It has been proposed that P-glycoprotein, the product of the human MDR
1 gene, may function not only as a drug transporter but, depending on
the conditions, as a volume-activated Cl- channel [Valverde, Diaz, Sep
ulveda, Gill, Hyde and Higgins (1992) Nature (London) 355, 830-833; Gi
ll, Hyde, Higgins, Valverde, Mintenig and Sepulveda (1992) Cell 71, 23
-32]. To verify this hypothesis, we have compared volume-activated Cl-
currents with the level of MDR1 mRNA and its protein product in the h
uman KB3 (epitheloid lung cancer) and HeLa cell lines. The related MDR
2 was also included to find out whether it could account for observed
discrepancies between Cl- current and MDR1 expression. A 40% decrease
in osmolarity evoked a Cl- current in both cell types (at + 80 mV: 50.
3 +/- 4.3 pA/pF in KB3, n = 13, 28.2 +/- 3.3 pA/pF in HeLa, n = 16). T
he blocking of this current in both cell types by 5-nitro-2-(3-phenylp
ropylamino)benzoic acid and by 1,9-dideoxyforskolin is similar to that
of the presumed P-glycoprotein associated Cl- channel. As measured by
reverse-transcriptase polymerase chain reaction, KB3 cells expressed
only an extremely small amount of the messengers for MDR1 and MDR2. Th
e signal observed for MDR1 in HeLa cells was at least an order of magn
itude more intense than in KB3 cells, while MDR2 mRNA was undetectable
. A clear difference in MDRI expression between KB3 and HeLa was also
observed at the protein level. These data are difficult to reconcile w
ith the hypothesis that in HeLa and KB3 cells MDR1- or MDR2-encoded P-
glycoproteins are associated with volume-activated Cl- channels.