackson ImmunoResearch Laboratories) at a concentration of 30 μg/ml and then mounted in Prolong® gold antifade reagent with DAPI (4′,6-diamidino-2-phenylindole; Invitrogen P36935). Labelled cells were analysed with an Olympus AH-2 microscope using 20× or 40× air-immersion objective lenses. When the cells were permeabilized, a similar procedure was used, except that the blocking and the antibody incubation solutions were supplemented with 0.1% saponin.

Protein stability assay

HEK-293 cells were transfected with Ostα–FLAG, Ostβ–c-Myc or both (0.1 μg of each plasmid per 3-cm-diameter plate). At 24 h after transfection, the cells were changed to medium with or without 100 μg/ml cycloheximide (Sigma–Aldrich), a protein-synthesis inhibitor. Cells were harvested at different time points and processed for immunoblotting using anti-FLAG or anti-c-Myc antibodies.

Real-time quantitative reverse-transcriptase PCR analyses

Gene-specific primers for Ostβ were as previously described [4,6]. Relative gene expression was determined on a Corbett Rotor-Gene 3000 real-time cycler. Samples were analysed using an iScript One-step reverse-transcriptase PCR kit with SYBR® Green (Bio-Rad).

Production of fluorescent fusion proteins

The sequences encoding YFP (1–154) (yellow fluorescent protein-1–154-peptide) and YFP (155–238) were kindly provided by Dr Tom K. Kerppola (University of Michigan Medical School, Ann Arbor, MI, U.S.A.) [17] and cloned into pCMV-HA (Clonetech) and pCMV-FLAG2 (Sigma–Aldrich). The sequences encoding YFP (1–158) and YFP (159–238) were generously provided by Dr Catherine H. Berlot (Weis Center for Research, Geisinger Clinic, Danville, PA, U.S.A.) [18] and cloned into pDNAI/Amp. To construct mammalian expression vectors for BiFC analysis, the mouse Ostα sequence was fused to the gene coding for the N-terminal end of YFP (1–154), YFP (155–238) or the C-terminal end of YFP (1–158) to produce Ostα-YN, Ostα-YC, and YN-Ostα. Mouse Ostβ was fused to the gene coding for the N-terminal end of YFP (155–238) or C-terminal end of YFP (159–238) to produce Ostβ-YC or YC-Ostβ. The constructs were verified by DNA sequencing.

Transient expression and assay for bile-acid-transport activity

HEK-293 cells were maintained under the conditions recommended by the A.T.T.C. Cells grown in six-well plates to 70–80% confluence were transfected with 0.1–1 μg of the plasmids expressing the proteins indicated in each experiment using FuGENE® 6 (Roche). After 24 h of culture, the cells were incubated at 37 °C for 30 min in KH (Krebs–Henseleit) buffer (118 mM NaCl, 4.7mM KCl, 1.2 mM KH2PO4, 25 mM NaHCO3, 0.6 mM MgSO4, 1.25 mM CaCl2 and 5 mM Hepes/Tris, pH 7.5) containing 25 μM [3H]taurocholic acid. After incubation, the culture medium was removed and each cell monolayer was washed three times with ice-cold KH buffer containing 0.2% (w/v) BSA and 1 mM taurocholate, and then once more with ice-cold KH buffer alone. The cell monolayer was dissolved in 0.1 M NaOH and aliquots were taken to determine cell-associated protein and radioactivity.

Imaging of fluorescent fusion proteins

HEK-293 cells grown in 35-mm-diameter glass-bottomed dishes (MatTek P-35G-0-14-C) to 50% confluence were transfected with 0.25–0.5 μg of various plasmids using FuGENE® 6. Transfected cells were incubated at 37 °C for 24 h and then switched to 30 °C for 2 h to promote fluorophore maturation. The cells were observed under a Leica TCS SP Spectral confocal microscope equipped with an inverted DMIRBE microscope. YFP fluorescence emission was measured at 555/15 nm. The fluorescence of DsRed2, a red fluorescent protein from the coral Discosoma, was measured at 610/30 nm.

Statistical analyses

Results are given as means±S.E.M. Mean values were considered to be significantly different (P<0.05) when one-way ANOVA followed by Bonferroni's multiple comparison test were used.


Mouse ileal Ostα and Ostβ proteins appear as multiple bands on Western blots

In the mouse, Ostα and Ostβ proteins are most abundant in kidney and small intestine, especially in the ileum [3]. To determine the relative molecular sizes of Ostα and Ostβ in the ileum and to test for possible protein complexes, affinity-purified polyclonal anti-peptide antibodies for mouse Ostα (mA315) and Ostβ (mB91) were used in Western-blot analysis of mouse ileum membrane proteins. Ostα was detected as a predominant protein band of ~40 kDa, as well as bands of ~50 and ~80 kDa, although the 50 kDa band was faint and not always detectable (Figure 1A). Comparable results were observed when the gel was run under non-reducing conditions (results not shown). The predicted molecular mass of mouse Ostα is 37.8 kDa, and thus the 40 kDa band is likely to represent the monomer form of Ostα [3]. The 80 kDa band, on the other hand, may represent Ostα that is either post-translationally modified or complexed with another protein, including perhaps Ostα itself (i.e., the Ostα homodimer).

Figure 1
Glycosidase sensitivity of Ostα and Ostβ in mouse ileum, and co-IP of these proteins in mouse ileum and transfected HEK-293 cells

To test whether the higher-molecular-mass band of Ostα is a glycosylated form, mouse ileum membrane proteins were treated with PNGase F. As expected [3], PNGase F treatment led to a decrease in the 40 kDa Ostα band and the appearance of a 36 kDa band (Figure 1A, lane 2). Interestingly, the size of 80 kDa band also decreased after treatment with PNGase F, and a new band was formed at about ~60 kDa (Figure 1A, lane 2). The 80 kDa band was not sensitive to Endo H treatment (results not shown).

For Ostβ, bands were detected at about 19 and 17 kDa, although these bands often overlapped (Figure 1B). The predicted molecular mass of mouse Ostβ is about 14.7 kDa. The Ostβ antibody was not able to detect the ~50 and 80 kDa bands stained by anti-Ostα, indicating that these higher-molecular-mass bands are not heteromeric complexes of Ostα and Ostβ. The 17 and 19 kDa bands of Ostβ were minimally affected by PNGase F (Figure 1B).

Co-immmunoprecipitation of Ostα and Ostβ

To examine whether Ostα and Ostβ associate directly, IP and immunoblot analyses were carried out using mouse ileum membrane proteins. The mA315 (anti-Ostα) antibody was used to immunoprecipitate the Ostα subunit, and the precipitated proteins were probed with mA315 and mB91 antibodies. The experiment was also performed in the opposite direction using the anti-Ostβ (mB91) to immunoprecipitate Ostβ. Mrp1, a transport protein that is also located in the basolateral membrane, was chosen as a negative control. IP of Ostβ from mouse ileum resulted in the co-IP of a 40 kDa protein that was detected by the anti-Ostα antibody (Figure 1C, lane 2). Likewise a 19 kDa Ostβ-reactive band was co-immunoprecipitated by the Ostα antibody (Figure 1D, lane 1), indicating that the two proteins are associated with each other. Anti-Mrp1 antibody did not pull down any band corresponding to Ostα (Figure 1C, lane 3, open arrow), and only a faint signal was detected with the anti-Ostβ antibody (Figure 1D, lane 3, open arrow). Because the antibodies used to immunoprecipitate and immunoblot were both rabbit polyclonal antibodies, some extra bands were created that obscured the ~50 and ~80 kDa bands. Thus the higher-molecular-mass Ostα bands which were detected by mA315 (Figure 1A) were not distinguishable in this experiment.

To confirm these findings and examine whether the higher-molecular-mass forms of Ostα also interact with Ostβ, co-IP studies were also performed in transfected HEK-293 cells. Mouse Ostα was labelled with the FLAG epitope and mouse Ostβ was tagged with the c-Myc epitope on their C-termini. The epitope tags allowed the use of different antibodies for IP and immunoblotting. When Ostα–FLAG was expressed by itself at relatively high levels (1 μg of DNA per 3-cm-diameter plate), bands at ~36, 60 and 80 kDa were formed (Figure 1E, lane 2); however, when co-expressed with Ostβ–c-Myc, a new ~40 kDa mA315-antibody-reactive band was formed (Figure 1E, lane 4), consistent with glycosylation of Ostα [3]. When the same samples were blotted with anti-Ostβ antibody, a prominent ~19 kDa band was seen only in cells co-expressing Ostα–FLAG and Ostβ–c-Myc (Figure 1F, lane 4), indicating that precipitation of Ostα resulted in the co-IP of Ostβ. When the anti-c-Myc antibody was used to precipitate Ostβ–c-Myc, bands were detected with the mA315 antibody at ~40 kDa and ~80 kDa, whereas the 36 kDa and the ~60 kDa were not detected (Figure 1E, lane 8). These results support the findings obtained in mouse ileum, and suggest that the ~80 kDa form of Ostα also interacts with Ostβ.

Visualization of Ostα–Ostβ heterodimers or heteromultimers in living HEK-293 cells

To confirm the IP results and to localize the putative Ostα-Ostβ heterodimer or heteromultimers in living cells, BiFC analysis was performed. Although HEK-293 cells express human OSTα and OSTβ mRNA, the levels are quite low (approx. 36 and 49 copies/ng of total RNA respectively) and did not interfere in the analysis of the heterologously expressed mouse genes. DNA coding for an N-terminal YFP fragment (residues 1–154) was fused to the C-terminus of mouse Ostα to produce Ostα-YN, and DNA coding for a C-terminal YFP fragment (residues 155–238) was fused to the C-terminus of mouse Ostβ to produce Ostβ-YC. The pDsRed2-ER plasmid (Clontech) was co-transfected as a marker of the endoplasmic reticulum, and the bJun-YN and bFos-YC plasmids were used as a BiFC-positive control. As reported previously [17], bJun-YN–bFos-YC associated preferentially in nuclei (Figures 2A–2C). When Ostα-YN was co-expressed with untagged Ostβ, or when Ostβ-YC was co-expressed with untagged Ostα, no fluorescence signal was detected, as expected (results not shown); however, when Ostα-YN and Ostβ-YC were co-expressed, a fluorescence signal was detected at both the plasma membrane and intracellularly (Figures 2E and and2F).2F). The intracellular Ostα-YN–Ostβ-YC green staining partially co-localized with the red endoplasmic reticulum staining, suggesting that Ostα-YN and Ostβ-YC interact in the endoplasmic reticulum and that the resulting complex(es) then proceed to the cell membrane.

Figure 2
Heterodimerization and functional activity of Ostα-YN and Ostβ-YC in HEK-293 cells

To confirm that the putative Ostα-YN–Ostβ-YC complex is functionally active, [3H]taurocholate transport activity was measured in the transfected cells. As expected [2,3], co-expression of untagged Ostα–Ostβ generated taurocholate transport activity (Figure 2G). Note that roughly comparable taurocholate transport activity was observed when Ostα-YN and Ostβ-YC were co-expressed (Figure 2G), indicating that the C-terminal YFP fragment constructs are able to reach the plasma membrane and generate a functional transporter.

Ostα forms homodimers or homomultimers, and Ostβ is required for delivery of the complex to the plasma membrane

To directly test the possibility that Ostα is forming a homodimer and/or homomultimers, BiFC analysis was performed using two different Ostα constructs: one tagged with the YFP C-terminal fragment (Ostα-YC) and the other with the N-terminal fragment (Ostα-YN). When these two constructs were co-expressed in HEK-293 cells, the total amount of YFP signal was relatively low, but was detected in a number of cells, indicating that Ostα-YN interacts with Ostα-YC (Figures 3B and and3C).3C). In contrast with the plasma-membrane localization of the heteromeric Ostα-YN–Ostβ-YC complex (Figure 2E and F), the YFP signal of Ostα-YN–Ostα-YC largely overlapped that of the endoplasmic-reticulum marker and there was no significant plasma-membrane staining (Figure 3C).

Figure 3
Ostα homodimer formation in HEK-293 cells

On the other hand, when Ostβ was transfected along with Ostα-YN and Ostα-YC, a strong YFP signal was now observed at the plasma membrane (Figures 3D–3F). In this experiment, pDsRed2-Mem, a marker of the plasma membrane, was used. This plasmid encodes a fluorescent fusion protein that is targeted to plasma membrane by the N-terminal 20 residues of neuromodulin [19] (Figure 3D). As illustrated in Figures 3(E) and and3(F),3(F), the fluorescence signal overlapped with that of the pDsRed2-Mem signal, confirming that the complex localizes to the plasma membrane.

Measurements of bile acid transport activity revealed that cells transfected with Ostα-YN and Ostα-YC showed no significant increase in taurocholate transport activity; however, when Ostβ was co-transfected along with Ostα-YN and Ostα-YC, taurocholate transport activity was now comparable with that generated by untagged Ostα and Ostβ (Figure 3G).

Membrane topology of Ostα and Ostβ

On the basis of the evolutionarily conserved seven-TM-domain-predicted architecture of Ostα, this protein is thought to have an extracellular N-terminus and an intracellular C-terminus [2]. On the other hand, Ostβ is thought to have a single TM domain [2], but its membrane topology is more difficult to predict. To gain insight into the membrane topology of Ostα and Ostβ, immunofluorescence studies were carried out in HEK-293 cells transfected with C-terminally tagged Ostα (Ostα–FLAG) along with Ostβ tagged at either its C- or its N-terminus with the c-Myc epitope (Ostβ–c-Myc or c-Myc–Ostβ respectively) (Figure 4). Both of these combinations (i.e., Ostα–FLAG with either Ostβ–c-myc or c-myc–Ostβ) generated taurocholate transport activity in transfected cells (results not shown). When cells were permeabilized with 0.1% saponin, both of the combinations were labelled by the anti-c-Myc antibody (Figures 4A and and4C);4C); however, when non-permeabilized cells were incubated with anti-c-Myc antibody, only cells transfected with the N-terminally tagged Ostβ construct were labelled (Figure 4D), supporting the conclusion that the C-terminus of Ostβ is intracellular. Likewise, when cells were incubated with anti-FLAG antibody, only cells permeabilized with saponin showed staining (Figures 4E and and4F),4F), indicating that the C-terminus of Ostα is also intracellular.

Figure 4
Ostβ is a type Ia transmembrane protein

To confirm this membrane topology, BiFC analysis was performed with various combinations of constructs, including YN-Ostα and YC-Ostβ, in which the two YFP fragments were placed on the N-terminal portions of Ostα and Ostβ. HEK-293 cells were transfected with either (a) Ostα-YN and Ostβ-YC, (b) Ostα-YN and YC-Ostβ, (c) YN-Ostα and Ostβ-YC or (d) YN-Ostα and YC-Ostβ, along with pDsRed2-ER as a control for transfection and as a marker of the endoplasmic reticulum (Figures 5A–5D). Figures 5(A)–5(D) show the results of the merged YFP and pDsRed2 fluorescence signal, and Figure 5(E) provides measurements of their functional activity (taurocholate transport).

Figure 5
Membrane topology of Ostα and Ostβ as assessed by BiFC analysis

Of these four combinations of BiFC constructs, only the Ostα-YN and Ostβ-YC pair was able to elicit a YFP signal (Figure 5A), indicating that the C-termini of both proteins are on the same side of the membrane and most likely the cytoplasmic one. This combination also elicited taurocholate transport activity (Figure 5E). No fluorescence signal was detected when one YFP fragment was placed on the N-terminal portion of Ostα and one on the C-terminal portions of Ostβ, and vice versa (Figure 5). It should be noted, however, that the addition of N-terminal epitope tags to Ostα invariably abolished taurocholate transport activity (Figure 5E; results not shown), indicating that such modifications to Ostα generate products that are either unstable, unable to be targeted properly to the plasma membrane or lack transport activity. Of significance, although the Ostα-YN and YC-Ostβ pair failed to generate a YFP signal, it generated taurocholate transport activity (Figure 5E), indicating a functional interaction. These observations provide additional direct support for the conclusion that the C-termini of both Ost proteins are on the same side of the plasma membrane, most likely the intracellular one.

Heterodimerization of Ostα and Ostβ increases stability of the proteins

To examine whether the interaction between Ostα and Ostβ influences the turnover rate of these proteins, HEK-293 cells were transfected with either Ostα–FLAG, Ostβ–c-Myc, or both, and protein abundance was assessed at different time intervals by Western blotting (Figure 6). For these studies, a lower level of DNA was used for the transfection, namely 0.1 μg of DNA per 3-cm-diameter dish as against 1 μg of DNA per 3-cm-diameter dish for the studies illustrated in Figures 1(E) and and1(F).1(F). When Ostα or Ostβ were expressed individually, the proteins were not detectable (Figure 6A, lane 2, and 6B, lane 3), but were easily detected when Ostα and Ostβ were co-expressed (Figures 6A and and6B,6B, lanes 4). To assess relative protein stability, the levels of transiently expressed proteins were compared at different time points in the presence (to inhibit protein synthesis) or absence of cycloheximide. In Ostα–Ostβ-co-expressing cells, the Ostα and Ostβ proteins continued to accumulate with time in culture (Figures 6C and and6D).6D). In the presence of cycloheximide, protein levels decreased only slightly during the 24 h period (Figures 6E and and6F),6F), but they were not detectable when expressed individually (Figures 6G and and6H),6H), suggesting that the proteins are relatively stable when co-expressed.

Figure 6
Relative turnover rates of the Ostα and Ostβ proteins in HEK-293 cells

If heterodimerization is required for protein stability, one would predict that the interacting partner should be present at lower levels in cells that are deficient in either Ostα or Ostβ. To test this possibility, our study took advantage of an Ostα-deficient mouse model recently developed in our laboratory. Ostα-null mice are viable and fertile, but exhibit growth retardation (F. Fang, W. Christian, N. Li, C. Hammond and N. Ballatori, unpublished work). Interestingly, Ostβ mRNA levels were maintained in tissues of Ostα−/− mice (Figure 7A); however, Ostβ protein was not detected in any of the tissues examined (Figure 7B). Taken together with the findings in HEK-293 cells (Figure 6), these results suggest that, in the absence of their heterodimerization partner, Ostα and Ostβ proteins are rapidly degraded.

Figure 7
In tissues from Ostα−/− mice, Ostβ mRNA levels are maintained, yet Ostβ protein is not detected


The present observations provide important insights into Ostα and Ostβ membrane topology, trafficking and protein stability. Ostα and Ostβ are known to be expressed in parallel in most tissues, such that cells which express high levels of Ostα also express relatively high levels of Ostβ, and vice versa [25]. Ostα and Ostβ are both localized to the basolateral plasma membrane of polarized intestinal, renal and biliary epithelial cells, and in vitro transfection studies show that co-expression of Ostα and Ostβ is required for delivery of the individual proteins to the plasma membrane [3,4]. Dawson et al. [3] demonstrated that co-expression is also required to convert the Ostα subunit into a mature N-glycosylated Endo H-resistant form, suggesting that co-expression facilitates the movement of Ostα through the Golgi apparatus [3]. The present results extend these findings by showing that an Ostα–Ostβ heteromeric complex is apparently formed in the ER, is modified as it transits through the Golgi apparatus, and is then targeted to the plasma membrane. By contrast, when Ostα and Ostβ are expressed individually, the proteins appear to be targeted for degradation.

Evidence for a direct interaction between Ostα and Ostβ was obtained from co-IP and BiFC analyses, which are two powerful approaches with which to examine protein–protein interactions [2022]. The use of primary antibodies to immunoprecipitate the native proteins in mouse ileum provides direct evidence for a physiologically relevant interaction, and the results of the BiFC analysis of transfected cells provides insights into the subcellular localization of the complex in intact cells.

The amino acid regions that are important for the interaction between Ostα and Ostβ have not yet been identified, although a recent report by Sun and co-workers [11] suggests that the N-terminal 50 amino acid residues of OSTα may be involved. These investigators demonstrated that truncation of this stretch of amino acids in human OSTα led to the intracellular accumulation of OSTα and OSTβ and to only background levels of taurocholate-transport activity. However, this truncation may have also affected OSTα biogenesis and folding, but these possibilities were not examined by Sun and co-workers [11]. Studies in our laboratory have revealed that although epitope tags are well tolerated on the C-terminus of Ostα (e.g., see Figures 2–5), tags placed on the N-terminus abolished taurocholate-transport activity (Figure 5), indicating that such modifications result in proteins that are either unstable, cannot be targeted properly to the plasma membrane or lack transport activity. Studies are currently in progress to distinguish among these possibilities.

Co-expression of Ostα and Ostβ was required not only for membrane targeting, but also to protect the proteins from degradation. Studies with other multimeric protein complexes have established that folding and assembly of such complexes are tightly controlled processes that ensure the expression of an appropriate number of fully assembled complexes on the cell surface. The assembly of multimeric protein complexes is often directly coupled to the trafficking of their individual components, preventing incompletely assembled complexes from reaching the cell surface [23]. In the present study, when Ostα-YN and Ostα-YC were co-expressed in the absence of Ostβ, the resulting homodimer was retained within the cell, and mainly in the ER. However, in the presence of Ostβ, the fluorescence was stronger and was mainly distributed at the plasma membrane (Figure 3). The altered subcellular YFP localization after transfection with Ostβ indicates that the trafficking of Ostα is regulated by Ostβ. Moreover, the smaller YFP signal from Ostα-YN–Ostα-YC indicates that this homodimer is not as stable in the absence of Ostβ, and it suggests that Ostβ somehow facilitates the movement of the Ostα homodimer through the endoplasmic-reticulum checkpoint. As demonstrated by Dawson et al. [3], co-expression facilitates the glycosylation of Ostα. Another possibility is suggested by studies showing that individual subunits of some ion channels or receptors complexes contain a discrete endoplasmic-reticulum retention signal that is hidden or overcome by forward trafficking signals in the correctly assembled protein complexes [2426]. The unassembled proteins would be retained in the endoplasmic reticulum. Interestingly, mouse Ostβ has an RXR (Arg-Xaa-Arg) motif in its C-terminus, whereas Ostα has an RRK (Arg-Arg-Lys) sequence, and both have dileucine motifs that have been demonstrated to function as endoplasmic-reticulum retention/retrieval signals, which prevent the cell surface expression of individual subunits and partially assembled complexes [23,2729]. Thus one possibility is that Ostβ functions to sequester this retrieval signal in the correctly assembled complex, but studies are needed to test this possibility.

Additional evidence that co-expression of Ostα and Ostβ increases protein stability was provided by studies with transfected HEK-293 cells and with Ostα−/− mice. When Ostα and Ostβ were expressed individually using moderate levels of DNA for the transfections (0.1 μg of plasmid DNA per 3-cm-diameter plate) the proteins were not detected, and were only seen when co-expressed. Interestingly, in Ostα−/− mice the Ostβ protein was also not detected, although the Ostβ mRNA levels were comparable with those of wild-type mice, indicating that the absence of Ostβ protein in the knockout mouse occurs post-transcriptionally and that Ostα expression is required for Ostβ protein stability.

The present study also advances our understanding of the membrane topology of Ostα and Ostβ by showing that C-termini of both Ostα and Ostβ are located intracellularly. This overall topology of Ostα-Ostβ is similar to that of the GPCR–RAMP complexes [1214]. Ostα–Ostβ and GPCR–RAMP both consist of a seven-helix TM protein (i.e. Ostα and GPCRs) and a single-TM accessory polypeptide (i.e. Ostβ and RAMPs) with a cytosolic C-terminus. GPCRs are known to be present as homodimers and heterodimers, whereas GPCR–RAMP complexes are present as heterodimers and hetero-oligomers [15,16]. The present results indicate that the Ostα–Ostβ complex is also likely to be a hetero-oligomer consisting of two Ostα and one or more Ostβ proteins, although the actual stoichiometry has yet to be defined.

Taken together, these findings propose a model in which Ostβ functions both as a chaperone and regulator of transport activity. Formation of the heteromeric complex is coupled to the trafficking and post-translational modification of the proteins and may be required for the formation of the functional transporter. However, additional studies are necessary to confirm this model and to characterize the amino-acid regions that are important for dimerization and transport function.


This work was supported in part by National Institute of Health Grants DK067214 and DK48823, and National Institute of Environmental Health Sciences Training Grant ES07026 and Center Grants ES03828 and ES01247. We thank Dr Tom K. Kerppola for providing C-terminal BiFC plasmids, Dr Catherine H. Berlot for providing N-terminal BiFC plasmids, Dr Paul Dawson (Wake Forest University School of Medicine, Winston-Salem, NC, U.S.A.) for providing the Ostα–FLAG–IRES–Ostβ–c-Myc construct, and Mr Michael Madejczyk, Mr Whitney Christian and Dr Sylvia Notenboom, all of this Institution, for their assistance with some of the analyses.


1. Wang W., Seward D. J., Li L., Boyer J. L., Ballatori N. Expression cloning of two genes that together mediate organic solute and steroid transport in the liver of a marine vertebrate. Proc. Natl. Acad. Sci. U.S.A. 2001;98:9431–9436. [PMC free article] [PubMed]
2. Seward D. J., Koh A. S., Boyer J. L., Ballatori N. Functional complementation between a novel mammalian polygenic transport complex and an evolutionarily ancient organic solute transporter, OSTα–OSTβ J. Biol. Chem. 2003;278:27473–27482. [PubMed]
3. Dawson P. A., Hubbert M., Haywood J., Craddock A. L., Zerangue N., Christian W. V., Ballatori N. The heteromeric organic solute transporter α–β, Ostα–Ostβ, is an ileal basolateral bile acid transporter. J. Biol. Chem. 2005;280:6960–6968. [PMC free article] [PubMed]
4. Ballatori N., Christian W. V., Lee J. Y., Dawson P. A., Soroka C. J., Boyer J. L., Madejczyk M. S., Li N. OSTα–OSTβ, a major basolateral bile acid and steroid transporter in human intestinal, renal and biliary epithelia. Hepatology. 2005;42:1270–1279. [PubMed]
5. Ballatori N. Biology of a novel organic solute and steroid transporter, OSTα–OSTβ Exp. Biol. Med. 2005;230:689–698. [PubMed]
6. Boyer J. L., Trauner M., Mennone A., Soroka C. J., Moustafa T., Zollner G., Lee J. Y., Ballotori N. Upregulation of a basolateral FXR-dependent bile acid efflux transporter OSTα–OSTβ in cholestasis in humans and rodents. Am. J. Physiol. Gastrointest. Liver Physiol. 2006;290:G1124–G1130. [PubMed]
7. Frankenberg T., Rao A., Chen F., Haywood J., Shneider B. L., Dawson P. A. Regulation of the mouse organic solute transporter α–β, Ostα–Ostβ, by bile acids. Am. J. Physiol. Gastrointest. Liver Physiol. 2006;290:G912–G922. [PubMed]
8. Zollner G., Wagner M., Moustafa T., Fickert P., Silbert D., Gumhold J., Fuchsbichler A., Halilbasic E., Denk H., Marschall H. U., Trauner M. Coordinated induction of bile acid detoxification and alternative elimination in mice: role of FXR-regulated organic solute transporter-α/β in the adaptive response to bile acids. Am. J. Physiol. Gastrointest. Liver Physiol. 2006;290:G923–G932. [PubMed]
9. Landrier J. F., Eloranta J. J., Vavricka S. R., Kullak-Ublick G. A. The nuclear receptor for bile acids, FXR, transactivates human organic solute transporter-α and -β genes. Am. J. Physiol. Gastrointest. Liver Physiol. 2006;290:G476–G485. [PubMed]
10. Lee H., Zhang Y., Lee F. Y., Nelson S. F., Gonzalez F. J., Edwards P. A. FXR regulates organic solute transporters α and β in the adrenal gland, kidney and intestine. J. Lipid Res. 2006;47:201–214. [PubMed]
11. Sun A. Q., Balasubramaniyan N., Xu K., Liu C., Ponamgi V. M., Liu H., Suchy F. Protein–protein interaction and membrane localization of human organic solute transporter (hOST) Am. J. Physiol. Gastrointest. Liver Physiol. 2007;292:G1586–G1593. [PubMed]
12. Morfis M., Christopoulos A., Sexton P. M. RAMPs: 5 years on, where to now? Trends Pharmacol. Sci. 2003;24:596–601. [PubMed]
13. Hay D. L, Poyner D. R., Sexton P. M. GPCR modulation by RAMPs. Pharmacol. Ther. 2006;109:173–197. [PubMed]
14. Bockaert J., Pin J. P. Molecular tinkering of G protein-coupled receptors: an evolutionary success. EMBO J. 1999;18:1723–1729. [PMC free article] [PubMed]
15. Angers S., Salahpour A., Bouvier M. Dimerization: an emerging concept for G protein-coupled receptor ontogeny and function. Annu. Rev. Pharmacol. Toxicol. 2002;42:409–435. [PubMed]
16. Bouvier M. Oligomerization of G-protein-coupled transmitter receptors. Nat. Rev. Neurosci. 2001;2:274–286. [PubMed]
16a. National Academy of Sciences. Bethesda: National Institutes of Health; 1985. Guide for the Care and Use of Laboratory Animals, NIH publication 86–23.
17. Hu C. D., Chinenov Y., Kerppola T. K. Visualization of interactions among bZIP and Rel Family proteins in living cells using bimolecular fluorescence complementation. Mol. Cell. 2003;9:789–798. [PubMed]
18. Hynes T. R., Tang L., Mervine S. M., Sabo J. L., Yost E. A., Devreotes P. N., Berlot C. H. Visualization of G protein βγ dimers using bimolecular fluorescence complementation demonstrates roles for both β and γ in subcellular targeting. J. Biol. Chem. 2004;279:30279–30286. [PubMed]
19. Zuber M. X., Strittmatter S. M., Fishman M. C. A membrane-targeting signal in the amino terminus of the neuronal protein GAP-43. Nature. 1989;341:345–348. [PubMed]
20. Phizicky E. M., Fields S. Protein–protein interactions: methods for detection and analysis. Microbiol. Rev. 1995;59:94–123. [PMC free article] [PubMed]
21. Kerppola T. K. Visualization of molecular interactions by fluorescence complementation. Nat. Rev. Mol. Cell Biol. 2006;7:449–456. [PMC free article] [PubMed]
22. Kerppola T. K. Complementary methods for studies of protein interactions in living cells. Nat. Methods. 2006;3:969–971. [PMC free article] [PubMed]
23. Margeta-Mitrovic M., Jan Y. N., Jan L. Y. A trafficking checkpoint controls GABAB receptor heterodimerization. Neuron. 2000;27:97–106. [PubMed]
24. Letourneur F., Klausner R. D. A novel di-leucine motif and a tyrosine-based motif independently mediate lysosomal targeting and endocytosis of CD3 chains. Cell. 1992;69:1143–1157. [PubMed]
25. Letourneur F., Hennecke S., Démollière C., Cosson P. Steric masking of a dilysine endoplasmic reticulum retention motif during assembly of the human high affinity receptor for immunoglobulin E. J. Cell Biol. 1995;129:971–978. [PMC free article] [PubMed]
26. Teasdale R. D., Jackson M. R. Signal-mediated sorting of membrane proteins between the endoplasmic reticulum and Golgi apparatus. Annu. Rev. Cell Dev. Biol. 1996;12:27–54. [PubMed]
27. Ellgaard L., Molinari M., Helenius A. Setting the standards: quality control in the secretory pathway. Science. 1999;286:1882–1888. [PubMed]
28. Zerangue N., Schwappach B., Jan Y. N., Jan L. Y. A new ER trafficking signal regulates the subunit stoichiometry of plasma membrane KATP channels. Neuron. 1999;22:537–548. [PubMed]
29. Ma D., Jan L. Y. ER transport signals and trafficking of potassium channels and receptors. Curr. Opin. Neurobiol. 2002;12:287–292. [PubMed]

Articles from Biochemical Journal are provided here courtesy of The Biochemical Society
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...