Cxlvi. Studies in Synthetic Immunochemistry

IN the first paper of this series [Clutton et al. 1937] a method was described by which simple carbohydrate residues could be introduced into a protein in the form of the O-fl-glycoside of tyrosine, the latter being coupled in peptide linkage with the free amino groups of the protein molecule. Application of this method to the preparation of O-fl-glucosidotyrosylgelatin was described and it was stated that a preliminary study of the product gave no indication of the possession of antigenic properties. The present communication deals with the extension of this work to the preparation of O-,-glucosidotyrosyl derivatives of other proteins and with a detailed serological investigation of the products; the possession of these new derivatives has made possible a reinvestigation of the serological properties of the gelatin compound itself, as a result of which the previous suggestion that it was non-antigenic has to be modified. The additional proteins selected for study were (a) insulin, (b) horse serum albumin and (c) horse serum globulin. The case of insulin presents features of general interest in common with that of gelatin. As was pointed out in our previous paper one of our main objects was to try to discover whether the reason for the generally accepted lack of antigenic power of gelatin might be the absence from its molecule of either carbohydrate or tyrosine or both; such a possibility seemed not unlikely in view of the known importance of both tyrosine and carbohydrate groups in immunological reactions. It is, however, doubtful whether it is fair to regard gelatin as a true protein and the possibility remains that the process of its preparation from collagen involves some destructive action which diminishes or abolishes antigenic capacity; in insulin, on the other hand, we have a compound which may be regarded as a complete protein in so far as its molecular weight and amino-acid composition are concerned; as is well known, however, insulin is non-antigenic and it differs chemically from proteins which are known to be antigenic in containing no carbohydrate. It was therefore of particular interest to examine whether the introduction of glucose residues into insulin would yield an antigenic substance. The derivatives of the serum proteins were prepared with the primary intention of ascertaining how powerfully determinant in an immunological sense the glucosidotyrosyl residue might be, i.e. to what extent the original immunological specificity of the proteins was masked by its introduction. In our previous paper we made a general criticism of earlier work on artificial protein-carbohydrate complexes on the ground that these contained an azo linkage which was foreign