Inhibitory glycinergic neurotransmission is mediated by the glycine receptor (GlyR), which is a ligand-gated chloride channel of the Cys-loop family. The heteromeric receptor consists of two α (α1–α4) and three β subunits (Grudzinska et al.,2005), having four glycine and two strychnine binding sites. A defining feature of this family is the presence of an extracellular “Cys-loop” structure, which is important for ligand binding. Also, Cys-loop receptors possess an intracellular loop linking TM3-TM4 that has several putative sites of modulation by protein kinases (Kittler and Moss,2003; Ruiz-Gómez et al.,1991; Swope et al.,1999; Wiesner and Fuhrer,2006). In this regard, Ruiz-Gómez et al. (1991) reported protein kinases C (PKC)-dependent serine S391 phosphorylation of the α1-subunit in the GlyR purified from the spinal cord. In addition, glycine currents are decreased by PKC activation in spinal cord neurons (Albarrán et al.,2001; Tapia et al.,1997), as well as in recombinant GlyRs (Albarrán et al.,2001; Tapia et al.,1997; Vaello et al.,1994); these currents increased in neurons isolated from different brain regions (Nabekura et al., 1996; Schönrock and Bormann,1995; Xu et al.,1996). Protein kinase A (PKA) activation increased glycine currents in the spinal cord (Song and Huang,1990; Tapia et al.,1997; Vaello et al.,1994) and decreased those obtained in substantia nigra neurons (Inomata et al.,1993). PKA down-regulates the open probability of the GlyR in ventromedial hippocampal neurons (Agopyan et al.,1993), exerting the opposite effect in trigeminus neurons (Song and Huang,1990). In retina, both PKC and PKA activation decreased the decay time constant of glycine currents (Han and Slaughter,1998); however, the mechanism is unknown. In order to better understand the mechanism of GlyR modulation in retinal synaptic transmission, we analyzed the effect of PKA and PKC activation on GlyR specific radioligand binding, its phosphorylation, and its internalization in the intact rat retina.
MATERIALS AND METHODS
Adult male and female Long-Evans rats (170–200 g) were maintained under dark-light conditions (12:12 h). Long-Evans rats were inbred from a colony originally obtained from Charles River (Raleigh). All animals (n = 100) were handled according to the Association Research in Vision and Ophthalmology Statement on the Use of Animals in Ophthalmic and Vision Research. All efforts were made to minimize the number of animals used and their suffering.
[3H]glycine (56.4–60.0 Ci/mmol) and [3H]strychnine (23 Ci/mmol) were purchased from NEN Life Science Products (Boston, MA). Monoclonal mouse antibody (Cat. No. 146 011) clone mAb4a, against rat spinal cord glycine receptor (recognizes all GlyR subunits), was obtained from Synaptic Systems (Göttingen, Germany); polyclonal rabbit anti-GlyR (GlyR α H-70, Cat. No. s.c.-33,611) Santa Cruz Biotechnology (Santa Cruz, CA); integrin β1 (M-106): s.c.-8978 is a rabbit polyclonal antibody that was also purchased from Santa Cruz Biotechnology; antiphosphoserine (mouse monoclonal IgG1, clone 4A4, Cat. No. 05-1000) from Millipore (Billerica, MA); mouse monoclonal antibody to rab11A (ab78337) was used (Abcam; Cambridge, UK). Protein A-Sepharose CL-4B was obtained from Sigma-Aldrich (St. Louis, MO).
Secondary antibodies, ECL antimouse IgG, peroxidase-linked, and ECL antirabbit IgG, peroxidase-linked were both purchased from GE Healthcare (Little Chalfont, Buckinghamshire, UK). Membranes of polyvinylidene difluoride (PVDF) were used (Millipore). Glycine, strychnine, cytochalasin D (Cyt D), phorbol 12-myristate 13-acetate (PMA), bisindolylmaleimide VII (Bis VII), 8-bromoadenosine-3′, 5′-cyclic monophosphate (8-Br-cAMP), forskolin, and H89 were purchased from Sigma-Aldrich.
Radioligand binding assay
Retinae from 2 h-dark-adapted rats were incubated at (20 ± 1)°C for 40 min in Krebs medium (mM: choline chloride 118; KCl 4.7; CaCl2 2.5; MgSO4 1.17; glucose 5.6; KHCO3 25) containing different concentrations of [3H]glycine (25–600 nM) or [3H]strychnine (10–200 nM) and then were incubated in the presence of 1 mM glycine or 100 μM strychnine for additional 40 min to determine the specific radioligand binding. After incubation, the tissue was rinsed with cold medium, weighed, and dissolved in 0.5 ml of 1% (v/v) of Sodium dodecyl sulfate (SDS). Radioactivity in the solubilized tissue was determined by liquid scintillation counting. Specific binding was defined as the binding displaced by 1 mM or 100 μM nonradioactive glycine or strychnine, respectively (Pérez-León and Salceda,1995).
Protein kinases effect on radioligands binding
To determine the effect of protein kinases on the specific binding, PKC or PKA activators (PMA, 8-Br-cAMP, and forskolin) and inhibitors (BisVII and H89) were added at the same time than radioligands and the binding was determined after 40 min. The whole experimental procedure was carried out under dim red light conditions.
Immunoprecipitation and immunoblotting
A retina (incubated as described above for the radioligand binding assay) was transferred into a lysis buffer (1:3 (p/v): RIPA-Tris buffer (mM: EGTA 2; NaCl 316; Na2MoO4 20; NaF 50; Tris-HCl 20; Na3VO4 100, PMSF 100 and EDTA 100; 0.1% of leupeptine and aprotinine; SDS 0.2% and Tritón-X100 2%) and maintained under constant shaking for 3 h at 4°C. Subsequently, the sample was centrifuged for 10 min at 20,800g and the supernatant (50 μg of protein) was incubated in the presence of 10 μg of anti-GlyR H-70 (immunoprecipitating antibody), previously coupled to sepharose pearls-protein A (0.5 mg/5 μl). Then, the sample was incubated at 4°C under constant shaking for 12–15 h and centrifuged at 20,800g for 2 min; the obtained pellet was washed three times with lysis buffer (without SDS but with 1% of triton X-100). The protein A-anti-GlyR H70-GlyR complex was denatured in Laemmli's sample buffer (Laemmli,1970), resolved through 12% SDS polyacrilamide gels and electroblotted to PVDF membranes. Blots were stained with Ponceau S to confirm that protein loading was the same in all lanes. Membranes were soaked in PBS to remove the Ponceau S and incubated for 90 min in Tris-buffered saline (TBS) containing 5% dried skimmed milk and 0.1% Tween 20 to block the nonspecific protein-binding sites. Afterwards, membranes were incubated for 14 h at 4°C with the primary antibody (mAb4a or antiphosphoserine) diluted in BSA 0.25%, Tween 20 0.1%, thimerosal 0.01% in TBS buffer. Then they were washed and incubated with the secondary antibodies (ECL antimouse or ECL antirabbit). The protein was detected using an ECL Western blot detection kit (Millipore). The mAb4a antibody revealed a band of 48–50 kDa (GlyR). The blots were subjected to a densitometry analysis and data were analyzed using GraphPad Prism5 software (San Diego, CA).
Subcellular fractionation was performed according to McKeel and Jarett's protocol (1970) with the following modifications. Retinae (n = 30), incubated under the same conditions described for the immunoprecipitation assay, were homogenized in buffer A (Tris-HCl 0.01 M, pH 7.6; sucrose 0.25 M, EGTA 1 mM, benzamidine 2 mM, bacitrin 1 mg/ml, soybean trypsine inhibitor 0.1 mg/ml, pepstatin 10 μg/ml, leupeptin 20 μg/ml, and aprotinin 20 U/ml), and centrifuged at 180g for 10 min; the pellet was washed three more times. The mixed supernatants were centrifuged for 10 min at 2000g. The resulting supernatant was centrifuged at 12,000g for 20 min to obtain: mitochondriae, synaptosomes, and plasma membranes (P1). The supernatant obtained was centrifuged at 61,000g for 25 min to obtain a plasma membrane component (P2) remaining in this fraction. Microsomal membranes were sedimented at 187,000g for 70 min; simultaneously, P1 was placed in a sucrose gradient (0.3 M: 0.9 M: 1.2 M) and centrifuged for 30 min at 53,000g to obtain the plasma membrane fraction, which was mixed with P2. Both, the microsomal and plasma membranes, were solubilized in RIPA-Tris buffer, immunoprecipated and immunobloted as described above. Subcellular fractions were characterized by determining the activity of marker enzymes: cytochrome-c reductase (Rickwood,1993), alkaline phosphatase (Sigma procedure no. 104), and glutamate dehydrogenase (Schmidt and Schmidt,1983), and the presence of plasma membrane (integrin β1) and microsomal (Rab 11a) marker proteins was demonstrated by Western blot.
Protein concentration was determined with a commercial kit (Bio-Rad Headquarters Hercules, CA) using bovine serum albumin as the standard (Lowry et al.,1951 ).
After Scatchard analysis (Segel,1975), the sigmoidal behavior of radioligand binding was studied by different equations; the best fit was obtained by the Hill equation (Y = BmaxXh/(Kdh + Xh) and the allosteric Monod–Wyman–Changeux model.
Statistical significance was determined by either the t-test or one-way ANOVA analysis, followed by the Dunnett's post hoc test.
Kinetic characterization of the specific [3H]glycine and [3H]strychnine binding in the intact rat retina
Since radioligand binding assay allows us to determine the proportion of surface membrane receptors, we characterized the specific [3H]glycine and [3H]strychnine binding to the GlyR in intact rat retina. To determine the conditions in which the specific binding reaches equilibrium, we incubated whole rat retina for different time periods in the presence of the radioligand. Under these conditions, specific [3H]glycine or [3H]strychnine binding increased linearly, reaching a plateau at 30 min and remained constant up to 90 min; therefore, kinetic assays were performed at 40 min incubation. Specific [3H]glycine binding constituted 20% of the total binding (Fig. 1A). The specific [3H]glycine kinetic binding established at different radioligand concentrations (25–600 nM) revealed a sigmoidal behavior with a Kd of 212 ± 11 nM and three binding sites (Fig. 2A).
Specific [3H]strychnine binding represented 25% of the total radioligand binding (Fig. 1B), exhibiting a sigmoidal shape with a Kd = 50 ± 5 nM and two binding sites (Fig. 2B).
Activation of PKC and PKA decreased the specific binding of [3H]glycine and [3H]strychnine to GlyR
Activation of PKC by PMA decreased the specific [3H]glycine and [3H]strychnine binding by 60% and 85%, respectively (Fig. 3). The specific PKC inhibitor BisVII (0.5 μM) prevented the PMA effect on radioligand binding. Likewise, PKA activation by 8-Br-cAMP (1 mM) or by forskolin (10 μM), also reduced specific radioligand binding significantly; 8-Br-cAMP effect was blocked with H89 (1 μM), a specific PKA inhibitor (Fig. 3).
The decrease of specific radioligand binding could be associated with changes on the cell surface in the GlyR proportion, as well as modifications in ligand binding properties; therefore, we proved if the GlyR affinity and the maximum binding sites (Bmax) for [3H]glycine and [3H]strychnine were modified in response to protein kinases activation. The Kd for the radioligands binding did not change in the presence of 8-Br-cAMP or PMA (Kd of 160 nM); however, activation of protein kinases led to a reduction in the Bmax of both radioligands, without affecting their sigmoidal behavior (Fig. 2). Moreover, Cyt D (10 μM), which inhibits actin polymerization required for endocytosis, blocked the effect of both kinases on the specific radioligands binding (Fig. 3). The Bmax reduction, as well as the effect of Cyt D, suggests changes in the number of GlyRs in the plasma membrane.
It is known that GlyR is phosphorylated in response to PKA and/or PKC activation in the spinal cord (Harvey et al.,2004; Ruiz-Gómez et al.,1991; Vaello et al.,1994). Our results revealed that retinal GlyR presented basal serine phosphorylation, which was enhanced in a time-dependent manner, following tissue incubation in the presence of PMA or 8-Br-cAMP. The phosphorylation state of GlyR increased as early as at 2 min incubation and reached its highest level at 10–20 min, decreasing thereafter (Fig. 4). Specific inhibition of PKA or PKC prevented GlyR phosphorylation without altering the basal one (Fig. 5).
Membrane receptor endocytosis is a well recognized mechanism that is involved in the down-regulation of transmembrane receptors. In order to determine whether GlyR is internalized in response to PKC and PKA activation, we analyzed the proportion of GlyR in plasma membrane and microsomal fractions. At 10 min incubation, we did not detect changes in the GlyR proportion in subcellular fractions compared with the control condition (data not shown). However, PKA activation decreased plasma membrane GlyR by (54 ± 8)% and increased it by (50 ± 3)% in the microsomal fraction at 40 min incubation (Fig. 6A). In addition, PKC activation also reduced the GlyR proportion in plasma membrane (41 ± 9)%, an effect that was blocked by BisVII (106 ± 9)%; however, microsomal GlyR proportion was neither altered in response to PKC activation nor in the presence of BisVII (Fig. 6B).
Glycine is the principal inhibitory neurotransmitter in the spinal cord and brain stem, exerting an inhibitory effect in the inner plexiform layer of the mammalian retina through the GlyR. In retinal ganglion cells, it was shown that PKC and PKA activation modified glycine currents (Han and Slaughter,1998; Zhao et al.,2010), a process that could affect the excitatory neurotransmission in the retina. In order to elucidate the mechanisms controlling GlyR in response to protein kinases activation, we studied the effect of PKA and PKC activation on the specific [3H]glycine and [3H]strychnine binding to GlyR, as well as receptor phosphorylation and internalization induced by these kinases in the intact retina. Our results showed that [3H]glycine and [3H]strychnine binding to the GlyR has similar affinity to that reported in the spinal cord (Kishimoto et al.,1981; Ruiz-Gómez et al.,1989; Young and Snyder,1973) and the one found in retinal membranes (Borbe et al.,1981; Pérez-León and Salceda,1995; Schaeffer and Anderson,1981). However, kinetic binding analysis revealed a sigmoidal shape that was fitted to the allosteric Wyman–Monod–Changeux model, as well as to the Hill equation, displaying Hill coefficients (nH) of 3 and 2 for glycine and strychnine binding, which suggests a cooperative system (Segel,1975). Hill coefficients might indicate at least two binding sites acting together (Segel,1975). These results are consistent with the electrophysiological response to glycine, in which the maximum gating efficacy of the GlyR is reached when three potential binding sites are occupied (Beato et al.,2004). These sites could also be related to the expression of different α GlyR subunits (α1–α4, β) identified in the mammalian retina (Greferath et al.,1994; Heinze et al.,2007; Wässle et al.,2009). Under physiological conditions in which glycine concentrations in the synaptic cleft can change significantly, the cooperative behavior of the GlyR would allowed a better control of the excitatory neurotransmission.
Currently, there are some studies that indicate that spinal cord GlyRs are phosphorylated by both PKA and PKC (Harvey et al.,2004; Ruiz-Gómez et al.,1991; Vaello et al.,1994). In addition, electrophysiological evidence indicates that glycine currents are regulated by PKA and PKC in spinal cord (Harvey et al.,2004; Vaello et al.,1994), different brain regions (Nabekura et al., 1996; Schönrock and Bormann,1995; Xu et al.,1996), and retina (Han and Slaughter,1998; Zhao et al.,2010). In this study, we show that activation of both kinases specifically decreased the radioligand binding to GlyR in the entire retina. Kinetic binding analysis revealed that PKA and PKC activation dramatically decreased the Bmax of the specific radioligands binding without altering the affinity of the GlyR for them, pointing to a decrease of the receptor on the cell surface. In agreement to this interpretation, Cyt D blocked the effect of kinases activation exerted on specific binding, indicating that GlyR is endocytosed in retina in response to PKA and PKC. In this respect, activation of both kinases resulted in the decrease of the GlyR in the plasma membrane with an increase in the microsomal fraction when PKA was activated, but not by PKC activation. Since PKA and PKC act on specific consense amino acid sequences, GlyR phosphorylation for these kinases could represent different regulatory mechanisms. Further studies are required to identify the significance of GlyR phosphorylation by PKC.
Phosphorylation of G protein-coupled receptors at specific sites is a process that regulates their function by different mechanisms such as internalization (Alcántara-Hernández and García-Sáinz,2005; Brandon et al.,2000; Kovacs et al.,2009; Pippig et al.,1993; Tang et al.,1998; Terunuma et al.,2010; Willoughby et al.,2007). It has also been shown that Cys-loop receptors, such as γ-aminobutyric acid (GABAA), are phosphorylated by protein kinases, inducing their internalization (Chen et al.,2006; Kumar et al.,2010). Likewise, our results indicate that GlyR is regulated by similar mechanisms; however, evidence demonstrating that GlyR phosphorylation is required to induce GlyR internalization is necessary. In a physiological manner, GlyR modulation by protein kinases could occur through a cross-talk stimulation of G protein-coupled receptors. In fact, melatonin MT2, prostaglandin E2, and serotonin type 1A receptors activation modify glycine currents in different systems (Ahmadi et al.,2002; Manzke et al., 2010; Zhao et al.,2010).
In summary, our results showed, for the first time, that mammalian retinal GlyR is phosphorylated and internalized in response to PKA activation. Physiologically, activation of PKA would decrease GlyR at the cell surface which in turn would control the synaptic action of glycine. Although GlyR is phosphorylated by PKC activation, its significance is not yet understood. Further studies are necessary to determine the cross-talk mechanisms activating these protein kinases.
The authors thank G. Sánchez-Chávez for his technical assistance, Dr. M.E. Torres-Márquez for the academic support provided during this study, and Dr. G. Reyes-Cruz and Dr. D. Escalante-Alcalde for helping in supplying some of the antibodies used. This work was performed in partial fulfillment of the requirements for the PhD degree in Biomedical Sciences of Miguel A. Velázquez-Flores at the Universidad Nacional Autónoma de México.