Characterization of a double homeodomain protein (DUX1) encoded by a cDNA homologous to 3.3 kb dispersed repeated elements

Target genes for the helicase-like transcription factor (HLTF), a member of the SNF/SWI family, were immunoprecipitated from HeLa chromatin fragments with an anti-HLTF antibody. A 182 bp fragment (HEFT1) presented 87% sequence identity with 3.3 kb dispersed repeats from the 4q35 D4Z4 locus linked to facioscapulohumeral muscular dystrophy (FSHD). The HEFT1 loci were, however, not genetically linked to FSHD. Transfection and in vitro binding studies identified within HEFT1 a promoter whose basal activity required a GC box activated by Spl or Sp3. A 4.4 kb homologous transcript was found mostly in human skeletal muscle and heart. A 1.2 kb cDNA fragment was cloned that encoded a 170 amino acid protein (DUX1) with two paired-type homeodomains. In vitro translated DUX1 specifically interacted in electrophoretic mobility shift assay (EMSA) with a P5 oligonucleotide (5'-GATCTGAGTCTAATTGAGAATTACTGTAC-3'). DUX1 co-expression activated up to 5-fold transient expression in insect cells of a minimal promoter-luciferase construct fused to P5. The presence of 20 kDa DUX1 in vivo in rhabdomyosarcoma TE671 cell extracts was shown by western blotting with a rabbit antiserum raised against a DUX1 peptide, This antiserum suppressed a TE671 protein-P5 complex in EWSA with identical migration as the in vitro translated DUX1-P5 complex. Genomic PCR experiments could not identify a gene fragment linking the HEFT1 and DUX1 sequences, which present one mismatch in their overlapping region. However, a similar gene was found in another 3.3 kb element comprising the HEFT1 promoter and a DUX1-like open reading frame. In addition, homologous gene sequences were identified in 3.3 kb elements of the D4Z4/FSHD locus, considered until now 'junk' DNA.