Flexibility in the second half-site sequence recognised by the c-Myb R2 domain--in vitro and in vivo analysis.

The oncoprotein c-Myb is a transcription factor that recognises its specific target sequences through two subdomains. The R3-domain binds the first half-site, YAAC, and plays a dominant role in sequence recognition, while the homologous R2-domain interacts with a more loosely defined sequence in the second half-site. The difficulty in precisely defining a preferred second half-site sequence might reflect the flexible nature of R2 which only attains its fully folded structure upon binding to DNA, a process that might allow the protein to adapt to different half-site sequences. Here we report that shifting the most conserved base in the second half-site, the G6, into position 5 resulted only in a minor reduction of complex stability in vitro. From an analysis of a series of second half-site variants by EMSA and DMS-interference, we conclude that the preferred recognition sequence should be revised to read [YAACNG or YAACGN]. Modeling the structure of c-Myb R2R3 in complex with a GT half-site variant revealed specific interactions with G5. When second half-site variants were tested in vivo using a sensitive yeast effector-reporter system, both the TG and GT half-site variants were functional mediating c-Myb-dependent transactivation. Unexpectedly, we observed large differences between the best second half-site variants at low levels of c-Myb-effector, the GG variant being five- to fifteen-fold more active in vivo than the single-G half-sites, the GH or HG variants.