|Home | About | Journals | Submit | Contact Us | Français|
Small parasites and larval stages pose a problem for molecular analyses because limited amounts of DNA template are available. Isothermal methods for faithfully copying DNA have the potential to revolutionize studies of such organisms. We evaluated the fidelity of multiple displacement amplification (MDA) for amplifying DNA extracted from a single miracidium of Schistosoma mansoni. To do this we genotyped DNA extracted from 28 F1 miracidia following MDA using 56 microsatellite markers. Because these miracidia were obtained from a cross between a male and female worm of known genotypes, we were able to predict the alleles present in the progeny and quantify the genotyping error rate. We found just 8/1568 genotypes deviated from Mendelian expectations. Furthermore, because one of these resulted from a genuine mutation, the error rate due to MDA is 7/1568 (0.45%). We conclude that many hundreds of microsatellites or other genetic markers can be accurately genotyped from a single miracidium using this method, greatly expanding the scope of population genetic, epidemiological and evolutionary studies on this parasite.
The genome sequence and genetic map of Schistosoma mansoni [1;2] open the door to genome-wide analyses involving hundreds or thousands of genetic markers. However, adult worms live in the mesenteric veins of human hosts. Thus, only larval stages, from which limited amounts of DNA can be extracted, are available for genetic analyses when working with samples from humans. Researchers have used two approaches to get around this problem. First, eggs obtained from human hosts can be hatched in the laboratory and the miracidia used to infect the appropriate freshwater snail. Clonally produced cercariae shed from snails or adult worms generated from passage through rodents can then be used for genetic analysis. This approach is cumbersome and time consuming, and has the additional problems of creating genetic bottlenecks and possibly causing host (snail or rodent) induced selection at pertinent loci . Hence, the parasites genotyped may be poorly representative of those initially sampled. Second, PCR based methods have been used to amplify and genotype DNA extracted from a single egg or miracidium [4–7]. Using this approach up to 21 loci have been genotyped from a single miracidium . However, PCR genotyping of miracidia using this approach requires careful optimization and pooled PCR reactions. Observations of strong deviations from Hardy-Weinberg proportions in the genotypes generated may indicate that template is limiting or of poor quality and suggest caution is needed when using these methods . Furthermore, these methods [4,5] do not leave enough template DNA for repeat genotyping, which precludes assessment of genotyping error rates in field studies.
Isothermal amplification of template DNA using Phi-29 DNA polymerase provides a simple approach for generating microgram amounts of DNA template from a few ng of starting material . This method is now widely used for copying DNA for use in human genetic studies and has been carefully validated for use in a variety of organisms . This approach has also been used to generate large amounts of DNA template from single sperm  and from a variety of parasite species, such as Plasmodium sp.  and Trypanosoma sp . We have evaluated the use of Phi-29 to faithfully replicate DNA extracted from a single miracidium. To measure the reliability of this approach we genotyped F1 miracidia generated from a genetic cross between parental worms in which genotypes were known. Therefore, we were able to predict the expected genotypes at each locus genotyped and quantify genotyping error rates. We observed just 7 discrepant genotypes due to whole genome amplification among 28 miracidia genotyped at 56 microsatellites markers (1568 genotypes) giving an error rate of 0.45%. While this validation experiment is specific to S. mansoni, we expect that similar success will be found for other schistosome species and for helminth parasites for which only eggs or larval stages are easily accessible.
We set up a genetic cross between single genotypes of a female schistosome from the NMRI (Puerto Rico) strain with a male schistosome from the LE (Brazil) strain to generate the F1 progeny. To do this we exposed snails (Biomphalaria glabrata) to a single miracidium. After 28–29 days, the cercariae were collected by exposure to light. We extracted DNA from 20–50 cercariae from each shedding snail in 50 μL of 6% chelex containing 0.2 mg/mL of proteinase K. This was incubated at 56°C for 2 hours and at 100 °C for 8 minutes. We sexed cercariae by amplifying cercarial DNA using the W1 primers, which are specific for a repetitive region in the W chromosome of female worms . As male worms are not expected to show any amplification, we ran a control PCR with the autosomal microsatellite locus sc18 (Forward CAGTTTGTCAACAAGTAACGAG and reverse AAAAGACAAATTCGTTGTGG) in order to confirm the DNA was extracted. PCR was performed with 15 μl reactions containing 2.4 μl of extraction supernatant, 1X PCR buffer, 1.5 mM MgCl2, 0.2 mM each of dNTP, 0.4 μM each primer, and 0.75 units (0.15 μl) Taq DNA polymerase (Takara Shuzo Co., Otsu, Shiga, Japan). PCR cycling was 95°C for 3 min, once; 94°C for 45 sec, 54°C for 30 sec, 72°C for 45 sec, 35 times; 72°C for 7 min, once. We then infected a hamster with LE male cercariae of a single clonal genotype and NMRI female cercariae of a single clonal genoytpe to set up the cross. After 45 days, the hamster was perfused and the parental worms collected. We removed the liver, hatched the eggs in freshwater, and collected single miracidia in 10μL H20 in 0.5mL microcentrifuge tubes for DNA extraction.
We extracted DNA from individual adult worms and 29 F1 miracidia in 200 μL and 20 μL of 5% chelex, respectively, as described above. We then used Phi-29 polymerase to expand the DNA template available using the GenomiPhi v2 DNA Amplification Kit (GE Healthcare) according to the manufacturer’s manual, using 2.5uL of the chelex extracted DNA in each reaction. The total reaction volume of 50 μL was incubated at 30°C for 4 hours and terminated at 65°C for 10 minutes. After amplification, the whole genome amplified DNA was purified through Sigma Spin Post Reaction Purification columns (Sigma, St Louis, Missouri). The final volume of the DNA product was adjusted to 150 μL with water. DNA concentration was determined using a Hoefer DyNS Quant 200 Fluorometer and diluted to a final concentration of 15 ng/μL. Two μL (30 ng) of product were used as template for genotyping of each microsatellite marker.
We used an economical method for fluorescent labeling of microsatellites , in which oligos are unlabeled, but attached to the 5′ end of the forward primer is the M13 oligo (5′-TGTAAAACGACGGCCAGT-3′). The PCR cocktail contains a fluorescent labeled M13 primer, the reverse primer, and the forward primer with an M13 tail. The concept underlying this approach is simple: the first rounds create the specific product and then the last rounds label the product. This method allows one to order a single labeled primer (i.e., M13) that can be used to label multiple loci. We also “pig-tailed” the reverse primers by adding GTTTCTT to the 5′ ends  to minimize stutter bands. Amplifications were performed in 5 μL reactions containing 1X buffer, 1.5 mM of MgCl2, 0.2 mM each of dNTP, 0.16 μM of fluorescent label M13(-21), 0.08 μM of the forward primer, 0.16 μM of the reverse primer, 0.15 units (0.03 μL) of Takara Taq (Takara Shuzo Co., Otsu, Shiga, Japan) and 30 ng of DNA template. For each reaction, the labeled M13(-21) primer was fluorescently marked with 6-FAM, VIC, NED, or PET. PCR cycling was as follows: 94°C (5 min), 30 cycles at 94°C (30 s)/56°C (45 s)/65°C (45 s), followed by 8 cycles 94°C (30 s)/53°C (45 s)/65°C (45 s), and 65°C for 10 min. The PCR products were pooled in groups of four and analyzed on an ABI 3130xl genetic automated analyzer with GENESCAN software (Applied Biosystems). Allele sizes were calculated using GENOTYPER software (Applied Biosystems) and visually confirmed.
For this study, we used 56 microsatellite markers distributed across the genome of the parasite (Supplementary Table 1). These markers have known chromosomal locations and are a subset of 243 microsatellite markers used to construct a genetic linkage map for S. mansoni . Forty–eight of the loci have autosomal inheritance (including 8 pseudoautosomal markers on the Z and W chromosomes) while 8 loci were located in the Z-specific region, which does not recombine with the W chromosome . When females are genotyped for these Z-specific loci, only the allele inherited from the male parent is present. Fifty-one loci were fully informative because the parental parasites did not share alleles. Thus, we could directly assess the genotyping error rate at these loci because at an autosomal locus, F1 offspring should have 2 alleles. Five loci were partially informative because the parents shared one of the alleles.
The final data set included parental genotypes for the 56 loci as well as for the 28 F1 miracidia, giving a total of 1568 F1 genotypes. We were able to chelex extract DNA from all 29 miracidia attempted. However, one sample contained 2 miracidia or was contaminated, because more than 2 alleles were seen at multiple loci. This sample was not included in the study. The results are summarized in Table 1. We observed a single allele from each of the two parents (or for Z-specific loci, a single allele from the male parent in F1 female offspring) as expected for Mendelian segregation for 1560/1568 genotypes, giving an overall success rate of 99.49% (see Fig. 1 for examples). This error rate is similar to that observed in studies in which the same individuals are re-genotyped to measure reproducibility , giving us considerable confidence in this method. Just 8/1568 (0. 51%) of genotypes showed patterns that differed from those expected under Mendelian segregation. Of these 8, discordant genotypes resulting from generation of additional non-parental alleles were observed in 4/1568 genotypes (Fig. 2a). In these cases the expected parental alleles were present, but a third non-parental allele was also observed. Discordant genotypes resulting from failure to genotype one of the parental alleles were observed in 4/1568 cases. In this situation one of the parental alleles was present but the other allele was shifted by 2 bp (2 case) (Fig. 2b) or absent (2 cases). These 8 discordant genotypes could have resulted from natural mutation events (e.g., occurring during clonal expansion of sporocysts within the snail host). Alternatively they may represent mutations occurring during MDA or slippage during the PCR reaction. To differentiate between these explanations, we repeated PCR reactions using miracidial DNA pre-MDA (i.e., a PCR using the chelex extracted DNA). In 7 cases, the discordant patterns were not observed when using extracted miracidial DNA as template for PCR. Hence, in a small number of genotypes (0.45% [95% confidence intervals: 0.22–0.92%]) MDA induced artifactual genotypes. We note that in about 6% of the genotypes the shorter allele had a lower peak height than the larger allele. These genotypes were not restricted to a subset of miracidia but were randomly distributed among individuals. This pattern appeared to result from the MDA because a repeat of the MDA from a few of these samples reversed this pattern (data not shown). A lower shorter allele could be problematic for scoring loci with a large degree of stutter. However, we feel this is not an insurmountable problem as there are methods that can reduce or eliminate microsatellite stuttering (e.g., using a polymerase that is highly processive). One could also repeat MDA on a subset of samples in order to determine if lower shorter alleles result in genotyping errors. Interestingly, one of the eight Mendelian discrepancies resulted from a genuine mutation in the miracidia genotyped. In this case both chelex extracted and WGA DNA generated genotypes with a 2bp shift relative to the allele from the LE parent (Fig. 2c). This gives an estimated mutation rate of 3.7 × 10−4 [95% confidence intervals: 1.0 × 10−4 – 2.2 × 10−3] per generation for autosomal loci [1/(48 autosomal loci × 28 miracidia × 2 chromosomes)], similar to that observed in humans and maize, but higher than in Drosophila . This mutation was not included in the overall error rate, because the Mendelian discrepancy observed was not due to WGA.
Only 2 markers (sc290, and sc19; p < 0.01) showed a deviation from expected Mendelian ratios. This was tested by comparing observed and expected numbers of genotypes using chi-squared tests. Both markers were fully informative and all offspring had 2 alleles. Thus, these deviations are not the result of MDA. Interestingly, sc290 and sc19 are only 2.4 cM apart in the linkage map . It is not uncommon to find genomic regions with segregation distortion when crossing diverged populations due to the evolution of coevolved gene complexes or of incompatible regions .
These results demonstrate that MDA can generate large amounts of DNA from single miracidia with high fidelity. In this work we used the high yield GenomePhi kit (GenomePhi HY DNA Amplification Kit) giving an average (± 1 s.d.) yield of 53.5μg ± 3.6 μg. Thus, 2.5 μL of chelex extracted miracidium can provide enough DNA to run up to 1500 microsatellite genotyping reactions (assuming 30 ng of template/reaction). Hence, up to 12000 reactions could be generated from the 20μL chelex extraction. The total amount of DNA available from a single MDA is also sufficient for SNP typing assays (e.g., Illumina Golden Gate) and next-generation sequencing using platforms such as the Illumina Genome analyzer (Solexa). An advantage of generating large amounts of DNA is that genotypes that fail to amplify at first attempt can be retried. In addition, samples can be regenotyped to provide reproducibility estimates in field-based studies. MDA currently costs between $4 and $16 depending on whether low or high yield is required. For most applications low yield kits generate ample DNA (approximately 5 μg). Hence the method is relatively inexpensive.
The Schistosoma mansoni genome sequence (v3.0 of the sequence assembly) contains approximately 16,757 di-nucleotide repeats (≥8 perfect repeats) and 4,976 tri (≥7 perfect repeats) and 1,422 tetranucleotide repeats (≥6 perfect repeats). Some of these have been previously used for population studies of S. mansoni, while 243 markers have recently been used to generate a genetic map for this organism . Furthermore, multiple SNP polymorphisms are present in the genome sequence data and from examination of EST sequences . MDA gives researchers the ability to genotype many hundreds of microsatellite or SNPs from single miracidia. Hence, genome-wide studies of genetic variation are now feasible. We anticipate that genotyping of miracidia after WGA will allow detailed studies of parasite population genetics at a variety of spatial scales, improving our understanding of mating systems and transmission dynamics of this parasite. Genome-wide genotyping of miracidia could also be used to systematically search for genome regions that have been exposed to strong selection, for example by drug treatment, and for genome-wide association studies of schistosome traits of biomedical importance.
Supported by NIH R21 AI072704 (TJCA), SFBR Forum Grant (CDC), NIH Training Grant D43TW006580 (PTL), and NIH schistosome supply grant AI30026. The molecular work at the Southwest Foundation for Biomedical Research was conducted in facilities constructed with support from Research Facilities Improvement Program Grant C06 RR013556 from the National Center for Research Resources, National Institutes of Health. We thank the following: the Wellcome Trust Sanger Institute for providing the genome sequence and repeat masked sequence; Matt Berriman and Najib El-Sayed for genome support; and Guilherme Oliveira for supplying the LE line.