SEARCH

SEARCH BY CITATION

Keywords:

  • fragile site;
  • FRA6H;
  • FRA13E;
  • genomic instability;
  • chromosome rearrangements

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Fragile sites are specific genomic loci that are especially prone to chromosome breakage. For the human genome there are 31 rare fragile sites and 88 common fragile sites listed in the National Center for Biotechnology Information database; however, the exact number remains unknown. In this study, unstable DNA sequences, which have been previously tagged with a marker gene, were cloned and provided starting points for the characterization of two aphidicolin inducible common fragile sites. Mapping of these unstable regions with six-color fluorescence in situ hybridization revealed two new fragile sites at 6p21 and 13q22, which encompass genomic regions of 9.3 and 3.1 Mb, respectively. According to the fragile site nomenclature they were consequently entitled as FRA6H and FRA13E. Both identified regions are known to be associated with recurrent aberrations in malignant and nonmalignant disorders. It is conceivable that these fragile sites result in genetic damage that might contribute to cancer phenotypes such as osteosarcoma, breast and prostate cancer. © 2007 Wiley-Liss, Inc.

Fragile sites are specific genomic loci that are especially prone to express genomic instability. They can be visualized as gaps and breaks on metaphase chromosomes after culturing cells under conditions of replication stress. Based on their incidence in the human population, they are divided into rare fragile sites, occurring in less than 5% of all individuals, and common fragile sites being a constitutive feature of the genome of probably all individuals.1 According to the National Center for Biotechnology Information (NCBI), there are 31 rare fragile sites and 88 common fragile sites in the human genome. However, the exact number of fragile sites remains unclear, since there are no stringent criteria for inclusion and there is no regular update of the database.2

The molecular basis for the expression of rare fragile sites is the dynamic mutation of expanding CGG trinucleotide or AT-rich minisatellite sequences that after reaching a certain threshold expansion account for the observed instability.3, 4, 5, 6 In contrast, the molecular basis for breakage of common fragile sites is still unclear. To date, there are 18 aphidicolin inducible common fragile sites characterized: FRA2G, FRA3B, FRA4F, FRA6E, FRA6F, FRA7E, FRA7G, FRA7H, FRA7I, FRA8C, FRA9E, FRA16D, FRAXB (reviewed by Ref.7), FRA1E,8FRA7K,9FRA11E,10FRA11G11 and FRA13A.12 However, analysis of the identified sequences did not reveal any particular sequence structure; an AT-richness and an enrichment of DNA flexibility structures seem to be the only shared features.13, 14 It has been hypothesized that the AT-richness leads to an accumulation of DNA secondary structures, which might cause a delayed replication at fragile sites.15, 16, 17, 18 This perturbed replication at fragile sites was shown to activate cell cycle checkpoints in an ATR-dependent manner.19 In line with this, several targets and modifiers of the ATR-pathway, such as BRCA1, SMC1, CHK1 and the Fanconi anemia pathway proteins, were reported to be involved in maintenance of fragile sites stability.20, 21, 22, 23 The replication perturbation may result in double strand breaks (DSB), which in turn will be subjected to repair by either homologous recombination or by nonhomologous end-joining pathways.24 In the rare case in which the cell manages to circumvent the cell cycle checkpoints and repair systems, the DSBs remain as unprotected DNA ends.

The consequences of genomic instability at fragile sites can be seen at an increased rate of sister chromatid exchanges, deletions, translocation breakpoints and amplifications of genes via breakage-fusion-bridge mechanism with the boundaries of the amplicons being situated in fragile sites.2, 25 Another traceable sign of the instability is the preferential integration of exogenous DNA, vector DNA in vitro as well as viral DNA in vivo, within fragile sites.10, 26, 27, 28

The expression of fragile sites is rarely seen as visible gaps or breaks on metaphase chromosomes without previous induction, and only few studies have reported a spontaneous expression of fragile sites.29, 30, 31 Recently, it has been shown that the expression of fragile sites can be induced by an siRNA induced knock down of genes involved in check point control19, 23, 32 and repair.24 In the breast cancer cell line MDA-MB-436, BRCA1 is heterozygously deleted in combination with a pathogenic mutation of the remaining allele.33 This mutation led to an aberrant cytoplasmic localization of the expressed mutant BRCA1 proteins33 that in turn might cause the spontaneous expression of gaps and breaks at fragile sites,34 which were observed in MDA-MB-436 to colocalize with common fragile sites.35

The combination of the preferential integration of exogenous DNA into fragile sites and the spontaneous expression of fragile sites in the breast cancer cell line MDA-MB-436 was the basis to genetically tag a substantial number of fragile sites with a marker gene, pSV2neo, by insertional mutagenesis.35 In total, 112 different pSV2neo integration sites were detected by fluorescence in situ hybridization (FISH), of which 44% coincided with already described common fragile sites.35 Here, we report the cloning of two pSV2neo integration sites at 6p21 and 13q22. We show that the integration sites are situated within or less than 1 Mb apart of aphidicolin inducible common fragile sites. The precise genomic localization of the two new fragile sites FRA6H and FRA13E was determined by six-color FISH on Epstein-Barr virus (EBV)-transformed lymphocytes and their sequences were analyzed for DNA flexibility.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Cell culture, fragile site induction and metaphase preparation

Two independent cell clones, previously named clone No. 21 and No. 22, of the pSV2neo transfected human breast cancer cell line MDA-MB-436 (ATCC number: HTB-130) were chosen for a detailed analysis of integration sites.35 Cells were cultured in RPMI 1640 supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 0.5 μg/ml amphotericin and 400 μg/ml G 418 sulfate.

Mapping of fragile sites were performed on EBV-transformed lymphocytes of two unrelated individuals. Cells were cultivated in RPMI 1640 supplemented with 12.5% FCS. Fragile sites were induced with 0.4 μM aphidicolin (Sigma, Munich, Germany) and 0.5% ethanol for 26 hr prior to metaphase preparation. For 6p21 breakage analysis the expression of fragile sites was partly enhanced with 5 μM camptothecin (LKT Laboratories, St. Paul, MN) instead of ethanol for 2 hr prior to metaphase preparation.36

Cells for metaphase preparation were subjected to a hypotonic treatment (0.55% KCl, 1% Na-Citrate, 1:1) and fixed in 3:1 methanol-acetic acid.

PCR analysis

DNA for PCR analysis was isolated by proteinase K digestion followed by phenol–chloroform extraction according to standard procedures.

Alu-PCR.

Alu-PCR was performed as described elsewhere.37 For the detection of pSV2neo-human junctions Alu-PCR was modified by the use of pSV2neo specific primers (Table I). The primers are specific for the 3′- or the 5′-end of the pSV2neo neomycin resistance gene, respectively, each having a second corresponding nested primer. The first round of PCR was performed with dUTP containing primers in 50 μl reactions: 100 ng genomic DNA, 100 pmol pSV2neo specific primer (Neo-3, Neo-5), 10 pmol Alu specific primer (Alu-A3, Alu-A5), 200 μM each dNTP and 1.5 U DNA polymerase mixture of Taq (Promega, Mannheim, Germany) and Pwo (Roche, Mannheim, Germany) in 1× PCR buffer (Pwo). Cycling conditions were: 1 min initial denaturation at 94°C, followed by 10 cycles of 30 sec denaturation at 94°C, 30 sec of annealing at 59°C and 3 min elongation at 72°C. One unit of Uracil-DNA Glycosylase (MBI Fermentas, St. Leon-Rot, Germany) was added to the reaction followed by an incubation for 30 min at 37°C. The subsequent incubation for 10 min at 95°C led to a breakdown of the dUTP containing DNA. This treatment in combination with a second round of PCR using Alu-Tag primers specific to the tag-sequence of the used Alu specific primers led to a negligible linear amplification of undesirable inter-Alu fragments. However, the specific target fragment is enriched via an exponential amplification. About 10 pmol of both the internal pSV2neo specific primer (Neo-3-ne, Neo-5-ne) and the Alu-Tag primer (Alu-Tag3, Alu-Tag5) were added for the second round of PCR with cycle conditions as follows: one initial denaturation of 1 min at 94°C, followed by 20 cycles of 30 sec denaturation at 94°C, 30 sec of annealing and 3 min elongation at 72°C. Annealing temperature was decreased by 1°C from 65°C every second cycle. Reaching 55°C, 20 cycles were added followed by an additional extension time of 10 min. All combinations between pSV2neo and Alu specific primers were used. To test for background amplification of inter-Alu products, PCRs including Alu specific primers only were conducted. PCR products were analyzed on 1% agarose gels.

Table I. Primer Sets Used for Alu-PCR
Primer1Sequence2
  • 1

    (1–4) Alu primers37, (5–8) pSV2neo primer, suffix “-ne” indicates nested primers.

  • 2

    Bold: added tag-sequences of Alu primers.

Alu-A55′-CAGUGCCAAGUGUUUGCUGACGCCAAAGUGCUGGGAUUACAG-3′
Alu-Tag55′-CAAGTGTTTGCTGACGCCAAAG-3′
Alu-A35′-AGUGCCAAGUGUUUGCUGACGACUGCACUCCAGCCUGGGCGAC-3′
Alu-Tag35′-CAAGTGTTTGCTGACGACTGCA-3′
Neo-35′-CGCAUCGAGCGAGCACGUACUCGGAUGG-3′
Neo-3-ne5′-TGAAGAGCTTGGCGGCGAATGGGCTGAC-3′
Neo-55′-AGUCCCUUCCCGCUUCAGUGCCAACGU-3′
Neo-5-ne5′-TGCCCAGTCATAGCCGAATAGCCTCTCCACC-3′
Standard PCR.

Standard PCR reactions were performed to complete the cloning of the pSV2neo integration site at 6p21 in clone No. 21, partially identified by Alu-PCR. pSV2neo specific primers were combined with different primers situated in the identified human sequence. After amplification of the missing human-vector junction (primers: AL356983-7980-for 5′-ACAAGGCAACACGAGGCAAGTG-3′; pSV2neo-3814-for 5′-CACAGAATCAGGGGATAACG-3′), the whole integration site was amplified and sequenced.

Sequencing

PCR fragments were extracted from an agarose gel (QIAquick Gel Extraction Kit, QIAGEN, Hilden, Germany) and sequenced by fluorescent dye terminator cycle sequencing (Big Dye Terminator v3.1 Cycle Sequencing Kit, Applied Biosystems, Darmstadt, Germany). Sequencing products were analyzed on an ABI PRISM 310 Genetic Analyser. Sequences were crossreferenced to sequence databases using the HUSAR program package (Heidelberg Unix Sequence Analysis Resources, Bioinformatics Lab, DKFZ, Heidelberg, Germany, http://genome.dkfz-heidelberg.de).

BAC, PAC and YAC clones

Bacterial artificial chromosome (BAC) clones (all RPCI-11 library) and P1 derived artificial chromosome (PAC) clones (RPCI-1 and RPCI-3 libraries, constructed at Roswell Park Cancer Institute by P.J. de Jong and P. Ioannou), were chosen according to their position in the Ensembl database at http://www.ensembl.org/index.html (Ensembl v39-Jun 2006).38 Yeast artificial chromosome (YAC) clones from the Human Mega YAC (CEPH) library were chosen according to their position in the database of the Whitehead Institute http://www-genome.wi.mit.edu/cgi-bin/contig/phys_map.39 BAC, PAC and YAC clones were obtained from German Resource Center for Genome Research, Berlin, Germany.

Fluorescence in situ hybridization

Six-color fluorescence in situ hybridization (FISH) was performed by using modified nucleotides, where either the haptens biotin (BIO) and dinitrophenol (DNP) (Molecular Probes, Eugene, OR), or the fluorescent dyes DEAC (Molecular Probes, Eugene, OR), Cy3, Cy3.5 and Cy5 (Amersham, Freiburg, Germany) were coupled to allylamine-dUTPs (Sigma, Deisenhofen, Germany) as described previously.40 YAC, BAC, PAC or plasmid DNA was labeled with BIO-, DNP-, DEAC-, Cy3-, Cy3.5- and Cy5-dUTPs by nick translation and was hybridized to metaphase chromosomes according to standard procedures. For the detection of BIO, avidin-FITC and biotinylated goat anti-avidin D (Vector Lab., Burlingame, CA) were used. The detection of DNP was performed with rat anti-DNP (Zymed, San Fransisco, CA) and Cy5.5 goat anti-rat (Rockland Immunochemicals, Gilbertsville, PA). Slides were counterstained with 4,6-diamidino-2-phenylindol (DAPI) (Sigma, Munich, Germany) and metaphases were analyzed with a Leica DMRA 2 microscope and Leica CW 4000 FISH software. The localization of gaps/breaks at 6p21 and 13q22 was determined relative to the hybridization signals of different DNA probes and was classified into: (i) distal (telomeric) to, (ii) within and (iii) proximal (centromeric) to the hybridization signal.

Databases, programs

The localization of BAC and PAC clones, as well as of markers and individual genes, was determined using the Ensembl database at http://www.ensembl.org/index.html (Ensembl v39-Jun 2006).38 The extraction of all genes with the status “known” within the FRA6H region was done with help of the Martview function at the Ensembl homepage.

The flexibility of DNA sequences was estimated with the TwistFlex program that calculates flexibility at the twist angle of the DNA. The calculations are made for overlapping windows along the sequence, for each window an average value of all twist angles between two consecutive dinucleotide steps was determined. The default settings were used. The size for the overlapping windows was 100 bp combined with a threshold for the calculated twist angle of >13.7° to be considered as a flexibility peak. Clusters of flexibility peaks are defined as at least three flexibility peaks with the distance between any two adjacent peaks being below or equal to 5 kb. As a comparative value the average number of clusters of unified peaks (c) per Mb (c/Mb) was calculated and compared with previously determined average values for G- and R-bands.14 Sequences for TwistFlex analysis were extracted from the Ensembl database using the “export data” function. The TwistFlex program can be downloaded at http://margalit.huji.ac.il/TwistFlex/.

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Cloning of pSV2neo integration sites

The breast cancer cell line MDA-MB-436 was previously transfected with the vector pSV2neo, and stably transfected cell clones were isolated.35 Two independent cell clones, previously named No. 21 and No. 22, were chosen to identify sequences flanking the integration site of the vector pSV2neo by Alu-PCR amplification. For both clones, the combination of the Neo-3′ primer set with the Alu-A5 primer set led to an amplification of discrete PCR fragments (Fig. 1a). PCR fragments were analyzed by sequencing. In clone No. 21, an integration of the vector pSV2neo was detected in the human sequence corresponding to PAC RP3-441G21 situated at chromosomal band 6p21.2. For clone No. 22, a pSV2neo integration was identified within sequences corresponding to the BAC RP11-267P7 at 13q22 (Fig. 1b).

thumbnail image

Figure 1. Cloning of pSV2neo integration sites. (a) Schematic representation of the loci with integrated pSV2neo in the vicinity of an Alu repeat. The basic principle of Alu-PCR amplification is exemplary depicted for primers Neo-3 and Alu-A5. An amplified DNA fragment contains specific human DNA flanked by pSV2neo and Alu sequences. (b) Sequenced PCR products identifying vector-human junctions in clones No. 21 and No. 22 are shown as bars. Numbers above indicate the exact bp position of sequences of pSV2neo (light grey). The identified human sequences flanking pSV2neo are displayed in dark grey, numbers below indicate the exact bp position of the sequence in the corresponding PAC RP3-441G21 (GeneBank accession no. AL356983) and BAC RP11-267P7 (GeneBank accession no. AL161717) for clone No. 21 and No. 22, respectively. An approximate position of the PAC/BAC is shown at chromosome ideograms. (c) Schema represents the PCR fragment obtained after amplification of the whole pSV2neo integration site at 6p21.2. Bar colored in light grey represents sequences corresponding to pSV2neo, bars in dark grey represent human sequences corresponding to sequences of PAC RP3-441G21 (GeneBank accession no. AL356983). Numbers above and below the bars represent the exact bp positions of the sequences in pSV2neo and PAC RP3-441G21. (d) FISH analysis of pSV2neo integration site at 13q22. Left: Partial metaphase with colocalization of the hybridization signals of BAC RP11-267P7 (red) with pSV2neo (green). Yellow signals result from an overlap of red and green signals. Right: Interphase nucleus showing the localization of pSV2neo signal (green) within the BAC signal (red). The splitting of the BAC signal into two unequal parts reflects the position of the pSV2neo integration site relative to the BAC RP11-267P7 sequence. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

To confirm the integration of the pSV2neo vector into 6p21.2, a DNA fragment of 3,611 bp including the pSV2neo vector was PCR amplified and sequenced. A total of 2,827 bp rearranged pSV2neo sequence replaced the genomic sequence represented by nucleotide positions 8,026–8,056 bp of the PAC RP3-441G21 (GeneBank accession no. AL356983), indicating a 30bp deletion of genomic DNA (Fig. 1c). The most convenient method to confirm pSV2neo integration sites is PCR-based, however, its limits are defined by the complexity of the target fragment. Intensive hybridization signals of pSV2neo probes in FISH analysis of clone No. 22, prompted us to assume that several copies of rearranged pSV2neo vector had integrated simultaneously at 13q22. To circumvent PCR problems due to complex target sequences, a FISH-based approach was chosen to confirm the integration site of the pSV2neo vector into 13q22. A colocalization of the hybridization signals of pSV2neo and the BAC RP11-267P7 was observed on metaphase chromosomes (Fig. 1d, left). In interphase nuclei, the pSV2neo signal was found to divide the BAC signal into unequal parts corresponding to the relative position of the integration site sequence in the BAC insert (Fig. 1d, right).

Determination of integration site relative to fragile sites

To investigate whether the observed pSV2neo vector integrations colocalize with regions of fragility, FISH was performed with probes PAC RP3-441G21 (6p21.2) and BAC RP11-267P7 (13q22) on EBV-transformed lymphocytes expressing aphidicolin-induced fragile sites. For both probes, gaps/breaks occurring both proximal and distal to the probe were recorded in different metaphases, indicating the presence of common fragile sites at these regions (Fig. 2).

thumbnail image

Figure 2. Localization of pSV2neo integration sites within regions of genomic instability on chromosomes with aphidicolin-induced fragile sites. Chromosomes are displayed as DAPI counterstained image (left) and as merged image with the hybridization signals (right). Chromosome 6 revealed breaks proximal (a) and distal (b) to the hybridization signal of the PAC RP3-441G21 (green), corresponding to the pSV2neo integration site at 6p21.2. Chromosome 13 showed breaks proximal (c) and distal (d) to the hybridization signal of the BAC RP11-267P7 (red) corresponding to the pSV2neo integration site at 13q22. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Mapping of a fragile site at 6p21

To further investigate the colocalization of the pSV2neo integration site at 6p21.2 with a common fragile site, FISH analysis was performed with DNA probes covering a 20 Mb region encompassing the pSV2neo integration site. Initially, seven BAC and PAC clones were chosen with distances of more than 1 Mb to each other to estimate the dimension of breakage at 6p21. The PAC RP3-441G21 was replaced by the BAC RP11-588I14 that covers the approximately same genomic region, however, having the sequence of pSV2neo integration situated in the centre. EBV-transformed lymphocytes were treated with aphidicolin, and gaps/breaks on metaphase chromosomes were scored in relationship to hybridization signals (Table II, Part A). Analysis of the 18 gaps/breaks observed at 6p21 showed that they were scattered along the initially analyzed 11 Mb genomic region. Since gaps/breaks could only be seen in ∼1% of all metaphases, further FISH analyses were performed on metaphases of EBV-transformed lymphocytes after a combined induction with aphidicolin and camptothecin. This combined induction led to an ∼10-fold increase of gaps/breaks at 6p21. To refine the topography of breakage, 17 additional DNA probes from 6p21.2–22 were used, and the localization of 116 gaps/breaks at 6p21 was determined in 862 metaphases analyzed (Table II, Part B). The region affected by breakage in 6p21 was identical in metaphases treated with aphidicolin alone and in metaphases treated with the combination of aphidicolin and camptothecin (Fig. 3b). This prompted us to assume that the addition of camptothecin is indeed enhancing the expression of aphidicolin inducible common fragile sites as was described previously.36 The graphical display of the percentage of breaks occurring distal to respective DNA probes (% distal) showed that the region affected of breakage at 6p21 is 9.3 Mb (Fig. 3b). The fragile region is defined by the proximal borders of PAC RP1-193B12 and BAC RP11-588I14 that spans the identified pSV2neo integration site at 37.2 Mb (pter = 0 Mb, according to Ensembl). Within this region there are 256 genes located, among them the genes of the human leucocyte antigen (HLA) locus (Fig. 3c).

thumbnail image

Figure 3. Map of 6p21.1-6p22 chromosomal region. (a) Genetic map of the 6p21.1-6p22 chromosomal region depicted with G-banding pattern (based on Ensembl database). (b) Percentage of gaps/breaks occurring distal to the hybridization signal of each BAC/PAC/YAC clone (rhombus) plotted against their genomic position in Mb from telomere (according to Ensembl). Black rhombi represent data from subset A (Table II, Part A); grey rhombi represent data from subset B (Table II, Part B). Dotted grey lines represent the BACs RP11-193B12 (28.0 Mb) and RP11-588I14 (37.2 Mb) that determine the borders of the fragile site region. Black dot at Mb scale represents the position of pSV2neo integration site at 37.2 Mb (according to Ensembl). (c) Approximate position of selected genes (20 out of 256) within the region of breakage. HLA class I, II and III gene regions are indicated with arrows.

Download figure to PowerPoint

Table II. Fish Analysis of Hybridization Signals on Chromosomes with Breaks in 6p21
 Clone id.MbTotal (T)Distal (D)1Within (W)2Proximal (P)3% Distal (%D)4
  • 1

    Number of gaps/breaks occurring distal to the hybridization signal.

  • 2

    Number of gaps/breaks occurring within the hybridization signal, resulting in a split signal.

  • 3

    Number of gaps/breaks occurring proximal to the hybridisation signal.

  • 4

    %Distal (%D = 100 × (D + 0.5W)/T). BAC, PAC and YAC clones are ordered from telomere (top) to centromere (bottom) according to their approximate Mb-position (according to Ensembl). (A) Fragile site induction with aphidicolin. (B) Fragile site induction with aphidicolin and camptothecin.

ARP1-139G21261801172.8
RP1-153G1427.51811168.3
RP11-234D1229.417311320.6
RP11-732O1931.31683559.4
RP11-600P333.418141380.6
RP11-174N335.217161097.1
RP11-588I1437.2181800100.0
BRP11-424H2321116001160.0
RP11-108N323102001020.0
RP1-139G2126102001020.0
RP1-45P2126.59710961.0
RP11-239L20279710961.0
RP1-153G1427.59720952.1
RP1-193B12289720952.1
RP11-519O1228.510270956.9
960-H-1029.510274918.8
RP11-192H1130981108711.2
RP11-23H0630.5941108311.7
RP11-331L1430.8981547917.3
RP11-732O1931.3681904927.9
RP11-666F0731.8583102753.4
RP11-598J932.2603302755.0
RP11-600P333.4593911966.9
RP11-107C834.649410883.7
RP11-388D1335.849410883.7
RP11-588I1437.11161131297.8
900-C-638.31161140298.3
963-F-2391161141198.7
873-B-3411161151099.6

Mapping of a fragile site at 13q22

To test whether the observed integration of pSV2neo vector in 13q22 occurred within a region of fragility in the human genome, FISH analyses were performed for this chromosomal band. Seven BAC clones (RP11-11C5, RP11-309J13, RP11-157E11, RP11-29G8, RP11-267P7, RP11-188A23 and RP11-338E20) being equally spread over the 6 Mb encompassing 13q22 were used to localize the region of genomic instability. Gaps/breaks at 13q22 were scored in relation to the hybridization signals of the BACs in metaphases of aphidicolin treated EBV-transformed lymphocytes (Table III). All gaps/breaks observed were localized within a 3.1 Mb region which was defined by the distal borders of the BACs RP11-11C5 and RP11-29G8. To refine the borders of the fragile region we conducted further FISH analyses with seven additional BACs (Table III). Gaps/breaks at 13q22 were observed distal or within BAC RP11-342J4 and proximal or within RP11-29G8, a result that corresponds to the previously performed rough mapping analysis. Thus, the fragile site at 13q22 spans 3.1 Mb, encompassing the genomic region from BAC RP11-342J4 to BAC RP11-29G8 that is situated 0.9 Mb proximal to the identified pSV2neo integration site at 76.2 Mb (pter = 0 Mb, according to Ensembl) (Fig. 4b). Altogether, 40 gaps/breaks at 13q22 were observed in 3,360 analyzed metaphases; an example of chromosome 13 with expression of the fragile site is shown in Figure 5a. According to the Ensemble database, eight genes are situated within the genomic region of the fragile site at 13q22 (Fig. 4c).

thumbnail image

Figure 4. Map of 13q21-13q22 chromosomal region. (a) Genetic map of the 13q21-13q22 chromosomal region depicted with G-banding pattern (based on Ensembl database). (b) Percentage of gaps/breaks occurring distal to the hybridization signal of each BAC clone (rhombus) (Table III) plotted against their genomic position in Mb from telomere (according to Ensembl). The region of instability was determined to a 3.1 Mb region encompassing BAC RP11-342J4 (72.2 Mb) and the BAC RP11-29G8 (75.2 Mb), showed by vertical grey dotted lines. Black dot at Mb scale represents the position of pSV2neo integration site at 76.2 Mb (according to Ensembl). (c) Localization of all genes within the region of fragile site.

Download figure to PowerPoint

thumbnail image

Figure 5. Fragile site at 13q22 (a) Mapping of fragile region at 13q22 with FISH. Break at 13q22 is displayed on DAPI counterstained chromosome 13 (left) and merged images of different BACs (from left to right): RP11-11C5 (green), RP11-505F3 (pink), RP11-309J13 (blue), RP11-157H4 (violet), RP11-132L12 (yellow), RP11-29G8 (red). Shown break localizes distal to RP11-11C5, RP11-505F3, RP11-309J13 and RP11-157H4 and proximal to RP11-132L12 and RP11-29G8. (b) Expression of four aphidicolin inducible common fragile sites on chromosome 13. FRA13A (13q12.2), FRA13C (13q21.2), FRA13D (13q32) and FRA13E at 13q22 (this study) are shown by arrows. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Table III. Fish Analysis of Hybridization Signals on Chromosomes with Breaks in 13q22
Clone id.MbTotal (T)Distal (D)1Within (W)2Proximal (P)3% Distal (%D)4
  • 1

    Number of gaps/breaks occurring distal to the hybridization signal.

  • 2

    Number of gaps/breaks occurring within the hybridization signal, resulting in a split signal.

  • 3

    Number of gaps/breaks occurring proximal to the hybridization signal.

  • 4

    % Distal (%D = 100 × (D + 0.5W)/T). BAC clones are ordered from centromere (top) to telomere (bottom) according to their approximate Mb-position (according to Ensembl).

RP11-11C572.1404000100.0
RP11-342J472.234322097.1
RP11-555G2272.328260292.9
RP11-505F372.529221677.6
RP11-309J1372.8322021065.6
RP11-157H473.423831241.3
RP11-459P2373.720401620.0
RP11-157E1174.230312611.7
RP11-132L1274.63130289.7
RP11-182M2074.93110303.2
RP11-29G875.24001391.3
RP11-267P776.24000400.0
RP11-188A2377.24000400.0
RP11-338E2077.94000400.0

DNA flexibility of fragile regions at 6p21 and 13q22

The majority of fragile sites characterized to date have been shown to be enriched for sequences with a high predicted DNA flexibility relative to nonfragile regions. To estimate the flexibility properties of the fragile DNA sequences identified here, the computer program TwistFlex was used.14

6p21.

For the fragile region at 6p21, being situated in an R-band, the predicted flexibility value was 0.5 c/Mb, which is in-between of the average values for nonfragile sequences (0.2 c/Mb) and fragile sequences (1.3 c/Mb) situated in R-bands.14 However, applying the TwistFlex program to the most frequently expressed fragile site FRA3B (region according to Ref.41) resulted in a flexibility value of 0.5 as well, indicating that according to the TwistFlex program both fragile sites have the same average DNA-flexibility. To estimate the DNA flexibility of sequences next to the fragile region at 6p21, flanking sequences, previously proven to be nonfragile by FISH analysis, were additionally analyzed with TwistFlex. The flexibility value for the 7 Mb flanking region distally was 1.3 c/Mb, the flexibility value for the 3.7 Mb flanking region proximally was 0.5 c/Mb. Since both sequences are located within G-bands, these values correspond to the average of 0.8 c/Mb for nonfragile sequences situated in G-bands (3.3 c/Mb for fragile sequences).14

13q22.

The flexibility value for the 3.1 Mb fragile region at 13q22 was 1.0 c/Mb. This is well in line with an average flexibility value of 1.3 c/Mb for fragile sequences of R-bands.14 As control, flanking sequences, similar in size to the fragile region (∼3 Mb) and proven to be nonfragile by FISH analysis, were also subjected to TwistFlex analysis. Flexibility values were 3.0 c/Mb for the distal flanking sequence (R-band) and 3.1 c/Mb for the proximal flanking sequence (G-band). These values correspond to the average of 1.3 and 3.3 c/Mb for fragile sequences situated in R- and G-bands, respectively.14

Collectively, our data show that the integration of pSV2neo vector occurred within or less than 1 Mb apart of the fragile site regions at 6p21 and 13q22. To determine whether the two identified regions of fragility correspond to the NCBI listed common fragile sites FRA6C (6p22.2) and FRA13C (13q21.2), respectively, we determined the cytogenetic localization of observed gaps/breaks on reversely stained DAPI metaphases. At 6p21 gaps/breaks were clustered within the defined region of the distal part at the light 6p21 G-band, in contrast to the predicted location of FRA6C at the dark 6p22 chromosomal band. Therefore, we suggest that the identified fragile region at 6p21 is a new aphidicolin inducible common fragile site, FRA6H. The gaps/breaks at 13q22 occurred within the light 13q22 G-band, which were clearly distinct from the proximal gaps/breaks at FRA13C, located at 13q21 (dark G-band). Moreover, the expression of both fragile sites on the same chromosome has been seen in several metaphases, and the rare case with a breakage at all four aphidicolin inducible common fragile sites located at chromosome 13 is depicted in Figure 5b. Hence, the revealed fragility at 13q22 is caused by expression of a so far undescribed aphidicolin inducible common fragile site, FRA13H.

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

One of the molecular approaches for the identification of common fragile sites is based on the cloning of integration sites of exogenous DNA, whose preferential integration into fragile site regions has been observed both in vitro and in vivo.10, 13, 42, 43, 44 In this study, spontaneously expressed fragile sites in the breast cancer cell line MDA-MB-436 were genetically tagged with the vector pSV2neo by insertional mutagenesis. Our data show that the identified DNA sequences of pSV2neo integration sites are situated within (6p21) or less than 1 Mb apart (13q22) of aphidicolin inducible common fragile sites in EBV-transformed lymphocytes. Here, we report that the integration sites at 6p21 and 13q22 correspond to fragile site regions, which are not listed in the NCBI database. These new common fragile sites are designated as FRA6H and FRA13E, respectively.

The identification of these two fragile sites was based on the cloning of the integration sites of exogenously supplied vector DNA. In vivo, viral DNA is known to preferentially integrate into regions of fragile sites. Two viral integrations of HPV16 within the band 6p21 have been previously reported in a head and neck cancer cell line.45 The location of these HPV16 integrations corresponds to the 9.3 Mb FRA6H region in our study, with one of them being situated less than 230 kb distal to the pSV2neo vector integration site. Chromosomal band 13q22 is frequently targeted by HPV16 integration in cervical cancer.44, 46, 47 All of the five HPV16 integration sites at 13q22 described so far cluster in a 300 kb region within the described 3.1 Mb FRA13E.

Besides the preferential integration of exogenous DNA, high DNA flexibility represents another typical attribute of fragile sites. Here, the TwistFlex program was used to estimate the DNA flexibility of the sequences encompassing FRA6H and FRA13E, as well as adjacent proximal and distal sequences. The sequence of FRA6H, situated in an R-band, has a DNA flexibility value ranging between the average values for fragile and nonfragile sequences.14 However, the same value was estimated for FRA3B, the “most fragile” common fragile site in human genome. The flexibility values of the flanking nonfragile sequences fit also to the corresponding expected average values. Thus, the sequence analysis of FRA6H and its nonfragile flanking sequences correlates to the breakage analysis with FISH. The DNA flexibility of the sequence encompassing the FRA13E fragile site corresponds to the average value for fragile sequences situated at R-bands.14 However, the flanking control sequences to FRA13E, which were proven to be nonfragile by breakage analysis, showed flexibility values similar to fragile DNA sequences. In silico calculations of DNA flexibility can reveal DNA sequences with a high potential of forming secondary structures that might account for the DNA fragility. However, to date only the results obtained by in vitro breakage analyses led to the identification of fragile regions in the genome.

Chromosomal lesions in the regions of bands 6p21 and 13q22 were described the first time in connection with the discovery of aphidicolin, an inhibitor of DNA-polymerases α, δ and ε, as an inducer of common fragile sites.48 Since then, breakage in 6p21 has been observed in several independent studies, in some, gaps/breaks in 6p21 occurred with >2% relative frequency of all observed gaps/breaks.49, 50, 51, 52 The observed incidence of the fragile site at 6p21 varies from 11%53 up to 94%54 in the human population. The nonrandom breakage at 13q22 has been observed by independent laboratories48, 55 and has been proposed to be caused by a new aphidicolin inducible common fragile site.56 Besides the induction with aphidicolin, a variety of mutagens induced breakage at 6p21 and 13q22.57 Moreover, spontaneous expression of their fragility was observed in chorionic villus cells.29

There is only one study that shows the use of camptothecin as an enhancer of the induction of common fragile sites.36 Here we demonstrated, exemplary for FRA6H, that the region affected by an aphidicolin induction of fragile sites and the region affected by an induction with both chemicals are identical. The observations that breaks at 6p21 can be induced with each of the chemicals alone (camptothecin,36 aphidicolin48) indicate that camptotecin may not only have a synergistic effect on aphidicolin but might also induce aphidicolin inducible common fragile sites on its own.

Breakage at fragile sites can lead to altered expression of the genes located within these regions being caused by deletions, translocations or amplifications triggered by the fragile site. For instance, large deletions in the steroid 21-hydroxylase (CYP21) gene, located within FRA6H, account for up to 18% of all known mutations causing congenital adrenal hyperplasia, a common autosomal recessive disorder.58 In addition to deletions and translocations,59 amplifications of the 6p21 region via breakage-fusion-bridge cycle have been shown in osteosarcoma.60, 61 In lymphoma cells, an amplicon boundary was localized exactly to the DNA sequence corresponding to FRA6H.61FRA6H harbors genes of the HLA locus, including the HLA class I, II and III genes, whose frequent loss of heterozygosity (LOH) has been observed in a variety of tumors (reviewed in Ref.62). In some studies, the proximal border of the LOH region was identified as being situated directly within FRA6H.62, 63, 64, 65, 66, 67 The frequently observed LOH in this region leads to a HLA haplotype loss of the cell, which in turn enables the cell to escape immunogenic surveillance mechanisms via cytotoxic T-lymphocytes.

The 3.1 Mb region of FRA13E encompasses only eight genes: the homolog of S. bombe DIS3 (KIAA1008) gene, the progesterone-induced blocking factor 1 (PIBF1) gene, the kruppel-like factor 5 and 12 (KLF5, KLF12) genes, the TBC1 domain family member 4 (TBC1D4) gene, the COMM domain containing protein 6 (COMMD6) gene, the ubiquitin carboxyl-terminal esterase L3 (UCHL3) gene and the LIM domain only 7 (LMO7) gene. The region of FRA13E is located within the core of the putative breast cancer susceptibility locus at 13q21-q22, which was defined by linkage analysis in BRCA1- and BRCA2-independent breast cancer families.68 In the Mitelman Database of Chromosome Aberrations in Cancer, the majority of the listed recurrent chromosomal aberrations for 13q22 are large scale deletions bordered by or being restricted to the chromosomal band 13q22 (http://cgap.nci.nih.gov/Chromosomes/Mitelman). Besides the observed deletions in breast cancer, 13q22 is often deleted in leukemia and lymphomas. An aggressive form of prostate cancer is associated with the smallest region of overlapping deletion (SRO) at 13q21-13q22,69 where the distal border is situated within the here identified FRA13E. In addition to somatic alterations in cancer cells, germline rearrangements involving 13q22 have been detected in patients with partial trisomy 13 or with distal 13q deletion syndrome, exemplary.70, 71

In summary, the approach of genetically tagging fragile site sequences by insertional mutagenesis has led to the cloning and molecular characterization of two new aphidicolin inducible common fragile sites, FRA6H and FRA13E. The identified regions of instability are associated with deletions, translocations and amplifications, which might contribute to both malignant and nonmalignant disorders.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The authors thank Mrs. Neta Ben-Porat for help with TwistFlex program and Dr. Kai-Oliver Henrich for critical reading of the manuscript.

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References