Am J Physiol Renal Physiol 307: F107–F115, 2014. First published May 14, 2014; doi:10.1152/ajprenal.00085.2014.

Mutation of a single threonine in the cytoplasmic NH2 terminus disrupts trafficking of renal betaine-GABA transporter 1 during hypertonic stress Eva S. Schweikhard,2* Stephen A. Kempson,3* Christine Ziegler,2 and Birgitta C. Burckhardt1 1

Institute of Systemic Physiology and Pathophysiology, University Medical Center Göttingen, Göttingen, Germany; Structural Biology Department, Max-Planck-Institute of Biophysics, Frankfurt am Main, Germany; and 3Department of Cellular and Integrative Physiology, Indiana University School of Medicine, Indianapolis, Indiana

2

Submitted 5 February 2014; accepted in final form 6 May 2014

Schweikhard ES, Kempson SA, Ziegler C, Burckhardt BC. Mutation of a single threonine in the cytoplasmic NH2 terminus disrupts trafficking of renal betaine-GABA transporter 1 during hypertonic stress. Am J Physiol Renal Physiol 307: F107–F115, 2014. First published May 14, 2014; doi:10.1152/ajprenal.00085.2014.— Betaine is an important osmolyte and is, compared with other organs, much more abundant in the kidneys, where it enters cells in the medulla by betaine-GABA transporter 1 (BGT1) to balance osmoregulation in the countercurrent system. In wild-type (wt-)BGT1-expressing oocytes, GABA-mediated currents were diminished by preincubation of oocytes with 100 nM PMA or 5 ␮M dioctanoyl-sn-glycerol, activators of PKC, whereas the application of staurosporine before the application of dioctanoyl-sn-glycerol restored the response to GABA. Four potential phosphorylation sites on BGT1 were mutated to alanine by site-directed mutagenesis. Three mutants (T235A, S428A, and S564A) evoked GABA currents comparable in magnitude to currents observed in wt-BGT1-expressing oocytes, whereas GABA currents in T40A were barely detectable. Uptake of [3H]GABA was also determined in human embryonic kidney-293 cells expressing enhanced green fluorescent protein (EGFP)-tagged BGT1 with the same mutations. T235A, S428A, and S564A showed upregulation of GABA uptake after hypertonic stress and downregulation by PMA similar to EGFP-wt-BGT1. In contrast, T40A did not respond to either hypertonicity or PMA. Confocal microscopy of the EGFP-BGT1 mutants expressed in Madin-Darby canine kidney cells revealed that T40A was present in the cytoplasm after 24 h of hypertonic stress. whereas the other mutants and EGFP-wt-BGT1 were in the plasma membrane. All mutants, including T40A, comigrated with wt-BGT1 on Western blots, suggesting that they are full-length proteins. T40A, however, cannot be phosphorylated, as revealed using a specific anti-phosphoantibody, and, therefore, T40 may be important for the trafficking and insertion of BGT1 in the plasma membrane. endocytosis; membrane trafficking; osmotic stress; protein kinase C

is a challenging environment for cells, especially during antidiuresis, when high concentrations of NaCl and urea are present (8, 14). One important requirement for survival is to balance the osmotic stress by the intracellular accumulation of organic osmolytes that do not disturb normal cell functions (2). Betaine is such an osmolyte, and cell uptake is mediated by betaine-GABA transporter 1 (BGT1), which belongs to the solute carrier (SLC)6 transporter family of Na⫹- and Cl⫺-dependent neurotransmitters (16, 18). BGT1 (also called SLC6A12) is mainly localized in the liver, kidneys, and brain (22). Although the apparent affinity of

THE KIDNEY INNER MEDULLA

BGT1 for GABA is at least fourfold greater than for betaine (10, 20), the primary substrate for renal BGT1 is betaine, which is present in the plasma at concentrations of ⬎100 ␮M (20). Up- and downregulation of BGT1 by changes in extracellular osmolarity is slow in cultured cells, requiring almost 24 h. However, we (5, 6) have previously demonstrated relatively rapid downregulation (within 30 min) of BGT1 transport in Madin-Darby canine kidney (MDCK) cells in response to extracellular ATP, adenosine, and Ca2⫹ (5, 6). Inhibition of BGT1 transport by Ca2⫹ was mediated by the phorbol ester PMA and the diacylglycerol analog dioctanoyl-sn-glycerol (DOG), in agreement with a previous study (17). This finding, suggesting a role for PKC, was accompanied by internalization of BGT1 from the plasma membrane (4, 9). BGT1 phosphorylation on serines and threonines has been previously reported in response to PKC activation in MDCK cells (9). In the present study, several potential PKC phosphorylation sites on BGT1 were mutated to alanine, and the responses to DOG and PMA were determined. Based on sequence alignments, these sites (T40, T235, S418, and S564) are highly conserved in the dog, human, mouse, and rat (http://genome.cs.mtu.edu/map.html) and are located at positions exposed to the cytoplasm (Fig. 1). BGT1 consists of 12 transmembrane domains (TMDs), 6 extracellular loops, and 5 intracellular loops. The Na⫹- and substrate-binding sites are located at or near TMD1 and TMD6 derived from the LeuT structure (21). One potential PKC phosphorylation site (T40) is located near the NH2 terminus; another (S564) is near the COOH terminus. Two other potential phosphorylation sites are situated in intracellular loops between TMD4 and TMD5 (T235) and TMD8 and TMD9 (S418). Since transport by BGT1 is electrogenic, the intrinsic transport activity can be monitored not only by measuring uptake of the radioactive-labeled lead substrates betaine or GABA in cells transfected with mutants but also in frog oocytes injected with cRNAs coding for wild-type (wt-)BGT1 as well as for mutants generated by substitution of the threonines and serines to alanines at the four intracellular PKC phosphorylation sites. To expand the knowledge of these sites, a specific BGT1 phosphoantibody and confocal microscopy were used to identify the localization of the mutants within the expression system, especially when uptake or current measurements showed no substrate-specific signals. METHODS

* E. S. Schweikhard and S. A. Kempson contributed equally to this work. Address for reprint requests and other correspondence: B. C. Burckhardt, Institute of Systemic Physiology and Pathophysiology, Univ. Medical Center Göttingen, Humboldtallee 23, Göttingen D-37073, Germany (e-mail: Birgitta. [email protected]). http://www.ajprenal.org

Chemicals and solutions. Unless specified otherwise, all chemicals were of analytic grade and purchased from Sigma-Aldrich (Deisenhofen, Germany, or St. Louis, MO) and Applichem (Darmstadt, Germany). Standard oocyte Ringer solution (ORi) was used for oocyte

1931-857X/14 Copyright © 2014 the American Physiological Society

F107

F108

DISRUPTED TRAFFICKING OF BGT1 BY MUTATION AT T40

Fig. 1. Topology of betaine-GABA transporter 1 (BGT1). Transporters of the solute carrier 6 family have 12 transmembrane helixes or domains (TMD) with both NH2 and COOH termini facing the cytoplasmic side of the membrane. The transporter core is formed by two inverted repeats of five TMDs with repeat 1 (TMD1–TMD5) and repeat 2 (TMD6 –TMD10). TMD11 and TMD12 do not belong to the inverted repeats. The position of the substrate GABA (G) and two Na⫹ are represented as a triangle and circles, respectively. [Modified based on the topology suggested by Ref. 21.] Four potential phosphorylation sites (T40, T235, S418, and S564) predicted by Prosite database (http://prosite.expasy.org), NetPhos (http://www.cbs.dtu.dk/services/NetPhos/), and NetPhosK (http:// www.cbs.dtu.dk/services/NetPhosK/) are shown as solid circles. EL, extracellular loop; IL, intercellular loop.

preparation, storage, and electrophysiological measurements. ORi contained (in mM) 110 NaCl, 3 KCl, 2 CaCl2, and 5 HEPES adjusted to pH 7.5 with Tris. The following PKC activators and inhibitors were used: DOG, PMA, 4␣-phorbol 12,13-didecanoate, and staurosporine. Stock solutions of these compounds were prepared in either DMSO or ethanol. In vitro transcription of cRNA. DNA of wt-BGT1 and of the mutants was used as a template for cRNA synthesis. Plasmids (pSPORT) were linearized with NotI, and in vitro cRNA transcription was performed using the T7 mMessage mMachine kit (Ambion, Austin, TX) according to the manufacturer’s instructions. After the purification of cRNAs by phenol-chloroform extraction, cRNAs were resuspended in purified, RNase-free water to a final concentration of 1 ␮g/␮l. Site-directed mutagenesis. Using the Prosite database (ExPASy Proteomics Server, SIB), four potential phosphorylation sites for PKC were identified on the cytoplasmic region of BGT1 (Fig. 1). Single mutations were created at each site by substituting either the threonine or serine to the neutral amino acid alanine or glutamate in both wt-BGT1 and enhanced green fluorescent protein (EGFP)-tagged wt-BGT1 (see below). All mutations were performed using the QuikChange site-directed mutagenesis kit (Stratagene, Cambridge, UK). To verify the success of site-directed mutagenesis, both strands of the mutants were sequenced by dye-terminated cycle sequencing using specific BGT1 primers (MWG Biotech, Munich, Germany). PCR products were sequenced using an ABI automatic sequencer (ABI 377, Applied Biosystems, Weiterstadt, Germany). Expression of wt-BGT1 and mutants. Stage V and VI oocytes from Xenopus laevis (Nasco, Fort Atkinson, WI) were prepared by an overnight treatment with collagenase (type CLSII, Biochrom, Berlin, Germany). After being washed with Ca2⫹-free ORi to remove adjacent follicle cells, oocytes were injected with 23 nl of either 1 ␮g/␮l wt-BGT1 cRNA or mutant cRNA or an equivalent volume of water as a control using a nanoliter injector (Nanoliter 2010, World Precision Instruments, Berlin, Germany). Water-injected oocytes were defined as mocks. Injected oocytes were incubated for 3 days in ORi supplemented with 50 ␮M gentamycin and 2.5 mM sodium pyruvate at 16 –18°C. The medium was changed daily, and damaged oocytes were discarded. Electrophysiological experiments. These experiments were carried out 3 days after injection of cRNA or an equivalent amount of water by two-electrode voltage clamp using a commercial amplifier (OC 725, Warner, Hambden, CT) in the voltage-clamp mode. Borosilicate glass microelectrodes (Biomedical Instruments, Zöllnitz, Germany) were filled with 3 M KCl and had resistances of ⬃1 M⍀. Steady-state currents were obtained at ⫺60 mV in the absence and presence of

GABA and betaine and in experiments testing the impact of several PKC activators or inhibitors on GABA-associated currents. Cell culture. Canine MDCK cells (CCL-34) were obtained from the American Type Culture Collection (Rockville, MD) and were cultured in a 1:1 mixture of DMEM-Ham’s F-12 containing 10% bovine calf serum, 15 mM HEPES, 25 mM NaHCO3 (pH 7.4), penicillin (100 IU/ml), and streptomycin (100 ␮g/ml). The same medium was used for the growth of human embryonic kidney (HEK)-293 cells, which were kindly provided by Dr. Paul Herring (Department of Physiology, Indiana University School of Medicine). For immunohistochemistry, MDCK cells were grown in monolayer culture on glass coverslips or collagen-coated Millicell-CM 12-mm filter inserts (Millipore, Billerica, MA). For tracer uptake experiments, HEK-293 cells were grown in six-well plastic plates. Following the manufacturer’s instructions, MDCK cells were transiently transfected using GeneJammer (Stratagene) within 3– 6 h after being plated, and HEK-293 cells were transfected with Fugene 6 (Roche, Indianapolis, IN) 24 h after being plated. Hypertonic stress to MDCK and HEK-293 cells was induced by a 24- or 16-h incubation in the above-mentioned medium to which sucrose was added to reach a final osmolality of 500 or 400 mosmol/ kg, respectively. All tested PKC activators were applied for 30 min in the concentrations indicated in the figures; an equivalent volume of solvent was added to the controls. Transport measurements. The osmolality of all solutions used for transport was matched to the osmolality of the growth medium. Transport activity of wt-BGT1 and mutants in MDCK monolayers was assessed as NaCl-dependent uptake of [3H]GABA (GE Healthcare, Pittsburgh, PA) or [14C]betaine (Moravek Biochemicals and Radiobiochemicals, Brea, CA). Monolayers were washed twice in 300 mM mannitol and 10 mM Tris·HCl (pH 7.4). Afterward, during gently shaking at 21°C for 10 min, the uptake of 0.01 mM [3H]GABA (or 0.1 mM [14C]betaine) in a total volume of 1 ml was assayed. Uptake buffer contained (in mM) 137 NaCl, 5.4 KCl, 2.8. CaCl2, 1.2 MgSO4, and 10 HEPES (pH 7.4). Uptake was terminated by four washes with ice-cold stop solution containing 137 mM NaCl and 14 mM Tris·HCl (pH 7.4). Cell monolayers were solubilized in 0.5% Triton X-100 for liquid scintillation counting and protein determination. All uptake data were corrected for surface binding. Immunostaining and Western blot analysis. MDCK cells were fixed in 4% paraformaldehyde in PBS and processed for antibody staining and confocal microscopy as previously described (5, 7). Cells were permeabilized by a 5-min incubation in PBS containing 0.2% Triton X-100 and 100 mM glycine and afterward rinsed in PBS. An affinitypurified BGT1 antibody (Proteintech Group, Chicago, IL) was used (1:200) after blockade with 2.5% BSA in PBS. The peptide antigen

AJP-Renal Physiol • doi:10.1152/ajprenal.00085.2014 • www.ajprenal.org

DISRUPTED TRAFFICKING OF BGT1 BY MUTATION AT T40

for this antibody was generated against residues 595– 613 of dog BGT1. The secondary antibody was goat anti-rabbit IgG conjugated to FITC (Jackson ImmunoResearch, Suffolk, UK) or alkaline phosphatase (Sigma-Aldrich), and both were diluted 1:100. Cells were washed and counterstained for 5 min in propidium iodide (2 ␮g/ml) and RNase A (50 ␮g/ml) in PBS to visualize nuclei. The trans-Golgi network was labeled by a 60-min incubation with an antibody (Santa Cruz Biotechnology, 1:100 dilution) to the endoprotease furin (19) followed by a 30-min incubation with goat anti-rabbit IgG conjugated with Cy5 (Jackson ImmunoResearch, 1:100 dilution). Triton extracts were prepared from MDCK cells expressing EGFP in wt-BGT1 as well as in the mutants and were processed for Western blot analysis as previously described (5, 7). Affinity-purified rabbit polyclonal antibody to GFP (Abcam, Cambridge, UK) was used at 1:10,000 dilution, and the secondary antibody was goat anti-rabbit IgG conjugated to

time (min)

A 0

20

40

60

GABA-mediated currents (nA)

20 0 -20 -40 wt-BGT1 mock

-60 -80

GABA-mediated currents (nA)

B

GABA-mediated currents (nA)

horseradish peroxidase at 1:5,000 dilution. Blots were stripped and reprobed sequentially for ␤-actin and E-cadherin. For specific detection of phosphorylation at T40, affinity-purified rabbit polyclonal antibody to PKC was generated against residues 32– 45 of dog BGT1 [antigenic sequence: CQVKDRGQW(pThr)NKMEF (GenScript, NJ, USA)] and used at 1:1,000 dilution, and the secondary antibody was goat anti-rabbit IgG conjugated to alkaline phosphatase at 1:5,000 dilution. Immunoprecipitation. Cell lysate (200 ␮l) was incubated with the phospho-PKC-specific antibody under gentle shaking overnight at 4°C. Protein G agarose beads (20 ␮l of 50% bead slurry) were added and incubated with gentle shaking for 1–3 h at 4°C and subsequently centrifuged for 30 s at 4°C. The pellet was washed five times with 500 ␮l cell lysis buffer [1% Triton X-100, 10% glycerol, 150 mM NaCl, 50 mM Tris (pH 7.5), and 5 mM EDTA] supplemented with oComplete protease inhibitor (Roche, Mannheim, Germany). The pellet was resuspended in 20 ␮l of 3⫻ SDS sample buffer [188 mM Tris·Cl (pH 6.8), 3% SDS, 30% glycerol, 0.01% bromophenol blue, and 15% ␤-mercaptoethanol]. After being vortexed, samples were centrifuged for 30 s, heated to 95–100°C for 2–5 min, and centrifuged for 1 min at 14,000 g. Samples (15–30 ␮l) were loaded on a SDS-PAGE gel (12%) and analyzed by Western blot analysis (see above). Confocal microscopy. Localization of wt-BGT1 and mutant proteins was visualized using a LSM 510 laser scanning confocal microscope (Carl Zeiss Microimaging, Thornwood, NY). The procedure has been previously described in detail in Ref. 7. Statistics. Michaelis-Menten parameters for GABA- and betaineinduced currents were calculated using SigmaPlot least-square fit analysis (GraphPad, San Diego, CA). Students’s t-test was applied to prove statistically significant differences of GABA-associated currents in the absence and presence of PKC activators or inhibitors. In the uptake measurements, different groups were compared by ANOVA and Tukey’s test for multiple comparisons (Systat version 3.06 software, GraphPad).

20 mock

wt-BGT1

0 -20 -40 -60

GABA GABA, staurosp. GABA, staurosp., DOG

n.s. n.s.

-80

C

F109

20 wt-BGT1

wt-BGT1

mock

0 -20 -40 -60 -80

n.s.

*

GABA GABA, PMA GABA, 4aPDD all conditions

Fig. 2. Effect of PKC activators and inhibitors on wild-type (wt-)BGT1expressing oocytes and water-injected oocytes (mocks). A: dioctanoyl-snglycerol (DOG) inhibited GABA-associated currents in a time-dependent manner. During a continuous clamp, the oocyte was first perfused with GABA [1 mM dissolved in oocyte Ringer solution (ORi)] in the absence of DOG (i.e., data point at time 0). After washout of GABA, superfusion with DOG (5 ␮M dissolved in ORi) started, and, after 15, 30, and 60 min, this perfusion was changed for 1 min to GABA plus DOG-containing ORi to determine the GABA-mediated current and turned again to ORi plus DOG perfusion until the next GABA application was performed. This protocol was applied to seven wt-BGT1 oocytes from four donors and to three mocks from three donors. In mocks only, data points at time 0 and at 60 min were obtained. B: pretreatment of oocytes with staurosporine reduced the inhibitory action of DOG. GABAassociated currents were first evaluated in the absence of either DOG or staurosporine to prove successful expression of wt-BGT1 (solid columns). Oocytes were then incubated for 1 h in ORi supplemented with 10 ␮M staurosporine. At the end of this period, GABA-associated currents were determined again (shaded columns). After washout of GABA, oocytes were incubated for additional 30 min in ORi supplemented with 5 ␮M DOG, and again at the end of this time period, GABA-associated currents were determined (open columns). This experimental procedure was applied to five wt-BGT1 oocytes from four donors and three mocks from three donors. C: PMA but not 4␣-phorbol 12,13-didecanoate (4␣-PDD) inhibited GABAinduced currents. In the same oocyte, currents induced by 1 mM GABA were first measured in the absence (solid columns) and then at the end of a 30-min pretreatment with either 100 nM PMA (3 oocytes from 3 donors; dark shaded column) or 100 nM 4␣-PDD (4 oocytes from 3 donors; light shaded column). Data for mocks were obtained from either three oocytes (100 nM PMA) or three oocytes (100 nM 4␣-PDD) from three donors and were pooled. Data are means ⫾ SD determined at a clamp potential of ⫺60 mV, which was continuously applied during all experimental procedures. *P ⬍ 0.01.

AJP-Renal Physiol • doi:10.1152/ajprenal.00085.2014 • www.ajprenal.org

F110

DISRUPTED TRAFFICKING OF BGT1 BY MUTATION AT T40

GABA-mediated currents (nA)

A wt-BGT1 T40A

T235A S418A S564A mock

0

*

*

-20

occurred after a 30-min preincubation of oocytes in ORi containing 5 ␮M DOG (Fig. 2A, solid circles). Under these conditions, GABA-mediated currents decreased by 93.6 ⫾ 5.1% (7 oocytes from 3 donors, P ⬍ 0.001). Extending the incubation time to 60 min did not enlarge the effect. DOG had no effect on the currents in mocks (Fig. 2A, open circles). When wt-BGT1-expressing oocytes were treated for 60 min with 10 ␮M staurosporine, there was only a small, insignificant

-40 -60

*

-80

GABA-mediated currents (nA)

B

20 wt-BGT1

T40A

mock

0 n.s.

-20 -40 -60

*

- PMA + PMA

-80 Fig. 3. Response of wt-BGT1- and mutant-expressing oocytes to GABA in the absence and presence of PMA. A: comparison of currents induced by 1 mM GABA in wt-BGT1-, T40A-, T235A-, S418A-, and S564A-expressing oocytes. For each of the three experiments performed, wt-BGT1 and two mutants were expressed in the same batch of oocytes to minimize interindividual changes in oocyte expression. B: effect of PMA on wt-BGT1- and T40A-expressing oocytes. Currents induced by 1 mM GABA were measured before (solid columns) and after a 30-min application of 100 nM PMA (shaded columns). Experiments in A were performed on six wt-BGT1-expressing, T40A-expressing, T235A-expressing, S418A-expressing, S564A-expressing, and mock oocytes from three donors. Experiments in B were performed on three wt-BGT1-expressing, T40A-expressing, and mock oocytes at a clamp potential of ⫺60 mV. *P ⬍ 0.01. RESULTS

At ⫺60 mV, superfusion of wt-BGT1-expressing oocytes with either 1 mM GABA or 1 mM betaine in ORi produced inward currents as long as the substrate was present in the superfusate. Application of increasing GABA or betaine concentrations induced substrate-mediated currents, which tended to saturate at concentrations of ⬎0.05 mM GABA and ⬎0.5 mM betaine, respectively (data not shown). From these results, Km values of 0.020 ⫾ 0.003 mM for GABA (5 oocytes from 4 donors) and 0.18 ⫾ 0.05 mM for betaine (3 oocytes from 3 donors) were calculated. The magnitudes of the maximal substrate-inducible currents for GABA and betaine were ⫺21.7 ⫾ 0.6 and ⫺18.3 ⫾ 1.1 nA, respectively. Due to its higher affinity to BGT1, in all further experiments using the two-electrode voltage-clamp device, only GABA-associated currents were monitored. These currents were sensitive to the PKC activator DOG. At ⫺60 mV, significant inhibition of GABA-induced currents

Fig. 4. Expression of mutants of enhanced green fluorescent protein (EGFP)wt-BGT1 in Madin-Darby canine kidney (MDCK) cells in monolayer cultures. A: response to hypertonic stress. After adaption to 500 mosM medium for 24 h, confocal micrographs of live cells revealed plasma membrane localization of all mutants except T40A. Micrographs are representative of a series of two identical experiments. Bars ⫽ 20 ␮m. B: confocal images of the T40A mutant of EGFP-BGT1 showing colocalization (yellow fluorescence, arrows) with furin (red fluorescence) a marker for the trans-Golgi network. The images represent two separate experiments, and the cell border is outlined in white. C: proof of expression of the full-length protein. Western blots of whole cell lysates using anti-GFP antibody revealed that all mutants expressed a protein of 95 kDa, similar to wt-BGT1. The housekeeping genes ␤-actin and E-cadherin were used to demonstrate equal loading of the single lanes and insensitivity to hyperosmolality. Note that T40A was investigated in duplicate.

AJP-Renal Physiol • doi:10.1152/ajprenal.00085.2014 • www.ajprenal.org

DISRUPTED TRAFFICKING OF BGT1 BY MUTATION AT T40

reduction in GABA-associated currents (Fig. 2B, shaded vs. solid columns). Application of 5 ␮M DOG in the continuous presence of staurosporine (Fig. 2B, open columns) did not change this result. PMA, another activator of PKC, affected also GABA-mediated currents (Fig. 2C). At ⫺60 mV, GABAmediated currents decreased by 53.0 ⫾ 19.9% (4 oocytes from 3 donors) upon a 30-min incubation in 100 nM PMA. The inactive analog 4␣-phorbol 12,13-didecanoate inhibited GABA current only by 13.9 ⫾ 4.9% (3 oocytes from 3 donors). These observations suggest a role for PKC in mediating downregulation of BGT1 transport by DOG and PMA. Therefore, four potential PKC phosphorylation sites on BGT1 (T40, T235, S418, and S564) were mutated to alanine by site-directed mutagenesis. Whereas T235A and S418A evoked GABA-sensitive currents comparable in magnitude to those evoked by wt-BGT1, those by S564A were larger (Fig. 3A). GABA-associated currents obtained by T40A were only ⫺7.8 ⫾ 2.2 nA (5 oocytes from 3 donors) compared with ⫺23.3 ⫾ 11.8 nA (6 oocytes from 3 donors) in wt-BGT1. T40A was also not sensitive toward PMA. PMA (100 nM) inhibited GABA currents in wt-BGT1 (Fig. 3B); the same concentration of PMA resulted in a nonsignificant reduction of GABA-associated currents in T40A (Fig. 3B). The putative PKC phosphorylation sites (T40A, T235A, S418A, and S564A) were mutated in EGFP-tagged MDCK cells as well to allow correlation of intracellular distribution and transport activity after adaptation to hypertonic medium. When expressed in MDCK cells in monolayer culture, all

F111

mutants except T40A were localized primarily in the plasma membrane after 24 h of hypertonic stress (500 mosM; Fig. 4A). In contrast, the T40A mutant showed a scattered distribution in the cytoplasm, in the region of the Golgi. Subsequently, intracellular T40A was found to be tightly correlated with furin (Fig. 4B, arrows), strongly suggesting that T40A resides in the trans-Golgi network (19). Anti-GFP antibody was used for Western blot analysis of whole cell lysates from HEK-293 cells expressing EGFP-tagged wt-BGT1 or one of the mutants (Fig. 4C). HEK-293 cells were used because of a much higher transfection efficiency compared with MDCK cells. A pronounced protein band was detected just below the 100-kDa marker in wt-BGT1 and all mutants (Fig. 4C), which corresponds to the 95-kDa protein identified recently in the kidney medulla (22). No free EGFP (26 kDa) was detected. These findings suggest indirectly that all mutants, including T40A, were expressed as full-length proteins. The intracellular localization of T40A was also confirmed after expression in polarized MDCK cells. As visualized in the micrographs shown in Fig. 5A, top, T40A remained intracellular and close to the nucleus, whereas T235A, S418A, and S564A (Fig. 5A, bottom) were located exclusively in the plasma membrane after 24 h of incubation in hypertonic medium, similar to wt-BGT1. This suggests that the cytoplasmic location observed in monolayer cultures (Fig. 4, A and B) was not due to the lack or disorganization of protein trafficking. The response of T235A, S418A, and S564A to PMA was

Fig. 5. Distribution of the mutants in polarized MDCK cells. A: comparison of the distribution pattern of the mutants transfected in MDCK cells grown on permeable supports. Polarized MDCK cells and mutant-transfected MDCK cells grown on filters were exposed for 24 h to an osmotic stress of 500 mosM. In contrast to T235A, S418A, and S564A (bottom), which showed staining of the plasma membrane, T40A (top) revealed even intracellular staining. B: internalization by PMA. MDCK cell monolayers grown on coverslips were exposed for 24 h to an osmotic stress of 500 mosM in the absence of PMA (⫺; top). After 24 h of osmotic stress, treatment of monolayers with 75 nM PMA (⫹; bottom) for 30 min led to an internalization of the protein in wt-BGT1, T235A, S418A, and S564A groups. Note the marked punctuate fluorescence signal in PMA-treated cells compared with the clear membrane staining in untreated cells. Nucleic acids were stained with propidium iodide (red) for better visualization of the intracellular staining. Hyperosmolality was induced by the addition of sucrose. Bars ⫽ 20 ␮m. Controls (EGFP-BGT1) were fixed and stained before imaging, but the confocal images of the mutants were obtained from live cells. Images of the untreated mutants are reproduced from Fig. 4A to allow direct comparison with the effect of PMA treatment, which was performed at the same time within the same live cell experiment. Micrographs are representative of a series of two identical experiments, as described in B, top.

AJP-Renal Physiol • doi:10.1152/ajprenal.00085.2014 • www.ajprenal.org

DISRUPTED TRAFFICKING OF BGT1 BY MUTATION AT T40

or DOG (Fig. 7B). This band was not observed using samples derived from cells that had been transfected with the vector alone after pretreatment with PMA or DOG (Fig. 7B, HEKvector) or wt-BGT1 and T40A without any treatment of the PKC activators PMA or DOG (Fig. 7B, wt-BGT1 and T40A). These findings strongly support the notion that the ⬃95-kDa immunoreactive band observed in the Western blots corresponds to a specific PKC phosphorylation at T40. Uptake of GABA was similar to wt-BGT1 when the threonine at position 40 was substituted by site-directed mutagenesis

A

GABA uptake (pmol/mg protein/10 min)

tested after expression in MDCK cell monolayers followed by a 24-h adaptation period to hypertonic medium. In the absence of PMA, EGFP-wt-BGT1 and mutants showed staining of the plasma membrane. This staining turned into punctuated intracellular staining upon 45-min exposure to PMA, consistent with endocytotic removal of the respective proteins from the plasma membrane (Fig. 5B, top and bottom). EGFP-wt-BGT1 and mutants were expressed in HEK-293 cells to demonstrate responses to hypertonicity and PMA. Nontransfected and vector-transfected HEK-293 cells showed only low uptake of GABA, which increased upon incubation in hypertonic medium (Fig. 6A), indicating that transfection with Fugene 6 had no significant effect on intrinsic GABA uptake. Under normotonic conditions, the uptake of GABA in T40A was similar to vector-transfected HEK-293 cells, i.e., 150.9 ⫾ 23.6 and 141.5 ⫾ 18.0 pmol·mg protein⫺1·10 min⫺1, respectively. Compared with nontransfected and vector-transfected HEK-293 cells, GABA uptake in wt-BGT1, T235A, and S418A was higher (895.9 ⫾ 290.9, 660.1 ⫾ 377.2, and 2,263.2 ⫾ 100.7 pmol·mg protein⫺1·10 min⫺1) and increased again upon adaption to hypertonicity by a factor of ⬃2.2– 6.4 (Fig. 6A, solid vs. shaded columns). The response to PMA was determined in HEK-293 cells immediately after the adaption of cells to a hypertonic medium for 16 h, and the uptake of GABA is expressed relative to that of EGFP-wt-BGT1 in the absence of PMA. In the absence of PMA, GABA uptake measured in T235A, S418A, and S564A was similar to that of wt-BGT1; the uptake of GABA in T40A was only 6.0 ⫾ 3.7% of that measured in wt-BGT1. A 30-min application of PMA (75 nM) reduced the uptake in wt-BGT1, T235A, S418A, and S564A by ⬃30%, whereas PMA had no impact on GABA uptake in T40A. These observations suggest that threonine at position T235 and serines at positions 418 and 564 did not interfere with PMA and that these sites may not be important sites for phosphorylation by PKC. The inhibitory action of PMA was substrate independent: after a 30-min incubation with PMA (75 nM), [3H]betaine uptake was reduced by 42.3 ⫾ 33.3% (P ⬍ 0.05), as observed in three independent experiments. A phosphospecific antibody against T40 was used to define PKC-specific interactions. This antibody detected phosphorylation by PKC (Fig. 7A). Western blots of whole cell lysates from HEK-293 cells expressing EGFP-wt-BGT1 and T40A adapted to hypertonic medium for 16 h and treated for 1 h with either PMA (75 nM) or DOG (5 ␮M) revealed, in both cases, a band at ⬃95 kDa, whereas no band at all was visible for T40A, indicating that EGFP-wt-BGT1 is phosphorylated by PKC, whereas T40A is not. To further confirm that the ⬃95kDa immunoreactive band observed in the Western blot from whole cell lysates does, in fact, represent a specific PKC interaction, immunoprecipitation experiments were carried out. Specifically, HEK-293 cells that had been transfected with EGFP-wt-BGT1 and T40A were switched to hypertonic medium for 16 h and treated again for 1 h with either PMA (75 nM) or DOG (5 ␮M). After membrane lysis, wt-BGT1 and T40A were immunoprecipitated with an antibody directed against PKC phosphorylated at T40, and this revealed that wt-BGT1, but not T40A, was coimmunoprecipitated when cells were pretreated with PMA or DOG (Fig. 7B). Interestingly, in all immunoprecipitation experiments, the ⬃95-kDa band was the only immunoreactive species that was consistently detectable by wt-BGT1 after treatment with either PMA

10000 isotonic hypertonic

8000

*

*

*

6000 4000 2000 0

* no tx vector wt-BGT1 T40A T235A S418A EGFP-tagged

B

GABA uptake (% of wt-BGT1 in the absence of PMA)

F112

200 - PMA + PMA

150

*

* *

100

*

50 n.s 0 wt-BGT1

T40A

T235A

S418A

S564A

Fig. 6. NaCl-dependent uptake of GABA in human embryonic kidney (HEK)293 cells transfected with wt-BGT1 and mutants. A: comparison of GABA uptake in wt-BGT1 and mutants. Cells were nontransfected (no tx), transiently transfected with the vector EGFP alone (vector), or transiently transfected with EGFP-wt-BGT1, EGFP-T40A, EGFP-T235A, or EGFP-S418A 48 h before the experiment begun. Cell batches were divided into two groups. One group was incubated in hypertonic medium for 16 h (shaded columns), whereas the other group remained in isotonic medium (solid columns). Afterward, the 10-min uptake of GABA in NaCl-containing medium was determined. Data are means ⫾ SD of 3 separate experiments. *Significant difference between uptake in isotonic versus hypertonic medium (P ⬍ 0.01). B: uptake of GABA after a 30-min incubation in the absence and presence of PMA (75 nM). The uptake of [3H]GABA in EFGP-wt-BGT1 and T235A, S418A, and S564A mutants was inhibited by PMA, whereas PMA showed no effect in T40A. Data are means ⫾ SD of 3 separate experiments. n.s., not significant. *P ⬍ 0.05.

AJP-Renal Physiol • doi:10.1152/ajprenal.00085.2014 • www.ajprenal.org

DISRUPTED TRAFFICKING OF BGT1 BY MUTATION AT T40

A

+ PMA + DOG kDa M wt-BGT1 T40A wt-BGT1 T40A 150

B kDa M

F113

T40A wt-BGT1 HEKvector +PMA +DOG +PMA +DOG +PMA +DOG wt-BGT1 T40A

150 100

100

75

75 50

50 37

37 25

25

Fig. 7. Phospho-specific Western blots. A: Western blots of whole HEK-293 cell lysates using a PKC phospho-specific antibody. After 16 h in hypertonic medium and treatment with either 75 nM PMA or 5 ␮M DOG for 1 h, EGFP-wt-BGT1 was phosphorylated by PKC but T40A was not. B: immunoprecipitation of EGFP-wt-BGT1 and T40A using an antibody recognizing PKC phosphorylated states in HEK-293 cells expressing wt-BGT1 and T40A. Immunoprecipitation experiments were carried out with lysates prepared from HEK-293 cells expressing wt-BGT1 and the T40A mutant after a 16-h adaption to hypertonic medium and treatment with either 75 nM PMA or 5 ␮M DOG for 1 h. As a control, HEK-293 cells were also transfected with EGFP vector alone (HEKvector) and additionally with wt-BGT1- and T40A-transfected cells not treated with PMA or DOG. Proteins were immunoprecipitated with a polyclonal-specific anti-PKC-phosphoantibody and then immunoblotted with EGFP antibody. In all immunoprecipitation experiments, the ⬃95-kDa band for wt-BGT1 after treatment with either PMA or DOG was the only immunoreactive band that was consistently detectable by the PKC-specific antibody. Data are representative of 2 independent experiments.

to glutamate (T40E) instead of alanine (T40A; Fig. 8). Uptake of GABA in wt-BGT1- and T40E-transfected HEK-293 cells revealed values of 584.5 ⫾ 22.5 and 573.2 ⫾ 247.3 pmol·mg protein⫺1·10 min⫺1, whereas uptake of GABA was 45.0 ⫾ 16.9 and 56.2 ⫾ 22.3 pmol·mg protein⫺1·10 min⫺1 in T40Aand vector-transfected HEK-293 cells. Confocal microscopy confirmed that T40E was localized in the plasma membrane (data not shown). DISCUSSION

GABA uptake (pmol/mg protein/10 min)

The combination of electrophysiological studies in the oocyte expression system with radiotracer uptake studies and immunolocalization of transfected proteins in different cell 1000 n.s. 800

*

600 400 200 0 vector

wt-BGT1

T40A

T40E

Fig. 8. Mimicking phosphorylation by substitution of threonine at position 40 to glutamate. The 10-min uptake of GABA was determined in HEK-293 cells transiently transfected with wt-BGT1, T40A, T40E, or vector. Data are means ⫾ SD of 3 separate experiments. *P ⬍ 0.01.

lines derived originally from the kidneys provided mutually supportive information and demonstrated that regulation of transporters seems, at least for BGT1, to be an intrinsic property of the transporter and independent of the expression system. The functional characteristics and kinetics of wt-BGT1 expressed in Xenopus oocytes, including the preference for GABA over betaine, are in agreement with a previous report (10). Furthermore, the inhibitory action of DOG on GABAassociated currents of wt-BGT1-expressing oocytes confirms results from our previous study (6) showing inhibition of GABA uptake in endogenous BGT1 in MDCK cells and suggests that the response to DOG is not unique for or restricted to MDCK cells but is an intrinsic property of BGT1. This is further supported by our previous observations (6) that in MDCK cells, the endogenous system A for amino acids was not sensitive to DOG, indicating a degree of specificity of MDCK cells in the response to DOG. According to the topology model (Fig. 1), the four potential canonic PKC phosphorylation sites face the cytoplasm. Three of them (T235A, S418A, and S564A) displayed GABA-dependent currents of similar magnitude as wt-BGT1 when expressed in Xenopus oocytes. When expressed as EGFP-BGT1 mutants in MDCK as well as in HEK-293 cells, their uptake of GABA was similar to wt-BGT1, indicating again independence of the expression system and of the methods used for studying transport. In addition, transfection with EGFP had no impact on GABA uptake, as seen by a comparison of the uptake in nontransfected and EGFP-transfected cells. Experiments on the localization of EGFP-BGT1 mutants in MDCK cells revealed localization in the plasma membrane during hypertonic stress and that internalization in response to PMA occurred likely by endocytosis. Inhibition of transport activity by PMA in HEK-293 cells is consistent with internalization, since substrate uptake is determined by the surface abundance

AJP-Renal Physiol • doi:10.1152/ajprenal.00085.2014 • www.ajprenal.org

F114

DISRUPTED TRAFFICKING OF BGT1 BY MUTATION AT T40

of the transport protein. All these observations suggest that introduction of mutations did not change the intrinsic transport activity of wt-BGT1 and EGFP-wt-BGT1 and that residues T235, S418, and S564 are not involved in the response to PKC. Compared with T235A, S418A, and S564A, the T40A mutant showed differences with respect to GABA-mediated currents in Xenopus oocytes and its localization in MDCK cells. In the oocyte expression system, GABA-induced currents in T40A are not comparable in magnitude to those observed in wt-BGT1. In addition, these currents were not sensitive to PMA. In MDCK cells adapted to hypertonic stress, EGFPtagged T40A was located in the trans-Golgi network, as evidenced by colocalization with furin. This explains why uptake of GABA in HEK-293 cells transfected transiently with T40A was small and why there was no measurable response to hypertonicity or PMA. A possible explanation for this unexpected behavior is that T40A may have trafficked normally to the plasma membrane during hypertonic stress but was not retained there, similar to what has been previously described for a truncated BGT1 mutant lacking the COOH terminus (9). This mutant was recovered in a recycling compartment of MDCK cells (9). T40A seems not to be truncated, as Western blots revealed a size similar to EGFP-wt-BGT1 and the other mutants without any degradation. An alternative possibility is that T40A may remain intracellular due to a trafficking defect or improper folding. The former may be unlikely because it has been previously established that signals for exit from the endoplasmic reticulum and sorting to the plasma membrane of MDCK cells are contained within the cytoplasmic COOH terminus of several proteins, including BGT1 (11, 13). In BGT1, this signaling information lies within a short segment of amino acids (565–572) rich in basic residues (15). The marked effect of alanine substitution at T40 in the cytoplasmic NH2 terminus might occur if the T40A mutant folds improperly at the NH2 terminus. This would likely result in the retention of T40A protein in the endoplasmic reticulum for degradation, but confocal microscopy revealed that this mutant was localized in the trans-Golgi, indicating that T40A might be faster internalized from the plasma membrane. In general, attachment of EGFP to the NH2 terminus does not change plasma membrane trafficking compared with wt-BGT1 (7). Finally, it should be considered that PKC-dependent phosphorylation of T40 may be required for normal trafficking, insertion, and/or retention of BGT1 at the basolateral plasma membrane of MDCK cells. Using an anti-phosphoantibody, we showed that T40A was not phosphorylated. This finding clearly demonstrated the importance of this threonine for correct plasma membrane trafficking and insertion. Substitution of the threonine by glutamic acid revealed a functional clone of BGT1 that was inserted properly into the plasma membrane, as seen by the similar uptake of GABA in the T40E mutant compared with wt-BGT1. In summary, BGT1 expression occurs at several levels, from the regulation of gene expression by the transcription factor tonicity-responsive enhancer binding protein in response to hypertonicity (2, 8) to the more rapid posttranslational regulation of plasma membrane abundance via vesicular insertion (7) and retrieval (4 – 6). Very recently, several other serine/threonine kinases affecting the response of BGT1 to osmotic stress have been identified. These are AMP-activated protein kinase

(12), JAK2 (3), and tau tubulin kinase 2 (1). AMP-activated protein kinase senses the cytosolic AMP-to-ATP concentration ratio and supports our finding of regulation of BGT1 by ATP (5). Apart from T612 (9), the function of specific residues within BGT1 has barely been explored, and an additional layer of complexity may be uncovered as the protein structurefunction relationships emerge. ACKNOWLEDGMENTS The authors gratefully acknowledge the technical assistance of Jeffrey S. Bonner. The authors thank Dr. G. Burckhardt, Dr. A. N. Rizwan, Dr.Cristina Fenollar-Ferrer and Dr. S. Rhodes for assistance and helpful discussions. GRANTS This work was supported by an American Heart Association, Midwest Affiliate, grant-in-aid, by an Indiana University-Purdue University Indianapolis research support funds grant (to S. A. Kempson), and by the German Research Foundation (to C. Ziegler). DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the author(s). AUTHOR CONTRIBUTIONS Author contributions: E.S.S., S.A.K., C.Z., and B.C.B. conception and design of research; E.S.S., S.A.K., and B.C.B. performed experiments; E.S.S., S.A.K., C.Z., and B.C.B. analyzed data; E.S.S., S.A.K., C.Z., and B.C.B. interpreted results of experiments; E.S.S., S.A.K., and B.C.B. prepared figures; E.S.S., S.A.K., and B.C.B. drafted manuscript; E.S.S., S.A.K., and B.C.B. edited and revised manuscript; B.C.B. approved final version of manuscript. REFERENCES 1. Almiaji A, Munoz C, Hosseinzadeh Z, Lang F. Upregulation of Na⫹, Cl⫺-coupled betaine/␥-amino-butyric acid transporter BGT1 by tau tubulin kinase 2. Cell Physiol Biochem 32: 334 –343, 2013. 2. Burg MB, Ferraris JD. Intracellular organic osmolytes: function and regulation. J Biol Chem 283: 7309 –7313, 2008. 3. Hosseinzadeh Z, Shojaiefard M, Bhavsar SK, Lang F. Up-regulation of the betaine/GABA transporter BGT1 by JAK2. Biochem Biophys Res Commun 420: 172–177, 2012. 4. Kempson SA, Beck JA, Lammers PE, Gens JS, Montrose MH. Membrane insertion of betaine/GABA transporter during hypertonic stress correlates with nuclear accumulation of TonEBP. Biochim Biophys Acta 1712: 71–80, 2005. 5. Kempson SA, Edwards JM, Osborn M, Sturek M. Acute inhibition of the betaine transporter by ATP and adenosine in renal MDCK cells. Am J Physiol Renal Physiol 295: F108 –F117, 2008. 6. Kempson SA, Edwards JM, Sturek M. Inhibition of the renal betaine transporter by calcium ions. Am J Physiol Renal Physiol 291: F305–F313, 2006. 7. Kempson SA, Parikh V, Xi L, Chu S, Montrose M. Subcellular distribution of the renal betaine transporter during hypertonic stress. Am J Physiol Cell Physiol 285: C1091–C1100, 2003. 8. Kwon MS, Lim SW, Kwon HM. Hypertonic stress in the kidney: a necessary evil. Physiology 24: 186 –191, 2009. 9. Massari S, Vanoni C, Longhi R, Rosa P, Pietini G. Protein kinase C-mediated phosphorylation of the BGT1 epithelial ␥-aminobutyric acid tansporter regulates its association with LIN7 PDZ proteins: a posttranslational mechanism regulating transporter surface density. J Biol Chem 280: 7388 –7397, 2005. 10. Matskevitch I, Wagner CA, Stegen C, Broer S, Noll B, Risler T, Kwon HM, Handler JS, Waldegger S, Busch A, Lang F. Functional characterization of the betaine/GABA transporter BGT-1 expressed in Xenopus oocytes. J Biol Chem 274: 16709 –16716, 1999. 11. Matter K, Yamamoto EM, Mellman I. Structural requirements and sequence motifs for polarized sorting and endocytosis of LDL and Fc receptors in MDCK cells. J Cell Biol 126: 991–1004, 1994. 12. Munoz C, Sopjani M, Dermaku-Sopjani M, Almilaji A, Föller M, Lang F. Downregulation of the osmolyte transporters SMIT and BGT1 by

AJP-Renal Physiol • doi:10.1152/ajprenal.00085.2014 • www.ajprenal.org

DISRUPTED TRAFFICKING OF BGT1 BY MUTATION AT T40

13.

14. 15.

16.

17.

AMP-activated protein kinase. Biochem Biophys Res Commun 422: 358 – 362, 2012. Muth TR, Ahn J, Caplan MJ. Identification of sorting determinants in the C-terminal cytoplasmic tails of the gamma-aminobutyric acid transporters GAT-2 and GAT-3. J Biol Chem 273: 25616 –25627, 1998. Neuhofer W, Beck FX. Cell survival in the hostile environment of the renal medulla. Annu Rev Physiol 67: 531–555, 2005. Perego C, Bulbarelli A, Longhi R, Caimi M, Villa A, Caplan MJ, Pietrini G. Sorting of two polytopic proteins, the GABA and betaine transporters, in polarized epithelial cells. J Biol Chem 272: 6584 –6591, 1997. Pramod AB, Foster J, Carvelli L, Henry LK. SLC6 transporters: structure, function, regulation, disease association and therapeutics. Mol Aspect Med 34: 197–219, 2013. Preston A, Yamauchi A, Kwon H, Handler J. Activators of protein kinase A and protein kinase C inhibit MDCK cell myo-inositol and betaine uptake. J Am Soc Nephrol 6: 1559 –1564, 1995.

F115

18. Rudnick G, Krämer R, Blakely RD, Murphy DL, Verrey F. The SLC6 transporters: perspectives on structure, functions, regulations, and models for transporter dysfunction. Pflügers Arch 466: 25–42, 2014. 19. Teuchert M, Berghofer S, Klenk HD, Garten W. Recycling of furin from the plasma membrane. Functional importance of the cytoplasmic tail sorting signals and interaction with the AP-2 adaptor medium chain subunit. J Biol Chem 274: 36781–36789, 1999. 20. Yamauchi A, Uchida S, Kwon HM, Preston AS, Robey RB, Garcia-Perez A, Burg MB, Handler JS. Cloning of a Na⫹- and Cl⫺-dependent betaine transporter that is regulated by hypertonicity. J Biol Chem 267: 649–652, 1992. 21. Yamashita A, Singh SK, Kawate T, Jin Y, Gouaux E. Crystal structure of a bacterial homologue of Na⫹/Cl⫺-dependent neurotransmitter transporters. Nature 437: 215–223, 2005. 22. Zhou Y, Holmseth S, Hua R, Lehre AC, Olofsson AM, Poblete-Naredo I, Kempson SA, Danboldt NC. The betaine-GABA transporter (BGT1, slc6a12) is predominantly expressed in the liver and at lower levels in the kidneys and at the brain surface. Am J Physiol Renal Physiol 302: F316–F328, 2012.

AJP-Renal Physiol • doi:10.1152/ajprenal.00085.2014 • www.ajprenal.org

Copyright of American Journal of Physiology: Renal Physiology is the property of American Physiological Society and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.