The ATP-binding cassette transporter family C 2 (Abcc2) is a member of efflux transporters involved in the biliary excretion of organic anions from hepatocytes. Posttranslational regulation of Abcc2 has been implicated, although the molecular mechanism is not fully understood. In the present study, we performed yeast two-hybrid screening to identify novel protein(s) that particularly interacts with the linker region of Abcc2 located between the NH2-terminal nucleotide binding domain and the last membrane-spanning domain. The screening resulted in the identification of a series of small ubiquitin-like modifier (SUMO)-related enzymes and their substrates. In yeast experiments, all of these interactions were abolished by substituting the putative SUMO consensus site in the linker region (IKKE) in Abcc2 to IRKE. In vitro SUMOylation experiments confirmed that the Abcc2 linker was a substrate of Ubc9-mediated SUMOylation. It was also found that the IKKE sequence is the target of SUMOylation, since a mutant with IKKE is substituted by IRKE was not SUMOylated. Furthermore, we demonstrated for the first time that Abcc2, endogenously expressed in rat hepatoma-derived McARH7777 cells, is SUMOylated. Suppression of endogenous Ubc9 by small interfering RNA resulted in a selective 30% reduction in Abcc2 protein expression in the postnuclear supernatant, whereas subcellular localization of Abcc2 confirmed by semiquantitative immunofluorescence analysis was minimally affected. This is the first demonstration showing the regulation of ABC transporter expression by SUMOylation.
- ABC transporter
the multidrug resistance-associated protein 2/ATP-binding cassette transporter family C2 (MRP2/ABCC2) is the second of nine members of the MRP family. It is located on the apical canalicular membrane of hepatocytes and is involved in the biliary excretion of a wide range of organic anions, including glutathione and glucuronide conjugates and sulfated conjugates of bile salts (5, 18, 35, 51). Extrahepatic expression was also found on the brush border membrane of the small intestine (33, 54), on renal epithelia of the proximal tubules (42, 43), on the luminal membrane of endothelial cells of the small blood capillaries in rat brain (32), and in the apical syncytiotrophoblast membrane of the term placenta (49). The defective expression of ABCC2 in the canalicular membrane in patients with Dubin-Johnson syndrome causes conjugated hyperbilirubinemia, and, consequently, the importance of the canalicular expression of human ABCC2 has been widely recognized in clinical situations (5, 18, 35, 51). Moreover, human ABCC2/rodent Abcc2 is important for the biliary excretion of many kinds of drugs. Indeed, alterations in the expression and function of human ABCC2 may lead to altered pharmacokinetic profiles and results in variations in clinical efficacy in individual patients treated with methotrexate, pravastatin, and irinotecan (7, 15, 17, 34).
It is known that the expression level of ABCC2/Abcc2 is affected by many factors at transcriptional and/or translational levels. Rat Abcc2 mRNA is decreased by bile duct ligation (52), anisoosmolarity, and administration of lipopolysaccharide (29), dexamethasone (28), and a variety of other drugs (22). In addition, rat Abcc2 protein is rapidly internalized from the canalicular membrane surface to the cytosolic pool by hyperosmolarity (27, 45), oxidative stress (19, 46, 48), and administration of endotoxin (29), phalloidin (40), and bile acids (4). Moreover, rat Abcc2 protein expression is regulated in pregnenolone-16α-carbonitrile-treated or pregnant rats by alterations in the polysomal distribution of mRNA (21).
As far as the regulatory molecules involved in the membrane surface expression of Abcc2 are concerned, radixin and PDZ-K1 have been reported to interact with Abcc2 (13, 14, 23, 24, 36). Indeed, impaired canalicular surface expression of Abcc2 and subsequent jaundice have been observed in radixin-knockout mice (23). Radixin is considered as the primary molecule anchoring Abcc2 to the filamentous (F)-actin. Despite the knowledge that such molecules directly interact with Abcc2, the precise mechanism governing Abcc2 protein expression and/or localization is not fully understood at the present moment.
Concerning the molecules that affect the expression and/or localization of ABC transporters, proteins that interact with the linker region of Abcc2 located between the NH2-terminal nucleotide-binding domain (NBD2) and the last membrane-spanning domain (MSD2) (Fig. 1) have been the focus of attention. Hax-1 and myosin light chain 2 (Mlc2) have been identified as a binding partner to rat Abcb11, a bile salt export pump expressed on the bile canalicular membrane of hepatocytes (6, 37). These cofactors particularly interact with Abcb11 on the linker region to regulate its apical localization (6, 37). Moreover, the linker region of CFTR/ABCC7, which is referred to as the R (regulatory) domain, is susceptible to phosphorylation by PKA and PKC, which regulate the opening of the Cl− channel (47). Although no factors have been reported to interact with the linker region of ABCC2/Abcc2, putative posttranslational sites are found in this region, including those for phosphorylation by PKC and casein kinase II, and for small ubiquitin-like modifier (SUMO)-ylation (Fig. 1).
In the present study, we performed yeast two-hybrid screening for the purpose of identifying the protein(s) that is involved in the posttranslational regulation of Abcc2.
MATERIALS AND METHODS
The linker regions of rat Abcc2 (amino acid residues 851F–958V) and human ABCC2 (amino acid 855F–962V) were amplified from Abcc2 cDNA plasmid (16) and cDNA from HepG2 cells, respectively, using the sense (5′-aagaattctttgctaggaactggaagac-3′) antisense (5′-aactcgagtcagaccttcccggtttccacaa-3′) primers (rat Abcc2) and sense (5′-aaggatccaatttgctaagaatctgaagacattt-3′) antisense (5′-aagcggccgctcacacctttccagtttctatgaattc-3′) primers (human ABCC2), which included the digestion sites by EcoRI/XhoI (rat Abcc2) and BamHI/NotI (human ABCC2) as indicated by the underlines. After digestion with restriction enzymes, the amplified fragment was inserted into the bait vector pEG202 (Takara Bio, Shiga, Japan) downstream of the LexA DNA binding domain, which was previously digested by EcoRI/XhoI (for rat Abcc2) and BamHI/NotI (for human ABCC2). Site-directed mutagenesis was performed to obtain rat Abcc2 K949R mutant and human ABCC2 K953R mutant, using a QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA).
A human liver cDNA library was constructed as described previously (1). Briefly, human liver poly(A)+ RNA purchased from Clontech (Shiga, Japan) was used to produce cDNA by using a cDNA synthesis kit (Stratagene) with modified oligo(dT) primers carrying an XhoI restriction site and EcoRI adaptor sequence to allow the unidirectional cloning in an appropriate vector. The cDNAs were then purified by using a CHROMA SPIN-400 column (Clontech) to pool cDNAs longer than 0.5 kb and ligated into the pJG4-5 prey vector previously digested with EcoRI and XhoI. After transformation in bacteria, 1.1 × 107 independent clones were obtained. The protein expression of the bait construct in the yeast was confirmed by Western blot analysis of yeast protein extracts using anti-LexA antibody (Invitrogen, Carlsbad, CA). The Abcc2 linker region was used as a bait to screen 4.8 × 106 clones of the human liver cDNA library with the LexA-based GFP two-hybrid system (Grow'n'Glow system; MoBiTec, Goettingen, Germany). The sequences of 96 positive clones obtained after secondary screening were confirmed by an automatic DNA sequencer (ABI Prism TM 377 DNA sequencer, Applied Biosystems, Foster City, CA).
Preparation of GST-linker fusion protein.
The cDNA fragments encoding the linker region of Abcc2 or its K949R (KR) mutant, in which lysine at 949 was substituted by arginine, was amplified with forward (5′-aaaaagaattctctatccatatgatgttccagattatgcttttgctaggaactggaagac-3′) and reverse (5′-ttttccttttgcggccgctcagaccttcccggtttccacaa-3′) primers with artificial restriction enzyme sites (EcoRI and NotI as underlined) and hemagglutinin (HA) tag (double underlined). PCR products were digested with EcoRI and NotI and then inserted into the same sites of pGEX-5X-2 vector (GE Healthcare Bio-Sciences, Little Chalfont, UK). Plasmids were transformed to Rosetta 2 Escherichia coli competent cells (EMD Chemicals, San Diego, CA) and cultured at 37°C. During the log phase, isopropyl-β-d-galactopyranoside (100 μM) was added and further cultured for 12 h at 15°C. Cells were collected by centrifugation, and the resulting pellets were solubilized with PBS containing 1% Triton X-100, 0.2 mg lysozyme, 7% glycerol, 0.5 M β-mercaptoethanol, 1 mM PMSF, 5 mg/ml leupeptin, 1 mg/ml pepstatin, and 5 mg/ml aprotinin and shock frozen at −80°C. After thawing, specimens were sonicated for 30 s and centrifuged at 10,000 rpm for 20 min. Glutathione S-transferase (GST)-fusion proteins were recovered from soluble supernatant by using glutathione-Sepharose 4B resin (GE Healthcare Bio-Sciences). The protein concentration was determined by the bicinchoninic acid (BCA) method with BSA as a standard.
In vitro SUMOylation.
In vitro SUMOylation assay was performed with the kit according to the manufacturer's instructions (BIOMOL, Goettingen, Germany). Briefly, GST-fusion linker proteins (200 nM) were mixed with E1 (50 nM), E2 (500 nM), SUMO-1 (250 nM), and Mg2+-ATP (5 mM) in the reaction buffer and incubated for 1 h at 30°C. Samples were separated by SDS-PAGE followed by probing with anti-HA antibody (Santa Cruz Biotechnology, Heidelberg, Germany) (1:800) and detected by the ECL-plus detection system (GE Healthcare Bio-Sciences). Antibody was washed out and reprobed with rabbit anti-SUMO-1 antibody (BIOMOL).
McARH7777 cells (Riken, Ibaraki, Japan), derived from rat hepatoma, were maintained in cultured in DMEM (Nacalai Tesque, Kyoto, Japan) supplemented with 10% FBS and 100 units/ml penicillin and streptomycin at 37°C in a humidified atmosphere supplemented with 5% CO2.
Detection of SUMOylated Abcc2 in McARH7777 cells.
Cells were seeded at a density of 2.4 × 106 cells/10-cm dish. After 48 h, cells were solubilized in 2.0 ml of RIPA buffer (150 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, 2 mM EDTA, 2 mM EGTA, 1 mM PMSF, 5 mg/ml leupeptin, 1 mg/ml pepstatin, 5 mg/ml aprotinin, 0.1% SDS, 1 mM Na2PO4, 50 mM NaF, 10 mM N-ethylmaleimide, 50 mM Tris·HCl, pH 8.0) for 10 min on ice. Cells were then scraped from the dish and homogenized by pipetting. After centrifugation at 21,000 g for 10 min, supernatants (0.6 ml) were mixed with control rabbit serum or rabbit anti-Abcc2 antiserum (5 μl) and then incubated at 4°C with rotation. After 1 h, protein G-coated magnetic beads (Adembeads, Ademtec, Pessac, France) were added and incubation was continued for 1 h. Specimens were washed five times with RIPA buffer, eluted by adding elution buffer (2% SDS, 5% glycerol, 0.1% bromophenol blue, 50 mM DTT, 62.5 mM Tris·HCl, pH 6.8), and heated for 15 min at 60°C. Then the specimens were separated on SDS-PAGE and analyzed by the Western blot with rabbit anti-SUMO-1 antibody (BIOMOL).
Suppression of Ubc9 expression via RNA interference.
Suppression of Ubc9 expression was accomplished by using a small interfering RNA (siRNA) duplex targeted to Ubc9 mRNA sequences (siUbc9). The siUbc9 sequences were 5′-GCAGAGGCCUAUACAAUUU-3′ (sense) and 5′-AAAUUGUAUAGGCCUCUGC-3′ (antisense). siRNA for Ubc9 and negative-control siRNA (siPerfect Negative Control; siControl) were purchased from Sigma. Cells were seeded on six-well plates at a density of 3.0 × 105 cells/well and immediately transfected with siRNA by using Lipofectamine RNAiMAX reagent (Invitrogen) according to the method recommended by the manufacturer.
Indirect immunofluorescence staining.
McARH7777 cells were seeded on glass-based 35-mm dishes at a density of 3.0 × 105 cells and transfected with siRNA as described previously. After the indicated culture period, cells were washed with PBS, then fixed with 4% paraformaldehyde at room temperature for 10 min and permeabilized for 10 min with PBS containing 1% Triton X-100. After being washed three times with PBS, specimens were blocked with PBS containing 1% BSA at room temperature for 30 min. Then specimens were incubated with anti-Abcc2 monoclonal antibody M2III-6 (Alexis, Lausen, Switzerland) (1:300) in PBS containing 1% BSA at 37°C for 1 h. After being washed three times with PBS, cells were incubated with Alexa 488-conjugated goat anti-mouse IgG (Invitrogen) (1:250), phalloidin-TRITC (2.5 μg/ml), and TO-PRO3 iodide (1:250) in PBS containing 1% BSA at 37°C for 1 h. After being washed three times with PBS, cells were mounted in VECTASHIELD mounting medium (Vector Laboratories, Burlingame, CA). Specimens were visualized by confocal laser scanning microscopy (FV1000, Olympus, Tokyo, Japan). The numbers of nuclei (NNuc), bile canaliculi (NBC), identified by the presence of dense F-actin, and bile canaliculi with Abcc2 staining (NAbcc2) were separately counted to calculate the following indexes: bile canaliculi formation = NBC/NNuc × 100 (%), and Abcc2-positive bile canaliculi formation = NAbcc2/NNuc × 100 (%). At least 500 cells (total 15 fields) were analyzed per specimen. Data are expressed as means ± SD of triplicate preparations.
Determination of protein expression in McARH7777 cells.
Cells were washed with PBS and collected by scraping and centrifugation. After addition of buffer A (2 mM EDTA, 2 mM EGTA, 1 mM PMSF, 5 mg/ml leupeptin, 1 mg/ml pepstatin, 5 mg/ml aprotinin, 50 mM Tris·HCl, pH 7.4) to cell pellets, specimens were homogenized by sonication. After the nuclear and cell debris were removed by centrifugation at 1,500 g for 15 min, the resulting postnuclear supernatants were subjected to protein assay (BCA method). Fifteen micrograms [for Abcc2 and 5′-nucleotidase (5′NT/CD73)] or 5 μg [for ezrin, radixin, moesin (ERM), and Na+-K+-ATPase] of each protein specimen was applied to the SDS-PAGE gel and detected by respective antibodies with ECL-plus detection system. For the analysis, rabbit anti-Abcc2 antiserum (1:1,000), rabbit anti-ERM antibody (TK89) (1:10,000) (26), mouse anti-5′NT monoclonal antibody (1:100) (Santa Cruz Biotechnology), and rabbit anti-Na+-K+-ATPase antibody (1:200) (Santa Cruz Biotechnology) were used.
Identification of SUMOylation-related enzymes as binding partners of Abcc2 linker region.
Human liver cDNA library (4.8 × 106) was screened with rat Abcc2 linker region (851-958) as a bait (Fig. 1). After secondary screening and sequencing of the insert DNA, 96 positive clones were finally isolated. Most of these were from SUMO-related enzymes or molecules, including ubiquitin-activating enzyme 2 (UBA2) (1 clone), ubiquitin-conjugating enzyme 9 (Ubc9) (21 clones), splicing variants of protein inhibitor of activated STAT (PIAS)x (8 clones) and PIASy (1 clone), and SUMO-1 itself (2 clones). These proteins are known to be involved cooperatively in the sequential SUMOylation cycle of the target protein (11, 44). Firstly, SUMO is activated by SUMO activating enzyme E1 [heterodimer of UBA2 and AOS1 (activation of Smt3p)], then covalently attached to the target lysine residue with the aid of SUMO conjugation enzyme E2 (Ubc9) and SUMO ligase E3 (such as PIASx and PIASy) (11, 44). Ubc9 is a key enzyme involved in conjugation of SUMO to the lysine residue in the consensus sequence composed of ΨKXE, where Ψ and X represent a hydrophobic and any amino acid, respectively, of the target protein (39). Indeed, Abcc2 has the expected consensus sequence only within the linker region (IKKE; residue 948-951) (Fig. 1). Supporting the importance of this sequence, mutation of the consensus lysine residue (K949; underlined lysine) to arginine (IRKE, KR mutant) in the linker of the bait construct resulted in the loss of the interaction with all of the preys examined, which include SUMO-1, UBA2, Ubc9, splicing variants of PIASx, and PIASy (Table 1). A positive interaction with these preys was also confirmed when the corresponding linker region from human ABCC2 (855-962; Fig. 1B) was used as bait (Table 1). These results suggest that human and rat SUMO-related molecules interact with the human ABCC2 and rat Abcc2 linker regions in a similar manner, where IKKE is prerequisite for interaction.
In vitro SUMOylation of Abcc2 linker protein.
To examine whether Abcc2 linker can be a target of the SUMOylation reaction, in vitro SUMOylation was performed using recombinant GST-Abcc2 linker protein and purified SUMO-related enzymes and substrate. If the linker was SUMOylated, the target protein should be shifted to a higher molecular weight region in the SDS-PAGE followed by immunoblot detection with anti-HA antibody, because SUMO-1 has a molecular weight of ∼12 kDa. As expected, the GST-Abcc2 linker fusion protein band was shifted from 43 kDa (without SUMOylation reaction mix) to 55 kDa (with SUMOylation reaction mix) (Fig. 2). No band shift was observed in the KR mutant. To support the result that the band appearing at the higher molecular weight corresponds to SUMOylated protein, membrane was reprobed with anti-SUMO-1 antibody. It was confirmed that the reactivity was observed only with the higher molecular weight band, but not with the original 43-kDa band (Fig. 2). Collectively, Abcc2 linker can be SUMOylated and K949 is a target of the SUMOylation reaction in vitro.
In vivo SUMOylation of Abcc2 in rat hepatoma McARH7777 cells.
We then tried to show the SUMOylation of Abcc2 full-length protein in mammalian cells. As a cell line, we used McARH7777, which is derived from rat hepatoma and endogenously expresses Abcc2. Abcc2 was immunoprecipitated with anti-Abcc2 antiserum and probed with anti-Abcc2 antibody and with anti-SUMO-1 antibody. Abcc2 was successfully detected with anti-Abcc2 antibody (∼190 kDa, Fig. 3, top). Moreover, a higher molecular weight band was observed with anti-SUMO-1 antibody in a specimen immunoprecipitated with anti-Abcc2 antiserum, whereas a tracer background signal of the corresponding size was observed in a specimen immunoprecipitated with control rabbit serum (Fig. 3, bottom). These results indicate that the fraction of Abcc2 is SUMOylated in McARH7777 cells.
Ubc9 knockdown does not affect Abcc2 localization but decreases Abcc2 protein expression in McARH7777 cells.
On the basis of the previously described results that SUMO-related molecules were identified in yeast two-hybrid screening and that Abcc2 is SUMOylated in both in vitro and in vivo cell systems, it is speculated that SUMOylation may be involved in the posttranslational regulation of Abcc2. To examine this possibility, Ubc9 was knocked down in McARH7777 cells and its effects on Abcc2 localization and expression were examined. The reason why Ubc9 was selected was that 1) it is the most abundant clone obtained in the screening and 2) it is generally considered to play a central role in the SUMOylation reaction. As shown in Fig. 4A, mRNA expression of Ubc9 was decreased in siUbc9-transfected cells compared with siControl-transfected cells. The bile canalicular structure and formation ratio were not significantly affected at least at 32 h after siUbc9 transfection (Fig. 4, B and C). Abcc2 localization was also not markedly affected at 32 h after siUbc9 transfection; i.e., Abcc2 was exclusively localized in the bile canalicular region of McARH7777 couplets as seen in control siRNA-transfected cells (Fig. 4B). The result of semiquantitative analysis of the Abcc2 localization in the bile canalicular region was also minimally affected by siUbc9 at 32 h after transfection (Fig. 4D).
Abcc2 protein expression was time dependently increased in cells transfected with control siRNA, whereas that in siUbc9-treated cells was decreased in a time-dependent manner (Fig. 5A). Consequently, the Abcc2 expression ratio, defined as the amount of Abcc2 in siUbc9-treated cells divided by that in control siRNA-transfected cells, decreased time dependently (Fig. 5B). The relative expression of Abcc2 protein in siUbc9-transfected cells was significantly decreased to 68 ± 5.6% of the control siRNA-transfected cells at 32 h after transfection (P < 0.05, n = 4, Student's t-test) (Fig. 5, C and D), whereas Abcc2 mRNA expression did not differ significantly between control siRNA and siUbc9-transfected cells (Fig. 5E).
Abcc2 protein expression in the bile canalicular membrane of hepatocytes is physically supported by interaction with ERM proteins, which indirectly connect Abcc2 to the cytosolic F-actin. Without ERM expression, Abcc2 apical membrane surface expression is reported to be decreased (23). However, in siUbc9-transfected cells, the expression level of ERM proteins was not significantly affected (Fig. 5C). Consequently, the decrease in Abcc2 protein expression may not be due to the secondary effect of ERM protein suppression. Moreover, the protein expression of 5′NT and Na+-K+-ATPase was not affected by siUbc9 transfection (Fig. 5C). Collectively, suppression of Ubc9 in McARH7777 cells selectively decreases Abcc2 protein expression without affecting its mRNA expression, the expression of interacting proteins such as ERM proteins, and the expression and/or localization of other membrane proteins.
Most of the clones identified in this screening were SUMO-related molecules, suggesting that Abcc2 may be SUMOylated in the linker region. In accordance with this speculation, we could only find a lysine in position 949 in a putative SUMO consensus sequence. In addition to E2 and E3 enzymes, which interact with the target protein during the SUMOylation cycle, UBA2 and SUMO-1, which are SUMO-activating enzyme and conjugation substrate itself, respectively, were isolated as interacting proteins with Abcc2. On the basis of the general scheme of the SUMOylation cycle, UBA2 and SUMO-1 do not necessarily interact with substrate protein but need to be activated prior to conjugation with the target protein (11, 55). One of the possible explanations for such an unexpected result is that Abcc2 may be SUMOylated in yeast by the endogenous SUMOylation system. Indeed, the SUMOylation system (Smt3p and the Ubc9p system corresponding to mammalian SUMO-1 and Ubc9, respectively) is preserved in yeast (20), and a considerable number of clones encoding another SUMO-interacting protein, Daxx (41), were isolated (data not shown). Supporting the SUMOylation of the Abcc2 linker region in yeast, interaction with all the clones tested [SUMO-1, UBA2, Ubc9, splicing variants of PIASx, PIASy, and Daxx (data not shown)] was abolished by substituting the K949 within the putative SUMO-targeting motif in Abcc2, to which SUMO-1 or its yeast homolog (Smt3p) is covalently linked via an isopeptide bond.
In the present study, we first demonstrated that Abcc2 can be a target of SUMO modification possibly mediated by Ubc9 and other upstream and downstream enzyme systems. After transfection of siUbc9, Abcc2 protein expression was decreased to 70% of the control siRNA-transfected cells at 32 h after transfection (Fig. 5, B and D). Although the Abcc2 protein expression level after transfection of siUbc9 relative to that transfected with control siRNA further decreased after an incubation period of more than 32 h after transfection, it was no longer selective for Abcc2 protein expression. Indeed, the bile canaliculi formation ratio was decreased at 48 h after transfection (data not shown), whereas it was not affected until 32 h after siUbc9 transfection (Fig. 4B). Since bile canalicular membrane formation is required for the expression of Abcc2 on the plasma membrane of McARH7777 cells, downregulation of Abcc2 protein at 48 h after transfection may be partially attributed to the secondary effect of Ubc9 knockdown. In other words, at 32 h after transfection, Abcc2 expression appears to be reduced to at least 70% of the control by the suppression of Ubc9.
The role of Ubc9 in the regulation of membrane protein has been reported in only a few cases. Glucose transporter 4 (GLUT4) is the most well-characterized membrane protein regulated by Ubc9, although modification of this transporter by SUMOylation is still controversial. Giorgino et al. (12) firstly reported that GLUT4 and GLUT1 bind to Ubc9 at the proximal COOH-terminus region of these transporters. GLUT4 and GLUT1 protein expression was increased and decreased, respectively, by overexpression of Ubc9, whereas the catalytically inactive mutant Ubc9-Ala93 did not affect the expression of both transporters in L6 myoblasts (12). Giorgino et al. concluded that Ubc9-mediated SUMOylation of GLUT4 and GLUT1 is important for the regulation of these transporters’ expression. In contrast, Liu et al. (31) claimed that Ubc9 regulates GLUT4 expression by a mechanism independent of its SUMOylation activity, since Ubc9-Ala93 has an effect indistinguishable from wild-type Ubc9 in upregulating the GLUT4 expression in 3T3-L1 adipocytes. It needs to be shown whether SUMOylation of Abcc2 and its regulation of protein expression really depend on Ubc9 catalytic activity in McARH7777 cells.
Concerning the role of SUMOylation of Abcc2, we found that the expression level of Abcc2 protein decreased in a time-dependent manner after transient Ubc9 knockdown (Fig. 5, A and B). The decrease was not attributed to the secondary effect of loss of ERM proteins, which are known to stabilize the apical expression of Abcc2 by anchoring Abcc2 to F-actin. The precise mechanism of the decrease in Abcc2 protein expression after Ubc9 knockdown is not clear at present. The Abcc2 expression level decreased to at least 70% of the control siRNA-transfected cells at 32 h after knockdown of Ubc9. Because of nonspecific cytotoxicity as revealed by the decrease in the bile canaliculi formation ratio at 48 h after knockdown, we could not chase the Abcc2 expression level longer than 32 h after transfection. It is expected that the effect of Ubc9 knockdown on the steady-state protein expression level of Abcc2 may be reduced even more after a longer incubation period. One of the possible mechanisms may be the accelerated degradation of Abcc2 in the absence of SUMO modification of Abcc2. To prove this hypothesis, the half-life of Abcc2 should be determined by monitoring the expression level of labeled Abcc2 after pulse chase labeling and/or the stability of Abcc2 after cyclohexamide treatment. However, we could not obtain any clear conclusions in McARH7777 cells treated with cyclohexamide after transfection with siUbc9 (data not shown). This may be due to multiple factors including the intrinsic toxicity of cyclohexamide, the limited culture period (<48 h) and the relatively long half-life of rat Abcc2 (∼27 h) as reported previously (21).
It has also been established that modification by SUMO competes with the modification by ubiquitin on the same lysine residue, thus protecting the protein from ubiquitin-dependent degradation (30, 50, 53). Ubiquitination of ABC proteins has been extensively studied in yeast (8, 9, 25). It is reported that ubiquitination of the linker region of yeast ABC protein Ste6, a yeast pheromone transporter, and Pdr5, a yeast ABCB1 homolog, are involved in endocytosis from the plasma membrane and the subsequent degradation in vacuoles. Consequently, these ABCs have shorter half-lives of 10–15 min (Ste6) and 60–90 min (Pdr5) (8, 9, 25). In contrast, Pma1, a plasma membrane ATPase, is not ubiquitinated and has a half-life of more than 10 h (2, 8). It is possible that SUMOylation competes with ubiquitination of the Abcc2 linker region, particularly at the IKKE motif, to protect Abcc2 from degradation.
We also focused on the role of SUMOylation on the localization of Abcc2. It was found that the effect of Ubc9 knockdown on the localization of Abcc2 in McARH7777 cells was minimal, at least from our semiquantitative analysis data (Fig. 4, B and D). It is likely that the surface Abcc2 after transfection of siUbc9 may be reduced to ∼70% of siControl cells at 32 h, since the cellular Abcc2 level in siUbc9-transfected cells was reduced to 70% of siControl (Fig. 5D). However, our analytical method to count the number of bile canaliculi positive or negative with Abcc2 is not sensitive enough to detect such a small change. The Abcc2 surface amount should be quantified by biotinylation assay, although this method was not applicable because of the difficulty in the access of the reagents through tight junction to the surface of bile canalicular membrane. Concerning the further role of SUMOylation, it is also possible that this modification affects the function of Abcc2. It is reported that SUMOylation of the K+ channels directly regulates their gate function (3, 38), although the concept has not been fully confirmed (10). The effects of Ubc9 knockdown on the SUMOylation status of Abcc2, its cellular localization, and transporter function need to be evaluated in future studies.
In conclusion, we have demonstrated that Abcc2 interacts with SUMO-related proteins in yeast and, indeed, is SUMOylated in cultured mammalian cells. Ubc9 may be involved in regulating the expression level of Abcc2. This is the first example showing the SUMO modification of ABC transporter family proteins.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society