Molecular cloning of a translocation breakpoint hotspot in 22q11

Abstract

It has been well documented that 22q11 contains one of the most rearrangement-prone sites in the human genome, where the breakpoints of a number of constitutional translocations cluster. This breakage-sensitive region is located within one of the remaining unclonable gaps from the human genome project, suggestive of a specific sequence recalcitrant to cloning. In this study, we cloned a part of this gap and identified a novel 595-bp palindromic AT-rich repeat (PATRR). To date we have identified three translocation-associated PATRRs. They have common characteristics: (1) they are AT-rich nearly perfect palindromes, which are several hundred base pairs in length; (2) they possess non-AT-rich regions at both ends of the PATRR; (3) they display another nearby AT-rich region on one side of the PATRR. All of these features imply a potential for DNA secondary structure. Sequence analysis of unrelated individuals indicates no major size polymorphism, but shows minor nucleotide polymorphisms among individuals and cis-morphisms between the proximal and distal arms. Breakpoint analysis of various translocations indicates that double-strand-breakage (DSB) occurs at the center of the palindrome, often accompanied by a small symmetric deletion at the center. The breakpoints share only a small number of identical nucleotides between partner chromosomes. Taken together, these features imply that the DSBs are repaired through nonhomologous end joining or single-strand annealing rather than a homologous recombination pathway. All of these results support a previously proposed paradigm that unusual DNA secondary structure plays a role in the mechanism by which palindrome-mediated translocations occur.

Figure 1.

Development of a PCR system specific for the PATRR22 region. (A) LCR22 in 22q11.2. There are at least four copies of LCR22, LCR22A, LCR22B, LCR22C, and LCR22D. One of the LCR22s, LCR22B, is the site of a genomic contig gap as well as the translocation breakpoint hotspot. (B) Location and orientation of the four PCR primers on the known structure appearing in other LCR22s and the t(11;22) translocation derivative chromosomes. There are multiple copies of primer “a” as a result of tandem repeats. Black lines indicate chromosome 22, while light-gray lines indicate chromosome 11. All of the genomic clones present in the database manifest the “known structure.” Only a der(22)-containing clone should be positive for the PCRab products. “(AT)n”s indicate the AT-rich region. Gray arrows indicate the putative PATRR22 arms. (C) Representative results of the PCRab reaction on genomic DNA. (Top) PCRab for eight normal individuals; (bottom) PCRab for patients with the 22q11.2 deletion syndrome. The additional bands originating from heteroduplexes appear as PCR artifacts in lanes 3 and 7. (D) The PCRbc (left) and PCRab (right) for somatic cell hybrids. The positive PCRbc product arises as a result of the presence of any of the LCR22s in the hybrids, while the positive PCRab product is derived only from LCR22B. (Lane 1) Normal human; (lane 2) GM10888, a human/hamster hybrid containing a normal chromosome 22 as its only intact human chromosome; (lane 3) Cl6-2/EG, a human/hamster hybrid containing a del(22q11.2) chromosome; (lane 4) Cl-4/GB, a human/hamster hybrid cell line containing the der(22) of a constitutional t(11;22); (lane 5) GM11685 (NF13/D3), a subclone of a human/mouse hybrid cell line containing the der(22) of a constitutional t(17;22). Black solid lines in the diagram below represent the chromosome 22q11 regions present in the hybrid lines analyzed.

Figure 2.

Molecular cloning and characterization of the PATRR22. (A) PCRad for four individuals. PCR products appearing as prominent bands did not contain the PATRR22. The DNA fraction indicated by the thick bracket contains the authentic PATRR22. Bands indicated by a thin bracket with an asterisk are derived from LCR22s with the standard structure. (B) Structure of the PATRR22. Hatched arrows indicate each unit of inverted repeats. Small arrows depict PCR primers. (C) Characteristics of the three translocation-related PATRRs. Arrows indicate each unit of inverted repeats of the PATRR. Black boxes indicate relatively non-AT-rich regions, while stippled boxes indicate the AT-rich regions flanking the PATRRs.

Figure 3.

Localization of the translocation breakpoints within the PATRR22. (A) Mapping of the t(11;22) breakpoints on the PATRR22. Each circle represents the breakpoints of the der(22) (top) and the der(11) (bottom). The scale indicates nucleotide number from the center of the PATRR22. The numbers in the circles indicate each individual translocation carrier. Black circles are from Kurahashi et al. (2000b), red from Edelmann et al. (2001), and blue from Tapia-Paez et al. (2001). Samples 38, 39, 43 (black), and 14 (red) have only the der(22) breakpoint, since they are from patients with the supernumerary-der(22) syndrome. It is noteworthy that the majority of balanced carriers show similar breakpoint locations on the PATR22 for the der(11) and the der(22). (B) Mapping the breakpoints of other translocations involving the 22q11 PATRR.

Figure 4.

Analysis of microhomologies at the breakpoints. (A) The number of identical nucleotides at the breakpoints of the t(11;22). The dotted line indicates the putative sample numbers from simulation of NHEJ. (B) A random simulation of the translocation. The details of the procedure are given in the Methods. The scale of the vertical axis is adjusted to reflect sample number relative to that shown in the graph (A). (C) Detail of one junction showing nine identical nucleotides (Case 3). The top indicates the junction fragment, while the middle and bottom indicate original breakpoint sequences. The arrows indicate each unit of the inverted repeats, while the arrowheads indicate the center of the PATRR.

Figure 5.

Structure of the region surrounding the PATRR22. (A) Predicted structure of LCR22B. There is still a contig gap in LCR22B; the end of the proximal contig is the telomeric end of cHK89, whereas the end of the more distal contig is the centromeric end of BAC RP11-562f10. Ensembl indicates BAC RP11-694E12 as a bridging clone possibly using partial sequence data. Two 45-kb modules, major components of LCR22, are indicated by hatched boxes. Both 45-kb modules contain segments previously designated “NF1L” and “VNTR-L” in the same orientation (Shaikh et al. 2000). Another 45-kb module could potentially reside in the middle. (B) The region surrounding PATRR22. AC288053 and AC074203 are derived from the der(11), whereas AC087065 is derived from the der(22). The bidirectional arrow indicates the longest PCR product identified in this study. Thick dotted arrows indicate the AT-rich region-HSAT1-Alu cassette.