The signal sequence of the p62 protein of Semliki Forest virus is involved in initiation but not in completing chain translocation

brane (Bos et al., 1984; Lipp and Dobberstein, 1986; Spiess and Lodish, 1986; Zerial et al., 1986 ; see also Shaw et al., 1988) . In addition, it is known from physical studies using synthetic signal peptides and artificial lipid membranes that the signal peptides readily insert into the membrane and there obtain an alpha-helical conformation (Briggs et al., 1985 (Briggs et al., , 1986 Batenburg et al., 1988; Cornell et al., 1989) . If the p62 signal sequence adapts such an orientation and conformation in the ER membrane it would mean that the glycosylation site at Asnt3 would locate inside the membrane (von Heijne, 1985) . In this location the site can hardly be accessible for the glycosylation machinery. (Note that in the related Ross River virus, the Venezuelan Equine Encephalitis virus and Eastern Equine Encephalitis virus the corresponding glycosylation site is even closer towards the NH2terminus, that is at Asn 11, see Fig. 7 .) According to several recent studies, glycosylation requires the exposure of the glycosylation site in the lumen of ER. Firstly, it has been shown that the binding protein for the glycosylation site of N-linked oligosaccharides is a lumenal 57-kD protein of the ER (Geetha-Habib et al., 1988) . Secondly, one study with the asialoglycoprotein receptor and another one with the Corona virus E1 membrane protein demonstrate that lumenally oriented glycosylation sites are not used on transmembrane polypeptides if they locate very close to the membranebinding segments of the chains (Mayer et al., 1988; Wessels and Spiess, 1988) . In the case of the asialoglycoprotein receptor a site was not used if located 12 residues apart from the membrane anchor, however, if moved 8 more residues apart from the anchor it became glycosylated. In the case of the Corona virus protein a site just adjacent to the combined signal sequence-anchor peptide remained unglycosylated, whereas an engineered site 24 residues further away was used for glycosylation. Such restrictions in glycosylation are most likely to be explained by sterical problems in attaching the very spacious sugar unit (Lee et al., 1984 ; see also Wier and Edidin, 1988) onto acceptor sites that are fixed in a position which is close to the membrane plane.