Segments of the POU domain influence one another''s DNA-binding specificity.

The ubiquitously expressed mammalian POU-domain protein Oct-1 specifically recognizes two classes of cis-acting regulatory elements that bear little sequence similarity, the octamer motif ATGCAAAT and the TAATGARAT motif. The related pituitary-specific POU protein Pit-1 also recognizes these two motifs but, unlike Oct-1, binds preferentially to the TAATGARAT motif. Yet in our assay, Pit-1 still binds octamer elements better than does the octamer motif-binding protein Oct-3. The POU domain is responsible for recognizing these diverse regulatory sequences through multiple DNA contacts that include the two POU subdomains, the POU-specific region, and the POU homeodomain. The DNA-binding properties of 10 chimeric POU domains, in which different POU-domain segments are derived from either Oct-1 or Pit-1, reveal a high degree of structural plasticity; these hybrid proteins all bind DNA well and frequently bind particular sites better than does either of the parental POU domains. In these chimeric POU domains, the POU-specific A and B boxes and the hypervariable POU linker can influence DNA-binding specificity. The surprising result is that the influence a particular segment has on DNA-binding specificity can be greatly affected by the origin of other segments of the POU domain and the sequence of the binding site. Thus, the broad but selective DNA-binding specificity of Oct-1 is conferred both by multiple DNA contacts and by dynamic interactions within the DNA-bound POU domain.     

Purification and characterization of Ku-2, an octamer-binding protein related to the autoantigen Ku

The octamer motif (ATTTGCAT) is an important regulatory element in eukaryotic gene expression. A previously unidentified protein that recognizes this motif has been isolated from the human B cell line, Daudi. The protein, which we term Ku-2, bears a close resemblance to the DNA-binding autoantigen Ku. Like Ku, it is a heterodimer with subunits of 83 and 72 kDa; antisera raised against either subunit of Ku cross-react with Ku-2. Two peptides have been sequenced and show a strong similarity to regions in the corresponding subunits of Ku. The sequences are not identical, however, suggesting that Ku-2 may be a B cell homologue of Ku. Both Ku and Ku-2 bind to the termini of DNA duplexes, but Ku-2 also binds to an internal octamer motif. It is not known whether Ku shares the latter property or whether the octamer binding is a consequence of sequence differences between the two proteins. Ku-2 does not react with antisera against the POU domain of the octamer-binding protein Oct-2, indicating that the DNA binding domains of the two proteins are dissimilar despite the ability of both to bind to the octamer motif. We discuss the evidence for the existence of a family of octamer-binding proteins related to Ku.     

Mechanisms for flexibility in DNA sequence recognition and VP16-induced complex formation by the Oct-1 POU domain.

DNA binding by the Oct-1 protein is directed by its POU domain, a bipartite DNA-binding domain made up of a POU-specific (POUS) domain and a POU-homeo (POUH) domain, two helix-turn-helix-containing DNA-binding modules that cooperate in DNA recognition. Although the best-characterized DNA target for Oct-1 binding is the octamer sequence ATGCAAAT, Oct-1 also binds a number of different DNA sequence elements. For example, Oct-1 recognizes a form of the herpes simplex virus VP16-responsive TAATGARAT element, called the (OCTA-)TAATGARAT site, that lacks octamer site similarity. Our studies suggest two mechanisms by which Oct-1 achieves flexible DNA sequence recognition. First, an important arginine found in the Oct-1 POUS domain tolerates substitutions of its base contacts within the octamer site. Second, on the (OCTA-)TAATGARAT site, the POUS domain is located on the side of the POUH domain opposite from where it is located on an octamer site. This flexibility of the Oct-1 POU domain in DNA binding also has an impact on its participation in a multiprotein-DNA complex with VP16. We show that Oct-1 POUS domain residues that contact DNA have different effects on VP16-induced complex formation depending on whether the VP16-responsive element involved has overlapping octamer similarity or not.     

Photoaffinity electrophoretic mobility shift assay using photoreactive DNA bearing 3-trifluoromethyl-3-phenyldiazirine in its phosphate backbone

Photoaffinity cross-linking enables the analysis of interactions between DNA and proteins even under denaturing conditions. We present a photoaffinity electrophoretic mobility shift assay (EMSA) in which two heterogeneous techniques―photoaffinity cross-linking using DNA bearing 3-trifluoromethyl-3-phenyldiazirine and sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE) analysis—are combined. To prepare the photoreactive DNA, which is an essential tool for photoaffinity EMSA, we first determined the optimal conditions for the integration of 4-(3-trifluoromethyl-3H-diazirin-3-yl)benzyl bromide to the specific site of oligonucleotide where phosphodiester linkage was replaced with phosphorothioate linkage. The photoaffinity EMSA was developed using the POU (initial letters of three genes: Pit-l, Oct-1,2, and unc-86) domain region of Oct-1 protein, which specifically bound to octamer DNA motif (ATGCAAAT). The affinity-purified recombinant POU domain proteins conjugated with glutathione-S-transferase (GST) contained three distinct proteins with molecular weights of 34, 36, and 45 kDa. The photoaffinity EMSA could clearly distinguish the individual binding abilities of three proteins on a single lane and showed that the whole POU domain protein specifically bound to octamer DNA motif by competition experiments. Using the nuclear extract of HeLa cells, the photoaffinity EMSA revealed that at least five specific proteins could bind to the octamer DNA motif. These results show that photoaffinity EMSA using 3-trifluoromethyl-3-phenyldiazirine can provide high-performance analysis of DNA-binding proteins.