Neonatal inoculation of the wild-mouse ecotropic retrovirus CasBrE (clone 15-1) causes a noninflammatory spongiform neurodegenerative disease with an incubation period of > or = 6 months. Introduction of sequences from Friend murine leukemia virus (clone FB29) into the genome of CasBrE results in a marked shortening of the incubation period. The FB29 sequences which influence the incubation period were previously localized to the 5' leader sequence of the viral genome (M. Czub, F. J. McAtee, and J. L. Portis, J. Virol. 66:3298-3305, 1992). In the current study, we constructed a series of chimeric viruses consisting of the genome of CasBrE containing various segments of the leader sequence from FB29. A 41-nucleotide element (positions 481 through 521) near the 3' end of the leader was found to have a strong influence on the incubation period. This element influenced the kinetics of virus replication and/or spread in nonneuronal tissues, a property which was shown previously to determine the extent of central nervous system infection (M. Czub, F. J. McAtee, and J. L. Portis, J. Virol. 66:3298-3305, 1992). Curiously, this sequence had no demonstrable effect on virus replication in vitro in a fibroblastic cell line from Mus dunni. This segment encodes 14 of the unique 88-amino-acid N terminus of pr75gag, the precursor of a glycosylated form of the gag polyprotein which is expressed at the cell surface. Previous in vitro studies of mutants of Moloney murine leukemia virus lacking expression of glycosylated Gag failed to reveal a function for this protein in virus replication. We mutated the Kozak consensus sequence around the initiation codon for this protein in the chimeric virus CasFrKP, a virus which induces neurologic disease with a short (18- to 23-day) incubation period. M. dunni cells infected with the mutants lacked detectable cell surface Gag, but, compared with CasFrKP, no effect on replication kinetics in vitro was observed. In contrast, there was a marked slowing of the replication kinetics in vivo and a dramatic attenuation of neurovirulence. These studies indicate that glycosylated Gag has an important function in virus replication and/or spread in the mouse and further suggest that the sequence of its N terminus is a critical, though likely indirect, determinant of neurovirulence.