NMDA and dopamine converge on the NMDA-receptor to induce ERK activation and synaptic depression in mature hippocampus

Abstract

The formation of enduring internal representation of sensory information demands, in many cases, convergence in time and space of two different stimuli. The first conveys the sensory input, mediated via fast neurotransmission. The second conveys the meaning of the input, hypothesized to be mediated via slow neurotransmission. We tested the biochemical conditions and feasibility for fast (NMDA) and slow (dopamine) neurotransmission to converge on the Mitogen Activated Protein Kinase signaling pathways, crucial in several forms of synaptic plasticity, and recorded its effects upon synaptic transmission. We detected differing kinetics of ERK2 activation and synaptic strength changes in the CA1 for low and high doses of neurotransmitters in hippocampal slices. Moreover, when weak fast and slow inputs are given together, they converge on ERK2, but not on p38 or JNK, and induce strong short-term synaptic depression. Surprisingly, pharmacological analysis revealed that a probable site of such convergence is the NMDA receptor itself, suggesting it serves as a detector and integrator of fast and slow neurotransmission in the mature mammalian brain, as revealed by ERK2 activation and synaptic function.

Figure 1. Time dependency of ERK2 activation by NMDA or dopamine.

Western blot analysis of the time dependency of ERK2 activation, as measured by the ratio of pERK2/ERK2, in response to NMDA or dopamine application to hippocampal slices. NMDA was co-applied with 10 µM glycine, and dopamine was co-applied with 1 mM ascorbic acid within various time frames. Hippocampal slices were harvested in the course of the various time frames, snap frozen on dry ice, homogenized in SDS sample buffer, and subjected to Western blot analysis of the different specimens. Antibodies against the phosphorylated ERK2 were directed to phosphorylation sites Thr183/Tyr185 (here and below). Each specimen comprised two hippocampal slices combined (i.e., 2 slices is n = 1) (this applies to all the following figure captions). A. NMDA was co-applied with glycine, each at a concentration of 10 µM (n = 6 for each time frame). B. NMDA was co-applied with glycine at concentrations of 100 and 10 µM respectively (n = 6 for each time frame). C. dopamine at a concentration 10 µM was co-applied with 1 mM ascorbic acid (n = 6 for each time frame). D. Dopamine at a concentration 100 µM was co-applied with 1 mM ascorbic acid (n = 6 for each time frame). Here and after representative immunoblots are depicted above the graphs.

Figure 2. Time dependency of ERK2 activation following co-application of low doses of dopamine and NMDA.

Western blot analysis of time dependency of ERK2 activation, as measured by the ratio of pERK2/ERK2, in response to co-application of dopamine and NMDA to hippocampal slices. Dopamine at 10 µM was co-applied with 1 mM ascorbic acid, 10 µM NMDA and 10 µM glycine for the different time frames (n = 6 for each time frame).

Figure 3. ERK2 but neither p38 nor JNK serves as a coincidence detector of dopamine and NMDA application.

Western blot analysis of MAPKs activation, as measured by the ratio of phospho-MAPK/total MAPK, induced by 10 min application of either 10 µM NMDA with 10 µM glycine (n = 6), or of 10 µM dopamine with 1 mM ascorbic acid (n = 6), or co-application of 10 µM dopamine, 1 mM ascorbic acid, 10 µM NMDA and 10 µM glycine (n = 6), or control (n = 6). A. ERK2 activation; B. p38 activation. Antibodies against the phosphorylated p38 were directed to phosphorylation sites Thr180/Tyr182; C. JNK activation. Antibodies against the phosphorylated JNK were directed to phosphorylation sites Thr183/Tyr185.

Figure 4. ERK2 activation by dopamine is NMDA receptor dependent, while ERK2 activation by NMDA is dopamine receptor independent.

A. Western blot analysis of ERK2 activation, as measured by the ratio of pERK2/ERK2, induced by 10 min application of either 20 µM NMDA with 10 µM glycine with and without APV 40 µM (n = 4, each), or of 20 µM dopamine with 1 mM ascorbic acid with and without APV 40 µM (n = 4, each). Also shown are control (n = 4) and APV 40 µM applied alone for 30 minutes (n = 4). Each application of either dopamine or NMDA, when co-applied with APV, was pre-perfused with APV 40 µM alone for 30 minutes, prior to their co-application with APV. B. Western blot analysis of ERK2 activation, as measured by the ratio of pERK2/ERK2, induced by 10 min application of either 20 µM NMDA with 10 µM glycine with and without co-application of SCH23390 40 µM and eticlopride 60 µM (n = 4, each), or of 20 µM dopamine with 1 mM ascorbic acid with and without co-application of SCH23390 40 µM and eticlopride 60 µM (n = 4, each). Also shown are control (n = 4), and co-application of SCH23390 40 µM and eticlopride 60 µM for 30 minutes (n = 4). Each application of either dopamine or NMDA, when co-applied with SCH23390 and eticlopride, was pre-perfused with SCH23390 40 µM and eticlopride 60 µM for 30 minutes, prior to their co-application with these antagonists.

Figure 5. The effect of high and low doses of NMDA and dopamine on efficacy of Schaffer-collateral synapses.

The effect of NMDA and dopamine on the efficacy of Schaffer-collateral synapses was determined. A. Strong NMDA stimulation (100 µM, 10 min) resulted in a robust and immediate depression of synaptic efficacy. After washout, synaptic efficacy returned to basal levels within 60 min of treatment (n = 21 slices). Grey bar indicates treatment. Inset illustrates representative traces 2 min before treatment, immediately after treatment and at the end of the experiment. B. Weak NMDA stimulation (10 µM, 10 min) resulted in a rapid depression of synaptic transmission that persisted for at least 120 min after treatment (n = 21 slices). C. Weak dopamine stimulation (10 µM, 10 min) resulted in a gradual potentiation of synaptic efficacy (n = 11 slices). The effect of weak dopamine stimulation was completely inhibited by the NMDA-R antagonist DL-APV (50 µM, n = 3 slices). D. Co-application of weak NMDA (10 µM) and dopamine (10 µM) resulted in a robust and immediate depression of synaptic efficacy. After washout, synaptic efficacy returned to basal levels within 60 min of treatment (n = 15 slices). Note the similarity with the effect of strong NMDA in panel A. E. The NMDA-R antagonist DL-APV (50 µM) completely inhibited the effect of co-application of weak NMDA (10 µM) and dopamine (10 µM) on synaptic efficacy (n = 4 slices). F. Summary statistics of the effect of treatments with neurotransmitters on Schaffer-collateral synaptic efficacy. ‘Baseline’ data were averaged over the first 20 min of recording where no treatment was administered. ‘Treatment’ data represent the average fEPSP during the 10 min where NMDA/Dopamine/APV were administered. ‘End of Experiment’ represents the average fEPSP from 130–140 min from the start of an experiment. No significant differences were observed between the experimental groups during the baseline. Significant differences across treatments were observed during treatment and at the end of experiment using pair-wise post-hoc Bonferroni comparisons. Asterisks indicate significant differences (* = p<0.05, ** = p<0.01, *** = p<0.001). In all panels, error bars indicate SEM.

Figure 6. Strong NMDA treatments do not alter paired-pulse facilitation.

Paired-pulse facilitation (PPF) at an inter-stimulus interval of 50 msec was measured in slices (n = 7) before and after treatment with NMDA (100 µM, 10 min). A. No significant differences were observed in PPF after treatment with NMDA (F_13,6 = 1.3, p<0.3). Grey bars indicate PPF before treatment with NMDA. Black bars indicate PPF after treatment with NMDA. B. Representative traces illustrating PPF.

Figure 7. Application of 100 µM NMDA or co application of NMDA and dopamine, 10 µM each induces reduction in the synaptosomal GluR1 content.

A. Western blot analysis of synaptosomal GluR1 and β-actin content, induced by 10 min application of either NMDA 100 µM with 10 µM glycine (n = 4), or co-application of 10 µM dopamine, 1 mM ascorbic acid, 10 µM NMDA and 10 µM glycine (n = 4), or control (n = 4). Synaptosomal fraction was performed as mentioned previously in the methods. Relative quantification of synaptosomal GluR1 content was measured in relation to β-actin within the same blot. B. Synaptosomal fraction validation. Western blot comparison of synaptosomal enriched fraction, non-synaptosomal fraction and total homogenate with PSD-95, as a synaptosomal marker, and β-actin within the same blot.

Figure 8. An illustration of possible fast and slow neurotransmission convergence.

A. shows that a strong input of fast neurotransmission (which means either large magnitude, long duration, or optimal spacing) causes GluR1 insertion and strong activation of ERK1/2, which may lead to increased protein synthesis and long-term changes (e.g., Late-LTP or Long Term Memory). B. shows that a weak input of fast neurotransmission (weak means small magnitude, too short duration, or non-optimal spacing) would lead to weak ERK1/2 activation, and would result in only short-term changes if any. C. shows that though the fast neurotransmission input is weak (as most ordinary daily physiological input is), the convergence of a slow neurotransmission (e.g., dopamine) upon the fast neurotransmission receptor induces GluR1 insertion and strong and fast ERK2 activation, thus gives it another meaning, and consequently causing long-term changes. The process is NMDA-R dependent and can be mediated via the activation of different types of AC's.