Cloning of novel PAX5 fusion partners by RCA-RACE. (A) RCA-RACE products were run on 1% agarose gel. Bands corresponding to amplified transcripts of fusion (i) or truncated genes (ii) are indicated by arrowheads. Additional bands correspond to amplification of fragments of wild-type PAX5 DNA. (i) Patient #101 amplified product corresponds to PAX5-NCoR1 fusion cDNA, #96 to PAX5-DACH2 fusion, #103 to PAX5-GOLGA6 fusion cDNA. (ii) Patient #108 product corresponds to truncated PAX5 cDNA, #102 to PAX5-TAOK1 fusion. (B) Schematic illustration of the fusion proteins predicted from cDNA sequencing, breakpoints (BP) are indicated by dashed lines. Ex1-10 indicates PAX5 exons. All fusion proteins retain the PAX5 DNA-binding domain (PBD, amino acids 16-142), the OP motif (amino acids 179-186), and NLS (amino acids 195-201). (1) PAX5-DACH2 contains the conserved coiled-coil domain of DACH2 (DACH-box C). (2) PAX5-GOLGA6 sequence contains a large part of a predicted coiled-coil domain of GOLGA6. (3) PAX5-NCoR1 retains part of the inhibitor of DNA-binding 2 (ID2) domain and the corepressor-nuclear receptor box 3 (CoRNR box 3). (4) PAX5-TAOK1 predictive protein contains an alternative amino acid sequence encoded by exon1B of PAX5 and an additional 17-amino acid tail, which does not correspond to any predictive functional domain. Truncated PAX5 contains N-terminal part of PAX5 (amino acids 1-201) and a 55-amino acid tail coded by the contiguous intron 5/intron 6 nucleotide sequence, which does not correspond to any predictive functional domain.