Introduction

Duchenne muscular dystrophy (DMD) is the most common lethal childhood neuromuscular disease, and it is caused by loss of function mutations in the X-linked DMD gene that codes for dystrophin protein. Boys with DMD undergo progressive weakness and muscle degeneration that result in loss of ambulation around the age of 12, and premature death in their 20s or early 30s1.

Dystrophin-null mice (mdx) have been widely used to study the pathogenesis of DMD, and the efficacy of novel therapies. However, compared to the severe clinical manifestations observed in Duchenne boys, mdx mice present a considerably milder phenotype, with only a minor reduction in lifespan2. The zebrafish DMD model sapta222a, namely sapje, carries a point mutation in the dystrophin orthologous gene3. Sapje zebrafish present severe muscle disorganization, progressive motor dysfunction and early death by 10–12 days post fertilization (dpf)4, and thus, compared to the mdx mice, sapje zebrafish recapitulate to a greater extent the human disease. This muscle disorganization can be easily detected by birefringence under polarized light in zebrafish larvae at 3–4 dpf5. Also, the early lethality of sapje at larvae stages greatly expedites survival experiments. These features make the sapje zebrafish a powerful model system for whole organism high throughput screening. Additional advantages of sapje zebrafish over other DMD animal models are their smaller size and transparency at larvae stages, a large number of offspring every one or two weeks, and lower maintenance costs6. These advantages have likely contributed to an increased use of this model, particularly in preclinical studies focused on testing the efficacy of therapeutic candidates for Duchenne4,7,8,9.

Since sapje mutants die at the larvae stage, heterozygous (Het) progenitors are needed, and therefore proper genotyping is useful in order to establish sapje colonies and subsequent crosses. Moreover, in drug screening assays in which the aim is to rescue the dystrophic phenotype, it is often mandatory to genotype the larvae4. The most common methods to genotype sapje mutants are Sanger sequencing, and the derived cleaved amplified polymorphic sequences method (dCAPS)10, a technique that introduces or destroys restriction enzyme recognition sites by using primers containing one or more mismatches to the template DNA. The modified PCR product is then subjected to restriction enzyme digestion and the presence or absence of single-nucleotide mutations is determined by the resulting restriction pattern. However, these techniques are expensive, tedious, and/or require special equipment. Thereby, a simple genotyping method would considerably facilitate the use of sapje zebrafish in preclinical studies in laboratories and biotech companies worldwide. In this study, we aimed to develop a simple, reliable, and cost-effective genotyping protocol that would facilitate studies in sapje zebrafish.

Results

Standard HRM assay

First, we tested a fast DNA extraction method based on an alkaline lysis protocol11. Using this method, we have been able to extract DNA from fresh and paraformaldehyde-fixed zebrafish samples (larvae and adults) with enough yield and quality for PCR amplification. Next, we aimed to use the High-Resolution Melt (HRM) assay for genotyping sapje zebrafish. HRM analysis is a post-PCR analysis method used to identify variations in nucleic acid sequences. This method is based on detecting small differences in PCR dissociation curves by measuring the change of fluorescence intensity in the transition from double-stranded to single-stranded DNA. This technique is highly convenient when analysing a large number of samples since it does not rely on electrophoresis for analysis. It is commonly used to discriminate single point mutations in PCR products up to 300 bp12. In fact, it has been previously used to genotype mdx mice carrying a C-to-T point mutation13, and therefore, we aimed to apply this method for genotyping sapje zebrafish.

To maximize the melting difference between wild-type (WT) and sapje samples, we designed a primer pair that amplifies a short PCR product of 166 bp (standard HRM primer pair, Table 1). Subsequently, we performed an HRM analysis on three samples with known genotypes: a WT, a sapje, and a Het sample (Fig. 1). This analysis was able to discriminate between the homozygous and Het samples, but it was not able to discriminate between sapje and WT homozygous samples (Fig. 1a). This result was not wholly unexpected since the single A/T point mutation present in sapje zebrafish is the most challenging base change for HRM detection due to the small melt curve shift (< 0.2 °C). In contrast, mdx mice13 carry a C/T point mutation that results in a broader temperature shift (0.8 °C).

Table 1 PCR primers and expected PCR products.
Full size table
Figure 1
figure 1

Representative high resolution melting graphs from standard PCR-HRM assay. (a) HRM difference melt plot of three zebrafish samples with known genotypes: a wild-type, a sapje, and a heterozygous sample. PCR-HRM analysis was performed in triplicate with the heterozygous cluster as a baseline. The samples were assigned to two different clusters: homozygous (wild-type and sapje, red) and heterozygous (green). (b) HRM difference melt plot of three DNA samples from zebrafish larvae diluted 1:1 with a sapje sample. This analysis was performed in triplicate, with the wild-type as a baseline. The samples were assigned to three different clusters corresponding to sapje (red), heterozygous (green), and wild-type (blue) genotypes.

Full size image

We then used a common strategy to identify the A/T base change by HRM, which is to dilute sample DNA with homozygous DNA14. Indeed, when we diluted the three DNA samples with a known sapje sample (1:1) and performed HRM analysis, we were able to identify the three different genotypes precisely (Fig. 1b). However, the addition of an extra PCR and dilution step per sample substantially increases the risk of contamination, the hands-on time, and the costs per sample, and thus, we decided to try a more straightforward approach.

Primer competition PCR

In order to avoid the extra dilution step, we decided to use an approach based on the primer competition PCR technique (pcPCR), in which allele-specific primers compete in a single PCR that result in genotype specific products that differ in length15. This technique has been previously optimized for genotyping several strains of DMD mice carrying different point mutations16. Hence, we designed a primer trio that produces different sized amplification products for WT and sapje alleles, by using the non-specific synthetic tails at the 5′ ends described in Shin et al. 16, and allele-specific nucleotides at the 3′ end of the reverse primers (A or T, Table 1). The optimal annealing and extension temperature was elucidated through gradient PCR with temperatures ranging from 60 to 65 °C and subsequent agarose gel electrophoresis. Lower annealing temperatures resulted in the amplification of non-specific bands in homozygous samples (see Supplementary Fig. S1 online). In contrast, at 65 °C, pcPCR products analysed by 3% agarose gel electrophoresis showed the expected pattern for WT, Het, and sapje samples with no detection of additional bands (Fig. 2a). Genotypes of several samples were also resolved by Sanger sequencing using HRM-for and HRM-rev primers (Table 1), with concordant results (Fig. 2b, Table 2). Likewise, birefringence analysis showed that zebrafish genotyped as sapje specifically presented a disorganized muscle pattern that is characteristic of the dystrophic phenotype, while zebrafish genotyped as WT or Het showed a regular muscle pattern (Fig. 2c).

Figure 2
figure 2

Validation of primer competition PCR for sapje zebrafish genotyping. (a) Genotypes of two wild-types, two heterozygous and two sapje mutants, were resolved by agarose electrophoresis after pcPCR. PCR products corresponded with the expected band sizes. No bands were observed in the non-template control (NTC), and no primer dimers were detected. (b) Representative Sanger analysis of wild-type, heterozygous, and sapje samples. (c) Images illustrate representative unaffected and sapje phenotypes of zebrafish larvae by birefringence analysis.

Full size image
Table 2 Genotype and phenotype analyses in a subset sample of zebrafish larvae.
Full size table

Next, we aimed to determine the sensitivity and reliability of this method. Additionally, we wanted to establish whether HRM analysis could also be used to resolve zebrafish genotypes after pcPCR amplification since this would substantially reduce the hands-on time required for agarose electrophoresis. To test this, we used DNA from 60 zebrafish larvae, 17 of which presented a dystrophic phenotype (Table 2; Supplementary Table S1 online). These samples underwent pcPCR as described above, and PCR products were analysed in triplicate by HRM assay (Fig. 3a). All samples tested were amplified and resulted in a corresponding melting curve. HRM assay was able to efficiently discriminate WT, Het, and sapje samples in an unbiased manner, based on pcPCR product melt curve profiles. After HRM analysis, 97% of samples were automatically assigned to one of the three clusters corresponding to WT (cluster 3), sapje (cluster 2), or Het genotypes (cluster 1). Two samples (#28 and #40) were termed as unresolved since their replicates were assigned to two different clusters. The HRM assay was re-run a second time, and this allowed resolution of sample #40 as Het since all the three replicates assigned to this cluster, with 91, 94, and 98% confidence. The genotype of each cluster was determined, by performing Sanger analysis in at least five samples per cluster, and later on by agarose electrophoresis of pcPCR products (Table 2; Supplementary Table S1 online). The reliability of the melt curve analysis in separating each genotype from the other two was 100% in 58 samples, with a 90% confidence threshold reached in at least one replicate per sample. The sensitivity of the melt curve analysis was 98%, due to one false negative in a replicate from sample #28. Sensitivity was calculated as the percentage of sapje replicates correctly identified as positive by the assay (50/51). The specificity of the pcPCR-HRM assay was 100%, calculated as the percentage of phenotypically unaffected replicates (Het or WT) correctly identified as negative by the assay (129/129).

Figure 3
figure 3

Genotyping sapje zebrafish by primer competition PCR followed by HRM and agarose electrophoresis. (a) HRM difference melt plots of 60 zebrafish larvae performed in triplicate with the wild-type cluster as a baseline. 58 samples were automatically assigned to three different clusters with 100% reliability, but two samples were inconclusive. Wild-type (blue), heterozygous (red), and sapje (green) clusters are identified on the melt plot. (b) 100% of the genotypes were resolved after running pcPCR products (same as in a) in a 3% agarose gel electrophoresis. Cropped images from two different gels are shown. White lines depict non-consecutive lanes, while yellow asterisks depict sapje genotypes.

Full size image

We then analysed the 60 PCR products by agarose electrophoresis (Fig. 3b), and these results were concordant with the genotypes previously resolved by HRM analysis. Gel electrophoresis of the pcPCR products from samples #28 and #40 was able to identify them unquestionably as a sapje and a Het, respectively. Moreover, these genotypes were concordant with data obtained by Sanger sequencing and birefringence phenotyping. In summary, the resolution of pcPCR products by combining HRM and agarose electrophoresis allowed the genotyping of all the zebrafish samples with 100% accuracy. Indeed, our technique resulted in 100% genotype–phenotype correlation, where the 17 samples identified as sapje by the birefringence analysis were genotyped as A/T mutants with the pcPCR method (Fig. 3; Supplementary Table S1 online).

Interestingly, we found that the melting temperature (Tm) of the pcPCR products from sapje samples (79.07 ± 0.08) were significantly higher compared to WT (78.70 ± 0.04) and Het samples (78.77 ± 0.05), with an increase in the mean Tm of 0.37 °C and 0.3 °C compared to WT and Het samples, respectively (p < 0.0001; Fig. 4). We also observed minor, although statistically significant differences in the Tm between WT and Het samples (0.07 °C, p = 0.0079). This result is in agreement with the Tm values calculated by Oligo Calc in silico calculator (https://biotools.nubic.northwestern.edu/OligoCalc.html), which is 0.3 °C higher in sapje PCR products (75.9 °C in WT vs. 76.2 °C in sapje). Overall, our results indicate that Tm analysis is able to discriminate sapje samples from healthy unaffected samples (WT or Het), without the need to perform a thorough HRM analysis or agarose electrophoresis. However, differentiation between WT and Het genotypes may be challenging when only Tm values are used for analysis. Genotype-specific Tm values were also used for determining the intra- and inter-assay variability on three runs on different days (Table 3). The intra-assay coefficients of variation (CVs) and inter-assay CVs showed values ranged from 0.04 to 0.13%, and from 0.03 to 0.07%, respectively. These values are well below the maximum acceptance criteria for precision in medicine (< 5–10%).

Figure 4
figure 4

Analysis of melting temperatures (Tm) from pcPCR products. Average Tm values of 60 zebrafish larvae measured in triplicate: wild-type (blue), heterozygous (red), and sapje (green) genotypes.. Data are expressed as mean ± SD. One-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis.

Full size image
Table 3 Determination of the intra- and inter-assay reproducibility.
Full size table

We next wanted to validate our method in another model with a different point mutation. To do this, we used the mdx mouse model of DMD, which carries a nonsense C/T mutation in exon 23. For this study, 9 mouse tails were used, from WT (n = 3), Het (n = 3) and mdx mice (n = 3). We used a primer trio specifically designed for pcPCR, described in a previous study16 (Fig. 5a), and we applied the same pcPCR protocol as the one optimized for zebrafish samples. We found that using a hot-start Taq and a 65 °C annealing/extension temperature resulted in robust amplification of DNA with expected sizes with no primer dimers or unspecific bands (Fig. 5b). HRM assay was able to efficiently discriminate WT, Het, and mdx samples in an unbiased manner, based on pcPCR product melt curve profiles (Fig. 5c). After HRM analysis, 100% of samples were automatically assigned to one of the three clusters corresponding to WT (blue), mdx (green), or Het genotypes (red). The melting temperature (Tm) of the pcPCR products from mdx samples (77.27 ± 0.06) were significantly lower compared to WT (77.67 ± 0.06) and Het samples (77.43 ± 0.06; p < 0.0001 for all comparisons), with a decrease in the mean Tm of 0.40 °C and 0.17 °C compared to WT and Het samples, respectively (Fig. 5d). We also observed statistically significant differences in the Tm between WT and Het samples (0.23 °C, p < 0.0001). This result is in agreement with the Tm values calculated by Oligo Calc in silico calculator, which is 0.3 °C lower in mdx PCR products (75.2 °C in WT vs. 74.9 °C in mdx). Overall, these results indicate that our method can be applied to a range of single-nucleotide substitution mutations in different species.

Figure 5
figure 5

Validation of the pcPCR-HRM technique to genotype mdx mouse. (a) Primer sequences and expected product sizes from Shin et al. 16. Bold letters represent allele-specific nucleotides at the position of the mdx mutation. Non-specific synthetic tails are underlined. (b) Genotypes of three wild-type, three heterozygous and three mdx mice resolved by 3% agarose electrophoresis after pcPCR. (c) HRM difference melt plots from 9 mice (same as in b) performed in triplicate with the mdx cluster as a baseline. Wild-type (blue), heterozygous (red), and mdx (green) clusters are identified on the melt plot. (d) Tm values from pcPCR products (same as in c): wild-type (blue), heterozygous (red) and, mdx (green) genotypes. Data are expressed as mean ± SD. One-way ANOVA with Tukey’s multiple comparisons test was used for statistical analysis.

Full size image

Discussion

Sapje zebrafish is becoming a preferred model for screening novel therapeutic candidates for DMD, and it is also used to study the pathogenic mechanisms of this disease4,7,8. Sanger sequencing is the classical genotyping method for sapje zebrafish, which carry an A/T point mutation, but this technique may prove costly and time-consuming. Here, we describe a highly reproducible PCR protocol that uses a primer trio for sapje genotyping. This method can be easily implemented in a routine laboratory setting, since it may be resolved by standard agarose electrophoresis. Moreover, it can also be used in combination with HRM analysis, which enables unbiased identification of a large number of samples. Indeed, in a single PCR-HRM setup, 128 larvae can be genotyped in triplicate in a 384-well plate with minimal hands-on time. We estimate that these samples could be processed for DNA extraction and analysed by HRM in less than 4.5 h. Given that a large number of samples is relatively common when working with zebrafish, we propose the pcPCR followed by HRM analysis as the most convenient method for genotyping the sapje DMD model. Compared to agarose electrophoresis, HRM is less time-consuming, and an automatic cluster assignment removes any potential bias in the interpretation of results.

In the current pcPCR-HRM assay, both the melting curve profiles and Tm values can be used to discriminate sapje samples. The accuracy of the melt curve analysis was 97% (2 unresolved samples), with a 94% sensitivity and 98% specificity. Subsequent agarose electrophoresis analysis was able to identify the unresolved samples, reaching 100% accuracy. The intra- and inter-assay variability analysis of Tm values rendered minimum CV values, which indicates an excellent precision and repeatability of the pcPCR-HRM method.

Our genotyping approach presents several advantages over currently used methods to genotype animal models. Compared to conventional pcPCR, our method has the capacity of high specificity and sensitivity. It is also more versatile, faster and unbiased, while allowing easy transition to high-throughput scale. Compared to standard HRM, our method is able to detect challenging point mutations, such as A/T substitutions. It is also more robust and versatile, since genotypes can be resolved through different analysis, alone or in combination, i.e., Tm analysis, HRM difference plots, and/or agarose electrophoresis. Most interestingly, our method is compatible with poor-quality DNA, while standard HRM is susceptible to variations in DNA quality, leading to unreliable genotyping17.

In summary, the genotyping method described in this study provides a precise and straightforward alternative to differentiate between WT, Het and mutant genotypes. Subsequent genotype identification may be performed by HRM assay, agarose gel electrophoresis, or even Tm analysis, which makes this method extremely versatile, user-friendly, and accessible at minimum cost and hands-on time. HRM analysis may be used in combination with pcPCR to genotype A/T and C/T point mutations present in the zebrafish and mouse models of DMD3,16. Furthermore, this approach is applicable to detect a variety of point mutations and even small insertions or deletions in different species. Implementation of this technique would require a careful primer design for specific PCR amplification, with non-specific tails to generate different-sized products. Using a hot-start Taq polymerase and optimizing the annealing temperature are critical measures for obtaining specific PCR products. More sophisticated PCR cycling protocols may be implemented if needed, such as touch-down or touch-up gradient amplification protocols18. Given its robust performance with crude DNA extracts, our approach may be particularly well-suited for detecting single nucleotide variants in poor-quality samples such as ancient DNA or DNA from formalin-fixed, paraffin-embedded material.

Methods

Animals

The dmd/sapta222a mutant strain (sapje) was obtained from the Tübingen Stock Collection (Tübingen, Germany19). Sapta222a heterozygotes were raised and maintained in the zebrafish facility at Biobide following the European Directive (2010/63/EU) for the protection of animals used for scientific purposes and standard procedures, as previously described20. C57BL/10ScSn-DMDmdx/J mice (mdx) and C57BL/10ScSnJ mice (wild-type), were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). All experiments were approved by the Ethical Committee for Animal Experimentation at Biodonostia. Crosses of Sapta222a heterozygotes results in WT (25%), Het (50%), and homozygous Sapta222a larvae (sapje, 25%). Larvae used for these experiments were 6 dpf or younger.

Birefringence analysis

Muscle damage in 3–4 dpf sapje larvae can be readily detected through a disrupted birefringence pattern5. Birefringence is the bright light pattern produced by the diffraction of polarized light through the muscle sarcomeres. For birefringence analysis, larvae were anesthetized with 0.02% tricaine, carefully aligned between two glass-polarizing filters, and viewed with a stereomicroscope while one filter was rotated to maximize birefringence. Larvae consistently showing bright, well-organized somites were phenotyped as unaffected, while those displaying patchy areas of disrupted and disorganized somites were phenotyped as sapje.

DNA extraction

Crude DNA was extracted from whole zebrafish larvae or mouse tails using fast alkaline lysis protocols11,21. Briefly, zebrafish larvae were lysed in 20 µl of 50 mM NaOH and heated at 95 °C for 10 min. Mouse tails (~ 2 mm) were lysed in 75 µl of 25 mM NaOH, 0.2 mM EDTA (pH 12) and heated at 100 °C for 30 min. Samples were cooled to 4 °C and neutralized with 2 µl of 1 M Tris–HCl at pH 8.0 (larvae), or 75 µl of 40 mM Tris–HCl at pH 5.0 (mouse tails).

Primers design

A primer pair was designed for High-Resolution Melt (HRM) and Sanger analyses, with the sapje point mutation located in the middle of the PCR product. For the primer competition PCR (pcPCR), a trio of primers was designed consisting of a common forward primer (pcPCR-For), a mutant allele-specific reverse primer (pcPCR-Mut-Rev), and a wild-type allele-specific reverse primer (pcPCR-WT-Rev). Specific primers were designed using Primer Express software (Thermo Fisher), with a 58-60 °C Tm criteria. Then, short GC-rich sequences were added to pcPCR-For and pcPCR-Mut-Rev primers, and a non-specific synthetic tail was added to the pcPCR-WT-Rev primer to allow size discrimination of WT and mutant alleles, based on a previous study16. Table 1 shows the sequences of all custom primers used in this study.

Standard PCR for HRM assay and Sanger analysis

Standard PCRs were carried out with Precision Melt Supermix (Bio-Rad), 300 nM of each primer (HRM-For and HRM-Rev, Table 1) and 1 ng/µl of DNA sample. PCR was run according to the following conditions: an initial denaturation step at 95 °C for 3 min followed by 39 cycles of 95 °C for 10 s and 60 °C for 1 min. HRM analysis was performed in triplicate on a CFX384 Touch PCR equipment (Bio-Rad), over a temperature gradient of 65–95 °C with a 0.1 °C increment, pausing for 5 s before each increment. Genotypes were resolved by examining normalized and difference melt plots using the Precision Melt Analysis software (Bio-Rad). Sanger sequencing was carried out by the Genomic Platform of Biodonostia Institute, using the HRM-Rev primer and a 16-capillary ABI 3130xl platform (Applied Biosystems) according to the manufacturer’s protocol. Some samples could not be resolved by Sanger sequencing, likely due to the crude DNA extraction method.

Primer competition PCR

Primer competition PCR (pcPCR) was carried out with Precision Melt Supermix (Bio-Rad) with a hot-start Taq DNA polymerase (iTaq), 200 nM of each custom-designed primer (one forward and two reverse primers), and 5% of DNA solution. The PCR protocol comprised an initial denaturation step at 95 °C for 3 min, followed by 34 cycles of 95 °C for 10 s and 65 °C for 1 min. Primer sequences and expected PCR products for WT and mutant samples are detailed in Table 1 for the zebrafish model, and in Fig. 5 for the mouse model. HRM analysis was performed as described above, and PCR products were subsequently resolved by 3% agarose gel electrophoresis. Repeatability of the pcPCR-HRM analysis in zebrafish samples was assessed by determination of the intra- and inter-assay CV of the Tm of three runs carried out on different days.

Statistical analysis

Data distribution was evaluated with D'Agostino & Pearson omnibus normality test (GraphPad Prism 6). Statistical significance was determined using One-Way ANOVA followed by unpaired Tukey’s posthoc test. The adjusted P values of less than 0.05 were considered statistically significant.