INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The electrosensory system of the gymnotiform electric fish Apteronotus leptorhynchus is a particularly useful model for studies on the neuronal mechanisms that mediate sensory processing (Berman and Maler 1999; Carr and Maler 1986). Studies on electrosensory processing have focused primarily on the electrosensory lateral line nucleus (ELL), where the initial stages of sensory processing occur. Critical sensory processes such as adaptation and attention are controlled from higher centers through feedback inputs to neurons in the ELL. Glutamate is the predominant excitatory neurotransmitter in the feedback and afferent projections to the ELL, with both N-methyl-D-aspartate (NMDA) and non-NMDA components to the synaptic responses (Bastian 1993; Berman et al. 1997). In situ hybridization and immunohistochemical analyses have demonstrated prominent expression of NMDA receptors in both primary afferent and feedback synapses of the ELL (Berman et al. 2001; Bottai et al. 1997). Furthermore, the plasticity of synaptic responses involved in the adaptive filtering of predictable electorsensory signals (Bastian 1996a,b; Bell et al. 1997) is inhibited by pharmacological blockade of these NMDA receptors (Han et al. 2000; J. Bastian, personal communication).

A variety of studies in other teleosts have implicated NMDA receptor currents as important functional elements. In the goldfish, synapses between the auditory afferents and the Mauthner neuron display a prominent NMDA component with an unusually rapid deactivation time course (Wolszon et al. 1997). NMDA receptors have also been shown to be critical for the activity-dependent refinement of the optic nerve projections to the tectum in both goldfish and zebrafish (Schmidt 1990; Schmidt et al. 2000). Zebrafish also display postsynaptic currents with prominent NMDA components on reticulospinal and motor neurons during the development of locomoter circuits (Ali et al. 2000). In all of these examples, the molecular and functional characteristics of the receptors that underlie these currents are not known.

To understand the roles for NMDA receptors in the processing of electorsensory signals, we have initiated a molecular analysis of apteronitid NMDA receptors. In mammals, the molecular analysis of the NMDA receptors has provided detailed information on their structures and properties (Dingledine et al. 1999). The extent to which these molecular structures and functional properties are shared by the NMDA receptors in the lower vertebrates has not been fully established. Ionotopic glutamate receptors are multimeric proteins proposed to contain four subunits, in the case of the NMDA receptor complex two each of the NR1 and NR2 subunits (Laube et al. 1998; Rosenmund et al. 1998). In both fish and mammals, a single NR1 gene encodes several isoforms of the NR1 protein that result from alternative mRNA splicing (Bottai et al. 1998; Hollmann et al. 1993). Diversity for the mammalian NMDA receptors is further developed from the complex patterns of expression of the four different NR2 genes, NR2A-D. The NR2 genes are expressed in cell-specific patterns that change during development of the CNS (Monyer et al. 1994). Studies with recombinant mammalian NMDA receptors expressed in heterologous cells have shown that different NR2 subunits impart specific properties to the resulting NR1/NR2 channel complex. The kinetics of the receptor response depends on the type of NR2 in the complex, with NR2A being faster than NR2B or NR2C, and NR2D being the slowest (Monyer et al. 1992; Vicini et al. 1998). The voltage dependence of the receptor also depends on the NR2 subunit through differences in the strength of the Mg²⁺ block. Receptors containing the NR2A or NR2B subunits are more sensitive to extracellular Mg²⁺ than are those containing NR2C or NR2D (Kuner and Schoepfer 1996; Monyer et al. 1994).

In mammals, several lines of evidence support a unique role for the NR2B subunit in the synaptic plasticity of a variety of central neurons. Deletion of the murine NR2B gene results in neonatal death due to an impairment of the neural systems that control the suckling response (Kutsuwada et al. 1996). Studies on neonatal survivors have shown that these mice fail to exhibit normal synaptic plasticity in hippocampal recordings. On the other hand, overexpression of NR2B in murine forebrain neurons results in enhanced synaptic potentiation in the hippocampus and improved memory function (Tang et al. 1999). If these special functions of the NR2B subunit are conserved in the neurons of lower vertebrates such as apteronitid fish, then the NR2B containing receptors may also be critical elements in the synaptic responses in the these organisms.

In A. leptorhynchus, sequence analysis of the NMDA receptor subunit NR1 (AptNR1) indicated strong sequence conservation with its mammalian orthologs, including some of the alternative splice variants (Bottai et al. 1997, 1998). Here, we present the sequence of the apteronitid NR2B subunit (AptNR2B). Whole-cell patch-clamp recordings were used to compare the electrophysiological and pharmacological properties of the apteronotid and mammalian NMDA receptors after transfection of AptNR1/NR2B receptor cDNAs into human embryonic kidney (HEK) cells. The results indicate strong evolutionary conservation of both amino acid sequence and functional properties of this teleost NMDA receptor.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolation of the apteronotus NR2B cDNA

An Apteronotus brain cDNA library (Bottai et al. 1998) was screened at reduced stringency with a cDNA for the rat NR2B gene. NR2B cDNAs were identified by sequencing both strands using the OpenGene Automated DNA Sequencing System (Visible Genetics, Toronto, Ontario, Canada). The 5' and 3' termini were obtained using the rapid amplification of cDNA ends (RACE) method developed by Chenchik et al. (1996) with minor modifications. Long AptNR2B cDNA clones were identified and sequenced.

Construction of expression vectors

Complete AptNR2B and AptNR1 cDNAs were obtained by PCR using rTth XL kit (Perkin-Elmer, Boston, MA). The PCR reactions were performed with 8 U rTth polymerase in a 100-µL reaction volume containing 0.20 mM dNTPs, 1 mM Mg(OAc)₂, 0.25 µM primers (AptNR2B primers: GGATCCTGACTGGCAGTACA AAAGGTGTTT and GGATCCTGCCATGTTCTGCACAACTTACTC; AptNR1 primers: CTCTAGATCCCGGCTGGCTGCACCTCAC and CTCTAGACTCTGTGCGTTTTGGCTCTCA) in XL PCR BUFFER II (Perkin-Elmer). PCR products were subcloned in pGemT (Promega, Madison, WI) and sequenced completely. The pGemT-AptNR2B clone was digested with Xho I, gel purified and subcloned in the pcDNA-zeo(+) vector (Invitrogen, Carlsbad, CA). This plasmid will be referred as pcDNA-AptNR2BWT.

To replace the AptNR2B signal peptide, the mNR2B signal peptide cDNA was first amplified from pcDNA.2 (kindly provided by M. Mishina, Tokyo, Japan) using rTth polymerase and primers CTCTGCAGGCTCTTTTGGGAACG and AATACGACTCACTATAGGGAGACC. The PCR product was subcloned in pGemT and sequenced. The pcDNA-AptNR2BWT plasmid was digested by Xho I and recircularized, which isolated the 5' end of aptNR2B ORF cDNA in the pcDNA vector (pcDNA-5'2B). The mouse signal peptide PCR product was digested and inserted at the Hind III and Pst I sites in the plasmid pcDNA-5'2B. The remaining 3' fragment of AptNR2B ORF cDNA was finally inserted at the Xho I sites to complete the aptNR2B-ps sequence in the pcDNA vector.

Cell culture and transfection of HEK293T cells

HEK 293T cells were grown in DMEM (Invitrogen) in an incubator containing 5% CO₂ at 37°C, transferred in a 35-mm dish (BD Biosciences Discovery Labware, Palo Alto, CA) at 40-60% confluency and transfected the next day in media containing 10 µM ketamine or 1 mM D-2-amino-5-phosphonovalerate (APV, Sigma, St. Louis, MO). CaPO₄ transfections were performed with 0.25 µg pEGFP-C1 (BD Biosciences Clontech), 1.5 µg pcDNA-NR1, and 1.5 µg pcDNA.NR2B. Cells exhibiting green fluorescence were used for electrophysiological analysis on the next day.

Electrophysiological recordings and analysis

External recording solution contained (in mM) 145 NaCl, 5 KCl, 0.5 CaCl₂, and 10 HEPES, pH 7.6. Internal recording solution contained (in mM) 145 CsCl, 5 KCl, 5 ATP-Na, 5 BAPTA, and 10 HEPES, pH 7.4. Osmolarity was adjusted with sucrose. Unless mentioned, 50 µM glycine was in the external solution and 1 mM NMDA was applied by a fast perfusion system (Warner Instrument, Hamden, CT). Small lifted cells were held at 80mV. Electrodes were between 3 and 6 M. Electrophysiological recordings were performed on an AxoPatch-200B (Axon Instruments, Union City, CA), filtrated at 1 KHz and acquired at 3 KHz using Clampex 6.0 (Axon Instruments). Two to eight traces were averaged. Their baselines were adjusted and the peak amplitude was determined. Dose-response data were fitted with the following equations: for agonist, I_n = 1/(1 + (EC₅₀/[A_g])ⁿ); for antagonist, I_n =1  1/(1 + (IC₅₀/[A_n])ⁿ), where I_n is the normalized peak current, EC₅₀ is the half-maximum agonist dose, [A_g] is the agonist concentration applied to stimulate NMDA receptor, IC₅₀ is the antagonist dose causing 50% of the NMDA receptor block, [A_n] is the antagonist concentration, and n is the Hill coefficient. To estimate the voltage dependence of the Mg²⁺ block, the Mg²⁺ affinity at 0 mV (K_0.5(0)) was estimated following the Woodhull equation (Woodhull 1973): K_0.5(V) = K_0.5(0) × exp(zVF/RT), where K_0.5(V), is the external [Mg²⁺] causing half inhibition at a given membrane potential,  is electric field fraction sensed by Mg²⁺, and z, V, F, R, and T have their usual physical significance. To determine the relative Ca²⁺ permeability of NMDA receptor, the external solution 1 contained 150 mM NaCl, 10 mM HEPES, pH 7.6, and 50 µM glycine while, in the external solution 2, NaCl was replaced by 110 mM CaCl₂. Osmolarity adjustments and stimulations were made as previously. Relative Ca²⁺ permeability (P_Ca/P_Na) was determined by the constant field theory with an adapted equation from Lewis equations (Lewis 1979): P_Ca/P_Na = [Na]₁/[Ca]₂ × exp ((E_r2  E_r1) × F/RT) × (1 + exp(E_r2 × F/RT))/4, where P_Ca/P_Na is the relative Ca²⁺ permeability, [Na]₁ and E_r1 are the Na⁺ activity and the reversal potential in external solution 1, respectively, while [Ca]₂ and E_r2 are the Ca²⁺ activity and the reversal potential in external solution 2, respectively.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Evolutionary conservation of the teleost NR2B amino acid sequence

Glutamate-induced synaptic responses mediated by the NMDA class of glutamate receptors have been demonstrated in the nervous systems of a variety of vertebrate species, including teleost fish (Berman et al. 1997; Curti et al. 1999; Han et al. 2000; Smeraski et al. 1999; Spiro 1997; Wolszon et al. 1997). To this date, the molecular analysis of the nonmammalian NMDA receptors has been limited to the NR1 subunit, with sequences reported for NR1 proteins from Xenopus (Soloviev et al. 1996), duck (Kurosawa et al. 1994), and apteronotid fish (Bottai et al. 1998). These sequences have indicated strong evolutionary conservation of critical elements in the NR1 protein, however, the direct functional comparisons with recombinant mammalian NMDA receptors have not be carried out due to the lack of a lower vertebrate NR2 receptor cDNA. To carry out a functional comparison of the teleost NMDA receptor with its mammalian orthologs, we have isolated a full-length cDNA encoding the A. leptorhynchus NR2B subunit.

Partial NR2B cDNAs were isolated from a cDNA library prepared from A. leptorhynchus brain mRNA in a screen with the rat NR2B cDNA probe. The 5' and 3' ends were obtained using a modified version of the RACE method described by Chenchik et al. (1996). The full-length AptNR2B cDNA encodes a protein of 1,617 amino acids that shows strong sequence and functional similarities to mammalian NR2B proteins. Figure 1A illustrates the sequence alignment of the fish and mouse NR2B proteins. In this alignment the two sequences share an overall value of 62% amino acid identity, with central regions including the transmembrane segments and the ligand-binding segments S1 and S2 sharing the highest levels of sequence identity. The phylogenetic comparison of AptNR2B to the family of murine NMDA receptor subunits confirms that AptNR2B is most closely related to the murine NR2B subtype (Fig. 1B).

View larger version (64K): [in this window] [in a new window]   Fig. 1. Comparison of the apteronotid and murine NR2B amino acid sequences. A: amino acid sequences of AptNR2B and mNR2B aligned with the Clustal V method in the Megalign program (DNAStar, Madison, WI). Amino acids in the mNR2B sequence that match residues in AptNR2B are shown as dots. Gaps in the sequences are identified by dashes. Transmembrane segments (I, III, and IV) and the reentrant loop region (II) are identified by bars over the sequence. Putative signal peptides are identified with a bar (SP) over both sequences. Proposed binding site for CaMKII is enclosed by the box. Stars indicate the predicted sites for phosphorylation at Ser by CaMKII (Strack et al. 2000) and at Tyr by Fyn (Nakazawa et al. 2001). B: phylogenetic comparison of AptNR2B to murine NR1 and NR2 subunits. Analysis by the parsimony method was performed using the PROTPARS program in the Phylogeny Inference Package (PHYLIP) (Felsenstein 1989). C: schematic representation of the relative sequence conservation within functional domains of AptNR2B. Amino acid residue numbers are on the top for AptNR2B on the bottom for mNR2B. Percentage values calculate the amino acid identity for each domain. Overall amino acid identity is 62%.

The high levels of amino acid similarity throughout the transmembrane segments (M1-M4) and ligand binding segments (S1-S2) predict strong evolutionary conservation of critical receptor properties, including glutamate binding and channel permeability. However, there are two regions where the level of sequence conservation between the fish and mammalian proteins is not as pronounced. The lowest level of sequence conservation occurs at the extreme N-terminus of AptNR2B, a sequence that includes the predicted membrane insertion sequence (Fig. 1, AptNR2B amino acids 1-52). This signal peptide includes the required hydrophobic segment and potential signal peptide cleavage site (von Heijne 1983), but is preceded by an unusual N-terminal extension of 23 amino acids. The sequence of the N-terminal region of AptNR2B was confirmed by amplification of the 5' sequence from brain mRNA and DNA sequence analysis of the products. The functional significance of this unusual signal peptide sequence is not known, but may involve the regulation of NR2B protein expression in fish neurons.

The second segment of relatively low sequence similarity occurs in the region that forms the carboxyl terminal intracellular tail sequence, with an overall amino acid identity of 36% (Fig. lC). In this region there are short segments of strong identities interspersed between nonhomologous segments, consistent with the evolutionary conservation of functionally critical sequences. The C-terminal segments of the NMDA receptor subunits are known to contain regulatory sites for NMDA receptors, either as the target sites for protein kinases or through direct protein interactions with the PDZ domain containing synaptic scaffolding proteins. Many of these regulatory sites are conserved in the fish and murine sequences. Two of the three tyrosine residues that are phosphorylated by the tyrosine kinase Fyn in the murine sequence (Nakazawa et al. 2001; Takasu et al. 2002) are conserved and fall within segments of high sequence identity in the fish sequence (Fig. 1A). Another well-characterized kinase target site is the Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) target site at Ser1303 in the murine NR2B sequence (Omkumar et al. 1996). The sequence adjacent to Ser1303 forms a binding site for the CaMKII enzyme that targets the kinase to postsynaptic regions of the neuron (Bayer et al. 2001; Strack et al. 2000). This sequence (boxed in Fig. 1A), which corresponds to 1404-1415 of AptNR2B, including the putative target Ser1409, is well conserved in AptNR2B.

The PDZ binding motifs at the carboxyl termini of the mammalian NR2 subunits constitute the binding sites for interaction with the PSD-95 family of synaptic scaffold proteins (Kornau et al. 1995; Niethammer et al. 1996). The C-terminal 17 residues of AptNR2B are identical to those of the mNR2B protein, except for one Ser to Thr substitution. Thus the PDZ binding site is well conserved in AptNR2B.

The functional properties of recombinant A. leptorhynchus NMDA receptors

A useful approach for studies to characterize NMDA receptors containing different combinations of subunits is heterologous expression of cloned cDNAs in cultured animal cells. To this date, these analyses have been restricted to mammalian NMDA receptors, except for the Xenopus NR1 subunit, which was coexpressed with a novel truncated subunit Xu1 to yield currents with both NMDA and non-NMDA properties (Soloviev et al. 1996). To study the teleost NR1/NR2 channels, the AptNR1 (Bottai et al. 1998) and AptNR2B cDNAs were coexpressed in the HEK cell line for electrophysiological and pharmacological analysis.

The cDNAs AptNR1 and AptNR2B were inserted into the expression vector pcDNA3.1. The AptNR1 cDNA encoded the splice variant lacking the N1 cassette (N1) and the carboxyl-terminal cassettes C1, C1', and C1" (pcDNA-AptNR1). This NR1 splice variant is the most common form expressed in apteronotid brain (Bottai et al. 1998). These cDNAs were introduced into HEK cells and glutamate-induced currents measured using whole-cell patch-clamp recording. Initial experiments produced very small currents, 10-100 pA. The unusual structure of the apteronotid NR2B signal sequence suggested failure to process AptNR2B precursor peptide. To overcome this problem, the AptNR2B signal peptide (amino acids 1-79) was replaced by the signal peptide from the mouse NR2B gene (amino acids 1-31). In the resulting precursor peptide, the predicted signal peptidase cleavage site yields a mature AptNR2B protein with a deletion of 21 amino acids from the N-terminus, however, this deletion does not affect the critical structural elements of the N-terminal LIVBP-like (leucine, isoleucine, valine binding) domain. Cotransfection of the modified AptNR2B with AptNR1 produced robust currents in response to applications of glutamate or NMDA (Fig. 2). In the presence of 50 µM glycine, NMDA-induced currents reached amplitudes of 3 nA in some cells.

View larger version (24K): [in this window] [in a new window]   Fig. 2. NMDA-stimulated currents result from the expression of AptNR1/NR2B and mNR1/NR2B receptors in HEK 293T cells. Cells were held at 80 mV and NMDA-stimulated currents recorded in whole-cell patch-clamp mode. For each cell, the membrane potential was adjusted to 80, 40, 0, or +40 mV for 1 s before the addition of NMDA (open bar) and held at these potentials for a total of 5 s. A: currents recorded from cells transfected with AptNR1/NR2B. B: currents recorded from cells transfected with mNR1/NR2B. Glycine (50 µM) was added in the external solution for the NMDA stimulation pulses. Time and current rulers are on the right.

An important property of the NMDA receptor is the relatively slow deactivation time course of the channel after removal of glutamate, which results in long-lasting synaptic currents (Lester et al. 1990). For mammalian NMDA receptors, the speed of deactivation is dependent on the type of NR2 subunit in the complex, with NR2B-containing receptors exhibiting slower currents than NR2A (Monyer et al. 1992; Vicini et al. 1998). We measured the deactivation kinetics of both apteronotid and murine NR2B-containing NMDA receptors expressed in HEK cells (Fig. 3).

View larger version (18K): [in this window] [in a new window]   Fig. 3. Comparison of the deactivation rates of recombinant AptNR1/NR2B and murine NR1/NR2B receptors in HEK 293T cells. HEK cells transfected with NMDA receptor cDNAs were pulsed with NMDA (1 mM) for either 25 ms or 1 s, and current responses recorded in whole-cell patch-clamp mode. Deactivation time constants for each trace were measured by the best fit from an exponential curve. Mean value for AptNR1/NR2B deactivation constant is 346.0 ± 32.9 ms with a 25-ms pulse (n = 14) and 228.0 ± 20.7 ms with a 1-s pulse. For recombinant mNR1/NR2B, the mean value is 324.7 ± 20.5 ms with a 25-ms pulse and 205.0 ± 32.5 ms with a 1-s pulse (n = 10). There is no significant difference between the deactivation times for the fish and the mouse receptors (Student's t-test, P < 0.05).

Deactivation time courses were measured after NMDA pulses of either short (25 ms) or long (1 s) pulses of NMDA to the cells. The fish and murine receptors displayed nearly identical deactivation kinetics in each protocol. When the agonist application was short, the AptNR1/2B receptor deactivated with a time constant of 346.0 ± 32.9 ms, not significantly different from the mNR1/2B time constant of 324.7 ± 20.5 ms. For the 1-s agonist applications, the values were 228.0 ± 20.7 and 205.0 ± 32.5 ms, respectively, again not significantly different. These results demonstrate that the mechanism controlling the deactivation time course of NR2B-containing receptors is well conserved between these species.

Another important feature of the NMDA class of glutamate receptors is their voltage-dependent sensitivity to channel block by extracellular magnesium ions. This voltage sensitivity endows NMDA receptors with a conditional response such that receptor activation requires a preceding synaptic depolarization. We tested for the extracellular magnesium sensitivity in HEK293 cells with a 1-s pulse of NMDA (1 mM). For the fish and mouse receptors, current-voltage (I-V) curves were plotted expressing the normalized peak currents as a function of the membrane potentials, ranging from 80 to +50 mV for the fish (n = 10) and the mouse (n = 4) receptors (Fig. 4, A and B, respectively). The fish receptor is less sensitive to Mg²⁺ at the higher magnesium concentrations (100 µM and 1 mM Mg²⁺). To evaluate the voltage dependency of Mg²⁺ inhibition, normalized currents were plotted as a function of the Mg²⁺ concentration at voltages ranging from 80 to 30mV for both fish and mouse receptors (Fig. 4, C and D, respectively).

View larger version (29K): [in this window] [in a new window]   Fig. 4. Comparison of the magnesium sensitivity of recombinant Apteronotus and murine NR1/NR2B receptors. Cells were stimulated as described in Fig. 3 with the membrane holding potentials increased at 10 mV increments from 80 to +50 mV. A: normalized current voltage curves (n = 10) for AptNR1/NR2B receptor in the presence of increasing concentrations of magnesium; 0 external Mg2+ , 10 µM Mg2+ (), 100 µM Mg2+ (), and 1 mM Mg2+ (). B: as in A with the mNR1/NR2B receptor (n = 4). C: Mg2+ inhibition of peak currents for the AptNR1/NR2B receptor measured at increasing membrane potentials; 80 mV (+), 70 mV (), 60 mV (), 50 mV (), 40 mV (), and 30 mV (). D: as in C with mNR1/NR2B receptor. E: half-maximal Mg2+ inhibition values for AptNR2/NR2B () and mNR2/NR2B (). EC50 at 0 mV, which represents the Mg2+ affinity at 0 mV, was extrapolated to 5.9 ± 1.2 mM for AptNR1/NR2B and to 2.7 ± 1.0 mM for mNR1/NR2B.

The Mg²⁺ IC₅₀ values were then determined by fitting a Hill curve for each voltage step and they were plotted as a function of voltage. The Hill coefficients of the fitted curves were an average of 0.81 ± 0.05 for the fish receptor and an average of 0.94 ± 0.12 for the mouse receptor. Finally, the Mg²⁺ IC₅₀ values were plotted as a function of the voltage between 80 and 30 mV for both fish and mouse receptor. Figure 4E shows the average Mg²⁺ IC₅₀ values for both receptors. From this figure, the Mg²⁺ affinity at 0 mV (K_0.5(0)) was extrapolated. The K_0.5(0) values are 5.90 ± 1.18 mM for AptNMDAR-2B and 2.72 ± 1.01 mM for the mouse receptor. The  values, which represent the fraction of the electric field sensed by the Mg²⁺ ion, were essentially similar (0.83 ± 0.14 for the fish receptor and 0.78 ± 0.18 for the mouse receptor). These results indicate a reduced affinity for Mg²⁺ for the apteronotid NMDA receptor, which will reduce the voltage-dependent block of these receptors, leading to increased currents at hyperpolarized potentials.

Finally, we determined the relative Ca²⁺ permeability (P_Ca/P_M) in high extracellular calcium by measuring the reversal potential (E_r). In external 150 mM NaCl, AptNR1/2B has an E_r of 0.6 ± 0.7 mV (n = 4) while the mouse receptor has an E_r of 0.9 ± 0.4 mV (n = 3). In external 100 mM CaCl₂, their E_r values are 32.1 ± 3.7 mV (n = 4) and 32 ± 0.8 mV (n = 3), respectively. After calculating ionic activity, P_Ca/P_M values are 8.12 ± 1.85 and 7.40 ± 0.42 for the fish and mouse receptors, respectively. These values for relative calcium permeabilities are not significantly different under these conditions.

Pharmacology of the A. leptorhynchus NMDA receptor

Agonist and antagonist responses for the apteronotid NMDA receptor were determined in whole-cell patch-clamp mode after drug applications to HEK cells expressing AptNR1/NR2B receptors, and the results were compared with identical experiments with cells expressing murine NR1/NR2B receptors. Cells were held at various potentials (80, 40, 0, and +40 mV) for 1 s before the drug application (1 s). Peak current amplitudes were measured and plotted as a function of the drug concentration. The results are shown in Fig. 5 and summarized in Table 1.

View larger version (25K): [in this window] [in a new window]   Fig. 5. Pharmacology of recombinant AptNR1/NR2B receptors. HEK cells transfected with AptNR1/NR2B () or mNR1/NR2B () were stimulated as previously (Fig. 3). Peak currents were measured in whole-cell patch-clamp mode. Dose-response curves are shown for glutamate (A), APV (B), glycine (C), ifenprodil (D), NMDA (E), and proton (F). For the application of antagonists (B, D, and F), currents were stimulated with 1-s pulses of NMDA (1 mM). Glycine (50 µM) was included in the external solution for recordings A, B, D, E, and F, while NMDA (50 µM) was included in C.

Mammalian NMDA receptors are characterized by a relatively high affinity for the neurotransmitter glutamate in comparison to the non-NMDA glutamate receptors. Figure 5A illustrates the glutamate response curves for the peak responses of the fish and murine NMDA receptors expressed in HEK cells. The EC₅₀ values for glutamate are nearly identical for the receptors from both species (Table 1) and are similar to those reported for NR2B-containing receptors in human and rodents (Kutsuwada et al. 1992; Varney et al. 1996). In addition to glutamate, NMDA receptors depend on the coincident binding of the coactivator molecule glycine. The concentration dependence for glycine was determined in presence of 50 µM NMDA (Fig. 5C). The results indicate that the affinity for glycine did not differ significantly between species and was similar to previously published values for mammalian NMDA receptors (Table 1) (Kutsuwada et al. 1992; Traynelis and Cull-Candy 1990).

The characteristic response to the synthetic agonist NMDA defines this specific class of glutamate current in the physiological study of neurons. Figure 5E illustrates that the NMDA concentration-response curves are very similar for fish and mammalian NMDA receptors expressed in HEK cells. Again, the calculated EC₅₀ values were not significantly different for the fish and mouse receptors (Table 1). This analysis of the recombinant apteronotid receptor has established that the agonist affinities for the NMDA class of glutamate receptor are highly conserved between teleost and mammalian species.

The availability of a variety of specific NMDA receptor antagonists has been instrumental in studies that isolate the different classes of glutamate synaptic currents in neurons. APV is a widely used competitive inhibitor for the glutamate site on NR2 (Dingledine et al. 1999). The concentration dependencies for APV for both fish and murine NMDA receptors were tested by perfusion of cells with APV and then activation of the receptors with 1-s test pulses of 1 mM NMDA at 80, 40, 0, and +40 mV in the presence of 50 µM glycine (Fig. 5B). The IC₅₀ values for APV are similar for the fish and the mouse receptors (Table 1). Our results demonstrate that fish and mammalian NMDA receptors display comparable affinities for glutamate, NMDA, and APV, all of which bind to the glutamate binding site on the NR2B subunit. The properties of the agonist site have not diverged in the evolutionary time since the separation of the teleost and mammalian lineages.

The noncompetitive antagonist ifenprodil has been used as a specific antagonist for NMDA receptors that incorporate the NR2B subunit (Williams 1993). The concentration dependencies for inhibition of the fish and murine NR2B containing receptors are shown in Fig. 5D. The IC₅₀ values were significantly different between the receptors from these two species (Student's t-test, P > 0.01). The ifenprodil IC₅₀ value for AptNR1/2B (IC₅₀ = 1.72 ± 0.43 µM) is three times higher than the one for the mouse receptor (IC₅₀ = 0.59 ± 0.06 µM) (Fig. 5D and Table 1).

NMDA receptors with the NR1 subunit lacking the N1 cassette are highly sensitive to hydrogen ions at physiological pH (Traynelis et al. 1995). The pH dose-responses for AptNR1/2B and mNR1/2B were determined from pH 6.5 to 8.5 (Fig. 5F). The IC₅₀ values for proton inhibition were not significantly different (fish: IC₅₀ pH 7.3 ± 0.11; mouse: IC₅₀ pH 7.5 ± 0.09).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The NMDA class of glutamate receptors make important contributions to the mechanisms of synaptic plasticity and neural development in all vertebrates. The main findings of this study establish that teleost NMDA receptors containing the NR2B subunit display highly similar functional properties to those of the mammalian NR2B receptors. NMDA receptors are unique among the glutamate receptor family in their requirement for simultaneous activation by both the neurotransmitter glutamate and the cofactor glycine. In the presence of glycine, expression of the apteronotid NR1/NR2B receptor cDNAs in cultured HEK cells yielded robust glutamate-activated currents of similar magnitude to those of the murine NR1/NR2B. The affinities for glutamate and glycine were nearly identical for receptors from both species, indicating that the concentration dependence of the response to neurotransmitter release has been highly conserved across evolutionary distances. The results also show that the time course of the receptor response is highly conserved. Studies in mammals have shown that the kinetics of receptor responses are dependent on the type of NR2 subunit in the complex, with NR2B/NR1 receptors deactivating about fivefold more slowly than NR2A/NR1 receptors, but much faster than the NR2D/NR1 complex (Monyer et al. 1992; Vicini et al. 1998). As a result, the expression of different NR2 subunits can determine the duration of neuronal postsynaptic currents and therefore impact on the ability of some synapses to integrate multiple synaptic signals. We find that the kinetics of the apteronotid NR1/NR2B receptor closely matched that of the murine NR1/NR2B, with essentially identical time courses for the channel deactivation. Although the deactivation time courses of the NR2A, C, and D subunits from Apteronotus have not been determined, the evolutionary conservation of the NR2B kinetics supports the idea that the relatively slow deactivation time course of the NR2B containing NMDA receptors is a critical functional property (Brockie et al. 2001; Lester et al. 1990; Silver et al. 1992; Tovar et al. 2000).

Measurements of the concentration dependence for Mg²⁺ inhibition indicated that the teleost NMDA receptor has a lower apparent affinity for Mg²⁺ and therefore a less complete Mg²⁺ block compared with the mammalian receptor. The concentration of Mg²⁺ in the plasma of freshwater fish is estimated to be between 0.8 and 1.7 mM (McDonald and Milligan 1992), only slightly lower than mammals. Thus the reduced Mg²⁺ affinity may contribute to the previous observation that NMDA receptors at feedback synapses in the electrosensory lateral line lobe are active at hyperpolarized potentials approaching 70 mV (Berman et al. 1997, 2001).

Ifenprodil is a noncompetitive NMDA receptor antagonist that shows high selectivity for the NR2B subunit (Williams 1993). Dose-response curves showed that ifenprodil has approximately threefold lower affinity for AptNR2B compared with the murine NR2B receptor. The arginine residue at position 337 in the murine sequence, which is required for high affinity interaction with ifenprodil (Gallagher et al. 1996), is conserved in AptNR2B (R385), thus other residues in the ligand binding region must be responsible for the reduction in ifenprodil affinity. A consequence of this result is that higher concentrations of ifenprodil will be required to inhibit NR2B receptors in teleost neurons. Further studies with the apteronotid NR2A, C, and D subunits will be required to establish the relative selectivity of ifenprodil for the NR2B subunit in this system.

Evolutionary conservation of regulatory sites on the C-terminal domain of NR2B

The large carboxyl terminal intracellular segment of the NR2B subunit is involved in the regulation and subcellular targeting of NMDA receptors. Evidence for the role of this C-terminal domain has come from directed mutations of the murine NR2B gene that removed the C-terminal domain. Although these mutations do not block the formation of functional receptors in vitro, mice bearing a targeted deletion of the C-terminal domain of NR2B suffer neonatal death and reduced NMDA currents in hippocampal synapses (Mori et al. 1998; Sprengel et al. 1998). This phenotype closely resembles the phenotype observed in mice with a complete NR2B gene ablation, which indicates that the carboxyl terminal segment is essential for NR2B activity. A comparison of C-terminal domains of the fish and mammalian NR2B subunits shows short segments of highly conserved amino acid sequence interspersed with segments of low sequence similarity (Fig. 1A). These short conserved segments are good candidates for functionally critical elements within the C-terminal domain.

A variety of regulatory motifs have been identified in the C-terminal domain of the rodent NR2B receptor. Prominent among these are the recognition sites for protein kinases. The nonreceptor tyrosine kinases Src and Fyn phosphorylate the NR2B subunit and potentiate the glutamate-activated currents of NMDA receptors (Chen and Leonard 1996; Kojima et al. 1998; Levine and Kolb 2000; Lin et al. 1998; Manabe et al. 2000; Rostas et al. 1996; Takasu et al. 2002; Wang and Salter 1994; Yu et al. 1997). Three tyrosine residues in the rodent NR2B sequence, Tyr1252, Tyr1336, and Tyr1472, have been identified as potential target sites for these kinases, with Tyr1472 proposed as the key regulatory site (Cheung and Gurd 2001; Nakazawa et al. 2001). Two of these tyrosines (Tyr1252 and Tyr1472) are located within segments of strong sequence conservation in the fish sequence (Tyr1362 and Tyr1608, respectively), suggesting that the regulation of NR2B receptors through the actions of the Src family tyrosine kinases is likely an important feature of teleost neurons.

In the rodent nervous system, the regulation of synaptic responses can involve Ca²⁺ currents through the NMDA receptor that activates the CaMKII (Lisman et al. 2002). This process is facilitated by a direct interaction of CamKII to the C-terminal domain of the NR2B subunit (Leonard et al. 1999; Omkumar et al. 1996; Strack and Colbran 1998). A short segment of the C-terminal domain of NR2B, amino acids 1290-1309 in the murine NR2B sequence, forms the principal site for binding of CaMKII to the receptor (Bayer et al. 2001; Strack et al. 2000). In addition to its role in binding CamKII, this segment also contains residue Ser1303, which, when phosphorylated by the CaMKII enzyme itself, inhibits the kinase-receptor interaction, thus providing a negative feedback for the kinase-receptor interaction (Bayer et al. 2001; Strack et al. 2000). The sequence of AptNR2B reveals strong conservation of the CamKII binding site residues 1298-1309 (AptNR2B residues 1404-1415), including key residues Leu1298, Arg1300, and Ser1303 identified by point mutagenesis (Strack et al. 2000). Thus the AptNR2B maintains the key elements for both CamKII binding and phosphorylation that have been identified in the rodent NR2B gene. CaMKII is abundant in many neurons of the apteronotid brain (Maler and Hincke 1999), which are also enriched in NMDA receptors, further supporting this interaction.

Especially prominent coexpression of NMDA receptors (NR1 subunit) and CaMKII is evident in ELL pyramidal cells and in neurons of the dorsal forebrain (Bottai et al. 1997; Maler and Hincke 1999); preliminary studies (in situ hybridization) have demonstrated that the AptNR2B subunit is also abundant at both sites (R. Dunn, unpublished observations.). It will therefore be interesting to determine whether the activity-dependent association of AptNR2B and CaMKII is part of the molecular basis of adaptive sensory filtering in ELL (Bastian 1999). Recent studies have demonstrated that the dorsolateral forebrain of teleost fish is essential for spatial memory (Rodriguez et al. 2002) and have suggested that this region is similar to mammalian hippocampus. In Apteronotus, this brain region has very high levels of NR1 (Bottai et al. 1997), AptNR2B (R. Dunn and L. Maler, unpublished observations), and CaMKII (Bottai et al. 1997). Both spatial memory and synaptic plasticity (long-term potentiation) in the hippocampus require intact interactions between NR2B and CaMKII (Malenka and Nicoll 1999); given the relative simplicity of teleost forebrain, our results suggest that this may prove to be a useful preparation for investigating the cellular basis of spatial memory.

NMDA receptors located at neuronal synapses are embedded in a dense, protein-rich structure referred to as the postsynpatic density (PSD). In the PSD, the C-termini of the NMDA receptor NR2 subunits interact with the PDZ (PSD-95/Dlg/Z0-1) domains of the PSD-95 family of synaptic scaffold proteins, interactions that are mediated by the sequence ESDV at the C-termini of the NR2 subunits, which forms the binding site for the PSD-95 proteins (Kornau et al. 1995; Niethammer et al. 1996). These interactions are thought to stabilize the NMDA receptors at the synapse (El-Husseini et al. 2000; Roche et al. 2001) and to link the receptor to signaling molecules such as NO synthase (Christopherson et al. 1999; Sattler et al. 1999). The C-terminal 17 residues of AptNR2B are almost identical to those of the mammalian NR2B, with only single conservative Ser-Thr substitution (Fig. 1). This sequence includes both the PDZ binding motif sequence ESDV and the adjacent receptor internalization motif YEKL (Roche et al. 2001), indicating that the mechanisms regulating synaptic interactions of the NR2B receptors are conserved between teleost and mammalian species. This idea is further supported by previous work showing high sequence similarities between the fish and mammalian PSD-95 family of proteins (Lee et al. 2000).

The activity of the NMDA receptor is critical for mechanisms that regulate synaptic function in the CNS. The data presented here provide evidence that the functional properties of this key receptor are highly conserved in a teleost organism. In future studies, the NR2B subunit cDNA will provide a valuable resource for studies on the specific roles for the different NR2 subunits in the teleost nervous system.