A novel strategy has been used to isolate a cDNA clone that encodes a DNA binding domain whose recognition properties overlap those of the mammalian transcription factors H2TF1 and NF-kappa B. These two factors are distinguished by their cell type distributions and their relative affinities for related sequence elements in the enhancers of the major histocompatibility complex (MHC) class I and immunoglobulin kappa chain genes. The human cDNA clone was detected by screening a lambda phage expression library with a binding site probe derived from the MHC enhancer. The phage encoded fusion protein binds specifically to both the MHC and kappa gene enhancers. The cDNA hybridizes to a single copy gene that is expressed as a 10 kb mRNA in both B and non-B cells. The strategy used in this study may prove generally useful in the cloning and analysis of sequence-specific DNA binding proteins.

Intergenic regions of polycistronic pre-mRNAs of trypanosomatid protozoans are the sites of two processing reactions: polyadenylation of the upstream gene and trans-splicing of the capped miniexon to the downstream gene. Their close proximity and the lack of consensus motifs at poly(A) sites led us to test whether poly(A) site selection is governed by the location of the downstream splice acceptor in the DHFR-TS locus of Leishmania major. Whenever the position of the downstream splice site was altered, the poly(A) site was shifted 400-500 nucleotides upstream of the new splice site. In contrast, when the wild-type poly(A) site was eliminated, the downstream splice site was unaffected, and polyadenylation was maintained 200-500 nucleotides upstream of the splice site. In a second set of experiments, T7 RNA polymerase expressed in Leishmania was used to direct the synthesis of artificial pre-RNAs in vivo whose expression was found to require the presence of a downstream splice acceptor. We conclude that poly(A) site selection in Leishmania is specified by the position of the downstream splice acceptor and propose a scanning model for poly(A) site selection after splice site recognition.

Transcription of promoters of immunoglobulin genes is controlled by an octanucleotide sequence element. The sequence of a eDNA encoding a B-cell-specific protein, Oct-2, has been determined. This protein specifically recognizes the octanucleotide element and is part of the previously identified NF-A2 family of proteins. The DNA-binding domain of Oct-2 is structurally related to the homeo box consensus and thus contains a potential helix-turn-helix sequence. Oct-2 also possesses a potential 'leucine zipper' domain, where four leucines are each separated by exactly seven residues. Comparisons of Oct-2 with protein Oct-l, which also recognizes the octanucleotide element but is constitutively expressed in all cell types, show high sequence conservation through the 60-residue DNA-binding domain, as well as an adjacent tract of 75 residues. The latter conserved region is also found in regulatory genes expressed in pituitary cells and nematodes and has been termed a POU box. Because two different cDNAs were isolated, it is proposed that the oct-2 gene is expressed as multiple mRNAs that vary in splicing patterns. Most interestingly, the oct-2 eDNA contains a second overlapping open reading frame, 278 residues in length, which might also specify a protein important for B-cell development.[Key Words: B-cell-restricted octanucleotide factorl Oct-2 protein sequencel helix-turn-helix DNA-binding domain~ leucine zipper domain~ POU domain]

The herpes simplex virus transactivator, alpha TIF, stimulates transcription of the alpha/immediate early genes via a cis‐acting site containing an octamer element and a conserved flanking sequence. The alpha TIF protein, produced in a baculovirus expression system, nucleates the formation of at least two DNA‐‐protein complexes on this regulatory element. Both of these complexes contain the ubiquitous Oct‐1 protein, whose POU domain alone is sufficient to allow assembly of the alpha TIF‐dependent complexes. A second member of the POU domain family, the lymphoid specific Oct‐2 protein, can also be assembled into similar complexes at high concentrations of alpha TIF protein. These complexes contain at least two cellular proteins in addition to Oct‐1. One of these proteins is present in both insect and HeLa cells and probably recognizes sequences in the cis element. The second cellular protein, only present in HeLa cells, probably binds by protein‐protein interactions.