The contribution of carbohydrate residues and peptide chain conformation to autoantibody binding sites on human thyroid peroxidase (TPO) and thyroglobulin (Tg) has been investigated. In addition the nature of carbohydrate residues associated with human TPO has been studied.

¹²⁵I-labelled human TPO and Tg were treated with the following glycosidases: EndoD, EndoH, neuraminidase, O-glycanase, neuraminidase followed by O-glycanase and PNGaseF. Thereafter binding to different sera containing TPO autoantibodies and Tg autoantibodies was assessed using solid phase protein A to separate antibody-bound and free labelled antigens. In addition, labelled Tg and TPO were treated with reducing agent (dithiotreitol) or sodium acetate buffer pH 7.5, 5.5 and 3.2 (followed by neutralisation with 2 M Tris pH 8.3) prior to antibody binding studies. Furthermore, the effect of deglycosylation and treatment with acid buffers on TPO enzyme activity was studied.

The nature of carbohydrate residues associated with hTPO was analysed by assessment of the effects of different glycosidases on ¹²⁵I-TPO mobility on SDS-PAGE followed by autoradiography and by the use of lectins.

Deglycosylation of labelled Tg and TPO had no clear effect on Tg and TPO autoantibody binding. Reduction of labelled Tg and TPO resulted in almost complete loss of autoantibody binding with all sera studied. Furthermore, adjusting the pH of labelled TPO or Tg transiently to pH 5.5 lowered autoantibody binding in the case of all the sera and the effect was more marked at pH 3.2. TPO enzyme activity (guaiacol assay) of unlabelled TPO was decreased after treatment with EndoH but not with other glycosidases. The low pH buffers affected unlabelled TPO enzyme activity measured by iodide assay.

Treatment of ¹²⁵I-labelled TPO with EndoH, neuraminidase and PNGaseF caused marked changes in the double band pattern characteristic of TPO on analysis by SDS gel electrophoresis (TPO doublet). Analysis of changes in the mobility of the 2 bands of the doublet after treatment with different glycosidases and binding studies with lectins indicated that both high mannose and complex type sugar residues were associated with hTPO. The high mannose type residues were associated mostly with the lower band of the hTPO doublet whereas complex type residues were associated mostly with the upper band.

Overall, our studies indicate that (1) the major autoantibody binding sites on hTPO and hTg are conformational, (2) sugar residues do not appear to be important in forming the autoantibody binding sites on hTPO and hTg, and (3) both high mannose type and complex type sugar residues are associated with hTPO.