DEK, an autoantigen involved in a chromosomal translocation in acute myelogenous leukemia, binds to the HIV-2 enhancer

Abstract

The product of the dek oncogene is the 43-kDa DEK nuclear protein. DEK was first identified in a fusion with the CAN nucleoporin protein in a specific subtype of acute myelogenous leukemia. DEK has also been shown to be an autoantigen in patients with pauciarticular onset juvenile rheumatoid arthritis. Further, the last 65 amino acids of DEK can partially reverse the mutation-prone phenotype of cells from patients with ataxia-telangiectasia. However, in spite of these significant disease associations, the function of DEK has remained unclear. The HIV-2 peri-ets (pets) site is a TG-rich element found between the two Elf-1 binding sites in the HIV-2 enhancer. The pets element mediates transcriptional activation whether the enhancer is stimulated by phorbol 12-myristate 13-acetate (PMA) alone, phytohemagluttinin (PHA) alone, PMA plus PHA, soluble antibodies to the T cell receptor, immobilized antibodies to the T cell receptor, or by antigen. Previously, we purified and characterized the pets factor, demonstrating that it is a 43-kDa nuclear protein. We now describe the identification of DEK as this 43-kDa pets factor. Using a modified Southwestern screening procedure, we find that DEK can recognize the pets element. We demonstrate the ability of recombinant DEK to bind specifically to the pets site using the electrophoretic mobility shift assay (EMSA) and DNase I footprinting. "Supershift" EMSA further confirms that DEK is the dominant protein binding to the pets site in T cell extracts. Our findings show that DEK is a site-specific DNA binding protein that is likely involved in transcriptional regulation and signal transduction. This has implications for multiple pathogenic processes, including hematologic malignancies, arthritis, ataxia-telangiectasia, and AIDS caused by HIV-2.

Figure 1

A nuclear factor(s) binds specifically to the HIV-2 pets site. EMSA using Jurkat nuclear extracts with the HIV-2 pets site probe. The shifted band seen in lane 1 (indicated by the arrow on the left) can be competed away by the addition of unlabeled self-competitor (lanes 3 and 4), but not by a mutant version of the pets site oligonucleotide (lanes 6 and 7) or by unrelated HIV enhancer elements (lanes 7–13).

Figure 2

Purified recombinant DEK–(HIS)₆ binds to the HIV-2 pets site in EMSA (lane 1). Addition of cold unlabeled pets oligonucleotide competed away the shifted band indicated by the arrow (lanes 3 and 4) but not when mutant or unrelated binding site oligonucleotides were added (lanes 5–12). The mutant oligonucleotide (mut-2) added in lanes 8–10, unlike mut-1 (lanes 5–7), did not have the downstream Elf-1 binding site mutated. In lane 13, human anti-DEK serum has been added to the reaction, leading to a supershift.

Figure 3

Identification of the nuclear factor binding to the HIV-2 pets site using supershift EMSA. The arrow on the left indicates a band shift seen in EMSA using the pets probe with Jurkat nuclear extract (lane 1). Addition of antibody recognizing DEK supershifted this complex (lanes 4 and 5) as indicated by the arrow to the right. Addition of control antibodies to NF-κB p50 or to rabbit immunoglobulin does not supershift the complex (lanes 2 and 3).

Figure 4

Binding of recombinant DEK to the HIV-2 pets site can be demonstrated using DNase I footprinting. No protection was seen when bovine serum albumin (BSA) was added to the probe (lane 2) or when control E. coli extract was used (lane 6). Addition of increasing amounts of E. coli extract containing GST–DEK (lanes 3–5) protects the TATACTTGG bases of the HIV-2 enhancer as indicated by the bracket on the right. A G+A ladder (lane 1) identifies the position of the footprint as the pets site.