ELECTROTONIC ARCHITECTURE OF CAT GAMMA-MOTONEURONS

1. Experimental measures of input resistance, R(N), and responses to brief hyperpolarizing current pulses were obtained in identified gamma-motoneurons in pentobarbital-anesthetized cats using conventional sharp micropipettes. The same cells were subsequently injected with horseradish peroxidase and completely reconstructed. In two cells, the electrophysiological and morphological data were of sufficient quality to permit estimation of specific membrane resistance, R(m), using biologically plausible ranges of specific cytoplasmic resistance, R(i), and membrane capacitance, C-m. 2. A combination of steady-state and dynamic computer models were employed to reconcile cell morphology with R(N) and the trajectories of the voltage decay following brief current pulses delivered to the soma. Simulated transient responses matched the tails of the observed transients when generated with the same current injections used experimentally. With C-m less than or equal to 1.0 mu F cm(-2) the most satisfactory fits were obtained when the values of R(m) assigned to the soma, R(ms), were much smaller than the spatially uniform value assigned to the dendrites, R(md) and R(i) = 60 - 70 Omega cm. With C-m = 1.0 mu F cm(-2), R(ms) ranged from 260 to 427 Omega cm(2), whereas R(md) was similar to 33 K Omega cm(2). With C-m = 0.8 mu F cm(-2), R(ms) ranged from 235 to 357 Omega cm(2) and R(md) was between 62 and 68 K Omega cm(2). When R(m) was constrained to be spatially uniform (i.e., R(m) = R(ms)), implausibly high values of C-m (2.5-5.0 mu F cm(-2) R(i) = 70 Omega cm) were required to match the observed tail time constant, tau(0,peel), but the simulated transients did not otherwise match those obtained experimentally. 3. With best fit values of R(ms) and R(md), both gamma-motoneurons were electronically relatively compact (80% of total membrane area within 0.85 length constants from the soma). However, the calculated average steady-state inward attenuation factor (AF(in)) for voltages generated at any point within the dendrites increased rapidly with distance from the soma, reaching levels of less than or equal to 90 and less than or equal to 45 for the proximal 80% of membrane area for the respective motoneurons in the presence of a somatic shunt (R(ms) much less than R(md)). If we assume that the somatic shunt is an artifact of sharp micropipette penetration (i.e., that R(ms) = R(md) for uninjured cells), then AF(i)n decreased to less than or equal to 20 and less than or equal to 15, respectively, for the proximal 80% of cell membrane. 4. The attenuation of synaptic potentials generated at various points within the dendritic tree was studied in simulations of representative, fully branched dendrites with and without somatic shunts and an alpha-function conductance of 5 pS peaking at 0.2 ms (reversal potential = 0 mV; resting membrane potential = -70 mV). The peak amplitudes of somatic excitatory postsynaptic potentials (EPSPs) decreased with increasing electrotonic distance [range similar to 1.0 mV at the soma (X = 0) to 70-100 mu V at X = 1.1], whereas the local EPSPs at the site of generation were as large as 35 mV at X > 0.8. The attenuation of peak synaptic potential amplitudes was related Linearly to the steady-state AF(in) but was 6-10 times larger in the absence of somatic shunt. Introduction of the leak conductance decreased the peak somatic EPSPs by 15-25% but the percent decrement increased with electrotonic distance. 5. We conclude that gamma-motoneurons, like alpha-motoneurons, exhibit large somatic shunts when penetrated with conventional micropipettes and that their passive R(md) values are 1.5- to 2-fold higher than alpha-motoneurons, when similarly estimated. Despite the relatively thin diameters and long lengths of gamma-motoneuron dendrites, these values of R(md) make the trees of gamma-motoneurons as electrotonically compact as those estimated for alpha-motoneurons.