MATERIALS AND METHODS

Patients

The clinical features of patient 1 with the (6;13) reciprocal translocation have previously been described in detail.²⁰ Briefly, the patient had ectrodactyly of the feet, microcephaly (OFC <3rd centile), prognathism, central cleft lip and palate, and severe mental retardation. With the exception of her fingerised thumbs, both hands were normal. She had bilateral microphthalmia and was totally blind. Cytogenetic analysis of the patient showed a balanced translocation, 46,XX, t(6;13)(q21;q12).²⁰

Patient 2 has also previously been described in detail.²⁶ He had a phenotype very similar to the first patient and both patients were described as having microcephaly, microphthalmia, ectrodactyly, and prognathism (MMEP).

Cell lines and DNA samples

Ethidium bromide treatment of a lymphoblastoid cell line of patient 1 was used to generate prometaphase spreads for FISH analysis.^27,28 DNA was isolated by high salt precipitation²⁹ and diluted to a concentration of 105 ng/μl.

RNA extraction and cDNA synthesis

Total RNA was extracted from lymphoblastoid cell lines (approximately 3 × 10⁶ cells) using TRIzol™ LS (Life Technologies) according to the manufacturer’s procedure. Samples were treated with DNase I/Amp (Life Technologies) for 15 minutes at room temperature and purified using the RNeasy Mini Kit (QIAGEN, Valencia, CA, USA). Approximately 3 μg of total RNA were reverse transcribed into cDNA using random hexamers (SuperScript™ Preamplification System, Life Technologies).

Fluorescence in situ hybridisation (FISH) analysis

PAC and BAC clones were used as FISH probes. The genomic clones were selected according to their microsatellite marker content, their genetic mapping position from the Cedar web site (http://cedar.genetics.soton.ac.uk/pub/chrom6/map), and their physical mapping position on the Whitehead radiation map (http://www-genome.wi.mit.edu/). Microsatellite markers mapping to chromosome 6q21 were used to screen the “non-redundant” (nr) database and the “high throughput genomic sequences” (htgs) database from the National Center of Biotechnology Information (NCBI), using BLASTN (http://www.ncbi.nlm.nih.gov/blast).

Human genomic BAC and PAC clones were purchased from Research Genetics (Huntsville, AL, USA) and DNA was isolated using the QIAGEN^™ plasmid maxi kit (QIAGEN, Valencia, CA, USA). Purified BAC or PAC DNA was labelled with digoxigenin-11-dUTP (Boehringer Mannheim Biochemical, Indianapolis, IN, USA) by standard nick translation procedures. FISH analysis was performed as described by Brkanac et al.³⁰

Sequencing of PCR products

PCR products were gel purified for direct sequencing with gene specific primers using the Thermosequenase Cy™5 Dye Terminator kit (Amersham Pharmacia Biotech, Piscataway, NJ, USA) according to the manufacturer’s procedures.

Southern blot analysis

SNX3 cDNA was amplified by RT-PCR from mRNA prepared from fresh lymphocytes, using cDNASNX3 primers (table 1). SNX3 exons 3 and 4 and their flanking intronic sequences were amplified from genomic DNA using primers designed for mutation analysis (table 1). BAC clone RP11-427M20 (GenBank accession No AC079380) was used as a basis for “ex4dup” primers (table 1) and probe design. PCR products were gel purified using QIAquick gel extraction kit; 25 ng were labelled with 50 μCi at 3000 Ci/mmol [α ³²P]-CTP (NEN, Boston, MA, USA), using the random prime labelling system Redprime™ II (Amersham Pharmacia Biotech, Piscataway, NJ, USA ). Southern blot analysis was performed as described previously.³¹

Densitometry

Densitometric analysis of Southern blot filters hybridised with SNX3 probes was performed as previously described.³² A single copy probe from chromosome 4, D4S12, was hybridised to each filter as a control. Hybridisation signals of both D4S12 and SNX3 gene specific probes were measured using the Molecular Dynamics 300A Computing Densitometer (Amersham Pharmacia Biotech, Piscataway, NJ, USA). Calculations of the SNX3/D4S12 probe ratios allowed comparisons between the controls and the patients, facilitating detection of deletions.

Inverse PCR

Inverse PCR was performed as described previously³¹ using “IPCR primers”, corresponding to the “complementary” primers of SNX3 exon 4 (table 1).

SNX3 IMAGE clone

A SNX3 Unigene cluster, Hs.12102, from the NCBI database (http://www.ncbi.nlm.nih.gov/unigene) allowed identification of SNX3 IMAGE clones No 1084159 and No 1977743 that were purchased from Research Genetics (Huntsville, AL, USA). IMAGE clone No 1977743 was sequenced using M13 universal Cy^TM5 labelled primers with the Thermosequenase fluorescent labelled primer cycle sequencing kit with 7-deaza-dGTP (Amersham Pharmacia Biotech, Piscataway, NJ, USA). Internal primers “nestedSNX3” (table 1) were designed and used for sequencing of the entire 1.6 kb insert. Alignment between the known SNX3 cDNA sequence (GenBank accession No NM_003795) and the IMAGE clone No 1977743 sequence was performed using the program Windows 32 SeqMan 4.05^© (DNASTAR Inc, Madison, WI, USA).

Mutation detection by dHPLC analysis

Individual exons of the SNX3 gene, along with flanking intronic sequences, were amplified and tested by dHPLC analysis as described by Han et al.³³

RESULTS

Mapping of the 6q21 breakpoint

In order to localise the breakpoint in band 6q21 of the t(6;13) translocation, we conducted a systematic series of FISH analyses. As shown in fig 1A, we could narrow the breakpoint region to less than 1 cM using seven BAC/PAC probes. In order to cover the area between clone RP3-429G5 (positive for marker D6S1594) that mapped to the der(6) and clone RP3-354J5 (positive for marker D6S1250) that mapped to the der(13) chromosome, we constructed a BAC/PAC contig map (data not shown). This contig was then used for further FISH analysis of the region of interest. In summary, two clones mapped proximal, nine mapped distal, and two clones (RP11-183F17 and RP11-144P8) hybridised to both derivative chromosomes in addition to the normal chromosome 6 (fig 1B). Since these two clones overlapped by 40 kb, the breakpoint location was narrowed to this overlapping region. Interestingly enough, FISH analysis using clones RP1-128O3 and RP3-429G5, which covered the same region, did not result in signals on both derivative chromosomes (fig 2). This might be explained by the fact that they only overlapped by 14 kb.

Molecular cloning of the 6q21 breakpoint

In order to detect a candidate gene mapping to the 40 kb breakpoint region, a genomic sequence analysis was performed by searching the NCBI database (http://www.ncbi.nlm.nih.gov/), which identified two already published genes: tailless (TLX), which is located in clone RP3-429G5, centromeric to the overlapping region, and sorting nexin 3 (SNX3), which exhibited homology to both RP3-429G5 and RP1-128O3 and was thus included in the overlapping region that spanned the breakpoint (fig 1B). The genomic structure of the SNX3 gene was determined by alignment between the known SNX3 cDNA sequence (GenBank accession No NM_003795) or the IMAGE clone No1977743 and the PAC clone RP1-128O3 using BLASTN 2 sequences (http://www.ncbi.nlm.nih.gov/gorf/bl2.html) (table 2). The SNX3 gene is composed of four exons that encompass 50 kb. Southern hybridisation with a SNX3 cDNA probe had shown a different pattern between the translocated patient and the controls. To determine precisely which exon was contained in the shifted DNA fragments, we hybridised with probes containing first exon 3 (fig 3A) and then exon 4 (fig 3B). Since both exons 3 and 4 hybridised to two different fragments in patient 1, but normally hybridise to a single DNA fragment for all the enzymes tested, the translocation breakpoint occurred somewhere between exons 3 and 4. Isolation of the 1.4 kb PstI fragment (fig 3B) by Inverse-PCR using exon 4 specific primers allowed us to sequence the chromosome 6 breakpoint. A BLASTN search (http://www.ncbi.nlm.nih.gov/blast/) against the “non-redundant” database confirmed the presence of SNX3 exon 4 and its adjacent intronic sequence in the shifted fragment and showed that the breakpoint occurred within a SINE/Alu repeat, 222 bp upstream from exon 4 (fig 4). This result was consistent with the fact that clones RP3-429G5 and RP1-128O3 did not give a split signal when hybridised to the derivative chromosomes. The signal was too weak to be detected since only 7 kb of RP3-429G5 hybridised to der(13) and only 7.8 kb hybridised to der(6) with probe RP1-128O3.

Molecular cloning of the 13q12 breakpoint

The BLASTN search against the “htgs” database also identified a BAC clone RP11-347L8 (GenBank accession No AL137250) containing several markers specific to chromosome 13q12. This placed the breakpoint on chromosome 13 close to marker SHCG-102422. Since the BAC clone sequence was deposited in the database, we could design primers from the flanking sequences of each chromosome at the breakpoints and amplify the der(13) chromosome breakpoint region. Sequencing of the PCR product showed that no material was deleted and that a two base duplication (TA) from the chromosome 6 Alu repeat sequence was present at both breakpoint sites (fig 4).

Characterisation of the human SNX3 gene

A 1201 bp mRNA sequence of the SNX3 gene (GenBank accession No NM_003795) was reported by Haft et al,³⁴ which encoded for a protein of 162 amino acids. Although SNX3 was widely expressed (NCBI Unigene database, Hs.12102), it was more strongly expressed in heart, skeletal muscle, and spleen.³⁴ A SNX3 Unigene cluster, Hs.12102, comprising 241 ESTs entries, included an IMAGE clone No 1084159 from a colon tumour library. However, this clone of 2.4 kb could not be characterised any further as it failed to grow, but sequencing of another IMAGE clone (No 1977743) from a brain library, with a 1.52 kb insert, showed a novel isoform, which we named SNX3.2. The isoform appears to have originated from use of a cryptic splice site within exon 1 that eliminates 22 amino acids (fig 5). In addition, the 3′UTR was extended by 288 bp with another putative polyadenylation site and the 5′UTR was extended by 135 bp upstream of the known sequence. Hybridisation of Southern filters with exon 4 of SNX3 had shown an extra fragment in the controls and patients with all three enzymes that was not expected from the genomic sequence of clones RP1-120O3 or RP3429G5 (fig 3B). Hybridisation of CEPH random subjects with SNX3 exon 4 showed the same fragments and a polymorphism of at least three alleles (data not shown). A BLASTN search against the htgs database (NCBI) showed 95% homology over 117 bp between exon 4 probe and BAC clone RP11-427M20. This 117 bp sequence was the only homology found between the SNX3 cDNA and this BAC clone. This clone was mapped to chromosome 4 by identification of several markers including D4S1575, using the electronic PCR from NCBI. The size of one of the extra fragments was consistent with the predicted restriction map from BAC clone RP11-427M20. According to the database sequence, the exon 4 homologous sequence on chromosome 4 was interrupted by a LINE element, which may cause this region to be unstable, possibly generating the three fragments shown by Southern hybridisation.

Mutation analysis of patient 2

To further our analysis of the SNX3 gene, we screened patient 2²⁶ for alterations of the gene as he had the MMEP phenotype and did not have any cytogenetically visible rearrangements. Amplification of all four SNX3 exons and their flanking sequence was used for dHPLC analysis and sequencing. No differences in dHPLC patterns were detected between a control and patient 2. Also, we sequenced the four exons of the SNX3 gene for patient 2 and failed to detect any point mutation. Southern analysis detected an extra fragment with the SNX3 exon 4 probe, which varied in size depending on the enzyme used for digestion. However hybridisation with a chromosome 4 specific probe “ex4dup” showed that all of the extra fragments originated from chromosome 4 and were present in normal subjects (data not shown).

DISCUSSION

In this study, we were able to map both breakpoints of a translocation t(6;13)(q21;q12) in a patient with MMEP. Using FISH analysis, we could map the breakpoint region on chromosome 6 between D6S1594 and D6S1250, while the der(13) breakpoint was subsequently mapped close to SHCG-102422. Molecular cloning of both breakpoints confirmed the cytogenetic karyotype previously reported by Viljoen and Smart²⁰ and also confirmed that the translocation was balanced.

The sorting nexin 3 (SNX3) gene was found to be disrupted within intron 3 by the der(6) breakpoint. Therefore, we suspect that the disruption of the SNX3 gene underlies the MMEP phenotype in patient 1. However, no alteration of the SNX3 gene could be detected in patient 2, a sporadic case with the MMEP phenotype but with no cytogenetically detectable rearrangement. It is possible that a mutation in the promoter region of the SNX3 gene could still be responsible for the phenotype, as we have not screened this portion of the gene. Also, an inversion of the gene could have been missed by Southern analysis if one or both breakpoints had occurred within an intron or involved the promoter region only. Both scenarios could result in haploinsufficiency as in the patient with the translocation.

When a chromosome breakpoint does not directly disrupt the disease causing gene, it is thought to alter the gene expression by a phenomenon called position effect, for which the mechanism is still unknown. Several hypotheses have been suggested and reviewed by Kleinjan and van Heyningen.³⁵ (1) The chromosomal rearrangement could separate the promoter from a distant regulatory element. (2) The chromosomal rearrangement may juxtapose a gene with a regulatory element from another gene. (3) The chromosomal rearrangement may bring a gene and its regulatory element close to another gene generating a competition for the regulatory element between the two genes. (4) And finally the rearrangement could give rise to position effect variegation (PEV), a phenomenon first described in Drosophila³⁶ and later reported in mammalian systems.^37,38 Several translocation cloning projects have shown that a gene causing a disorder is not always disrupted within the coding sequence³⁵ or that, even though a gene is disrupted, it can still be another gene mapping proximal or distal to the breakpoint that causes the phenotype.³⁹ Therefore, the MMEP phenotype in patient 1 could in fact be cause by the impaired expression of another gene or by contiguous genes in the vicinity of the der(6) breakpoint. The necessity of altering the expression of two genes in order to underlie the MMEP phenotype would be consistent with its rarity. Patient 2 could possibly carry an inversion involving contiguous genes, while in patient 1 SNX3 is disrupted by the breakpoint, and perhaps another gene is affected by position effect.

SNX3 is a member of the sorting nexin family, composed of at least 14 other members (NCBI database).^34,40 None of the sorting nexin family members has yet been associated with any human disorder. Sorting nexin family members are identified by a conserved phox homology (PX) domain of about 100 amino acids, first characterised in NADPH oxidase, p47phox and p40phox.^41,42 Function of the SNX3 protein remains to be characterised to determine its putative role in development. The PX domain is highly conserved through evolution, therefore suggesting a similarity in biological function between species.³⁴ The yeast Grd19p is the most closely related in sequence to the human SNX3. Grd19p was shown to maintain late Golgi enzymes in their proper locations by retrieving mislocalised molecules from the prevacuolar compartment and Grd19p is not required for vacuolar protein sorting, unlike the other studied yeast sorting nexin proteins.⁴³ Study of the SNX3 function may help in deciding if SNX3 is or is not a good candidate gene for this disorder. SNX3 knockout mice would be most useful in providing this information.