In isolated perfused rat liver, urea synthesis from ammonium ions was dependent on extracellular HCO3- and CO2 concentrations when the HCO3-/CO2 ratio in the influent perfusate was constant (pH 7.4). Urea synthesis was half-maximal at HCO3- = 4 mM, CO2 = 0.19 mM and was maximal at HCO3- and CO2 concentrations above 20 mM and 0.96 mM, respectively. At physiological HCO3- (25 mM) and CO2 (1.2 mM) concentrations in the influent perfusate, acetazolamide, the inhibitor of carbonic anhydrase, inhibited urea synthesis from ammonium ions (1 mM) by 50-60% and led to a 70% decrease in citrulline tissue levels. Acetazolamide concentrations required for maximal inhibition of urea synthesis were 0.01-0.1 mM. At subphysiological HCO3- and CO2 concentrations, inhibition of urea synthesis by acetazolamide was increased up to 90%. Inhibition of urea synthesis by acetazolamide was fully overcome in the presence of unphysiologically high HCO3- and CO2 concentrations, indicating that the inhibitory effect of acetazolamide is due to an inhibition of carbonic-anhydrase-catalyzed HCO3- supply for carbamoyl-phosphate synthetase, which can be bypassed when the uncatalyzed intramitochondrial HCO3- formation from portal CO2 is stimulated in the presence of high portal CO2 concentrations. With respect to HCO3- supply of mitochondrial carbamoyl-phosphate synthetase, urea synthesis can be separated into a carbonic-anhydrase-dependent (sensitive to acetazolamide at 0.5 mM) and a carbonic-anhydrase-independent (insensitive to acetazolamide) portion. Carbonic-anhydrase-independent urea synthesis linearly increased with the portal 'total CO2 addition' (which was experimentally determined to be CO2 addition plus 0.036 HCO3- addition) and was independent of the perfusate pH. At a constant 'total CO2 addition', carbonic-anhydrase-dependent urea synthesis was strongly affected by perfusate pH and increased about threefold when the perfusate pH was raised from 6.9 to 7.8. It is concluded that the pH dependent regulation of urea synthesis is predominantly due to mitochondrial carbonic anhydrase-catalyzed HCO3- supply for carbamoyl phosphate synthesis, whereas there is no control of urea synthesis by pH at the level of the five enzymes of the urea cycle. Because HCO3- provision for carbamoyl phosphate synthetase increases with increasing portal CO2 concentrations even in the absence of carbonic anhydrase activity, susceptibility of ureogenesis to pH decreases with increasing portal CO2 concentrations. This may explain the different response of urea synthesis to chronic metabolic and chronic respiratory acidosis in vivo.