Abstract
The stability of the heme-globin interaction of chemically modified human hemoglobin (Hb) was tested by measuring rates of heme loss from methemoglobin. Heme transfer from methemoglobin to human serum albumin was measured by rapid-scanning spectrophotometry, and the resulting absorption matrices were analyzed by singular value decomposition. Unmodified human HbA0, hemoglobin cross-linked between beta subunits with either 2-nor-2-formylpyridoxal 5'-phosphate or 3,5-(dibromosalicyl)fumarate (DBBF), hemoglobin cross-linked between alpha subunits with DBBF, and pyridoxalated hemoglobin polymerized with either glycolaldehyde or glutaraldehyde were tested. Initial rates were evaluated by fitting the time courses to a biexponential equation using a matrix least squares curve-fitting algorithm. Reaction rates fell into two classes: (1) HbA0 and the site-specifically cross-linked hemoglobins, with biphasic rates of heme loss of 0.02 and 0.004 min-1, and (2) polymerized hemoglobins, with 10-20-fold higher rates at 0.5 and 0.03 min-1. The total fitted amplitudes of the reaction depended upon the specific modification: beta beta-cross-linked Hbs < alpha alpha-cross-linked Hb approximately glycolaldehyde polymerized Hb < glutaraldehyde polymerized Hb < HbA0.