While cystic fibrosis transmembrane conductance regulator (CFTR) is well known to function as a Cl(-) channel, some mutations in the channel protein causing cystic fibrosis (CF) disrupt another vital physiological function, HCO(3)(-) transport. Pathological implications of derailed HCO(3)(-) transport are clearly demonstrated by the pancreatic destruction that accompany certain mutations in CF. Despite the crucial role of HCO(3)(-) in buffering pH, little is known about the relationship between cause of CF pathology and the molecular defects arising from specific mutations. Using electrophysiological techniques on basolaterally permeabilized preparations of microperfused native sweat ducts, we investigated whether: a) CFTR can act as a HCO(3)(-) conductive channel, b) different conditions for stimulating CFTR can alter its selectivity to HCO(3)(-) and, c) pancreatic insufficiency correlate with HCO(3)(-) conductance in different CFTR mutations. We show that under some conditions stimulating CFTR can conduct HCO(3)(-). HCO(3)(-) conductance in the apical plasma membranes of sweat duct appears to be mediated by CFTR and not by any other Cl(-) channel because HCO(3)(-) conductance is abolished when CFTR is: a) deactivated by removing cAMP and ATP, b) blocked by 1 mM DIDS (4,4'-diisothiocyanostilbene-2,2'-disulfonic acid) in the cytoplasmic bath and, c) absent in the plasma membranes of DeltaF508 CF ducts. Further, the HCO(3)(-)/Cl(-) selectivity of CFTR appears to be dependent on the conditions of stimulating CFTR. That is, CFTR activated by cAMP + ATP appears to conduct both HCO(3)(-) and Cl(-) (with an estimated selectivity ratio of 0.2 to 0.5). However, we found that in the apparent complete absence of cAMP and ATP, cytoplasmic glutamate activates CFTR Cl(-) conductance without any HCO(3)(-) conductance. Glutamate activated CFTR can be induced to conduct HCO(3)(-) by the addition of ATP without cAMP. The non-hydrolysable AMP-PNP (5'-adenylyl imidodiphosphate) cannot substitute for ATP in activating HCO(3)(-) conductance. We also found that a heterozygous R117H/DeltaF508 CFTR sweat duct retained significant HCO(3)(-) conductance while a homozygous DeltaF508 CFTR duct showed virtually no HCO(3)(-) conductance. While we suspect that the conditions described here are not optimal for selectively activating CFTR Cl(-) and HCO(3)(-) conductances, we surmise that CFTR may be subject to dramatic alterations in its conductance, at least to these two anions under distinctly different physiological conditions which require distinctly different physiological functions. That is physiologically, CFTR may exhibit Cl(-) conductance with and/or without HCO(3)(-) conductance. We also surmise that the severity of the pathogenesis in CF is closely related to the phenotypic ability of a mutant CFTR to express a HCO(3)(-) conductance.