The three isoforms of TGF-beta expressed in mammals are functionally interchangeable in most in vitro assay systems, but they have distinctive activities on certain cells such as endothelial cells where the activities of TGF-beta1 and TGF-beta3 on inhibition of growth are approximately 100- fold more potent than that of TGF-beta2. To define specific regions of the TGF-beta molecule responsible for these differences in activities, we developed a system for expression and purification of recombinant TGF- betas engineered to have portions of different isoforms spliced into a single chimeric molecule. Study of chimeras in which a region of the amino acid sequence of mature TGF-beta1 was spliced into the TGF-beta2 molecule demonstrated that the middle third of the TGF-beta molecule was sufficient to confer isoform-specific biological activity on endothelial cells. We have now shown that only two amino acids (45 and 47) in this region define the specificity of TGF-beta1 or 2 on endothelial cells and that these two amino acids determine the ability of TGF-beta to be sequestered by alpha-2-macroglobulin. By using colorectal carcinoma cells which show selectivity for TGF-beta1 that is independent of alpha-2- macroglobulin, we hope to identify other regions of the molecule which may be involved in isoform-specific receptor binding. We are using this same expression system to produce monomeric TGF-beta and to study possible agonist and antagonistic activities of this monomer. In a related aspect of this problem, we are attempting to characterize the expression of TGF- beta receptors and possible downstream signalling intermediates to determine whether these will be common to all TGF-beta isoforms or whether they will mediate isoform-specific effects.
