Envelope glycoprotein E1 (gp51 to gp54) is the most antigenic protein of hog cholera virus or classical swine fever virus (CSFV). Four antigenic domains, A to D, have been mapped on E1 with a panel of monoclonal antibodies (MAbs) raised against CSFV strain Brescia. The boundaries of these domains have been established by extensive studies on binding of MAbs to transiently expressed deletion mutants of E1 (P. A. van Rijn, E. J. de Meijer, H. G. P. van Gennip, and R. J. M. Moormann, J. Gen. Virol. 74:2053-2060, 1993). In this study, we used neutralizing MAbs of domains A, B, and C to isolate MAb-resistant mutants (MAR mutants) of CSFV strain Brescia and Chinese vaccine strain ("C"). The E1 genes of MAR mutants were cloned in a eukaryotic expression vector, and the effects of MAR mutations on epitopes were studied with a panel of 19 MAbs by immunostaining of COS1 cells transiently expressing these mutant E1s. Except for the MAR mutation Cys-->Arg at position 792, which abolished binding of all MAbs of domains A and D, amino acid substitutions affected only MAbs belonging to the same domain as the MAb used to select the MAR mutant. However, a MAR mutation in a particular domain did not per se abolish binding of all MAbs recognizing that domain. Furthermore, MAR mutants possessed conservative as well as nonconservative amino acid substitutions. To investigate the significance of a secondary structure for the binding of MAbs, all cysteine residues in the N-terminal antigenic part of E1 were mutated to serine. We found that the cysteines at positions 693 and 737 were essential for binding by MAbs of domains B and C, whereas those at positions 792, 818, 828, and 856 appeared to be essential for the binding of most MAbs of domains A and D. These results fully comply with the previously proposed two-unit structure of the N-terminal half of E1. One unit consists of antigenic domains B and C, whereas the other unit consists of the highly conserved domain A and domain D. We conclude that the first six cysteines are critical for the correct folding of E1. A model of the antigenic structure of E1 is presented and discussed.