Report

Asymmetrical Allocation of NMDA Receptor ε2 Subunits in Hippocampal Circuitry

See allHide authors and affiliations

Science  09 May 2003:
Vol. 300, Issue 5621, pp. 990-994
DOI: 10.1126/science.1082609

Abstract

Despite its implications for higher order functions of the brain, little is currently known about the molecular basis of left-right asymmetry of the brain. Here we report that synaptic distribution of N-methyl-D-aspartate (NMDA) receptor GluRϵ2 (NR2B) subunits in the adult mouse hippocampus is asymmetrical between the left and right and between the apical and basal dendrites of single neurons. These asymmetrical allocations of ϵ2 subunits differentiate the properties of NMDA receptors and synaptic plasticity between the left and right hippocampus. These results provide a molecular basis for the structural and functional asymmetry of the mature brain.

The NMDA receptor plays important and diverse roles in a number of brain functions (1). It is composed of hetero-oligomers of GluRζ (NR1), GluRϵ (NR2), and, occasionally, GluRχ (NR3) subunits (2, 3). The ϵ subunit family contains four distinct subtypes, ϵ1 to ϵ4 (NR2A to -2D). Because the four ϵ subunits differ in distribution and development in the brain, the subunit compositions of the NMDA receptors also differ depending on the brain regions and developmental stages (47). The actual subunit composition of the NMDA receptors in vivo is not known at present. NMDA receptors with distinct subunit combinations differ in physiological and pharmacological properties (5, 811). Recent studies suggested the possibility that NMDA receptors with distinct properties are distributed to the synapses in an input-selective manner (12, 13).

The adult mouse hippocampus expresses three distinct NMDA receptor subunits (ζ1, ϵ1, and ϵ2) (4, 6, 7). Hippocampal CA1 pyramidal neurons receive major excitatory inputs from Schaffer collateral (sch) fibers originating from ipsilateral CA3 pyramidal neurons and commissural (com) fibers, from contralateral CA3 pyramidal neurons (14). To characterize the NMDA receptors selectively in the sch-CA1 pyramidal neuron (sch-CA1) synapses, we have developed a method by which com fibers are denervated in live mice. Electrophysiological and immunoblot analyses using hippocampal slices from the com fiber–denervated mice revealed unexpected asymmetries in NMDA receptor–mediated responses and synaptic allocation of ϵ2 subunits in the hippocampus.

The bilateral hippocampi are connected by the ventral hippocampal commissure (VHC), and com fibers innervate contralateral hippocampal neurons via the VHC (14). We transected the VHC, thereby denervating com fibers (15). Selective denervation of com fibers was confirmed by the lack of retrograde transportation of the fluorescent dye, Fast blue (Fig. 1A), and by measuring the monosynaptic field excitatory postsynaptic potentials (fEPSPs) evoked on the basal dendrites of CA3 pyramidal neurons (Fig. 1B). In hippocampal slices from VHC transected (VHCT) mice 5 days after surgery, electrical stimulation at the stratum oriens of area CA3 evoked fEPSPs and presynaptic fiber volleys (PFVs), indicating that associational (asc) fiber-synapses on the basal dendrites of CA3 pyramidal neurons are functional in VHCT slices (VHCT and Ori, Fig. 1B). Electrical stimulation in the ventral fimbria at the same or much higher intensity elicited no measurable fEPSP or PFV, indicating that com fibers were essentially nonfunctional in the VHCT mice (VHCT and Fim, Fig. 1B). In the experiments that follow, we used hippocampal slices prepared from VHCT mice 5 days after surgery.

Fig. 1.

Selective denervation of commissural fibers by VHC transection. (A) Fluorescence photomicrographs of hippocampal slices prepared from a naïve mouse (Naïve) and from a VHCT mouse (VHCT) injected with the fluorescent dye, Fast blue, into the left hippocampus (arrowheads). Retrograde labeling of the right CA3 pyramidal neurons was absent in the VHCT mouse. CA1, field CA1; CA3, field CA3; DG, dentate gyrus. Bar, 200 μm. (B) Schematic diagram shows synaptic inputs on the basal dendrites of a CA3 pyramidal neuron and the electrode arrangement. A stimulating electrode was placed in the stratum oriens of area CA3 [Stim.(Ori)] or in the ventral fimbria [Stim.(Fim)], and an extracellular electrode [Rec.(field)] was placed in the stratum oriens of area CA3. Asc, associational fibers; Com, commissural fibers. Sample superimposed traces show fEPSPs and PFVs (arrows) in response to electrical stimulation at the stratum oriens (Ori) and at ventral fimbria (Fim), respectively. Each trace is the average of three consecutive responses recorded in hippocampal slices from a naïve mouse and from a VHCT mouse 5 days after the surgery. Bars are 1.0 mV (vertical) and 10 ms (horizontal).

We characterized NMDA EPSCs at the CA1 pyramidal neuron synapses using Ro 25-6981, an ϵ2 subunit–selective antagonist (1618). To record NMDA EPSCs, whole-cell recordings were made from CA1 pyramidal neurons in the presence of 6,7-dinitroquinoxaline-2,3-dione (DNQX, 20 μM) and bicuculline (30 μM) at a holding potential of +10 mV (Fig. 2). Because excitatory synapses on CA1 pyramidal neurons are located on both apical and basal dendrites, NMDA EPSCs were independently elicited by electrical stimuli applied either at the stratum radiatum or at the stratum oriens of area CA1.

Fig. 2.

Inhibitory effects of Ro 25-6981 on NMDA EPSCs in the left and right CA1 pyramidal neuron synapses. (A) Schematic diagrams show synaptic inputs on the apical dendrites of CA1 pyramidal neurons and arrangement of electrodes. In slices from naïve mice and from VHCT mice, electrical stimulation was applied at the stratum radiatum [Stim.(SR)] of area CA1. Sch, schaffer collateral fibers; Com, commissural fibers. Whole-cell recordings [Rec.(WC)] were made from CA1 pyramidal neurons. Sample superimposed traces show NMDA EPSCs recorded in the absence (Control) and presence of Ro 25-6981 (Ro, 0.6 μM). The levels of inhibition were maximal after exposure to Ro 25-6981 for 40 to 50 min (fig. S1). Left and Right indicate recordings in the left and right hippocampal slices, respectively. Each trace is the average of five consecutive recordings. Bars are 25 pA (vertical) and 100 ms (horizontal). Relative amplitudes of NMDA EPSCs in the presence of Ro 25-6981 were expressed as percentages of control responses. Columns and error bars represent means and SEM, respectively (n = 5 each, **P < 0.01, absence of an asterisk indicates P > 0.05). (B) Effects of Ro 25-6981 on NMDA EPSCs evoked on the basal dendrites of CA1 pyramidal neurons. Electrical stimulation was applied at the stratum oriens [Stim.(SO)] of area CA1. The others are the same as those described in (A) (n = 5 each, **P < 0.01, absence of an asterisk indicates P > 0.05). Bars are 25 pA and 100 ms. (C) Differential Ro 25-6981 sensitivities between apical (Api) and basal (Bas) dendrite synapses examined in the same CA1 pyramidal neuron. Using VHCT slices, electrical stimuli were applied alternately to the stratum oriens [Stim.(SO)] and to the stratum radiatum [Stim.(SR)] every 5 s. The others are the same as in (A) (for both left and right, n = 5 from 5 animals, **P < 0.01, ***P < 0.001). Bars are 25 pA and 100 ms. (D) Ratios of NMDA EPSCs to non-NMDA EPSCs evoked on the apical and basal dendrites of CA1 pyramidal neurons in VHCT slices. Upward and downward responses in each superimposed trace are NMDA EPSC and non-NMDA EPSC, respectively. Each trace is the average of five consecutive recordings. Bars are 50 pA, 100 ms. The others are the same as in (A) and (B) (n = 5 each, absence of an asterisk indicates P > 0.05).

Figure 2A illustrates the results obtained with the stimulation at the stratum radiatum. In naïve mice, Ro 25-6981 (0.6 μM) reduced peak amplitudes of NMDA EPSCs to a similar extent in the left and right hippocampal slices (left, 64 ± 7% of control, n = 5 slices from 5 animals; right, 63 ± 4% of control, n = 5 from 5 animals; P > 0.90, t test) (Naïve, Fig. 2A). By contrast, in VHCT mice, Ro 25-6981 reduced NMDA EPSCs in the left hippocampal slices more intensely than it did in the right (left, 39 ± 3% of control, n = 5 from 4 animals; right, 81 ± 8% of control, n = 5 from 5 animals; P < 0.01) (VHCT, Fig. 2A). Perforant path (pp) fibers from entorhinal cortex form synapses on CA1 pyramidal neurons in the stratum lacunosum moleculare (14) and pp-CA1 synaptic responses are suppressed by the activation of presynaptic group II metabotropic glutamate receptors (mGluRs) expressed in these fibers (1922). Application of the group II selective mGluR agonist L-CCG-1 (20 μM) did not depress EPSCs evoked by stimulation at the stratum radiatum of area CA1 (102 ± 7% of control, n = 5 from 5 animals), verifying that the currents were not significantly contaminated by pp inputs.

An opposite asymmetrical effect was observed in response to stimulation at the stratum oriens (Fig. 2B). In naïve mice, Ro 25-6981 diminished NMDA EPSCs to the same extent in the left and right slices (left, 63 ± 4% of control, n = 5 from 5 animals; right, 67 ± 3% of control, n = 5 from 4 animals, P > 0.40) (Fig. 2B), whereas in the VHCT mice its effect was a mirror-image asymmetry of that found with stimulation at the stratum radiatum (left, 81 ± 7% of control n = 5 from 4 animals; right, 42 ± 5% of control, n = 5 from 5 animals, P < 0.01) (Fig. 2B).

The asymmetrical effect of Ro 25-6981 at the apical and basal dendrite synapses in VHCT slices was confirmed in the same CA1 pyramidal neurons (left apical, 43 ± 4% of control, left basal, 76 ± 5% of control, n = 5 from 5 animals, P < 0.01; right apical, 72 ± 3% of control, right basal, 48 ± 2% of control, n = 5 from 5 animals, P < 0.001) (Fig. 2C). The amplitude ratios of NMDA EPSCs to DNQX-sensitive non-NMDA EPSCs, evoked at the same stimulation intensity, were indistinguishable between these synapses in VHCT mice (left apical, 62 ± 7%, n = 5 from 4 animals, left basal, 57 ± 7%, n = 5 from 4 animals; right apical, 61 ± 4%, n = 5 from 4 animals, right basal, 64 ± 6%, n = 5 from 5 animals) (Fig. 2D). Therefore, the asymmetry observed in VHCT slices appears to be specific to the ϵ2 subunit–mediated component rather than to the whole NMDA current.

In VHCT slices, NMDA EPSCs evoked by electrical stimuli at the stratum radiatum and at the stratum oriens are attributable to the sch-CA1 synapses. Thus, our results indicate that Ro 25-6981 sensitivities of NMDA EPSCs in the sch-CA1 synapses are asymmetrical between the left and right hippocampus and between the apical and basal dendrites of single neurons. Subsequently, Ro 25-6981 sensitivities of NMDA EPSCs in the com-CA1 synapses are expected to have opposite asymmetries to those in the sch-CA1 synapses, because Ro 25-6981 sensitivity in naïve mice that have both sch- and com-CA1 synapses was apparently identical between the left and right hippocampus (Fig. 2, A and B). Because it is not possible to stimulate com fibers selectively in area CA1, we examined Ro 25-6981 sensitivities of NMDA EPSCs in the com fiber synapses formed on the basal dendrites of CA3 pyramidal neurons by stimulating the ventral fimbria in hippocampal slices from naïve mice. To reduce contamination of com fiber responses on the apical dendrites of CA3 pyramidal neurons and to avoid antidromic activation of ipsilateral CA3 axons by stimulation at the ventral fimbria, we optimized the cutting angles for preparation of hippocampal slices (15). In hippocampal slices prepared with our procedure, we confirmed that electrical stimulation at the ventral fimbria rarely induced fEPSPs and antidromic population spikes in the stratum radiatum and stratum pyramidale of area CA3, respectively. When multiple peaks in EPSCs were generated by stimulation at the ventral fimbria, we rejected the slices.

With the ventral fimbria stimulation, Ro 25-6981 reduced NMDA EPSCs in the left hippocampal slices more intensely than it did in the right (left, 43 ± 6% of control, n = 6 from 6 animals; right, 85 ± 7% of control, n = 6 from 6 animals, P < 0.01) (Com, Fig. 3, A and B). Half of these experiments (n = 3 each) were performed in a “blind” fashion, and data were pooled because results were not significantly different (left, P > 0.75; right, P > 0.85) between the blind and non-blind experiments. The amplitudes of NMDA EPSCs, estimated by the NMDA/non-NMDA EPSC ratios, were comparable between the left and right hippocampus (left, 44 ± 4%, n = 9 from 9 animals; right, 48 ± 5%, n = 9 from 8 animals, P > 0.47) (Fig. 3C). By contrast, when we applied electrical stimuli at the stratum oriens of area CA3 and activated both com and asc fibers, Ro 25-6981 sensitivity of NMDA EPSCs was apparently identical between the left and right hippocampus (left, 67 ± 10%, n = 5 from 5 animals; right, 60 ± 3%, n = 5 from 5 animals, P > 0.5) (Asc/Com, Fig. 3, A and B). This result further supported the assumption that we selectively activated com fibers with the ventral fimbria stimulation without significant contamination of asc inputs. Thus, the left-right asymmetry in the effect of Ro 25-6981 was also present at the synapses formed by com fibers in an opposite manner to that found at the synapses formed by sch fibers (Fig. 2B).

Fig. 3.

Left-right asymmetry of com-CA3 synapses. (A) Effects of Ro 25-6981 on NMDA EPSCs in the left and right com-CA3 synapses. Schematic diagrams show synaptic inputs on the basal dendrites of CA3 pyramidal neurons and arrangement of electrodes. In slices from naïve mice, whole-cell recordings were made from CA3 pyramidal neurons. A stimulating electrode was placed in the ventral fimbria or in the stratum oriens of area CA3 to activate com fibers or to activate both associational and commissural fibers (Asc/Com), respectively. Superimposed traces show NMDA EPSCs recorded in the absence and presence of Ro 25-6981 (Ro, 0.6 μM). Bars are 25 pA, 100 ms. Each trace is the average of five consecutive recordings. (B) Relative amplitudes of NMDA EPSCs in the presence of Ro 25-6981 are expressed as percentages of the control responses. Columns and error bars represent means and SEM, respectively (Com, n = 6 each, **P < 0.01; Asc/Com, n = 5 each, absence of an asterisk indicates P > 0.05). (C) Ratios of NMDA EPSCs to non-NMDA EPSCs in the left and right com-CA3 synapses. Columns and error bars represent means and SEM, respectively (n = 9 each, absence of an asterisk indicates P > 0.05). (D) Effects of Ro 25-6981 on LTPs in the left and right com-CA3 synapses. Schematic diagrams show arrangement of electrodes for extracellular recording. fEPSPs were recorded by an extracellular electrode placed in the stratum oriens of area CA3 and electrical stimuli were applied at ventral fimbria. A tetanic stimulation (arrows) was given in the presence (closed diamonds) and absence (open circles) of Ro 25-6981 (0.2 μM). Ro 25-6981 was applied to the bath from 60 min before the tetanic stimulation to 10 min after (thick bars). Symbols and error bars represent means and SEM, respectively (n = 6 to 10). (E) Developmental asymmetry in LTPs at the left and right com-CA3 synapses. Experimental conditions were the same as those described in (D). Open circles represent 9W mice, and filled triangles represent 2W mice. Symbols and error bars represent means and SEM, respectively (n = 5 to 7).

To examine how this asymmetrical property of NMDA receptors is reflected in the property of synaptic plasticity, we examined the effect of Ro 25-6981 on long-term potentiation (LTP) at these synapses. Similar to Ro 25-6981 sensitivities of NMDA EPSCs, LTP at the left com-CA3 synapses was significantly reduced by Ro 25-6981 (0.2 μM) (relative fEPSP slope 90 min after tetanic stimulation were as follows: control, 158 ± 2%, n = 8 from 8 animals; Ro, 130 ± 2%, n = 10 from 10 animals, P < 0.001) (Fig. 3D), but LTP at the right com-CA3 synapses was apparently unaffected (control, 151 ± 5%, n = 6 from 5 animals; Ro, 151 ± 3%, n = 6 from 6 animals, P > 0.90) (Fig. 3D). To further confirm the asymmetrical contribution of ϵ2 subunits to LTP, we also examined LTP in 2-week-old (2W) mice and compared it with LTP in 9-week-old (9W) mice. The ϵ2 subunit is the major ϵ subunit in the hippocampus of 2W mice (23). LTP in 2W mice was completely suppressed by Ro 25-6981 in both the left and right com-CA3 synapses (left, 99 ± 2 %, n = 5 from 5 animals; right, 101 ± 6 %, n = 5 from 5 animals). In the left com-CA3 synapse, the amplitudes of LTP were similar between 2W and 9W mice (2W mice, 192 ± 10 %, n = 6 from 6 animals; 9W mice, 182 ± 9%, n = 5 from 4 animals, P > 0.46) (Fig. 3E). In the right com-CA3 synapse, however, the amplitude of LTP in 2W mice was smaller than that in 9W mice (2W mice, 154 ± 10 %, n = 6 from 6 animals; 9W mice, 189 ± 5 %, n = 7 from 6 animals, P < 0.01) (Fig. 3E), consistent with the smaller degree of contribution of ϵ2 subunits to LTP in the right.

The preceding observations were complemented by quantitative measurements of ϵ2 and ζ1 subunits in postsynaptic density (PSD) fraction (24) purified from the left and right CA1 stratum radiatum and stratum oriens dissected from VHCT slices. In these preparations, remaining synapses should be mostly made by heavily innervating sch fibers (14), though in the stratum oriens some weak innervation is made by axon collaterals from CA1 pyramidal cells (14).

In both the stratum radiatum and stratum oriens, total amounts of ζ1 and ϵ2 subunits in the homogenate were equal between the left and right hippocampus (Fig. 4A). In PSD fractions, however, we found a significantly higher quantity of ϵ2, but not ζ1 subunit in the left stratum radiatum than in the right and, conversely, in the right stratum oriens than in the left (Fig. 4A). In naïve mice, no significant difference was detected for ϵ2 and ζ1 subunits between PSD fractions from the left and right stratum radiatum (fig. S2) (Supporting Online Material Text).

Fig. 4.

Asymmetrical distribution of ϵ2 subunits in hippocampal synapses. (A) Western blot analyses of ϵ2 subunit proteins in sch-CA1 synapses of VHCT mice. Homogenates from the left and right stratum radiatum (Rad) and stratum oriens (Ori) of the CA1 area were blotted to PVDF membranes, and each membrane was separated into two parts including proteins larger and smaller than 135 kD and were reacted with antibodies to ϵ2 and ζ1 subunits, respectively. Molecular mass markers are indicated (in kD) on the left. In PSD fractions from the left and right stratum radiatum and stratum oriens, immunoreactivity for ϵ2 is stronger in the left Rad and right Ori than it is in the right Rad and left Ori, respectively. Ratios of optical densities (bar graph, left/right ratio in stratum radiatum and right/left ratio in stratum oriens) for ϵ2 and ζ1 immunoreactive bands in the PSD fractions were calculated in three independent experiments and averaged. The ratios for ϵ2 but not ζ1 are significantly different from 1.0. Columns and error bars represent means and SD, respectively (n = 3 each, *P < 0.05, **P < 0.01, t -test). (B) Hippocampal asymmetry proposed here. Left and right pyramidal neurons and their axons are colored red and blue, respectively. Closed and open circles represent ϵ2-dominant and ϵ2-nondominant synapses, respectively. Straight and wavy lines represent inputs from the ipsilateral and contralateral CA3 pyramidal neurons, respectively. Apical, apical dendrites; basal, basal dendrites.

On the basis of the asymmetrical allocation of ϵ2 subunits, hippocampal synapses formed by the inputs from CA3 pyramidal neurons are classified into two populations: one where the ϵ2 subunit is dominant, and the other where it is not. These two populations of synapses are considered to be located asymmetrically on pyramidal neurons (Fig. 4B). Because these populations, with complementary properties formed by ipsilateral and contralateral inputs, are located together on both the apical and basal dendrites of pyramidal neurons, it is not possible to detect asymmetrical distribution of ϵ2 subunits without stimulation selective to ipsilateral or contralateral input. One intriguing property of our model is its input-side selective arrangement of the ϵ2-dominant and ϵ2-nondominant synapses. For example, axons from the left CA3 pyramidal neurons (red lines, Fig. 4B) form ϵ2-dominant synapses (closed circles) on the apical dendrites of pyramidal neurons in bilateral hippocampi. By contrast, axons from the right CA3 pyramidal neurons (blue lines) form ϵ2-dominant synapses on the basal dendrites of pyramidal neurons in bilateral hippocampi. This organization might allow postsynaptic neurons to identify the side of the input source, especially in animals during the early developmental period in which the two populations of synapses differ in their ability to express NMDA receptor–dependent synaptic plasticity (Fig. 3E).

Mechanisms for generating such asymmetrical allocation of ϵ2 subunits may be explained by two distinct hypotheses, based on the differing properties of presynaptic and postsynaptic pyramidal neurons. First, presynaptic signals characterizing the left and right CA3 pyramidal neurons may have opposite properties for recruiting ϵ2 subunits in postsynaptic pyramidal neurons. Second, left and right postsynaptic pyramidal neurons may have opposite properties for sorting ϵ2 subunits into synapses, depending on the identities of ipsilateral or contralateral inputs and apical or basal dendrites. In either case, generation of the asymmetry requires interaction between presynaptic and postsynaptic elements originating from left and right pyramidal neurons with distinct properties.

Although the exact subunit compositions of NMDA receptors in the adult hippocampus are not clear at present, the asymmetrical allocation of ϵ2 subunits seems to differentiate the subtype but not the total number of NMDA receptors between these synapses, because the amount of ζ1 subunits and the amplitudes of NMDA EPSCs, estimated by the NMDA/non-NMDA EPSC ratios, are apparently identical between the left and right. By contrast, the developmental asymmetry observed in LTP at com-CA3 synapses can be explained by the difference in the number of functional NMDA receptors in these synapses. In early postnatal animals such as 2W mice, the expression of ϵ1 subunits in the hippocampus is still absent or low, whereas the ϵ2 and ζ1 subunits are already expressed at a high level (4, 6, 7, 23). In this case, the asymmetrical allocation of ϵ2 subunits may produce distinct numbers of NMDA receptors in these synapses, resulting in differential ability to express synaptic plasticity (Fig. 3E). Hippocampal pyramidal neurons, thus, might regulate the development of synaptic plasticity in a side-selective manner by controlling the synaptic allocation of ϵ2 subunits.

The left-right asymmetry is a fundamental concept of brain science (2527). Our present findings suggest that the brain can involve asymmetries not only at a macroscopic level of left and right hemispheres but also at microscopic levels of neurons and synapses, and they may provide an initial step for elucidating the molecular basis of brain asymmetry.

Supporting Online Material

www.sciencemag.org/cgi/content/full/300/5621/990/DC1

Materials and Methods

SOM Text

Figs. S1 and S2

  • * These authors contributed equally to this work.

References and Notes

View Abstract

Navigate This Article