Frequent allelic loss at chromosome arm 3p is distinct from genetic alterations of the Von-Hippel Lindau tumor suppressor gene in head and neck cancer

Previous molecular genetic studies revealed that allelic loss of chromosome arm 3p is a frequent event in upper aerodigestive tract squamous cell carcinoma (UADT SCC). Recently, the Von-Hippel Lindau (VHL) tumor suppressor gene was identified at chromosome band 3p25-26. To determine if the VHL locus is altered in these tumors, a paired series of 26 tumors and blood from patients with UADT SCC that were previously shown to exhibit allelic loss of 3p were tested for LOH surrounding the VHL locus using four different polymorphic markers. All of the samples (100%) exhibited LOH for at least 1 marker. However, no LOH was detected using a polymorphism within exon 1 of the VHL gene which was informative for 18 of the 26 cases. Furthermore, mutations of the VHL gene could not be identified by single-strand conformation polymorphism, dideoxyfingerprint or direct DNA sequence analysis. In addition, the VHL gene was not inactivated by hypermethylation in any of the 26 tumor samples studied. These findings demonstrate that allelic loss of chromosome arm 3p in UADT SCC involves regions surrounding the VHL locus but does not include the VHL gene. The VHL gene, therefore, does not appear to be involved in the pathogenesis of UADT SCC.

p53 mutations, O6-alkylguanine DNA alkyltransferase activity, and sensitivity to procarbazine in human brain tumors

BACKGROUND

In human brain tumors, sensitivity to procarbazine as measured by sensitivity in a xenograft tumor model correlated inversely with amounts of the DNA repair enzyme O6-alkylguanine DNA alkyltransferase (AT).

METHODS

To test the hypothesis that mutations of the p53 tumor suppressor gene in human tumors also can correlate with the response to chemotherapy, p53 mutations2 were identified in primary human malignant brain tumors and cell lines in which AT activity and procarbazine sensitivity in a xenograft model was ascertained.

RESULTS

Mutations were identified in 7 of 21 (33%) specimens tested. Specimens containing p53 mutations tended to exhibit an increased growth delay in procarbazine-treated xenografts and lower amounts of AT.

CONCLUSIONS

p53 mutations in brain tumors may contribute to procarbazine sensitivity by failing to induce arrest at the G1/S cell-cycle checkpoint, thereby preventing the repair of procarbazine-induced genetic alterations.

Allelic loss at chromosomes 3p, 8p, 13q, and 17p associated with poor prognosis in head and neck cancer

BACKGROUND

Little is known about the molecular genetic events that contribute to the pathogenesis of squamous cell carcinoma of the upper aerodigestive tract. Previous molecular genetic studies have been limited to the identification of mutations of the p53 (also known as TP53) tumor suppressor gene, activation of a limited set of oncogenes, allelic loss at 3p and other locations, and occasional association with human papillomavirus infection.

PURPOSE

Our purpose was to screen tumor tissue and blood from patients with squamous cell carcinoma of the upper aerodigestive tract for loss of heterozygosity at polymorphic loci corresponding to each of the autosomal chromosomes and to identify the locations of additional putative tumor suppressor genes, other than RB (also known as RB1) and p53, that may contribute to the pathogenesis of this disease.

METHODS

Tumor tissue and blood were obtained from 68 consecutive patients with squamous cell carcinoma of the upper aerodigestive tract. In all cases, tumor tissue was obtained from the center of the surgical specimen. The relative absence of non-neoplastic tissue was confirmed by frozen-section histologic examination of immediately adjacent tissue. Initially, 30 paired tissue and blood samples were tested for loss of heterozygosity by polymerase chain reaction (PCR) to amplify 43 different highly polymorphic sequences containing small oligonucleotide repeats. After PCR amplification, with unique oligonucleotides flanking the repeat, visualization and sizing of the alleles on DNA sequencing gels were performed. Specific loss of heterozygosity was distinguished from random genetic loss due to generalized chromosomal instability if it occurred in more than 20% of specimens tested for a particular marker.

RESULTS

Significant loss of heterozygosity (> 20%) occurred at alleles at chromosome bands 3p21 (32%), 3p25-26 (56%), 8pter-21.1 (31%), 13q14 (27%), and 17p12 (45%). Loss of heterozygosity at more than two loci was significant with a poor prognosis (P = .039).

CONCLUSIONS

These findings demonstrate that squamous cell carcinoma of the upper aerodigestive tract exhibits genetic alterations at multiple loci and that allelic loss at more than two locations is indicative of a poor prognosis (the likelihood of the patient dying of disease).

IMPLICATIONS

While tumor suppressor genes at 3p (VHL), 13q (RB), and 17p (p53) have been identified, altered genes at other loci on 3p and on 8p have not yet been characterized. Furthermore, the genotype at these loci for squamous cell carcinoma of the upper aerodigestive tract has prognostic importance and may identify the patients who should receive the most aggressive treatment.

Infrequency of ras, p53, WT1, or RB gene alterations in Wilms tumors

BACKGROUND

Alteration of the ras family of oncogenes and of the tumor suppressor genes p53 and RB are the most common genetic events in human tumors. Although there have been no reports of the prevalence of these alterations in Wilms tumors, overexpression of the N-myc and insulin-like growth factor-II (IGF-II) genes have been observed, and alteration of another tumor suppressor gene (WT1) has been demonstrated.

METHODS

Forty-four Wilms tumor specimens were tested for the presence of N-, K-, and H-ras mutations in codons 12, 13, and 61 by single-strand conformation polymorphism (SSCP) analysis and direct DNA sequence analysis. Sixteen tumors were tested for abnormalities of WT1 by Southern and northern blot analysis and reverse transcriptase polymerase chain reaction (RT-PCR). N-myc, c-myc, WT1, and IGF-II mRNA expression was measured in 16 tumors by Northern blot analysis. Thirty-eight tumors were screened for p53 mutations by SSCP analysis and direct DNA sequence analysis. Nine tumors were analyzed for loss of heterozygosity (LOH) of RB.

RESULTS

Although the authors confirmed that N-myc and IGF-II are overexpressed in Wilms tumors, no mutations of ras family, p53, or RB genes were identified, and no gross alterations of WT1 were detected by Southern or Northern blot analysis.

CONCLUSIONS

These findings suggest that H-ras, K-ras, N-ras, p53, and RB are not involved in the pathogenesis of Wilms tumor.

Infrequency of MDM2 gene amplification in pediatric solid tumors and lack of association with p53 mutations in adult squamous cell carcinomas

Loss of function of the p53 tumor suppressor gene by point mutation is the most commonly detected genetic alteration in human cancer. There is growing evidence that amplification and overexpression of the MDM2 gene are alternative mechanisms that also lead to functional inactivation of p53. While p53 mutations and MDM2 amplification have been reported to occur in rhabdomyosarcoma and osteogenic sarcoma, the incidence of MDM2 in other pediatric solid tumors is not known. We therefore tested a series of other pediatric solid tumors for MDM2 gene amplification. MDM2 amplification could not be detected in specimens from 40 Wilms' tumors, 15 neuroblastomas, 12 sarcomas, or 4 hepatoblastomas tested. To determine whether MDM2 amplification was an alternative mechanism of p53 inactivation in adult carcinomas that frequently possess p53 mutations, 68 samples of squamous cell carcinomas of the upper aerodigestive tract, 24% of which were previously shown to contain p53 mutations, were also tested for MDM2 amplification. MDM2 amplification did not occur in any of the tumor specimens tested. These findings suggest that MDM2 amplification may only occur in a limited subset of human tumors. Loss of function of p53 may be an essential event in human tumorigenesis. If so, then other mechanisms of p53 inactivation must occur in those tumors that exhibit neither p53 mutation nor MDM2 amplification.

p53, retinoblastoma, and human papillomavirus in squamous cell carcinoma and adjacent normal mucosa of the upper aerodigestive tract

OBJECTIVE

The primary objective of this study was to determine the incidence of p53 and retinoblastoma tumor suppressor gene mutations and human papillomavirus infection in squamous cell carcinoma and adjacent normal mucosa of the upper aerodigestive tract. The secondary objective was to associate these findings with clinical and histopathologic features.

DESIGN

Point mutations of p53 were identified by single-strand conformation polymorphism analysis and confirmed by direct DNA sequence analysis. Polymerase chain reaction-based methods were used to identify loss of heterozygosity of the retinoblastoma tumor suppressor gene and the presence of human papillomavirus sequences.

SETTING

University-based tertiary care center.

PATIENTS OR OTHER PARTICIPANTS

Forty-five consecutive cases of upper aerodigestive tract squamous cell carcinoma.

RESULTS

Eleven point mutations of p53 were identified in tumor samples (24%). No functional p53 mutations were detected in adjacent normal tissue from eight of these individuals nor was there evidence of p53 alteration in normal tissue adjacent to 12 of 30 additional tumors tested that demonstrated conformational alterations by single-strand conformation polymorphism analysis. The p53 mutations were significantly associated with local invasion. Loss of heterozygosity (which has a 20% chance of random occurrence in tumors) was detected at the retinoblastoma locus in 15% of the tumors tested. Five of the specimens (11%) were positive for human papillomavirus sequences (two of which also contained p53 mutations).

CONCLUSIONS

These findings suggest that p53 but not retinoblastoma or human papillomavirus is an important prognostic factor and is involved as a late event in the pathogenesis of upper aerodigestive tract squamous cell carcinoma.

Nonrandom distribution of N-myc oncogene genotypes in neuroblastoma

The distributions of Pvu II and Sph I alleles of the N-myc oncogene (also known as MYCN) were studied in a series of normal individuals and pediatric patients with solid tumors. In the case of Pvu II, where the polymorphic site is located 3' of the gene, the frequencies of the allele were 0.27 (11-kilobase fragment) and 0.73 (8-kilobase fragment) in 43 unrelated normal Caucasians. The frequencies of the allele were similar in 40 non-N-myc-amplified neuroblastomas, 47 Wilms' tumors, and 31 other pediatric tumors. In these cases, the genotypes were in Hardy-Weinberg equilibrium. In 18 N-myc-amplified neuroblastomas, however, the observed genotype frequencies deviated from Hardy-Weinberg equilibrium (P less than .005). Similar observations were made with an Sph I restriction fragment length polymorphism where the polymorphic site is located in intron 2. The differences between amplified and nonamplified neuroblastomas suggest a possible involvement of sequences at or near N-myc in the progression of tumors where the N-myc gene is amplified.

Nonrandom X chromosome DNA methylation patterns in hemophiliac females

Molecular X chromosome inactivation analysis was used to characterize three females (and their families) with severe hemophilia. First, the maternal and paternal X chromosomes were distinguished by restriction fragment length polymorphisms (RFLPs). Second, the patterns of methylation of X chromosome genes using methylation-sensitive restriction endonucleases were determined. Of the six X chromosome probes tested, only the phosphoglycerol-kinase (PGK) and hypoxanthine-phosphoribosyl-transferase (HPRT) clones were informative, indicating that other X chromosome probes are not useful for X inactivation analysis. After digestion with Hpa II or Hha I, the hybridization intensity of the RFLPs of all three mothers and an unaffected sister were diminished by 50%, consistent with random X chromosome inactivation. The methylation patterns of the X chromosomes of the affected females, however, were clearly nonrandom. Depending upon the probe and the patient, HPRT and PGK sequences were either completely methylated or unmethylated. These findings are extremely suggestive that nonrandom X chromosome inactivation (lyonization) is the basis for severe hemophilia in these females.

N-myc oncogene RNA expression in neuroblastoma

Tumor specimens from 33 patients with neuroblastoma were assayed for amplification of the N-myc oncogene and RNA expression to determine whether N-myc RNA expression levels correlated with N-myc gene amplification and clinical outcome. N-myc gene amplification was detected in one stage II tumor, one stage IV-S tumor, and seven stage III or IV tumors. In each case, N-myc RNA expression roughly paralleled N-myc gene amplification. However, enhanced N-myc RNA expression was not confined to tumors with N-myc gene amplification: all of the early (stage I and II) tumors, five stage IV-S tumors, and 12 advanced (stage III and IV) tumors had levels of N-myc RNA that were elevated up to 50-fold. While N-myc gene amplification correlated with prognosis, there was no such correlation with levels of N-myc RNA expression. The precise role of the N-myc gene in the pathogenesis of neuroblastoma remains unclear.

Beta-thalassemia in China: a systematic molecular characterization of beta-thalassemia mutations

In order to initiate a program of prenatal diagnosis for the prevention of beta-thalassemia in China, we have begun systematic studies of the beta-thalassemia mutations among the Chinese. DNA polymorphisms in the beta-globin gene cluster were examined in 46 beta-thalassemia chromosomes. Six different haplotypes were observed. One beta-thalassemia gene associated with a new haplotype was cloned and sequenced. The mutation was a single base substitution (A----G) at position -29 within the highly conserved proximal promoter element (the "TATA" box). This mutation was not observed previously in the Chinese. The beta-thalassemia genes were further screened with oligonucleotide probes specific for all known mutations in the Chinese. Five mutations were identified and accounted for 35 beta-thalassemia alleles.

The spectrum of beta-thalassemia genes in China and Southeast Asia

To make possible prenatal diagnosis of beta-thalassemia in China and Southeast Asia by direct detection of mutant beta-globin genes, we have determined the spectrum of mutations producing the disorder in this region of the world. Seventy-eight beta-thalassemia genes from Chinese and Southeast Asians were randomly obtained, and the relevant mutation was characterized in 76 (98%) of them. Seven different point mutations were found among the 78 genes studied. Of these seven beta-thalassemia alleles, two constitute 62%, and two others account for 29% of the total. Since only four alleles make up 91% of the mutant genes, prenatal diagnosis of beta-thalassemia in China and Southeast Asia should be feasible by simplified techniques for direct detection of point mutations.

Use of oligonucleotide hybridization in the characterization of a beta zero-thalassemia gene (beta 37 TGG----TGA) in a Saudi Arabian family

Analysis of restriction site polymorphisms in the beta-globin gene cluster of a Saudi Arabian female with beta zero-thalassemia demonstrated that both of her beta-globin genes were missing a nonpolymorphic AvaII site in exon 2. Examination of the normal nucleotide sequence surrounding this AvaII site revealed that either of two nucleotide substitutions, TGG----TAG or TGG----TGA, could produce a nonsense codon at codon 37 and eliminate the AvaII site. Consequently, two oligonucleotides (19-mers spanning codons 36 through 41 and containing either TAG or TGA at codon 37) were synthesized and hybridized against genomic DNA of the proband and her family. Specific hybridization with one of the oligomers demonstrated that the patient's beta o-thalassemia was the result of homozygosity for the TGG----TGA mutation at codon 37. In certain cases, oligonucleotide hybridization using genomic DNA may obviate the need for gene cloning and sequencing in the characterization of point mutations.

Concordance of a point mutation 5' to the A gamma-globin gene with A gamma beta + hereditary persistence of fetal hemoglobin in Greeks

In the Greek A gamma beta + type of hereditary persistence of fetal hemoglobin (HPFH), adult heterozygotes produce about 20% fetal hemoglobin (HbF), which is predominantly of the A gamma chain variety. The affected beta-globin gene cluster produces near normal amounts of beta-like globin, but in a A gamma to beta ratio of 20:80 instead of 0.5:99.5. Gelinas et al and Collins et al have shown a G to A change 117 nucleotides 5' to the A gamma gene in two Greeks with A gamma beta + HPFH. To demonstrate that this change is not a neutral polymorphism, we carried out hybridization with oligonucleotide probes (19mers) specific for the normal and the mutant sequences. While normal probe identified the A gamma fragment in genomic DNA of all subjects studied, mutant probe was positive only in Greeks with A gamma beta + HPFH. In sum, 108 beta-globin gene clusters of individuals without HPFH were negative when tested with mutant probe, but all 11 affected individuals of six families with Greek A gamma beta + HPFH (two previously sequenced and four new families) were positive with mutant probe. These data support the conclusion that the -117 mutation is causative of A gamma beta + HPFH in Greeks.

Hemophilia A. Detection of molecular defects and of carriers by DNA analysis

To understand the molecular basis of hemophilia A and to provide heterozygote detection and prenatal diagnosis by DNA analysis, we used cloned factor VIII:C DNA fragments to study 10 affected families. In four of these families, inhibitors of factor VIII:C had developed in affected persons. In one such family a deletion of approximately 80 kb within the factor VIII:C gene was identified. Carriers of the deletion were identified through detection of an abnormal DNA fragment located at the deletion end points. In another family a single nucleotide change in the coding region of the factor VIII:C gene produced a nonsense codon leading to premature termination of factor VIII:C synthesis. Carrier detection was performed in eight female members of this four-generation family. In a third family a small change in the size of a restriction-endonuclease fragment correlated with the presence of the mutant gene, and in the other seven families the molecular defect has not yet been identified. In addition, we used two common polymorphic sites in the factor VIII:C gene to differentiate the normal from the defective gene in four of six obligate female carriers from families with patients in whom inhibitors did not develop. Carrier detection was possible in other members of these families. These data suggest that DNA analysis of the factor VIII:C gene provides an accurate method of carrier detection and, potentially, of prenatal diagnosis in at least 50 per cent of the pedigrees affected by hemophilia A.

Huntington's disease: two families with differing clinical features show linkage to the G8 probe

To test the hypothesis that interfamily variability in Huntington's Disease (HD) is due to mutation at different loci, linkage analysis was undertaken in two large HD kindreds that differed in ethnicity, age-at-onset, and neurologic and psychiatric features. Both families showed linkage of the HD locus to the G8 probe. Several recombinants were documented in each family, and the best estimate of the recombination fraction for the two families was 6 percent with a 95 percent confidence interval of 0 to 12 percent. Although the data support the existence of a single HD locus, use of the G8 probe for presymptomatic testing in these kindreds would have resulted in a 12 percent error rate in genotype assignment at the HD locus.

Concordance of a point mutation 5' to the G gamma globin gene with G gamma beta +. Hereditary persistence of fetal hemoglobin in the black population

Hereditary persistence of fetal hemoglobin (HPFH) is a genetically heterogeneous and clinically benign condition characterized by persistent expression of fetal hemoglobin (Hb F) into adulthood. In the G gamma beta + type, no major deletions in the globin gene cluster occur; adult heterozygotes produce approximately 20% Hb F, which results from overproduction of G gamma chains, with no apparent increase in production from the adjacent A gamma gene. We have recently described a point mutation 202 base pairs 5' to the cap site of the G gamma gene in an individual with G gamma beta + HPFH. This mutation abolishes a normal ApaI restriction endonuclease site, and thus can be detected by blotting of genomic DNA. We present here further data on the ApaI mutation: (1) It occurs in six of seven families with G gamma beta + HPFH. (2) In three families, detailed haplotype analysis using 11 polymorphic restriction sites in the beta globin cluster has been done. The two that carry the missing ApaI site are identical but the third, which has a normal ApaI pattern, differs from the other two in at least two sites, one of which is a new polymorphic Nco I site between the delta and beta globin genes. This suggests the possibility of a different HPFH mutation in the third family. (3) The haplotype of the G gamma beta + HPFH chromosome carrying the ApaI mutation is different from that of 108 beta A chromosomes of black individuals that have been tested. (4) The G gamma ApaI site is normal in 61 beta A and 109 beta S alleles from non-HPFH black individuals, including 22 who share the same haplotype for the intragenic G gamma, A gamma HindIII polymorphisms. These data add support to the possibility that the -202 mutation is actually causative of the G gamma beta + HPFH phenotype.

Nonuniform recombination within the human beta-globin gene cluster

Population genetic analysis of 15 restriction site polymorphisms demonstrates nonuniform recombination within the human beta-globin gene cluster. These DNA polymorphisms show two clusters of high nonrandom associations, one 5' and another 3' to the beta-globin structural gene, with no significant linkage disequilibrium between the two clusters. The 5'- and 3'-association clusters are 34.6 kilobases (kb) and 19.4 kb long, respectively, and are separated by 9.1 kb of DNA immediately 5' to the beta-globin gene. For each of these three DNA regions, we have observed a relationship between nonrandom associations and physical distance between the polymorphisms. However, this relationship differed for each of these regions. On the assumption that the effective population size (Ne) is 5,000-50,000, we estimate the total recombination rate to be 0.0017%-0.0002% in the 5' cluster, 0.0931%-0.0093% in the 3' cluster, and 0.2912%-0.0219% in the 9.1-kb region between them. The beta cluster thus shows nonuniformity in recombination. Moreover, the recombination rate in the 9.1-kb DNA segment is 3-30 times greater than expected and is thus a hot spot for meiotic recombination.

Quantification of the close association between DNA haplotypes and specific beta-thalassaemia mutations in Mediterraneans

It has been suggested that there is a close linkage between specific restriction fragment polymorphism patterns, defined as haplotypes, in the beta-globin gene cluster and specific mutations in Mediterranean people with thalassaemia. This association formed the basis of a strategy for the efficient characterization of beta-thalassaemia mutations from the DNA sequence of one or two beta-thalassaemia genes derived from each haplotype in each ethnic group. Subsequently, Robertson and Hill argued that this strategy greatly underestimates the number of mutations on haplotypes which are frequent among normal chromosomes. We have therefore now analysed the proposed association and strategy quantitatively by the use of oligonucleotide hybridization and direct restriction analysis. Our results suggest that: (1) the association of specific haplotypes with specific mutations is high, but not invariant; (2) a different beta-thalassaemia mutation has arisen within each haplotype in Mediterraneans; and (3) mutation spread from one haplotype to another occurs mainly through meiotic recombination within a 9-kilobase region 5' to the beta-globin gene.

Molecular characterization of seven beta-thalassemia mutations in Asian Indians

To characterize systematically the mutations which produce beta-thalassemia in Asian Indians, we first determined the DNA polymorphism haplotype in the beta-globin gene cluster of 44 beta-thalassemia chromosomes in the ethnic group. Nine different haplotypes were observed. Upon molecular cloning and partial DNA sequencing of one beta-gene from each of eight haplotypes and two from the ninth, seven different mutations were found. None of these have been identified in Mediterranean patients, even among the five haplotypes which appeared identical in the two groups. Asian Indian mutations included one nonsense and three frameshift mutations, one deletion affecting an acceptor splice site, and two mutations affecting a donor splice site. The correlation of a specific mutation with a specific haplotype was high but not invariant. Two mutations were associated with more than one haplotype but, in each instance, the mutation spread to a new haplotype could be explained most simply by recombination 5' to the beta-globin gene. In addition, four mutations, one reported here and three others previously reported, have been observed on two chromosome backgrounds that are identical except for the status of a polymorphic HinfI site 5' to the beta gene. This HinfI site does not show significant linkage disequilibrium with markers both 5' and 3' to it, suggesting that it lies within a region of relative sequence randomization.

Hemoglobin E in Europeans: further evidence for multiple origins of the beta E-globin gene

We have determined haplotypes for the known restriction site polymorphisms in the beta-globin gene cluster in two families of European ancestry containing individuals who are heterozygous for hemoglobin E. In both families, the beta E mutation is associated with a haplotype not previously found among the haplotypes of beta E chromosomes in Southeast Asia. Moreover, in one family, the mutation is present in a beta-gene framework not found in beta E chromosomes of Southeast Asia. These data provide further evidence of multiple independent origins of the beta E mutation in human populations.

Beta-globin locus is linked to the parathyroid hormone (PTH) locus and lies between the insulin and PTH loci in man

Using a parathyroid hormone (PTH) cDNA probe we found a common Pst I polymorphic restriction site 3' to the PTH gene in all ethnic groups examined. Because the PTH, insulin, and beta-globin loci have been localized to the short arm of chromosome 11 (11p) we used DNA polymorphisms adjacent to each of these three loci to determine whether they are genetically linked and to determine their order. We found that the PTH and beta-globin loci are closely linked (estimated recombination fraction, 0.07; 95% confidence limits, 0.05-0.10; lod score, 4.63; odds favoring linkage, 42,000:1). Furthermore, our findings strongly indicate that the beta-globin gene cluster lies between the PTH and insulin loci. Therefore, the gene order on 11p is centromere-PTH-beta-globin-insulin.

The entire beta-globin gene cluster is deleted in a form of gamma delta beta-thalassemia

We have used restriction endonuclease mapping to study a deletion involving the beta-globin gene cluster in a Mexican-American family with gamma delta beta-thalassemia. Analysis of DNA polymorphisms demonstrated deletion of the beta-globin gene from the affected chromosome. Using a DNA fragment that maps greater than 40 kilobases (kb) 5' to the epsilon-gene as a probe, reduced amounts of normal fragments were found in the DNA of affected family members. Similar analysis using radiolabeled DNA fragments located 3' to the beta-globin cluster has shown that the deletion extends more than 17 kb 3' to the beta-gene, but terminates before the 3' endpoint of the Ghanian HPFH deletion. Hence, this gamma delta beta-thalassemia deletion eliminates over 105 kb of DNA and is the first report of a deletion of the entire beta-globin gene cluster.

Evidence for multiple origins of the beta E-globin gene in Southeast Asia

To investigate whether recurrent mutation has contributed to the high frequency of the beta E-globin gene in Southeast Asia, we used the haplotypes at three polymorphic restriction sites within and to the 3' side of the beta-globin gene to predict the framework of 23 beta E-globin genes. These haplotypes suggested that beta E-globin genes are present in two different beta-globin gene frameworks. DNA sequence determination of one gene representing each framework demonstrated that the same mutation (GAG leads to AAG at codon 26) was present in both frameworks. Moreover, the frameworks differed at three nucleotide positions known to be polymorphic in Mediterraneans. These polymorphic sites are located 70 nucleotides to the 5' side of the beta E mutation and 382 and 1032 nucleotides to the 3' side of it. The existence of the beta E mutation in these two beta-globin gene frameworks can be explained by (i) recurrent mutation giving rise to beta E-globin, (ii) a double crossing-over event, or (iii) two single crossing-over events. Mathematical analysis suggests that the first alternative, recurrent mutation of G leads to A at the first nucleotide of codon 26, is most likely.