Long-term specification of AMPA receptor properties after synapse formation

Abstract

AMPA receptors expressed at auditory nerve synapses in the mammalian and avian cochlear nuclei display exceptionally rapid channel gating, an adaptation well suited for acoustic processing. We examined whether cellular interactions during development might determine the subunit composition of these receptors. After synapse formation in the avian nucleus magnocellularis (nMag) in vivo, the rate of receptor desensitization increased threefold, sensitivity to channel block by polyamines increased, and sensitivity to cyclothiazide, an inhibitor of desensitization, increased, indicating a reduction in glutamate receptor subunit 2 and of flip splice variants. This phenotypic switch was prevented, but not reversed, by isolating nMag neurons in a cell-culture environment. We propose that the switch in receptor kinetics is an outcome of cellular interactions during a critical period that result in the long-term determination of receptor phenotype.

Fig. 1.

Developmental acceleration in AMPA receptor desensitization kinetics. Single exponential time constant of desensitization (τ_des) versus embryonic age (E) of development. Glutamate (10 mm) was rapidly applied to patches excised from nMag neurons in slices at various stages of development. Shown below the axis are major developmental events during this time period. Fromleft, points are mean ± SEM of 6, 3, 9, 10, 8, 11, and 8 patches. Horizontal bars denote variation in developmental time points pooled for each data point.Inset, Scaled representative current traces excised from neurons at E11 (gray), E13.5 (dotted), and E18 (black).Traces are averages of 10–30 events. Time course of exchange (junction potential change) is illustratedabove the current traces.

Fig. 2.

Developmental control of subunit composition indicated by changes in spermine sensitivity.A, B, Families of AMPA receptor-mediated currents at potentials from −75 to +60 mV in 15 mV increments in patches from E11 and E18 neurons. All patches were in the continuous presence of 1 mm intracellular spermine and 60 μm extracellular CTZ. Slowed deactivation inA was inconsistently observed. C, Current– voltage relationship showing, that at positive potentials, there is less inward rectification in E11 (open circles) than in E18 (filled circles) neurons. Currents were normalized to the current amplitude at −60 mV. D, Rectification index, as defined by the ratio of conductance (G) at +30 mV relative to conductance at −60 mV for both E11 (n = 4) and E18 (n = 5) neurons. I–V relationships in patches from E18 neurons (gray bar) are slightly outwardly rectifying (rectification index >1) in the absence of intracellular spermine (n = 3). * indicates rectification ratio significantly less at E18 than E11 (p < 0.001). E11 and spermine-free control data also differed significantly (p < 0.01).

Fig. 3.

Developmental changes in flip–flop splice variants indicated by changes in effectiveness of cyclothiazide.A, Long pulses (2 sec) of 10 mm glutamate in the continued presence of 60 μm CTZ show marked desensitization in E11 (gray trace) and E18 (black trace) patches. Traces scaled to same peak value. Double-exponential fits to the traces with the following parameters are shown: E11, τ_fast 86 msec, τ_slow 588 msec, 28% fast phase, 27% steady-state current; E18, τ_fast 51 msec, τ_slow 462 msec, 38% fast phase, 7% steady-state current. B, The current remaining after desensitization, measured as steady-state/peak ratio in which steady-state current was obtained from best single- or double-exponential fits to the traces during onset of desensitization. n = 5 and 11 patches for E11 and E18, respectively. *p < 0.05 indicates significant difference.

Fig. 4.

Culture environment arrests the developmental switch to fast AMPA receptor kinetics. A, Fluorescent image of contralateral (uninjected) side of slice after retrograde labeling of nMag neurons with TMR. Arrow, Labeled fibers in dorsal acoustic commissure. Double arrow, Labeled nMAG cell bodies in situ. Inset, Representative fluorescent E11 neuron retrogradely labeled and acutely dissociated. Scale bar, 10 μm. B, AMPA receptor desensitization time constant from neurons in slices (filled triangles; data from Fig. 1), acutely dissociated TMR-labeled neurons (open circles;n = 3–9 cells), unlabeled acutely dissociated neurons (gray filled circles;n = 8–9), and neurons cultured from E10 embryos (black filled circles; n = 9–16).C, Representative scaled traces of desensitizing currents in patches from E17 neurons either acutely dissociated (gray trace) or grown for 14–16 d in culture (black trace). Ten millimolar glutamate was applied during the time period marked by the horizontal bar.Inset shows average desensitization time constants of acutely dissociated (n = 10) or cultured E17 (n = 22) neurons. Time constants were not significantly different (p = 0.49).D, Scaled traces from E10 neurons either acutely dissociated (gray trace) or grown for 14–16 d in culture (black trace). Inset shows average desensitization time constants of acutely dissociated (n = 6) or cultured E10 (n = 20) neurons. *p < 0.05 indicates significance.

Fig. 5.

AMPA receptor kinetics are uniform over the entire neuron. Amplitude versus time constant of desensitization for 32 patches from E16–E18 nMag neurons. There was no correlation between patch amplitude and desensitization at these ages, indicating that synaptic and nonsynaptic receptors were indistinguishable in kinetic characteristics. Correlation coefficient is 0.05. Inset, Representative traces from data points, marked by grayand black circles in main figure, which showed a 50-fold variation in amplitude (left), are identical in time course when peak responses are normalized (right).