The MAL6 locus is one of five closely related unlinked loci, any one of which is sufficient for fermentation of maltose in Saccharomyces. Previous genetic analysis indicated that this locus is defined by two complementation groups, MALp and MALg. MALp reportedly is a regulatory gene required for inducible synthesis of the two enzymatic functions needed for fermentation: maltose permease and maltase. We have investigated the physical and genetic structure of the MAL6 locus, which has been isolated on a recombinant DNA plasmid. One subclone of the region, pDF-1, was found to encode a single transcribed region and to contain the MALp gene. A second subclone, pl, was shown to contain the MALg function but surprisingly had not one but two maltose-inducible transcripts. Subclones having only one of these transcribed regions lacked MALg activity. The three transcribed regions have been named MAL61 and MAL62, which correspond to MALg, and MAL63, which corresponds to MALp. This clustered arrangement of a regulatory gene adjacent to the sequences it controls has not previously been described in eukaryotes and is reminiscent of bacterial operons except that the messenger RNA molecules are not polycistronic.

Individual mononegavirus genes terminate with a short cis-acting element, the gene-end (GE) signal, that directs polyadenylation and termination and might also influence the efficiency of reinitiation at the next downstream gene. The 12-13 nucleotide (nt) GE signals of human respiratory syncytial virus (RSV) consist of a conserved pentanucleotide (3'-UCAAU, negative sense), followed by a 3-nt middle region that is AU-rich but otherwise not conserved, followed by a 4- or 5-nt poly(U) region that is thought to generate the poly(A) tail of the encoded mRNA by reiterative copying. Most of the naturally occurring differences in the GE signals of the various RSV genes occur in the "middle" and "poly(U)" regions. We mutated a copy of the fusion protein (F) GE signal that was positioned at the end of the promoter-proximal gene of a tricistronic minigenome and evaluated the effect of these mutations on RSV transcription in a plasmid-initiated, intracellular assay. Mutations confirmed the importance of the middle region's AU-rich nature and 3-nt length, and the poly(U) tract's 4-nt minimum functional length, with maximal termination efficiency observed at five U residues. Nt assignments other than U at position 13 also affected the efficiency of termination, showing that this position is part of the functional 13-nt GE signal. These results indicate that differences in nt assignments in the middle and poly(U) regions of the GE signal, which occur frequently in nature, affect the efficiency of termination. Unexpectedly, the ability of certain mutations to inhibit termination was completely dependent on coexpression of the M2-1 protein, and in many other cases the inhibitory effect of the mutation was greatly enhanced in the presence of M2-1. Thus, M2-1 appears to have the effect of altering the polymerase such that it ignores suboptimal GE signals. Interestingly, certain mutations that greatly decreased the efficiency of termination in the absence of M2-1 did not have much effect on the expression of the second gene, implying that correct termination and/or polyadenylation at the upstream gene is not obligatory for reinitiation at the next downstream gene.

We have previously demonstrated that a plasma membrane-enriched fraction isolated from human liver is capable of binding recombinant hepatitis B surface antigen (rHBsAg) (P.