RESUMO
1. Hexosaminidases were studied by electrophoresis with different human fibroblast extracts. We found in the same conditions of detection and culture three bands from the cathode to the anode, namely Hex B, Hex A, Hex C for the normal fibroblast, Hex B for the two different Tay-Sachs and Hex C for the two unrelated Sandhoff patients. 2. The analysis of man-rodent hybrids (hamster and mouse with normal and Sandhoff human fibroblasts) indicates a probable synteny between MPI, Hex C, "Hex A fast", and "Hex A-like". "Hex A fast" is probably a man-hamster hybrid enzyme, "Hex A-like" a man-mouse enzyme. Our data agree with the model of Ropers and Schwantes (Hex C = (alphaalpha)n; Hex A = (alphabeta)n; Hex B = (betabeta)n). Probably Hex A-fast = (alphabeta')n with hamster Hex B' = (beta'beta')n; and Hex A-like = (alphabeta1)n with mouse Hex B1 = (beta1beta1)n; and probably n = 2 according to the tetrameric structure model of Tallman et al. (1974). 3. As an explanation of the results given by Poenaru et al. (anti Hex A reacts with Hex A and Hex B but not with Hex C) we propose the existence of a compound antigen (alphabeta) for Hex A. Anti Hex A specific = anti (alphabeta); anti Hex A non-specific = anti Hex B = anti B, anti alpha being absent or negligible. 4. In our opinion, the Tay-Sachs mutation opposes the alphaB association while the alphaalpha association is possible at a low rate or unstable; it is thus possible to observe Hex C in certain conditions, e.g. in foetal brain. 5. We present a discussion about the genetic control of hexosaminidases, GM2 gangliosidosis, and the possible localization of the different mutations in the variants.