Reviews in Endocrine & Metabolic Disorders 2000;1:69±77 # 2000 Kluwer Academic Publishers, Boston. Manufactured in The Netherlands. Thyroglobulin as Autoantigen: Structure±Function Relationships * Murtaza Vali, 1 Noel R. Rose, 1,2 and Patrizio Caturegli, 2 1 Department of Molecular Microbiology and Immunology and 2 Department of Pathology, The Johns Hopkins Medical Institutions, Baltimore, MD 21205, USA Key Words. thyroglobulin, experimental thyroiditis, Hashimoto's disease, autoantigen Introduction Thyroglobulin (Tg) represents up to 75% of the protein content in mammalian thyroids [1]. It comprises the bulk of the follicular colloid as well as a substantial part of the intracellular material [2]. A large glycoprotein, it consists of two monomeric polypeptide chains that, together, make up a mature 660 kDa, 19 S dimer. The known physiological functions of Tg are to (i) provide a matrix for the synthesis of the thyroid hormones, thyroxine (T4) and triiodothyronine (T3, and (ii) act as a storage vehicle for iodine, in the form of iodinated tyrosyl residues; i.e., monoiodotyrosine (MIT) and diiodotyrosine (DIT) which are the inactive hormone precursors. It is also one of the autoantigens implicated in the pathogenesis of Hashimoto's thyroiditis (HT) and is able to induce disease in animal models described in a number of different species, experimental autoimmune thyroiditis (EAT) [3]. Like Hashimoto's disease, the hallmarks of EAT are lymphocytic in®ltration of the thyroid gland, disruption of the normal thyroid archi- tecture, and the presence of high levels of Tg-speci®c autoantibodies [4]. Thyroglobulin: Primary Structure To date, the complete primary structures of human (h) [5] and bovine (b) (Tg) [6] have been deduced from their respective mRNAs; partial sequences have also been reported for several other species, including rat [7,8], goat [9], sheep [10], rabbit [11], and turtle [12]. The mRNAs for hTg and bTg code for polypeptide chains of 2767 and 2769 residues, respectively. The ®rst 19 residues in both species constitute the signal peptide followed by the mature Tg monomer. Identity at the amino acid level among the different species is also high: 77.3% between human and bovine; 71.8% between human and mouse, and 73.5% between bovine and mouse. An analysis of the Tg amino acid sequence reveals a highly organized internal structure. The N-terminal portion of the molecule (approximately, residues 1- 2170) shows a high degree of internal homology. There are three types of repetitive domains, in which the positions of cysteine residues (and some proline and glycine residues) are highly conserved [5,6]. The type I domain is about 60 residues long and is repeated 10 times between residues 1±1200 (in hTg). Single copies of the domain have been found in other proteins [13], including the major histocompatibility complex (MHC)-class II- associated invariant chain (Ii) [14]. The type II domain is shorter (14±17 residues) and is repeated 3 times between residues 1436±1483 (in hTg). The type III domain, of which there are 2 subtypes, IIIa and IIIb, is repeated 5 times (between residues 1583±2170 in hTg), with IIIa and IIIb repeated twice in tandem followed by a copy of IIIa. In contrast to the N-terminal portion, the C-terminal portion, approximately 550 residues long, shows no internal homology. It does, however, show striking homology to the type B carboxylesterase family, of which acetylcholinesterase is a member (AchE) [15,16,17]. In addition, this part of the molecule has a lower cysteine and higher tyrosine content compared to the monomer as a whole. There is a cluster of tyrosines near the carboxy-terminus, three of which have been shown to be hormonogenic. The lack of homology between the N-terminal and C-terminal portions of the Tg monomer and the difference in the size of introns * This article is based on M.Vali's Sc.M. thesis at The Johns Hopkins University 69