Single amino acid residue in the M4 domain of GluN1 subunit regulates the surface delivery of NMDA receptors

Authors


Address correspondence and reprint requests to Martin Horak, Institute of Physiology, Academy of Sciences of the Czech Republic v.v.i., Videnska 1083, 14220 Prague 4, Czech Republic.

E-mail: mhorak@biomed.cas.cz

Abstract

N-methyl-d-aspartate (NMDA) receptors are glutamate ion channels that are critically involved in excitatory synaptic transmission and plasticity. The functional NMDA receptor is a heterotetramer composed mainly of GluN1 and GluN2 subunits. It is generally thought that only correctly assembled NMDA receptors can pass the quality control checkpoint in the endoplasmic reticulum (ER) and are transported to the cell surface membranes. The molecular mechanisms underlying these processes remain poorly understood. Using chimeric and mutated GluN1 subunits expressed in heterologous cells, we identified a single amino acid residue within the fourth membrane domain (M4) of GluN1 subunit, L830, that regulates the surface number of NMDA receptors. Our experiments show that this residue is not critical for the interaction between GluN1 and GluN2 subunits or for the formation of functional receptors, but rather that it regulates the forward trafficking of the NMDA receptors. The surface expression of both GluN2A- and GluN2B-containing receptors is regulated by the L830 residue in a similar manner. We also found that the L830 residue is not involved in the trafficking of individually expressed GluN1 subunits. Our data reveal a critical role of the single amino acid residue within the GluN1 M4 domain in the surface delivery of functional NMDA receptors.

Abbreviations used
AChR

acetylcholine receptor

AMPA

2-amino-3-(5-methyl-3-oxo-1,2-oxazol-4-yl)propanoic acid

BSA

bovine serum albumin

COS-7

African Green Monkey kidney fibroblast cell line

DOC

deoxycholate

ER

endoplasmic reticulum

FBS

fetal bovine serum

GA

Golgi apparatus

GFP

green fluorescent protein

HEK293

Human Embryonic Kidney 293 cells

HRP

horseradish peroxidase

M

membrane domain

NMDA

N-methyl-D-aspartate

PBS

phosphate-buffered saline

TBS

Tris-buffered saline

TX-100

Triton X-100

YFP

yellow fluorescent protein

N-methyl-d-aspartate (NMDA) receptors are a subclass of the family of ionotropic glutamate receptors that play critical roles in excitatory synaptic transmission (Traynelis et al. 2010). The abnormal functioning of NMDA receptors is implicated in various neurological and psychiatric disorders such as epilepsy and schizophrenia. The functional NMDA receptors are likely heterotetramers composed of two GluN1 subunits and two GluN2 and/or GluN3 subunits (Lau & Zukin 2007, Traynelis et al. 2010). There are eight different splice variants of the GluN1 genes, four GluN2 genes (GluN2A-D), and two GluN3 genes (GluN3A-B). All subunits share the same membrane topology with four membrane domains (M1-M4), an extracellular N-terminal region, and an intracellular C-terminal region. Although much is known about the functional and pharmacological properties of surface NMDA receptors, the molecular mechanisms underlying their assembly, endoplasmic reticulum (ER) processing, and intracellular trafficking are poorly understood.

The most abundant GluN1 splice variant, GluN1-1a, and the GluN2 subunits are retained in the unassembled form in the ER of heterologous and neuronal cells, while expressed on the cell surface as GluN1/GluN2 heterotetramers (McIlhinney et al. 1998; Okabe et al. 1999; Fukaya et al. 2003). This suggests the presence of the ER retention signals within the NMDA receptor subunits as well as the specific regulatory mechanism(s) controlling the release of correctly folded receptors from the ER. Previous studies have shown that multiple regions control the ER retention of NMDA receptor subunits including the C-termini of some GluN1 splice variants (Standley et al. 2000; Horak et al. 2008), the C-terminus of the GluN2B subunit (Hawkins et al. 2004), the glycine binding site within the GluN1 subunit (Kenny et al. 2009), the N-terminal domain within the GluN2A subunit (Qiu et al. 2009), and the third membrane domains of both GluN1 and GluN2B subunits (Horak et al. 2008). However, it remains unclear how these regions collaborate so that only correctly assembled NMDA receptors are trafficked to the cell surface.

We have shown previously that the presence of the M4 domain of the GluN1 subunit is critical for release of NMDA receptors from the ER, likely by negating the ER retention signals present within the GluN1 and GluN2B M3 domains (Horak et al. 2008). In this study, we investigated that the mechanism by which the GluN1 M4 domain regulates the surface expression of NMDA receptors. Using numerous chimeric/mutated GluN1 subunits expressed in heterologous cells, we identified a single amino acid residue within the second half of the GluN1 M4 domain, L830 that regulates the surface targeting of full-length NMDA receptors. Our biochemical and electrophysiological experiments showed that the presence of the L830 residue is not critical for the GluN1/GluN2 interaction or for the formation of functional receptors. We conclude that the L830 residue is involved in the ER processing and/or forward trafficking of functional NMDA receptors.

Materials and methods

All experiments were performed in accordance with relevant guidelines and regulations of the Institute of Physiology, Academy of Sciences of the Czech Republic v.v.i.

Molecular biology

The yellow fluorescent protein (YFP)-tagged GluN1-1a subunit and the corresponding variant with stop codon located several amino acid residues after the M4 domains (YFP-GluN1ΔCt) and untagged and MYC-tagged versions of GluN2B subunits were described previously (Horak et al. 2008). To generate chimeric and mutated GluN1 subunits, the Quick-Change site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) was used according to the manufacturer's instructions. Amino acid residues are numbered as published (Ishii et al. 1993). All constructs were verified by DNA sequencing.

Heterologous cell culture

The African Green Monkey kidney fibroblast (COS-7) cells were maintained in the Minimum Essential Medium with Earle's salts (MEM) containing 10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA, USA; v/v). The COS-7 cells were used in most experiments as they remain attached well to plastic culture dishes while performing extensive washing procedures. Human Embryonic Kidney 293 (HEK293) cells, used for electrophysiology, were cultured in Opti-MEM I (Invitrogen) containing 5% FBS (v/v).

Quantitative assay of surface and total expression

Confluent COS-7 cells grown in 12-well plates were transfected with a total 1.8 μg of cDNAs (equal amounts of cDNAs were used in the case of co-transfection of two different DNA vectors) and 4 μL Lipofectamine 2000 (Invitrogen), as described (Horak and Wenthold 2009). After 5 h, media containing Lipofectamine/DNA mixture were replaced with fresh media containing NMDA receptor inhibitors - 20 mM MgCl2, 1 mM d,l-2-amino-5-phosphonopentanoic acid and 3 mM kynurenic acid. The quantitative assays were performed 38–40 h after the end of transfection. Briefly, COS-7 cells were fixed in 4% paraformaldehyde (PFA) in PBS (w/v) for 15 min and incubated for 1 h in PBS containing 0.2% bovine serum albumin (BSA; w/v) without (surface expression) or with (total expression) 0.1% Triton X-100 (TX-100; w/v). Then, cells were incubated in primary rabbit anti-green fluorescent protein (GFP) (Millipore, Billerica, MA, USA; 1 : 500 for surface expression and 1 : 1000 for total expression) and secondary antibodies (horseradish peroxidase-conjugated donkey anti-rabbit IgG; GE Healthcare, Chalfont, St. Giles, UK; 1 : 1000) diluted in PBS with 0.2% BSA. Finally, 400 μL of ortho-phenylenediamine (final concentration 0.4 mg/mL) dissolved in phosphate-citrate buffer containing sodium phosphate (Sigma, St. Louis, MO, USA) was added to each well for 30 min (surface expression) or 15 min (total expression). The color reaction was terminated with 100 μL of 3 M HCl and the optical density was determined at 492 nm using the Personal Densitometer SI (Molecular Dynamics, Sunnyvale, CA, USA). The signal measured from cells transfected with empty vector was subtracted from the signal obtained from cells transfected with NMDA receptor subunits. In each experiment, data obtained from three different wells for surface and three different wells for total expression measurements for each subunit combination were normalized to average data obtained from the cells expressing control NMDA receptor subunit(s). Three independent experiments were performed for each NMDA receptor subunit combination. Data are expressed as the mean ± SEM; statistical comparisons were made using a one-way anova test.

Electrophysiology

The HEK293 cells grown in a 24-well plate were transfected with 0.9 μg of cDNAs coding for GluN1 (0.35 μg), GluN2 (0.35 μg), and GFP (0.2 μg; pQBI 25, Takara, Otsu, Shiga, Japan) mixed with 0.9 μl Matra-A Reagent (IBA, Göttingen, Germany) in 50 μL of Opti-mem I, as described (Cais et al. 2008). After 20 min on a magnet, cells were trypsinized and resuspended in Opti-MEM I containing 1% FBS (v/v) supplemented with 20 mM MgCl2, 1 mM d,l-2-amino-5-phosphonopentanoic acid, and 3 mM kynurenic acid, and plated on 30 mm poly-l-lysine-coated glass coverslips. Whole-cell voltage-clamp recordings were performed with an amplifier Axopatch 200B (Axon Instruments, Union City, CA, USA) after a capacitance and series resistance (< 10 MΩ) compensation of 80%. Glutamate-induced responses were low-pass filtered at 2 kHz with an eight-pole Bessel filter, digitally sampled at 5 kHz and analyzed using pCLAMP software version 9 (Axon Instruments). Patch pipettes (3–5 MΩ) were pulled from borosilicate glass and filled with intracellular solution (in mM: 125 gluconic acid, 15 CsCl, 5 EGTA, 10 HEPES, 3 MgCl2, 0.5 CaCl2, and 2 ATP-Mg salt; pH adjusted to 7.2 with CsOH). The extracellular recording solution (in mM: 160 NaCl, 2.5 KCl, 10 HEPES, 10 glucose, 0.2 EDTA, and 0.7 CaCl2; pH-adjusted to 7.3 with NaOH) contained the NMDA receptor co-agonist glycine (10 μM). A multibarrel fast application system, with a ~10 ms time constant of solution exchange around the cell was described previously (Petrovic et al. 2009). Electrophysiological experiments were performed at 23–25°C.

Immunoprecipitation

The immunoprecipitation experiment was performed as described (Horak et al. 2008). Briefly, COS-7 cells grown on 10 cm plates were transfected using the calcium phosphate coprecipitation method (CalPhos Mammalian Transfection kit; Clontech, Mountain View, CA, USA). Two days later, cells were collected by centrifugation (1000 g for 10 min at 4°C), solubilized in 2 mL of solubilization buffer [1% sodium deoxycholate (DOC) in 50 mM Tris-HCl, pH 7.3; w/v] for 30 min at 37°C and the supernatant (400 μL) obtained after centrifugation (100 000 g for 30 min at 4°C) was incubated with 5 μg of mouse IgG or anti-MYC (9E10) antibody and a 50 μL aliquot of protein G agarose beads (Pierce, Rockford, IL, USA) at 4°C overnight. Then, the beads were washed with TBST buffer (0.1% TX-100 in TBS; w/v) and boiled in 2× sodium dodecyl sulfate loading buffer (25 μL). Proteins were loaded onto 7% polyacrylamide gels, transferred to polyvinylidene difluoride membranes, incubated with primary anti-GFP (1 : 1000; Millipore) and secondary horseradish peroxidase-conjugated donkey anti-rabbit IgG (GE Healthcare; 1 : 5000) antibodies, and detected with ECL using BioMax MR x-ray films (Eastman Kodak, Rochester, NY, USA). The intensities of the protein bands were quantified using ImageJ imaging software (National Institutes of Health, Bethesda, MD, USA).

Endoglycosidase H digestion

Endoglycosidase H digestion was performed similarly as described (Kenny et al. 2009). The cells expressing NMDA receptor subunits were collected by centrifugation and solubilized in Denaturating Buffer (New England Biolabs, Hitchin, UK) at 100°C for 10 min. For digestion, 10× G5 reaction buffer (7 μL) and 500 units of endoglycosidase H were added to final volume of 70 μL. Samples were incubated for 1 h at 37°C and analyzed by immunoblotting.

Microscopy

For co-localization studies, COS-7 cells were washed in PBS, fixed in 4% PFA in PBS (w/v) for 20 min, permeabilized by 0.25% TX-100 in PBS (w/v) for 5 min and labeled with primary rabbit anti-GFP (Millipore; 1 : 1000) and primary mouse anti-PDI (Abcam, Cambridge, UK, 1 : 200; ER marker) or mouse anti-58K Golgi protein (Abcam, 1 : 200; GA marker) and secondary goat anti-mouse Alexa Fluor® 647 and anti-rabbit Alexa Fluor® 488 (Invitrogen) antibodies. Cells were then mounted with ProLong Antifade reagent (Invitrogen). For internalization studies, COS-7 cells were washed in ice-cold PBS and incubated with rabbit anti-GFP antibody for 30 min on ice to label surface-expressed receptors. After removing the primary antibody, cells were washed in PBS and returned to conditioned media for 30 min at 37°C to allow receptor internalization. The cells were washed in ice-cold PBS and the surface population of receptors was saturated with unconjugated goat anti-rabbit antibody (Invitrogen) for 30 min on ice to minimize subsequent receptor turnover. After fixation with 4% PFA in PBS (w/v) and permeabilization with 0.25% TX-100 in PBS (w/v), cells were incubated with 10% normal goat serum and then with goat anti-rabbit Alexa Fluor® 647 antibody (Invitrogen) to label the internalized population of receptors. Cells were washed and mounted. Images were taken on a Leica SPE confocal fluorescence microscope with a 63× objective (optical sections: 0.38 μm) and their maximum projections were analyzed using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Results

The presence of an L830 residue within the GluN1 M4 domain is critical for surface delivery of NMDA receptors

Our previous data showed that the deletion of the GluN1 M4 domain or its replacement with the acetylcholine receptor (AChR) M3 domain abolishes the surface trafficking of NMDA receptors (Horak et al. 2008). The molecular mechanism underlying the role that the GluN1 M4 domain plays in the surface delivery of NMDA receptors remains unknown. We hypothesized that there is a specific structural determinant within the GluN1 M4 domain that is critical for forward trafficking of NMDA receptors.

To identify this structural determinant, we first generated two chimeric GluN1 subunits having replacements of the amino acid residues within the first (YFP-GluN1ΔCt-AChR/GluN1-M4) or the second (YFP-GluN1ΔCt-GluN1/AChR-M4) half of the M4 domain for the amino acid residues of the AChR M3 domain (Fig. 1a and b). We used the sequence of the AChR M3 domain for generating the chimeric GluN1 subunits because it has the same orientation in the membrane as the GluN1 M4 domain and does not contain any ER retention signal (Wang et al. 2002). The wild type and chimeric GluN1 subunits used in our experiments were truncated after their M4 domains to remove the C-terminal ER retention and export signals, as described previously (Horak et al. 2008). We assessed the surface targeting of NMDA receptors composed of GluN2BΔCt and chimeric GluN1ΔCt subunits, expressed in heterologous COS-7 cells, using a quantitative assay of surface and total expression. Our data showed that the substitution of the second half, but not the first half, of the GluN1 M4 domain for the AChR M3 domain inhibited surface targeting of NMDA receptors while the total expression of GluN1 subunits remained unchanged (Fig. 1c). Thus, the presence of the second half of the GluN1 M4 domain is critical for the surface delivery of the C-terminally truncated NMDA receptors.

Figure 1.

The replacement of specific amino acid residues within the second half of the GluN1 M4 domain decreases the surface targeting of NMDA receptors without C-termini. (a) Schematic drawing of membrane topology of GluN1 subunit is shown with numbers indicating the membrane domains (M1–M4). (b) The sequences of the wild type GluN1 M4 domain (GluN1 M4), the chimeric GluN1 M4 domain with replacement of the first 11 amino acid residues (N812-G822; AChR/GluN1 M4) and the last 11 residues (G823-I833; GluN1/AChR M4) with the AChR M3 domain are shown. The AChR sequence is colored gray. (c) Bar graph shows quantification of surface (black) and total (white) expression of indicated yellow fluorescent protein (YFP)-GluN1ΔCt/GluN2BΔCt receptors expressed in African Green Monkey kidney fibroblast cell line (COS-7) cells using a quantitative colorimetric assay. Each vertical bar represents the mean ± SEM; n = 9 in three experiments. *< 0.05 relative to control (YFP-GluN1ΔCt/GluN2BΔCt), anova. (d) The amino acid sequence of the wild type GluN1 M4 domain is shown. The underlined amino acid residues were individually replaced with alanine residues by PCR mutagenesis. (e) Bar graph represents surface (black) and total (white) expression of NMDA receptors determined using a quantitative colorimetric assay performed on COS-7 cells transfected with the indicated NMDA receptor subunits. Data show mean ± SEM; n = 9 in three experiments. *< 0.05 relative to control (YFP-GluN1ΔCt/GluN2BΔCt), anova.

Next, we aimed to identify the specific structural determinant within the second half of the GluN1 M4 domain that is responsible for the surface targeting of NMDA receptors. We generated 10 single alanine substitutions of the amino acid residues within the second half of the GluN1 M4 domain in the C-terminally truncated GluN1 subunit (Fig. 1d). Alanine scanning mutagenesis is routinely used for the identification of trafficking motifs within membrane domains, such as in the case of the identification of an ER retention motif within the first membrane domain of the AChR subunit (Wang et al. 2002). We co-expressed the GluN1ΔCt subunits carrying the single mutations with the GluN2BΔCt subunit in COS-7 cells and assessed the surface targeting of GluN1/GluN2B complexes by a quantitative assay of surface and total expression (Fig. 1e). These experiments showed that co-expressing the GluN2BΔCt subunit with three of the mutated GluN1ΔCt subunits (YFP-GluN1ΔCt-F829A, -L830A, -I831A) resulted in reduced surface expression of GluN1/GluN2B complexes; other mutations within the GluN1 M4 domain did not alter the surface delivery of NMDA receptors (Fig. 1e). The total expression of mutated GluN1ΔCt subunits was not significantly different from control GluN1ΔCt subunits (Fig. 1e and FigureS1b). These results show that there are at least three specific amino acid residues within the GluN1 M4 domain (F829, L830, I831) that are critical for the surface delivery of NMDA receptors with truncated C-termini.

The C-termini of NMDA receptors contain multiple ER retention and export signals that may interfere with the membrane domain ER retention signals in the trafficking of NMDA receptors (Stephenson et al. 2008; Petralia et al. 2009). To investigate whether the specific amino acid residues within the GluN1 M4 domain, identified in our previous experiments with truncated NMDA receptors, play similar roles in the full-length NMDA receptors, we generated three full-length GluN1 constructs having these alanine substitutions (YFP-GluN1-1a-F829A, -L830A, -I831A; Fig. 2a). We co-expressed these mutated GluN1-1a subunits with the full-length GluN2B subunit in heterologous COS-7 cells and then examined using a quantitative assay of their surface and total expression. These experiments showed that two NMDA receptor combinations (YFP-GluN1-1a-F829A/GluN2B, YFP-GluN1-1a-I831A/GluN2B) trafficked to the cell surface similarly as the wild type NMDA receptors. However, one subunit combination (YFP-GluN1-1a-L830A/GluN2B) exhibited reduced surface expression (Fig. 2b). These observations suggest that the C-terminal region of the GluN1-1a subunit may interfere with the M4 domain-mediated regulation of the surface expression of the NMDA receptors. It is also possible that the GluN1 subunits truncated after the M4 domain exhibit slightly different conformation of the M4 domain that leads to the reduction of the surface expression of the C-terminally deleted NMDA receptors containing the GluN1-F829A and GluN1-I831A mutations. Furthermore, the reduction in the surface targeting of YFP-GluN1-1a-L830A/GluN2B receptors (~40%) was similar to the reduction observed in the case of the YFP-GluN1ΔCt-L830A/GluN2BΔCt receptors (Fig. 1e). Thus, the presence of the C-terminal region of the GluN1-1a subunit is not likely to be critical for the reduction in surface delivery of NMDA receptors mediated by the L830A mutation. The total expression of the full-length non-mutated and mutated GluN1 subunits was not significantly different in these experiments (Fig. 2b and FigureS1c). Thus, our data show that the L830 residue is critical for the surface targeting of the full-length NMDA receptors.

Figure 2.

The presence of an L830 residue within the GluN1 M4 domain is critical for surface delivery of full-length GluN1/GluN2B receptors. (a) The amino acid sequence of the wild type GluN1 M4 domain is shown. The underlined amino acid residues were individually replaced with alanine residues in the full-length GluN1-1a subunit. (b) A quantitative colorimetric assay was used to determine surface (black) and total (white) expression of indicated full-length GluN1/GluN2B receptors expressed in heterologous African Green Monkey kidney fibroblast cell line (COS-7) cells. Data show mean ± SEM; n = 9 in three experiments. *< 0.05 relative to control [yellow fluorescent protein (YFP)-GluN1-1a/GluN2B], anova. (c) The replacement of the GluN1 L830 residue with alanine does not alter the GluN1/GluN2B interaction. COS-7 cells transfected with YFP-GluN1-1a/MYC-GluN2B and YFP-GluN1-1a-L830A/MYC-GluN2B receptors were solubilized with 1% deoxycholate (DOC), immunoprecipitated with mouse IgG or anti-MYC antibody, and probed with rabbit anti- green fluorescent protein (GFP) antibody. Densitometric analysis revealed no significant differences in the normalized amounts of the bound fractions between the studied combinations of the subunits. (d–e) The distribution of indicated NMDA receptors closely matches the distribution of an ER marker (d) but not a golgi apparatus (GA) marker (e). Images were taken on fixed COS-7 cells using a confocal microscope, as described in the Methods. Scale bar, 10 μm. (f–g) COS-7 cells were transfected with the indicated NMDA receptor subunits, cell homogenates were incubated in the presence or absence of endoglycosidase H, and samples were analyzed by immunoblotting using anti-GluN1 (f) or anti-MYC (g) antibodies.

Our results showing that the L830 residue is critical for the surface delivery of NMDA receptors might be explained by an altered interaction between the GluN1 and GluN2 subunits in the presence of the L830A mutation. In the next set of experiments, we co-expressed the wild type and mutated YFP-GluN1-1a subunits with the MYC-tagged GluN2B subunit in heterologous COS-7 cells and performed the co-immunoprecipitation experiments to study the interaction between the GluN1 and GluN2B subunits. We did not observe statistical difference in the level of interaction between the wild type and mutated GluN1-1a subunits with the GluN2B subunit (ratio between the YFP-GluN1-1a-L830A/MYC-GluN2B and YFP-GluN1-1a/MYC-GluN2B: 1.06 ± 0.04; n = 3; Fig. 2c). Therefore, we suggest that the GluN1-L830 residue is not critical for the interaction between the GluN1 and GluN2 subunits, but rather that it is likely involved in the ER processing and/or forward trafficking of the functional NMDA receptors. In this case, the majority of YFP-GluN1-1a-L830A/GluN2B receptors would be present in the ER. We performed immunofluorescence experiments on heterologous COS-7 cells expressing YFP-GluN1-1a/GluN2B and YFP-GluN1-1a-L830A/GluN2B receptors and stained them with antibodies recognizing an ER and a Golgi apparatus (GA). Indeed, our immunofluorescence data revealed that both receptor combinations profoundly co-localize with the ER, but not with the GA marker (Fig. 2d and e). Together, our observations suggest that the L830 residue is involved in the ER processing and/or forward trafficking of the NMDA receptors. To substantiate the co-localization studies, we performed the deglycosylation assays with endoglycosidase H on the wild type and mutated receptors (Fig. 2f and g). However, our data did not reveal any endoglycosidase H-resistant population of the NMDA receptors, suggesting that the early trafficking of NMDA receptors is regulated differently from, for example, the AMPA receptors (Penn et al. 2008).

Does the replacement of the L830 residue within the GluN1 M4 domain alter the formation of the functional NMDA receptors? In the next experiments, we performed whole-cell patch-clamp recording from HEK293 cells expressing YFP-GluN1-1a/GluN2B and YFP-GluN1-1a-L830A/GluN2B receptors. NMDA receptor-mediated currents were evoked by 5 s application of 1 mM glutamate and 10 μM glycine at a membrane potential of 60 mV. Both combinations of GluN1/GluN2 subunits were able to form functional receptors and were completely inhibited with 1 mM MgCl2 or 1 mM d,l-2-amino-5-phosphonopentanoic acid and 3 mM kynurenic acid (Fig. 3a, b and FigureS2). Next, we asked if the L830A mutation alters the functional properties of NMDA receptor responses. First, we calculated the degree of desensitization as the ratio of the steady state current measured at the end of the glutamate application (Iss) over the peak current (Ip). Similar to previous data, the wild type YFP-GluN1-1a/GluN2B receptors exhibited Iss/Ip ~0.75 (Fig. 3c) (Vicini et al. 1998). The responses of YFP-GluN1-1a-L830A/GluN2B receptors did not exhibit significantly different Iss/Ip ratios when compared with wild type receptors (Fig. 3c). Second, the deactivation time course of the responses was not significantly different between both combinations of the NMDA receptor subunits (Fig. 3d). Together, our data show that the L830 residue is not critical for the formation of functional receptors. Indeed, the reduced surface expression of the mutated NMDA receptors could be caused by their increased internalization rates. We performed the internalization assays with the YFP-GluN1-1a/GluN2B and YFP-GluN1-1a-L830A/GluN2B receptors expressed in COS-7 cells (Fig. 3e, f, and g). Our data showed that the mutated receptors exhibit slower internalization rate than the wild type receptors, indicating that the L830 residue regulates the forward trafficking of the NMDA receptors rather than their surface stability.

Figure 3.

The L830 residue does not alter the functional properties of NMDA receptors or increase their internalization. (a) Amino acid sequence of the GluN1 M4 domain with underlined amino acid residue that was individually substituted with alanine residue is shown. (b) Whole-cell patch-clamp recordings were performed on Human Embryonic Kidney 293 cells (HEK293) cells expressing the yellow fluorescent protein (YFP)-GluN1-1a/GluN2B and YFP-GluN1-1a-L830A/GluN2B receptors. Currents were elicited with a 5 s long application of 1 mM glutamate (indicated by filled bar); representative traces are shown. (c–d) Quantitative analysis of currents mediated by NMDA receptors revealed that the degree of desensitization (c), calculated as the ratio of the steady state current measured at the end of the glutamate application (Iss) over the peak current (Ip), and the weighted deactivation time constant (τw; d), calculated from a double exponential function fit, were not significantly different between YFP-GluN1-1a/GluN2B and YFP-GluN1-1a-L830A/GluN2B receptors; n = 6; p > 0.05, t-test. (e–g) Internalization of NMDA receptors in the African Green Monkey kidney fibroblast cell line (COS-7) cells. The cells were incubated live for 30 min at 37°C with anti-green fluorescent protein (GFP) antibody before fixation and incubation with secondary antibody conjugated with a fluorophore. (e) Representative cells are shown. Scale bar, 10 μm. (f–g) Data represent mean ± SEM (n = 10) of the YFP expression (f) and the average number of vesicular puncta per cell (g) for the wild type and mutated NMDA receptors. *< 0.05 relative to control (YFP-GluN1-1a/GluN2B), t-test.

The GluN2A-containing NMDA receptors exhibit different transport regulation as well as different functional properties from the GluN2B-containing receptors (Petralia et al. 2009; Qiu et al. 2009). In the next experiments, we studied whether the L830 residue within the GluN1 M4 domain is critical for the surface delivery of GluN1/GluN2A receptors. We were also interested to see if mutations of the surrounding residues, that decreased the surface expression of C-terminally deleted GluN1/GluN2B receptors (F829A and I831A; Fig. 1), alter the surface targeting of GluN1/GluN2A receptors. We co-expressed the YFP-GluN1-1a or YFP-GluN1-1a-F829A, -L830A, and -I831A subunits with the GluN2A subunit in COS-7 cells and assessed the surface targeting of GluN1/GluN2A complexes by a quantitative assay of surface and total expression (Fig. S3a). These experiments showed that co-expressing the GluN2A subunit with two of the mutated GluN1-1a subunits (YFP-GluN1-1a-F829A, -I831A) did not result in altered surface expression of GluN1/GluN2A receptors (FigureS3b). On the other hand, YFP-GluN1-1a-L830A/GluN2A receptors exhibited reduced surface expression, although total expression among the GluN1/GluN2A receptors remained unchanged. Interestingly, the degree of the reduction in surface expression of YFP-GluN1-1a-L830A/GluN2A receptors (~40%) is similar to that observed with YFP-GluN1-1a-L830A/GluN2B receptors (Fig. 2b). These results show that the L830 residue regulates the surface delivery of both GluN1/GluN2A and GluN1/GluN2B receptors, likely by a similar mechanism.

To ensure that the L830A mutation does not alter the formation of functional GluN1/GluN2A receptors, we performed the whole-cell patch-clamp recording from HEK293 cells expressing YFP- GluN1-1a/GluN2A and YFP-GluN1-1a-L830A/GluN2A receptors, by the method described above. Indeed, both combinations of GluN1/GluN2A subunits were able to form functional receptors (FigureS3c). Furthermore, we calculated the degree of desensitization as well as deactivation time course of the responses of both combinations of NMDA receptor subunits (FigureS4d and e). We did not observe any significant difference between wild type and mutated NMDA receptors for either of these values. Thus, our observations support the hypothesis that the L830 residue is not involved in the formation of functional receptors but rather in their forward trafficking.

The L830 residue is not critical for the surface expression of GluN1 subunits

The GluN1 subunits contain multiple export/retention signals that control their surface delivery even in the absence of the GluN2 subunit (Okabe et al. 1999; Standley et al. 2000; Horak and Wenthold 2009; Petralia et al. 2009). Next, we asked whether the substitutions of the amino acid residues within the GluN1 M4 domain, that reduced the surface delivery of C-terminally truncated NMDA receptors, alter the surface delivery of individually expressed GluN1 subunits (Fig. 4a). We performed a quantitative assay of surface and total expression on heterologous COS-7 cells expressing the YFP-GluN1-1a, -F829A, -L830A, and -I831A subunits (Fig. 4b). Our data showed that neither of studied mutated GluN1-1a subunits exhibited significantly different surface nor total expression when compared with the wild type GluN1-1a subunit. Thus, our results show that the L830 residue within the GluN1 M4 domain does not regulate the trafficking of individually expressed GluN1 subunits but is likely involved in the forward trafficking of functional NMDA receptors.

Figure 4.

The L830 residue does not regulate the surface expression of GluN1 subunits. (a) Amino acid sequence of the GluN1 M4 domain with underlined amino acid residues that were individually substituted with alanine residues is shown. (b) Bar graph represents surface (black) and total (white) expression of wild type and mutated NMDA receptor subunits determined using quantitative colorimetric assay on heterologous African Green Monkey kidney fibroblast cell line (COS-7) cells. Data show mean ± SEM; n = 9 in three experiments. *p < 0.05 relative to control [yellow fluorescent protein (YFP)-GluN1-1a], anova.

The GluN1-L830A mutation does not alter the reduction in the surface expression of NMDA receptors mediated by the M3 domain

Our recent data show that there are multiple trafficking determinants within the GluN1 M3 domain including W636 residue that regulate the trafficking of the functional NMDA receptors (Kaniakova et al. 2012). Next, we asked whether the structural determinants within the GluN1 M3 and M4 domains decrease the surface delivery of NMDA receptors by a shared mechanism. Indeed, in this case the GluN1 construct carrying both W636A and L830A mutations would traffic to the cell surface similarly to the GluN1 construct carrying the single W636A mutation (Fig. 5a and b). As our data showed no significant difference between the surface expression of YFP-GluN1-1a-W636A/GluN2B and YFP- GluN1-1a-W636A+L830A/GluN2B receptors (Fig. 5c), we suggest that there is a shared mechanism for the M3 and M4 domain-mediated regulation of surface expression of NMDA receptors.

Figure 5.

The GluN1-L830A mutation does not alter the reduction in the surface expression of NMDA receptors mediated by the GluN1-W636A mutation within the M3 domain. (a) Amino acid sequences of the GluN1 M3 and M4 domains, with underlined amino acid residues that were individually substituted with alanine residues, are shown. (b) A structural model of the membrane domains of the GluA2 receptor (Sobolevsky et al. 2009), with marked homologous residues to GluN1-W636 and GluN1-L830 residues. (c) A quantitative assay was used to determine surface (black) and total (white) expression of the indicated full-length GluN1/GluN2B receptors expressed in heterologous African Green Monkey kidney fibroblast cell line (COS-7) cells. Data show mean ± SEM; n = 9 in three experiments. *< 0.05 relative to control [yellow fluorescent protein (YFP)-GluN1-1a-L830A/GluN2B], anova.

Discussion

Assembly and forward trafficking of NMDA receptors to the surface membranes are likely regulated by multiple mechanisms. These mechanisms ensure that only correctly assembled receptors are released from the ER and transported to the cell surface.In this study, we investigated the mechanism by which the GluN1 M4 domain regulates the surface targeting of NMDA receptors. Using quantitative assays, biochemistry and electrophysiology on heterologous cells expressing recombinant NMDA receptors, we show that the second half of the GluN1 M4 domain is critical for the surface delivery of NMDA receptors. Furthermore, we identified within this region a single amino acid residue, L830, that is important for surface expression of NMDA receptors containing both GluN2A or GluN2B subunits. The L830 residue does not seem to be critical for the interaction between GluN1 and GluN2 subunits or for the formation of functional receptors. We conclude that the L830 residue regulates the forward trafficking of NMDA receptors.

Role of GluN1 M4 domain in the forward trafficking of NMDA receptors

We have shown previously that the GluN1 M4 domain is critical for the surface delivery of NMDA receptors (Horak et al. 2008). Our initial finding that the replacement of the second half of the GluN1 M4 domain with the appropriate part of the AChR M3 domain dramatically reduces the surface expression of the C-terminally truncated NMDA receptors indicated that this region contains a critical structural determinant that controls trafficking of the receptors. Our mutagenesis studies revealed that there are three surrounding amino acid residues within the GluN1 M4 domain, F829/L830/I831, that are important for normal surface delivery of NMDA receptors without C-termini. Out of these residues, the presence of only the L830 residue was critical for the surface targeting of full-length GluN1/GluN2B receptors, indicating that the C-terminus may, in some cases, interfere with the M4 domain-mediated regulation of their trafficking.

What role does the L830 residue play in the regulation of the surface expression of NMDA receptors? The membrane domains of NMDA receptors have been shown previously to be critical for the functioning of NMDA receptors. For example, specific mutations within membrane domains influence agonist potency and channel gating (GluN2A M3 domain) as well as sensitivity to ethanol (GluN2A M3 and M4 domains) (Ren et al. 2007; Salous et al. 2009). Interestingly, the presence of GluN1 and GluN2 M4 domains has been also shown to be critical for the formation of functional receptors from NMDA receptor subunits truncated before M4 domains (Schorge and Colquhoun 2003). This indicates that specific amino acid residues within M4 domains are critical for inter-membrane domain interaction(s). Recent data showed that it is likely that highly specific inter-membrane domain interactions of the M4 domain with the M1/M3 domains are required for surface expression of AMPA receptors (Salussolia et al. 2011). Moreover, some of the mutations within the GluA1 M4 domain, including the V805 residue in a homologous position to the L830 residue in the GluN1 subunit, exhibit greatly reduced current responses. Our electrophysiological data show that the L830A mutation of the GluN1 M4 domain does not alter the kinetic properties of GluN1/GluN2A and GluN1/GluN2B receptors. Thus, this mutation seems to alter only the trafficking of NMDA receptors but not their functioning.

However, it is still not clear what mechanism accounts for the reduction in the surface expression of NMDA receptors without L830 residues. We suggest that specific inter-membrane domain interactions of the GluN1 M4 domain, likely mediated or at least regulated by L830 residue, control the ER processing and/or forward trafficking of NMDA receptors. Our recent data showed that there are at least two specific structural determinants within GluN1 M3 domain including W636 residue that regulate the trafficking of NMDA receptors (Kaniakova et al. 2012). One possibility is that the L830 residue is involved in the physical interaction with the M3 domain(s). As double mutated GluN1 subunits within both M3 (W636A) and M4 (L830A) domains trafficked to the cell surface to a similar level as the GluN1 subunit with a single mutation in the M3 domain (W636A), we suggest that there is a common mechanism shared by M3 and M4 domains that regulate the surface delivery of NMDA receptors. Using fluorescence resonance energy transfer measurements, it has been proposed that transmembrane regions are required for the assembly of NMDA receptors (Cao et al. 2011). Indeed, if the L830 residue is involved in the assembly of NMDA receptors, we would expect to see an effect on the GluN1/GluN2 interaction or the formation of functional receptors containing the L830A mutation.

Physiological implications

The ER retention signals within the C1 cassette of some GluN1 splice variants are considered to be major trafficking determinants of NMDA receptors (Stephenson et al. 2008; Petralia et al. 2009). It is interesting that the absence or presence of the C-terminus of the GluN1-1a subunit does not alter the reduction in surface expression of NMDA receptors with the L830A mutation. This suggests that the C-termini and the GluN1 M4 domain regulate the trafficking of NMDA receptors by a different mechanism. Indeed, it is not clear at the moment whether the cell assesses the specific regions within the membrane domains and/or specific inter-membrane domain interactions. Likely, specific transmembrane proteins present within the ER interact with the membrane domains of the NMDA receptor subunits; the proper assembly of the receptor might disrupt this interaction and the GluN1 M4 domain might contribute to this process.

Why are there multiple regions within the functional NMDA receptors that regulate their trafficking? One possibility is that there are multiple levels of control during the assembly and trafficking of NMDA receptors that ensures that only properly folded receptors are transported to the cell surface. Indeed, various mechanisms might be employed in specific circumstances, for example, during activity-driven stimulation of neurons, so that adequate receptor numbers are localized to the appropriate cellular compartments. In conclusion, regulation of surface expression of NMDA receptors via their membrane domains involves a unique mechanism among the glutamate ionotropic receptors.

Acknowledgments

We thank Magda Kuntosova for technical assistance and Ronald S. Petralia for critical comments on the manuscript. This work was supported by the Grant Agency of the Czech Republic (P303/11/0075, P304/12/G069, P303/12/1464) and Marie Curie International Reintegration Grant (PIRG-GA-2010-276827) and Research Project of the AS CR AV0Z50110509 and RVO:67985823. The authors declare no conflict of interests.

Ancillary