Hydrolyzing a protein in acid for a single hydrolysis interval, normal
ly 24 h, will lead to inaccurate estimates of the amino acid compositi
on of that protein due to an effect of the time of hydrolysis on pepti
de bond cleavage and amino acid degradation. The simultaneous yield an
d decay of amino acids during the hydrolysis of a protein can be descr
ibed by a compartmental model with parameters for the hydrolysis and l
oss rates specific to each amino acid in a protein. The amino acid com
position of the protein prior to hydrolysis can be determined by nonli
near regression of data derived from multiple hydrolysis intervals. In
the present study egg-white lysozyme was hydrolyzed in 6 M HCl using
18 hydrolysis intervals (range, 2-141 h) using the conventional duplic
ate hydrolyses/interval system. Hydrolysis and loss rates were determi
ned for each amino acid. Increasing the number of hydrolysis intervals
prior to the maximum point on the hydrolysis curve, and including an
hydrolysis interval greater than 100 h increased the accuracy with whi
ch the hydrolysis and loss rates were estimated. Most of the amino aci
ds underwent some degree of loss during hydrolysis, Of particular note
was the loss rate for cysteic acid, which was greater than that found
for serine which is commonly regarded as an acid-labile amino acid. T
he determined amino acid composition of the protein, based on the nonl
inear regression of the data from four different series of hydrolysis
intervals, was compared with the known amino acid composition (sequenc
ing). Using the routine duplicate sampling system, a nonlinear regress
ion including 10 hydrolysis intervals (2, 6, 10, 14, 18, 22, 26, 30, 6
0, and 141 h) resulted in a mean amino acid recovery of 100% (range, 9
4-110%) and provided an acceptable compromise between accuracy and the
cost of analysis. (C) 1996 Academic Press, Inc.