Control of serotonin transporter phosphorylation by conformational state

Significance

Mutations in serotonin transporter (SERT) affecting regulation of a cGMP-dependent signaling pathway have been associated with psychiatric disorders. Earlier studies provided preliminary evidence that cGMP stimulates phosphorylation of SERT in a transmembrane helix. Transporters like SERT function by undergoing a cycle of conformational changes as they move their substrate (in this case the neurotransmitter 5-HT) across the membrane. This work shows that SERT conformation dramatically affects phosphorylation, which is stimulated when the transporter is in a conformation known to increase during 5-HT transport. The results suggest a novel mechanism of regulation in which transport of 5-HT increases the level of an inward-open SERT conformation that provides accessibility of the phosphorylation site to a kinase.

Abstract

Serotonin transporter (SERT) is responsible for reuptake and recycling of 5-hydroxytryptamine (5-HT; serotonin) after its exocytotic release during neurotransmission. Mutations in human SERT are associated with psychiatric disorders and autism. Some of these mutations affect the regulation of SERT activity by cGMP-dependent phosphorylation. Here we provide direct evidence that this phosphorylation occurs at Thr276, predicted to lie near the cytoplasmic end of transmembrane helix 5 (TM5). Using membranes from HeLa cells expressing SERT and intact rat basophilic leukemia cells, we show that agents such as Na⁺ and cocaine that stabilize outward-open conformations of SERT decreased phosphorylation and agents that stabilize inward-open conformations (e.g., 5-HT, ibogaine) increased phosphorylation. The opposing effects of the inhibitors cocaine and ibogaine were each reversed by an excess of the other inhibitor. Inhibition of phosphorylation by Na⁺ and stimulation by ibogaine occurred at concentrations that induced outward opening and inward opening, respectively, as measured by the accessibility of cysteine residues in the extracellular and cytoplasmic permeation pathways, respectively. The results are consistent with a mechanism of SERT regulation that is activated by the transport of 5-HT, which increases the level of inward-open SERT and may lead to unwinding of the TM5 helix to allow phosphorylation.

In 2003, Ozaki et al. (1) reported a rare missense mutation in the coding region of serotonin transporter (SERT) in two unrelated families. Within these families, individuals with the mutation (Ile425Val, or I425V) were more strongly affected with several psychiatric disorders, including obsessive-compulsive disorder and autism spectrum disorders. Subsequently, additional families with this mutation and others in SERT have been identified (2–8).

In transfected cells, the I425V mutation led to enhanced 5-HT transport relative to WT SERT (5, 9). The enhanced transport was similar to the up-regulation of 5-HT transport that had been observed in rat basophilic leukemia (RBL-2H3) cells on stimulation of adenosine A3 receptors with subsequent NO formation and cGMP production (10). Results with heterologously expressed SERT suggest that the I425V mutation interferes with this pathway either by directly increasing basal SERT phosphorylation or by inhibiting SERT dephosphorylation (11).

Mutation of serine and threonine residues on the cytoplasmic surface of SERT led to the identification of Thr276 as a potential cGMP-dependent phosphorylation site (12). From homology models based on prokaryotic and eukaryotic homologs, this residue is located near the cytoplasmic end of transmembrane helix 5 (TM5), a location that agrees with cysteine scanning accessibility studies (13) that also have shown this position to be poorly accessible to aqueous reagents. The unlikely location of a phosphorylation site in a transmembrane helix, and the inaccessibility of this position, cast some doubt on the proposal that Thr276 is the site of cGMP-stimulated phosphorylation. A possible mechanism for phosphorylation of SERT at this site is that the cytoplasmic end of the TM5 helix undergoes a conformational change in which the region containing Thr276 unwinds and extends into the cytoplasm, where it would be accessible to a kinase. A recent X-ray structure of the bacterial SERT homolog, MhsT, showed a partial unwinding of the cytoplasmic half of TM5 (14), suggesting that such a conformational change might be possible.

Here we present additional support for Thr276 as the site of cGMP-stimulated SERT phosphorylation and examine the relationship between this phosphorylation and SERT conformational states. The results suggest a novel mechanism by which the normal function of SERT as a transporter acts to enhance its phosphorylation.

Materials and Methods

Materials.

HeLa and RBL-2H3 cells were obtained from American Type Culture Collection. Rabbit polyclonal antibody against a rat SERT peptide containing phospho-Thr276 (p1700-276) was obtained from PhosphoSolutions. Rabbit polyclonal antibody against the human SERT C terminus (C-20) was a generous gift from Dr. F. Kilic, University of Arkansas, Little Rock, AR. Antibody against the γ-subunit of the Fcε receptor (06-727) was obtained from Millipore. 2-Aminoethyl methane thiosulfonate hydrobromide (MTSEA) and (2-sulfonatoethyl) methanethiosulfonate (MTSES) were purchased from Anatrace. 8-(4-Chlorophenylthio)-guanosine 3′,5′-cyclic monophosphorothioate, Rp isomer (Rp-8-pCPT-cGMPS), S-nitroso-N-acetyl-DL-penicillamine (SNAP), 5′-(N-ethylcarboxamido)adenosine (NECA), 8-bromoguanosine-3′,5′-cyclic monophosphate (8-Br-cGMP), monoclonal anti-FLAG M2 affinity agarose, and 3×FLAG peptide were purchased from Sigma-Aldrich. Protein kinase G (PKG)-specific peptide substrate (RKRSRAE) was obtained from GenScript. Ni-NTA agarose was obtained from Qiagen. [³H]5-HT (27.1 Ci/mmol), [γ-³³P]ATP (3,000 Ci/mmol), and [¹²⁵I]2β-carbomethoxy-3β-(4-iodophenyl)tropane ([¹²⁵I]RTI-55, β-CIT; 2,200 Ci/mmol) were obtained from PerkinElmer. All other materials and chemicals were reagent grade and obtained from commercial sources.

SERT mutants were generated using the QuikChange Site-Directed Mutagenesis Kit (Stratagene). N-terminal His₆ and C-terminal FLAG-tagged human SERT expression constructs were generated by PCR. In brief, forward and reverse primers containing unique restriction sites were designed to introduce DNA sequences encoding six His residues immediately after the initiator methionine and a FLAG sequence immediately before the stop codon and used to amplify a full-length SERT. The PCR products were then digested with restriction enzymes, excised from agarose gels, and ligated into pcDNA3.1. The regions amplified by PCR were confirmed by DNA sequencing. Double-tagged SERT (His-SERT-FLAG) did not differ in transport rate or kinetics from WT SERT.

Cell Culture and Transfection.

HeLa cells were cultured in DMEM supplemented with 10% (vol/vol) FBS, 2 mM l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin and RBL-2H3 in MEM with 15% (vol/vol) FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a humidified 5% (vol/vol) CO₂ incubator. SERT was expressed in HeLa cells transfected using lipofectin with pcDNA3.1 containing His-SERT-FLAG cDNA under control of the CMV promoter. Transfected cells were incubated for 24–36 h at 37 °C with 5% (vol/vol) CO₂ before the transport assays. Protein concentration was determined using the Pierce Micro BCA Protein Assay Reagent Kit (Thermo Fisher Scientific).

Transport and Binding Assays.

5-HT transport into RBL cells or SERT-transfected HeLa cells was measured as described previously (13). In brief, cells growing in 96-well plates were washed once with 100 µL of PBS/CM (PBS containing 0.1 mM CaCl₂ and 1 mM MgCl₂) and incubated in PBS/CM for the indicated times at 20 °C with or without various modulators. 5-HT uptake assays were initiated by the addition of [³H]5-HT (20 nM final concentration), and continued for 10 min for HeLa cells and 20 min for RBL-2H3 cells. The assays were terminated by three rapid washes with ice-cold PBS. The cells were then solubilized in 30 μL of 0.1 M NaOH for 30 min, after which 120 µL of Optifluor (PerkinElmer) was added. Accumulated [³H]5-HT was determined by liquid scintillation spectrometry in a PerkinElmer Microbeta plate counter.

Binding of the high-affinity cocaine analog [¹²⁵I]β-CIT was measured in crude membrane preparations from transfected HeLa cells as described previously (13). For membrane binding assays, frozen membranes from cells expressing SERT mutants were thawed on ice, applied to Multiscreen-FB 96-well filtration plates (Millipore; ∼100 µg per well), and washed five times by filtration with 100 μL of binding buffer (10 mm Hepes buffer, pH 7.4, containing 150 mM NaCl). MTSEA was subsequently added to the membranes in the presence of modulators at the indicated concentrations and incubated for 15 min at 20 °C, after which the membranes were washed five times with by filtration to remove unbound MTSEA. β-CIT binding was then initiated by the addition of 100 µL of binding buffer containing 0.1 nm [¹²⁵I]β-CIT. Binding was allowed to proceed for 1.5 h at 20 °C with gentle rocking. The reaction was stopped by filtration and three washes with 100 μL of ice-cold binding buffer. The filters were removed from the plate and counted with a PerkinElmer Microbeta plate counter in 150 μL of Optifluor.

Kinase Assay.

PKG activity was measured by determining the amount of ³³P radioactivity incorporated from [γ-³³P]ATP into a PKG-specific peptide substrate (RKRSRAE) as described previously (15). The 75-μL assay mixture contained 0.15 µCi of [³³P]ATP, 10 µM ATP, 15 µM PKG peptide substrate, 2 µM PKI (a synthetic peptide inhibitor of cAMP-dependent protein kinase), 1 µg of purified PKG-1α, 100 µM 8-Br-cGMP, and the indicated modulator in 50 mM Hepes buffer, pH 7.4, containing 10 mM MgCl₂, 0.1% Tween 20, and 1 mM DTT. After incubation at 30 °C for 2 min, the reaction was cooled on ice. Then 20 µL of the assay mixture was spotted onto P81 phosphocellulose paper and quenched in 0.42% H₃PO₄. The paper was washed three more times in 0.42% H₃PO₄ for 10 min with gentle agitation and then rinsed once with acetone. After air-drying, radioactivity on the paper was measured with a Beckman LS6500 liquid scintillation counter.

SERT Phosphorylation in Vitro.

SERT phosphorylation in vitro was measured using membranes prepared from HeLa cells expressing N-terminal His₆- and C-terminal FLAG-tagged WT (or mutant) SERT as described previously (15). The standard 50-μL phosphorylation mixture contained 0.5 µCi of [γ-³³P]ATP, 50 µg of membrane protein, 1 µg of purified PKG-Iα, and 100 µM 8-Br-cGMP in 150 mM NaCl, or N-methyl-d-glucamine (NMDG)-Cl, containing 10 mM Hepes buffer, pH 7.4, 5 mM MgCl₂, 0.2 mM EDTA, and 1 mM DTT. Modulators (e.g., cocaine, ibogaine) were added to the phosphorylation mixture at the indicated concentrations, and the reaction was started by the addition of ATP. The mixture was incubated at 30 °C for 30 min, then cooled on ice and solubilized in lysis buffer (1 mL of 150 mM NaCl containing 25 mM Tris buffer, pH 8.0, 1 mM EDTA, and 1% Triton X-100). The detergent extracts were incubated with 100 μL of anti-FLAG M2 affinity agarose [50% (vol/vol) suspension in lysis buffer] at 4 °C overnight. After five washes with ice-cold lysis buffer, SERT was eluted with 50 μL of 150 ng/μL 3×FLAG peptide. The eluate was then denatured with 500 μL of 150 mM NaCl containing 25 mM Tris buffer, pH 8.0, 1 mM EDTA, and 8 M urea and incubated with 80 µL of Ni-NTA agarose [50% (vol/vol) suspension in the same buffer] at 4 °C for 3 h. After three washes with ice-cold 150 mM NaCl containing 25 mM Tris buffer, pH 8.0, 1 mM EDTA, 8 M urea, and 30 mM imidazole, SERT was eluted with 50 µL of 150 mM NaCl containing 25 mM Tris buffer, pH 8.0, 1 mM EDTA, 8 M urea, and 250 mM imidazole. A small portion of the eluate was counted with a Beckman LS6500 liquid scintillation counter, and the remainder was resolved by SDS/PAGE, and radiolabeled SERT was detected and quantified using a PharosFX Molecular Imager (Bio-Rad). To normalize for sample size, a parallel sample was subjected to immunoblot analysis using anti-SERT antibody, as described previously (16).

Biotinylation and Immunoblotting.

Surface expression of SERT in RBL-2H3 was determined using the membrane-impermeant biotinylation reagent sulfo-NHS-SS-biotin (Thermo Fisher Scientific) as described previously for HeLa cells (13). In brief, RBL cells were incubated twice with sulfo-NHS-SS-biotin for 20 min on ice. After labeling, the cells were rinsed with 100 mM glycine in PBS/CM for 20 min on ice to quench excess reagent. The cells were then lysed, and biotinylated proteins were recovered using streptavidin-agarose beads (Thermo Fisher Scientific) in an overnight incubation at 4 °C with gentle agitation. The beads were washed, and the biotinylated proteins were eluted with 50 µL of SDS/PAGE sample buffer. The eluate was applied to a 4–15% (wt/vol) gradient SDS-polyacrylamide gel and visualized by Western blot analysis. SERT phosphorylated at Thr-276 was detected using a rabbit polyclonal anti-SERT pThr276 (1:1,000). For total surface SERT expression, the blot was stripped with Restore Western Blot Stripping Buffer (Thermo Fisher Scientific) and reprobed with anti-SERT (C-term) antibody. Anti–γ-subunit of Fcε receptor antibody as a loading control was detected on a separate blot using a reduced amount of sample. An IRDye 680RD donkey anti-rabbit IgG (LI-COR; 1:10,000) was used to visualize the signal, which was determined by quantitative fluorescence imaging using the LI-COR Odyssey CLx Infrared Imaging System.

Accessibility Measurements.

Conformational changes were measured using the accessibility of cysteine residues placed in the cytoplasmic (S277C) and extracellular (S404C) permeation pathways as described previously (17). For measurement of extracellular pathway accessibility, intact cells expressing SERT C109A-S404C growing in 96-well plates were incubated for 15 min with 0.01 mM MTSEA and then washed with PBS/CM. Transport rates were then measured to determine the reactivity of S404C under the experimental conditions tested. For cytoplasmic pathway accessibility, membranes prepared from cells expressing SERT S277C-X5C were incubated on glass fiber filters for 15 min with 0.015 mM MTSEA and then washed with PBS/CM. The background construct, SERT X5C (C15A/C21A/C109A/C357A/C622A) (18), was designed to have minimal reactivity toward MTSEA. In this construct, five reactive Cys residues were replaced with alanine to decrease sensitivity MTSEA sensitivity relative to WT SERT. Binding of the high-affinity cocaine analog β-CIT was measured on the filters as described previously (13). Modification of Cys277 inactivates β-CIT binding by preventing closure of the cytoplasmic pathway and opening of the extracellular pathway (19).

Data Analysis.

Nonlinear regression fits of experimental and calculated data were performed with Origin software (OriginLab), which uses the Marquardt–Levenberg nonlinear least squares curve-fitting algorithm. The statistical analysis given was from multiple experiments. Data with error bars represent the mean ± SE for multiple experiments. Statistical analyses comparing control and experimental conditions were performed using Student's paired t test.

Results

PKG Activation Stimulates SERT Phosphorylation at Thr276.

Whereas mutation of SERT Thr276 to Ala abolishes PKG-dependent regulation of transport, direct evidence for phosphorylation of Thr276 has been lacking. To examine the characteristics of Thr276 as a phosphorylation site in SERT, we mutated this residue to alanine, serine, glycine, cysteine, aspartate, and glutamate, and tested the ability of these SERT mutants to respond to PKG activation by an increase in transport activity, phosphorylation, or both. The results, shown in Table 1, indicate that of the tested amino acid replacements, only serine was able to substitute for the native threonine. Although these results are consistent with the action of a serine/threonine kinase at Thr/Ser276, they do not rule out the possibility that a β-hydroxy amino acid is required in this position to allow phosphorylation at other positions. Consequently, we used an antibody raised against a SERT peptide containing phospho-Thr276.

Table 1.

Effect of amino acid substitutions at Thr276 on stimulation of transport activity and phosphorylation by 8-Br-cGMP

Amino acid	5-HT influx stimulation by cGMP, %	Phosphorylation, % of WT
Thr (WT)	36 ± 5	100 ± 6
Ser	39 ± 6	106 ± 13
Ala	5 ± 2	8.3 ± 1.3
Asp	4 ± 3	ND
Glu	−2 ± 3	ND
Gly	ND	9.7 ± 3.9
Cys	ND	8.4 ± 1.0

Transport and phosphorylation were measured in intact HeLa cells, and using membranes from HeLa cells, expressing the indicated mutants, where the amino acid shown replaced the native threonine at position 276. Transport and phosphorylation were measured as described in Materials and Methods. ND, not determined. Results are from three or four experiments for phosphorylation and transport measurements, respectively, and uncertainties are SEs. Relative to WT (threonine), stimulation of both transport and phosphorylation was significantly decreased (P < 2 ×10⁻⁴) for all amino acid replacements except serine, and the serine mutant was not significantly different from WT (P > 0.6).

As shown in Fig. 1A, using membranes from HeLa cells expressing FLAG-tagged WT SERT, purified PKG-Iα stimulated SERT phosphorylation from [γ-³³P]-ATP in the presence of cGMP. FLAG-tagged SERT T276A was not phosphorylated under the same conditions. The anti-SERT pThr276 antibody labeled SERT in Western blots from parallel incubations with unlabeled ATP when WT SERT was phosphorylated in the presence of cGMP, but not in the case of the T276A mutant (Fig. 1B). These results provide further evidence that Thr276 is the site of phosphorylation in response to PKG activation.

Fig. 1.

Thr-276 is essential for PKG-mediated phosphorylation. (A) In vitro [γ-³³P]ATP incorporation. Membranes prepared from HeLa cells expressing N-terminally His₆ and C-terminally FLAG-tagged hSERT (WT or T276A mutant) were incubated with [γ-³³P]ATP and purified PKG-1α in the absence or presence of 8-Br-cGMP at 30 °C for 30 min and then solubilized. SERT was captured by anti-FLAG agarose, eluted with FLAG peptide, and subsequently subjected to the second purification using Ni-NTA agarose after denaturation with 8 M urea. The 300 mM imidazole eluate was resolved by SDS/PAGE, and radiolabeled SERT was detected by phosphorimaging. (B) Western blot analysis of in vivo phosphorylation with anti-SERT pThr276. HeLa cells were transfected with cDNAs encoding hSERT (WT or T276A) and PKG-1α and then treated with 8-Br-cGMP at 37 °C for 30 min as indicated. SERT was solubilized and successively purified using anti-FLAG and Ni-NTA agarose. After treatment with PNGase F at 4 °C overnight, phosphorylated SERT was detected by immunoblotting with anti-SERT pThr276. The results shown in the bar graphs represent band density of radiolabeled SERT or SERT recognized by anti-SERT pThr276, respectively, from three or two independent experiments, respectively, for ³³P incorporation or antibody reactivity (mean ± SEM). Asterisks indicate values significantly different (P < 0.04) from control (no PKG). Other values were not significantly different from control (P > 0.20).

Because HeLa cells do not endogenously express SERT, and the results in Table 1 and Fig. 1 were obtained using an in vitro phosphorylation assay dependent on added PKG, we tested the ability of anti-SERT pThr276 to detect phosphorylation in RBL-2H3, a neurosecretory cell line that endogenously expresses SERT and all cellular components required for its phosphorylation in response to activation of adenosine A3 receptors on the cell surface. Treatment of RBL cells with the A3 agonist NECA, the NO donor SNAP, or 8-Br-cGMP led to an increase in phosphorylation as measured by reactivity toward anti-SERT pThr276 in Western blots (Fig. 2). 5-HT transport into the treated cells was increased as well (Fig. 2). The effect of each stimulus on both transport and phosphorylation was blocked by the PKG inhibitor Rp-8pCPT-cGMPS, indicating that each of these agents stimulates SERT transport and phosphorylation through activation of PKG.

Fig. 2.

Activation and PKG dependence of SERT phosphorylation and 5-HT influx in RBL-2H3 cells. RBL cells were treated with NECA (10 µM), SNAP (10 µM), or 8-Br-cGMP (10 µM) for 30 min, after a 30-min preincubation with or without Rp-8-pCPT-cGMPS (10 µM). Cell surface proteins were biotinylated with sulfo-NHS-SS-biotin, isolated with immobilized streptavidin and separated by SDS/PAGE. SERT, P-Thr276, and Fcε-γ were detected by quantitative immunoblotting using anti-SERT (C-term), anti-SERT pThr276, and anti-Fcε-γ antibody, respectively. The results represent surface SERT expression (open bars) and phosphorylation (light-gray bars) normalized to Fcε-γ surface levels, from three independent experiments (mean ± SEM). 5-HT influx was initiated by addition of [³H]5-HT to a final concentration of 20 nM. Transport (dark-gray bars) was measured in a 20-min incubation as described in Materials and Methods. Influx rates represent data combined from two experiments, each with triplicate measurements for each sample. Asterisks indicate statistically significant increases in phosphorylation or 5-HT influx (P < 0.05).

Stimulation of SERT activity by cGMP signaling in RBL cells and synaptosomes has been suggested to result from either an increase in SERT abundance on the cell surface (20) or an increase in catalytic efficiency (12). We examined the level of SERT cell surface expression in RBL and its response to NECA, SNAP, and cGMP. We determined surface expression by biotinylating surface proteins with the impermeant reagent sulfo-NHS-biotin, extracting labeled proteins using streptavidin agarose, and performing Western blot analysis with a C-terminal SERT antibody and, as an internal control, an antibody against the Fcε receptor γ-subunit. These treatments, which increased 5-HT transport activity and the level of SERT phosphorylation, did not induce a corresponding increase in surface expression. These results are consistent with a signaling pathway starting with the activation of A3 receptors, leading to production of NO by NOS, subsequent activation of guanylyl cyclase, and the action of cGMP on PKG, as originally proposed by Miller and Hoffman (10). Moreover, they indicate that SERT phosphorylation increases catalytic activity without an increase in SERT surface localization. Our previous results showed that PKG does not directly phosphorylate SERT, implicating the action of an unidentified kinase and potentially additional signaling components downstream from PKG (15).

Agents that Affect SERT Conformation Influence PKG-Dependent Phosphorylation.

To examine the effects of SERT conformation on PKG-dependent phosphorylation, we measured the phosphorylation of exogenously expressed SERT in purified membrane preparations in vitro with [γ-³³P]ATP as in Fig. 1A. Although added PKG is required for SERT phosphorylation under these conditions, our previous results indicate that PKG does not directly phosphorylate the transporter. Phosphorylation was determined under conditions that influence the distribution of SERT between inward-open and outward-open conformations. Na⁺ is known to stabilize SERT and other transporters in the NSS family in outward-open conformations (21–25). Fig. 3 shows that Na⁺ inhibited cGMP and PKG-dependent Thr276 phosphorylation to approximately one-half of the level found in the absence of Na⁺, with NMDG used as a substitute cation. Cocaine, which also stabilizes outward-open conformations, further decreased phosphorylation in the presence of Na⁺ (Fig. 3). In contrast, ibogaine, a noncompetitive SERT inhibitor that stabilizes inward-open conformations (17, 26, 27), dramatically increased Thr276 phosphorylation in the presence of Na⁺ (Fig. 3). 5-HT, the substrate for SERT, stimulated conversion of the transporter to inward-open conformations (19, 27, 28) and also stimulated phosphorylation, although not as strongly as ibogaine (Fig. 3).

Fig. 3.

Effect of cations and ligands on SERT phosphorylation. Membranes prepared from HeLa cells expressing WT hSERT were preincubated with cocaine (5 µM), ibogaine (20 µM), or 5-HT (1 µM) for 5 min at 20 °C before the addition of [γ-³³P]ATP, purified PKG-1α, and 8-Br-cGMP (100 µM) in buffer containing 150 mM NaCl or NMDG-Cl, as indicated. SERT was then subjected to two-step purification and resolved by SDS/PAGE, and radiolabeled SERT was detected by phosphorimaging as described in Materials and Methods. The results represent the relative integrated density of radiolabeled SERT from four independent experiments (mean ± SEM). Asterisks indicate statistically significant (P < 0.015) changes in integrated density of radiolabeled SERT compared with the sodium control (dashed line).

The effects of Na⁺, cocaine, ibogaine, and 5-HT on SERT phosphorylation could not be attributed to a direct effect on PKG. We measured in vitro phosphorylation of the PKG-specific peptide substrate RKRSRAE by purified PKG Iα with [γ-³³P]ATP as the phosphate donor. Na⁺, 5-HT, and SERT inhibitors, (cocaine, ibogaine) had minimal effects on PKG activity (Table 2).

Table 2.

Effect of agents that affect SERT phosphorylation on in vitro PKG kinase activity with peptide substrate

Condition	Kinase activity, pmol mg−1 min−1
Control (no cGMP)	84 ± 4
100 µM 8-Br-cGMP alone	745 ± 18
100 µM Rp-8pCPT-cGMPs	89 ± 2
150 mM Na+	705 ± 34
5 µM cocaine	652 ± 27
20 µM ibogaine	684 ± 24
1 µM 5-HT	719 ± 32

PKG activity against the peptide substrate (RKRSRAE) was determined in the presence of the indicated concentrations of various agents as described previously (15) and in Materials and Methods. Results are from two measurements in each of two experiments, and uncertainties are SEs. Relative to peptide phosphorylation in the presence of 8-Br-cGMP, PKG activity was significantly different only in the absence of cGMP and in the presence of the PKG inhibitor (P < 5 × 10⁻⁵), with all other values not significantly different from control (P > 0.2).

We tested the influence of SERT conformation on Thr276 phosphorylation in RBL cells to confirm that the effects were preserved in an intact cell system in which all components of the cGMP signaling pathway are endogenously expressed. This experiment used reactivity of anti-SERT pThr276 as an indicator of phosphorylation. The results in Fig. 4 show that 8-Br-cGMP modestly, but significantly increased SERT Thr276 phosphorylation in RBL cells in Na⁺ medium (Fig. 4, leftmost two columns). The phosphorylation level was further increased by ibogaine and 5-HT or replacement of Na⁺ by NMDG, and was inhibited by cocaine. We noticed an elevated basal level of phosphorylation in RBL cells (Fig. 4) relative to in vitro studies with SERT expressed in HeLa cell membranes (Figs. 1 and 3). Prolonged incubation with 10 µM Rp-8-pCPT-cGMPS or 10 µM 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one, a highly selective irreversible inhibitor of soluble guanylyl cyclase, did not decrease SERT phosphorylation in RBL cells below the level seen with Na⁺ buffer. These results suggest that PKG does not mediate basal phosphorylation of SERT in these cells, and imply the existence of one or more additional pathways leading to SERT Thr276 phosphorylation.

Fig. 4.

Conformation-dependent SERT phosphorylation in RBL-2H3 cells. RBL cells plated in 6-well plates were preincubated, where indicated, with ibogaine (20 µM), cocaine (5 µM), or 5-HT (1 µM) for 5 min before incubating with or without 8-Br-cGMP (10 µM) at 37 °C for 30 min in buffer containing 150 mM NaCl or NMDG-Cl, as indicated. The cells were then biotinylated with sulfo-NHS-SS-biotin and surface proteins isolated with immobilized streptavidin. Surface-expressed total SERT and P-Thr276 were detected by quantitative immunoblotting after SDS/PAGE using anti-SERT (C-term) and anti-SERT pThr276, respectively. The results represent Thr276 phosphorylation normalized to SERT surface expression from three independent experiments (mean ± SEM) and are expressed as a percentage of the unstimulated control. Asterisks indicate statistically significant phosphorylation changes in the absence (dashed line) or presence of 8-Br-cGMP (leftmost two columns; P = 0.004) or between Na⁺ plus cGMP (dotted line) and NMDG, Ibogaine, cocaine, or 5-HT (P < 0.03).

Cocaine and ibogaine compete for sites on SERT that are mutually exclusive (17, 26, 27). This competition is also apparent in the opposing effects of these drugs on PKG-induced SERT phosphorylation in vitro. Ibogaine (4 µM) was sufficient to increase phosphorylation by fivefold in the presence of 150 mM NaCl. The addition of cocaine antagonized this increase, with half-maximal inhibition occurring at concentrations <2 µM (Fig. 5A). Conversely, 5 µM cocaine was sufficient to inhibit phosphorylation by almost 90% in the absence of Na⁺. This inhibition was reversed by ibogaine, with half-maximal stimulation at <20 µM (Fig. 5B). These results support the idea that the competitive influence of cocaine and ibogaine on transporter conformation led to competitive antagonism in their effects on phosphorylation.

Fig. 5.

Competition between ibogaine and cocaine influences SERT phosphorylation. Membranes prepared from HeLa cells expressing WT hSERT were preincubated with ibogaine (4 µM in NaCl) (A) or cocaine (5 µM in NMDG-Cl) (B) in the presence of the indicated concentrations of cocaine (range, 0–20 µM) (A) or ibogaine (0–50 µM) (B), respectively, at 30 °C for 5 min, before the addition of [γ-³³P]ATP, purified PKG-1α, and 100 µM 8-Br-cGMP and continued incubation for 30 min. SERT was solubilized and purified with anti-FLAG agarose and Ni-NTA agarose as described in Materials and Methods. The 300 mM imidazole eluate was counted by liquid scintillation spectrometry. The results represent data combined from three experiments (mean ± SEM).

Na⁺, cocaine, ibogaine, and 5-HT are all known to influence SERT distribution between outward-open and inward-open states. Nevertheless, the experimental results described above do not rule out the possibility that other proteins in the cGMP-dependent signaling pathway leading to SERT phosphorylation could be sensitive to these agents and might mediate their effects. Accordingly, we compared the ability of Na⁺ and ibogaine to influence SERT conformation with their effect on PKG-induced phosphorylation in isolated membranes (Fig. 6). Conformation was measured by the reactivity of cysteine residues introduced at sites in the extracellular or cytoplasmic permeation pathways. In intact cells, we measured the accessibility of a cysteine residue at position 404 in the extracellular pathway. The solid line in Fig. 6A shows the extent of modification by the aqueous reagent MTSEA at varying [Na⁺], as measured by inactivation of 5-HT transport. The dashed line shows the influence of Na⁺ on phosphorylation under the same conditions. In Fig. 6B, we measured the conformational effect of ibogaine in isolated membranes by following the reactivity of a cysteine residue in the cytoplasmic pathway at position 277. Ibogaine stabilized inward-open SERT conformations in which Cys277 was more accessible to modification by MTSEA, leading to decreased SERT binding activity. The concentration dependencies for the conformational responses to Na⁺ and ibogaine were remarkably similar to those of phosphorylation (Fig. 6). These observations are consistent with the proposal that SERT conformation controls PKG-dependent phosphorylation at Thr276.

Fig. 6.

Parallel effects of Na⁺ and ibogaine on phosphorylation and conformation. (A) Effect of Na⁺ on phosphorylation and extracellular pathway conformation. HeLa cells expressing SERT C109A-S404C were treated with 0.01 mM MTSEA in Hepes buffer, pH 7.4, containing the indicated Na⁺ concentrations at 20 °C for 15 min. Residual MTSEA was removed by washing with PBS/CM, and 5-HT uptake was measured as described in Materials and Methods. Results are shown for one of three experiments with similar results. The data are mean ± SD values of triplicates in the experiment shown. Phosphorylation was measured using membranes prepared from HeLa cells expressing WT SERT that were incubated with [γ-³³P]ATP, purified PKG-1α, and 100 µM 8-Br-cGMP and the indicated concentrations of Na⁺ at 30 °C for 30 min. SERT phosphorylation was determined as described in Materials and Methods and represent combined data from three experiments (mean ± SEM). (B) Effect of ibogaine on phosphorylation and cytoplasmic pathway conformation. Membranes from HeLa cells expressing SERT S277C-X5C were treated with 0.01 mM MTSEA and the indicated ibogaine concentrations for 15 min at 20 °C, washed, and assayed for residual β-CIT binding as described in Materials and Methods. Results are shown for one of three experiments with similar results. The data are mean ± SD values of triplicates in the representative experiment. Phosphorylation (combined results from three experiments) was measured as in A, in the presence of the indicated ibogaine concentrations.

To evaluate the effect of NaCl and ibogaine on the region of TM5 surrounding T276, we measured the rate constant for inactivation of SERT mutants by MTSES. The results, illustrated in Fig. 7, show that G273C, with a cysteine residue in the second intracellular loop, and S277C react fairly rapidly with this cysteine reagent. Cysteine replacements of V274 through T276 were much less reactive, as was V278C. The reactivity of most of these mutants was increased by 20 µM ibogaine or by replacement of NaCl with NMDG-SO₄; however, the increase was similar through this region, and the helical pattern of reactivity was not measurably altered by these treatments.

Fig. 7.

Cysteine accessibility scan of residues in the vicinity of Thr276. Membranes from HeLa cells expressing SERT mutants with Cys replacing each residue from Gly273 through Val278 were incubated with a range of MTSES concentrations to determine the rate constant for Cys modification (17). Rates were determined in 200 mM NaCl with and without 20 µM ibogaine and in 200 mM NMDG-SO₄. Asterisks indicate values significantly different (P < 0.05) from NaCl. (Inset) Mutants K275C and T276C shown on an expanded scale.

Discussion

The principal conclusion that can be drawn from our results is that SERT is more susceptible to cGMP-stimulated phosphorylation at Thr276 when in an inward-facing, as opposed to an outward-facing conformation. Because SERT shuttles between such conformations when transporting 5-HT across the membrane, this conclusion has direct implications concerning the effect of 5-HT transport on the phosphorylation process and its consequences for SERT activity.

In the absence of extracellular 5-HT, SERT is likely to be in an outward-facing conformation. The presence of extracellular Na⁺ stabilizes outward-facing states of SERT (24), and the subsequent addition of 5-HT leads to a synchronous conformational change associated with a transient current (28). By analogy with prokaryotic SERT homologs, this substrate-dependent change represents opening of a cytoplasmic substrate permeation pathway (21, 29) that also has been observed with SERT (19) and leads to a release of substrate to the cytoplasm. Subsequently, SERT returns to an outward-facing conformation in a K⁺-stimulated step that is rate-limiting for transport (28, 30). Because the transition from inward-facing to outward-facing conformations is the slow step in the transport cycle, SERT will accumulate in an inward-facing conformation when transporting 5-HT. Thus, SERT is more likely to be in an outward-facing conformation in the absence of 5-HT and in an inward-facing conformation in the presence of 5-HT.

The consequence of cGMP-dependent SERT phosphorylation is to increase SERT activity. This increase has been proposed to result from either increased catalytic activity (9, 12) or increased expression of SERT on the cell surface (31). The data in Fig. 2 show that stimulation of cGMP signaling increased SERT transport and phosphorylation in RBL cells with little effect on surface expression. Moreover, Table 3 clearly shows that SERT surface expression in RBL cells was not affected by 8-Br-cGMP alone or by agents (i.e., Na⁺, 5-HT, cocaine, or ibogaine) that modified its effect on SERT phosphorylation. These results suggest that the increase in SERT activity associated with cGMP-stimulated phosphorylation results from an increase in the transport activity of SERT molecules preexisting on the cell surface. We do not yet understand the mechanism by which phosphorylation of Thr276 increases transport activity, but TM5, which is the likely location of this residue, lines the substrate permeation pathway, which opens and closes during the transport cycle, and its modification is likely to alter conformational dynamics of the transporter.

Table 3.

Effect of agents that affect SERT phosphorylation on SERT surface expression

Condition	SERT surface expression, relative to no cGMP
8-Br-cGMP	108 ± 7
cGMP, no Na+	100 ± 1
cGMP + ibogaine	101 ± 7
cGMP + cocaine	96 ± 7
cGMP + 5-HT	97 ± 6

Measurements of SERT surface expression in RBL cells were made as in Fig. 2 and described in Materials and Methods. The indicated reagents were used at the following concentrations: 8-Br-cGMP, 10 µM; ibogaine, 20 µM; cocaine, 5 µM; 5-HT, 1 µM. In the Na⁺-free sample, NMDG was used to replace 150 mM Na⁺. Results represent four experiments and uncertainties are SEs. None of the treatments significantly altered surface expression relative to the control (P > 0.25).

Fig. 2 also shows that phosphorylation stimulated by an adenosine A3 receptor agonist, an NO donor, or by 8-Br-cGMP was inhibited in each case by the PKG inhibitor Rp-8-pCPT-cGMPS, implicating PKG activation as a key step in the phosphorylation process. The results are consistent with A3 receptor activation leading to activation of nNOS, presumably through increased cytosolic Ca²⁺ levels. The resulting NO would stimulate soluble guanylyl cyclase to produce cGMP, thereby activating PKG, as proposed by Miller and Hoffman (10). Our previous work showed that PKG does not directly phosphorylate SERT and that an additional, unidentified kinase is responsible (15). The significant level of basal SERT phosphorylation in RBL cells, shown in Figs. 2 and 4, was insensitive to the inhibition of cGMP signaling. These observations may result from another pathway leading to SERT phosphorylation that is independent of cGMP signaling and may converge with the cGMP pathway at the level of the unidentified SERT kinase.

Both in vitro and in RBL cells, ibogaine was more effective than 5-HT in stimulating phosphorylation. We can only speculate on the basis of this difference. However, at concentrations that saturate the transport rate, 5-HT may generate the phosphorylatable intermediate only transiently, whereas ibogaine likely forms a dead-end complex with the transporter, trapping it more effectively in an inward-open conformation. Moreover, there are many potential inward-open SERT conformations, because 5-HT, Na⁺, and Cl⁻ all need to dissociate and K⁺ needs to bind before SERT returns to an outward-open state. We do not know which of these ibogaine stabilizes to increase phosphorylation.

The effect of SERT conformation on cGMP-dependent phosphorylation suggests a novel role for this process in the control of 5-HT homeostasis. An increase in serotonergic neurotransmission will produce more extracellular 5-HT in terminals and varicosities where 5-HT is released. As the primary substrate for SERT, 5-HT converts the transporter from an outward-facing to an inward-facing conformation, increasing its availability to serve as a substrate for phosphorylation, as shown in Figs. 3 and 4 in assays in vitro and in intact RBL cells. The increased activity of phosphorylated SERT allows more rapid reuptake of released 5-HT, potentially compensating for the increased release.

Another unique aspect of this signaling process comes from the mechanism by which SERT conformation controls phosphorylation. The phosphorylation site was proposed to be Thr276, and data presented in Table 1 and Fig. 1 provide additional evidence confirming this site. However, evidence suggests that this residue is in a relatively inaccessible position in TM5 (13, 26), where it would seem unlikely to serve as a kinase substrate. Structural analysis of more than a dozen kinase-substrate complexes has repeatedly shown that peptide substrates bind in an extended conformation (32, 33), unlike the helical structure that we identified for the region of SERT TM5 containing Thr276 (13). TM5 is also one of the helices lining the permeation pathway by which 5-HT and ions are delivered to the cytoplasm, however (19), and as such, its accessibility to solvent increases dramatically in inward-facing conformations. The homologous bacterial transporter MhsT was recently crystallized in an inward-occluded conformation in which the cytoplasmic half of TM5 had lost considerable helical character between Gly171 and Pro181 (14), which are conserved as Gly278 and Pro288 in SERT. Thus, a probable mechanism for Thr276 phosphorylation emerges in which unwinding of the TM5 helix in inward-facing conformations allows the sequence containing Thr276 to extend into the cytoplasm, where it can be phosphorylated.

Although phosphorylation sites typically fall within unstructured regions and exposed loops (34), the requirement for a conformational change to facilitate accessibility to a phosphorylation site has been observed occasionally in other systems. For example, Ser51, the regulatory phosphorylation site in eukaryotic initiation factor 2α (eIF2α), is sequestered within an α-helix, but interaction with its kinase PKR promotes melting of the helix to allow access to the phosphoacceptor residue (35, 36). Similarly, bifunctional isocitrate dehydrogenase kinase/phosphatases in bacteria appear to induce a conformational rearrangement in their substrates that liberates a phosphorylation site residue that is normally buried and thus inaccessible (37). In these cases, the presence of the phosphorylation site in a folded helix serves as a mechanism for achieving a high degree of specificity for a particular phosphorylating kinase. In contrast, conformational rearrangement of SERT leading to exposure of its phosphorylation site is facilitated by substrate binding and translocation, potentially as a mechanism to promote SERT activity in the presence of substrate. Ibogaine and NaCl removal increased MTSES reactivity in this region (Fig. 7), and interaction with the SERT Thr276 kinase also may contribute to an unwinding of the helix to facilitate phosphorylation. However, the overall pattern of accessibility did not change as accessibility increased (Fig. 7), suggesting that helix unwinding affects only a small fraction of SERT at any given time.

Because SERT mutations, such as I425V, that affect cGMP-dependent phosphorylation are associated with several psychiatric disorders, it seems likely that this signaling pathway plays an important role in 5-HT homeostasis in vivo. The knowledge that agents controlling SERT conformation can modulate cGMP signaling provides potential ways to manipulate the pathway pharmacologically. For example, SERT-selective compounds that act like ibogaine could enhance SERT phosphorylation, and inhibitors of the as-yet unidentified kinase that phosphorylates SERT at Thr276 could be used to inhibit the process. Because SERT is the target of so many therapeutic drugs, all of which are competitive inhibitors that stabilize outward-facing conformations, it may be worth considering agents that influence the function of SERT by favoring alternative conformations that enhance cGMP-dependent phosphorylation.