Introduction

Hypomelanosis of Ito (HI) is a sporadic multisystem disorder of heterogeneous aetiology which commonly shows hypopigmentation of the skin along the lines of Blaschko, in many cases associated with abnormalities of the central nervous system, eyes, hair and musculoskeletal system. Phenotypically there is overlap between sporadic incontinentia pigmenti (IP) cases and HI although the IP locus segregating in families seems to be phenotypically distinct and maps by linkage to Xq28 [1]. Every described HI case exhibits a balanced X;autosome translocation with breakpoints in the pericentromeric region of Xp21.1-q13. In a few sporadic cases described as IP (which is most likely the same phenotype as HI) there is evidence for a common cytogenetic breakpoint in Xp11 [for a review, see ref. 2]. The heterogeneity of the translocation breakpoints mapping in Xp21.1-q13 may explain the clinical heterogeneity of associated findings in HI. It also confirms that it is unlikely that the interruption of a single locus by the breakpoint on the X chromosome is a basic mechanism of this disease [2].

The autosomal breakpoints in HI/IP cases involve chromosomes 5, 9, 10, 13, 14, 22 as described elsewhere [310] with two cases on chromosome 17 [11, 12]. However, in the case of the X;17 translocation originally reported by Hodgson et al. [12] there is evidence against a direct locus interruption since the breakpoints are thought to be in alpha satellite DNA [13]. The cytogenetic analysis of Hodgson et al. [12] and Steichen-Gersdorf et al. [11] has shown the chromosome 17 breakpoints to be distinct and a single locus on chromosome 17 giving rise to HI is not very likely. The translocation breakpoint reported on the short arm of chromosome 17 (17p13) by Steichen-Gersdorf et al. [11] is in a region harbouring the p53 tumour suppressor gene. Mutation or interruption in p53 are the most common tumour-specific genetic changes seen in many different types of cancer [14, 15]. The breakpoint described in this region must also be near the FMR1L2 gene since p53 and FMR1L2 both map to the same radiation hybrid [16].

The hypothesis suggested by Hatchwell [2] argues that the HI phenotype in constitutional X;autosome translocations results not from the interruption of specific X-linked genes but from the presence of mosaic functional disomy of X sequences above the breakpoint. In this paper we isolated the translocation breakpoint in an HI patient with (X;17)(q13;p13) translocation and choroid plexus papilloma [11] and searched for known genes localized in the breakpoint regions mostly to find genes possibly implicated in the patient’s brain tumour and other independent brain tumours. The same translocation was detected in both the DNA from the patient’s brain tumour and peripheral blood lymphocytes and was classified as constitutional.

Materials and Methods

Patient Material

Peripheral blood lymphocytes from the patient with HI and choroid plexus papilloma [11] were immortalised by Epstein-Barr virus (EBV) transformation. The established lymphoblastoid cell line was used for DNA extraction and fluorescence in situ hybridisation (FISH) experiments.

In situ Hybridisation

FISH of cosmids and YACs to normal and patient metaphase spreads were performed essentially as described [17, 18]. Usually, 1 µg of YAC and cosmid DNA was labelled by the BRL Bionick Kit at 16° C for 1 h (cosmids) and 1 h 30 min (YACs). Biotin-labelled probes were purified through Select B columns (Pharmacia). Hybridisation to metaphase spreads proceeded at 37° C for 1 day (cosmids) or for 3 days (YACs). Detection was carried out at 37° C with FITC-conjugated avidin and the signals were amplified with one additional layer of biotinylated goat anti-avidin, followed by an avidin-FITC layer. Chromosomes were counter-stained with propidium iodide in antifade medium (Vector). Metaphases were examined on a Nikon Optiphot microscope and visualised with an MRC 1000 Confocal Imaging System (BioRad). Comos software was used to capture and analyse images.

YAC and Cosmid Clones

YACs, including YAC 4849 (ICRFy900A01103) positive by hybridisation with the DXS1 probe, were isolated from the ICRF human YAC Reference library [19] and CEPH I library [20]. Cosmids A07109 (ICRFc104A07109) and E05106 (ICRFc104E05106) were identified from the ICRF chromosome X specific library filters by hybridisation with the DXS1 probe. Primers for PCR amplification of a 179-bp STSfortheDXS1 marker were: DXS1 -R TTTGGT-AAGACACTTGGGTGC; DXS1-F TTTGAGAGTGGAAACCA-CACC.

PACs (RPCI-1 151.J.10; RPCI-I 59.L.13) were identified by hybridisation screening of the human PAC library [21] using the DXS1 STS. Cosmids H06119 (ICRFc105HO6119D1), D10164 (ICRFc105D10164D1) and G1 14 (ICRFc105G114D1) were identified by screening of the ICRF chromosome 17 cosmid library with the 1.2-kb BamHI/EcoRI fragment isolated from the cloned translocation junction fragment.

Subcloning of Cosmids

Cosmids A07109 and E05106 were digested separately with BamHI and EcoRI, ligated into BamHI and EcoRI digested and dephosphorylated pBluescript (Stratagene) and transformed into DH5a competent cells. Recombinants were isolated and digested DNA was probed sequentially by hybridisation with total human DNA, a LINE specific probe, pLawrist cosmid vector probe and the DXS1 STS probe. EcoRI subclone E-B6 was found to be positive with the DXS1 STS and was then subjected to double digestion with EcoRI and BamHI, HindIII, PstI or XbaI. Unique fragments E-B6 fr0 and E-B6 fr1 were selected as probes as they were negative for total human DNA and LINE probes.

Isolation and Characterisation of the Translocation Junction Fragment

A genomic phage library was prepared from complete EcoRI digests of patient DNA ligated into EcoRI digested EMBL 4 vector DNA (Stratagene). According to the Stratagene protocol, in vitro packaging was performed using freeze-thaw lysate and sonic extract (Gigapack II XL, Stratagene). 500,000 recombinants of the phage library were plated on XL1-MR(P2) host strain and filter lifts were prepared on Hybond-N+ membranes (Amersham), according to the manufacturer’s protocol. The phage library filters were hybridised with a mixture of the probes DXS1 STS, E-B6 fr0 and E-B6 fr1. Four positive phage clones were purified, digested with EcoRI and ligated into EcoRI-digested and dephosphorylated pBluescript and transformed into DH5a-competent cells.

Positive phage clone 3.1.9, bearing the 4.0 kb normal fragment and 4.1.4, bearing the 5.2 kb translocation junction fragment were ligated into EcoRI digested and dephosphorylated pBluescript and transformed into DH5a-competent cells. Recombinants were digested and hybridised with E-B6 fr.0, E-B6 fr.1 and DXS1 probes. Two subclones from 3.1.9 and 4.1.4 were subjected to double digestion with EcoRI and BamHI, HindIII, PstI, SacI or SacII to construct a restriction map.

Exon Amplification

Cosmids AO7109, EO5105 and PACs 151.J.10, 59.L.13 bearing human sequences adjacent to the translocation breakpoint Xq12 and cosmids HO6119, D10164, G114 bearing human sequences close to the breakpoint in 17p13 were double digested with BamHI and BglII and ligated into BamHI digested pSPL3b-CAM dephosphorylated vector (Integrated Genetics). The ligation reaction was transformed into Escherichia coli DH5a host, larger aliquots of each transformation were plated on selective medium and colonies were subsequently pooled for plasmid DNA isolation. COS7 cells were propagated in DMEM medium supplemented with 10% fetal calf serum (FCS). For transformation, COS7 cells were grown to 85% confluence, trypsin-ised, collected by centrifugation and after washing resuspended in ice-cold PBS to have 4 × 106 cells for one electroporation. For electroporation, cells in cuvettes were mixed with 1 µg of plasmid DNA and elec-troporated (1.2 kV; 3 kV/cm; 25 µF). After 10 min, the cells were transferred to a tissue culture dish, containing prewarmed DMEM and 10% FCS. Cytoplasmic RNA was isolated 72 h after transfection and first-strand cDNA synthesis was performed with 5 µg of each RNA, SA2 primer [22], 5 × 1st strand buffer, dNTP (1 µl of 10 mM stock), Superscript RT (200 U; 1 µl). The mixture was incubated 1 h at 37° C, heated at 55° C 5 min and the entire cDNA synthesis reaction was then converted to double-stranded DNA by PCR reaction using SA2 and SD6 primers [22]. To eliminate vector-only splice products and false-positive products, 50 U of BstXI (New England Biolabs) was added and incubated overnight at 55° C 10 µl of the digest was used in the secondary PCR amplification with SA4 and SD2 primers [22]. PCR products were analysed, cloned into DH5a-competent cells. Plasmid DNA from the colonies was isolated and after PCR reaction using the SA4 and SD2 primers, products were analysed on a 2% agarose gel. PCR products of interest were sequenced using the dideoxy chain termination method or by automated TAQ FS cycle sequencing on the ABI 373 DNA Sequencer (Applied Biosystems). Obtained exon sequences were analysed and compared with the non-redundant nucleotide and protein database and EST database using the BLASTN and BLASTX programs on the NCBI server.

Physical Mapping of the Cosmids HO6119 and D10164

1 µg of cosmid DNA was digested with HindIII (H), EcoRI (R) and BamHI (B) in the following combinations: H; H+R; B; B+R; R; H+B+R, transferred onto Hybond N+ membrane (Amersham) and hybridised with PCR products of FMR1L2 gene, using primers:

  • FMR1L2-1 FTCTACAAGGGCTTTGTGAAGGATG; (301–324)

  • FMR1L2-1 R AGGTGGCATCACAGGCAGCATATT; (564–541)

  • FMR1L2-2 F GGAAGATGAATCAAGACCTCAACG; (1955–1978)

  • FMR1L2-2 R AAACCCCATTCACCATACTACCCA; (2244–2221)

and cDNAs 108234 5′ and 3′ (GenBankID T70769 and T69784), 158456 5′ (GenBankID H27109). For detection of SHBG gene (GenBankID M31651) PCR products amplified with these primers were used:

  • SHBG-1 F AGATACCCCGCGGTTCAAAGGC;

  • SHBG-1 R CTGAGACAGGAGGATCGCTTGA;

  • SHBG-2 F TAAGAATGACTGGTTTATGCTG;

  • SHBG-2 R CTAGTACTATAGTGGAAAGAAC;

  • SHBG-3 F CTTCAGCTGAAGCTGAGTATGT;

  • SHBG-3 R ACTTTTCCACTCCTCCAACCTG.

The probes for detection of the junction fragment were 1.2-kb EcoRI/BamHI junction fragment representing the part of chromosome 17- and 265-bp PCR product generated from this fragment using primers:

  • 4.1.4 F AGTGATAACAGAAATAATTATCTG;

  • 4.1.4 R GGATTACACGGTGAGCCGCCGCAC

Results

Localising the Translocation Breakpoint by FISH

Approximately 30 different YACs and cosmids covering different regions in Xq12-q13 were analysed by FISH (fig. 1). Sizes of the YACs ranged from 200 to 1,100 kb, and were estimated by pulsed-field gel electrophoresis and hybridisation with total human DNA probe. FISH analysis of YAC 4849 (1),000 kb) indicated that this clone was non-chimaeric and consistent signals were seen on the patient’s metaphase spreads at Xq12 on both normal and der(X) chromosomes as well as the der(17) chromosome (fig. 2). This observation confirmed that YAC 4849 crossed the translocation breakpoint and this YAC was used to further refine the location of the breakpoint. YAC 4849 was positive for the DXS1 locus and DXS1 was used as a probe to screen the ICRF X chromosome cosmid library. Six positive cosmids were identified and verified using Southern blot analysis, from which A07109 and E05106 were selected for FISH analysis. Both cosmids showed signals at Xq12 on normal X and der(X) chromosomes as well as on the der(17) chromosome, indicating that these cosmids crossed the X chromosome translocation breakpoint (data not shown).

Fig 1
figure 1

YACs and cosmids with appropriate markers covering the region Xq12-q13 used in the FISH analysis. YAC 4849 contains the marker DXS1 and the translocation breakpoint.

Fig 2
figure 2

Detection of the X;17 translocation on metaphase chromosome spreads from the patient’s lymphocytes by FISH analysis using: YAC 4849 covering the Xq12 region and positive for DXS1; three hybridisation sites were observed (A) and confirmation by sequential hybridisation with an X chromosome paint (Cambio) that YAC 4849 hybridises to the normal X and the two translocation products (B).

Detection of the Translocation Junction Fragment

The cosmids A07109 and E05106 were digested with EcoRI and BamHI and subcloned into pBluescript. Forty recombinant clones from both ligations were analysed to search for unique fragments for Southern blot analysis. One subclone (E-B6) was found to hybridise with the DXS1 probe. Two unique subfragments (E-B6 fr0 and E-B6 fr1) identified an altered sized restriction fragment on Southern blots of EcoRI-digested DNA from the patient (fig. 3). To exclude this altered 5.2-kb fragment in the patient as an EcoRI polymorphism, twenty different human samples were tested and all showed only the 4.0-kb EcoRI fragment (data not shown). Junction fragments were also identified with PstI, HindIII, BglII and BclI digests of the patient’s DNA (fig. 3) eliminating the possibility that the additional fragments were due to restriction fragment length polymorphism.

Fig 3
figure 3

Hybridisation of X chromosome DNA probes E-B6 fr0 and fr1 to a Southern blot of control (N) and patient (P) DNA digested with BamHI, EcoRI, PstI, HindIII, BglII, BclII. Only a single restriction fragment is identified in normal DNA lanes (N) while two hybridising fragments are seen in each lane with patient DNA (P) except with BamHI.

The two unique subfragments (E-B6 fr0 and E-B6 fr1) were used to isolate the translocation junction fragment by screening a complete EcoRI digest genomic phage library constructed from the patient’s DNA. Four strongly hybridising phage clones were identified, and after digestion with EcoRI, two clones had the normal 4.0-kb fragment and two clones contained the 5.2-kb junction fragment. All four phage clones were subcloned into pBluescript and clones 3.1.9, containing the normal fragment and 4.1.4, bearing the translocation junction fragment were mapped by restriction enzyme digestion and hybridisation of the probes E-B6 fr0, E-B6 fr1 and DXS1 (fig. 4). The resulting restriction map shows that there was a new SacI site generated in the junction fragment clone 4.1.4 indicating that the translocation breakpoint is localised between the HindIII and SacI sites. After EcoRI and BamHI digestion, a 1.2-kb fragment is present that is specific to the 5.2-kb translocation junction fragment (4.1.4) and is not present in the normal 4.0-kb fragment (3.1.9). This 1.2-kb fragment (JF clone 1.2), therefore represents chromosome-17-specific DNA from the other side of the translocation and was used as a probe to screen the ICRF chromosome 17 cosmid library. Three cosmids, HO6119, D10164 and G114, containing this fragment were identified. The cosmids were shown to map to 17p13 by FISH analysis on normal metaphase chromosomes (data not shown) and were analysed further.

Fig 4
figure 4

Restriction map as determined by complete digestion with EcoRI (E), HindIII (H), BamHI (B), SacI (S), and PstI (P). The upper section shows the restriction map of the 4.0-kb EcoRI fragment (clone 3.1.9) from the normal chromosome X; middle section shows the map of the 5.2-kb EcoRI fragment (clone 4.1.4) from the junction fragment containing sequences from both sides of the X; 17 translocation breakpoint. The lowest section localises the position of E-B6 fr0, E-B6 f1 and DXS1 probes, which were used for identification of the normal and translocation junction fragments.

Identification of Genes Close to the Translocation Breakpoint

Two cosmids and two PACs from Xq12 and three cosmid clones from 17p13 localised on both sides of the translocation breakpoint were analysed for coding sequences by exon amplification [22]. Around 100 exon trap clones, containing inserts larger then 175 bp (vector splice product), were isolated. From these, 33 unique clones were identified and sequenced and all mapped back by hybridisation to the input genomic clones. The sequences were analysed on the NCBI server using both BLASTN and BLASTX searches of the nucleotide (including ESTs) and protein databases.

The sequences of exons 17-2-23 and 17-2-20 were identical with the 5′-end of FMR1L2 (human fragile X mental retardation syndrome like protein 2) gene and the sequence of exon 17-2-27 was identical with the 3′-end of the FMR1L2 gene. Using these exons and representative EST cDNA clones 108234 and 158456 as probes, the FMR1L2 gene was found to be localised on cosmid HO6119 (fig. 5). Exon 17-1-1 was completely homologous with the 5′-end of SHBG (human sex hormone-binding globulin) gene and exon 17-2-3 with the 3′-end of this gene. Hybridisation of exon 17-1-1 and 17-2-3 probes to digested cosmids confirmed the localisation of the SHBG gene within cosmids D10164 and G114 (fig. 5). Both genes were identified by exon trapping on cosmids cloned from chromosome 17. Both of these genes have cytogenetic localisations reported in the Genome DataBase as 17p13. Approximately 50% of the trapped exons were isolated from cosmids cloned from chromosome X. However, no known genes or ESTs were found by nucleotide sequence comparisons (BLASTN) nor were there any proteins found to be significantly similar or homologous to ORFs in these exons (BLASTX). This of course does not rule out a gene from this region but we did not find any obvious candidate exons from the products amplified, sequenced and searched against the EST, nucleotide and protein databases. The putative exons have not yet been screened against cDNA libraries, and exon-linking techniques have not been performed.

Fig 5
figure 5

Restriction map of the cosmids HO6119, Gl14 and D10164 selected from the chromosome 17 cosmid library as containing the translocation junction fragment. A Hybridisation of EcoRI/BamHI digested cosmids HO6119, G114 and D10164 to a 1.2-kb EcoRI/BamHI probe (clone 1.2) from the cloned junction fragment. B Physical map of cosmids HO6119 and D10164. Partial restriction digest maps of HindIII (H), and BamHI (B) are shown. Position of FMR1L2 (5′ and 3′ ends indicated) and SHBG genes is shown related to localisation of the translocation junction fragment (JF clone 1.2).

By hybridisation of the junction fragment clone 1.2 (and a 265-bp STS amplified from this fragment) to digests of the chromosome 17p13 cosmids HO6119, G114, D10164, an overlapping 15- to 18-kb region was found between the three cosmids (fig. 5). From the restriction mapping of cosmids HO6119 and D10164 after hybridisation of 5′ and 3′ specific probes from the two genes identified by exon trapping, it was possible to estimate their position. SHBG appears to be closer to the translocation breakpoint than the FMR1L2 gene, but there is no evidence that either of these two genes was disrupted by the translocation rearrangement (fig. 5). The genomic sequence for SHBG has been determined (M31651) and covers 6 kb. Therefore, the translocation breakpoint did not disrupt this locus nor the coding region of FMR1L2 which was oriented with its 3′ end closer to the breakpoint. However we could not rule out effects on any putative distant regulatory regions of these genes by the translocation.

Discussion

The YAC 4849 (fig. 1) and cosmids A07109 and E05106 which were isolated by DXS1 were shown to cross the translocation breakpoint by FISH analysis. This finding confirmed and refined the localisation of the breakpoint of this patient which was previously analysed by G-banding and hybridisation with a biotinylated chromosome-17-specific library probes [11]. These results also correlate with all described cases of HI where the translocation is localised within the pericentromeric region Xp21.1-q13. As described previously, at least three of the breakpoints in HI/IP patients span a distance of 5 Mb in Xp11 and this case extends the region by at least 10–15 Mb. Therefore, it is unlikely that a single locus or gene is interrupted by all these translocations. As hypothesised [2] the common factor in many cases of HI is the presence of chromosomal mosaicism, although this is not an invariable finding [6, 23]. The presence of a discordant pheno-type in two females with the same breakpoint (X;10) implies that it is the functional status of the derived X that is important, rather than the precise site of interruption. We analyzed the sequences of 17 different exons trapped from the cosmids and PACs spanning the Xq12 breakpoint region and found no strong evidence for the presence of any known or homologous gene or EST in this region. These exon trap sequences are currently being analysed for novel transcripts.

On the chromosome 17p13 side of the translocation, exon amplification revealed two known genes, FMR1L2 and SHBG located very close to the breakpoint. Restriction mapping of cosmids HO6119 and D10464 using gene specific probes was consistent with the coding regions of these two genes not being interrupted by the translocation.

The FMR1L2 gene encodes a protein that is very similar to FMR1 protein, thought to be an RNA-binding protein [24]. The fragile X mental retardation syndrome results from lack of expression of the FMR1 protein or expression of a mutant protein. FMR1L2 and FMR1 interact tightly with a third autosomal homologue FMR1L1 [25] and all three proteins may play important roles in the pathogenesis of this mental retardation syndrome. SHBG and the androgen binding protein produced by Sertoli cells are encoded by the same gene [26] and are generally known to bind steroids. If the genes FMR1L2 and SHBG are correlated with the multisystem disorder of the patient, it is necessary to perform detailed analysis of the breakpoint and the sequences in the neighbourhood of both genes. It remains to be seen whether the translocation disrupted the expression of either of these genes by a position effect but at least these results rule out fusion genes with FMR1L2 or SHBG crossing the translocation that could give rise to the choroid plexus papilloma in this patient.

Chromosome 17 is frequently altered in many human cancers, and allelic loss often coincides with mutations in the p53 tumor suppressor gene, located in 17p13. The translocation in our patient is located within this region, but the p53 gene was not disrupted by the breakpoint. Several overlapping cosmids for p53 were analysed by FISH to rule out any rearrangement (data not shown). It has been shown previously that there is a 17p locus involved in pediatric primitive neuroectodermal tumours that is distinct from p53 [27] and our translocation breakpoint may coincide with this locus. Another factor, HIC-1 (hypermethylated in cancer) found at 17p13.3 [15] which contains a consensus p53 binding site 4 kb upstream from the transcription site is activated by p53 in at least one human tumour cell line. It is possible that there are several regions that are adjacent to p53 related to its activation. Therefore, it is important to test whether this region of chromosome 17 and these particular genes are important in the formation of the X;17 patient’s brain tumour and other forms of brain tumour.