Low solubility of unconjugated bilirubin in dimethylsulfoxide--water systems: implications for pKa determinations

Aqueous pKa values of unconjugated bilirubin are important determinants of its solubility and transport. Published pKa data on an analog, mesobilirubin-XIIIalpha, studied by 13C-NMR in buffered solutions containing 27 and 64 vol% (C2H3)2SO because of low aqueous solubility of mesobilirubin, were extrapolated to obtain pKa values in water of 4.2 and 4.9. Previous chloroform-water partition data on bilirubin diacid led to higher estimates of its pKa, 8.12 and 8.44, and its aqueous solubility. A thermodynamic analysis, using this solubility and a published solubility in DMSO, suggested that the systems used to measure 13C-NMR shifts were highly supersaturated. This expectation was assessed by measuring the residual concentrations of bilirubin in the supernatants of comparable DMSO-buffer systems, after mild centrifugation to remove microprecipitates.

Extensive sedimentation was observed from numerous systems, many of which appeared optically clear. The very low supernatant concentrations at the lowest pH values (4.1-5.9) were compatible with the above thermodynamic analysis. Extensive sedimentation and low supernatant concentrations occurred also at pH as high as 7.2.

The present study strongly supports the validity of the aqueous solubility of bilirubin diacid derived from partition data, and, therefore, the corresponding high pKa values. Many of the mesobilirubin systems in the 13C-NMR studies were probably supersaturated, contained microsuspensions, and were not true solutions. This, and previously documented errors in pH determinations that caused serious errors in pKa values of the many soluble reference acids and mesobilirubin, raise doubts regarding the low pKa estimates for mesobilirubin from the 13C-NMR studies.