INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Cryptosporidium parasites infect the microvillus border of the gastrointestinal epithelium of a wide range of vertebrate hosts, including humans. To date, at least 21 Cryptosporidium species have been named (20). However, only six to eight species (Cryptosporidium parvum, C. muris, C. wrairi, C. felis, C. meleagridis, C. baileyi, C. serpentis, and C. nasorum) are considered valid by most researchers (9, 20). C. parvum is the species that infects immunocompetent humans and most mammals. Cross-transmission studies performed with various mammals have suggested the possibility that zoonotic transmission of C. parvum to humans occurs (13, 20, 24). However, person-to-person transmission of this parasite has also been demonstrated (2, 8). In recent years, C. parvum has been reported to cause waterborne disease outbreaks worldwide (25).

Workers have described many PCR-based protocols for detection of Cryptosporidium parasites. The primers of these PCR protocols are based on undefined genomic sequences (14, 18, 29) or specific genes (2, 4, 12, 15). PCR methods have been shown to be more sensitive and specific than traditional microscopic techniques for detecting Cryptosporidium spp. in clinical and environmental samples (16, 19). Recently, workers have compared the performance of some PCR techniques commonly used for detection of Cryptosporidium parasites (7, 22, 23).

In recent years, researchers have also developed PCR-based techniques for differentiating C. parvum of human origin and C. parvum of animal origin. These techniques are based on the polymorphic nature of C. parvum strains that infect humans and most animals at the -tubulin, oocyst wall protein (COWP), dihydrofolate reductase (DHFR), thrombospondin-related adhesive protein 1 (TRAP-C1), thrombospondin-related adhesive protein 2 (TRAP-C2), internally transcribed spacer 1 (ITS1), polythreonine repeat (Poly-T), small-subunit (SSU) rRNA, and undefined genomic sequences (2, 3, 5, 6, 11, 15, 17, 21, 26-28, 30-32). These genotyping tools, however, were designed to analyze clinical specimens, and the comparative performance of the protocols has not been evaluated previously. In this study, we compared eight genotyping protocols (3, 5, 6, 11, 17, 21, 26-28), two species-differentiating protocols (2, 15), and one species-differentiating and genotyping protocol (31) for detecting and differentiating Cryptosporidium parasites.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Oocysts and DNA extraction. Fecal samples containing genotype 1 and genotype 2 of C. parvum (based on a DNA sequence analysis of the TRAP-C2, -tubulin, and SSU rRNA genes), C. muris, C. serpentis, Eimeria neischulzi, E. papillata, and Giardia duodenalis oocysts were obtained from infected humans and animals and were stored at 4°C in 2.5% potassium dichromate solutions. Oocysts and cysts were purified following sucrose and Percoll flotation (1). Genomic DNA was isolated from the purified oocysts as described previously (27, 30) and was stored at 20°C until it was used. The concentrations of DNA samples were determined by UV absorption at 260 nm.

PCR. Usually, the PCR mixture contained 50 ng of genomic DNA, each deoxynucleoside triphosphate (Perkin-Elmer, Foster City, Calif.) at a concentration of 200 µM, 1× PCR buffer (GeneAmp 10× PCR buffer; 500 mM KCl, 100 mM Tris-Cl [pH 8.3], 15 mM MgCl₂, 0.01% gelatin; Applied Biosystems, Branchburg, N.J.), 5.0 U of Taq polymerase (GIBCO BRL, Frederick, Md.), and the required amounts of forward and reverse primers in a total volume of 100 µl. The Mg²⁺ concentration was adjusted by adding MgCl₂. DNA amplification was carried out with a Perkin-Elmer model PCR 9700 Gene Amp thermocycler for each pair of primers by using the previously described cycling conditions (initial hot start, denaturation, annealing, elongation, final extension, and total number of cycles). A negative control, consisting of a reaction mixture without the DNA template, was included in each experiment. Each PCR product was analyzed by electrophoresis in a 1.5% agarose gel and was visualized after ethidium bromide staining. The same DNA stock preparation was used as the template in all evaluation experiments.

PCR-RFLP analysis. Restriction digestion and PCR-restriction fragment length polymorphism (RFLP) techniques based on Poly-T repeats, COWP, DHFR, TRAP-C1, TRAP-C2, and the SSU rRNA gene were carried out by using previously described protocols (6, 11, 26-28, 31). For the restriction fragment analysis, 20 µl of amplified products was digested by using 10 U of BpuAI (Boehringer Mannheim, Indianapolis, Ind.), BstEI (Boehringer), HaeIII (New England BioLabs, Beverly, Mass.), RsaI (Promega, Madison, Wis.), SspI (New England BioLabs), or VspI (GIBCO BRL) and 5 µl of the appropriate restriction buffer for 1 h under the conditions recommended by the supplier. The digested products were fractionated by agarose gel electrophoresis and were visualized by ethidium bromide staining.

DNA sequencing and data analysis. When desired, the amplified PCR product was purified by using the Wizard PCR Preps DNA purification system (Promega) and was sequenced with a model ABI 377 automated sequencer by using a dRhodomine terminator cycle sequencing kit (Applied Biosystems, Foster City, Calif.). A multiple alignment of the sequences was prepared by using the Genetics Computer Group program (10).

Nucleotide sequence accession numbers. The COWP sequences determined in this study have been deposited in the GenBank database under accession no. AF161577 to AF161580.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Genotyping techniques evaluated in the present study. In the present study we evaluated 16 primer pairs used for 11 previously described Cryptosporidium species differentiation and C. parvum genotyping protocols (Table 1). Of these 11 protocols, 4 are based on the rRNA gene (2, 5, 15, 31), 2 are based on unknown sequences (3, 17), and the remainder are based on COWP, DHFR, TRAP-C1, TRAP-C2, and Poly-T repeats (6, 11, 26-28); 10 of the 11 techniques amplified C. parvum DNA efficiently, and PCR fragments of the expected sizes were produced. The technique of Bonnin et al. (3) amplified DNA from C. parvum genotype 2 isolates. However, we were not able to amplify the DNA of C. parvum genotype 1, C. muris, and C. serpentis isolates under the conditions used in this study. No amplification was observed even after the Mg²⁺ concentration was modified and the amount of template was increased or decreased.

Specificity of Cryptosporidium genotyping tools. The specificity of the primers was examined by using DNA from three Cryptosporidium species (genotype 1 and genotype 2 of C. parvum, C. muris, and C. serpentis), two Eimeria species (E. neischulzi and E. papillata), and one Giardia species (G. duodenalis). The specificity data are summarized in Table 2. Three SSU rRNA gene-based protocols were used, and the primers of two of these protocols, those developed by Awad-El-Kariem et al. (2) and Leng et al. (15), amplified the DNA of Eimeria species in addition to the DNA of Cryptosporidium species, generating bands of similar sizes (Fig. 1). In contrast, the Cryptosporidium-specific primers developed by us (31) produced an ~1,325-bp band during primary PCR and an ~825-bp band during nested PCR from genomic DNA of all Cryptosporidium species but not from genomic DNA of Eimeria species or G. duodenalis.

View larger version (48K): [in this window] [in a new window]   FIG. 1.   Ethidium bromide-stained gel showing amplification products obtained from various members of the Apicomplexa by using the primers of protocols described by Awad-El-Kariem et al. (2) and Leng et al. (15). Lanes 1 and 9, 100-bp marker; lane 2, C. parvum; lane 3, C. muris; lane 4, C. serpentis; lane 5, G. duodenalis; lane 6, E. neischulzi; lane 7, E. papillata; lane 8, negative control. (A) Primers of Awad-El-Kariem et al. (2), which amplified a ~556-bp product. (B) Primers of Leng et al. (15), which amplified a ~1,750-bp product.

Evaluation of PCR-RFLP protocols and nucleotide sequencing. We also evaluated the restriction digestion procedures used in five different PCR-RFLP protocols (6, 11, 26-28, 31). The two species-differentiating protocols of Awad-El-Kariem et al. (2) and Leng et al. (15) were not evaluated further due to nonspecific amplification of DNA from other Apicomplexa species, such as E. neischulzi and E. papillata. An RFLP analysis of DHFR (11), Poly-T (6), TRAP-C1 (27), and TRAP-C2 (28) PCR products showed that the restriction patterns for genotype 1 and genotype 2 of C. parvum were distinct, as reported in the original papers. Restriction digestion of SSU rRNA products could be used to differentiate the Cryptosporidium species, as well as the two genotypes of C. parvum, as reported previously (31).

Sensitivity of the genotyping primers, as determined by using genomic DNA extracted from known numbers of oocysts from C. parvum genotype 2 isolates. Nine PCR protocols were also evaluated to determine their sensitivity for detecting Cryptosporidium parasites by using genomic DNA extracted from known numbers of oocysts of genotype 2 C. parvum isolates. Genotype 2 C. parvum oocysts were used in the sensitivity experiments because of their abundance. Sensitivity tests were not performed for the two protocols (2, 15) that were not specific. Our data on the sensitivity of primers revealed that the nested PCR protocols of Gibbons et al. (11) and Xiao et al. (31) were the most sensitive protocols as the two primer pairs amplified the DNA of even a single oocyst. The results of a comparison of the sensitivities of the genotyping primers are shown in Table 3.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Results of recent molecular characterization studies have shown that there are extensive genetic differences among different Cryptosporidium species, as well as within C. parvum. These inter- and intraspecies differences have been used in the development of molecular diagnostic tools for Cryptosporidium spp. and for genotype differentiation. So far, at least two species-specific (2, 15) and seven genotyping-specific (3, 5, 6, 11, 17, 21, 26-28) protocols have been described. One PCR-RFLP protocol has been developed for both species and genotype differentiation (31). Most of these techniques have not been subjected to evaluation. Recently, several PCR techniques were evaluated to determine whether they detected C. parvum in environmental samples (7, 22, 23), but most of the species and genotype differentiation techniques were not included. The nature of environmental samples suggests that multiple Cryptosporidium spp. and C. parvum genotypes may be present. Therefore, the species differentiation and C. parvum genotyping techniques should be evaluated before they are used for analysis of environmental samples.

Two of the three Cryptosporidium species differentiation techniques based on the SSU rRNA gene were found not to be Cryptosporidium specific. The primers used in these two protocols (2, 15) amplified DNA from two Eimeria species (E. neischulzi and E. papillata). As RFLP was used for species differentiation and the taxon Apicomplexa contains thousands of species, the nonspecificity of the two techniques makes diagnosis of Cryptosporidium species unreliable. The major reason for the nonspecificity is the fact that conserved 18S rRNA sequence data were used when the primers were designed. The primers used by Leng et al. (15) are the universal primers used by most researchers to clone the SSU rRNA gene of eukaryotic organisms. However, another SSU rRNA-based PCR-RFLP protocol, the protocol of Xiao et al. (31), was shown to be Cryptosporidium specific. The primers used in this protocol did not amplify the DNA of non-Cryptosporidium species. This protocol differentiated three Cryptosporidium species and two genotypes of C. parvum.

Six C. parvum genotyping protocols, which were based on DHFR, TRAP-C1, TRAP-C2, ITS1, Poly-T repeats, and an unknown genomic sequence (5, 6, 11, 17, 27, 28), were found to be C. parvum specific. The primers used in these protocols amplified the DNA of genotype 1 and genotype 2 isolates of C. parvum but not the DNA of non-C. parvum species, Eimeria spp., or G. duodenalis. In the techniques of Carraway et al. (5) and Morgan et al. (17) genotype 1- and genotype 2-specific primers are used. The genotype 1-specific primers amplify DNA of C. parvum genotype 1 and not DNA of genotype 2 isolates, whereas the genotype 2-specific primers amplify DNA of C. parvum genotype 2 and not DNA of genotype 1 isolates. Our studies revealed that these four genotype-specific primers are indeed C. parvum genotype 1 or genotype 2 specific, as they amplified and distinguished the two genotypes of C. parvum, as claimed previously. The DHFR-, TRAP-C1-, TRAP-C2-, and Poly-T-based PCR-RFLP techniques (6, 11, 27, 28) were found to be C. parvum specific, and following restriction digestion of single-round PCR and nested PCR products two distinct band patterns were obtained for the genotype 1 and genotype 2 C. parvum isolates. However, a PCR protocol based on an unknown genomic sequence (3) was shown to be less sensitive, as it amplified only C. parvum genotype 2 isolates even after the Mg²⁺ concentration and the amount of DNA template were modified.

We found that the PCR-RFLP protocol based on the COWP gene (26) was not C. parvum specific, as the primers also amplified genomic DNA from other Cryptosporidium spp. However, the primers did not amplify the DNA of non-Cryptosporidium parasites. The COWP gene-based PCR-RFLP protocol has been reported to differentiate C. wrairi (some workers consider this organism C. parvum) from the two genotypes of C. parvum (26). We evaluated this technique by using C. muris and C. serpentis and identified distinct band patterns for these parasites. Furthermore, nucleotide sequencing of the PCR products also revealed similar distinct sequences for different species and genotypes. Therefore, the set of primers may be utilized for species- and genotype-specific diagnosis of Cryptosporidium parasites after the primer sequences are modified.

In summary, the PCR-RFLP protocols of Awad-El-Kariem et al. (2) and Leng et al. (15) for Cryptosporidium species differentiation described previously are not Cryptosporidium specific. We found that the set of primers developed by Bonnin et al. (3) has poor diagnostic sensitivity for genotype 1 C. parvum isolates. The PCR-RFLP method of Spano et al. (26) has the potential to distinguish different Cryptosporidium species (C. muris and C. serpentis), as well as the two genotypes of C. parvum. We found that the genotype 1-specific and genotype 2-specific primers of Carraway et al. (5) and Morgan et al. (17) are C. parvum specific, as these primers detected C. parvum genotype 1 and genotype 2 isolates. Similar results were obtained for the DHFR-, TRAP-C1-, TRAP-C2-, and Poly-T-based PCR-RFLP techniques of Gibbons et al. (11), Spano et al. (27), Sulaiman et al. (28), and Carraway et al. (6). The nested PCR-RFLP protocols developed by Gibbons et al. (11) and Xiao et al. (31) are the most sensitive protocols evaluated in this study. Additional evaluations with other apicomplexan parasites and Cryptosporidium species may be needed before these techniques can be used for analysis of environmental samples.