N. Spruston et al., DENDRITIC GLUTAMATE-RECEPTOR CHANNELS IN RAT HIPPOCAMPAL CA3 AND CA1 PYRAMIDAL NEURONS, Journal of physiology, 482(2), 1995, pp. 325-352
1. Properties of dendritic glutamate receptor (GluR) channels were inv
estigated using fast application of glutamate to outside-out membrane
patches isolated from the apical dendrites of CA3 and CA1 pyramidal ne
urons in rat hippocampal slices. CA3 patches were formed (15-76 mu m f
rom the soma) in the region of messy fibre (MF) synapses, and CA1 patc
hes (25-174 mu m from the soma) in the region of Schaffer collateral (
SC) innervation. 2. Dual-component responses consisting of a rapidly r
ising and decaying component followed by a second, substantially slowe
r, component were elicited by 1 ms pulses of 1 mM glutamate in the pre
sence of 10 mu M glycine and absence of external Mg2+. The fast compon
ent was selectively blocked by 2-5 mu M 6-cyano-7-nitroquinoxaline-2,3
-dione (CNQX) and the slow component by 30 mu M D-2-amino-5-phosphonop
entanoic acid (D-AP5), suggesting that the fast and slow components we
re mediated by the GluR channels of the pha-amino-3-hydroxy-5-methyl-4
-isoxazolepropionate (AMPA) and NMDA type, respectively. The peak ampl
itude ratio of the NMDA to AMPA receptor-mediated components varied be
tween 0.03 and 0.62 in patches from both CA3 and CA1 dendrites. Patche
s lacking either component were rarefy observed. 3. The peak current-v
oltage (I-V) relationship of the fast component was almost linear, whe
reas the I-V relationship of the slow component showed a region of neg
ative slope in the presence of 1 mM external Mg2+. The reversal potent
ial for both components was close to 0 m V. 4. Kainate-preferring GluR
channels did not contribute appreciably to the response to glutamate.
The responses to 100 ms pulses of 1 mM glutamate were mimicked by app
lication of 1 mM AMPA, whereas 1 mM kainate produced much smaller, wea
kly desensitizing currents. This suggests that the fast component is p
rimarily mediated by the action of glutamate on AMPA-preferring recept
ors. 5. The mean elementary conductance of AMPA receptor channels was
about 10 pS, as estimated by non-stationary fluctuation analysis. The
permeability of these channels to Ca2+ was low (similar to 5% of the p
ermeability to Cs+). 6. The elementary conductance of NMDA receptor ch
annels was larger, with a main conductance state of about 45 pS. These
channels were 3.6 times more permeable to Ca2+ than to Cs+. 7. AMPA r
eceptor-mediated currents activated rapidly in response to 1 ms pulses
of 1 mM glutamate and deactivated with a predominant, fast time const
ant and a smaller, slower component (tau(1) approximate to 2 ms, tau(2
) approximate to 8 ms, contributing similar to 80 and similar to 20% t
o the total decay amplitude, respectively). Desensitization of the cur
rent during a 100 ms pulse was best fitted by two time constants (tau(
1) approximate to 10 ms, similar to 60%; tau(2) approximate to 34 ms,
similar to 40%). 8. NMDA receptor-mediated currents in response to 1 m
s pulses of 1 mM glutamate activated and deactivated much more slowly
than AMPA receptor-mediated currents. The time course could be describ
ed by a single exponential rising phase (tau approximate to tau ms) fo
llowed by a double exponential decay (tau(1) approximate to 200 ms, si
milar to 80%; tau(2) approximate to 1-3 s, similar to 20%). 9. Mg2(+)
blocked the NMDA component in a voltage-dependent manner, with a half-
maximal inhibitory concentration (IC50) of 21 mu M at -80 mV. At physi
ological Mg2+ concentrations, block of the NMDA component could be rap
idly relieved with voltage jumps from negative to positive potentials.
Block of the current upon return to negative potentials occurred almo
st instantaneously. 10. Zn2+ also selectively blocked the NMDA recepto
r-mediated current with an IC50 of 22 mu M, but this block differed fr
om that of Mg2+ in that it. showed little voltage dependence. Rapid ap
plication of Zn2+ together with glutamate produced partial block of th
e current. More block was observed if Zn2+ and glutamate were co-appli
ed when NMDA receptor channels were already open. 11. The functional p
roperties of dendritic GluRs were similar to those found at the soma.
Knowledge of these properties facilitated simulations investigating th
e contribution of coactivated AMPA and NMDA receptors to synaptic depo
larization and Ca2+ entry into dendritic spines. Because of its slow d
eactivation, the NMDA receptor-mediated current contributes substantia
lly to depolarization and Ca2+ entry and is susceptible to modulation
over a period of seconds, either by backpropagating action potentials
or by the release of Zn2+ from presynaptic boutons.