Comprehensive mutational analysis of the ATM gene by denaturing HPLC. Y.R. Thorstenson¹, P.J. Oefner¹, V. Goss², R.W. Davis¹, R.A. Gatti³, G. Chu². 1) Stanford Genome Tech Ctr, Palo Alto, CA; 2) Stanford Univ Medical Ctr, Stanford, CA; 3) UCLA Medical Ctr, Los Angeles, CA.
   ATM is the gene mutated in the autosomal recessive disease Ataxia-telangiectasia (A-T). ATM is a complex gene with 66 exons and 9168 base pairs of coding sequence in a genomic region of 150,000 base pairs. A worldwide effort to identify all of the mutations responsible for A-T has been hindered by the large size of the gene, as well as the heterogeneity of the causative mutations. Furthermore, most A-T patients are compound heterozygotes - a challenging situation even for conventional dideoxy sequencing protocols. Given these difficulties, it may not be surprising that published reports indicated a detection rate of only about 75%. Here, a new mutation detection strategy using denaturing HPLC (DHPLC), is compared with conventional protein truncation test (PTT), and single-strand conformation polymorphism analysis (SSCP). A set of 60 genomic DNA samples from 55 A-T patients and 5 obligate carriers was blinded to the DHPLC operator. The samples included mutations that had proven difficult to detect; only 69 out of 115 expected alleles had been identified previously (60%). DHPLC analysis confirmed all 69 and added 27 more alleles, for a total of 96 alleles (83%). Given that the sensitivity of DHPLC is close to 100%, it is likely that the remaining 19 alleles involve intron or flanking sequences in the ATM gene. The 96 alleles included 65 unique disease-causing mutations, including 23 not reported before. Of 19 unique mutations that were detected only by DHPLC, one was a single base pair insertion, and 18 were single nucleotide changes, clearly demonstrating a bias for detection of insertion and deletion mutations using conventional methods. Furthermore, the 18 newly detected single base pair changes were functionally relevant, including five premature stops, six non-conservative amino acid changes at residues conserved in the mouse ATM homolog, three alterations at canonical splice donor or acceptor sites, and four disruptions of potential splice enhancer motifs. Hence, many more of the previously undetected mutations in A-T patients may be caused by single base pair changes.