Two-stage, input-specific synaptic maturation in a nucleus essential for vocal production in the zebra finch

Abstract

In most songbirds, vocal learning occurs through two experience-dependent phases, culminating in a reduction of behavioral plasticity called song crystallization. At ends of developmentally plastic periods in other systems, synaptic properties change in a fashion appropriate to limit plasticity. Maturation of glutamatergic synapses often involves a reduction in duration of NMDA receptor (NMDAR)-mediated synaptic responses and a coincident reduction in the contribution of NMDARs to synaptic transmission. We hypothesized that similar changes in the zebra finch song system help limit behavioral plasticity during song development. Nucleus robustus archistriatalis (RA) is a key nucleus in the forebrain song motor pathway and receives glutamatergic input from the motor nucleus HVc. RA also receives glutamatergic input, mediated primarily by NMDARs, from the lateral magnocellular nucleus of the anterior neostriatum, which is part of a circuit essential for learning but not song production. We examined whether synaptic maturation occurs in either input to RA by recording synaptic currents in brain slices prepared from zebra finches of different ages. We find the motor input from HVc to RA uses both AMPA receptors (AMPARs) and NMDARs, and synaptic maturation occurs in two phases: an early reduction in duration of NMDAR-mediated synaptic currents in both inputs, and a later reduction in the NMDAR contribution to synaptic responses in the motor pathway. Although NMDAR kinetics change too early to account for crystallization, the reduction of the relative NMDAR contribution to synaptic transmission could contribute to the onset of crystallization. Thus, synaptic maturation events can be temporally distinct and input-specific and may play different roles in behavioral plasticity.

Fig. 1.

Developmental time line and experimental preparation. A, Developmental time line of song learning showing the four age groups of birds examined in relation to the phases of song learning. B, Schematic drawing of a sagittal view of the male zebra finch brain, depicting a highly simplified song system. There are two major pathways: the motor pathway (HVc, RA, and nXIIts; open nuclei) is necessary for song production; the anterior forebrain pathway (area X, DLM, and lMAN;shaded nuclei) is necessary for song learning but not for song production. C, Schematic drawing of a coronal slice preparation. We stimulated afferent pathways from HVc and lMAN and recorded whole-cell synaptic currents from neurons in nucleus RA. These inputs converge on single RA neurons such that EPSCs from each pathway can be recorded in a single postsynaptic neuron.DLM, Medial portion of the dorsolateral nucleus of the anterior thalamus; HVc, used here as a proper name;lMAN, lateral magnocellular nucleus of the anterior neostriatum; nXIIts, tracheosyringeal portion of the hypoglossal motor nucleus; RA, robust nucleus of the archistriatum; X, area X of the parolfactory lobe.

Fig. 2.

EPSCs from HVc and lMAN inputs are mediated by different complements of glutamate receptor subtype. EPSCs elicited by HVc afferent stimulation (A, C) are mediated by both AMPARs and NMDARs, whereas EPSCs elicited by lMAN afferent stimulation (B, D) are mediated almost exclusively by NMDARs. Each panel shows three EPSCs (each an average of 5–20 traces), overlaid and aligned at the time of stimulation. Some stimulus artifacts have been clipped for clarity. APV blocked the NMDA component, revealing a substantial AMPA component in the HVc–RA EPSC (A), and a negligible AMPA component in the lMAN–RA EPSC (B); AMPA components were subsequently blocked by CNQX. Initial application of CNQX blocked the AMPA component in the HVc–RA EPSC, revealing a substantial NMDA component (C), but had minimal effect on the lMAN–RA EPSC (D); NMDA components in both pathways were subsequently blocked by APV. EPSCs in Aand B are from a single neuron from a 43 DPH bird; EPSCs in C are from an adult bird; EPSCs in Dare from a 42 DPH bird.

Fig. 3.

Comparison of NMDA:AMPA ratio between HVc–RA and lMAN–RA EPSCs. N:A ratio values (and median plus interquartile range) for all ages examined, plotted on a log scale. The dotted line indicates a 1:1 ratio. In this and subsequent figures, each individual symbol represents data from one cell. Median N:A ratio for lMAN–RA EPSCs is approximately tenfold higher than that for HVc–RA EPSCs, illustrating that lMAN–RA EPSCs are mediated predominantly by NMDARs, whereas HVc–RA EPSCs are mediated by both NMDARs and AMPARs.

Fig. 4.

Developmental reduction in NMDAR contribution to total synaptic current. A₁, Scatterplot of NMDA:AMPA ratio versus age for HVc–RA EPSCs. In this and subsequent figures, cells from adult birds are plotted at 100 DPH.A₂, Mean N:A ratio values (± SD) for HVc–RA EPSCs. The N:A ratio decreased during development, but post hoc tests did not identify differences between groups. B₁, Scatterplot of Late:Early ratio, an independent measure of the contribution of NMDAR to the total EPSC, versus age.B₂, Mean Late:Early ratio values (± SD) for HVc–RA EPSCs. Late:Early ratio also decreased during development. Post hoc tests revealed that the decrease in Late:Early ratio occurred between the juvenile and adult stages, but inspection of the scatterplot (B₁) suggests that the decrease may occur during the juvenile stage. See Table 1 for statistics.NS, Not significant.

Fig. 5.

Example traces showing a developmental decrease in duration of NMDAR-mediated EPSCs. Normalized EPSCs from different ages for both lMAN (A) and HVc (B) inputs to RA. The duration of the isolated NMDAR-mediated current decreased dramatically between the nestling stage (20 DPH in A, 23 DPH in B) and the fledgling stage (34 DPH in A, 28 DPH inB). There was no further change between fledglings and adults. Juvenile examples are omitted for clarity, but the time course resembles that of the fledgling and adult traces.

Fig. 6.

e-fold decay time decreases early in song development. A₁, Scatterplot of e-fold decay times versus age for lMAN–RA EPSCs. For this and subsequent figures, filled symbols are from males,open symbols are from females or birds of unknown sex (see Materials and Methods). A₂, Mean e-fold decay time (± SD) for lMAN–RA EPSCs.B₁, Scatter plot of e-fold decay times versus age for HVc–RA EPSCs.B₂, Mean e-fold decay time (± SD) for HVc–RA EPSCs. For both inputs, nestlings had significantly longer e-fold decay times than any of the other age groups (***p < 0.001, comparing nestlings with each of the other groups). There were no significant differences among fledglings, juveniles, and adults, indicating that e-fold decay time stabilized by the fledgling stage. See Table 1 for statistics.

Fig. 7.

Decay time constants decrease early in song development. Scatterplots of three variables (τ_f, τ_s, % slow) from the double-exponential fits of lMAN EPSCs (A) and HVc EPSCs (B) versus age. Horizontal lines indicate the mean for each value within an age group. Except for adults, the length of each line indicates the age range of each group used for ANOVA analysis. Groups are as in Figure 6. For both input pathways, both the fast (τ_f) and slow (τ_s) decay time constants of NMDAR-mediated EPSCs decreased with development. The most influential effect on the overall time course of EPSCs was the decrease in τ_s between the nestling and fledgling stages. The relative amplitude of the slow component (% slow) showed a small but significant decrease between the nestling and fledgling stages in the lMAN pathway, but did not change during development in the HVc pathway. See Table 1 for statistics.

Fig. 8.

Schematic representation of two-stage, input-specific synaptic maturation in RA neurons. A, In nestlings, NMDARs with slow decay kinetics (filled ovals) mediate synaptic transmission at lMAN–RA synapses. At this age, HVc inputs are rare and are not shown here; however, the few NMDAR-mediated currents we have recorded had slow decay times.B, At the fledgling and juvenile stages, NMDARs mediating both HVc–RA and lMAN–RA EPSCs have fast decay kinetics (open ovals). HVc inputs are mediated by both NMDARs and AMPARs (filled rectangles) with an ∼2:1 ratio.C, By adulthood, the relative contributions of NMDARs and AMPARs have decreased to an ∼1:1 ratio, and the decay kinetics of NMDARs remain fast in both input pathways. The symbol key,inset in A, applies toA–C. D, Summary of our observations, depicting the spatial and temporal dissociation of two forms of synaptic maturation in nucleus RA.