With the discovery of point mutations as a cause of disease, the need arises for methods to detect them with high sensitivity. Especially when the mutated allele is present in extremely low amounts, such as in DNA extracted from toxic thyroid nodules, which is often contaminated with surrounding tissue, blood and connective tissue, the sensitivity of the method to detect the mutation becomes increasingly important. Although numerous methods for mutation detection such as denaturing gradient gel electrophoreses (DGGE), temperature gradient gel electrophoreses (TGGE), single-strand conformation polymorphism (SSCP), constant denaturing gel electrophoreses (CDGE), chemical mismatch cleavage (CMC) and direct sequencing exist, they differ in sensitivity, ease of application and cost effectiveness. Moreover, these basic techniques are subject to constant improvement [eg, doublegradient (DG)-DGGE](1), heteroduplex analysis (HA)(5) and improved fluorescent dye chemistry (8). It has already been shown that DGGE is more sensitive than SSCP for the detection of point mutations, and recent studies comparing the sensitivity of DGGE, SSCP and CMC found that DGGE and CMC were slightly more sensitive than SSCP (9, 10). Our goal was to compare the sensitivity of DGGE to direct sequencing in the detection of mutated alleles among wild-type (WT) alleles, which to our knowledge, has not been carried out so far. First, the capability of DGGE in detecting the thyrotropin (TSH) receptor mutations Val656Phe, Asp633Glu, Thr632Ile, Ile486Phe, Asp619Gly, Ile630Leu, Ala623Val or the deletions 613–621 and 658–661 were determined. Secondly, the detection limit of DGGE and direct sequencing for identifying the mutant allele among WT alleles was compared utilizing the point mutations Val656Phe, Asp633Glu, Thr632Ile and Ile486Phe.

The mutant DNA was prepared by means of commercially available kits (QIAamp® Blood and Tissue Kit; Qiagen, Hilden, Germany) from toxic thyroid nodules or plasmids containing the mutated receptor and the WT DNA from healthy thyroid tissue surrounding the nodule or from plasmids with the WT TSH receptor sequence. The template consisted of a mixture of WT DNA containing decreasing concentrations of DNA from a toxic thyroid nodule or mutated plasmid DNA (50%, 25%, 12.5%, 6%, 3%, 1.5% or 0.8%). Mutant plasmid DNA is not subject to the contamination with WT DNA mentioned above and can therefore be better quantitated, while the genomic DNA better reflects the situation in vivo, in which a 1: 1 ratio of WT to mutant allele rarely exists. The mutation contents of the DNA derived from toxic thyroid nodules described in this article are referred to as putative because the exact mutation content cannot be determined due to unknown amounts of contamination with WT DNA. The DNA concentration in the template of either plasmids or genomic DNA was measured photometrically and then mixed in the appropriate ratios before the