B-Cell coactivator OBF-1 exhibits unusual transcriptional properties and functions in a DNA-bound Oct-1-dependent fashion

Abstract

Eukaryotic transcriptional activators generally comprise both a DNA-binding domain that recognizes specific cis-regulatory elements in the target genes and an activation domain which is essential for transcriptional stimulation. Activation domains typically behave as structurally and functionally autonomous modules that retain their intrinsic activities when directed to a promoter by a variety of heterologous DNA-binding domains. Here we report that OBF-1, a B-cell-specific coactivator for transcription factor Oct-1, challenges this traditional view in that it contains an atypical activation domain that exhibits two unexpected functional properties when tested in the yeast Saccharomyces cerevisiae. First, OBF-1 by itself has essentially no intrinsic activation potential, yet it strongly synergizes with other activation domains such as VP16 and Gal4. Second, OBF-1 exerts its effect in association with DNA-bound Oct-1 but is inactive when attached to a heterologous DNA-binding domain. These findings suggest that activation by OBF-1 is not obtained by simple recruitment of the coactivator to the promoter but requires interaction with DNA-bound Oct-1 to stimulate a step distinct from those regulated by classical activation domains.

FIG. 1

OBF-1 exhibits transcriptional activity only in conjunction with other transactivation domains. The activity of a lacZ reporter gene integrated at the HIS3 locus and bearing six octamer motifs (OCTA) upstream of the his3 TATA element was assessed in yeast strains expressing (+) or not expressing (−) OBF-1 and the indicated proteins from plasmid DNAs. The Oct-1 variants (and the RFX derivatives depicted in Fig. 3 and 5) were placed under control of the TBP promoter on a single-copy plasmid, and OBF-1 was expressed from the multicopy vector pYES2 (Invitrogen) under the control of the GAL1 promoter. Transformed cells were grown in selective medium containing galactose and assayed for β-galactosidase activity. In this and the other figures, values are relative to the level of β-galactosidase activity seen with either VP16–Oct-1 or VP16-RFX alone, which was assigned a value of 100. A protein immunoblot of whole-cell extracts from strains expressing VP16–Oct-1 either alone (lane 5) or together with OBF-1 (lane 6), using antibodies directed against human Oct-1 or yeast TFIIB, is shown at the top.

FIG. 2

OBF-1 does not stabilize Oct-1 on the octamer motif in vitro, and it forms an unstable complex with Oct-1 in solution. The effect of OBF-1 on the kinetics of Oct-1 binding to (A) or dissociating from (B) its cognate site in the absence and presence of OBF-1 was determined by gel retardation analyses. Nuclear extracts from HeLa cells expressing endogenous Oct-1 were incubated with a radiolabeled oligonucleotide encompassing the same wild-type Ig(κ) octamer site (ATGCAAAT) and flanking sequences as inserted upstream of the lacZ reporter gene depicted in Fig. 3. Where indicated, OBF-1 expressed in rabbit reticulocyte lysates was added in excess such that only the ternary complex containing both OBF-1 and Oct-1 is formed. The association rates were measured by incubating the binding reaction mixtures at 4°C for 1, 2, 5, and 20 min before loading them onto a 4% polyacrylamide gel (A, lanes 4 to 1 and 5 to 8, respectively). The dissociation rates were measured by incubating the binding reaction mixtures for 30 min at ambient temperature before the addition of a 100-fold excess of competitor DNA. The oligonucleotides used as competitor contained either the wild-type Ig(κ) octamer site (ATGCAAAT) or a mutated site (ATGCTAAT) that allows normal binding of Oct-1 but not ternary complex formation (5, 12). Aliquots of the reactions were either directly loaded onto a running gel (B, lanes 4 and 5) or further incubated for 2, 5, and 10 min (lanes 3 to 1 and 6 to 8, respectively). (C) Dissociation rates obtained with either the wild-type Ig(κ) (ATGCAAAT) or the mutated (ATGCTAAT) octamer site as a competitor. The signals were quantitated by PhosphorImager analysis. Similar results were obtained in three independent experiments carried out with different batches of proteins.

FIG. 3

OBF-1 mediates transcriptional synergy without enhancing transcription factor binding to the promoter. The activities of the octamer-containing lacZ reporter genes diagrammed above each panel were assessed in yeast strains expressing the indicated proteins from plasmid DNAs. (A) The promoter bears a single octamer element (OCTA) upstream of the his3 TATA box in place of the six reiterated octamer motifs depicted in Fig. 1. Note that the values are relative to the level of β-galactosidase activity seen with VP16–Oct-1, but the absolute levels of transcription from that promoter are decreased by a factor of 5 to 10 compared to the levels of transcription from the promoter bearing six octamer motifs. (B) The promoter contains a single octamer site surrounded by poly(dA-dT) sequences that impair nucleosome assembly or stability. (C) The promoter contains a single RFX-binding site (X) inserted upstream of six octamer motifs.

FIG. 4

Mapping of the OBF-1 domains critical for interaction with Oct-1 and coactivation function. (A) Strategy used for the high-resolution mapping of the smallest carboxy-terminal OBF-1 truncation capable of interacting with Oct-1 in a one-hybrid screen in yeast. The collection of OBF-1 truncations fused to VP16 was generated by digesting plasmid DNA linearized at the 3′ end of the OBF-1 coding region with exonuclease III for increasing lengths of time. The library was introduced together with a centromeric plasmid expressing Oct-1 into a tester strain carrying an integrated copy of an octamer-dependent his3 allele. Transformants were first grown in medium selecting for both plasmids. Cells expressing OBF-1 derivatives that retain the ability to associate with promoter-bound Oct-1 were subsequently selected in synthetic medium containing 10 mM AT, a competitive inhibitor of the HIS3 gene product. Oct-1BD, Oct-1-binding domain. The native protein is 256 amino acids long, and the Oct-1-binding domain ends at residue 99. (B) Size analysis of the inserts encoding OBF-1 excised from plasmids recovered from large pools of transformants before (w/o AT) and after (+AT) selection for induced HIS3 expression and separated on a 6% acrylamide–urea denaturing gel after end labeling. Lanes G and C show the sequencing reactions of an unrelated DNA fragment of known sequence that were used as markers to determine the boundary of the OBF-1 domain required for interaction with Oct-1 factor in vivo at single-codon resolution. (C) The indicated OBF-1 carboxy-terminal truncations devoid of a VP16 activation domain were examined for the ability to synergize with VP16–Oct-1 on the octamer-dependent lacZ reporter as depicted in Fig. 1.

FIG. 5

OBF-1 exhibits DNA-bound Oct-1-dependent transcriptional activity. (A) The ability of native OBF-1 to potentiate the activity of a VP16–Oct-1-RFX chimera was examined on lacZ promoters bearing either six octamer elements (OCTA) or a single RFX-binding site (X) upstream of the his3 TATA box. OBF-1 may or may not interact with Oct-1 on the RFX-dependent promoter. (B) OBF-1 was noncovalently recruited to the RFX-dependent promoter by fusing to the amino termini of OBF-1 and RFX (or RFX-VP16) the complementary dimerization domains of the human c-Myc oncoprotein and its partner Max, respectively (22). Max-RFX-VP16 is expressed from the GAL1 promoter instead of the TBP promoter. As a control, Myc–OBF-1 was also tested for synergy with VP16–Oct-1 on the octamer-dependent promoter. (C) The indicated amino-terminally truncated forms of OBF-1 were examined for their transactivation potential when recruited to the RFX-dependent promoter by Max-RFX-VP16 through their Myc dimerization domains.