Cellular and structural biology of the deiodinases

The three iodothyronine deiodinases catalyze the initiation (D1, D2) and termination (D3) of thyroid hormone effects in vertebrates. A recently conceived three-dimensional model predicts that these enzymes share a similar structural organization and belong to the thioredoxin (TRX) fold superfamily. Their active center is a selenocysteine-containing pocket defined by the beta 1-alpha 1-beta 2 motifs of the TRX fold and a domain that shares strong similarities with the active site of iduronidase, a member of the clan GH-A fold of glycoside hydrolases. All three deiodinases form homodimers through disulfide bridges when transiently expressed but because these enzymes are present at such low levels in vivo, it is not clear if deiodinase dimers are formed at endogenous levels. At least for D1 and D2, dimers are catalytically active but only one monomer partner is required for catalytic activity. While D1 and D3 are long-lived plasma membrane proteins (t1/2 10-12 hour), D2 is an endoplasmic reticulum resident protein with a half-life of approximately 40 minutes. Exposure to thyroxine (T-4) shortens D2 half-life even further (similar to 10 min) while during hypo-thyroidism D2 activity disappears with a half-life of approximately 5 hours. This D2 inactivating mechanism is mediated by selective conjugation to ubiquitin, a process that is accelerated by T4 catalysis and thus maintains local triiodothyronine (T-3) homeostasis. Remarkably, D2 ubiquitination is reversible and activity restored after deubiquitination. This is because D2 interacts with and is a substrate of the pVHL-interacting deubiquitinating enzymes (VDU1 and VDU2), and thus the ubiquitination-deubiquitination cycles regulates the supply of active thyroid hormone in D2-expressing cells.