Introduction

Myoclonus epilepsy with ragged-red fibers (MERRF) is a progressive encephalomyopathy characterized by myoclonia, seizures, ataxia and dementia as well as ragged-red fibers in the Gomori trichrome staining of muscle [1]. Maternal inheritance of the disease suggested that a mutation in the mitochondrial DNA (mtDNA) could be associated with the disease. This was first shown by Shoffner et al. [2], who found an A→G transition at nucleotide (nt) 8344 [3] of the mitochondrial tRNALys gene (tRNA8344Lys) in MERRF patients. The same mutation has subsequently been identified in MERRF patients from several populations [4–13].

A typical feature of mtDNA mutations is heteroplasmy. The tissues of the patients with an mtDNA mutation contain both mutated and normal, ‘wild-type’, mtDNA molecules. It is still unclear how much mutated mtDNA can be present without affecting the energy metabolism of the cell, or if there is any correlation between the percentage of mutated genomes and the severity and onset of the disease.

The tRNA8344Lys mutation has been shown to exist in various tissues, including leukocytes, of MERRF patients [4–13]. Detection of the mutation has so far been based on DNA sequencing or amplification of the gene by the polymerase chain reaction (PCR) [14] followed by restriction-site analysis of the PCR product. The tRNA8344Lys has been found to be heteroplasmic, but determining the proportions of wild-type and mutated mtDNA by densitometry of electrophoretically separated bands [2, 5, 9, 11] is not an ideal procedure for a quantitative analysis.

Maternal transmission of a syndrome fulfilling the criteria for MERRF in a large Belgian family encouraged us to study the leukocytic mtDNA from nine of the family members. We found the tRNA8344Lys mutation of mtDNA in all the individuals of maternal lineage. To detect the mutation we used a new method, solid-phase minisequencing, which is based on primer-guided incorporation of a single nucleotide [15]. The result of this method is obtained as a numeric value reflecting the proportion of wild-type and mutated mtDNA initially present in the sample. The tRNA8344Lys mutation was found to be heteroplasmic in all the samples of maternally related individuals, and the solid-phase minisequencing technique allowed us to accurately determine the ratio of the mutated and wild-type genomes in these samples.

Material and Methods

Case Reports

The pedigree and the main clinical symptoms of the members of the Belgian MERRF family are described in figure 1 and table 1. As controls we analyzed leukocyte mtDNA of a patient with progressive external ophthalmoplegia (control patient 1), and one with distal dystonia (control patient 2), as well as leukocytes of three healthy individuals on the laboratory staff.

Fig. 1
figure 1

Pedigree of the MERRF family. Open symbols indicate the healthy family members, closed symbols indicate the affected ones. A slash in the symbol indicates deceased individuals.

Full size image
Table 1 Main clinical symptoms of the members of the MERRF family
Full size table

Oligonucleotides

The oligonucleotides were synthesized on an Applied Biosystems 381A DNA synthesizer. The upstream PCR primer (primer no. 1) was homologous to the sequence between nt 8313–8328 [3] in the tRNALys gene of mtDNA. This primer was biotinylated as previously described [16]. The downstream PCR primer (no. 2) was complementary to the region between nt 8354–8373. A detection primer for the minisequencing reaction was designed to hybridize to nt 8345–8365, immediately adjacent to the tRNA8344Lys mutation. In addition, two 61-mer oligonucleotides spanning the region between nt 8313 and 8373 were prepared to serve as standards for quantitative analysis by the solid-phase minisequencing method. One of them contained the nucleotide corresponding to the wild-type (A) sequence at nt 8344 and the other contained the mutated G at nt 8344. The 61-mer oligonucleotides were purified by reverse-phase high-performance liquid chromatography as previously described [16].

PCR Amplification

50 ng of DNA extracted from leukocytes or muscle tissue samples according to standard methods [17] were amplified in each PCR reaction. The reaction mixture contained 10 pmol of the biotinylated primer no. 1 and 50 pmol of the primer no. 2, the four dNTPs at 0.2 µM and 1.25 U of Taq DNA polymerase (Promega) in a total volume of 50 µl consisting of 20 mM Tris-HCl, pH 8.8, 15 mM (NH4)2SO4, 1.5 mM MgCl2, 0.1% Tween 20 and 0.01% gelatin. The denaturation was carried out at 94 °C for 1 min, annealing at 55 °C for 1 min and extension at 72°C for 1 min for 25 cycles in a programmable heat block (Techne PHC1).

Solid-Phase Minisequencing

The principle of the method is described in figure 2. For each minisequencing reaction a 10-µl aliquot of the PCR mixture and 40 µl of 20 mM sodium phosphate buffer, pH 7.5, 0.1% Tween 20 were added to microtitration wells (Maxisorb Nunc) that had been coated with streptavidin by passive absorption. The samples were incubated with gentle shaking at 37 °C for 1.5 h and washed three times with 200 µl of 40 mM Tris-HCl, pH 8.8, 1 mM EDTA, 50 mM NaCl, 0.1% Tween 20, at 22 °C. 100 µl of 50 mM NaOH were added twice to the wells with 5 min incubation at room temperature, and the wells were washed as above.

Fig. 2
figure 2

Principle of the solid-phase minisequencing method. PCR is carried out with one biotinylated primer, and the biotinylated PCR product is captured into a streptavidin-coated microtiter well and denatured. In the minisequencing reaction a primer hybridizing adjacent to the site of the mutation is elongated with a 3H-labelled test nucleotide by Taq DNA polymerase. The primer is released and the incorporated label is measured in a liquid scintillation counter.

Full size image

The reaction mixture consisting of 0.2 µM detection primer, 0.4 µM [3H]dTTP (Amersham, TRK 576, 100 Ci/mmol) to detect the wildtype A8344, or 0.4 µM [3H]dCTP (TRK 625: 64 Ci/mmol, Amersham) to detect the mutated G8344, and 1 U Taq DNA polymerase in 50 µl of PCR buffer was added to each well, the samples were incubated at 50 °C for 10 min and washed as above. The microtitration plates were treated with 60 µl of 50 mM NaOH for 5 min at room temperature and the eluted radioactivity was measured in a liquid scintillation counter (1209 Rackbeta; LKB).

Results

We analyzed leukocytic DNA samples from nine members of a large MERRF family using the solid-phase minisequencing method [15]. In addition, one patient with progressive external ophthalmoplegia and one with distal dystonia as well as three healthy control individuals were also analyzed. We identified the A→G transition at nt 8344 in all the diseased members of the family, as well as in two healthy men and a young child (II/4, III/1, III/7) of maternal lineage (table 2). A healthy man (III/5), whose father (II/4) carried the mutation, did not show mutated genomes. In all these individuals the presence of the mutation resulted in a significant incorporation of [3H]dCTP in the minisequencing reaction. The incorporation of [3H]dTTP also in all these samples indicated that the patients were heteroplasmic for the mutation. Two other patients with neurological symptoms did not carry the tRNA8344Lys mutation, nor did the healthy control individuals.

Table 2 Detection and quantification of the A→G tRNA8344Lys mutation of mtDNA in MERRF patients and control individuals by solid-phase minisequencing
Full size table

The result of the minisequencing method is obtained as cpm values of incorporated [3H]dNTP. The ratio between the incorporated nucleotides reflects the ratio of the corresponding mutated and wild-type mtDNA. Figure 3 shows a standard curve prepared by mixing, in known ratios, two synthetic 61-mer oligonucleotides, one corresponding to the mutant sequence with G8344 and the other corresponding to the normal sequence with A at this position. After PCR amplification and analysis by the solid-phase minisequencing method the obtained Ccpm/Tcpm ratios were plotted as a function of the initial ratio between the wild-type and mutant sequences in the mixtures. A linear standard curve between 5 and 95% was obtained (fig. 3). The curve was reproduced in several subsequent amplifications.

Fig. 3
figure 3

A standard curve of solid-phase minisequencing. Two synthetic oligonucleotides were mixed in varying proportions, one containing the wild-type A at nt 8344 of mtDNA, the other carrying the MERRF mutation (G) at the same position. The PCR products were quantitated by solid-phase minisequencing and the resulting CcpmTcpm ratios were plotted as a function of the ratio of the sequences in the initial mixture.

Full size image

Comparing the Ccpm/Tcpm ratios from the patients’ samples with the standard curve allowed determination of the ratio of the mutant and wild-type mtDNA present in the original samples (table 2).

In addition to the leukocyte samples, a muscle tissue sample of patient II/3 was studied. The muscle had 76% of the mutated mtDNA, which is more than was found in his leukocytes (57%).

Discussion

In this study we have shown the presence of the tRNA8344Lys point mutation in leukocyte samples from a large MERRF family. Because the majority of patients filling the clinical criteria for MERRF carry the tRNA8344Lys mutation, a convenient and accurate diagnostic test to detect the mutation is needed. In previous studies, the tRNA8344Lys mutation has been detected by restriction site analysis of DNA amplified by the PCR. The A→G transition at nt 8344 creates a restriction site for the restriction enzyme CviJI [2, 7], which is, however, not commercially available. An approach to circumvent this problem is to create a new restriction site mediated by a mismatched PCR primer [5, 6, 8, 9, 11, 12].

The tRNA8344Lys mutation has been found to be heteroplasmic and a correlation between the amount of mutated mtDNA and the severity of the disease has been suggested [2, 8, 9]. To more clearly investigate this interesting suggestion a diagnostic test for the MERRF disease should allow accurate quantification of the proportion of mutated and wild-type mtDNA. Although PCR itself is not quantitative, the ratio between the mutant and wild-type sequences in the PCR product reflects the initial ratio between the two mtDNA populations. The ratio has so far been determined by restriction enzyme cleavage of the PCR product carrying the mutation [2, 5, 9, 11]. This method may result in overestimation of the amount of the normal sequence due to incomplete cleavage of heteroduplexes formed between the two sequences during PCR [18, 2, 9]. This problem was avoided by Tanno et al. [9] who added a radioactively labelled primer into the last cycle of the PCR reaction. However, all these approaches require densitometric analyses of restriction fragments separated by gel electrophoresis, which involves several steps that may introduce errors into the quantitative analysis.

We applied a recently described method, solid-phase minisequencing [15], to detect the tRNA8344Lys A→G transition of mtDNA. In this method the amplified DNA is immobilized on a solid support, and the mutated and normal nucleotides are detected by separate primer extension reactions, in which a detection primer located immediately adjacent to the test nucleotide is elongated by a single labelled dNTP. The microtitration well format of the assay allows convenient and rapid analysis of large numbers of samples. The result of the test is obtained as numeric values making the interpretation of the result objective and easy.

In other studies, the solid-phase minisequencing technique has been used qualitatively to detect point mutations causing cystic fibrosis [19] and aspartylglucosaminuria [20], and it proved to be very suitable for routine diagnostics. In these studies the alleles that were present in predefined ratios (2:0 homozygote, 1:1 heterozygote), were unequivocally determined by incorporation of labelled dNTPs in the minisequencing reaction. The ratio between the incorporated nucleotides directly reflects the ratio between the two sequences present in the PCR product. In the present study this property of the minisequencing method was used to determine the proportions of the mutated and wild-type mtDNA in the MERRF-samples.

In the present study the highest percentage of mutated mtDNA, 72%, was detected in the leukocyte samples from a child (patient III/6), who already had the first symptoms of MERRF at the age of 1 year. The patients with severe disease and relatively early ages of onset (II/2, II/3, II/5, III/4) also had high percentages of mutated mtDNA, up to 57%. Patient III/7, who had 31% mutated mtDNA, is still considered healthy at the age of 1 year.

Interestingly, in the third generation, the symptoms began at an earlier age than in the second generation, and two out of the three children studied (III/4, III/6) had more mutated genomes than their mothers. This observation may suggest a genetic advantage of the mutated population of genomes over the normal ones, but needs to be studied in more families. The lowest percentage of mutated mtDNA, 9%, was detected in the samples of a healthy man (II/4) from the maternal lineage. In the samples analyzed here, the percentages of the mutated mtDNA detected were generally lower than usually reported in MERRF [2, 5, 9, 11]. However, according to our result from analysis of patient II/3 and a previous report [11], the leukocytes seem to harbor less mutated mtDNA than the muscle tissue. Despite the rather low number of patients studied, our results suggest a correlation between the amount of mutated mtDNA in the leukocytes and the severity and age of onset of the MERRF syndrome.

In conclusion, we present here a convenient method suitable for detecting the tRNA8344Lys mutation for the diagnosis of MERRF. The method is also applicable for detection of other mitochondrial point mutations, independently of restriction site variations. The method allows simultaneous identification of the mutation and accurate quantification of the mutated mtDNA in the same assay.