Development of TaqMan MGB real-time PCR assays for quantitative typing of major cattle and human-pathogenic Cryptosporidium species.
Development of TaqMan MGB real-time PCR assays for quantitative typing of major cattle and human-pathogenic Cryptosporidium species.
Three specific TaqMan MGB real-time PCRs, based on the SSU rRNA gene, were directed towards livestock-restricted Cryptosporidium andersoni and Cryptosporidium bovis as well as both human-pathogenic Cryptosporidium parvum and Cryptosporidium hominis. A generic TaqMan assay further identified all known Cryptosporidium species and simultaneously monitored PCR inhibition through an external amplification control. The generic and specific assays were highly reproducible, and all displayed a detection limit of one oocyst per reaction. The specific TaqMan protocols also proved valuable for specifically detecting and quantifying target DNA in the presence of non-target DNA in environmental samples.
All TaqMan MGB real-time PCR assays fulfilled the required specificity and sensitivity criteria, both on laboratory strains and on a surface water matrix.
No molecular-based method was yet available for the quantitative detection of C. andersoni and the cluster formed by C. bovis, Cryptosporidium ryanae and Cryptosporidium xiaoi. This work provides a novel tool to evaluate the parasite load from domestic ruminants and humans, and to improve assessment and management of microbial risk through better appraisal of the origin and fate of faecal pollutions.
The protozoan Cryptosporidium is a major etiological agent of waterborne outbreaks that displays worldwide distribution (Baldursson and Karanis 2011). More than 20 species and 60 genotypes of Cryptosporidium have been described, the closely related Cryptosporidium parvum and Cryptosporidium hominis being the major pathogenic species for humans (Fayer 2010; Fayer et al. 2010; Robinson et al. 2010). Cryptosporidium is widespread in animals, notably in cattle, and generates major production losses in the livestock industry (De Graaf et al. 1999; Fayer et al. 2007; Silverlas et al. 2010). Although most Cryptosporidium species show limited cross-species transmission, some infect multiple hosts (Xiao and Fayer 2008). For instance, C. parvum infects a wide range of mammals and can affect humans through zoonotic transmission, thereby underpinning a relevant public health issue (Chalmers and Giles 2010). Cryptosporidium oocysts contained in human and animal faeces are released in the environment and potentially reach water resources that are used for human activities. The low infectious dose and the high resistance to conventional water treatment processes increase the contamination risk for humans (Betancourt and Rose 2004).
Over two decades, monitoring of Cryptosporidium in surface and drinking water has relied on microscopic methods that do not allow identification of species composition in a sample. Molecular methods now provide the tools to explore the genetic diversity of Cryptosporidium found in the environment. A majority of data on Cryptosporidium species occurrence in environmental samples has been gathered with PCR-restriction fragment length polymorphism (RFLP) analysis as well as sequencing of PCR amplification products, but these methods are time-consuming (Xiao et al. 1999; Jellison et al. 2002; Ruecker et al. 2007). Real-time PCR has evolved as a rapid and sensitive tool for pathogen detection and quantification in both clinical and environmental microbiology (Espy et al. 2006; Gilbride et al. 2006; Girones et al. 2010). For Cryptosporidium, detection by real-time PCR has been relying on several target genes (Fontaine and Guillot 2002; Guy et al. 2003; Di Giovanni and LeChevallier 2005; Jothikumar et al. 2008; Hadfield et al. 2011; Homem et al. 2011). Among these, the best described one encodes the small subunit ribosomal RNA (SSU rRNA) and has the advantage to display 20 copies of the target locus per oocyst (Abrahamsen et al. 2004), thus increasing the probability of detecting it from small amounts of Cryptosporidium oocysts that are common in environmental samples (LeChevallier et al. 2003; Smith et al. 2006).
Most real-time PCR assays have been directed towards Cryptosporidium species of public health importance and have targeted single, or a group of Cryptosporidium species. Currently, no probe-based assay is available for specific identification of Cryptosporidium species other than the most common human-pathogenic species C. parvum, C. hominis and Cryptosporidium meleagridis (Higgins et al. 2001; Stroup et al. 2006; Hadfield et al. 2011). More recently, Hadfield and Chalmers (2012) developed a real-time PCR for specific detection of Cryptosporidium cuniculus, an emerging human-pathogenic species. There is a growing need, though, to develop such systems for environmental purposes as reliable detection and identification of Cryptosporidium species in the environment are crucial for the establishment of an accurate risk assessment and elucidation of contamination sources and transmission routes. Because Cryptosporidium species that infect cattle represent a major input to water, we have developed a two-step TaqMan real-time PCR toolbox based on the SSU rRNA gene. The developed toolbox quantifies total Cryptosporidium loads in a generic assay, before differentiating livestock-restricted Cryptosporidium andersoni and Cryptosporidium bovis from both major human-pathogenic species by using a common primer pair but discriminatory TaqMan MGB probes (Kutyavin et al. 2000). Potential PCR inhibition is further monitored by an external amplification control.
Purified preparations of viable C. parvum oocysts (Iowa Harley Moon isolate) were obtained from Waterborne Inc (New Orleans, LA, USA). Suspensions of C. andersoni oocysts were provided by Dr N. Neumann (Provincial Public Health Laboratory, Calgary, AB, Canada), and DNA samples from C. andersoni, C. bovis and Cryptosporidium ryanae were obtained from Dr R. Fayer (Agricultural Research Service, US Department of Agriculture, Beltsville, MD, USA). DNA from C. hominis, Cryptosporidium canis, Cryptosporidium felis, Cryptosporidium ubiquitum, C. meleagridis and C. cuniculus was provided by the UK Cryptosporidium Reference Unit (Swansea, Wales, UK), whereas Cryptosporidium muris DNA was obtained from the Wisconsin State Laboratory of Hygiene (Madison, USA). All DNA samples were stored at −20°C until use in real-time PCR assays.
Stock suspensions of C. parvum and C. andersoni oocysts were enumerated by immunofluorescence microscopy using fluorescein isothiocyanate-labelled monoclonal antibodies (Waterborne Inc.). For C. parvum oocysts, serial 10-fold dilutions were prepared for further DNA extraction. After enumeration of the C. andersoni oocysts, the suspension was used as a whole for DNA extraction followed by log-dilutions of the extract. As stock concentrations of provided DNA samples (C. andersoni, C. bovis, C. ryanae, C. hominis, C. canis, C. felis, C. ubiquitum, C. cuniculus and C. meleagridis) were unknown, templates of those species were amplified with the generic real-time PCR assay developed in this study. The generated Ct value was used to quantify stock concentrations of the respective DNA solutions, based on the assumption that the amplification performance was identical for DNA from C. parvum and other Cryptosporidium species. For sensitivity assays, serial 5- and 10-fold dilutions were performed with C. bovis and C. andersoni stock DNA, respectively.
DNA from C. parvum and C. andersoni oocysts was released using the QIAamp DNA minikit (Qiagen, Venlo, the Netherlands) based on the protocol described by Helmi et al. (2011). Briefly, 180 μl of ATL buffer was added to the 200 μl-suspension, followed by five freeze/thaw cycles (freezing in liquid nitrogen during 30 s and thawing at 95°C during 3 min), overnight proteinase K digestion at 56°C and incubation in lysis buffer at 70°C during 10 min. DNA was purified on silica-gel columns according to the instructions of the manufacturer and eluted in DNase-free water before storage at −20°C.
The gene coding for the SSU rRNA was chosen for the design of a generic TaqMan assay and three specific TaqMan MGB assays targeting, respectively, (i) C. parvum and C. hominis, (ii) C. andersoni and (iii) C. bovis (Table 1; Fig. 1a). Both the generic and the specific assays were designed using Primer Express® software v2.0 (Applied Biosystems). Designed primers as well as the generic TaqMan probe were synthesized by Eurogentec (Liége, Belgium). TaqMan MGB probes were synthesized by Life Technologies (Halle, Belgium).
|Target||Type||Name||Sequence (5′–3′)||Position (nt)a|
|Cryptosporidium parvum/Cryptosporidium hominis||Probe||Cp/Ch18S||FAM/GTTAATAATTTATATAAAATATTTTGATG/NFQ-MGB||621–649a|
Available complete and partial sequences of the SSU rRNA gene for 21 Cryptosporidium species were retrieved from GenBank (National Centre for Biotechnology, http://www.ncbi.nlm.nih.gov/genbank/). For each species, an alignment of the collected sequences was performed using ClustalW2 (European Bioinformatics Institute, http://www.ebi.ac.uk/Tools/msa/clustalw2/) and a representative sequence was chosen. For most species, type sequences (Plutzer and Karanis 2009) were chosen as representative sequence, except for those that were too short to include the region targeted by the generic assay. In total, sequences from 21 Cryptosporidium species were used: C. parvum (AF161856), C. hominis (AF108865), C. bovis (EF514234), Cryptosporidium xiaoi (GQ337962), C. ryanae (HQ179574), Cryptosporidium baileyi (L19068), C. meleagridis (AF112574), C. cuniculus (FJ262726), Cryptosporidium wrairi (AF115378), Cryptosporidium suis (AF115377), C. canis (AF112576), C. felis (AF108862), Cryptosporidium galli (HM116388), C. muris (AB089284), Cryptosporidium serpentis (AF151376), C. andersoni (AF093496), C. ubiquitum (JN247403), Cryptosporidium macropodum. (AF513227), Cryptosporidium fayeri (AF108860), Cryptosporidium fragile (EU162752) and Cryptosporidium varanii (AF112573). It should be noted that for C. parvum, the type sequence (accession no. AF112574) refers to the recently described Cryptosporidium tyzzeri (Ren et al. 2012). As we target zoonotic C. parvum (former C. parvum bovine genotype), we have therefore chosen the corresponding GenBank sequence (accession no. AF161856). On the basis of the alignment of the 21 sequences (Fig. 1b), the generic TaqMan assay was designed in a conserved region of the SSU rRNA gene. The size of the generated amplicon was ~107 bp (depending on the sequence).
Based on the same multiple sequence alignment of SSU rRNA gene sequences (Fig. 1b), the hypervariable region of the gene (~100 bp) was selected for the design of the specific assays (Fig. 1a). Universal primers (ChvF18S and ChvR18S) that are genus-specific for Cryptosporidium were designed in conserved regions flanking the polymorphic one (Fig. 1b). The size of the generated amplicon was ~170 bp (depending on the sequence). Within the latter, three distinct TaqMan MGB probes were designed. The first probe (Cp/Ch18S) was based on GenBank accession number AF161856 for amplification of both C. parvum and C. hominis. The second probe (Ca18S) was based on GenBank accession number AF093496 for specific amplification of C. andersoni. The third probe (Cb18S) was based on GenBank accession number EF514234 for specific detection of C. bovis. Because of 100% sequence homology, the Cb18S assay is expected to amplify DNA not only from C. bovis but also from C. xiaoi (Fig. 1b).
Real-time PCR was performed with the Applied Biosystems 7500 FAST instrument. Each 25-μl-reaction mixture contained 5-μl-template DNA, 12·5-μl-qPCR MasterMix Plus Low ROX (2× reaction buffer with 5 mmol l−1 MgCl2 final concentration) (Eurogentec). For the generic assay, primers (CcF18S and CcR18S) and TaqMan probe (Csp18S) were used at final concentrations of 300 nmol l−1 and 100 nmol l−1, respectively. For the specific assays, final primer concentrations were 500, 300 and 500 nmol l−1 for the Cp/Ch18S, Ca18S and Cb18S assays, respectively. The three specific probes were all used at a final concentration of 100 nmol l−1.
Real-time thermocycling conditions were as follows: HotGoldStar DNA polymerase activation and initial denaturation during 10 min at 95°C followed by 50 cycles including denaturation at 95°C during 15 s and annealing/extension at 60°C for 1 min. For the Cp/Ch18S assay, the temperature of the annealing/extension step was 58°C. Fluorescence data were collected during the annealing/extension step and analysed using sds v2.0.5 software (Applied Biosystems).
To assess PCR inhibition and prevent from false-negative results due to inhibition, an external amplification control (EAC) was incorporated into the generic assay (D'Agostino et al. 2011). The EAC consisted in an optimized TaqMan assay including EAC-primers and a Yakima Yellow-BHQ-1 TaqMan probe (Universal Exogenous qPCR Positive Control for TaqMan Assays, Eurogentec) contained in the IPC (Internal Positive Control) mix. The IPC DNA template was also provided in the kit. The resulting duplex EAC-generic assay was used to identify inhibition by comparing the Ct value between control wells (containing DNase-free water as template) and wells containing the sample template. The concentration of IPC mix and IPC DNA to be added to the reaction mixture was adapted in a way to generate a standard curve with a slope and limit of detection (LOD) as close as possible to that of the simplex generic assay. Following the instructions of the manufacturer, the concentration of the EAC was adapted for low amounts (Ct > 30) of target DNA. Slope values and LOD were compared between both simplex and duplex assays.
To assess the analytical sensitivity of the generic real-time PCR assay, DNA extracts from log-diluted C. parvum oocysts were used with a dynamic range from 104 to 10−1 oocysts reaction−1. For the Cp/Ch18S assay, the dynamic range of log-diluted DNA was between 103 and 10−1 eq.oocysts reaction−1. Similarly, for both Ca18S and Cb18S assays, standard curves were constructed with DNA concentrations ranging from 103 to 10−1 eq.oocysts reaction−1 and from 102 to 10−1 eq.oocysts reaction−1, respectively. For each assay, the cycle threshold values (Ct) of dilution points were plotted as a function of the logarithm of the oocyst input quantity. The slope of the standard curves was used for determining the PCR efficiency (E), in conformity with E = 10(−1/s)−1. The LOD of the real-time PCR assays was defined as the last dilution point being detected with a probability of 95% (Bustin et al. 2009).
A first assessment of assay specificity was performed in silico by subjecting selected primers and probes to a Blast test (Basic Local Alignment Search Tool, http://blast.ncbi.nlm.nih.gov/), to check for cross-reactions with non-target sequences. Further evaluation of assay specificity was performed with DNA templates from 11 Cryptosporidium species (C. parvum, C. hominis, C. andersoni, C. muris, C. bovis, C. ryanae, C. canis, C. felis, C. ubiquitum, C. meleagridis and C. cuniculus). Concentrations of DNA stock solutions were estimated using the generic TaqMan assay. Specificity was evaluated for both generic and specific assays using 5 μl of DNA stock solutions. In a last step, competition was studied for the three specific TaqMan MGB assays. For this purpose, a constant concentration of DNA from each target species (C. parvum, C. andersoni or C. bovis) was amplified either in absence or in presence of various DNA concentrations from a non-target Cryptosporidium species. Non-target C. parvum DNA was mixed with C. andersoni and C. bovis DNA for evaluation of cross-reactions using the Ca18S and Cb18S assays, respectively, while non-target C. andersoni DNA was mixed with C. parvum DNA for assessment of the Cp/Ch18S assay. Concentration of target DNA was set at ~100 eq.oocysts reaction−1 for both C. parvum and C. andersoni and at ~25 eq.oocysts reaction−1 for C. bovis, because of a lower initial DNA stock concentration. Non-target DNA was added either at same concentrations than target DNA (ratio 1 : 1) or at 5- and 10-fold higher concentrations (1 : 5 and 1 : 10 ratios, respectively). Potential competition was assessed by comparing Ct values of mixed samples to that of reference samples free of non-target DNA (ratio 1 : 0). Significant differences were questioned using one-way anova, followed by Dunn's multiple comparison post hoc test (SigmaStat 2.03, Systat Software Inc., Chicago, IL, USA).
Four surface water samples (40 l) were simultaneously collected at the pre dam of the Upper-Sûre reservoir in Luxembourg. These samples contained naturally occurring C. andersoni oocysts at a concentration of 2·5 eq.oocysts l−1 (as determined with the Ca18S assay). Three of them were seeded with different amounts of C. parvum. Seed doses of 50, 200 and 1000 C. parvum oocysts were tested, resulting in final concentrations of about 1, 5 and 25 C. parvum oocysts l−1, respectively. An unspiked sample was analysed as control sample. After thorough mixing by mechanical agitation (IKA Eurostar equipped with R1382 Propeller stirrer) at 300 rev min−1 during 10 min, seeded samples were concentrated and purified following ISO 15553 standard protocol (ISO 2006) including the following modifications (Helmi et al. 2011). First, Envirochek high volume filter capsules (Pall Corp., Port Washington, NY, USA) were eluted with standard Laureth-12 elution buffer without a pre-elution step with polyphosphate buffer. Secondly, after concentration of the eluate by centrifugation and purification by immunomagnetic separation (IMS), the magnetic bead–oocyst complex was dissociated through the addition of 100 μl instead of 50 μl of HCl 0·1 mol l−1. The final suspension containing the purified oocysts was then centrifuged at 10 000 g during 3 min and the supernatant discarded. The pellet was resuspended in 40 μl of distilled water, vortexed during 30 s and split into two 20-μl-analytical replicates for DNA extraction according to the initial protocol. The DNA extract was further purified by the drop dialysis technique. For this purpose, 15 ml DNase-free water was poured in a Petri plate and a nitrocellulose membrane (diameter = 25 mm, mean pore size = 0·025 μm, VSWP type, Millipore Inc., Molsheim, France) was deposited on the water solution. The DNA extract was then transferred on the membrane and recovered after an incubation of 30 min at room temperature. During each extraction, a control sample (DNase-free water) was included to identify any potential DNA contamination.
Standard curves were constructed with serial dilutions of oocysts or DNA stock and they are illustrated in Fig. 2. For the generic assay, a linear relationship was observed over the dynamic range (104 to 10−1 oocysts reaction−1) using C. parvum oocysts. For the specific assays, a linear relationship between Ct value and input DNA quantity was observed from 103 to 100 eq.oocysts reaction−1 for the Cp/Ch18S- and Ca18S-specific assays and from 102 to 100 equivalent (eq.)oocysts reaction−1 for the Cb18S assay. As indicated in Fig. 2, slope values ranged between −3·302 and −3·760 and square correlation coefficients varied from 0·999 to 0·968. When no Cryptosporidium DNA was added to the PCR well, no amplification was observed.
The lowest dilution point at which target DNA was always amplified was 1 (eq.)oocyst reaction−1 for all assays. Using the generic assay, one of three replicates (33%) was amplified for 0·1 oocyst reaction−1. Knowing that an oocyst can contain up to 20 copies of the target gene, this corresponds to a detection of two gene copies per PCR. According to these results, the LOD of the generic assay lies between 0·1 and one oocyst reaction−1, that is, between two and 20 gene copies reaction−1, respectively. Considering a LOD defined by a 95% detection rate, all of the assays displayed a constant and reproducible LOD of one (eq.)oocyst reaction−1 (Fig. 2).
Reproducibility (intra- and inter-assay variance) was high, as illustrated by the low standard deviations (Fig. 2). Intra- and inter-assay standard deviations were always below one Ct, except for the one oocyst reaction−1 dilution point of the Ca18S assay (standard deviation of 1·3 Ct). Resulting coefficients of variation (CV) did not exceed 4% for all developed assays.
According to the results of in silico BLAST tests, no organism other than Cryptosporidium is recognized by both primers and probe. Concerning the specific assays, BLAST tests indicated that all known Cryptosporidium species and genotypes could be recognized by the universal primer set. For the Cp/Ch18S assay, only C. parvum and C. hominis sequences were homologous to the probe, while for the Ca18S assay, only C. andersoni was reported. Finally, the Blast test of the Cb18S assay returned sequences from both C. bovis and C. xiaoi. The in vitro tests confirmed that the Cp/Ch18S assay only recognizes C. parvum and C. hominis and no other species tested in the present study, while the Ca18S assay only amplified DNA from C. andersoni (Table 2). The Cb18S assay amplified DNA from C. bovis but also from C. ryanae (Ct of 38·6 ± 0·3), despite the presence of a single mismatch in the probe region (Fig. 1b). It was, however, not possible to test the analytical specificity in vitro for C. xiaoi because no DNA was commercially available. Finally, the generic assay consistently amplified DNA templates from the 11 Cryptosporidium species tested in this study (Table 2).
|DNA template||Generic assay||Cp/Ch18S assay||Ca18S assay||Cb18S assay|
As shown in Table 3, no significant competition was observed for the Cp/Ch18S and Ca18S assays (P > 0·05). The Ct values for C. parvum and C. andersoni remained constant over the different ratios and did not differ from the control (ratio 1 : 0) with the exception of the 1 : 10 ratio. Here, the higher standard deviation resulted from the later amplification of one of the three replicates. For the Cb18S assay, a significant difference between the control and both the 1 : 5 and 1 : 10 ratios was observed (P < 0·05), resulting in a Ct increase of about 1·5 unit.
|TaqMan MGB assay||Non-target DNA||Ratio ‘target DNA: non-target DNA’|
|1 : 0||1 : 1||1 : 5||1 : 10|
|Cp/Ch18S||C. andersoni||32·1 ± 0·3||32·2 ± 0·2||32·1 ± 0·1||33·3 ± 2·3|
|Ca18S||C. parvum||31·4 ± 0·1||31·3 ± 0·2||31·0 ± 0·3||32·5 ± 1·3|
|Cb18S||C. parvum||34·3 ± 0·1||34·1 ± 0·3||35·8 ± 0·1a||35·4 ± 0·3a|
The addition of the external amplification control (EAC) to the generic assay did not impact on assay performance. The LOD of both simplex (generic) and duplex (EAC-generic) PCR assays was one oocyst.reaction−1 and slope values were −3·333 and −3·302, respectively (Fig. 2a). Amplification of IPC DNA generated a Ct value that remained constant over four points of the dynamic range. From 102 to 10−1 oocysts reaction−1, the average Ct value of the EAC was 31·0 ± 0·1. It increased to 37·9 ± 0·6 when amplified in the presence of a target DNA concentration of 103 oocysts reaction−1. Above that concentration, IPC DNA was not amplified.
To test the performance of the designed assays for the detection of Cryptosporidium occurring in environmental samples, surface water was seeded with C. parvum and processed according to the ISO 15553 standard protocol. For each seeded sample, both 20-μl analytical replicates (cf. section ‘Materials and methods’) were used for DNA extraction. DNA extracts were then purified by membrane dialysis prior to the analysis using the designed TaqMan assays. Surface water samples already contained C. andersoni as testified by the Ct values of around 37 consistently generated for spiked and unspiked samples using the Ca18S assay. A screening with the generic assay generated average Ct values of 32·6 ± 0·1, 33·2 ± 0·3 and 30·1 ± 0·2 in samples seeded with one, five and 25 C. parvum oocysts l−1, respectively. The Ct value of the unspiked sample was similar (31·6 ± 0·2) to that observed in the samples seeded with one and five oocysts l−1, but higher than that of the samples seeded with 25 oocysts l−1. When performing the Cp/Ch18S assay, amplification occurred in samples containing five and 25 C. parvum oocysts l−1, in one of two analytical replicates. No amplification was observed in the samples spiked with 0 or one oocyst l−1. No amplification was generated in any sample with the Cb18S assay, indicating either absence of target DNA or levels below LOD.
Sensitive detection and reliable identification of Cryptosporidium species are key steps towards improved risk assessment and elucidation of contamination sources and transmission routes of this pathogen. This can be achieved by molecular diagnostic tools, which should ideally be able to detect all species/genotypes (Xiao and Ryan 2008). A series of real-time PCR protocols have been described for the identification of human-pathogenic C. parvum or C. hominis. However, the occurrence of other species/genotypes has already been reported by PCR-RFLP at several occasions in water samples, C. andersoni often being the dominant species (Ruecker et al. 2007; Yang et al. 2008). These observations call for techniques allowing rapid and sensitive differentiation of those species from human-pathogenic ones. In the present work, we have developed real-time PCR assays using TaqMan probes that target a portion of the SSU rRNA gene for the specific detection of cattle-infecting species, without the need for post-treatment of PCR products. Ruminant-restricted species can be distinguished from those that are at risk for humans, which is relevant when analysing water samples from various origins. Identification of C. andersoni and C. bovis also constitutes a practical tool for veterinary-oriented studies.
The genetic polymorphism of the region delimited by the universal primers was high enough to differentiate between the cluster C. parvum/C. hominis, C. andersoni and C. bovis, but the design of probe-based assays within the A-T-rich hypervariable region of the SSU rRNA gene required the use of TaqMan MGB probes (Kutyavin et al. 2000). The Ca18S assay did only amplify DNA from C. andersoni, and not from C. muris, although both are closely related species (Lindsay et al. 2000). The three mismatches within the C. muris sequence might account for such discrimination. It should be noted that the BLAST of the Ca18S probe did report two sequences falsely annotated as C. muris, AF093496 and L19069. The first one is annotated as C. muris calf genotype and corresponds to the type sequence of C. andersoni. In the GenBank database, some entries corresponding to the former C. muris calf genotype have indeed not yet been updated as C. andersoni and, as recently highlighted by Stensvold et al. (2011), the nomenclature in GenBank does not always follow the changing taxonomy. Hence, this should be taken into consideration when designing new assays.
The Cb18S probe can recognize the cluster formed by C. bovis, C. ryanae and C. xiaoi (former C. bovis-like genotype), as shown on the multiple sequence alignment (Fig. 1b). Those three species have indeed a close phylogenetic relationship and are known to infect ruminants, especially cattle (C. bovis and C. ryanae) and sheep (C. xiaoi) (Fayer and Santin 2009). The Cb18S PCR amplified DNA from C. ryanae, despite the presence of a single mismatch; other authors also reported amplification with MGB probes for DNA sequences containing one or several mismatches (Yao et al. 2006; Hadfield et al. 2011).
The Cp/Ch assay did not amplify DNA from Cryptosporidium species other than C. parvum and C. hominis (Table 2). We did not aim at discriminating between C. parvum and C. hominis because other real-time PCR assays have been developed for this purpose (Jothikumar et al. 2008; Hadfield et al. 2011). The Cp/Ch18S PCR is therefore intended for the detection of both human-infecting species and reports the presence of species posing a significant risk for public health.
Sensitivity of molecular assays is of key relevance because low amounts of Cryptosporidium DNA are commonly encountered in environmental samples (LeChevallier et al. 2003; Smith et al. 2006). Upon optimization, both Ca18S and Cb18S assays displayed a LOD of one eq.oocyst reaction−1. Also, the Cp/Ch18S PCR could detect one eq.oocyst reaction−1, which is similar to and lower than the LOD reported in other C. parvum-specific real-time PCR protocols (Johnson et al. 1995; Fontaine and Guillot 2002; Limor et al. 2002; Jothikumar et al. 2008; Hadfield et al. 2011; Homem et al. 2011). Besides sensitivity, the resolution of a mixture of species constitutes another challenging issue that PCR-based techniques have to deal with (Reed et al. 2002; Stensvold et al. 2011). In particular, environmental matrices are known to commonly contain more than one Cryptosporidium species or genotype (Jellison et al. 2002; Yang et al. 2008; Ruecker et al. 2011). We therefore evaluated the performance of the specific assays for amplification of target DNA in the presence of non-target DNA. Although no significant competition was observed for the Cp/Ch18S and Ca18S systems, the presence of 5- and 10-fold higher amounts of non-target DNA did cause a slight increase in Ct values for the Cb18S assay. The lower target DNA concentration (25 oocysts reaction−1) available for C. bovis might account for this observation. More generally, the fact that one out of three replicates was less efficiently amplified at the 1 : 10 ratio for the Cp/Ch18S and Ca18S assays suggests that a 10-fold excess of non-target DNA begins to induce PCR competition. Other authors have come to similar conclusions (Reed et al. 2002; Tanriverdi et al. 2002; Jothikumar et al. 2008; Ruecker et al. 2011). Using repetitive nested PCR-RFLP analysis, Ruecker et al. (2011) observed a reduction or loss of detection of the less abundant DNA template at ratios above 1 : 10. Although mixed samples are a key issue for molecular analysis, especially with environmental samples, available information on preferential amplification of DNA templates is limited and would therefore require more thorough investigation.
In environmental settings, the presence of inhibitory compounds (e.g. humic substances and salts) that remain after sample preparation and DNA extraction can cause PCR inhibition and misinterpretation of the results (Sluter et al. 1997; Jiang et al. 2005). Therefore, the inclusion of an amplification control is strongly recommended (Hoorfar et al. 2004; D'Agostino et al. 2011). To date, only a few real-time PCR protocols have included an amplification control for the detection of Cryptosporidium (Higgins et al. 2001; Ramirez and Sreevatsan 2006; Hadfield et al. 2011). In the present study, the generic assay consistently amplified Cryptosporidium DNA extracted from 1 oocyst reaction−1, which is similar to recently developed generic protocols (Miller et al. 2006; Jothikumar et al. 2008; Hadfield et al. 2011; Li et al. 2011). Importantly, the inclusion of a non-competitive external amplification control did not impact on the LOD of the generic assay. The duplex EAC-generic TaqMan protocol developed in this study is informative on PCR inhibition for a range of target DNA concentrations expected to be encountered in the environment, and for which PCR inhibition would have the most significant consequence on data interpretation.
After validation and verification of the above-discussed sensitivity, specificity, competition and inhibition, the use of both the generic and specific assays was shown to enable identification and quantification of mixed Cryptosporidium DNA in an environmental sample. For surface water samples from the pre-dam of the Upper-Sûre reservoir, Luxembourg, a Ct value of 30·2 ± 0·3 was generated with the EAC duplex assay, indicating that no apparent inhibition was observed. DNA from naturally occurring oocysts was detected with the generic assay and identified as C. andersoni with the Ca18S PCR. This was achieved along with the detection of seeded C. parvum oocysts with the Cp/Ch18S assay, positive results being observed for samples spiked with C. parvum at a concentration of five and 25 oocysts l−1. With the developed TaqMan MGB systems, it is therefore possible to identify C. andersoni oocysts in surface water, even in the presence of 10-fold higher amounts of C. parvum oocysts.
In summary, the TaqMan protocols developed in the present study detect Cryptosporidium at both genus and species level, based on the amplification of a conserved and hypervariable region of the SSU rRNA gene, respectively. The Cb18S assay detects ruminant-infecting species C. bovis, C. xiaoi and C. ryanae and a positive result with this assay would strongly suggest contamination originating from agriculture. The Ca18S PCR can inform even more accurately on the pollution source by directly pointing out bovine contamination with C. andersoni. Using the same primer pair, another TaqMan MGB assay specifically detects zoonotic C. parvum and human-restricted C. hominis thereby adding epidemiological value to the developed molecular toolbox. In future work, the specific systems could be combined into a multiplex PCR and hence increase sample throughput. The generic and specific assays might prove useful for rapid and quantitative screening of animal faeces for veterinary purposes. They can further be used along with detection techniques targeting other pathogens or indicator organisms towards a better appraisal of the origin and fate of faecal pollutions.
The authors want to thank Dr Sébastien Bonot for valuable discussions, Dr R. Fayer as well as Dr M. Santin for their advice, and Dr L. Hoffmann for final reading of the manuscript. The authors also acknowledge Dr N. Neumann and Dr R. Fayer for their kind support through providing Cryptosporidium oocysts and DNA samples. The present work was carried out in the framework of the PATHOS project co-funded by the National Research Fund of Luxembourg (FNR) (contract number C08/SR/08). Jean-Baptiste Burnet received an AFR grant from the FNR (reference TR-PHD BFR08-097).