Whole – cell potassium currents… , behavioral neurobiology

1108

• volume 2 no 12 • december 1999

As with somatic current injections, activation of AMPA recep-

tor-mediated synaptic inputs (Fig. 3c) elicited EPSPs that were

enhanced by 4-AP (by 6.4 ± 1.7 mV in 6 cells), whereas the single

spike typically elicited by longer dual-component synaptic inputs

occurred with a much shorter lag in 4-AP (Fig. 3d). Dual-com-

ponent EPSPs under control conditions had a slowly rising depo-

larizing phase that was analogous to the upward creep (due to

inactivation of IA) observed in the control voltage responses to

long somatic current injections. This slow phase produced a

∼100-millisecond delay to the peak of the EPSP (99 ± 18 ms;

n = 4), as well as a pronounced delay (71 ± 11 ms; n = 15) and

desynchrony in synaptically evoked spiking. Thus, the NMDA

receptor-mediated depolarization in granule cells outlasts IA,

thereby eliciting spikes after a pronounced lag, whereas the AMPA

receptor-mediated EPSP is counterbalanced by IA.

Intrinsic membrane mechanisms in granule cells

The activation and inactivation properties of IA in granule cells as

described above were largely conventional18, as was the amplitude

of the whole-cell IA current (Ipeak = 1220 ± 120 pA, +2 mV; n = 24;

see refs. 21, 22). We thus considered whether the localization of IA

could contribute to its powerful effect on the excitation of gran-

ule cells. Although the small diameter of the distal dendrites of

granule cells precluded a direct measure of dendritic IA channels, we

made an indirect assessment by comparing whole-cell potassium

currents with those in somatic patches (Fig. 4a and b). The ratio

of the transient (IA) to the steady-state component (IK) of the

whole-cell current (IA/IK = 3.7 ± 0.3, n = 25) was much larger than

in outside-out patches (IA/IK = 1.0 ± 0.2, n = 26) or in cell-attached

somatic patches (IA/IK = 1.1 ± 0.2, n = 4). The whole-cell IA and IK

(n = 9) had activation properties indistinguishable from the patch

currents. Two cell-attached patches taken from proximal dendrites

also showed a small transient component (Fig. 4b). The small tran-

um currents in somatic patches had typical

properties, with an activation threshold of –41 ± 3 mV

(n = 7), a steady-state inactivation midpoint voltage of –63 mV

(n = 2) and rapid inactivation decay kinetics (τ = 0.90 ± 0.11 ms at

–18 mV; n = 8). However, the maximum rate-of-rise of synapti-

cally evoked somatic spikes in granule cells (116 ± 14 V per s;

n = 7) was lower than in other cells (300–600 V per s; refs. 26, 27),

consistent with a relatively low density of sodium channels.

IA drives long-lasting inhibition of mitral cells

The unique mechanism by which IA regulates the excitation of

granule cells at dendrodendritic synapses would also be expected

to affect the output of granule cells, that is, GABA release onto

mitral cells. As observed for granule cell spiking, the GABAA recep-

tor-mediated IPSC in mitral cells is blocked by AP5, implying that

GABA release depends on the activation of granule cell NMDA

receptors11,12. However, in the presence of 4-AP, IPSCs were only

modestly reduced by AP5 (by 29 ± 7%; n = 5), but completely

blocked by subsequent addition of NBQX (n =5; Fig. 5a), indi-

cating that AMPA receptors can support GABA release when IA

is blocked (Fig. 5b). 4-AP also altered the kinetics of the IPSCs.

In 11 mitral cells, 4-AP shortened the IPSC duration (Fig. 5c),

reducing both its decay time constant (by 30 ± 5%) and time to

peak (from 60 ± 5 ms to 27 ± 2 ms). These kinetic changes were

associated with an increased peak amplitude (A4-AP/AControl =

1.53 ± 0.12; n = 11), although no change was detected in the IPSC

charge (Q4-AP/QControl = 1.17 ± 0.20). 4-AP did not change the

kinetics of the unitary synaptic events (τdecay = 13 ± 1 and 13 ± 1

ms before and after 4-AP, respectively; n = 5), implying that the

faster kinetics of the composite IPSC reflects disinhibition of the

early AMPA receptor-mediated EPSP, leading to more synchro-

nous GABA release. Thus, by controlling which receptor type elic-

its activation of granule cells, IA effectively regulates the kinetics

of mitral cell inhibition.

articles

a

b

© 1999 . •

Fig. 2. Properties of IA in granule cells. (a) Whole-

cell potassium currents in granule cells had tran-

sient and steady-state components that were

differentially sensitive to bath application of 4-AP (6

mM) and TEA (10 mM). Currents were induced by

voltage pulses between –58 mV and –8 mV. (b) The

transient current IA measured in nucleated patches

had an inactivation decay time constant near 20 ms.

The threshold for activation of IA in the patch

experiments was near –45 mV, whereas the mid-

point voltage for steady-state inactivation for IA was

near –65 mV. Each point in the activation and

steady-state inactivation curves reflects 3–8 exper-

iments. Steady-state inactivation was evaluated dur-

ing a 40-ms test pulse (+2 mV), preceded by a

200-ms prepulse of varying amplitude.

sient current in patches suggests a relatively high

density of IA channels in distal dendrites of gran-

ule cells, where they would be in a good posi-

tion to attenuate excitatory inputs.

The morphology of granule cells could also

help explain the strong effect of IA on spiking.

Whereas action potentials in many neurons ini-

tiate in the axon hillock or initial segment23,

which contain high densities of sodium chan-

nels24,25, spike initiation in the axonless granule

cell must occur in the soma or dendrites. Sodi-

Connected themes: neuroscience lab, neuroscience lab testing, neuroscience laboratories, Whole – cell potassium currents… , behavioral neurobiology, neuroscience laboratory, neuroscience labs, neuroscience language.