Abstract
Attention deficit/hyperactivity disorder (ADHD) is the most commonly diagnosed disorder of school-age children. Although genetic and brain-imaging studies suggest a contribution of altered dopamine (DA) signaling in ADHD, evidence of signaling perturbations contributing to risk is largely circumstantial. The presynaptic, cocaine- and amphetamine (AMPH)-sensitive DA transporter (DAT) constrains DA availability at presynaptic and postsynaptic receptors following vesicular release and is targeted by the most commonly prescribed ADHD therapeutics. Using polymorphism discovery approaches with an ADHD cohort, we identified a hDAT (human DAT) coding variant, R615C, located in the distal C terminus of the transporter, a region previously implicated in constitutive and regulated transporter trafficking. Here, we demonstrate that, whereas wild-type DAT proteins traffic in a highly regulated manner, DAT 615C proteins recycle constitutively and demonstrate insensitivity to the endocytic effects of AMPH and PKC (protein kinase C) activation. The disrupted regulation of DAT 615C parallels a redistribution of the transporter variant away from GM1 ganglioside- and flotillin1-enriched membranes, and is accompanied by altered CaMKII (calcium/calmodulin-dependent protein kinase II) and flotillin-1 interactions. Using C-terminal peptides derived from wild-type DAT and the R615C variant, we establish that the DAT 615C C terminus can act dominantly to preclude AMPH regulation of wild-type DAT. Mutagenesis of DAT C-terminal sequences suggests that phosphorylation of T613 may be important in sorting DAT between constitutive and regulated pathways. Together, our studies support a coupling of DAT microdomain localization with transporter regulation and provide evidence of perturbed DAT activity and DA signaling as a risk determinant for ADHD.
Introduction
The neurotransmitter dopamine (DA) provides critical modulatory influences over circuits subserving reward, locomotor activity, and attention (Carlsson, 1987; Robbins, 2003). As such, alterations in DA signaling contribute to multiple neurological and psychiatric disorders including Parkinson's disease (Chase et al., 1998), attention deficit/hyperactivity disorder (ADHD) (Mazei-Robison et al., 2005), and addiction (Ritz et al., 1987). The reuptake of DA through presynaptic DA transporters (DATs) is a primary mechanism for terminating DA action at presynaptic and postsynaptic receptors and is a major target for psychostimulants, such as cocaine and amphetamine (AMPH).
Multiple studies point to a contribution of variation in the genes encoding DAT, COMT (catechol-O-methyl transferase), and/or DA receptors as influencing risk for ADHD (Gill et al., 1997; Qian et al., 2003; Bobb et al., 2005; Mazei-Robison et al., 2005), the most commonly diagnosed disorder of school age children in the United States. A further link between DAT function and ADHD is suggested from the therapeutic utility of DAT-interacting psychostimulants, including methylphenidate (Ritalin) and AMPH preparations (e.g., Adderall). Brain-imaging studies also point to deficits in DA signaling as a key feature of ADHD (Swanson et al., 2007). Genetic elimination of DAT expression in mice reduces presynaptic DA stores, elevates extracellular DA, and produces hyperactivity in a novel environment (Giros et al., 1996). However, humans that are homozygous for loss-of-function DAT (SLC6A3) alleles exhibit infantile neonatal dystonia rather than ADHD (Kurian et al., 2009), raising questions as to the relevance of compromised DAT function and DA signaling as a key feature in ADHD.
Reasoning that genetic alterations that perturb DAT regulation, as opposed to DAT elimination, could reconcile data from rodent and human studies, we screened ADHD subjects for rare variation producing coding variation in DAT. Previously, we described the anomalous, nonvesicular release of DA by the DAT A559V coding variant (Mazei-Robison et al., 2008), leading to the hypothesis that dysregulated DA availability may be a determinant of risk for ADHD. Here, we provide additional evidence for this idea through study of the properties of a second DAT variant, R615C. Our studies reveal novel contributions of the DAT C terminus in orchestrating the membrane compartmentalization, trafficking pathways, and protein associations of the transporter. The functional impact of the DAT R615C variant on DAT localization and trafficking encourages further evaluation of perturbations of DAT regulatory networks in ADHD.
Materials and Methods
Materials
DMEM and FBS were from Invitrogen. Mouse anti-calcium/calmodulin-dependent protein kinase II (CaMKII) antibody (MA1-048), tris(2-carboxyethyl)phosphine (TCEP), sulfosuccinimidyl1-2-(biotinamido)ethyl-1,3-dithioproprionate-biotin (Sulfo-NHS-SS-Biotin), and streptavidin agarose resin were from Thermo Fisher Scientific. [3H]DA and [32P]orthophosphoric acid were from PerkinElmer. Alexa Fluor 647 cholera toxin B (CTxB) was from Invitrogen. 2-[N-(2-Hydroxyethyl)]-N-(4-methoxybenzenesulfonyl)]amino-N-(4-chlorocinnamyl)-N-methylbenzylamine) (KN-93), 2-[N-(4-methoxybenzenesulfonyl)]amino-N-(4-chlorocinnamyl)-N-methylbenzylamine, phosphate (KN-92), and β-PMA were from EMD Biosciences. Rat and rabbit anti-hDAT antibodies were from Millipore (MAB369 and AB5803, respectively). Mouse anti-flotillin-1 antibody was from BD Biosciences (catalog #610820). Secondary antibodies were obtained from Jackson ImmunoResearch Laboratories. Other reagents were obtained from Sigma-Aldrich.
ADHD subject collection and ascertainment
Subject recruitment and ascertainment for United States ADHD probands has been previously described (Mazei-Robison et al., 2005, 2008). Additional subjects from Ireland were also enrolled from child guidance clinics and ADHD support groups around Ireland. The age range of the probands was between 5 and 14 years, with males accounting for 88% of the cases. To establish Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) diagnoses, one or both parents were interviewed using the Child and Adolescent Psychiatric Assessment (CAPA). To fulfill DSM-IV ADHD criteria for symptom pervasiveness, information regarding ADHD symptoms at school was obtained from teachers using a semistructured teacher telephone interview. ADHD symptom dimensions, as well as comorbid disorders such as ODD (oppositional defiant disorder) and CD (conduct disorder) were obtained from the CAPA. A positive family history of ADHD was defined by at least one parent scoring 36 or greater on the 25-item subscale of the WURS (Wender Utah Rating Scale). A 25-item subscale was used in with a cutoff score of 36 or greater being 96% sensitive and specific for a retrospective diagnosis of childhood ADHD in the parent.
PCR amplification of hDAT exons and polymorphisms screening via temperature gradient capillary electrophoresis
PCR amplification, screening, and sequence confirmation were performed as described previously (Mazei-Robison et al., 2005).
Cell culture, transfections, and stable cell line generation
Flp-In 293 cells were used to generate stable lines expressing either DAT 615R or DAT 615C following the manufacturer's protocol (Invitrogen). Selection media contained 100 μg/ml Hygromycin B. HEK 293T cells used for transient transfections were cultured, maintained, and transfected as described previously (Mazei-Robison and Blakely, 2005).
DA transport assays
[3H]DA transport assays were performed in triplicate or quadruplicate in 24-well plates as described previously (Apparsundaram et al., 1998). Briefly, cells were washed with Krebs'–Ringer's–HEPES (KRH) (130 mm NaCl, 1.3 mm KCl, 2.2 mm CaCl2, 1.2 mm MgSO4, 1.2 mm KH2PO4, 10 mm HEPES, pH 7.4) buffer and were incubated in the uptake assay buffer (KRH, 10 mm glucose, 100 μm pargyline, ascorbic acid, and 10 μm tropolone) before incubation with [3H]DA for 10 min followed by three 4°C washes. To determine nonspecific uptake, parallel wells were incubated with 1 μm 1-[2-[bis(4-fluorophenyl)methoxy]ethyl]-4-(3-phenylpropyl)piperazine hydrochloride (GBR 12909) 5 min before addition of [3H]DA. For single point uptake assays, 50 nm [3H]DA was used. For saturation kinetic analyses, six different DA concentrations were used (0.05, 0.1, 0.5, 1, 2, and 3 μm, or 0.05, 0.1, 0.5, 1, 3, and 6 μm). Serial dilution from a stock made with 5% [3H]DA was used for saturation analyses, except for the 50 nm concentration where 100% tritiated substrate was used. For all inhibitors, six concentrations (0.001, 0.01, 0.1, 0.5, 1, 10 μm) were tested to determine IC50 values. With AMPH treatments, cells were rapidly washed with ice-cold KRH buffer and warmed back up to 37°C using uptake assay buffer for 5 min before addition of [3H]DA for a further 3 min incubation. The shorter uptake time was used to reduce confounds from DA efflux due to the presence of intracellular AMPH. MicroScint scintillation mixture was added to the wells at the end of the assay and DA uptake was measured using a TopCount Scintillation Counter. Mean ± SEM values reported derive from three to six independent experiments.
Cell surface biotinylation, biotinylation internalization, and biotinylation recycling assays
Cells were seeded in six-well plates and incubated for 36–48 h before experiments. Cells were washed twice with warm KRH buffer before drug treatments. At the end of the treatment time, cells were placed on ice and washed rapidly twice with ice-cold PBS (136 mm NaCl, 2.5 mm KCl, 1.5 mm KH2PO4, 6.5 mm Na2PO4, 2.8 mm glucose) containing 0.1 mm CaCl2 and 1 mm MgSO4. Cells were incubated with Sulfo-NHS-SS-Biotin for 30 min at 4°C. Excess biotin was quenched by two washes with 0.1 m glycine in PBS, and cells were solubilized using radioimmunopreciptation assay (RIPA) (100 mm Tris, pH 7.4, 150 mm NaCl, 1 mm EDTA, 0.1% SDS, 1% Triton X-100, 1% sodium deoxycholate) buffer for 30 min at 4°C.
To determine the internalization rate, we used a modified version of a previously published protocol (Holton et al., 2005). Briefly, cells seeded on two different plates were labeled with Sulfo-NHS-SS-Biotin for 30 min at 4°C and quenched with 0.1 m glycine in PBS. One plate was mainlined at 4°C to determine total surface expression and stripping efficiency. The other plate was warmed by three rapid washes with warm PBS, and cells were incubated with warm PBS at 37°C for 30 min either with or without AMPH to allow trafficking of the surface proteins. At the end of incubation, plates were transferred onto ice and cells were washed three times with ice-cold PBS followed by three washes with NT buffer (150 mm NaCl, 1 mm EDTA, 0.2% BSA, 20 mm Tris, pH 8.6). The biotin signal remaining on the surface from the 37°C plates and stripping control wells from 4°C plates were stripped using 50 mm TCEP in NT buffer for 60 min at 4°C. Fresh stripping solution was added after the first 30 min to ensure optimal stripping efficiency. Cells were washed rapidly three times with ice-cold NT buffer followed by ice-cold PBS and solubilized using RIPA buffer for 30 min at 4°C. Protein concentration was determined using the BCA protein assay and equal protein amounts were incubated with prewashed streptavidin beads for 2 h or overnight at 4°C. Streptavidin beads were washed three times with RIPA buffer and bound proteins were eluted using LSB (Laemmli sample buffer) and analyzed using SDS-PAGE, followed by Western blotting. A rat anti-hDAT antibody (Millipore; MAB 369; 1:5000) was used to visualize DAT and mouse anti-β-actin (Sigma-Aldrich; a5316; 1:5000) antibody was used to detect β-actin as a loading control. HRP-conjugated goat anti-rat and mouse secondary antibodies (1:10,000) were used.
Recycling rates were estimated using a biotinylation-recycling assay. Cells were labeled with Sulfo-NHS-SS-Biotin for 60 min at 37°C to achieve labeling of both surface and recycling intracellular proteins. Cells were placed on ice and cooled by washing with ice-cold PBS, followed by NT buffer three times each. Surface biotin labeling was stripped using 50 mm TCEP in NT buffer for 90 min at 4°C. After an initial 60 min of stripping, fresh TCEP solution was added for the remaining 30 min to ensure efficient stripping, assessed via Western blots. Following stripping, cells were washed rapidly three times with ice-cold NT buffer followed by three times with ice-cold PBS buffer. One plate was maintained at 4°C, and the other plate was warmed by washing three times with 37°C PBS. Cells were then incubated for 30 min either with or without AMPH at 37°C. At the end of the incubation, cells were again placed on ice and washed three times with ice-cold PBS followed by NT buffer. Biotin signal from intracellular DAT that recycled to the surface was stripped by incubating cells in 50 mm TCEP in NT buffer on ice for 60 min, with fresh stripping buffer added after 30 min. Cells were then washed three times with ice-cold NT buffer followed by three washes with ice-cold PBS buffer and solubilized using RIPA buffer for 30 min at 4°C. Streptavidin pulldown and Western blots were performed as described above.
Coimmunoprecipitation assays
DAT-CaMKII coimmunoprecipitation.
HEK 293T cells were seeded on 10 cm dishes and cotransfected with pEYFP-HA-hDAT-WT (615R) (gift from Dr. Alexander Sorkin, University of Pittsburgh, Pittsburgh, PA) or the pEYFP-HA-hDAT 615C variant and CaMKIIα (gift from Dr. Roger Colbran, Vanderbilt University, Nashville, TN) 24 h later. The pEYFP-HA-hDAT-WT construct has an N-terminal YFP tag and the HA tag derives from modifications of extracellular loop 2, as described previously (Sorkina et al., 2003, 2006). The presence of YFP- or HA-tag does not interfere with expression, function, and trafficking of DAT (Sorkina et al., 2006). Cells were lysed using coimmunoprecipitation (co-IP) buffer (50 mm Tris, pH 7.4, 150 mm NaCl, 1% Triton X-100) containing protease inhibitors 48 h after transfection. Immunoprecipitation was performed overnight using goat anti-CaMKII antibody (gift from Dr. Roger Colbran), and coimmunoprecipitated DAT was visualized using rat anti-hDAT antibody as for Western blots noted above. Mouse anti-CaMKII antibody (1:5000) was used to visualize immunoprecipitated CaMKII. Goat IgG was used in parallel co-IP experiments to insure specificity of immunoprecipitation.
DAT-flotillin-1 co-IP.
Flp-in HEK stable cells were seeded on 10 cm dishes. Cells were lysed using co-IP buffer 48 h after plating. Immunoprecipitation was done using rat anti-hDAT antibody, and coprecipitated flotillin-1 was visualized using a mouse anti-flotillin-1 antibody (1:1000). Rabbit anti-hDAT antibody (1:5000) was used to visualize immunoprecipitated DAT. Extracts from Flp-In HEK parent line were used as a negative control.
Metabolic labeling to assess DAT phosphorylation
[32P]Orthophosphate metabolic labeling was performed as described previously (Bowton et al., 2010). Briefly, Flp-In HEK stable cells were seeded on six-well plates and cultured for 48 h. Cells were washed once with phosphate-free DMEM and incubated in the same medium for 1 h pool at 37°C to deplete the intracellular ATP. Cells were then supplemented with 1 mCi/ml [32P]orthophosphoric acid and incubated for 4 h at 37°C. Cells were then washed three times with ice-cold PBS and lysed with RIPA buffer containing protease inhibitors and 1 μm okadaic acid and microcystin each for 30 min at 4°C. Lysates were immunoprecipitated with rat anti-DAT antibody overnight at 4°C. Immunoprecipitated DATs were visualized using SDS-PAGE, followed by autoradiography and densitometry.
CTxB labeling and confocal microscopy
HEK 293T cells were plated on glass-bottom MatTek plates. Cells were transfected with pEYFP-HA-hDAT-WT (615R) or pEYFP-HA-hDAT 615C 24 h later and cultured for 48 h. Cells were washed three times with ice-cold PBS containing 0.1 mm CaCl2 and 1 mm MgSO4 and incubated with 1 μg/ml Alexa 647-conjugated CTxB for 30 min at 4°C. Cells were washed three times with ice-cold PBS and fixed using 4% paraformaldehyde for 10 min at room temperature. Fixed cells were washed with PBS and stored in PBS until imaging at 4°C. Confocal imaging was performed in the Vanderbilt University Medical Center Cell Imaging Shared Resource. Laser output was adjusted to assure that all fluorescence monitored was nonsaturating with respect to fluorophore emission. Intensity correlation quotient (ICQ) to assess colocalization of DAT with CTxB-labeled membranes was obtained as previously described (Steiner et al., 2009; Baucum et al., 2010) using the Integration Colocalization Analysis section of WCIF ImageJ (http://www.uhnresearch.ca/facilities/wcif/imagej/). This analysis computes a correlation in the intensity of two probes across pixels in regions of interest, with ICQ values distributed between −0.5 and +0.5. ICQ values that report evidence for colocalization fall between 0 and +0.5 and are evaluated for significance using a one-sample t test comparing against a zero ICQ value. Genotype differences in ICQ values were determined by a two-tailed, Student's t test.
Detection of AMPH using HPLC
Flp-In HEK cells were seeded in six-well plates and incubated for 36–48 h before experiments. Cells were washed twice with warm KRH buffer before incubation with 10 μm AMPH for 5 min at 37°C. Cells were rapidly washed three times using ice-cold KRH buffer and lysed using 10 mm Tris/1 mm EDTA, pH 8.0, buffer (TE) containing protease inhibitors. Total protein concentration was determined using BCA protein assay and used to normalize the AMPH uptake. AMPH uptake was determined using HPLC analysis, and data are expressed as nanomoles of AMPH transported per microgram of protein.
Amperometry
Unpatched amperometric currents were recorded as previously described (Bowton et al., 2010) using Flp-In HEK stable cells. Briefly, cells were plated at a density of 103 per 35 mm culture dish. To preload cells with DA, attached cells were washed with KRH assay buffer containing 10 mm d-glucose, 100 μm pargyline, 1 mm tropolone, and 100 μm ascorbic acid, and then incubated with 1 μm DA in assay buffer for 20 min at 37°C. Dishes containing DA-loaded cells were then washed three times with external solution (130 mm NaCl, 10 mm HEPES, 34 mm d-glucose, 1.5 mm CaCl2, 0.5 mm MgSO4, 1.3 mm KH2PO4, adjusted pH to 7.35, and 300 mOsm). A carbon fiber electrode (ProCFE; fiber diameter of 5 μm; obtained from Dagan Corporation), juxtaposed to the plasma membrane and held at +700 mV (a potential greater than the oxidation potential of DA), was used to monitor basal and AMPH-evoked DA efflux through DAT as a consequence of DA oxidation. To determine basal DAT-dependent efflux, cells were treated with 10 μm cocaine following establishment of a stable recording baseline. Cells were not voltage clamped in basal or AMPH (10 μm)-evoked DA efflux experiments. Amperometric currents in response to an addition of AMPH were recorded using an Axopatch 200B amplifier (Molecular Devices) with a low-pass Bessel filter set at 1 kHz. Traces were digitally filtered offline at 1 Hz using Clampex9 software (Molecular Devices). DA efflux was quantified as the mean peak amperometric current (in picoamperes) ± SEM.
Quantification and statistics
Western blots were quantified using NIH ImageJ software, with multiple exposures taken to insure linearity of signal detection. Student's t test or one-way ANOVA with Bonferroni's post hoc test was used wherever appropriate. GraphPad Prism was used to determine Vmax and Km for saturation kinetic analyses and IC50 for inhibition analyses. Significant differences between DAT 615R and 615C values of Vmax, Km, and IC50 were determined using Student's t test, and a p value of 0.05 was considered as evidence of significance.
Results
Identification of a functional DAT coding variant in an ADHD subject
Using polymorphism discovery methods (Li et al., 2002), we screened coding exons and intron–exon junctions of the SLC6A3 gene in 417 ADHD subjects. Although a number of variants have been identified that alter the coding of DAT, all known variants are rare (allele frequency, <1%) (Mazei-Robison et al., 2005). In an Irish cohort of 100 subjects receiving a DSM-IV diagnosis of ADHD (Bellgrove et al., 2009), we identified a nonsynonymous, single-nucleotide polymorphism (2026 T/C) that converts a highly conserved Arg residue at position 615 to Cys (R615C). R615 is completely conserved in mammalian DATs and resides in a generally well conserved region of the DAT C terminus (Fig. 1A), a region implicated in transporter somatic export, surface trafficking, and protein–protein interactions (Torres et al., 2001; Bjerggaard et al., 2004; Fog et al., 2006). The subject, a Caucasian male of European origin, 13 years of age at the time of assessment, had a strong clinical and research diagnosis of combined type ADHD (Connor's parent rating scale-inattentive symptoms, T = 70; hyperactive symptoms, T = 75; total score, 74; WISK-III UK IQ, 141) and was found to be heterozygous for the R615C variant. Pedigree genotyping revealed that the mutation was transmitted from the proband's mother who is retrospectively suspected to have suffered from ADHD as a child (Wender Utah, 59), although no clinical diagnosis is currently available. The subject's father and two older sisters who do not carry the variant are unaffected (Fig. 1B). The R615C proband has been successfully treated with methylphenidate.
Reduced surface expression and DA transport Vmax associated with the R615C variant. A, Sequence alignment showing conservation of arginine 615 across mammalian DATs. B, Pedigree of ADHD proband carrying the R615C variant. The subject (black box) was heterozygous for the R615C-DAT variant. Transmission of the mutant allele occurred from mother (gray circle). Other family members (father and two older sisters) were unaffected. C, DA transport Vmax and Km are expressed as picomoles · microgram protein−1 · minute−1 and micromolar, respectively. Km and Vmax values are expressed as ±SEM. A significant decrease in the DA transport Vmax was observed for DAT 615C without a significant change Km. D, Left, Representative Western blot showing reduction in the surface and total DAT for the R615C variant compared with 615R. Total protein levels of β-actin, an intracellular loading control, were similar for both 615R and 615C. Right, Quantification from six independent experiments indicating significant reduction in the R615C variant compared with 615R surface and total expression. Total protein samples were diluted fivefold compared with the surface protein fraction to retain signal linearity. Error bars indicate SEM. Asterisks indicate statistically significant difference.
To search for functional evidence that the R615C variant may be causal in the ADHD of the proband, we compared the activities of DAT 615R and DAT 615C in stably transfected HEK cells. To limit confounds associated with different sites of integration, we created stable Flp-In HEK lines that express either of the two transporters from a common locus, under the control of the same promoter (Sauer, 1994). Saturation kinetic analysis revealed a significant reduction in DA transport Vmax for the R615C variant (615R, 1.6 ± 0.2, vs 615C, 1.0 ± 0.1 pmol · μg protein−1 · min−1; p < 0.005, two-tailed Student's t test) without a significant change in DA Km (1.3 ± 0.4 vs 1.0 ± 0.3 μm) (Fig. 1C). The reduced DA transport capacity was associated with a reduction in steady-state DAT protein levels (86 ± 3% of DAT 615R; n = 6; p < 0.005, two-tailed Student's t test) and diminished surface transporter expression (50 ± 7% of DAT 615R; n = 6; p < 0.0001, two-tailed Student's t test) (Fig. 1D), the latter determined by whole-cell biotinylation. However, we observed no significant differences in the IC50 values for cocaine (615R, 7.7 ± 2.0, vs 615C, 10.0 ± 2.1 nm), GBR 12909 (615R, 9.7 ± 3.0, vs 615C, 10.9 ± 2.0 nm), or AMPH (615R, 329 ± 135, vs 615C, 226 ± 101 nm), and a small, but statistically significant loss of potency for methylphenidate (615R, 585 ± 290, vs 615C, 821 ± 160 nm; p < 0.05, two-tailed Student's t test). Thus, although surface levels and DA transport capacity are reduced, the R615C variant exhibits normal interactions with DA and DAT antagonists.
Anomalous modulation of DAT 615C by AMPH
AMPH produces two alterations in DAT-expressing cells (Sulzer et al., 2005). First, AMPH induces (sec-min) nonvesicular DA release (efflux), mediated by reverse transport of cytoplasmic DA supported by CaMKII-mediated phosphorylation of the DAT N terminus (Khoshbouei et al., 2003; Fog et al., 2006). Second, AMPH treatment of DAT-expressing cells has been reported to produce both a rapid translocation of the transporter to the cell surface (Furman et al., 2009) and, with more prolonged treatment (10–30 min), produces net transporter internalization (Saunders et al., 2000; Chen et al., 2010). Using carbon fiber amperometry (Bowton et al., 2010), we monitored basal efflux from DAT 615R and DAT 615C cells preloaded with DA (data not shown), and found no significant difference. We also monitored AMPH-triggered DA efflux from DA-preloaded cells, and, as shown in Figure 2A, 615R- and 615C-expressing cells exhibited comparable AMPH-evoked DA efflux, measured using the peak amplitude of amperometric responses. This finding was surprising due to the reduced steady-state DAT surface expression of the R615C variant noted above. We found a possible answer to this conundrum in examining the impact of AMPH on transporter surface expression. Whereas treatment of 615R-expressing cells with 10 μm AMPH for 30 min produced the expected reduction in transporter surface expression, without any change total DAT protein levels (Fig. 2B), the same treatment failed to reduce surface expression of DAT 615C. Consistent with these findings, AMPH reduced DA transport from DAT 615R cells, but not from DAT 615C cells (Fig. 2C). Equivalent findings were also obtained following transient transfection of DAT 615R or DAT 615C into HEK 293T cells (data not shown) or CAD cells, a catecholamine-producing neuronal cell line (Qi et al., 1997) (Fig. 2D,E).
Anomalous modulation of DAT 615C by AMPH. A, Flp-In 293 stable cells were preloaded with 1 μm DA for 20 min and AMPH-induced DA efflux was measured using carbon fiber amperometry. AMPH-induced DA efflux from DAT 615C was not significantly different from 615R (1.0 ± 0.22 pA for 615R vs 0.996 ± 0.18 pA for 615C). B, Top, Representative Western blot showing surface DAT protein levels after AMPH treatment (10 μm; 30 min) for DAT 615C were significantly greater than AMPH-treated 615R surface levels. Total β-actin is used as an intracellular loading control. Bottom, Quantification of AMPH-induced decrease in DAT surface levels expressed as surface/total. Data are normalized to basal DAT surface/total expressions as 100%. AMPH treatment caused a significant (62 ± 7%) decrease in the DAT 615R expression, while no change was detected in 615C expression (95 ± 8%). Significant differences were calculated using one-way ANOVA followed by Bonferroni's post hoc test (n = 7; p < 0.0001). C, AMPH-mediated (10 μm; 30 min) reduction in the DA transport seen in DAT 615R (white bar; AMPH) is lost for 615C (black bar; AMPH). Basal DA transport levels were normalized to respective cell lines and AMPH-mediated reduction was measured. AMPH caused a significant decrease (66 ± 5% of basal) in the DA transport of DAT 615R, while no significant changes were detected for 615C (93 ± 6% of basal). Significant differences were calculated using one-way ANOVA followed by Bonferroni's post hoc test (n = 3; p < 0.001). D, Quantification of AMPH-induced decrease in DAT surface levels expressed as surface/total from transiently transfected CAD cells. Data are normalized to basal DAT surface/total expressions as 100%. AMPH treatment caused a significant (63 ± 6%) decrease in the DAT 615R expression, while no change was detected in 615C expression (104 ± 9%). Significant differences were calculated using one-way ANOVA followed by Bonferroni's post hoc test (n = 4; p < 0.0001). E, Transiently transfected, neuronally derived CAD cells display similar loss of AMPH-mediated reduction in DA transport seen in Flp-In HEK stable cells lines. Basal DA transport levels were normalized to respective cell lines, and AMPH-mediated reduction was measured. AMPH caused a significant decrease (72 ± 4% of basal) in the DA transport of DAT 615R, while no significant changes were detected for 615C (93 ± 9% of basal). Significant differences were calculated using one-way ANOVA followed by Bonferroni's post hoc test (n = 4; p < 0.01). Error bars indicate SEM. Asterisks indicate statistically significant difference.
DAT internalization is known to occur following protein kinase C (PKC) activation via phorbol esters, such as β-PMA (Loder and Melikian, 2003). We also found that the DAT 615C displayed no net endocytosis (109 ± 9% of vehicle-treated cells) and displayed no reduction in DA transport activity (98 ± 2% of control), following β-PMA treatment (200 nm, 30 min). However, similar β-PMA treatment caused internalization of DAT 615R (72 ± 4% of vehicle-treated cells; p < 0.005, two-tailed Student's t test) and a significant reduction in DA transport activity (61 ± 6% of control; p < 0.001, two-tailed Student's t test).
AMPH modulates DAT trafficking upon entering the cell, as demonstrated by studies of a transport-incompetent DAT mutant where intracellular injection of AMPH causes transporter internalization (Kahlig et al., 2006). Possibly, the R615C variant could transport AMPH more efficiently than wild-type DA, leading to the appearance of equivalent DA efflux despite reduced surface expression. However, HPLC assays of DAT 615C cells treated with AMPH under the same conditions used for efflux assays revealed similar levels of intracellular AMPH for DAT 615R and DAT 615C cells (615R, 4.9 ± 1.5 nm, vs 615 Cm, 5.1 ± 1.5 nm, n = 4; p > 0.05, two-tailed Student's t test).
DAT R615C exhibits accelerated rates of constitutive endocytosis and recycling
To define the mechanisms supporting loss of AMPH-induced surface trafficking of the DAT R615C variant, we implemented kinetic surface biotinylation assays that report rates of membrane protein endocytosis and recycling (Deken et al., 2003; Holton et al., 2005). By monitoring the amount of surface biotinylated DAT that displays resistance to a biotin-stripping reagent, we quantified the extent of transporter internalization under basal and AMPH-treated conditions. Here, we discovered that the DAT R615C variant displays an accelerated, constitutive internalization rate relative to the wild-type transporter (Fig. 3A). Moreover, whereas AMPH significantly increased the rate of internalization for wild-type DAT, we observed no change in DAT 615C internalization rate over that seen under basal conditions.
DAT 615C exhibits accelerated rates of constitutive endocytosis and recycling. A, Top, Representative Western blot showing increased basal endocytosis associated with the R615C variant. AMPH treatment (10 μm; 30 min) caused significant increase in internalization of DAT 615R without affecting 615C internalization rate. Bottom, Quantification of internalization rates normalized to stripping efficiency. Basal internalization rate of DAT 615C was significantly increased compared with 615R (16.7 ± 0.9% for 615C vs 8.5 ± 1.8% for 615R; n = 3; p < 0.05, Student's t test). AMPH treatment significantly increased internalization rate of DAT 615R (8.5 ± 1.8% for basal vs 18 ± 2.6% for AMPH; n = 3; p < 0.05, Student's t test) without affecting the R615C variant (16.7 ± 0.9% for basal vs 16.3 ± 2.9% for AMPH; n = 3; p > 0.05, Student's t test). B, Top, Representative Western blot showing accelerated recycling of R615C variant. AMPH treatment (10 μm; 30 min) did not impact the recycling rate for both DAT 615R and 615C. Bottom, Quantification of recycling rate normalized to intracellular DAT fraction. A bigger fraction of DAT still remains intracellularly after constitutive recycling for DAT 615R compared with 615C (42.9 ± 8% for 615R vs 11.2 ± 2.3% for 615C; n = 6; p < 0.005, Student's t test). AMPH treatment did not alter fractional DAT remaining inside after recycling (basal, 42.9 ± 8%, vs AMPH, 43.4 ± 11.7% for DAT 615R; and basal, 11.2 ± 2.3%, vs AMPH, 14.8 ± 4.1% for DAT 651C; n = 3–6; p > 0.05, Student's t test). Error bars indicate SEM. Asterisks indicate statistically significant difference.
Possibly, accelerated internalization of DAT 615C under basal conditions might explain its significantly diminished basal surface expression. Alternatively, a change in surface recycling could also explain this difference, and also account for insensitivity to AMPH. We therefore determined the extent of DAT 615R and 615C recycling by biotinylating the full pool of recycling transporter molecules at 37°C, followed by stripping away any surface resident transporters at 4°C. Upon warming the cells back to 37°C, we determined the rate of reappearance of either wild-type or DAT 615C. Thirty minutes after stripping and shifting back to 37°C, ∼40% of wild-type DAT remained intracellular under basal conditions, and the extent of recycling was not further decreased by AMPH. In contrast, ∼90% of the DAT R615C variant recycled under basal conditions. This capacity for recycling was also not diminished by AMPH (Fig. 3B). Thus, AMPH treatment produced a net internalization of wild-type DAT protein by enhancing transporter endocytosis rates without a compensatory increase in recycling rates. DAT 615C constitutively endocytosis and recycles at a much faster rate than wild-type transporters, and this difference is not impacted by AMPH treatment. Thus, the failure of DAT 615C to exhibit AMPH-triggered internalization and a commensurate reduction in DA uptake (Fig. 2B,C) is due to the efficient recycling of the transporter through a pathway that cannot support acceleration of trafficking rates, which are already elevated.
DAT 615C reveals a CaMKII-dependent state of functional inactivation
AMPH acts to mobilize intracellular calcium (Ca2+) in a CaMKII-dependent manner (Gnegy et al., 2004; Wei et al., 2007) and Ca2+ mobilization/CaMKII activation has been shown to be critical for AMPH evoked DA efflux, likely through phosphorylation of the DAT N terminus (Khoshbouei et al., 2004; Fog et al., 2006). Most relevant to our trafficking studies, CaMKII inhibition has also been shown to preclude net AMPH-mediated DAT surface redistribution (Wei et al., 2007). Since DAT 615C effluxes DA in response to AMPH, yet fails to traffic, we investigated whether CaMKII/DAT interactions, which depend on residues 615–617, remain intact (Fog et al., 2006). As shown in Figure 4A, we detected an approximately twofold increase in recovery of DAT 615C/CaMKII versus 615R/CaMKII complexes. Since CaMKII binding to the DAT C terminus results in phosphorylation of DAT N-terminal serines (Fog et al., 2006), we asked whether the increased CaMKII association of the R615C variant is paralleled by a change in transporter phosphorylation. Indeed, immunoprecipitation of DAT 615R or 615C from [32P]orthophosphate-labeled cells revealed a significantly increased basal phosphorylation of the R615C variant (Fig. 4B) when data were normalized by total protein input. Given that, in the stable cell lines used, total DAT 615C protein levels are modestly, but significantly, reduced when compared with total WT DAT protein, the elevation in DAT 615C phosphorylation is likely to be actually somewhat higher than illustrated.
Increased basal CaMKII association and phosphorylation of the R615C variant. A, Top, Representative Western blot showing increased association between DAT 615C and CaMKII compared with 615R. Bottom, Quantification of DAT co-IP with CaMKII normalized to respective inputs. Significantly larger pool of 615C is coprecipitated with CaMKII (179 ± 26% of 615R; n = 3; p < 0.05, Student's t test). B, Top, Representative autoradiograph indicating increased basal phosphorylation of DAT 615C compared with 615R. Bottom, Quantification of DAT bands from the autoradiograph showing significant increase in basal phosphorylation levels of DAT 615C (125 ± 1.6% of 615R; n = 3; p < 0.05, Student's t test). Equal total protein amounts were used for IP following the metabolic labeling. Error bars indicate SEM. Asterisks indicate statistically significant difference.
To assess whether the increased CaMKII association and basal phosphorylation associated with the R615C variant has functional consequences, we treated DAT 615R or 615C cells with the CaMKII inhibitor KN-93, either by itself, or as a pretreatment before AMPH addition. Treatment of cells with KN-93 alone did not impact the DA transport activity of wild-type, 615R cells (Fig. 5A). However this treatment blocked the reduction imposed on DA transport activity by AMPH treatment. As shown above, AMPH imposed no significant effect on DA transport activity for DAT 615C cells (Fig. 5A). Surprisingly, KN-93 alone significantly decreased the DA transport activity of 615C cells, and this effect was more pronounced combining KN-93 with AMPH. The specificity of these drug treatments was confirmed through the use of KN-92, a structural analog of KN-93 that does not inhibit CaMKII (Fig. 5B). Thus, it is possible that excessive, constitutive CaMKII regulation of DAT 615C contributes to the shift of transporters to an AMPH-insensitive endocytosis and recycling pathway. Indeed, biotinylation studies demonstrated that CaMKII blockade with KN-93 antagonized AMPH-induced reductions in cell surface expression cell of wild-type DAT (Fig. 5C), similar to its ability to block AMPH-induced reductions in DA uptake. In contrast to its effect on DAT 615C-supported DA uptake, KN-93 treatment produced no effect on transporter internalization alone or in the presence of AMPH (Fig. 5C). These data reveal a CaMKII-dependent capacity to modify the function of DAT 615C in a trafficking-independent manner.
DAT R615C reveals a CAMKII-dependent state of functional inactivation. A, KN-93 (1 μm; 30 min) treatment (black bar, KN-93) caused a significant DA transport reduction in 615C (79 ± 4% of basal) while showing no effect on WT-DAT (white bar, KN-93, 95 ± 9% of basal). Pretreatment with KN-93 before AMPH (10 μm; 30 min) addition blocked AMPH-mediated DA transport reduction in DAT 615R [white bars, AMPH (70 ± 4%) vs KN-93 plus AMPH (98 ± 10%)], while DAT 615C showed significant reduction upon KN-93 plus AMPH treatment [black bars, AMPH (89 ± 3%) vs KN-93 plus AMPH (63 ± 5%)]. The reduction upon KN-93 plus AMPH treatment was not significantly different from the reduction obtained by treatment with KN-93 alone in the R615C variant (79 ± 4% for KN-93 vs 63 ± 5% for KN-93 plus AMPH). Data were normalized to basal DA uptake of respective cell lines. Significant differences were calculated using one-way ANOVA followed by Bonferroni's post hoc test (n = 3–6; p < 0.0001). B, An inactive analog of KN-93, KN-92, by itself did not affect DA uptake in either DAT 615R (white bar, KN-92, 102 ± 10% of basal) or 615C (black bar, KN-92, 108 ± 12% of basal). Pretreatment with KN-92 (1 μm; 30 min) did not block AMPH-mediated (10 μm; 30 min) DA transport reduction in DAT 615R (white bar, KN-92 plus AMPH, 65 ± 8% of basal) and had no effect on 615C (black bar, KN-92 plus AMPH, 90 ± 7% of basal). Data were normalized to basal DA uptake of respective cell lines. Significant differences were calculated using one-way ANOVA followed by Bonferroni's post hoc test (n = 3; p < 0.05). C, Top, Representative Western blots showing KN-93 plus AMPH treatment did not cause reduction in DAT surface levels for both DAT 615R and 615C. Bottom, Quantification showing KN-93 plus AMPH (1 μm KN-93 for 30 min followed by 10 μm AMPH for 30 min) treatment rescued the reduction in the DAT surface expression of DAT 615R (101 ± 7% of basal) similar to results obtained by DA uptake assays. KN-93 plus AMPH treatment showed no change (113 ± 10% of basal) in the surface levels of DAT 615C. The differences obtained from the KN-93 plus AMPH treatment were not statistically significant from their basal correlates assessed by one-way ANOVA (n = 7; p > 0.05). D, KN-93 plus AMPH (1 μm KN-93 for 30 min followed by 10 μm AMPH for 30 min) treatment resulted in significant reduction in DA transport Vmax for the R615C variant. DA transport Vmax and Km are expressed as picomoles · well−1 · minute−1 and micromolar, respectively. A significant decrease in the DA transport Vmax was observed after KN-93 plus AMPH treatment (65 ± 10, basal, vs 42 ± 3, KN-93 plus AMPH; n = 4; p < 0.05, Student's t test) without a significant change in Km (5.4 ± 1.5, basal, vs 3.8 ± 0.6, KN-93 plus AMPH; n = 4; p > 0.05, Student's t test). Error bars indicate SEM. Asterisks indicate statistically significant difference.
The ability of CaMKII to enhance DAT 615C DA transport activity could arise from either a shift of transporters to high-affinity DA recognition or from a concentration-independent increase in DA transport capacity. To address these possibilities, we conducted saturation kinetic analysis for the DAT R615C variant, in the presence or absence of KN-93 plus AMPH (Fig. 5D). These drug treatments produced no significant change in DA transport Km (basal, 5.4 ± 1.5, vs KN-93 plus AMPH, 3.8 ± 0.6 μm), but generated a significant reduction in DA transport Vmax (basal, 65 ± 10, vs KN-93 plus AMPH, 42 ± 4 pmol · well−1 · min−1; p < 0.05, two-tailed Student's t test). These findings support the hypothesis that CaMKII activity is required to sustain R615C in an active state.
DAT 615C demonstrates altered localization to membrane microdomains
Membrane microdomains have been reported to associate with (Foster et al., 2008) and influence DAT conformations (Hong and Amara, 2010), to constrain DAT lateral mobility (Adkins et al., 2007), to be required for PKC-mediated internalization, and to dictate CaMKII-mediated DA efflux (Cremona et al., 2011). DAT has been shown to localize to microdomains enriched for the protein flotillin-1 (Cremona et al., 2011). Coimmunoprecipitation studies revealed reduced levels of the flottilin-1 in extracts of cells stably transfected with DAT 615C versus extracts from cells transfected with wild-type DAT (Fig. 6A). Since reduced association of flotillin-1 could reflect diminished surface expression of DAT (no data has been published as to whether DAT/flotillin-1 associations are present in recycling endosomes or limited to the plasma membrane), as opposed to a change in microdomain localization, we used confocal imaging to compare the colocalization of YFP-tagged DAT 615R or 615C proteins with Alexa 647-conjugated CTxB. The latter probe detects GM1 ganglioside, another molecule known to localize to discrete membrane subdomains (Simons and Ikonen, 1997). Moreover, in transfected N2a cells, DAT has been found to exhibit a greater degree of colocalization with CTxB-labeled membranes, when compared with the transferrin receptor (Adkins et al., 2007). As shown in Figure 6B, colocalization of the DAT R615C variant with CTxB-labeled membranes was significantly lower than that seen with wild-type DAT.
Altered membrane raft microdomain association of the R615C variant. A, Top, Representative Western blot showing reduced association of DAT 615C with flotillin-1. Bottom, Quantification of flotillin-1 co-IP with DAT normalized to respective inputs. DAT 615C displays significantly reduced association with flotillin-1 (33 ± 8% of 615R; n = 5; p < 0.001, Student's t test). B, Top, HEK 293T cells are transiently transfected with YFP-HA-hDATs (green). Alexa 647-conjugated cholera toxin B is used to mark GM1 ganglioside (red). Representative cells are shown with white arrows indicating lack of DAT association with the raft fraction of R615C variant. Bottom, ICQ analysis is performed to define colocalization of YFP-DAT with CTx-B-labeled lipid rafts. Quantification demonstrated that, although both 615R and 615C DAT are significantly colocalized with CTx-B, 615C exhibits a significantly reduced ICQ value compared with 615R (615R, 0.1 ± 0.006, vs 615C, 0.06 ± 0.006; n = 22; p < 0.01, Student's t test). Error bars indicate SEM. Asterisks indicate statistically significant difference.
DAT 615C acts dominantly via generation of local negative charge to disrupt AMPH actions
The ADHD subject with whom we initiated our studies is heterozygous for the R615C variant, suggesting that, if the variant is a significant risk determinant for the disorder, it likely acts dominantly to alter DAT function. One mechanism by which dominance could be established is from an ability of the R615C variant to preclude association of molecules needed for normal DAT trafficking and function, possibly as a consequence of the ability of the transporter to dimerize (Sorkina et al., 2003; Torres et al., 2003). To explore this idea, we synthesized peptides that comprise the last C-terminal 24 aa of either DAT 615R or 615C attached to the membrane-permeable TAT sequence (Schwarze et al., 1999). Incubation of DAT 615R cells with the TAT-C24615R peptide failed to impact AMPH-mediated downregulation of DA uptake (Fig. 7A). However, incubations of these cells with the TAT-C24615C peptide eliminated the effects of AMPH on the wild-type transporter (Fig. 7A). Addition of either TAT-C24615R or TAT-C24615C peptide (or no peptide) produced no effects on either basal or AMPH-modulated DA transport of DAT 615C cells (Fig. 7B). These findings suggest that, rather than attracting molecules that can drive DAT 615C out of GM1 and flotillin-1-containing membrane microdomains, the 615C-substituted C terminus may preclude interactions needed for successful residency in these compartments.
DAT 615C acts dominantly via generation of local negative charge to disrupt AMPH actions. A, B, Exogenous addition of TAT-C24615C peptide blocks AMPH-mediated reduction in the DA transport for DAT 615R and had no effect on 615C. Addition of either TAT-C24615R or TAT-C24615C peptide (10 μm; 18 h) did not alter basal DA uptake compared with no peptide control for both DAT 615R and 615C (for DAT 615R, 104 ± 6% with TAT-C24615R and 102 ± 3% with TAT-C24615C; for DAT 615C, 96 ± 3% with TAT-C24615R and 103 ± 4% with TAT-C24615C). Addition of TAT-C24615R peptide or no peptide control produced a significant reduction in DA uptake for DAT 615R upon AMPH (10 μm; 30 min) treatment (72 ± 6% of basal no peptide for TAT-C24615R and 77 ± 5% of basal no peptide for TAT-C24615C) and had minimal effect on DAT 615C (84 ± 2% of basal no peptide for TAT-C24615R and 85 ± 1% of basal no peptide for TAT-C24615C). Addition of AMPH (10 μm; 30 min) to TAT-C24R615C peptide-treated DAT 615C cells also did not produce a significant reduction in DA uptake (88 ± 4% basal no peptide control). AMPH addition to TAT-C24615C-treated WT-DAT cells significantly blocked AMPH-mediated reduction in the DA transport seen in TAT-C24615R-treated and no peptide control conditions (91 ± 3% basal no peptide control). Significant differences were calculated using one-way ANOVA followed by Bonferroni's post hoc test (n = 3–4; p < 0.05). C, Presence of DAT 615C or a negatively charged aspartic acid (613E) blocked AMPH-mediated reduction in the DA transport. Specific DA uptake values are expressed as percentage of basal control. Only DAT 615C (103 ± 13% of basal) and 615E (102 ± 12% of basal) mutations did not display reduction in the DA uptake upon AMPH treatment (10 μm; 30 min). All other mutations at R615 position and DAT 615R showed a significantly reduced DA uptake upon AMPH treatment (615R, 68 ± 3%; 615K, 75 ± 6%; 615A, 66 ± 5%; 615Q, 78 ± 5%; 615S, 73 ± 5% of basal). Significant differences were calculated using Student's t test for each mutation (n = 3–4; p < 0.05). D, Introduction of a negative charge by T613E mutation abolishes AMPH-mediated reduction in the DA transport. AMPH treatment (10 μm; 30 min) caused a significant reduction in the DA transport for DAT 615R and 613A (DAT 615R, 77 ± 5%; 613A, 81 ± 3% of basal) T613E mutation completely blocked AMPH-mediated reduction in the DA transport (97 ± 4% of basal). Mutating T613 to either A or E on DAT 615C background and 615C alone did not produce any significant reduction in the DA uptake upon AMPH treatment (10 μm; 30 min) reduction (DAT 615C, 92 ± 5%; 613A-615C, 86 ± 3%; 613E-615C, 92 ± 4% of basal). Significant differences were calculated using Student's t test for each mutation (n = 5; p < 0.01). Error bars indicate SEM. Asterisks indicate statistically significant difference.
Next, we sought to determine whether the 615C substitution per se perturbs DAT regulation, or whether the loss of Arg residue at this position confers AMPH insensitivity. Following site-directed mutagenesis of the wild-type DAT C terminus at residue 615, we found that cells transfected with DAT 615A presented a pattern of AMPH regulation indistinguishable from those of DAT 615R (Fig. 7C), indicating that Cys addition versus Arg loss determines AMPH insensitivity. Cytoplasmic Cys residues can be modified by palmitoylation or nitrosylation (Nagahara et al., 2009), modifications that have both been suggested to occur with catecholamine transporters (Kaye et al., 2000; Foster and Vaughan, 2011). However, we detected no differences in palmitoylation between wild-type DAT and the R615C variant, and NOS inhibition did not restore AMPH regulation (data not shown). Cys residues also participate in disulfide bond formation, although this seems unlikely given the reducing environment of the cytosol where the DAT C-terminus resides. Cytoplasmic Cys residues can form acidic thiolates (Nagahara et al., 2009) and as such could confer a charge inversion compared with the wild-type 615R residue. Therefore, we asked whether DAT 615E would also confer insensitivity to AMPH. Indeed, DAT 615E displayed no reductions in DA uptake in response to AMPH treatment (Fig. 7C). Since the mutants 615K, 615A, 615Q, and 615S all respond to AMPH (Fig. 7C), we suggest that a negative thiolate arising from Cys substitution at the 615 position may be responsible for the shift of the DAT R615C variant to AMPH insensitivity.
The Arg residue at 615 of wild-type DAT forms a canonical phosphorylation site for multiple kinases with Thr613 (Amanchy et al., 2007). However, since the R615A substitution is still AMPH sensitive, phosphorylation at this site is not required for the impact of the psychostimulant on DAT regulation. Consistent with this finding, cells transfected with the T613A mutant retained AMPH sensitivity (Fig. 7D). Moreover, cells transfected with a T613E mutant on the wild-type 615R background to mimic the charge that would be induced by Thr613 phosphorylation lost AMPH sensitivity (Fig. 7D). These findings suggest that Thr613 may need to remain dephosphorylated to sustain psychostimulant (and PKC) regulation and that a nearby 615C thiol/thiolate may sufficiently mimic the structure or negative charge of phosphorylated Thr613 to preclude DAT regulation by AMPH.
Discussion
ADHD is the most commonly diagnosed disorder of childhood (Smith et al., 2009) and multiple lines of evidence suggest that alterations in DA signaling, including changes in DAT expression or function (Mazei-Robison et al., 2005), can contribute to both cognitive and hyperactive traits of this disorder. A meta-analysis of association studies examining DAT polymorphisms in ADHD (Gizer et al., 2009) concluded that, although replicable associations are evident, the SLC6A3 locus is likely to harbor multiple functional variants whose variable influence across families can account for differences in effect sizes detected across studies. To date, however, the majority of genetic studies implicating DAT gene variation with ADHD, or the efficacy of ADHD medications, derive from analysis of a variable number tandem repeat (VNTR) in the 3′-untranslated region. As the functional impact of this VNTR remains ill defined (Winsberg and Comings, 1999; Kirley et al., 2003), aligning these studies with specific alterations in DA signaling is difficult. Rather than focus on common DAT variation of uncertain function, we sought insights into the DA contributions to ADHD via a search for highly penetrant, DAT coding variants. Studies of rare, heritable forms of Alzheimer's disease (Lemere et al., 1996) and Parkinson's disease (Groen et al., 2004) provide two cogent examples of how the elucidation of rare gene variation can lead to novel pathophysiological concepts. Moreover, intensive study of rare variants is justified as such studies can help define broader networks that may elucidate a common underlying pathophysiology.
Recently, we described properties of the DAT coding variant, A559V, identified in two male siblings with ADHD (Mazei-Robison et al., 2005). We found A559V displays increased DAT channel activity and spontaneous, nonvesicular DA release that can be greatly enhanced by membrane depolarization (Mazei-Robison et al., 2008; Bowton et al., 2010). Moreover, spontaneous DA efflux can be attenuated by the ADHD therapeutic AMPH. Here, we describe a second, rare, ADHD-associated DAT coding variant, R615C, where substitution establishes profound basal and regulatory alterations. Although the pedigree harboring the R615C variant is small, with only a single affected carrier, the regulatory disruption we report provides further evidence that changes in DAT-dependent DA signaling contributes to risk for ADHD. Studies with knock-in mice that harbor the R615C variant, similar to those underway in our laboratory involving the A559V variant (M. A. Mergy and R. D. Blakely, unpublished observations) should allow us to assess which of the multiple in vitro alterations in trafficking and regulation we have identified occurs in vivo, and how these or other changes impact synaptic DA inactivation.
Multiple elements of the cytoplasmic N and C termini of DAT have been implicated in basal and regulated control of DAT surface expression, stability, and activity. Whereas the DAT N terminus has been the focus of much research investigating AMPH-induced DA efflux (Khoshbouei et al., 2004; Fog et al., 2006; Binda et al., 2008), Holton et al. (2005) have proposed that sequences in the C terminus (residues 587–596) influence basal and PKC-modulated transporter trafficking. The distal C terminus (residues 618–620) bears a type II PDZ domain interaction motif that has been shown to dictate interactions of the transporter with the PDZ domain protein, PICK1, and possibly enhance DAT surface expression (Torres et al., 2001). Interestingly, Bjerggaard et al. (2004) found that substitution of AAA for the RHW sequence immediately upstream of the PDZ recognition motif leads to ER retention, but preserves the ability of mutant DAT to bind PICK1. The R615C mutation lies in the RHW sequence, and thus an effect on ER/Golgi export may contribute to the reduced surface expression of the transporter and DA transport Vmax.
As previously reported, we found AMPH treatments produce net transporter endocytosis (Saunders et al., 2000; Kahlig et al., 2004; Boudanova et al., 2008a). Activation of PKC isoforms with β-PMA also induces net transporter internalization (Miranda et al., 2007; Boudanova et al., 2008b; Cremona et al., 2011), an effect proposed to be mediated by DAT C-terminal sequences (Holton et al., 2005). Remarkably, we found that the R615C variant demonstrates complete resistance to the trafficking effects of both AMPH and PKC activation. These findings led us to monitor the kinetics of DAT 615C basal and regulated surface trafficking. We found that DAT 615C displays a significantly accelerated rate of both endocytosis and recycling compared with wild-type DAT. In our cells, treatment with AMPH accelerates the endocytic rates of wild-type DATs, with no detectable change in transporter recycling. Thus, it seems likely that DAT 615C lacks the dynamic range needed to display regulated endocytosis/exocytosis due to its high rate of constitutive trafficking.
DAT proteins have been found to reside within cholesterol and GM1 ganglioside-enriched membrane microdomains often referred to as “lipid or membrane rafts” (Sandvig and van Deurs, 2000; Adkins et al., 2007; Foster et al., 2008) and to associate with the raft-associated protein flotillin-1 (Cremona et al., 2011). Also, DATs localized to CTxB-labeled membrane microdomains demonstrate restricted mobility and can be mobilized by membrane cholesterol extraction (Adkins et al., 2007). We found a reduced colocalization of DAT 615C with CTxB-labeled membranes as well as a reduced association of the membrane raft-associated protein flotillin-1. Since quantitation of our coimmunoprecipitation data was normalized for total DAT protein, the reduced flotillin-1 association of the R615C variant may reflect, at least in part, the reduced surface expression of the transporter. However, at present we do not know whether DAT and flotillin-1 have constitutive or dynamic interactions during normal transporter recycling. In the context of our immunofluorescence findings of a surface redistribution of R615C between GM1+ and GM1− compartments, we should also consider that redistribution on the surface might result in targeting DAT to a compartment in which opportunities for flotillin-1 interactions are lost. Cremona et al. (2011) provided evidence that interactions with flotillin-1 are required for both PKC-dependent DAT trafficking and AMPH-induced DA efflux. Consistent with these findings, we observed a loss of PKC-induced transporter trafficking with DAT 615C, although AMPH-induced DA efflux was maintained. These findings indicate that DAT/flotillin-1 associations support, but may not be necessary, for AMPH-triggered DA efflux. Since CaMKII associates with DAT through C-terminal sequences that overlay the R615 residue and is critical for AMPH-triggered DA efflux (Fog et al., 2006), the constitutively elevated DAT/CaMKII association may compensate for the loss of flotillin-1 associations, placing mutant transporters in a more DA efflux-competent state. Alternatively, since DAT responds rapidly to AMPH action via increased surface expression (Furman et al., 2009), the R615C substitution may also enhance DAT levels transiently in response to AMPH, during the period of DA efflux measurements. Regardless, these findings raise the possibility that flotillin-1 and CaMKII interactions may be exclusive, with their exchange being a key feature of dynamic DAT regulation. Recently, a Ras-like GTPase, Rin, was shown to associate with DAT C terminus in membrane rafts and regulate PKC-mediated DAT downregulation (Navaroli et al., 2011). Possibly, the mislocalization and regulatory perturbations we observe with DAT 615C may derive from altered Rin associations, but this issue requires further studies.
As we further explored the role of CaMKII in the insensitivity of the R615C variant to AMPH-induced internalization, we discovered a capacity for the kinase to support a trafficking-independent mode of transporter functional regulation. Specifically, CaMKII activity appears to maintain basal activity of DAT 615C when transporters are localized away from flotillin-1-rich membrane microdomains. Possibly, wild-type DAT residence in membrane microdomains may interfere with the activity of CaMKII, providing a mechanism whereby physiological stimuli could produce enhanced DAT activity, possibly during states of high DA release. Both norepinephrine and serotonin transporters exhibit similar states of trafficking-independent catalytic regulation (Apparsundaram et al., 2001; Steiner et al., 2008).
Studies examining the physical requirements for AMPH regulation of wild-type DAT in relation to the R615C variant provide key insights into mechanisms supporting transporter regulation. Using TAT-C24615R and TAT-C24615C peptides, we demonstrated that the mutant C terminus acts dominantly to eliminate AMPH actions on wild-type DAT. These findings are important given the carrier status of our ADHD proband from which the R615C variant was identified. They also indicate that the R615C may compete in the hemizygous state with the wild-type DAT C terminus for key interactions needed for AMPH-induced transporter regulation. Additionally, we found that, whereas 615C precludes AMPH-induced transporter downregulation, Ala, Lys, Gln, and Ser substitutions for R615 fail to perturb regulation. In further pursuit of the structural basis for the 615C effect, we found that 615E recapitulated the impact of 615C, suggesting that the generation of negative charge in the distal C terminus may play a role in disrupting transporter regulation. Cys residues can exist as thiolates and thereby create a local negative charge (Nagahara et al., 2009). Whereas the pKa of a free Cys thiol is estimated at ∼8, leaving the Cys side chain of 615C largely protonated, the pKa of proteinaceous Cys residues is known to be quite sensitive to its local environment, where the thiolate anion can be stabilized by surrounding residues (Netto et al., 2007).
Since R615 forms a phosphorylation site motif with T613, we considered the possibility that 615C gains its capacity to disrupt regulation by interfering with T613 phosphorylation. However, we found that a T613E mutation, created to produce the negative charge associated with Thr phosphorylation, recapitulates the behavior of 615C. These findings suggest that dephosphorylation of T613 may be a critical step in sorting DAT into regulated versus constitutive endocytic pathways. Further studies are needed to determine whether, and under what conditions, kinases and phosphatases may target this residue, either constitutively or as part of a rapid, regulatory mechanism, or whether other mechanisms, such as charge-dependent protein associations, impart sorting decisions.
In toto, our findings lead us to propose a two-compartment model that organizes our findings (Fig. 8). DAT exists at the cell surface in GM1/flotillin-1-enriched microdomains that restrict transporter lateral mobility and provide for regulated trafficking. Alternatively, the transporter can reside in membrane depleted of these raft components and here traffics largely via constitutive endocytosis and rapid recycling. Decisions dictating the localization and trafficking of DAT are dependent on sequences in the distal C terminus of DAT, which we suspect involves phosphorylation/dephosphorylation reactions at T613. Our model has parallels with the differential trafficking of insulin-responsive (GLUT4) and nonresponsive (GLUT1) glucose transporters. GLUT4 and GLUT1 equivalently support glucose uptake, but GLUT4 traffics through a limited capacity, regulated pathway, whereas GLUT1 traffics in the same cells through higher-capacity, constitutive mechanisms (Zorzano et al., 1997). Wild-type DAT has the capacity to occupy either a regulated or constitutive pathway, but in our model systems is biased normally toward GM1/flotillin-1-enriched domains. As shown by our studies with the DAT615C peptide, the DAT C terminus can play a dominant role in the capacity of the transporter for regulated trafficking. The DNA encoding DAT 615 and the other mutants generated in this study, as well as the DAT615C peptide should be useful in further elucidating the mechanisms by which DAT sorts between constitutive and regulated trafficking pathways. Although the R615C variant is rare, our findings suggest that a more intensive analysis of proteins that influence DAT localization and function within membrane microdomains may provide important insights for idiopathic ADHD.
Model describing differential trafficking of DAT 615R and DAT 615C to the regulated and constitutive endocytic pathways and biased localization toward GM1/flotillin-1 rich or depleted membrane microdomains. DAT molecules are depicted as located in trafficking vesicles or at the plasma membrane as homomeric or heteromeric dimers (although other configurations are acceptable). Transporters localize to two types of membrane domains, one domain enriched for GM1 ganglioside and flotillin-1, with the other domain relatively depleted of these molecules. Homo-multimers of WT DAT (615R) subunits target preferentially to GM1 ganglioside/flotillin-1-enriched compartment compared with DAT 615C subunit containing dimers. The heterodimer is shown preferentially targeting domains that are depleted of GM1 and flotillin-1, in keeping with the dominant action of the 615C mutation in the ADHD proband and the dominant action of the DAT C-terminal peptide containing 615C. DAT 615C subunits are diminished in surface levels relative to WT subunits. Microdomains enriched for GM1 and flotillin-1 support regulated trafficking of DATs and under basal conditions exhibit slow rates of endocytosis, whereas membranes depleted of GM-1 and flottilin-1 support more rapid, constitutive endocytosis and recycling.
Footnotes
This work was supported by National Institutes of Health (NIH) Predoctoral Fellowships MH067472 (M.S.M.-R.), MH090738 (M.A.M.), MH064913 (E.B.), and National Science Foundation Fellowship DGE0909667 (P.J.H.). R.D.B. was supported by NIH Award HL56693. A.G. was supported by NIH Awards DA013975 and DA014684. The Vanderbilt University Medical Center Cell Imaging Shared Resource was supported by NIH Grants CA68485, DK20593, DK58404, HD15052, DK59637, and EY08126. We thank Mark Stein (University of Illinois, Chicago, IL) and Irwin Waldman (Emory University, Atlanta, GA) for contribution of DNAs from ADHD subjects for this study. We thank Chris Svitek, Jane Wright, and Angela Steele for laboratory oversight and technical assistance. D.S. thanks the Vanderbilt Brain Institute for graduate training support and Roger Colbran for critical review of this manuscript.
- Correspondence should be addressed to Dr. Randy D. Blakely, Vanderbilt Brain Institute, Suite 7140, MRBIII, Vanderbilt University Medical Center, Nashville, TN 37232-8548. randy.blakely{at}vanderbilt.edu
