The amino acid composition of proteins

Abstract

Published data on the amino acid analyses of various proteins have been recalculated to give sets of values representing single or average determination on 80 separate proteins, as residues of amino acid/100 residues in each protein. These values, when plotted as histograms, suggested that the occurrence of an individual amino acid in proteins in general could be specified by a mean value and its standard deviation. This was tested by comparing the histograms with those expected from the calculated mean, assuming a Gaussian distribution. The agreement was generally satisfactory when non-occurrences were taken into account. In the case of tryptophan the agreement was not good because there were abnormally large numbers of non-occurrences. It was felt that this probably reflected the known difficulties in analysing for this amino acid. The occurrence of specific amino acids in individual proteins did not follow any discernible pattern, nor could any correlation be made between amino acid composition and function. However, it was possible to rank the 80 proteins according to their deviation from an “average protein” deduced from the mean values for each amino acid. Those proteins in the middle of the ranking were average in the more usual sense of being neither close to the “average protein” nor unusually far from it. These results are interpreted as showing that all proteins are based upon a common pattern of amino acid composition. On the basis of the present data this can be fairly well represented by the following values: Asp 25, Glu 24, Leu 20, Ala 19, Gly 17, Ser 17, Val 16, Lys 15, Pro 13, Thr 13, Ben 11, Arg 10, Phe 10, Tyr 8, His 5, Met 4, CyS 3, Try 3. These are the numbers of residues making up an “average protein” of 233 residues and of molecular weight 25,700. The bearing of these results on protein function and evolution is discussed. The method of analysis gives a means of comparing the amino acid compositions of different proteins in a manner which is open to statistical analysis. The results obtained are of theoretical interest for protein chemists, and of practical interest in studies where the amino acid composition of a protein can be of importance.