Isolation and characterization of a human DNA binding protein which contains an ANTP homeobox sequence.

Abstract

Protein factors have been identified that bind DNA and regulate mammalian transcription. Several of these factors bind at, or around, the TATA box. To isolate additional TATA-recognizing factors, we screened a human B-cell $\lambda$gt11 expression library with a radiolabeled TATAA repeated probe. One recombinant clone bound specifically to the TATAA repeated sequence. The binding clone contained a class I homeobox that was identical in 59/60 residues to the Drosophila Antp homeobox. The homeobox cDNA that was isolated from the B-cell library also recognized a 1.6 kB message in T-cells. This is the first example of an Antp homeobox protein in lymphoid cells. The specificity of the homeobox protein for the TATA box was investigated. Single base mutations in the TATAA sequence determined that the last three bases (TAA) were essential for homeobox binding. The first base (T) was important, but not critical, for binding. Although the human Antp homeobox protein demonstrated affinity for the TATAA sequence, it does not appear to be the mammalian TATA-binding factor TFIID. A polyclonal antibody to denatured human Antp protein did not recognize native TFIID. Furthermore, human Antp is approximately 29 kD in size, whereas, mammalian TFIID is 120-140 kD. Cloned yeast TFIID, which is functionally interchangable with mammalian TFIID, is 27-29 kD but contains no sequence homology to Antp. Whether the homeobox protein plays any role with the TATA box in vivo remains to be determined. The role of the pentapeptide in homeotic gene function was investigated in the human Antp protein. Site directed mutagenesis experiments demonstrated that four of the five residues in the pentapeptide are not required for the homeodomain to bind to DNA. Mutation of one residue (proline to glycine) did drastically reduce the ability of the protein to bind DNA. This residue may not interact with DNA directly, but rather determine the conformation of the protein such that the "recognition helix" in the homeodomain is able to interact with DNA.