Identification of Germline Mutations in Hereditary Nonpolyposis Colorectal Cancer Using Base Excision Sequence Scanning Analysis

Hereditary nonpolyposis colorectal cancer (HNPCC) is an autosomal dominantly inherited disease caused by loss of function of DNA mismatch repair genes. Defects in MLH1 and MSH2 account for ∼98% of the mutations in HNPCC families (1). Identification of gene carriers within these families is of great importance because surveillance may be restricted to genetically affected relatives. Identification of mutations by direct sequencing is time-consuming and not feasible in a large-scale clinical setting. Molecular screening strategies, including single-strand conformation polymorphism analysis (2), denaturing gradient-gel electrophoresis (3), constant denaturant gel electrophoresis (4), or in vitro transcription/translation assays (5), have been described and may facilitate the detection of mutations. However, these techniques often have low sensitivity with mutation detection rates of only 35–70% (6), or they are highly accurate but are technically difficult to perform (3)(4)(5). In the present study, we developed and evaluated a modified base excision sequence scanning (BESS) protocol (7) for the detection of MLH1 and MSH2 germline mutations. This simple method is based on the incorporation of dUTP into the PCR products. Subsequent digestion with uracil N-glycosylase, which releases uracil from both single-stranded and double-stranded DNA and thus creates apyrimidinic sites, and endonuclease IV, which cleaves the phosphodiester bond at these sites, generates a defined series of fragments (7)(8). Lymphocytes were prepared from whole blood of patients with HNPCC and healthy subjects using Vacutainer cell preparation tubes (Becton Dickinson). After extraction of total RNA (Tri-Star-Kit; AGS), complementary DNA synthesis was performed with reverse transcriptase (Superscript; Life Technologies) and random hexamer oligonucleotides or 2.5 μmol/L reverse primers (Table 1). The PCR amplification was carried out in a Perkin-Elmer 9700 PCR system in a total volume of 50 μL containing 5 U of AmpliTaq Gold polymerase (Perkin-Elmer); 60 mmol/L Tris-HCl, pH 8.5; 15 mmol/L (NH4)2SO4; 3.5 mmol/L MgCl2; 200 μmol/L dATP, dTTP, dGTP, and dCTP; 16 μmol/L dUTP (Biozym Diagnostik); and 2.5 μmol/L forward and reverse primers (Table 1). Either the forward or the reverse primer was labeled with 6-carboxy-fluorescein. Amplification conditions were optimized and applied as follows: 10 min at 95 °C; 50 cycles of 30 s at 95 °C and 30 s at 50, 55, or 65 °C, depending on the amplified fragments (see Table 1), and 1 min at 72 °C; and final extension 10 min at 72 °C. PCR products were purified on 1.5% agarose gels using the Qiaquick Gel purification kit (Qiagen). The eluted DNA was digested in 20 μL of a solution with 2 μL of excision enzyme mixture containing uracil N-glycosylase and endonuclease IV (Biozym Diagnostik), 50 mmol/L Tris-HCl (pH 9.0), 20 mmol/L (NH4)2SO4, and 10 mmol/L EDTA at 37 °C for 45 min. The fragments of the digested PCR products, ranging in size from 22 to 480 bp, were mixed with 11.5 μL of formamide and 1.5 μL of TAMRA size marker (N,N,N′,N′-tetramethyl-6-carboxyrhodamine; Perkin-Elmer) and electrophoresed for 40 min on an automated ABI 310 DNA sequencer with laser scanning and linear detection characteristics (Perkin-Elmer). The peak pattern represented fragments ending with dUTP and was comparable with the “T” lane of a conventional sequencing reaction. BESS analyses of the MLH1 and MSH2 genes were performed on samples from four patients with known sequence-confirmed mutations and subsequently prospectively on samples from four patients who fulfilled the Amsterdam criteria. The appearance, disappearance, or change in intensity of a peak in comparison with a control, which indicated the presence of a mutation, was assessed by an investigator who was unaware of the direct sequence data. All patients consented to participate in the study, which was approved by the Ethics Committee for Medical Research in Frankfurt a.M. in accordance with the Declaration of Helsinki. Primers for amplification and BESS analysis of overlapping MLH1 and MSH2 fragments.1 BESS analysis of both strands using forward and reverse 6-carboxy-fluorescein-labeled primers, respectively. RT, reverse transcriptase. Primers for amplification and BESS analysis of overlapping MLH1 and MSH2 fragments.1 BESS analysis of both strands using forward and reverse 6-carboxy-fluorescein-labeled primers, respectively. RT, reverse transcriptase. In families with HNPCC, 214 different germline mutations have been described to date by direct sequence analysis: 127 mutations were located in the MLH1 gene and 81 in the MSH2 gene, whereas only 6 mutations were detected in the MSH6, the PMS1, or the PMS2 gene (9). To establish and evaluate BESS as a screening method, three representative mutations and one polymorphism of MLH1 and MSH2 were analyzed. Initially, we investigated a heterozygous ACG→ATG missense mutation in codon 117, exon 4 of the MLH1 gene leading to a change from threonine to methionine (MLH1, Thr117Met). BESS analysis identified this mutation by an additional peak at nucleotide position 350 (Fig. 1 A). In addition, we performed BESS analysis on a patient with a heterozygous TGT→CGT missense mutation in codon 697, exon 13 of the MSH2 gene (MSH2, Cys697Arg). As shown in Fig. 1B , this mutation was detectable by a 50% reduction of the T peak at nucleotide position 2089. As a third representative mutation, we investigated an as yet unidentified 1-bp deletion within codon 782, exon 14 of the MSH2 gene (MSH2, DEL782FS). This frameshift mutation produced a complex fragment pattern in the BESS analysis, which was caused by the superimposed band pattern of the wild-type and mutant alleles. When compared with the BESS pattern of a healthy control subject, the deletion at nucleotide position 2345 was clearly identified (Fig. 1C). However, in a patient with a known GGC→GGG polymorphism in codon 713 of the MSH2 gene (MSH2, Gly713Gly), a BESS pattern identical to those of healthy control subjects (n = 7) was observed (data not shown). Subsequently, we investigated the MLH1 and MSH2 genes of patients fulfilling the Amsterdam criteria prospectively by the BESS protocol. In two patients, missense mutations in the MSH2 gene, one in codon 322 (MSH2, Gly965Asp) and one as yet unidentified mutation in codon 388 (MSH2, Pro1165Leu), were detected. In the third patient, BESS analysis showed another as yet unidentified missense mutation in codon 618 of the MLH1 gene (MLH1, Glu1853Asp). All mutations detected by the BESS protocol were confirmed by direct sequencing. In the remaining patient, BESS analysis and direct sequencing revealed no mutation in MLH1 and MSH2. BESS analysis of the MLH1 and MSH2 genes in patients with HNPCC. RNA was obtained from patients with HNPCC and healthy controls. After reverse transcription, cDNA was amplified using dUTP in addition to dATP, dTTP, dGTP, and dCTP. PCR products were purified and digested by uracil N-glycosylase and endonuclease IV; the resulting fragments were subsequently resolved on an automated DNA sequencer. (A, top scan), wild-type sequence of MLH1 exon 4 (codons 114–126) and the corresponding BESS analysis band pattern. (A, bottom scan), the heterozygous ACG→ATG missense mutation (MLH1, Thr117Met) is clearly identified by an additional peak of the BESS analysis band pattern (arrow). (B, top scan), wild-type sequence and the corresponding BESS analysis pattern of MSH2 exon 13 (codons 684–698). (B, bottom scan), the heterozygous TGT→CGT missense mutation (MSH2, Cys697Arg) can be identified by a 50% reduction of the T peak at nucleotide position 2089 (arrow). (C, top scan), wild-type sequence and the corresponding BESS analysis pattern of MSH2 exon 14 (codons 774–788). A 1-bp deletion within codon 782 (MSH2, DEL782FS) leads to a complex fragment pattern in the BESS analysis (bottom scan, arrow), which is caused by the superimposed peak patterns of the wild-type and mutant alleles. RNA-based techniques provide an interesting approach for mutation screening because 81% of these genetic defects are located in exons and the remaining 19% are relevant intronic mutations that lead to detectable splicing variants. BESS analysis identifies missense mutations, deletions, insertions, repeat expansions, and frameshift mutations at sites involving dTTP, which account for 96% of known MLH1 and MSH2 mutations (1)(9). To date, only 8 of 208 MLH1 and MSH2 mutations have been identified as G→C or C→G missense mutations (9), which cannot be detected by the BESS method. Thus, the sensitivity of BESS analysis (7) for detecting MLH1 and MSH2 mutations theoretically should be higher compared with in vitro transcription/translation (62%) (5) and single-strand conformation polymorphism techniques (35–70%) (6). However, for the definite determination of accuracy of the BESS method, a large prospective evaluation is required. In healthy controls, we did not observe aberrant BESS peak patterns, indicating high specificity. BESS analysis of MLH1 and MSH2 can be performed on an automated sequencer in less than 24 h with a hands-on time of 6 h. In addition, a potential mutation must be confirmed by sequence analysis. Nevertheless, the costs of BESS analysis are <50% compared with complete genomic sequence analysis. BESS is considerably less labor-intensive than genomic sequencing and several other screening methods, such as constant denaturant gel electrophoresis, denaturing gradient-gel electrophoresis, and in vitro transcription/translation assays (3)(4)(5). With the described BESS method, we could analyze fragments of up to 500 bp. To further accelerate mutation screening, we attempted to analyze PCR fragments of up to 750 bp. Despite extension of the electrophoresis time, we were not able to resolve distinct fragment peaks; this problem might be overcome by the use of longer capillaries or an optimized polymer. In summary, the described modified BESS method allows rapid, efficient, and simple detection of MLH1 and MSH2 germline mutations in HNPCC. Its application can improve the genetic diagnosis of hereditary cancer susceptibility syndromes caused by germline mutations of large genes without mutation hotspots. fax 49-69-6301-4807, e-mail Zeuzem@em.uni-frankfurt.de