Cloning and characterization of the common fragile site FRA6F harboring a replicative senescence gene and frequently deleted in human tumors

Abstract

The common fragile site FRA6F, located at 6q21, is an extended region of about 1200 kb, with two hot spots of breakage each spanning about 200 kb. Transcription mapping of the FRA6F region identified 19 known genes, 10 within the FRA6F interval and nine in a proximal or distal position. The nucleotide sequence of FRA6F is rich in repetitive elements (LINE1 and LINE2, Alu, MIR, MER and endogenous retroviral sequences) as well as in matrix attachment regions (MARs), and shows several DNA segments with increased helix flexibility. We found that tight clusters of stem-loop structures were localized exclusively in the two regions with greater frequency of breakage. Chromosomal instability at FRA6F probably depends on a complex interaction of different factors, involving regions of greater DNA flexibility and MARs. We propose an additional mechanism of fragility at FRA6F, based on stem-loop structures which may cause delay or arrest in DNA replication. A senescence gene likely maps within FRA6F, as suggested by detection of deletion and translocation breakpoints involving this fragile site in immortal human-mouse cell hybrids and in SV40-immortalized human fibroblasts containing a human chromosome 6 deleted at q21. Deletion breakpoints within FRA6F are common in several types of human leukemias and solid tumors, suggesting the presence of a tumor suppressor gene in the region. Moreover, a gene associated to hereditary schizophrenia maps within FRA6F. Therefore, FRA6F may represent a landmark for the identification and cloning of genes involved in senescence, leukemia, cancer and schizophrenia.

Introduction

Gaps, breaks and other chromosome aberrations are induced in cells cultured under specific conditions, generally in the presence of drugs inhibiting or delaying DNA replication. Regions that show a statistically significant increase in the frequency of such abnormalities, above that occurring by chance throughout the human genome, are defined as fragile sites (Sutherland et al., 1998). In recent years, fragile sites have raised considerable interest, because they appear to participate in chromosomal rearrangements involved in cancer, leukemia, mental retardation and aging (Le Beau and Rowley, 1984; Yunis and Soreng, 1984; Smith et al., 1998; Sutherland et al., 1998).

Fragile sites are classified into two categories, rare and common or constitutive, depending on the frequency observed in the population and on the culture conditions inducing their expression (Smith et al., 1998; Sutherland et al., 1998). The 89 common fragile sites are expressed constitutively in all individuals, although the level of expression can vary from 1 to 70% of metaphases for different fragile sites. Variations in expression may also be observed in different individuals for the same fragile site. Most common fragile sites (84 out of 89) are induced by aphidicolin, a drug which inhibits DNA polymerases α and δ, but other common fragile sites are induced by 5-azacytidine or bromo-deoxyuridine (Glover et al., 1984; Smith et al., 1998; Sutherland et al., 1998).

The structure and molecular features of common fragile sites are not well characterized. However, those which have been better defined or cloned, such as FRA3B at 3p14.2, FRA7G at 7q31.2, FRA7H at 7q32.3, FRA16D at 16q23.2 and FRAXB at Xp22.1, share some characteristics. Constitutive fragile sites are generally rich in repetitive sequences (LINE, Alu, MER). They may also contain small polydispersed circular DNA sequences, endogenous retroviral sequences and integration sites for plasmids or for DNA tumor viruses such as HPV16 and SV40 (Rassool et al., 1996; Wilke et al., 1996; Boldog et al., 1997; Inoue et al., 1997; Huang et al., 1998b; Huebner et al., 1998; Mishmar et al., 1998; Mimori et al., 1999). Studies carried out on fragile sites FRA3B (Mimori et al., 1999) and FRA7H (Mishmar et al., 1998) have shown coincidental regions of greater DNA flexibility and lower DNA stability within the fragile sites compared to genomic regions not harboring sites of fragility. Areas of high flexibility and low stability colocalized with clusters of DNA sequences with non-B and triple-helix conformation (Mishmar et al., 1998) which are known to inhibit DNA replication and enhance homologous recombination. Finally, all the common fragile sites are regions of late DNA replication, and this delay in replication is enhanced by aphidicolin (Le Beau et al., 1998; Wang et al., 1999; Hellman et al., 2000). It has been suggested that late-replicated DNA regions would affect locally the chromatin structure, leading to the unstable and recombinogenic properties of fragile sites (Smith et al., 1998; Sutherland et al., 1998; Wang et al., 1999). Indeed, since DNA replication must be completed before the initiation of mitosis, delayed replication may cause the incomplete packaging and collapse of the chromosome structure during metaphase, resulting in expression of fragile sites (Simonic and Gericke, 1996). In spite of this knowledge on the characteristics of common fragile sites, a molecular basis and a clear explanation for the mechanisms of their fragility are still lacking. Hence, the need to clone further fragile sites in order to clarify what DNA sequence elements and molecular organization give rise to such regions in the human genome.

Common fragile sites seem to play a functional role and to be involved in cancer. In fact, cancer associated deletions and translocations often affect chromosome regions containing fragile sites. Recently, it was shown that expression of fragile sites triggers intrachromosomal gene amplification (Coquelle et al., 1997). Through these chromosomal aberrations, fragile sites may play a role both in loss of tumor suppressor genes and in amplification of oncogenes. These hypotheses have recently gained more credit by the observation that FRA3B, the most frequently expressed human common fragile site, maps within the sequence of the tumor suppressor gene FHIT at 3p14.2. The FHIT gene has abnormal transcripts in colorectal, lung and breast carcinomas as well as in other human tumors, resulting primarily from deletions generated by recombination between or within LINE sequences in the region of FRA3B (Huebner et al., 1998). The common fragile site FRA7G maps at 7q31.2, in a region of frequent loss of heterozygositey (LOH) in breast and prostate cancer (Huang et al., 1998a). A detailed LOH analysis in ovarian cancer showed the highest frequency of deletion breakpoints within the sequences of FRA7G (Huang et al., 1999). Recently, homozygous deletions have been detected at 16q23.2, within the common fragile site FRA16D, in tumor cell lines derived from colonic, gastric, ovarian and small cell lung carcinoma (Mangelsdorf et al., 2000; Paige et al., 2000; Ried et al., 2000; Krummel et al., 2000). FRA16D associated deletions affect the transcripts of the FOR gene, also named WWOX (Bednarek et al., 2000, 2001), spanning FRA16D and belonging to the family of oxidoreductases (Ried et al., 2000).

The process of immortalization of human diploid fibroblasts by simian virus 40 (SV40) involves non-random chromosomal aberrations induced by SV40 T antigen (Ray et al., 1990). Most of the chromosomal alterations are gaps and breaks identical to those induced by aphidicolin and may therefore involve fragile sites. The highest frequency of chromosomal aberrations in human diploid fibroblasts immortalized by SV40 is detected in the long arm of chromosome 6 (Hubbard-Smith et al., 1992; Ray and Kraemer, 1992; Sandhu et al., 1994) which harbors five common fragile sites. We have mapped a gene inducing replicative senescence in a 4 Mb region at 6q21, where we have constructed a complete YAC contig and a partial BAC/PAC contig (Morelli et al., 1997a,b). This region is the most common breakpoint in a panel of translocation, deletion and radiation monochromosome human/mouse cell hybrids containing human chromosome 6 (Pappas et al., 1995; C Morelli and G Barbanti-Brodano, unpublished). In preliminary fluorescence in situ hybridization (FISH) experiments with YACs of the 4 Mb contig, we detected breaks at 6q21 in metaphases of aphidicolin-treated human lymphocytes, suggesting that the common fragile site FRA6F, cytogenetically mapped at 6q21 (Smith et al., 1998), lies within the region harboring the senescence gene. We therefore decided to finely map and clone FRA6F as a possible landmark of the gene at 6q21 inducing replicative senescence.

Results

Localization of FRA6F

A senescence gene was mapped at chromosome region 6q21 by chromosome transfer experiments. Introduction of an intact human chromosome 6 into immortalized mouse cells by microcell fusion induced senescence in most recipient cells. The few clones that grew up and maintained the immortalized phenotype of the original cell line invariably showed a common region of deletion involving chromosome bands 6q21-q22 (Gualandi et al., 1994). Subsequently, the region containing the senescence gene was restricted to 4 Mb by molecular analysis and localized between markers D6S1499 and D6S266 at chromosome region 6q21 (Morelli et al., 1997a,b). During our studies on the senescence gene, we characterized several translocation, deletion and radiation monochromosomic human-mouse somatic cell hybrids containing a rearranged chromosome 6 or portions of the chromosome (Pappas et al., 1995; Morelli et al., 1997a; Karayianni et al., 1999). The majority of the hybrids show breakpoints within the 4 Mb region containing the senescence gene. Since fragile sites are regions particularly prone to breaking and the 4 Mb region at 6q21 is frequently deleted in immortalized cells (Ray and Kraemer, 1992; Morelli et al., 1997b), it is possible that FRA6F, cytogenetically mapped at 6q21, lies within the senescence gene region. We therefore used three YAC clones, belonging to the contig covering the 4 Mb portion of the 6q21 region (Morelli et al., 1997b), in FISH experiments on aphidicolin-treated human lymphocyte metaphases to assess whether FRA6F is located in the senescence gene region. We chose YAC clones 852d6, 933c4 and 856g2 because they are not chimeric and cover the whole 4 Mb region (Morelli et al., 1997b). Since FRA6F is expressed at low frequency (1–2%) on human chromosome 6, the location of the signal relative to the position of the fragile site (proximal, crossing or distal) was scored on 15 metaphases expressing FRA6F, corresponding to a total of 800 metaphases, for each probe. YAC clones 852d6 and 933c4 show proximal, crossing and distal signals relative to the fragile site, whereas YAC clone 856g2 gives only distal and a single crossing signal (Table 1 and Figure 1). The ratio of the number of distal plus crossing over proximal plus crossing signals (D+C/P+C) is also reported in Table 1. This value indicates that YAC clones 852d6 and 933c4, despite giving proximal, crossing and distal signals, do not contain the whole fragile site, since the D+C/P+C ratio is 0.5 and 1.7 respectively instead of 1. From these data we argue that FRA6F may extend proximally in a region larger than the overlapping region of the 852d6 and 933c4 YAC clones.

Table 1 Localization of FRA6F by FISH analysis
Figure 1
figure1

Physical and transcriptional map of the FRA6F region. The physical map and the whole region of breaking are reported in the top part of the figure with the markers distributed along the region. Dotted lines indicate the two hot-spots of breaking, revealed by FISH experiments with PAC probes. Thick bold lines represent the YAC clones used for initial FRA6F localization. PAC clones used in FISH experiments are shown in the figure with plain lines. PAC clones subjected to sequence analysis are indicated with dashed lines. PAC clones marked with both continuous and dashed lines have been used for FISH experiments and sequence analysis. The transcriptional map of FRA6F is shown in the lower part of the figure with genes distributed in the regions proximal to, within and distal to FRA6F

Characterization of FRA6F

In order to address this question and to precisely define FRA6F, we assembled a contig of 14 PAC clones covering the whole region from marker CMFE7 to marker D6S1066 and extending centromerically beyond the overlapping region of YAC clones 852d6 and 933c4 (Figure 1). The continuity of the PAC contig was assessed by known STS content and by developing new STSs from PAC end sequences. The correct orientation of PAC clones in relation to centromere and telomere was defined in triple color FISH experiments using a proximal or distal probe as the anchor point. Each PAC clone was used as a probe in FISH experiments on aphidicolin-treated metaphases of human lymphocytes (Figure 2). At least 11 metaphases expressing FRA6F were scored for each probe as previously described for YAC clones. The number of proximal, crossing and distal signals was recorded. Based on the number of crossing signals given by each PAC (Table 1), two hot spots of breaking, measuring each about 200 kb, were detected, one proximal involving PACs RP3-442M11, RP3-415N13 and RP3-487J7, the other distal involving PACs RP1-139A3, RP1-96N13, RP1-97J1 and RP1-276A6 (Table 1; Figure 1). The value of D+C/P+C ratio was calculated (Table 1), although it is less significant as compared to the same value obtained with YAC clones. However, the small size of PAC probes allowed a more detailed analysis which detected two regions of breakage.

Figure 2
figure2

Chromosomal localization of FRA6F by FISH analysis. (a) In this experiment FRA6F was detected on metaphases of aphidicolin-treated normal human lymphocytes, using PAC RP1-97J1 (see Figure 1) as the probe. Only one of the two copies of chromosome 6 shows a break at FRA6F (right green dots). Due to the rare expression of FRA6F, this was the general pattern observed in most of the metaphases analysed. (b) The propidium iodide staining of the same metaphase is shown in black and white. In a and b arrows indicate the position of FRA6F at 6q21

Transcription map of the FRA6F region

Ten known genes map within the fragile site region. In a proximal to distal position they are: dj1112D6.1, REV3L (Gibbs et al., 1998; Morelli et al., 1998), DIF13, C6UAS and C6ORF4-6 (Morelli et al., 2000), FYN (Karayianni et al., 1999), H3F3A, FKHRL1 (Hillion et al., 1997), dj142L7.3, LAMA4, (Figure 1). Furthermore, nine other genes map close to FRA6F, in a proximal or distal position (Figure 1). Three CpG islands, indicating the presence of putative new genes and corresponding to markers D6S1499, D6S302 and D6S416, were detected within FRA6F.

Sequence analysis of FRA6F

To analyse the nucleotide sequence of FRA6F, PACs RP5-1112D6, RP3-442M11, RP3-415N12, RP3-487J7, RP1-66H14, RP1-139A3, RP1-97J1, RP1-276A6 and RP1-142L7 were sequenced. These PACs cover essentially all the region of FRA6F in the contig (Figure 1), since a total of 1100 kb were sequenced. In particular, the regions corresponding to the two hot spots of breakage were completely sequenced on PACs RP3-422M11, RP3-415N12, RP3-487J7, RP1-97J1 and RP1-276A6 (Figure 1). The nucleotide sequence of the contig was used to measure the length of FRA6F. By subtracting the overlapping regions of the clones, the length of the whole fragile site and of each of the two hot spots was found to be 1200 and 200 kb, respectively.

The DNA sequence spanning FRA6F was analysed for the type and quantity of DNA repeats (Figure 3), since DNA repeats have been proposed to have a role in common fragile site instability (Inoue et al., 1997; Mimori et al., 1999). Total repetitive elements represent 35.7% of the entire DNA sequence in FRA6F, an amount which is not significantly different from a standard region of the genome (37.4%) with a similar GC content (Genome 1 in Figure 3). Notable features of FRA6F were a high content of Alu sequences in the proximal hot spot and the regions outside the hot spots and a considerable amount of DNA transposons of the MER type in the proximal hot spot (Figure 3). Although LINE (LINE1+LINE2) sequences were consistently represented in the two hot spots and in the other regions of FRA6F, they were less abundant than in the region of the genome (Genome 1) with the corresponding GC content (Figure 3). Twenty-two sequences with the typical characteristics of matrix attachment regions (MARs), which have been suggested to participate in chromosomal instability (Wang et al., 1997; Mishmar et al., 1998), were detected in FRA6F (Table 2). While the average density of MARs in the human genome is of one every 80–90 kb (Luderus et al., 1992), the density in FRA6F, corresponding to 22 MARs in 1200 kb, is of one every 54 kb. Moreover, in some portions of FRA6F MARs were clustered, such as in PAC RP5-1112D6 (five MARs in 67 kb) or in PAC RP1-66H14 (four MARs in 57 kb) (Table 2). No CCG repeats or long AT-rich sequences were detected in FRA6F.

Figure 3
figure3

DNA repeat composition of FRA6F proximal (prox) and distal (dist) hot spots of breakage and of the regions outside (out) the hot spots. The DNA repeat composition of the other fragile sites FRA3B, FRA6E, FRA7H, FRA16D and of three regions of the whole genome with increasing GC content (Genomes 1, 2 and 3) are also shown for comparison. Values of percentage refer to the percentage of repetitive elements relative to the entire sequence of FRA6F. Values for FRA3B, FRA6E, FRA7H and FRA16D are from Ried et al. (2000), values for Genomes 1, 2 and 3 are from Smit (1996)

Table 2 Analysis for MAR sequences within FRA6F

In the FRA6F sequence we detected stem-loop structures with a stem of 20–50 bp and a loop of 20–200 bp, often aggregated in clusters of more than 80 stem-loops in 2 kb of DNA sequence (Table 3). Clusters with the highest density were found in the proximal and distal hot spots of FRA6F (PAC clones RP3-415N12, RP1-139A3 and RP1-97J1; Figure 1 and Table 3), whereas the regions outside the hot spots, (PAC clones RP5-1112D6, RP3-487J7, RP1-66H14 and RP1-142L7; Figure 1 and Table 3), a genomic region (9p23) selected at random and other fragile sites, with the exception of FRA6E, had a much lower density of stem-loop structures (Table 3). Interestingly, the genomic region 22q11, frequently involved in chronic myeloid leukemia translocations, with consequent truncation of the BCR gene, shows a high density of stem-loops, similarly to the unstable hot spot regions of FRA6F (Table 3). Contrary to FRA6F and FRA6E, in fragile sites FRA3B, FRA7G, FRA7H and FRA16D clusters of stem-loops were detected in connection to the presence of dinucleotide repeats (Table 3).

Table 3 Distribution of stem-loop structures in FRA6F, other fragile sites and control genomic DNA

The analysis of DNA helix flexibility in the FRA6F sequence was carried out with the bend.it program. This program measures the flexibility parameter which is expressed as fluctuation in the twist angle. The analysis revealed six regions with potential high flexibility (predicted curvature parameter greater than 14) in the proximal hot spot and five regions in the distal hot spot of FRA6F. The analysed sequences of other fragile sites FRA3B (180 kb), FRA6E (330 kb) and FRA16D (270 kb) showed 14, 10 and 11 regions of greater flexibility, while the control genomic DNA showed no regions of high flexibility (Figure 4). Stem-loop clusters in FRA6F and FRA6E were located in regions of greater stability, close to high flexibility sites of the DNA helix (Figure 4).

Figure 4
figure4

Analysis of DNA helix flexibility in the sequence of FRA6F proximal and distal regions of maximum breakage (proximal hot spot and distal hot spot) as well as in the other fragile sites FRA3B, FRA6E, FRA16D and in a non fragile control region (20q11) of the genome. The FRA3B sequence analysed for flexibility corresponds to the 180-kb region encompassing FHIT exon 4 (Boldog et al., 1997; Inoue et al., 1997; Mimori et al., 1999). The values of the curvature of the DNA helix are expressed in the y axis in degrees per helical turn. Values of the curvature above 14 (marked by dots) are considered indicative of high flexibility. The x axis reports the size (kb) of the region subjected to the flexibility analysis. The arrows indicate the positions where stem-loop clusters were detected

Detection of deletion and translocation breakpoints involving FRA6F in immortal cells, leukemia and cancer

A gene inducing replicative senescence was mapped in a 4 Mb region at 6q21 containing FRA6F (Morelli et al., 1997a,b). Therefore, to assess directly whether breaks occur at FRA6F and are involved in the process of immortalization, we carried out a deletion analysis by PCR (see Materials and methods) on a series of immortal cell lines bearing translocation and deletions involving region 6q21 (Figure 5). In all the eight cell lines examined (two translocations human-mouse fusion cell hybrids, one monochromosomic human-mouse hybrid containing a human chromosome 6, two SV40-immortalized intraspecific human fibroblast hybrids carrying an additional exogenous chromosome 6 and three human fibroblast cell lines immortalized by SV40; see Materials and methods) one breakpoint end was detected within FRA6F, either in the distal hot spot, in the region between the two hot spots or in the telomeric portion of the fragile site (Figures 5 and 6). These results indicate that FRA6F is a true region of instability in the human genome and that a senescence inducing gene likely maps within FRA6F. Moreover, deletions involving breaks at FRA6F are detected in human leukemias and lymphomas, in breast, gastric, ovarian, prostate and salivary gland carcinoma, in melanoma and mesothelioma (Table 4Table 4).

Figure 5
figure5

Breakpoint ends within FRA6F in cell lines (left column) showing translocations and deletions involving region 6q21 (see the sections on Materials and methods and Results for details). FRA6F is depicted with the two hot spots of breakage represented by dotted lines. Markers below FRA6F were used for the molecular analysis of cell lines by PCR. Solid lines corresponding to the cell line names indicate the conserved portion of chromosome 6. Markers at breakpoint ends indicate the chromosome 6 region of break in each cell line

Figure 6
figure6

LOH analysis for markers A116zg9, D6S404 and D6S302, located within FRA6F, in immortal human cell lines (see Materials and methods). XP6+C1.11 and XP6+C1.12 are two clones of immortal cells derived from transfer of a normal human chromosome 6 from the donor cell line 262A1D6 to the recipient cell line XP12ROSV. The two clones are compared with the donor and recipient cell lines. MRC5-SV-TG1, GM04429 and GM04312 are human diploid fibroblasts immortalized by SV40. They are compared with the normal parental fibroblast strains. Note that polymorphic markers present in the donor cell line 262A1D6 are lost in clones XP6+C1.11 and XP6+C1.12. Likewise, markers present in normal parental fibroblasts are lost in the derived SV40-immortalized cell lines

Table 4 Deletions within FRA6F in human leukemias

Discussion

We have cloned and characterized the common fragile site FRA6F in a 1200 kb region at 6q21. It is notable that numerous genes map within FRA6F, in contrast to other fragile sites such as FRA3B and FRA16D that are associated each to only one gene, FHIT and FOR respectively (Huebner et al., 1998; Ried et al., 2000). The richness of genes in FRA6F may be relevant to the biological consequences of DNA instability at this fragile site.

Molecular mechanisms of FRA6F instability

Repetitive elements may constitute the molecular basis of common fragile site expression, as the expansion of repeat elements is the only known cause for the instability of rare fragile sites (Sutherland et al., 1998). The DNA, affected at the fragile region by aphidicolin, carcinogens or other damaging agents, may break and be repaired during replication. If the break falls within or close to a LINE repeat, annealing of the homologous sequences of the repeat element may generate a stem-loop structure inducing deletion of a DNA segment, by nonhomologous or unequal homologous recombination, during the process of replication/repair in the damaged region (Mimori et al., 1999). In agreement with this model, association of LINE1 elements and cancer deletion breakpoints was observed at the proximal end of FRA3B (Inoue et al., 1997). However, no association was found between repetitive sequences and aphidicolin-induced or cancer deletion breakpoints located at the distal end of FRA3B (Wang et al., 1997) or at FRA16D (Ried et al., 2000). A considerable degree of variation in the type and quantity of repeat DNA sequences was observed in the three regions of FRA6F. In particular, LINE1 elements were significantly under-represented when compared with the normal human genome. Variation in repeat element composition and under-representation of LINE1 elements was observed also in FRA16D (Ried et al., 2000). On the whole, these observations suggest that a specific repeat composition is unlikely to be related to the mechanism of DNA instability of fragile sites.

DNA helix flexibility correlated with DNA instability at FRA3B and FRA16D (Mimori et al., 1999; Ried et al., 2000). Regions of greater DNA helix flexibility influence protein-DNA interactions (Mishmar et al., 1998, 1999), and may therefore affect chromatin structure and condensation, generating DNA instability. The high flexibility peaks detected in FRA6F proximal and distal hot spots of breakage are therefore candidate regions for the DNA instability of this fragile site. Delay in DNA replication plays a crucial role in DNA instability at fragile sites. Cytogenetic analysis indicated that aphidicolin-induced breaks at FRA3B preferentially occurred on the chromosome 3 with a late replicating allele (Wang et al., 1999). In situ hybridization and FISH analysis showed that the FRA7H region has a replication delay during S phase in about 35% of the nuclei. The results of this study indicated that replication delay was due to perturbation of replication fork progression (Hellman et al., 2000). This effect could depend on two peculiar structures present in FRA6F: MARs and stem-loop clusters.

Due to the attachment of MARs to the nuclear matrix, progression of the replication fork through these DNA elements may be hindered and these regions may normally replicate late in S phase. They could replicate even later under aphidicolin treatment and fail to complete replication by the time chromatin is packaged into chromosomes. Further replication delay may derive from the greater density and aggregation of MARs in clusters at FRA6F as well as from generation at MARs of independent loop domains, due to unwinding of the DNA helix and continuous base unpairing (Kohwi-Shigematsu and Kohwi, 1990; Bode et al., 1992). MARs were detected in FRA3B (Wang et al., 1997) and FRA7H. In FRA7H, six MARs colocalize in a cluster with a high flexibility region and an SV40 integration site (Mishmar et al., 1998), strongly suggesting that they participate in chromosomal instability at this fragile site. We detected clusters of stem-loops in FRA6F. Similar multiple hairpin structures were detected close to the eight breakpoints at the distal region of breakage in FRA3B, and in FRA16D in coincidence with some of the breakpoint locations in AGS and HCT116 tumor cell lines (Ried et al., 2000). The expanded trinucleotide repeats and the AT-rich minisatellite repeats, responsible for the DNA instability at the rare fragile sites FRA10B and FRA16B, are also capable of forming hairpin structures (Sutherland et al., 1998). Stem-loop and hairpin structures are candidates to directly induce deletions by nonhomologous recombination, according to a mechanism akin to that proposed for LINE1 sequences. Moreover, they may generate localized sequence elements with the properties of DNA polymerase pause sites, representing areas of the genome where progression of the DNA polymerase is hindered or delayed. These sites would play an important role in inhibiting or delaying DNA replication. Since clusters of stem-loops were detected in FRA6F and FRA6E exclusively at regions of high stability, adjacent to flexibility zones of the DNA helix, clustered stem-loops may cooperate with high DNA flexibility in inducing DNA breaks. It will be interesting to search for clustered stem-loops in future cloned fragile sites to confirm if such structures really represent a common mechanism of fragility.

Involvement of FRA6F in senescence, oncogenesis and schizophrenia

Convincing data exist linking expression of common fragile sites to leukemia and cancer. Activation of common fragile sites by aphidicolin was detected in lymphocytes from colon, rectum, head and neck cancer patients and in their first-degree relatives, with a significant difference compared to lymphocytes of normal controls (Egeli et al., 2000; Tunca et al., 2000a,b). Moreover, several tumor suppressor genes have been cloned from regions containing fragile sites: FHIT from FRA3B (Huebner et al., 1998), TESTIN from FRA7G (Tatarelli et al., 2000) and FOR from FRA16D (Ried et al., 2000). Our study shows that breaks at FRA6F are involved in cell immortalization and oncogenesis, suggesting the presence of a senescence gene and perhaps of a tumor suppressor gene within or close to this fragile site. Hereditary schizophrenia has also been associated with fragile sites. Lymphocytes from schizophrenia patients show activation of a rare fragile site at 2q11.2 and of a common fragile site at 9q12 with a significant difference compared to lymphocytes of normal controls (Chen et al., 1998). A gene involved in hereditary schizophrenia was localized at 6q21. The greatest allele sharing in members of affected families was defined by marker D6S416 (Cao et al., 1997) which maps within FRA6F (Morelli et al., 1997a,b; Karayianni et al., 1999; Morelli et al., 2000 and this study). Therefore, FRA6F may be a landmark for the identification and cloning of genes involved in senescence, leukemia, cancer and schizophrenia.

Materials and methods

Fluorescent in situ hybridization (FISH)

Whole blood was cultured in the presence of PHA for 72 h, followed by colcemid treatment (20 min at 37°C) and standard procedures for harvesting and fixation. For interphase physical mapping of PAC clones, 2.5 mM of sodium borate was added to the hypotonic solution. For fragile site induction, cells were cultured in the presence of 0.4 μM aphidicolin for 24 h before harvesting. After spreading, slides were fixed for 1 h at room temperature in 3 : 1 methanol : acetic acid solution and then dehydrated in an ethanol series. Slides were used for hybridization after aging overnight at 42°C and a 10 min treatment in acetone. YAC and PAC DNA was used to prepare probes for FISH analysis. DNA was extracted from broth cultures by standard techniques. One μg of YAC DNA or 400 ng of PAC DNA were labeled by nick-translation with biotin-16-dUTP or 11-digoxigenin-dUTP. After precipitation in the presence of a 15-fold excess of human Cot1 DNA and a 30-fold excess of sonicated salmon sperm DNA, probes were resuspended in 10 μl of hybridization buffer (50% formamide, 10% dextran sulfate in 2×SSC). For hybridization, chromosomal DNA was denatured by placing the slides in 2×SSC, containing 70% formamide, at 75°C for 5 min. The preparations were then immediately dehydrated through a cold ethanol series. Probes were denatured at 75°C for 5 min and preannealed on slides for 30 min before hybridization. Hybridization was carried out at 37°C for 72 h for YAC probes and once overnight for PAC probes. After washing in stringent conditions (50% formamide in 2×SSC twice at 42°C, 1×SSC twice at 42°C), biotin-labeled probes were visualized with avidin-FITC or avidin-Texas Red, performing a cycle of signal amplification with biotinylated anti-avidin antibody. Digoxigenin-labeled probes were detected using a sheep anti-digoxigenin antibody, a rabbit-FITC anti-sheep antibody and goat-FITC anti-rabbit antibody (Southern Biotechnology Associates, Birmingham, AL, USA). Cell metaphases were counterstained with propidium iodide or DAPI and observed with a Zeiss Axioplan microscope equipped with a triple-pass filter (VYSIS) using the IPLAB Spectrum P software.

Assembly of YAC and PAC contigs and DNA sequence determination

Direct sequencing of PAC clone ends was performed on 1 μg of purified PAC DNA with T7 or Sp6 standard primers. PAC end amplification was carried out by PCR (33 cycles with 30′′ at 94°C, 30′′ at the appropriate annealing temperature and 30′′ at 72°C) in a 10 μl reaction mixture, using 50 ng of YAC DNA or 25 ng of PAC DNA. Complete sequence of all PAC clones in the contig at 6q21 was carried out using established shotgun methods, available at the Sanger Center (URL: http://www.sanger.ac.uk/HGP/overview.shtml).

Sequence analysis

Sequence analysis of FRA6F, other fragile sites and control genomic DNA was carried out on DNA sequences with the following GenBank accession numbers. FRA3Bex4: AF152363; FRA3Bin5: U66722; FRA3Bex5: AF020503; FRA3Bex6: AF152365; FRA6E: AB016897; FRA6F: RP5-1112D6: AL080317; RP3-442M11: AL512325; RP3-415N12: AL136310; RP3-487J7: AL008730; RP1-66H14: Z97989; RP1-139A3: AL109916; RPI-97J1: AL158035; RP1-276A6: AL512299; RP1-142L7: Z99289; FRA7G: AC002066; FRA7H: AF017104; FRA16D: AF217490; region 9p23: AL135923; region 20q11: AL355392; region 22q11: U07000. The composition in DNA repeats was analysed using the RepeatMasker program (URL: http://repeatmasker.genome.washington.edu) which compares DNA sequences to a library of repetitive elements and detects low complexity regions. For the analysis of DNA helix flexibility, the bendability/curvature propensity plot were calculated with the bend.it server (URL: http://www2.icgeb.trieste.it/dna/bend_it.html) using the DNase I based bendability parameters (Brukner et al., 1995) and the consensus bendability scale (Gabrielian and Pongor, 1996). This server predicts DNA curvature from DNA sequences. The curvature is calculated as a vector sum of dinucleotide geometries (roll, tilt and twist angles) using the BEND algorithm of Godsell and Dickerson, and is expressed as degrees per helical turn (10.5 degrees/helical turn=1 degree/base pair). The STEMLOOP program of the Genetics Computer Group (GCG) at the UK Human Gene Mapping Project (URL: http://www.hgmp.mrc.ac.uk) was used to determine the presence and distribution of stem-loop structures within the sequences analysed. MARs were identified using the MarFinder program (URL: http://ncgr.org/MarFinder). This computational approach to find MARs is based on the simultaneous presence of several motifs occurring in the neighborhood of MAR sequences. The density of these motifs in a region of DNA indicates the presence of a MAR in that region.

Cell lines

The following cell lines, cultured in DMEM/F12 medium (GIBCO–BRL) supplemented with 10% fetal calf serum, were subjected to molecular analysis by PCR to detect breakpoints within FRA6F: 5184.4 and GM610, two human-mouse cell fusion hybrids where portions of chromosome 6 are translocated to the X chromosome and to other human and mouse chromosomes, with breaks at 6q21 (Pappas et al., 1995); p6.1, a monochromosomic human-mouse somatic cell hybrid containing, as a sole human genetic complement, a human chromosome 6 deleted at q16.3-q21 (Morelli et al., 1997b); XP6+C1.11 and XP6+C1.12, two clones derived from the SV40-immortalized human fibroblast cell line XP12ROSV after transfer of an exogenous normal human chromosome 6 by microcell fusion from the donor cell line 262A1D6; GMO4312, GMO4429 (NIGMS Human Genetic Mutant Cell Repository, Coriell Institute for Medical Research, Camden, NJ, USA) and MRC5-SV-TG1 (Huschtscha and Holliday, 1983), derived by immortalization of human diploid fibroblasts with SV40. The molecular analysis of cell lines was carried out by PCR with markers located within FRA6F. The markers used were D6S1698, D6S1499, CMFE8, CMFE3, CMFE4, CMFE5, D6S404, D6S302, A116zg9, D6S418, D6S416, D6S1259 and D6S1066. Only polymorphic markers were used to characterize the human cell lines. XP6+C1.11 and XP6+C1.12 cells were compared in a LOH analysis with the donor (262A1D6) and the recipient (XP12ROSV) cell lines, while GMO4312, GMO4429 and MRC5-SV-TG1 cells were compared with the normal parental fibroblast strains (Figure 6).

References

  1. Bednarek AK, Laflin KJ, Daniel RL, Liao Qiaoyin, Hawkins KA, Aldaz CM . 2000 Cancer Res. 60: 2140–2145

  2. Bednarek AK, Keck-Waggoner CL, Daniel RL, Laflin KJ, Bergsagel PL, Kiguchi K, Brenner AJ, Aldaz CM . 2001 Cancer Res. 61: 8068–8073

  3. Bell DW, Jhanwar SC, Testa JR . 1997 Cancer Res. 57: 4057–4062

  4. Bode J, Kohwi Y, Dickinson L, Joh T, Klehr D, Mielke C, Kohwi-Shigematsu T . 1992 Science 255: 195–197

  5. Boldog F, Gemmill RM, West J, Robinson M, Robinson L, Li E, Roche J, Todd S, Waggoner B, Lundstrom R . 1997 Hum. Mol. Genet. 6: 193–203

  6. Brukner I, Sanchez R, Suck D, Pongor S . 1995 EMBO J. 14: 1812–1818

  7. Cao Q, Martinez M, Zhang J, Sanders AR, Badner JA, Cravchik A, Markey CJ, Beshah E, Guroff JJ, Maxwell ME . 1997 Genomics 43: 1–8

  8. Chen C-H, Shih H-H, Wang-Wuu S, Jen Tai J, Wuu K-D . 1998 Hum. Genet. 103: 702–706

  9. Cooney KA, Wetzel JC, Consolino CM, Wojno KJ . 1996 Cancer Res. 56: 4150–4153

  10. Coquelle A, Pipiras E, Toledo F, Buttin G, Debatisse M . 1997 Cell 89: 215–225

  11. Egeli U, Ozkan L, Tunca B, Kahraman S, Cecener G, Ergul E, Engin K . 2000 Head Neck J. 22: 591–598

  12. Gabrielian A, Pongor S . 1996 FEBS Lett. 393: 65–68

  13. Gibbs PEM, McGregor WG, Maher VM, Nisson P, Lawrence CW . 1998 Proc. Natl. Acad. Sci. USA 95: 6876–6880

  14. Glover TW, Berger C, Coyle J, Echo B . 1984 Hum. Genet. 67: 136–142

  15. Gualandi F, Morelli C, Pavan JV, Rimessi P, Sensi A, Bonfatti A, Gruppioni R, Possati L, Stanbridge EJ, Barbanti-Brodano G . 1994 Genes Chrom. Cancer 10: 77–84

  16. Hellman A, Rahat A, Scherer SW, Darvasi A, Tsui L-C, Kerem B . 2000 Mol. Cell. Biol. 20: 4420–4427

  17. Hillion J, Le Coniat M, Jonveaux P, Berger R, Bernard OA . 1997 Blood 90: 3714–3719

  18. Huang H, Qian C, Jenkins RB, Smith DI . 1998a Genes Chrom. Cancer 21: 152–159

  19. Huang H, Qian J, Proffit J, Wilber K, Jenkins R, Smith DI . 1998b Oncogene 16: 2311–2319

  20. Huang H, Reed CP, Mordi A, Lomberk G, Wang L, Shridhar V, Hartmann L, Jenkins R, Smith DI . 1999 Genes Chrom. Cancer 24: 48–55

  21. Hubbard-Smith K, Patsalis K, Pardinas JR, Jha KK, Henderson AS, Ozer HL . 1992 Mol. Cell. Biol. 12: 2273–2281

  22. Huebner K, Garrison PN, Barnes LD, Croce CM . 1998 Annu. Rev. Genet. 32: 7–31

  23. Huschtscha LI, Holliday R . 1983 J. Cell Sci. 63: 77–99

  24. Inoue H, Ishii H, Alder H, Snyder E, Druck T, Huebner K, Croce CM . 1997 Proc. Natl. Acad. Sci. USA 94: 14584–14589

  25. Jackson A, Panayiotidis P, Foroni L . 1998 Genomics 50: 34–43

  26. Karayianni E, Magnanini C, Orphanos V, Negrini M, Maniatis GM, Spathas DH, Barbanti-Brodano G, Morelli C . 1999 Cytogenet. Cell Genet. 86: 263–266

  27. Kohwi-Shigematsu T, Kohwi Y . 1990 Biochemistry 29: 9551–9560

  28. Krummel KA, Roberts LR, Kawakami M, Glover TW, Smith DI . 2000 Genomics 69: 37–46

  29. Le Beau MM, Rassool FV, Neilly ME, Espinosa III R, Glover TW, Smith DI, McKeithan TW . 1998 Hum. Mol. Genet. 7: 755–761

  30. Le Beau MM, Rowley JD . 1984 Nature 308: 607–608

  31. Luderus MEE, De Graaf A, Mattia E, Den Blaauwen JL, Grande MA, De Jong L, Van Driel R . 1992 Cell 70: 949–959

  32. Mangelsdorf M, Ried K, Woolatt E, Dayan S, Eyre H, Finnis M, Hobson L, Nancarrow J, Ventner D, Baker E . 2000 Cancer Res. 60: 1683–1689

  33. Menasce LP, Orphanos V, Santibanez-Koref M, Boyle JM, Harrison CJ . 1994 Genes Chrom. Cancer 10: 286–288

  34. Miele ME, Jewett MD, Goldberg SF, Hyatt DL, Morelli C, Gualandi F, Rimessi P, Hicks DJ, Weissman BE, Barbanti-Brodano G, Welch DR . 2000 Int. J. Cancer 86: 524–528

  35. Mimori K, Druck T, Inoue H, Alder H, Berk L, Mori M, Huebner K, Croce CM . 1999 Proc. Natl. Acad. Sci. USA 96: 7456–7461

  36. Mishmar D, Mandel-Gutfroind Y, Margalit H, Rahat A, Kerem B . 1999 Am. J. Hum. Genet. 64: 908–910

  37. Mishmar D, Rahat A, Scherer SW, Nyakatura G, Hinzmann B, Kohwi Y, Mandel-Gutfroind Y, Lee JR, Drescher B, Sas DE . 1998 Proc. Natl. Acad. Sci. USA 95: 8141–8146

  38. Morelli C, Cardona F, Boyle JM, Negrini M, Barbanti-Brodano G . 1997a Cytogenet. Cell Genet. 79: 97–100

  39. Morelli C, Magnanini C, Mungall AJ, Negrini M, Barbanti-Brodano G . 2000 Gene 252: 217–225

  40. Morelli C, Mungall AJ, Negrini M, Barbanti-Brodano G, Croce CM . 1998 Cytogenet. Cell Genet. 83: 18–20

  41. Morelli C, Sherratt T, Trabanelli C, Rimessi P, Gualandi F, Greaves MJ, Negrini M, Boyle JM, Barbanti-Brodano G . 1997b Cancer Res. 57: 4153–4157

  42. Paige AJW, Taylor KJ, Stewart A, Sgouros JG, Gabra H, Sellar GC, Smyth JF, Porteous DJ, Watson JEV . 2000 Cancer Res. 60: 1690–1697

  43. Pappas GJ, Polymeropoulos MH, Boyle JM, Trent JM . 1995 Genomics 25: 124–129

  44. Queimado L, Reis A, Fonseca I, Martins C, Lovett M, Soares J, Parreira L . 1998 Oncogene 16: 83–88

  45. Queimado L, Seruca R, Costa-Pereira A, Castedo S . 1995 Genes Chrom. Cancer 14: 28–34

  46. Rassool FV, Le Beau MM, Shen ML, Neilly ME, Espinosa III R, Ong ST, Boldog F, Drabkin H, McCarroll R, McKeithan TW . 1996 Genomics 35: 109–117

  47. Ray FA, Kraemer PM . 1992 Cancer Genet. Cytogenet. 59: 39–44

  48. Ray FA, Peabody DS, Cooper JL, Scott-Cram L, Kraemer PM . 1990 J. Cell. Biochem. 42: 13–31

  49. Ried K, Finnis M, Hobson L, Mangelsdorf M, Dayan S, Nancarrow JK, Woollatt E, Kremmidiotis G, Gardner A, Venter D . 2000 Hum. Mol. Genet. 9: 1651–1663

  50. Sandhu AK, Hubbard-Smith K, Kaur GP, Jha KK, Ozer HL, Athwal RS . 1994 Proc. Natl. Acad. Sci. USA 91: 5498–5502

  51. Sheng ZM, Marchetti A, Buttitta F, Champeme MH, Campani D, Bistocchi M, Lidereau R, Callahan R . 1996 Br. J. Cancer 73: 144–147

  52. Shridhar V, Staub J, Huntley B, Cliby W, Jenkins R, Pass HI, Hartmann L, Smith DI . 1999 Oncogene 18: 3913–3918

  53. Simonic I, Gericke GS . 1996 Hum. Genet. 97: 524–531

  54. Smit A . 1996 Curr. Opin. Genet. Dev. 6: 743–748

  55. Smith DI, Huang H, Wang L . 1998 Int. J. Oncol. 12: 187–196

  56. Starostik P, Greiner A, Schultz A, Zettl A, Peters K, Rosenwald A, Kolve M, Muller-Hermelink HK . 2000 Blood 95: 1180–1187

  57. Sutherland GR, Baker E, Richards RI . 1998 Trends Genet. 14: 501–506

  58. Talwalkar VR, Scheiner M, Hedges LK, Butler MG, Schwartz HS . 1998 Cancer Genet. Cytogenet. 104: 111–114

  59. Tatarelli C, Linnenbach A, Mimori K, Croce CM . 2000 Genomics 68: 1–12

  60. Tunca B, Egeli U, Zorluoglu A, Yilmazlar T, Yerci O, Kizil A . 2000a Cancer Genet. Cytogenet. 119: 139–145

  61. Tunca B, Egeli U, Zorluoglu A, Yilmazlar T, Yerci O, Kizil A . 2000b Cancer Lett. 152: 201–209

  62. Utada Y, Haga S, Kajiwara T, Kasumi F, Sakamoto G, Nakamura Y, Emi M . 2000 Jpn. J. Cancer Res. 91: 293–300

  63. Wang I, Darling J, Zhang JS, Huang H, Liu W, Smith DI . 1999 Hum. Mol. Genet. 8: 431–437

  64. Wang L, Paradee W, Mullins C, Shridhar R, Rosati R, Wilke CM, Glover TW, Smith DI . 1997 Genomics 41: 485–488

  65. Wilke CM, Hall BK, Hoge A, Paradee W, Smith DI, Glover TW . 1996 Hum. Mol. Genet. 5: 187–195

  66. Yunis JJ, Soreng AL . 1984 Science 226: 1199–1204

Download references

Acknowledgements

We thank all members of the chromosome 6 project group at the Sanger Centre, funded by the Wellcome Trust, for continuous support in sequence determination. We also thank Annalisa Peverati, Augusto Bevilacqua, Pietro Zucchini and Iva Pivanti for excellent technical assistance. C Morelli was supported by a fellowship from Fondazione Italiana per la Ricerca sul Cancro (FIRC). This work was supported by grants from Associazione Italiana per la Ricerca sul Cancro (AIRC) and from MURST ex 60%.

Author information

Correspondence to Giuseppe Barbanti-Brodano.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • FRA6F
  • fragile sites
  • chromosome 6q21
  • senescence
  • human tumors
  • schizophrenia

Further reading