Mutation screening of the mitochondrial genome using denaturing high-performance liquid chromatography

https://doi.org/10.1016/j.ymgme.2004.09.011Get rights and content

Abstract

Over 170 known mutations of the mitochondrial genome are responsible for disease. Due to the unique features of mitochondrial genetics, such patients are clinically diverse and difficult to diagnose. As pathogenic mitochondrial DNA (mtDNA) mutations are mostly heteroplasmic, denaturing high-performance liquid chromatography (DHPLC) could be used to detect these heteroplasmic species and therefore act as a rapid screening test for mtDNA mutations. The entire mitochondrial genome was amplified by PCR in 40 overlapping regions. In addition, known mtDNA mutants were constructed for each of these regions using a PCR-based site-directed mutagenesis approach. These mutants were used as positive controls and showed a detection limit of 3–10% heteroplasmy by DHPLC (depending on the specific mutation) compared to 40% for conventional sequencing. To further validate the screening test, mtDNA from 17 patients with seven different pathogenic mutations was used to compare mutation detection by DHPLC and conventional sequencing. DHPLC had a sensitivity of 88% compared to 82% for sequencing. This increased to 100% sensitivity for DHPLC when excluding the m.8993T>G mutation. DHPLC analysis is therefore a sensitive, rapid and cost-effective method to screen for mutations in the mitochondrial genome. The role of pyrosequencing in the quantitation of mutant load for known mtDNA mutations was highlighted using the m.3243A>G mutation as an illustrative example. Pyrosequencing analysis was able to discriminate samples containing as little as 5% heteroplasmy and proved to be an accurate and reproducible method for estimation of mutant load.

Introduction

Mitochondrial DNA (mtDNA) is a 16,569 bp circle of maternally inherited, extra-nuclear, double-stranded DNA [1]. It contains 37 genes encoding two ribosomal RNAs, 22 transfer RNAs and 13 subunits of the respiratory chain [2]. In addition to its maternal inheritance, mtDNA differs from nuclear DNA in that there are several hundreds to thousands of copies per somatic cell [3]. The exact amount of mtDNA per cell is tissue-dependent with large copy numbers found in ‘metabolically active’ tissues such as the central nervous system, skeletal muscle, and gastrointestinal system [1].

Another contrast to nuclear genes is the existence of heteroplasmy and the threshold effect [1]. The presence of a single species of mtDNA is termed homoplasmy with a mixture of wild-type and mutant mtDNA termed heteroplasmy. As the proportion of mutant mtDNA species increases beyond a threshold amount, clinical manifestations of the mutation become apparent—often with the severity of phenotype in proportion to the mutant load [4].

To date there have been over 170 reported mutations and over 1000 known polymorphisms of mtDNA [5]. Non-pathogenic polymorphisms are normally homoplasmic [6] whilst pathogenic mutations are mostly heteroplasmic [4].

Some mitochondrial disorders are associated with a characteristic set of clinical and pathological features [1]. However, the factors of heteroplasmy, threshold effect and differential tissue distribution present a diagnostic challenge for clinicians presented with a patient harboring a mtDNA mutation. The variable age of onset, mode of presentation and rate of progression of many mitochondrial disorders makes diagnosis particularly difficult [3].

The use of molecular genetic approaches to determine the presence of mtDNA mutations has had recent success but has often been limited to mutations that have been previously identified [7]. Other groups have also had limited success using techniques such as single-stranded conformation polymorphism [8] and two-dimensional electrophoresis [9]. Issues of sensitivity, specificity and labour intensive methodologies inherent in these techniques have been overcome by the use of denaturing high-performance liquid chromatography (DHPLC) [10]. This automated technology allows rapid detection of heterozygous and heteroplasmic mutations and has been extensively used for the diagnosis of cystic fibrosis [11], breast cancer [12], and acute lymphoblastic leukemia [13]. In addition to these strategies, pyrosequencing has emerged as a new and accurate method for the detection of single nucleotide polymorphisms [14] and has lent itself to the detection of specific mtDNA mutations [15].

Several groups have recently described the use of DHPLC in screening for mitochondrial DNA mutations [16], [17]. The van den Bosch study screened the entire mtDNA genome in 13 fragments and performed ‘multiplex’ DHPLC on each fragment following restriction enzyme digestion [16]. Liu et al. [17] used a more conventional DHPLC approach to investigate m.3243A>G mutations and variants of the non-coding D-loop region of mtDNA. Both of these studies illustrated the usefulness of DHPLC in mtDNA mutation detection, but were limited by the low number of individual mutations used to validate their screening strategies.

Here, we describe the development and implementation of a DHPLC-based screening approach for the detection of mutations in mtDNA. The entire mitochondrial genome was divided into 40 overlapping regions and site-directed mutants were constructed for each region allowing a comprehensive assessment to be made of DHPLC sensitivity. A group of 17 patients covering seven known mtDNA mutations were used to further validate this screening strategy. In addition, the important role of pyrosequencing in the quantitation of mutant load was highlighted using the m.3243A>G mutation as an illustrative example.

Section snippets

DNA samples

Seventeen DNA samples from patients with known mtDNA mutations, covering seven individual mutations with varying mutant loads, were used to determine the sensitivity and specificity of the DHPLC screening test (see Table 1 for details). In addition, total DNA was isolated from healthy anonymous male donors using a salting-out method [18]. This DNA was used as a negative control and as a template to construct site-directed mutants that would provide positive-control specimens for all 40

DHPLC of site-directed mutants

DHLPC analysis was performed at a range of temperatures on samples containing a mixture of wild-type and site-directed mutant in a 1:1 stoichiometry. Fig. 1 compares the profiles formed with these samples compared to wild-type at their established melting temperatures. These data illustrate the formation of heteroduplex species with mutant/wild-type mixtures compared to the homoduplex profiles of wild-type samples. Heteroduplex formation was consistent and pronounced for all regions of the

Discussion

The development of a DHPLC-based screening test for the detection of mtDNA mutations requires stringent validation and assessment of both specificity and sensitivity. Validation of each of the 40 regions of the screening test was undertaken by the successful detection of specific site-directed mutants to each of these regions. Where possible, site-directed mutagenesis was based on currently identified mtDNA mutations with only nine regions having no associated mutation. Mutants for these

Acknowledgments

This research was supported by an Australian National Health and Medical Research Council (NHMRC) project grant, the Cecilia Kilkeary Foundation, and the Children’s Hospital Fund of the Children’s Hospital at Westmead. D.R.T. is a NHMRC Senior Research Fellow.

References (34)

  • M. Zeviani et al.

    Mitochondrial disorders

    Mol. Hum. Reprod.

    (1997)
  • C.-Y. Lu et al.

    Detection of DNA mutations associated with mitochondrial diseases by Agilent 2100 bioanalyzer

    Clin. Chim. Acta

    (2001)
  • C. Eng et al.

    Genetic testing

    Nat. Biotechnol.

    (1997)
  • W. Xiao et al.

    Denaturing high-performance liquid chromatography: a review

    Hum. Mut.

    (2001)
  • C. Le Marechal et al.

    Complete and rapid scanning of the cystic fibrosis transmembrane conductance regulator (CFTR) gene by denaturing high-performance liquid chromatography (D-HPLC)

    Hum. Genet.

    (2001)
  • C. Eng et al.

    Interpreting epidemiological research: blinded comparison of methods used to estimate the prevalence of inherited mutations in BRCA1

    J. Med. Genet.

    (2001)
  • U. Zur Stadt et al.

    Denaturing HPLC for identification of clonal T-cell receptor gamma rearrangements in newly diagnosed acute lymphoblastic leukemia

    Clin. Chem.

    (2001)
  • Cited by (45)

    • Colorimetric detection of single nucleotide polymorphisms in the presence of 10<sup>3</sup>-fold excess of a wild-type gene

      2015, Biosensors and Bioelectronics
      Citation Excerpt :

      One evident advantage of the proposed assay is that it can detect both homozygous and heterozygous SNPs, and particularly ultralow levels of heterozygous SNPs, capable of detecting SNPs present at as low as 0.01% heterozygy. In contrast, when the heteroplasmic level is lower than 20% in a DNA sample extracted from leukocytes, SNP detection based on direct sequencing may yield ambiguous results due to considerable responses from the co-existing WT gene (Biggin et al., 2005). The proposed assay was applied to the detection of SNPs in genomic DNA samples extracted from tissue and culture cells for the polymorphic site codon 12 of Exon 2 in KRAS oncogene.

    • Mitochondrial respiratory chain disorders in childhood: Insights into diagnosis and management in the new era of genomic medicine

      2014, Biochimica et Biophysica Acta - General Subjects
      Citation Excerpt :

      Screening for less frequent or unknown mutations can be undertaken using a number of molecular methods including: direct sequencing, RFLP, denaturing high performance liquid chromatography (dHPLC) analysis, Southern-blot, and chip arrays [61,92,93]. Although considered the gold standard for the identification of mutations, direct sequencing of mtDNA should be interpreted with caution, as it does not detect large deletions, and, importantly because of the low sensitivity of sequencing in detecting low level heteroplasmic (< 20%) mtDNA mutations [9,93,94]. In addition, care should be taken when annotating a variant as pathogenic because of the high frequency of single nucleotide polymorphisms which may not be of pathogenic significance [59].

    • The link between mitochondrial DNA hypervariable segment I heteroplasmy and ageing among genetically unrelated Latvians

      2011, Experimental Gerontology
      Citation Excerpt :

      However, both techniques do not specify the position of the mutated nucleotide. In contrast, pyrosequencing is able to identify the level of heteroplasmy up to 5% and the mutated nucleotide position (Biggin et al., 2005; Rose et al., 2007). He et al. (2010) used the parallel sequencing-by-synthesis approach for the analysis of the whole mitochondrial genome from normal and colorectal tumor cells.

    View all citing articles on Scopus
    View full text