DNA usually exists as double stranded macromolecule in living organisms but it can form, transiently, a single-stranded structure during vital cell processes such as replication, transcription and DNA repair [1,2]. Single stranded DNA (ssDNA) is much more flexible than double stranded DNA and can be folded easily [3,4]. It is proposed that, remarkable characteristics of the ssDNA such as enzymatic activities are as a result of its secondary structure [5-7]. This unique property of the ssDNA has made it a suitable tool for different applications in molecular genetic studies [8]. In recent years its application in medical diagnosis, treatment of disorders and nanotechnology have attracted attention and is expected to increase [9-11]. Despite the importance of knowledge regarding the ssDNA secondary structure, in clarifying precise mechanism of different cellular processes and increasing the efficiency of it in various applications, our understanding of thermodynamic properties of ssDNA secondary structure is very limited [12,10].

It is suggested that elongation of short oligodeoxynucleotides to a very long nucleic acids, in the presence of only dNTPs and DNA polymerase, is one of the results of secondary structure formation in the single stranded end of these oligodeoxynucleotides [13,14]. Liang et al. have proposed a terminal hairpin formation and self-priming extension (THF-SPE) model for efficient DNA elongation in this condition [14]. According to this model, a terminal hairpin structure at the 3′-end of DNA can act as a self-priming complex and lead to elongation of the molecule in the absence of specific primers. It is suggested that defining the precise mechanism of this event can be helpful to increase our knowledge of molecular evolution in the past [12].

During our investigation on hearing loss, trying to link the disease to a particular gene in different families, we found that TMC1 gene may be defected in a family. While sequencing patient’s DNA, we encountered an unusual sequencing result for exon 5 of this gene. This unusual sequencing result suggested amplicon secondary structure construction and elongation during the sequencing process. This is the first report of its kind in that ssDNA can fold and elongate during cycle sequencing.

Several families with hearing loss were admitted into our study. Informed consent forms were filled for each family. From each person 5-10 ml peripheral blood was taken in EDTA as anticoagulant. DNA was extracted using salting out method [15]. Sequencing was done by KBC (KBC, Tehran, Iran) using Big dye terminator kit™ (Life Technologies, USA) according to manufacturer’s recommendation and the reaction product was run on a 3130XL Genetic Analyzer (Life Technologies, USA). The sequence of forward and reverse primers used to perform PCR and sequencing TMC1 exon 5 were: TMC1 E5F 5'-GTACTCCGTCCCTTAAATGGC-3' and TMC1 E5R 5'-GTCTTTCTGAGTGAAAGCGAGG.

BLASTN was used to find the origin of the 270-bp extra segment seen in the sequencing result [16]. Secondary structure prediction was performed by Mfold web server (version 3.5, online: http://www.bioinfo.rpi.edu/applications/mfold/old/dna/) [17].

During our hearing loss study, linkage analysis using short tandem repeats was performed for several investigated families. We found linkage between the hearing loss and TMC1 gene in a family. PCR reaction was used to amplify all 24 exons of TMC1 gene in DNA sample of a patient suspected of having mutation in this gene followed by sequencing. During sequence analysis we encountered an unusual sequencing result for exon 5 of the gene. Analysis of the sequencing result showed that a 270 bp extra segment is formed. We checked this phenomenon by repeating PCR and running the PCR product on agarose gel to see if it has been present in the patient’s DNA or it may have occurred during PCR or sequencing. No size increase was observed by gel electrophoresis (Figure 2). This process was repeated few times and no change was seen. This suggested that the elongation has happened during the sequencing reaction.

Further investigation and alignment of the DNA sequence showed that the 270 bp extra segment had been initiated just after the last nucleotide of reverse primer (Figure 1). BLASTN analysis showed that this extra sequence is an inverted repeat of the first part of this amplicon. The position of forward and reverse primers and extra segment on sequenced segment are shown in Figure 1.

Figure 1: The position of forward and reverse primers, extra segment and homologous regions on the sequenced segment. Forward and reverse primers are shown by arrows. The sequence of homology regions 1, 2 (HR1, 2) and their borders are shown in boxes. Homologous regions are highlighted. Forward and reverse primers are shown by arrows. As shown in this figure, homologous region 2 is a part of the reverse primer sequence. Dashed arrows indicate the direction of sequences.

We re-sequenced the same exon ten times, using the forward primer, to confirm the initial results. We expected to see a 509 bp length sequencing result if no elongation had occurred. However, we observed that the length of three out of ten sequences obtained were longer than expected. The length for the other seven were as expected (i.e. 509 bp). Interestingly no size increase was observed using the reverse primer.

Comparing the DNA sequences of the extra segment and its origin borders led us to find two homologous regions (HR1, 2). One of them was located 279 bp after the beginning of the forward primer with a length of 11 bp. The second one was at the 11 bp end of the 509 bp fragment (Figure 1). To find out if formation of secondary structure can explain the unusual sequencing result, DNA sequence of this region was submitted to Mfold web server. None of the suggested secondary structures was what we had expected. However, when RNA equivalent of the sequenced region was submitted, a secondary structure was predicted which could explain such an elongation during sequencing amplification. In this structure, 11 consecutive nucleotides at the 3' end of the nucleic acid (HR2) have been paired with another 11 bp homologous nucleotides just after the replicated segment (HR1) (Figure 3).

Figure 2: Gel electrophoresis result. The result of gel electrophoresis shows that the size of the amplicon is between 500-600 bp. As regards the sequenced segment is about 780 bp, it can be suggested that elongation of amplicon has happened during the sequencing process.

Figure 3: One of the predicted structures for the RNA equivalent of amplicon. In this structure, 11 nucleotides in the 3' end of the nucleic acid have been paired with nucleotides just after the replicated segment. Pairing of these regions is magnified in the right-hand image.

Based on our findings, we are proposing presence of a potential mechanism which explains the unusual sequencing observed in our study. Figure 4 shows schematic diagram of our hypothetical amplicon secondary structure and potential mechanism for the formation of the additional segment. As shown in Figure 4, terminal hairpin formation can be followed by elongation using the first part of the amplicon as a template.

Figure 4: Schematic diagram of our hypothetical amplicon secondary structure and mechanism of its elongation during sequencing. As shown in this figure, terminal hairpin formation can be followed by elongation using the first part of the amplicon as a template.

In this study we described an unusual sequencing result of exon 5 of the TMC1 gene in a patient with autosomal recessive non-syndromic hearing loss. This sequencing result showed that the sequenced segment is longer than was expected. The results of the BLASTN showed that the extra sequence is an inverted repeat of the first part of the amplicon.

At first, this result indicated an inverted duplication which may be present in the patient’s DNA. This hypothesis was rejected since gel electrophoresis finding didn’t support size increase during PCR. Observation of this unusual result, in normal DNA samples, showed that this event cannot be related to the hearing loss. These results, suggested that the extra sequence could be the result of a secondary structure formation in the single stranded amplicon and elongation. This could have occurred through terminal hairpin formation and selfpriming extension (THF-SPE) model suggested by Liang et al. [14]. One explanation for observation of this result in the sequencing but not in the PCR result may be that during PCR both forward and reverse primers are present and the normal size can participate in exponential amplification but the elongated types will amplify linearly which will not be detectable by agarose gel. However, absence of reverse primer and presence of only forward primer during the sequencing can explain formation of this molecular event during sequencing. One possible explanation is that only non-exponential amplification is generated during cycle sequencing.

Repeating sequencing with similar condition, with forward primer, showed that only three out of ten sequences were longer than we expected. It shows that structural stability of this hypothetical construction is not enough to form most of the time. However this structure can be one of the predicted constructions for the single stranded amplicon.

Inability of Mfold web server to predict this structure can be explained by indefinite thermodynamic parameters that are used by this program [12,10]. It seems that much more knowledge of the ssDNA properties is needed to predict all likely structures. Evidence for ssDNA folding in special conditions obtained from designed experiments or by sheer chance, like what was obtained in this study, can be used to investigate different parameters needed to identify accurate conformation of the ssDNA in different situations.

Although none of the predicted secondary structures for this ssDNA by Mfold web server was similar to our hypothetical structure, our hypothetical construction was predicted for the RNA equivalent. Similarity between properties of RNA and ssDNA can explain this result. Therefore the parameters used in the Mfold for DNA cannot be verified.

Continuing this study by performing experimental and computational investigations can prove formation of our hypothetical secondary structure and help us to determine actual conformation of this amplicon as a clue for increasing our knowledge about the ssDNA secondary structure formation.

From some years ago it has been shown that the nucleic acid secondary structure is an important interfering factor of different nucleic acid hybridization assays [18,19]. Lane et al. showed that amplicon secondary structure can interfere with hybridization of targets to their respective microarray oligonucleotide probes [20]. Measuring of hybridization kinetics for some designed probe-target pairs by Gao et al. also showed that probe and target conformations can affect the DNA hybridization kinetics in different situations [21].

Previous studies also have shown that secondary structure formation in the sequencing template can affect DNA polymerase activity and lead to no signal or weak signal in the electropherogram [22]. Here, we presented another result of ssDNA folding during DNA sequencing. It seems that specific forms of folding in the template can cause elongation during sequencing. More information about the detailed mechanism of DNA polymerase is needed to understand in what situations the secondary structure will result in halting the DNA polymerase and when it lead to elongation of template.

Until now several studies have been performed to predict folding of nucleic acids and some computational and experimental methods are suggested for this purpose but none of these methods can certainty predict ssDNA secondary structure [17,23-26]. Dong et al. demonstrated that using of both experimental method by using Taq Exo 5'-nuclease and Mfold web server can help to increase the probability of correct folding prediction [10]. Studies of Liang et al. also have shown that the investigation of catenanes formed from two ssDNA molecules can provide worthwhile information about the ssDNA folding [12]. Our finding suggests a new way for studying the ssDNA folding. Investigation of the ssDNA elongation in the presence of dNTPs and DNA polymerase can help to detect terminal hairpin loop formation.

Elongation studies by using DNA segments with different sequences and various conditions can increase our knowledge about factors affecting ssDNA secondary structure formation.

This work was supported by the Kawsar Human Genetics Research Center. We gratefully acknowledge all staffs in Kawsar Center and Medical Genetic laboratory of Dr. Zeinali that contribute to this study.