\: wee uk pe RAG Det De veeAEE en Porr/Baicirs/! fiesceH/CONRAD 214 Discussion The procedures used to purify human hemoglobin were chosen because nonhemoglobin proteins of higher and lower molecular weights than hemoglobin can be removed by molecularseiving and nonhemoglobin proteins of molecular weights similar to hemoglobin can be separated by repeated chromatographyof dlerivatized hemoglobin on ion exchangers. Proteins that do not contain heme should chromatograph similarly on carboxymethyl cellulose each of the three times because the same ion exchanger, buffers and gradient arc employed; however, the chromatographic properties of the derivatized hemoglobins are different. Thus, nonhemoglobin proteins that contaminate onc form of hemoglobin should be removed during the next chromatography. Human fetal hemoglobin is a protein that does contain isoleucine and its chromatographic behavior on carboxymethyl cellulose is changed by deri. atization much like that of hemoglobin A. However, hemoglobin F has a different isoelectric point andit has different chromatographic properties than adult hemoglobin; 95% of any hemoglobin Fpresent is removed by the repeated chromatography over carboxymethyl cellulose. The potential contamination by hemoglobin F was determined by starting with a chemically determined mixture of 90% adult and 10% fetal hemoglobin. Chromatographyover carboxymethyl cellulose removed 80, 50, and 50% of the fetal hemogiobin at each successive step using these procedures (unpublished). Jf human hemoglobin on the average contains 1% hemoglobin F, the hemoglobin F remaining after the third chromatography could contribute 7 parts of isolettcine per million amino acid residues. This is fourfold less than the isoleucine content found in highly purified human hemoglobin A (table 1). Moreover, hemoglobin F cannot be the principal source of the isoleucine in the globin samples because isoleucine is found in both the «- and f-chain polypeptides (table 1); the «-chain is, but the B-chain is not, cleanly separated from the y-chain by the procedures used. Furthermore, in marmoset and shecp the *H-isoleucine and optical density profiles of the separated o- and B-chains are coincident [unpublished], which indicates that the iso- leucineis in hemoglobin rather than in nonhemoglobin protein contaminants. co MD <7 fwd co cH Chemical analyses show that highly purified human hemoglobin A has a very low content of isoleucine. A slight, but insignificant increase with age in the isoleucine substitution frequency is suggested on the basis of a least squares treatment of the data; the individual data points, except for sample 1547, fit well with a linear increase of 0.0296 x 10-*/year between ages 20 and 51 years. Sample 1547 deviated significantly from linearity; this sample