The ATM gene is responsible for the autosomal recessive disorder ataxia-telangiectasia (A-T), characterized by cerebellar degeneration, immunodeficiency and cancer predisposition. A-T carriers were reported to be moderately cancer-prone. A wide variety of A-T mutations, most of which are unique to single families, were identified in various ethnic groups, precluding carrier screening with mutation-specific assays. However, a single mutation was observed in 32/33 defective ATM alleles in Jewish A-T families of North African origin, coming from various regions of Morocco and Tunisia. This mutation, 103C→T, results in a stop codon at position 35 of the ATM protein. In keeping with the nature of this mutation, various antibodies directed against the ATM protein failed to detect this protein in patient cells. A rapid carrier detection assay detected this mutation in three out of 488 ATM alleles of Jewish Moroccan or Tunisian origin. This founder effect provides a unique opportunity for population-based screening for A-T carriers in a large Jewish community.
Ataxia-telangiectasia (A-T) is a multisystem autosomal recessive disorder characterized by progressive cerebellar degeneration, developmental defects in specific tissues, immunodeficiency, chromosomal instability, acute predisposition to lymphoid malignancies and radiosensitivity (reviewed in refs 1–3). The cellular phenotype of A-T suggests defective regulation of cell cycle progression and increased apoptosis in response to specific types of DNA damage (2,4). A more general defect in responding to oxidative stress was recently proposed (5). A-T has attracted significant public attention due to reports of increased frequency of malignancies, particularly breast cancer, among A-T carriers (6–8).
The gene responsible for A-T, ATM, was recently cloned by Savitsky et al. (9,10) using a positional cloning approach. The ATM gene spans 150 kb on chromosome 11q22–23, and is transcribed in all tissues tested into a large transcript of ∼13 kb with an open reading frame predicting a protein of 3056 amino acids (9–11). The ATM protein is a member of an expanding family of large proteins identified in several organisms ranging from yeast to mammals. All these proteins contain a C-terminal region resembling the catalytic domain of phosphatidylinositol 3-kinase (PI 3-kinase), and are involved in various cellular processes ranging from regulation of telomere length to responses to DNA damage, such as cell cycle arrest or DNA repair (reviewed in refs 10,12 and 13).
An extremely wide spectrum of A-T mutations has been observed in various ethnic groups. The great majority of these mutations are expected to yield inactive or dysfunctional protein due to truncations or large deletions (9,14–16). Because most of these mutations are unique to single families or to a small number of geographically clustered families, large population-based carrier detection cannot be performed using conventional mutation-specific assays. On the other hand, the identification of a founder effect in a specific community would provide such an assay for that group and yield valuable information on the role of ATM in cancer predisposition in the general population.
All Jewish A-T families in Israel originated in North African countries, most of them in Morocco (17). We report here a significant founder effect for A-T in this particular ethnic group.
A total of 17 Jewish North African A-T families have been identified to date. The parents in all of these families originated in North Africa, except for one Ashkenazic parent born in Poland. The 33 parents of North African origin came from 12 widely separated locations in Morocco (Fig. 1), and the cities of Tunis and Nabil in Tunisia.
Early in our search for the A-T gene, we generated a high-density map of microsatellite markers across the A-T region (18). One of these markers on the proximal portion of this region, D11S2179, was subsequently mapped within one of the introns of the ATM gene (9,10). All patients except the one with an Ashkenazic parent and one of Jewish Moroccan descent were homozygous for the same allele at this locus. Thus, 32/33 (0.97) of the affected chromosomes of Jewish North African origin contained this allele, while its frequency in the general population of this origin was 25/87 (0.29). These results strongly suggested a significant founder effect for A-T in this community.
Systematic screening of the open reading frame (ORF) of the ATM transcript in patient cells was performed using the restriction endonuclease fingerprinting (REF) method (19), as previously described (15). A sequence alteration indicated by REF very close to the 5′ end of the ORF turned out to be a C→T transition at nt 103 of the ORF (Fig. 2), resulting in a stop codon at position 35 of the ATM protein. No other sequence alterations were indicated by REF analysis over the entire ATM ORF in these patients. Genomic sequencing showed that all patients homozygous at D11S2179 were also homozygous for this mutation (32 chromosomes); again, only the affected chromosome with the different D11S2179 allele did not show this mutation. Interestingly, in this compound heterozygote patient, F-2016, the 103C→T mutation appeared in genomic DNA in the heterozygous state, while only the mutant sequence was detected in the ATM transcript using RT-PCR. We assumed that the other ATM allele of this patient bears a mutation that precludes its expression. Complete sequencing of the promoter region did not reveal any sequence alterations in this patient. Thus, the other mutant allele from F-2016 may not be expressed or may produce an extremely unstable transcript. Patients from five non-Jewish A-T families of North African origin did not show the nonsense mutation at codon 35, suggesting that this mutation might be unique to the Jewish community originating from that geographic region.
The 103C→T mutation is expected to truncate the ATM protein very close to its N-terminus, essentially eliminating its synthesis. Immunoblotting analysis was used to confirm this prediction. Antisera were raised in rabbits against four peptides derived from various regions across the ATM protein (see Materials and Methods), and two previously described antisera against other ATM epitopes (20,21) were also used. All six antisera detected the predicted band of ∼350 kDa in cellular extracts from normal lymphoblasts, but failed to detect this band in a cell line from a patient homozygous for the 103C→T mutation (Fig. 3A). Immunofluorescence analysis with the antiserum pAb 132 (20) yielded nuclear staining in normal fibroblasts, but failed to give a signal with fibroblasts from another homozygous patient (Fig. 3B).
The 103C→T mutation was predicted to eliminate an HinfI site within exon 5 of the ATM gene. This prediction was used to design a rapid assay for the presence of the mutation in genomic DNA. Primers flanking the mutation site were used to amplify a 254 bp fragment that contains three HinfI sites one of which is destroyed by the mutation, giving rise to a 216 bp band unique to the mutant allele (Fig. 4). This assay can be applied conveniently to buccal smear cells or dry blood spots obtained from Guthrie cards.
We obtained a preliminary estimate of A-T carrier frequency among North African Jews by screening a random sample of this community. Two hundred archival DNA samples and 56 Guthrie cards were examined. (The samples were provided by another laboratory without any personal identification, making it impossible to trace their origin.) Initial screening was performed using the HinfI-based assay, and positive samples were subjected to genomic sequencing in order to verify the presence of the 103C→T mutation. This sample contained 488 ATM alleles of Jewish Moroccan or Tunisian origin, of which three presented the 103C→T mutation, indicating a frequency of ∼0.6% of the mutant allele in this community.
The classical A-T phenotype is caused largely by ATM null mutations (9,14–16). The 103C→T mutation identified in North African Jews falls under this category, eliminating the ATM protein (Fig. 3). Truncation mutations further downstream have recently been found to eliminate the appearance of any protein product of the ATM gene, indicating that most truncated derivatives of this protein are probably unstable (20,21 and unpublished results). No evidence has been obtained in those studies and in the present one, for ATM isoforms created by alternative initiation or alternative splicing within the ORF. Thus, complete absence of any protein product of the ATM gene is a common phenotype in A-T cells. The clinical phenotype of A-T patients of Jewish North African origin is also typical (17 and unpublished clinical observations).
A-T is characterized by a wealth of unique mutations spread throughout the ORF of the ATM transcript (14–16). A local founder effect was reported by Gilad et al. (15) in six Italian A-T families, who share a common mutation, all living in the Naples region. Telatar et al. (16) reported several mutations common to two or three North American A-T families, some with common origin. A haplotype common to British patients with a milder variant of A-T (22), and a haplotype common to 24/54 mutant chromosomes in the Costa Rican population (23) may flag mutations common to those patients. The founder effect reported here is the most significant one identified to date for A-T in a specific ethnic group. It approximates the founder effect observed in Yemenite Jews with phenylketonuria, where all patients are homozygous for a single mutation unique to this population (24).
A different mutation was found in only one ATM allele, in patient F-2016. Since none of the other patients showed this mutation, and the patient's parents are from the same village, we believe the mutation probably appeared in the patient's father or in a previous generation of this family. The probability of introduction of this mutation to the family by intermarriage is extremely remote in view of the religious nature and the well recognized ethnic isolation of this community.
The 103C→T mutation probably appeared in the North African Jewish community, since it was not found in five non-Jewish A-T families from that region. The relatively large size of the haplotype common to most of the mutant alleles, ∼3 Mb (unpublished data), indicates that the mutation may be relatively recent. Based on the names of the affected families and the presence of the same ATM allele in two Tunisian Jewish families, they probably descended from a founder individual who lived in Spain. When the Jewish community was expelled from that country at the end of the 15th century, many families settled in Morocco and Tunisia. The wide geographic distribution of the affected families in Morocco (Fig. 1) also argues for the existence of this mutation for at least several hundred years.
Genetic disorders within defined ethnic groups may show strong founder effects (24–26), or may have multiple origins even within small geographic areas (27,28). Factors influencing these patterns include the mutation rate for specific genes, the age of the examined mutations, and the degree of isolation of the investigated community from other ethnic groups. Tay-Sachs and phenylketonuria, for example, have several origins among Moroccan Jews (29,30), while a significant founder effect for Familial Mediterranean Fever is predicted in this community by haplotype analysis (31). In many cases, Jewish communities that lived for many generations in various parts of the world, relatively isolated from each other and from the surrounding population, have specific sets of mutations (32).
The variety of ATM mutations in North American and European A-T families certainly poses serious obstacles to general carrier detection tests for these populations. The instability of truncated ATM proteins probably precludes western blotting analysis as a rapid assay for carrier detection, leaving general, sequence-based methods, which might be tedious given the size of the ATM transcript. The founder effect reported here should enable carrier screening in a community of ∼500 000 individuals. The ability to identify A-T carriers in this population will facilitate genetic counseling and enable epidemiologic follow-up of identified carriers to correlate any carrier susceptibility risks. Adequate counseling and support should accompany such testing in the event that a correlation is found.
Materials and methods
Identification of ATM mutations
RT-PCR followed by restriction endonuclease fingerprinting (REF) analysis was used to search for sequence alterations in the ATM transcript as described in detail by Gilad et al. (15). When the REF pattern indicated a sequence alteration, the region spanning that site was reamplified from mRNA and genomic DNA and the products were sequenced directly (15).
Antibodies against the ATM protein
Four ATM peptides were selected as immunogens on the basis of hydrophilicity profiles and predicted surface probability, using algorithms from the software package ‘GeneWorks’ from Intelligenetics: ATM-ORP3 (positions 13–24 of the ATM protein); ATM-PRP5 (positions 368–380), P2R2 (positions 1299–1313) and ATM-ORP11 (positions 1353–1366). Peptides were synthesized by standard methods and purified by reverse phase chromatography. Purified peptide (10–20 mg) was conjugated to keyhole limphet hemocyanine (KLH) and used to immunize New Zealand White Rabbits. Rabbit sera were tested for antigen reactivity by capture ELISA using free peptides, purified over Protein A-agarose and dialyzed against PBS. The antisera pAb 132 (20) and ATM-3BA (21) were raised against peptide representing ATM positions 819–844 and 2929–2958, respectively.
107 cells were rinsed twice with cold phosphate-buffered saline and lysed in 150 µl of solution consisting of 150 mM NaCl, 20 mM HEPES, 1% NP-40, 0.1% SDS, 10% glycerol and protease inhibitors (Boehringer-Mannheim, Germany). The lysates were left on ice for 30 min and spun at 15 000 g 10 min at 4°C. The supernatant was mixed with Laemmli's loading buffer (33), and the mixture was heated at 65°C for 10 min before loading onto a 6% polyacrylamide gel (100–150 µg protein/lane). Following electrophoresis with molecular weight standards (Rainbow Coloured Molecular Weight Marker, Amersham), the proteins were electroblotted to nitrocellulose membranes (Pharmacia) for 16 h 300 mA in transfer buffer (50 mM Tris-HCl, 384 mM glycine, 20% methanol). The blots were stained by soaking for 1 min in 0.2% Ponceau (Sigma), 3% trichloroacetic acid, 3% sulfosalicylic acid, in order to visualize the protein bands. Prior to reacting them with antibodies, the blots were blocked for 6 h at room temperature in 5% BSA (Sigma) in TTBS (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.1% Tween). The primary antibody was added for 2 h at room temperature or overnight at 4°C. The blots were rinsed six times, 5 min each time, with 0.5%BSA in TTBS. A horseradish peroxidase conjugated secondary antibody was then added for 1 h, and the blots were washed five times with 1%BSA, TTBS, twice with 1% NP-40, 0.1% SDS, TTBS, and twice with TTBS, all at room temperature. Immunoreactive bands were visualized using an ECL system (western Blotting Detection Reagents, Amersham).
This analysis was performed as described before (19). Briefly, fibroblasts were cultured on sterilized glass coverslips, fixed in 3% formaldehyde and permeabilized with 0.1% Triton-X100. Following blocking, the coverslips were incubated with affinity purified pAb 132 in PBS/5% BSA for 30 min at 37°C, rinsed extensively with PBS and incubated with FITC-labeled goat anti-rabbit (Kirkegaard and Perry Laboratories, Gaithersburg, MD). Prior to mounting, some of the slides were counterstained with Hoechst 33258 to visualize the DNA. Cells were viewed with a Zeiss Optihot microscope and images captured with a Photronix color CCD camera.
Carrier detection assay
Amplification of a 254 bp fragment containing the 103C→T mutation from genomic DNA was performed using the primers 5′-GTGTGTTCTGAAATTGTGAACC-3′, and 5′-CCTTGTTTGGAATCTGAATG-3′, for 30 cycles at an annealing temperature of 58°C. The products were purified using the Qiaquick PCR Purification System (Qiagen, Hilden, Germany). The purified fragment was digested with 2 U/µg HinfI (New England Biolabs), and the reaction products were separated by electrophoresis on 3% NuSieve/1% agarose gels (FMC).
This assay was applied to dry blood spots obtained from Guthrie cards. One square millimeter was cut from the blood spot, washed twice with distilled water for 30 min each time, and added to 100 µl of PCR mixture containing 0.2% gelatin.
The authors are grateful to Drs Francis S. Collins and Aravinda Chakravarti for helpful discussions, to Dr Dianne Watters for developing the ATM-3BA antibody, and to Drs Gilbert Lenoir and Marc Fellous for cells and samples of non-Jewish Moroccan A-T patients. This study was supported by research grants from the A-T Children's Project, the A-T Medical Research Foundation, the Thomas Appeal (A-T Medical Research Trust), The United States-Israel Binational Science Foundation and the National Institute of Neurological Disorders and Stroke (NS31763). This work was carried out in partial fulfillment of the requirements for the PhD degree of S.G.