Finding new human minisatellite sequences in the vicinity of long CA-rich sequences

Microsatellites and minisatellites are two classes of tandem repeat sequences differing in their size, mutation processes, and chromosomal distribution. The boundary between the two classes is not defined. We have developed a convenient, hybridization-based human library screening procedure able to detect long CA-rich sequences. Analysis of cosmid clones derived from a chromosome 1 library show that cross-hybridizing sequences tested are imperfect CA-rich sequences, some of them showing a minisatellite organization. All but one of the 13 positive chromosome 1 clones studied are localized in chromosomal bands to which minisatellites have previously been assigned, such as the 1pter cluster. To test the applicability of the procedure to minisatellite detection on a larger scale, we then used a large-insert whole-genome PAC library. Altogether, 22 new minisatellites have been identified in positive PAC and cosmid clones and 20 of them are telomeric. Among the 42 positive PAC clones localized within the human genome by FISH and/or linkage analysis, 25 (60%) are assigned to a terminal band of the karyotype, 4 (9%) are juxtacentromeric, and 13 (31%) are interstitial. The localization of at least two of the interstitial PAC clones corresponds to previously characterized minisatellite-containing regions and/or ancestrally telomeric bands, in agreement with this minisatellite-like distribution. The data obtained are in close agreement with the parallel investigation of human genome sequence data and suggest that long human (CA)s are imperfect CA repeats belonging to the minisatellite class of sequences. This approach provides a new tool to efficiently target genomic clones originating from subtelomeric domains, from which minisatellite sequences can readily be obtained. [The sequence data described in this paper have been submitted to the EMBL data library under accession nos. AJ000377-AJ000383.]