Prader-Willi syndrome (PWS) and Angelman syndrome (AS) are distinct neurobehavioral disorders that most often arise from a 4-Mb deletion of chromosome 15q11-q13 during paternal or maternal gametogenesis, respectively. At a de novo frequency of approximately.67-1/10,000 births, these deletions represent a common structural chromosome change in the human genome. To elucidate the mechanism underlying these events, we characterized the regions that contain two proximal breakpoint clusters and a distal cluster. Novel DNA sequences potentially associated with the breakpoints were positionally cloned from YACs within or near these regions. Analyses of rodent-human somatic-cell hybrids, YAC contigs, and FISH of normal or rearranged chromosomes 15 identified duplicated sequences (the END repeats) at or near the breakpoints. The END-repeat units are derived from large genomic duplications of a novel gene (HERC2), many copies of which are transcriptionally active in germline tissues. One of five PWS/AS patients analyzed to date has an identifiable, rearranged HERC2 transcript derived from the deletion event. We postulate that the END repeats flanking 15q11-q13 mediate homologous recombination resulting in deletion. Furthermore, we propose that active transcription of these repeats in male and female germ cells may facilitate the homologous recombination process.

Recent studies have implicated alpha-satellite DNA as an integral part of the centromere, important for the normal segregation of human chromosomes. To explore the relationship between the normal functioning centromere and alpha-satellite DNA, we have studied eight accessory marker chromosomes in which fluorescence in-situ hybridization could detect neither pancentromeric nor chromosome-specific alpha-satellite DNA. These accessory marker chromosomes were present in the majority of or all cells analyzed and appeared mitotically stable, thereby indicating the presence of a functional centromere. FISH analysis with both chromosome-specific libraries and single-copy YACs, together with microsatellite DNA studies, allowed unequivocal identification of both the origin and structure of these chromosomes. All but one of the marker chromosomes were linear mirror image duplications, and they were present along with either two additional normal chromosomes or with one normal and one deleted chromosome. Indirect immunofluorescence analysis revealed that the centromere protein CENP-B was not present on these markers; however, both CENP-C and CENP-E were present at a position defining a 'neo-centromere'. These studies provide insight into a newly defined class of marker chromosomes that lack detectable alpha-satellite DNA. At least for such marker chromosomes, alpha-satellite DNA at levels detectable by FISH appears unnecessary for chromosome segregation or for the association of CENP-C and CENP-E at a functional centromere.