The reversibility and cation selectivity of the K+-Cl- cotransporter (KCC),
which normally extrudes Cl- out of neurons, was investigated in dissociate
d lateral superior olive neurons of rats using the gramicidin perforated pa
tch technique. Intracellular Cl- activity (a[Cl-](i)) was maintained well b
elow electrochemical equilibrium as determined from the extracellular Cl- a
ctivity and the holding potential, where the pipette and external solutions
contained 150 mM K+ ([K+](pipette)) and 5 mM K+ ([K+](o)), respectively. E
xtracellular application of 1 mM furosemide or elevated [K+](o) increased a
[Cl-](i). When the pipette solution contained 150 mM Cs+ ([Cs+](pipette)),
a[Cl-](i) increased to a value higher than the passive a[Cl-](i). An increa
se of a[Cl-](i) with the [Cs+](pipette) was not due to the simple blockade
of net KCC by the intracellular Cs 1 since a[Cl-](i), with the pipette solu
tion containing 75 mM Cs+ and 75 mM K+, reached a value between those obtai
ned using the [K+](pipette) and the [Cs+](pipette). The higher-than-passive
a[Cl-](i) with the [Cs+](pipette) was reduced by 1 mM furosemide, but not
by 20 mu M bumetanide or Na+-free external solution, indicating that the ac
cumulation of [Cl-](i) in the [Cs 1]pipette was mediated by a KCC operating
in a reversed mode rather than by Na+-dependent, bumetanide-sensitive mech
anisms. Replacement of K+ in the pipette solution with either Li+ or Na+ mi
micked the effect of Cs+ on a[Cl-](i). On the other hand, Rb+ mimicked K+ i
n the pipette solution. These results indicate that K+ and Rb 1, but not Cs
1,Li 1,or Na 1, can act as substrates of KCC in LSO neurons.