Molecular analysis of the reciprocal chromosomal translocation t(12;22)(p11.2;q13.3) cosegregating with a complex type of synpolydactyly showed involvement of an alternatively spliced exon of the fibulin-1 gene (FBLN1 located in 22q13.3) and the C12orf2 (HoJ-1) gene on the short arm of chromosome 12. Investigation of the possible functional involvement of the fibulin-1 protein (FBLN1) in the observed phenotype showed that FBLN1 is expressed in the extracellular matrix (ECM) in association with the digits in the developing limb. Furthermore, fibroblasts derived from patients with the complex type of synpolydactyly displayed alterations in the level of FBLN1-D splice variant incorporated into the ECM and secreted into the conditioned culture medium. By contrast, the expression of the FBLN1-C splice variant was not perturbed in the patient fibroblasts. Based on these findings, we propose that the t(12;22) results in haploinsufficiency of the FBLN1-D variant, which could lead to the observed limb malformations.
- limb development
- Approved gene symbols were obtained from the HUGO Nomenclature Committee: FBLN1 = human fibulin-1 gene
- FBLN1 = human fibulin-1 protein
- fbln1 = mouse fibulin-1 gene
- fbln1 = mouse fibulin-1 protein
- ECM, extracellular matrix
Statistics from Altmetric.com
- Approved gene symbols were obtained from the HUGO Nomenclature Committee: FBLN1 = human fibulin-1 gene
- FBLN1 = human fibulin-1 protein
- fbln1 = mouse fibulin-1 gene
- fbln1 = mouse fibulin-1 protein
- ECM, extracellular matrix
The formation of digits is a complex, tightly regulated process which involves apoptosis and in which several pathways and signalling molecules are involved. The presentation of a small family with a very specific congenital hand malformation associated with an apparently balanced autosomal reciprocal translocation provided us with the opportunity to identify novel genes involved in limb development. Previously, we described the clinical phenotype of three patients from one family sharing a complex type of synpolydactyly (fig 1) associated with a constitutional balanced t(12;22)(p11.2;q13.3) and the exclusion of synpolydactyly associated HOXD13 mutations in these patients.1–3 All affected subjects carry the constitutional balanced translocation, whereas unaffected family members have a normal karyotype. The cosegregation of the phenotype with this translocation suggested that a synpolydactyly associated gene might be localised at one or both of the chromosomal breakpoints involved in this translocation. As a first step towards the identification of the genes involved in the translocation, we were able to pinpoint the breakpoint region on chromosome 12 to a 2.6 kb BamHI fragment.4
Here, we describe the cloning of the translocation breakpoint and the identification of two genes directly targeted by this translocation. One of these genes, FBLN1, turned out to be the ideal positional and functional candidate for the observed phenotype. We therefore examined the consequences of the translocation induced disruption on the expression of FBLN1-D, one of the four variants of FBLN1, in the possible genesis of the limb malformations observed in our patients.
MATERIALS AND METHODS
Vectorette PCR experiments were performed essentially as described by Schoenmakers et al.5 Genomic DNA was digested with BamHI and a vectorette linker with BamHI compatible cohesive ends was ligated to the digested DNA. This ligated DNA was used in two consecutive rounds of (hemi-nested) PCR amplification. In a first round a universal vectorette primer (5′-CGAATCGTAACCGTTCGTACGAGAATCGCT-3′) was used in combination with a specific primer (5′-CTTGAAGGAGAGAGGCAGGGAAACTGAAGTGG-3′), based on sequence data obtained during random sequence sampling, and mapping within the rearranged 2.6 kb BamHI fragment on chromosome 12p. Part of the DNA produced in this first round was then reamplified using the same universal vectorette primer and the following (nested) chromosome 12 specific primer (5′-GGTAATGCATGCTACATGGTGGCTTCA GGG-3′). PCR products were analysed on, and recovered from, standard 1% agarose gels and subcloned using the pGEM-T Easy system (Promega).
Sequencing of the wild type FBLN1-D and construction of FBLN1 probes
The sequence of the presumed wild type FBLN1-D was determined as follows: first, the wild type FBLN1-D cDNA was isolated with the FBLN1-D specific cDNA synthesis primer (5′-CCACCTATAAGAAAAGAT-3′). This primer is located downstream of the translocation breakpoint, so only the presumed wild type transcript, that is, originating from the non-disrupted gene, was isolated. Subsequently, overlapping fragments covering the whole FBLN1-D cDNA were amplified by PCR and subcloned in the pGEM-T Easy vector.
The following primer sets were used to amplify the overlapping cDNA fragments: RT1-up: 5′-CTCATTTTTTAATGCG AAGGCTAAG-3′; RT1-low: 5′-AAGAAA AGATGGATGATG CAGAGTG-3′; RT2-up: 5′-TGCGGGACTCTTTTGACATCATC-3′; RT2-low: 5′-TCTTTCTTTCCTTGGCCCTTAGC-3′; RT3-up: 5′-CGC AACTGCCAAGACATTGATGA-3′; RT3-low: 5′-CAGCTT CAGGACGGCATGAAATG-3′; RT4-up: 5′-AGAACACGCTGGG CTCCTACCTC-3′; RT4-low: 5′-CGGACCGTGTCTGTCTTCT CCTG-3′; RT5-up: 5′-GCCCTATTGGGCATACATGCATC-3′; RT5-low: 5′-CTCGCCGGCAGTAACACTGGTAG-3′; RT6-up: 5′-GACGAGGTGGTCTGCTCCTGCTT-3′; RT6-low: 5′-CGCACT CGTCCACATCAACACAG-3′; RT7-up: 5′-TCAAGAGCC AGGAGACCGGAGA-3′; RT7-low: 5′-CTGGTAGCCCACGAA GCAGGAG-3′; RT8-up: 5′-CCCGCCGCCCATGGAGCGCGC-3′; RT8-low: 5′-CGGCACTGCTGCTTGCAGGGCCCGC-3′.
Annealing temperatures of 60°C were used for all primer combinations. DMSO (10% final concentration) was added when using sets RT5, RT7, and RT8. Sequencing was performed using standard protocols and T7 and SP6 primers.
FBLN1 probes for northern blot analysis were constructed by PCR. The FBLN1-D specific probe was amplified using primers 5′-ACCATCTCCCACA-3′ (position 1793-1805 in FBLN1-D mRNA) and 5′-CAGCAATGATTTG-3′ (position 2162-2174). Similarly, a FBLN1-C specific probe was constructed using primers 5′-CCAAGCTGCCTCTGAGAATA-3′ (position 1749-1768 in FBLN1-C mRNA) and 5′-GCTCTGCAGACACAAAGATG-3′ (position 2038-2057). For the probe residing within the part common to all FBLN1 variants, primers RT5-up and RT5-low were used.
3′ RACE analysis
3′ RACE was performed according to the method described by Schoenmakers et al6 with minor modifications. RNA used for the 3′ RACE analysis was derived from skin fibroblasts. First strand cDNA synthesis was performed using the poly (A) specific primer AP2 (5′-AAGGATCCGTCGACATC(T)17-3′) or a FBLN1-D specific primer (5′-CCACCTATAAGAAAAGAT-3′). Specific 3′ RACE primers were designed on the published FBLN1-D sequences. In the first PCR round, a specific FBLN1-D primer (5′-CCAACGATGTCACATGCGTGT-3′, position 1760-1780 in FBLN1-D mRNA), located in the first FBLN1-D specific exon (exon 18) was used in combination with the AP2 specific primer UAP2 (5′-CUACUACUACUAAAGGATCCGTCGACATC-3′). In the second round, a tailed, nested primer (5′-CAUCAUCAUCAUAAGGATCCGTCGACA-3′, position 1795-1808), also situated within the first FBLN1-D specific exon (exon 18) was used in combination with the UAP2 primer.
Detection of fibulin-1 in conditioned culture medium and extracellular matrix of cultured fibroblasts
Age matched, normal foreskin fibroblasts (ATCC CRL 2056, Rockville, MD) and synpolydactyly fibroblasts isolated from skin biopsy (obtained during surgery after informed consent) were grown in 100 mm tissue culture plates (Corning, Corning, NY) in Dulbecco's Minimal Essential medium (DMEM) supplemented with 10% iron supplemented bovine calf serum (Hyclone, Logan, UT) until they reached 90% confluency. The medium was then aspirated and replaced with DMEM supplemented with insulin, transferrin, and selinious acid (5 μg/ml, 5 μg/ml, and 5 ng/ml, respectively, Becton Dickinson Labware Bedford, MA). After an additional three days in culture, the conditioned culture medium (CCM) was collected and the cell layer washed with phosphate buffered saline (pH 7.4). The cell layer was then extracted with 10 mmol/l EDTA for five minutes with gentle agitation. The extraction buffer was collected and centrifuged for 10 minutes at 2000 × g. A 400 μl aliquot of the supernatant was concentrated with 10 μl of Strataclean resin (Stratagene, La Jolla, CA) by vortexing for one minute and centrifugation at 13 000 × g for 10 seconds. Similarly, a 500 μl aliquot of CCM was absorbed with Strataclean resin and 10 μl of 2 × Laemmli sample buffer containing β-mercaptoethanol was added to each sample and extracts were run on SDS-PAGE using 4-20% gels (Invitrogen/Novex, Carlsbad, CA). After SDS-PAGE, proteins were transferred to PVDF membrane and probed with antibodies against FBLN1 (1 μg/ml) in Tris buffered saline (pH 7.4) containing 0.1% Tween 20 and 5% non-fat milk. Bound antibody was detected using horseradish peroxidase (HRP) labelled secondary antibody (1:5000 dilution, Amersham Pharmacia, Piscataway, NJ) and ECL+ detection system (Amersham Pharmacia). The mouse monoclonal antibody 5D12 (mAb5D12) was produced as described previously.7 The epitope of mAb5D12 antibody maps to the C-terminal region of FBLN1-C (residues 567-683). The rabbit polyclonal antibody (rb811) was raised against the C-terminal region of FBLN1-D (residues 567-703).8
The 5′UTR of C12orf2 is disrupted by the translocation
Using a vectorette-PCR mediated approach, we isolated a 1.2 kb breakpoint encompassing BamHI fragment. Sequencing of the insert showed ectopic sequences fused to known sequences (obtained during the random sequence sampling of chromosome 12p derived cosmid LL12NCO1-149H4 which spans the chromosome 12 breakpoint).4 In fig 2B the sequence of the breakpoint region on the der(12) is depicted. Further analysis of the chromosome 12 derived sequences showed that the breakpoint on chromosome 12p was located between two exons belonging to a cDNA contig constructed starting from the anonymous EST R72964, which in its turn was isolated with STS PD108.4 EST R72964 corresponds to a gene designated C12orf2 located at 12p, also known as HoJ-1 (GenBank accession number U82396). In silico mapping of the C12orf2 cDNA, showed that the breakpoint is located in the 5′UTR (fig 2A). At least four alternatively spliced exons were detected by sequencing intron-exon boundaries using cDNA derived sequencing primers (data not shown). Northern blot analysis showed that C12orf2 is widely expressed as a 6.2 kb transcript.4 A 2.2 kb alternatively spliced transcript is also expressed exclusively in testis.
FBLN1-D is disrupted by the translocation
A BLAST search conducted with the ectopic sequences (210 bases) from the 1.2 kb vectorette clone showed a 100% sequence identity with BAC CTA-941F9 (GenBank accession number Z95331), which had previously been assigned to chromosome 22q13.3.9 BAC 941F9 was used in FISH experiments on metaphase spreads from the patients, which confirmed its localisation across the breakpoint (data not shown). To exclude the possibility that additional deletions flanking the breakpoint were present, we also determined the reciprocal breakpoint flanking sequences on the derivative chromosome 22. Sequencing of a 900 bp PCR product containing the reciprocal breakpoint and comparison with the wild type sequences obtained showed that the sequences from chromosome 12p were perfectly (that is, without any breakpoint associated deletions) fused to those derived from chromosome 22q. In this way, we could show that the translocation was truly reciprocal.
BAC CTA-941F9 contains part of the gene encoding the ECM protein FBLN1. The fibulin-1 gene consists of 20 exons. Four different fibulin-1 variants result from alternative splicing of exons 15-20 (FBLN1-A-D, GenBank accession numbers X53741, X53742, X53743 and U01244 respectively). Exact positioning of the breakpoint within the BAC showed that it was located in the intron between the last two specific exons of the FBLN1-D isoform (exons 19-20) (fig 2A). In order to investigate whether the alterations in FBLN1-D were restricted to one allele, we also sequenced the wild type FBLN1-D, but no mutations could be identified.
Northern blot analysis and 3′ RACE
The chromosomal orientation of the two genes, which are both widely expressed, was such that, theoretically, the reciprocal translocation could not lead to the formation of reciprocal fusion transcripts between (sense) C12orf2 and (sense) FBLN1-D (fig 2C). This was confirmed by Northern blot analysis and 3′ RACE. Northern blot analysis of RNA from patient skin fibroblasts was performed using probes specific for C12orf2 (probe PD108), the FBLN1-C and D isoform, and a probe common to all FBLN1 isoforms (fig 3). From these Northern blot experiments it can be concluded that compared to FBLN1-D expression in normal control fibroblasts, the expression of FBLN1-D is clearly diminished in our patient. Since no clear differences in C12orf2 expression levels are apparent on this Northern blot, no firm conclusions can be drawn about possible alterations in the expression levels of this transcript. In these Northern blot experiments, we were also not able to identify aberrant FBLN1 or C12orf2 transcripts in fibroblasts from our synpolydactyly patients (data not shown). 3′ RACE analysis using the UAP-2 primer and two specific FBLN1-D primers also failed to show the presence of any fusion transcript.
Fbln1 is expressed in the developing limb
Human FBLN1 and mouse fibulin-1 (fbln1) are widely expressed during embryonic development.10,11 We used immunohistochemical staining, as described by Roark et al,12 to examine its expression in the developing mouse limb, focusing on the handplate region. As shown in fig 4, fbln1 is expressed in interdigital regions of the handplate of a ∼12 dpc mouse embryo and in the lateral perichondrial regions (fig 4A). Similar staining persists in the ∼13 dpc handplate, particularly in the perichondral regions and apical aspects of the developing digits (fig 4B).
FBLN1-D expression is altered in fibroblasts from a synpolydactyly patient
Skin fibroblasts derived from a t(12;22) synpolydactyly patient and a normal age matched control were grown in vitro and the level of FBLN1 incorporation into the ECMs of each was immunologically evaluated. Using antibodies specific for the C (mAb5D12) and D (rb811) variants, it was observed that FBLN1-D levels were significantly lower in extracts of the ECM of synpolydactyly fibroblasts as compared to normal fibroblasts (fig 5). By contrast, little or no differences were apparent in the level of FBLN1-C in the ECM of synpolydactyly fibroblasts as compared to normal skin fibroblasts. The immunological analysis also indicated that FBLN1-C and D differ in the way in which each are apparently incorporated into the ECM of skin fibroblasts. While full length FBLN1-D of ∼95 kDa is incorporated into the ECM, only proteolytic fragments (∼55 kDa) and multimeric forms (∼170 kDa) of FBLN1-C are incorporated into the ECM.
This paper describes the molecular cloning of the breakpoint of a t(12;22)(p11.2;q13.3) translocation involved in a novel complex type of synpolydactyly observed in three patients in one family. Using a positional cloning approach, we identified two genes targeted by this translocation, FBLN1 and C12orf2. C12orf2, also known as HoJ-1, is located on chromosome 12p12.3. The gene is widely expressed as a 6.2 kb transcript and a 2.2 kb transcript in testis. Interestingly, the protein encoded by C12orf2 has a 56% homology with B44478, a putative cell growth or differentiation regulator.13 Since, at present, there is no further information regarding the function of C12orf2, its role in the genesis of the observed phenotype cannot be completely excluded. Since we do not have data on C12orf2 expression at the protein level, it remains difficult to speculate whether or not the expression of C12orf2 is affected by this translocation.
FBLN1 encodes a calcium binding ECM and plasma glycoprotein. Four fibulin-1 variants, designated FBLN1A-D, all differing at their carboxyl termini, are encoded by alternatively spliced transcripts.7,8,14,15FBLN1-A and FBLN1-B are only expressed in human placenta, whereas FBLN1-C and FBLN1-D are expressed in a variety of tissues and cultured cells. FBLN1 has been shown to interact with nidogen, NOVH, fibronectin, laminin, aggrecan, versican, tropoelastin, and amyloid precursor protein.12,14–21 FBLN1 may also play a role in the regulation of cell migration22 as well as in haemostasis and thrombosis owing to its ability to bind fibrinogen and incorporate into clots.7,23,24 Expression studies in human and mouse embryos suggest a role for FBLN1 in organogenesis and embryonic development. Zhang et al11 showed fbln1 expression in the developing mouse limb. More specifically, fbln1 was expressed in precartilage condensations of the phalanges. FBLN1 expression in human embryos was analysed by Miosge et al.10 In human embryos of gestational week (gw) 4, FBLN1 staining was observed in the early mesenchymal bone anlagen. In embryos of gw 6.5 and 8, all perichondrial structures showed FBLN1 expression but the chondrocytes themselves showed no staining. Later in development, in embryos of gw 10, FBLN1 expression is prominent in the interterritorial matrix surrounding the hypertrophic chondrocytes. Our immunohistochemical stainings confirm that fbln1 is expressed in the developing handplate of mouse embryos. We observed fbln1 expression in the interdigital regions of a 12 dpc developing murine handplate and in the lateral perichondrial regions. In the 13 dpc handplate, fbln1 staining is more prominent in the apical aspects of the developing digits. The expression pattern of fbln1 in the developing handplate correlates with the phenotype observed in our patients in whom the deformities are also localised in the distal structures (metacarpals and phalanges) of the developing upper limb.
Our findings show that the t(12;22) translocation results in disruption of the “fibulin type module” of the FBLN1-D variant (fig 6). We investigated several obvious consequences of the translocation. Using RACE and Northern blot analysis no fusion transcripts were detectable in RNA preparations derived from patient fibroblasts so it is unlikely that the production of mutant proteins results in a dominant negative effect. Similarly, no aberrant polypeptides were detected in extracts of cultured patient fibroblasts. Therefore, we favour the possibility that the t(12;22) results in a haploinsufficiency for FBLN1-D. Haploinsufficiency has indeed been implicated in limb malformations. For example, deletions of a large part of the genes in the HOXD cluster result in monodactylous limbs and abnormal genitalia.25 Devriendt et al26 described a patient with hand-foot-genital syndrome, velopharyngeal insufficiency, and persistent patent ductus Botalli, associated with haploinsufficiency for the HOXA cluster. A submicroscopic deletion of chromosome 2q31 which removes the most 5′ HOXD genes and the adjacent EVX2 gene results in a limb phenotype similar to synpolydactyly.27 Other examples of hemizygosity for the HOXD cluster resulting from deletions of 2q31 associated with limb and genital malformations have been described.28,29 These observations show that the presence of both alleles of certain Hox genes is necessary for normal growth and limb patterning. Similarly, we speculate that a diploid dose of FBLN1-D is necessary for normal limb morphogenesis.
Our findings show that cultured fibroblasts from patients with the t(12;22) had significantly reduced levels of FBLN1-D polypeptide incorporated in the ECM and present in the conditioned culture medium as compared to control fibroblasts. By contrast, levels of FBLN1-C were similar both in the ECM and conditioned culture medium fractions from the patient fibroblasts. The latter results are consistent with the fact that the exon encoding the FBLN1-C specific carboxy terminal module (exon 17) is upstream from the D specific exon disrupted by the translocation (exon 20). Therefore, it is predictable that the production of the C specific variant might be spared by the mutation found in this family. The basis for the reduced incorporation of FBLN1-D in synpolydactyly fibroblasts is uncertain, but may be related to quantity, and, consequently, may be the result of haploinsufficiency. However, we cannot at present rule out the possibility that the mutation results in the production of a form of FBLN1-D that is more susceptible to proteolytic degradation than the wild type polypeptide. Taken together, the results from our analysis support the hypothesis that haploinsufficiency for FBLN1-D might have caused the abnormal phenotype. It must be noted, however, that in this context the term haploinsufficiency should be used with the understanding that only loss of the D isoform is considered since no changes were observed in the expression of FBLN1-C. In the 22q13 deletion syndrome (deletion of chromosomal band 22q13.3), the whole FBLN1 cluster is contained within the deletion interval. In this syndrome the hand malformations are very mild (ranging from syndactyly to fifth finger clinodactyly or relatively fleshy hands).30 In our patients, on the other hand, alterations in FBLN1-D expression alone seem to be responsible for the observed limb deformities. A plausible explanation for this phenomenon might be that there is a critical balance in the stoichiometry of the different FBLN1 isoforms. We believe that the translocation might result in a disturbance of this critical balance and lead to the observed phenotype.
The formation of the digits is a complex, tightly regulated process which involves apoptosis and in which several pathways and signalling molecules are involved. Solursh et al31–33 and Archer et al34 showed that the ECM plays an important role in determining the fate of cells in the interdigital region during morphogenesis. Surgical manipulations of the interdigital tissue of the chick embryo leg bud result in the formation of ectopic digits.35 Miner et al36 showed that mice lacking the laminin α5 chain exhibit multiple developmental defects, including exencephaly and syndactyly. Mutations in the fibrillin-2 gene, which results in congenital contractural arachnodactyly (CCA, MIM 121050) in humans, can cause syndactyly in mice.37–39 These data show that a “deficient” ECM can lead to anomalous limb development, both in humans and in mice. We believe that haploinsufficiency of FBLN1-D results in similar ECM changes and causes the observed synpolydactyly.
To our knowledge, this is the first human disorder linked to alterations in FBLN1. Corroboration of this conclusion requires that mutations in this gene are found in other families with complex synpolydactyly. In this regard, Percin et al40 have described a Turkish family with a complex type of syndactyly and synostoses. In this family, HOXD13 involvement was excluded. Furthermore, linkage analysis, using polymorphic markers of the region involved on chromosomes 12p and 22q turned out to be negative (A N Akarsu, personal communication; data not shown). Therefore, the genetic defects in the Turkish family and the family described in the present study appear different. Alternative proof that FBLN1 mutations may indeed cause hand and foot malformations may be obtained by the construction of a FBLN1-D specific knock-out mouse and by analysis of its phenotype. Recently, Kostka et al41 described the phenotype of fibulin-1 deficient mice. Heterozygous embryos developed normally and were phenotypically indistinguishable from wild type litter mates. The homozygous embryos all died in the perinatal period because of massive haemorrhages in several tissues. The cause of the early death was probably kidney and lung defects combined with abnormalities in some endothelial compartments. These mice were deficient for all the fibulin-1 isoforms and no limb deformities were mentioned. We are currently constructing a FBLN1-D specific knock-out in order to find out whether an isoform specific knock-out mouse develops limb abnormalities and whether a critical dose of FBLN1-D is essential for proper limb development. The involvement of FBLN1-D in a specific human disorder might also shed light on functions of the FBLN1 isoforms, which are currently poorly understood.
The authors would like to thank Christel Huysmans, Marleen Willems, and Reinhilde Thoelen for excellent technical assistance. The cooperation of all the family members is also greatly appreciated. This work was supported by NIH grants GM42912 and HL52813 to WSA.
Electronic database information. Accession numbers and URLs for data in this article are as follows: GenBank, http://www.ncbi.nlm.nih.gov/Genbank/index.html (for the fibulin1-A-D genes (accession numbers X53741, X53742, X53743, and U01244, respectively), the C12orf2 gene (accession number U82396), and BAC CTA-941F9 (accession number Z95331)). Online Mendelian Inheritance in Man (OMIM), http://www.ncbi.nlm.nih.gov/Omim/ (for CCA (MIM 121050)).
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.