Background

Human intestinal protozoan infections can lead to significant morbidity and mortality when not diagnosed timely and treated appropriately. Such infections are typically more associated with low-income countries. But with worldwide traveling and migration increasing in importance, there is also a continuous increase in the number of cases in industrialized countries. The diagnosis of intestinal protozoan diseases has remained a challenge, especially in low-endemicity countries where practitioners and health institutions are less familiar with the respective causative agents [1].

Cryptosporidium spp., Entamoeba histolytica, and Giardia intestinalis are the most important parasitic protozoans causing diarrhea [2, 3]. Giardiasis and cryptosporidiosis are major causes of moderate to severe diarrhea in both, developing and developed countries, leading to considerable mortality worldwide [2, 3]. Amoebiasis is the third most frequent cause of death from protozoan parasitic diseases, with high prevalence in developing countries [4, 5].

Blastocystis spp. and Dientamoeba fragilis are very common protozoan commensals whose pathogenic potential is still under evaluation [6]. Human infection rates with Blastocystis spp. can range between 1–60% and can even go up to 100%, depending on the geographic distribution and also the economic status of a given country [10, 11]. D. fragilis is a trichomonad parasite [4] of the human gastrointestinal tract and has, since 1919, been repeatedly reported in symptomatic patients [6]. The incidence of D. fragilis is equal to or even exceeds that of G. intestinalis [7].

As several intestinal protozoans can cause diarrhea, it is important to use diagnostic methods capable of easily and accurately identifying the infectious agents. A multiplex PCR approach can address challenges as presented by microscopy and immunofluorescence assays, such as the dependence on well-trained microscopists, and time or sensitivity issues. Despite the limitations, microscopy is still considered as the reference standard method for routine diagnostics due to its low cost and practicability in endemic countries. The main benefits of a multiplex PCR include cost-effectiveness, time-saving, and higher throughput. Nevertheless, the design and evaluation of a multiplex assay require an investment in terms of time and costs for optimization and validation.

In this study, we established a cost-effective and practical tandem multiplex—high resolution melting analysis for simultaneous detection and quantification of five intestinal protozoa: Cryptosporidium spp., E. histolytica, Giardia intestinalis (assemblages A and B), Blastocystis spp., and D. fragilis.

Results

Assay design and optimization

The individual primer sets used for the multiplex PCR were designed based on sequences obtained from the literature for the generation of long fragments (Table 1). These fragments were submitted to BLAST and used for the design of shorter fragments for the multiplex PCR. Each short fragment was selected according to its specificity, fidelity, and predicted melting temperature. The short fragment primers were designed using the online Primer 3 software (http://bioinfo.ut.ee/primer3-0.4.0/) [8]. Table 1 is presenting details of all primers from the long and the short fragments from the respective target gene for each parasite species.

Table 1 List of primers for amplification of the long original and the short new fragments, with fragment sizes and target genes for each parasite species
Full size table

The 5-plex qPCR is specifically differentiating Blastocystis spp. (Blastocystis), Cryptosporidium spp. (Cryptosporidium), D. fragilis (Dientamoeba), E. histolytica (Entamoeba), G. intestinalis assemblage A (Giardia assemblage A), and G. intestinalis assemblage B (Giardia assemblage B) through the melting temperature of their selected and amplified fragment (Fig. 1). We also used individual primer sets in a single-plex run for the identification of each parasite species.

Fig. 1
figure 1

Melting curves a, melting peaks b, normalized melting peaks c, and differential curves d of reference plasmids of Blastocystis spp., Cryptosporidium spp., D. fragilis, E. histolytica, G. intestinalis assemblage B and G. intestinalis assemblage A. The following melting temperatures were observed: 71.5 °C for D. fragilis (brown), 75.5 °C for E. histolytica (green), 78.5 °C for Cryptosporidium spp. (lila), 80 °C for Blastocystis spp. (blue), 81.5 °C for G. intestinalis assemblage B (light red), and 83.5 °C for G. intestinalis assemblage A (red). The distance between the melting curves is well specified in the difference curve graph d

Full size image

Interestingly the plasmids were presenting the same melting peak in both multiplex and single-plex runs. The melting temperature (Tm) values were 71.50 ± 0.00 °C (Dientamoeba), 75.20 ± 0.25 °C (Entamoeba), 78.23 ± 0.25 °C (Cryptosporidium), 79.84 ± 0.23 °C (Blastocystis), 81.51 ± 0.08 (Giardia assemblage B) and 83.50 ± 0.00 °C (Giardia assemblage A) (Fig. 2).

Fig. 2
figure 2

Boxplots of the melting temperatures (Tm) of Blastocystis spp. (red), Cryptosporidium spp. (brown), D. fragilis (green), E. histolytica (turquoise), Giardia intestinalis assemblage B (pink) and Giardia intestinalis assemblage A (blue). The box indicates the likely range of melting temperature variation. The melting temperature ranges were 71.5–71.5 °C for Dientamoeba, 75–75.5 °C for Entamoeba, 78–78.5 °C for Cryptosporidium, 79.5–80 °C for Blastocystis, 81.5–82 °C for Giardia assemblage B, and 83–83.5 °C for Giardia assemblage A. The melting temperature difference (∆Tm) between the closest melting curves were 3.7 °C (Dientamoeba/Entamoeba), 3.03 °C (Entamoeba/Cryptosporidium), 1.61 °C (Cryptosporidium/Blastocystis), 1.67 °C (Blastocystis/Giardia assemblage B) and 2 °C (Giardia assemblage B/A)

Full size image

Efficiency and limit of detection (LOD) of the assay

The sensitivity of the multiplex was tested by amplifying individually tenfold serial dilutions of plasmids containing copies (≤ 107 copies) of the respective targeted gene fragment each. The multiplex assay detected target gene numbers as low as 100 copies/µl (for Giardia assemblage B) or even ten copies/µl (for Blastocystis, Cryptosporidium, Dientamoeba, Entamoeba, and Giardia assemblage A). The ranges of efficiency, slope, and R2 were 95.77 to 103.11%, − 3.2497 to − 3.4277 and 0.9942 to 0.9998 respectively (Table 2).

Table 2 Multiplex melting temperatures (Tm), specificities and limits of detection (LOD)
Full size table

Further dilutions were prepared (between 200 and 1 copy/µl) to determine the LOD for each parasite. The calculated LODs at 95% confidence for the multiplex method using probit analysis varied from 8.78 (6.06–15.99) copies/µl for Cryptosporidium to 30.08 (18.02–69.13) copies/µl for Entamoeba (Table 2).

The Welch’s unequal variances t test showed that all Tm’s were significantly different from each other as in the predicted melting, with ΔTm (between the closest Tms) value of 3.7 °C (Dientamoeba/Entamoeba), 3.03 °C (Entamoeba/Cryptosporidium), 1.61 °C (Cryptosporidium/Blastocystis), 2.67 °C between (Blastocystis/Giardia assemblage B), and 1.99 °C (Giardia assemblage B/Giardia assemblage A) (Table 2).

Sample genotyping and multiplex specificity

The discriminatory power of the multiplex PCR was tested using 143 DNA isolates from our routine diagnostic laboratory archive. We tested blinded samples using the multiplex method, and the results were compared to previous diagnostic results obtained through microscopy, immunoassays, and/or FRET-based qPCR. The DNA of Entamoeba coli, Entamoeba hartmanni, Iodamoeba buetschlii and 13 other organisms listed in the methods section was included in the tests to confirm the specificity of the method. The negative samples included parasite-free human stool samples and Entamoeba dispar samples. Figure 3 juxtaposes the multiplex PCR results with the original results from the routine diagnostic unit. Up to 95.8% (137) of our results were matching with the routine diagnostics. The 5-Plex method was able to detect four (2.8%) more Cryptosporidium infections, which were confirmed by DNA sequencing and also with Verweij et al. primers (Table 1) [9], to be C. parvum, and six (4.2%) samples were found to be positive for D. fragilis. Details on non-concordant samples with (8.4%) and without (4.2%) Dientamoeba data are presented in Table 3.

Fig. 3
figure 3

Multiplex (red) results of 143 samples as compared to routine diagnostics (blue) (microscopy (all parasites), immunoassay (Giardia and Entamoeba), FRET-qPCR analysis (Blastocystis, Cryptosporidium, E. histolytica, and G. intestinalis). The multiplex results of 137 samples (95.8%) matched with the routine diagnostics. The multiplex method detected four (2.8%) more Cryptosporidium samples and six (4.2%) samples were positive for D. fragilis

Full size image
Table 3 List of non-concordant from the 143 samples with their results in both routine diagnostics and the developed 5-Plex-PCR
Full size table

The selected primer sets were highly specific for the target species, and each parasite species was well differentiated by the individual primer set and also by the 5-plex primers mix. The non-pathogenic amoeba species E. dispar Entamoeba coli, Entamoeba hartmanni, Iodamoeba buetschlii and the 13 organisms listed in the section Materials and methods tested negative with our designed primer sets.

The following cases of double and triple infections were identified within the studied samples: Dientamoeba/Blastocystis (n = 5), Dientamoeba/Giardia (n = 1), Entamoeba/Blastocystis (n = 4), Blastocystis/Giardia (n = 12), and Blastocystis/Entamoeba/Giardia (n = 1) Fig. 4. Each mix infection was confirmed by the relative single-plex runs using the target plasmids as reference.

Fig. 4
figure 4

Melting peaks of mixed infection samples presenting 4 cases of double infection and one case of triple infection. All qPCR runs were done with the 5 primers mix and one color is attributed to each parasite species and the samples (black and grey): Dientamoeba (brown)/Giardia B (pink) a, Entamoeba (green)/Blastocystis (Blue) b, Dientamoeba/Blastocystisc, Entamoeba/Blastocystis/Giardia assemblage A (red) d, and Blastocystis/Giardia assemblage A e. Singleplex runs were done with the target primers to confirm each mixedinfection

Full size image

Discussion

PCR and real-time PCR (qPCR) are nowadays well-established in routine diagnostics for fecal parasites. As compared to microscopy, PCR has the advantage to be fast, more sensitive, easy to perform, and not to require a long period of training, as commonly necessary for a good microscopist. PCR also has the advantage of detecting and quantifying multiple pathogens within one reaction.

Although PCR and qPCR are popular in-house testing methods for parasitological stool diagnostics, they are, in most cases, only used as a complementary method to microscopy and/or immunofluorescent analysis [9, 10].

In this study, we developed a multiplex real-time PCR (5-plex qPCR) followed by high resolution melting (HRM) curve analysis targeting specifically protists associated with diarrhea in humans. The five targeted protozoa were Blastocystis spp., Cryptosporidium spp., Dientamoeba fragilis, Entamoeba histolytica, and Giardia intestinalis.

To our knowledge, this method is the first non-probe-based multiplex targeting up to 5 human diarrheal parasites with well-separated melting temperatures (∆Tm > 1.5 °C). Lalonde and Gajadhar (2011) previously developed PCR melting curve analysis method targeting 8 coccidian oocysts (Cyclosopora cayetanensis, C. parvum, Cryptosporidium muris, Toxoplasma gondii, Eimeria bovis, Eimeria acervulina, Isospora suis, and Sarcocystis cruzi), with eight Tm’s ranging between 81.66 and 87.19 °C. In this case, a sample with multiple parasites would be difficult to screen using the melting peaks. Our 5-Plex, however, could be used for the screening of six different types of mixed infection samples since our melting peaks are well separated (Fig. 4). Presently, our assay is the only 5-Plex assay currently available as other multiplex tests for diarrheal parasites target either two [11], three [1, 9, 12, 13], or maximal four [14, 15] parasites.

Our novel approach harnesses well-differentiated melting temperatures (ΔTm ≥ 1.6 °C, Fig. 1), which allows the simultaneous detection and identification of all five parasites including the assemblages A and B of Giardia. The fact that our multiplex method is detecting more Cryptosporidium species (four) than established standard protocols suggests that our assay may be more sensitive as compared to the standard diagnostic qPCR method for C. parvum. Our method identified the same number of Blastocystis-positive samples as compared to established diagnostic methods used at our institute. However, four Blastocystis samples, one mixed Blastocytis + Giardia sample, and one Giardia sample were also positive for D. fragilis, highlighting the great advantage of our multiplex method in detecting co-infections (Table 3). The ability to simultaneously detect the five relevant parasites in a single reaction could drastically reduce the costs of parasitological stool diagnostics. Indeed, our assay is inexpensive (approximatively 6.65 USD per sample including DNA extraction) as compared to other methods (12 USD for 4 targets, Stark et al. [15]) and can be performed within 1 h 25 min. The method is also quantitative and sensitive, with Cryptosporidium presenting the highest sensitivity (8.78 copy/µl) and E. histolytica presenting the lowest sensitivity (30.08 copy/µl). We used EvaGreen which seems to be more suitable for multiplex PCR applications as compared to other dyes that are either expensive or may inhibit PCR [11, 16,17,18].

The limitation of our 5-plex method is the detection of mixed infections with close melting temperatures (ΔTm) and one dominant parasite infection. In such a case, only the melting peak of the major parasite would be identified.

However, this limitation is compensated by the possibility to use individual primer sets in single-plex runs to confirm and identify suspected mixed infections.

Conclusions

We developed a fast, efficient, cost-effective and easy-to-perform 5-plex qPCR-HMR assay for simultaneous detection of Cryptosporidium spp., E. histolytica, both G. intestinalis assemblages (A and B), Blastocystis spp., and D. fragilis in stool samples. Moreover, the assay is suited for diagnosing mixed infections and could be expanded to widen its diagnostic potential by the inclusion of further intestinal protozoans (such as Entamoeba coli). It is also adaptable for epidemiological surveys and could contribute significantly to the improvement of patient management and infection control. However, the developed method still requires interlaboratory validation before its implementation in routine diagnostics.

Methods

Samples and DNA extraction

The method was evaluated using DNA from 143 stool samples submitted to the Institute of Specific Prophylaxis and Tropical Medicine at the Medical University of Vienna between 2011 and 2018 for diagnostic analysis. The standard diagnostic test of the study samples at our institute included microscopic analysis of stool for all intestinal parasites, immunoassay analysis (Antigens) for G. intestinalis and E. histolytica, and FRET-based qPCR (DNA) for Blastocystis spp., C. parvum, C. hominis, E. histolytica, E. dispar and G. intestinalis. All samples were anonymized. The Qiagen kit for stool DNA extraction (Qiagen GmbH, Vienna, Austria) was used for the samples’ DNA extraction according to manufacturer instructions.

Target selection and primer design

A fragment of the 18S small subunit ribosomal RNA (SSU rRNA) gene (Blastocystis spp., E. histolytica, D. fragilis, Cryptosporidium spp.) and a fragment of the NADP-dependent glutamate dehydrogenase (gdh) gene (G. intestinalis) were selected as targets for the assay due to their frequent use in the literature (Table 1) and were tested with one or two samples of each species (positive samples from the routine diagnostic lab) to confirm their effectiveness. The primers of Cryptosporidium and Blastocystis were selected to cover all species of these parasites as there are many pathogenic species in these genera. As D. fragilis was not included in the routine diagnostics at our institute, the long fragment of one of the positive sample (using Verweij et al. primers [9]) was sequenced by Eurofins MWG Synthesis GmbH (Ebersberg, Germany) to confirm the species before designing the short fragment primers.

New forward and/or reverse primers were designed for each parasite to target a shorter fragment with a melting temperature matching with those of the other parasites for a multiplex qPCR-HRM design (Table 1). All primers were designed using the Primer3 online tool (http://bioinfo.ut.ee/primer3-0.4.0/) and checked for specificity using basic local alignment search. We aligned the sequences with the BioEdit 7.2 Sequence Alignment Editor and produced predictive melting profiles of all fragments using the uMelt BATCH online tool (https://www.dna.utah.edu/umelt/umelt.html) before primers synthesis by Eurofins MWG Synthesis GmbH (Ebersberg, Germany).

PCR and HRM assay

The long fragments were amplified in the same conditions with 0.5 µM primers and the following amplification program: 45 cycles of 95 °C for 15 s, 60 °C for 30 s and 68 °C for 30 s. The PCR products were resolved in 2% agarose gels and visualized by MIDORI Green Advance (Biozym Scientific GmbH, Hess Oldendorf; Germany) staining.

The multiplex PCR was performed in a volume of 20 µl containing primers, 1 × SsoFast EvaGreen Supermix (BioRad, Hercules, CA) and 4 µl DNA (≤ 20 ng/µl). Samples with DNA concentration > 20 ng/µl were diluted 10 times and retested. The optimized primers concentrations were 0.25 µM of forward and reverse primers for D. fragilis and E. histolytica, and 0.15 µM forward and reverse primers for Cryptosporidium spp., Blastocystis spp., and G. intestinalis.

The PCR was performed on a CFX Connect real-time PCR detection system (BioRad Laboratories, Inc., Singapore) with an initial denaturation step at 95 °C for 3 min, followed by 40 cycles of 95 °C for 5 s, 57 °C for 3 s and 68 °C for 5 s. The PCR products were subjected to the following melting program: denaturation at 95 °C for 1 min, cooling to 65 °C (held for 30 s) and continuous heating at 0.5 °C/s with fluorescence acquisition from 65 °C to 90 °C.

HRM plasmids preparation and sequencing

The short fragments of each parasite were modified and inserted into pET17b through EcoRI/XhoI (Fermentas) restriction sites. The ligated plasmids were used to transform competent XL-1 Blue E. coli competent bacteria. Selected positive clones were cultured, the plasmids purified and sequenced by Eurofins MWG Synthesis GmbH (Ebersberg, Germany). The sequencing data were analyzed using Vector NTI.10 (Invitrogen) software, and the sequences were checked by using the Basic Local Alignment Search Tool (Nucleotide BLAST, https://blast.ncbi.nlm.nih.gov/Blast.cgi) to confirm their identity.

Assay specificity and sensitivity

We evaluated the method specificity in situ with negative (for all tested protozoans included in this study) stool samples and mixed infections samples, and also with DNA (≤ 20 ng/µl) from the following organisms (using identical PCR conditions): Toxoplasma gondii, Leishmania infantum, Trypanosoma brucei, Trypanosoma cruzi, Babesia divergens, Enterocytozoon bieneusi, Enzephalitozoon cuniculi, Pneumocystis jirovecii, Echinococcus granulosus, Strongyloides stercoralis, Dirofilaria repens, Toxocara canis, Entamoeba dispar, Entamoeba coli, Entamoeba hartmanni, Iodamoeba buetschlii and Ascaris suum. The assay was performed in duplicate with each DNA sample.

The limit of detection (LOD) for each parasite was determined individually under the conditions of the multiplex PCR. The LOD was defined as the measured concentration producing at least 95% positive replicates [19].

The LOD was assessed by amplifying seven different concentrations (120, 80, 40, 20, 10, 5 and 1 copies/µl for Blastocystis spp., Cryptosporidium spp. and D. fragilis; 160, 120, 80, 40, 20, 10 and 5 copies/µl for E. histolytica and G. intestinalis) of each plasmid, in twenty (20) replicates on two separate occasions.

The total proportion of positive tests were recorded and subjected to probit regression analysis using R version 3.4.2 (2017-09-28) via RStudio version Version 1.1.383. Similarly, the boxplots of the melting temperatures (Tm) were also produced using R via the RStudio version. The Welch’s unequal variances t-test was used to compare the difference between arithmetic means of the respective Tm of the amplicons of all five parasite species using R.