The 1991 amino acid analysis study examined three subjects: accuracy of absolute quantitation, accuracy of relative yields of amino acids, and the analysis of the problem amino acids tryptophan and cystine/cysteine. The sample chosen for the study was bovine serum albumin, a protein that has been used as a standard for nearly 50 years. Relative to previous ABRF AAA studies, a large amount of protein (150 ug) was sent to each facility in order to allow for a number of analyses and to enhance the overall quality of data. A questionnaire asked for a general strategy of analysis, a single answer of how many nmoles of each amino acid residue was present in the vial, and a number of questions relating to the methodology and calibration.
Data were received from 51 laboratories, seven of which analyzed the sample by two different methodologies for a total of 58 "sites". Precolumn methods were slightly favored over postcolumn methods (57 % vs. 43 %) with PTC (30 sites) and ninhydrin (21 sites) chemistries being most popular. After discarding 18 sites as outliers (differing more than 2 standard deviations from the mean), the average yield of protein relative to the theoretical value of 150 ~g was 91.2 + 6.1 % . This could signify an error in the "theoretical" yield or a pervasive bias. As in previous ABRF AAA studies, the average of the absolute percent error of each residue extended from very good, 2.8%, to very bad, 25%. The best 14 sites showed, < 5% average error, 23 sites showed 5-10% error, and 21 sites showed > 10% error. The latter sites often had difficulties with problem amino acids, e.g. methionine, or glycine contamination, whereas the best sites did not.
Cystine/cysteine was determined by 44 sites using the following methods: oxidation (23 sites); alkylation (4 sites); disulfide exchange (4 sites); and direct determination (12 sites). All methods showed a wide range of errors. Although the data were limited, the alkylation and disulfide exchange methods appeared to be the best. Only 16 sites reported data for tryptophan indicating that its determination is not even attempted by most facilities. In general, the errors were very large, however the fact that BSA only contains 2 tryptophans probably compounded the difficulty of this analysis. Five methods were reported, with methane sulfonic acid hydrolysis being most popular (8 sites). No method was statistically superior, although the thioglycolic acid method was successful in two of the facilities in which it was attempted.
Return to the The ABRF Home Page