Abstract

In larger mammals, thyroid hormone-binding plasma proteins are albumin, transthyretin (TTR) and thyroxine (T4)-binding globulin. They differ characteristically in affinities and release rates for T4 and triiodothyronine (T3). Together, they form a 'buffering' system counteracting thyroid hormone permeation from aqueous to lipid phases. Evolution led to important differences in the expression pattern of these three proteins in tissues. In adult liver, TTR is only made in eutherians and herbivorous marsupials. During development, it is also made in tadpole and fish liver. More intense TTR synthesis than in liver is found in the choroid plexus of reptilians, birds and mammals, but none in the choroid plexus of amphibians and fish, i.e. species without a neocortex. All brain-made TTR is secreted into the cerebrospinal fluid, where it becomes the major thyroid hormone-binding protein. During ontogeny, the maximum TTR synthesis in the choroid plexus precedes that of the growth rate of the brain and occurs during the period of maximum neuroblast replication. TTR is only one component in a network of factors determining thyroid hormone distribution. This explains why, under laboratory conditions, TTR-knockout mice show no major abnormalities. The ratio of TTR affinity for T4 over affinity for T3 is higher in eutherians than in reptiles and birds. This favors T4 transport from blood to brain providing more substrate for conversion of the biologically less active T4 into the biologically more active T3 by the tissue-specific brain deiodinases. The change in affinity of TTR during evolution involves a shortening and an increase in the hydrophilicity of the N-terminal regions of the TTR subunits. The molecular mechanism for this change is a stepwise shift of the splice site at the intron 1/exon 2 border of the TTR gene. The shift probably results from a sequence of single base mutations. Thus, TTR evolution provides an example for a molecular mechanism of positive Darwinian evolution. The amino acid sequences of fish and amphibian TTRs are very similar to those in mammals, suggesting that substantial TTR evolution occurred before the vertebrate stage. Open reading frames for TTR-like sequences already exist in Caenorhabditis elegans, yeast and Escherichia coli genomes.