Functional dissection of the human spumaretrovirus transactivator identifies distinct classes of dominant-negative mutants.

The bel1 gene of human spumaretrovirus (HSRV) codes for a 300-amino-acid nuclear protein, termed Bel1, that can strongly activate transcription from the cognate long terminal repeat (LTR) by at least 200-fold. Bel1 can also activate human immunodeficiency virus type 1 (HIV-1) LTR expression. By using site-directed mutagenesis, we have identified distinct regions of Bel1 essential for HSRV LTR activation. The amino-terminal 55 residues, which comprise a highly acidic region followed by a short basic stretch, were dispensable for activation. The distribution of functionally defective mutants indicates that two distinct regions between residues 56 and 300 cooperate to confer full activator function. The larger, more amino-terminal region between residues 56 and 227 is sufficient to minimally activate the HSRV LTR. It contains a region between residues 88 and 110 that is strongly conserved between the simian and human spumavirus transactivators but otherwise lacks obvious homology to known transcriptional activators except for an Arg-rich nuclear localization sequence (NLS) between residues 211 and 225 that can be functionally substituted for by the NLS of the simian virus 40 large T antigen. The carboxy-terminal 73 residues contain two functionally redundant regions that can independently augment the activity of the more N-terminal minimal activator domain by 30- to 90-fold. Comparative analysis of the effect of Bel1 mutations on HSRV and HIV-1 LTR expression revealed a similar requirement of Bel1 domains for activation of the two LTRs. Bel1 is phosphorylated in vivo, and a nuclear localization-defective mutant lacking residues 211 to 222 was severely defective for phosphorylation, whereas various deletion mutations in residues 228 to 300 resulted in a four- to eightfold reduction in phosphate incorporation. When functionally defective bel1 mutants were examined for a dominant-negative phenotype, only mutants lacking a proline-rich basic region between residues 194 and 200 or the NLS between residues 211 and 222 that were found to occupy predominantly nuclear and cytoplasmic locations, respectively, could suppress wild-type Bel1 function efficiently. In identifying two classes of dominant-negative mutants with distinct subcellular localization phenotypes, the mutational analysis of Bel1 has revealed a feature unusual for known transcriptional activators.