Monovalent anions such as isethionate (HO-C,H,-SO,-), methyl sulfate (CH,-SO,-), acetate, and others are commonly used as substitutes for chloride in physiological salt solutions (l-6). However, it has been recently reported by Christoffersen and Skibsted that a number of chloride substitutes can bind calcium quite strongly (7). For example, the log K,,, for isethionate and calcium was stated to be +l. 15, while the value for methyl sulfate was stated to be +0.5. On this basis they suggested that some physiological results previously attributed to low chloride might instead be due to dramatically lowered calcium levels in the medium. The conclusions were based on measurements of calcium concentration in 0.15 M sodium salt solutions with a calcium selective microelectrode and a saturated KC1 calomel electrode.
It was evident to us that some previous in vitro studies performed in chloride-substituted physiological solutions might have to be reevaluated if these results were correct. Therefore, we reexamined the presumed calcium binding ability of these and other anions.
In a preliminary study with arsenazo III, a chromogenic calcium chelating agent (a), we were unable to detect calcium binding by isethionate (9). However, the presumed interaction between isethionate and calcium would not be detected by this method if the association constant for arsenazo III and calcium were substantially greater than that for isethionate. On this basis, arsenazo III would simply strip calcium from its possible association with isethionate.
In this paper we have utilized a Ca2+-specific electrode to measure free calcium and now report that these salts have quite weak but detectable affinities for calcium. Our primary result is that the calcium ion activity is only slightly lowered in the presence of the chloride substitutes and that the single ligand association constants are substantially lower than those previously reported by Christoffersen and Skibsted (7).
In the present studies, calcium ion activity was measured in 0.5ml samples with an Orion Biomedical ionized calcium analyzer Model SS-20, and data were recorded on a Texas Instruments Servo/Riter II. A constant sodium ion concentration, as indicated below, was maintained for all experiments, and the ionic strength of all solutions was constant within lo%, including standards. Thus, since the activity coefficients are the 311 Copyright