Background

Zoonoses are often considered as infectious diseases (IDs) acquired by humans via (in)direct contact with animal species that act as carriers of the infective agents. However, there is increasing evidence for the transmission of IDs from humans to livestock and wildlife species [1]. Here, etiological agents of concern include viruses, bacteria and protists. For instance, the diarrhoeal disease caused by Cryptosporidium parvum, that is transmitted from cattle to humans and vice versa, is responsible for economic losses in livestock animals, particularly calves, linked to mortality, morbidity, and subsequent human (re)infections as a consequence of poor hygiene [2]. Wild animals are also at risk from diseases originating in humans, e.g. human Ebola virus [3] or Yersinia pestis [4], or domesticated animals, e.g. canine distemper virus (morbillivirus) [5]. Given the ‘threatened’ status of many wildlife species which are already at risk from anthropogenic activities (i.e. illegal hunting, habitat modification), the increased threat of disease transmission from humans and livestock animals, and subsequent changes to host-parasite dynamics because of smaller habitat ranges imposes unnecessary risks on their continued survival.

The mountain gorilla (Gorilla beringei beringei) is critically endangered [6], and lives in two isolated regions, the Bwindi Impenetrable National Park (BINP) in Uganda and the Virunga Volcanoes Conservation Range, bounded by Uganda, Rwanda and the Democratic Republic of Congo [7, 8]. In 1993, several mountain gorilla groups were habituated to humans to promote wildlife tourism and behavioural research. In addition to the increased human contact as a direct result of these activities, the habituation process has led to gorillas venturing outside protected regions to forage. The areas surrounding BINP are subject to extreme ecological imbalances with 300–500 people/km2 and high numbers of livestock, both with low-quality health services. As a result, gorilla conservation is now also threatened by the increased risk of disease transmission from humans and livestock. Precautions are required to avoid interspecies transmission of ‘novel’ pathogens [9, 10].

The accurate identification of parasites from animals and environmental samples (i.e. soil, water) underpins a holistic approach to disease control. Given the limitations of conventional microscopic methods and host origin to the specific diagnosis of many parasites (i.e. Cryptosporidium, Giardia, Cyclospora and Entamoeba [11,12,13,14]), tools based on PCR have been used to characterise samples. Genetic characterisation has benefitted our understanding of epidemiology, host and geographical ranges, and assessing the risk infected hosts pose as reservoirs for interspecies infection. Despite the availability of these molecular techniques, substantial gaps in our knowledge remain. Here, we carried out a molecular study of protists infecting mountain gorillas, livestock and humans, from sites in and around BINP. We used PCR and targeted amplicon sequencing to detect and characterise parasites. The genotypes defined here were compared with published resources to provide insights into the epidemiology of disease in and around BINP and the potential for interspecies transmission.

Methods

Bwindi impenetrable National Park, Uganda

The Bwindi Impenetrable National Park (1°4′50″S, 29°39′41″E), Uganda, covers 32,092 ha and is located on the eastern edge of the Albertine Rift Valley, sharing a border with the protected Sarambwe forest in the Democratic Republic of Congo. The National Park was created in 1991 to protect the critically endangered mountain gorilla. The Park experiences a wet and mild climate with a mean temperature range of 11–23 °C, no real dry season, and provides diverse habitats ranging from 1160 to 2706 m in altitude. BINP is renowned as a biodiversity hotspot (see http://whc.unesco.org/en/list/682), and is home to ~340 of the critically endangered mountain gorilla. Surrounded by one of the poorest and most densely populated rural areas in Uganda with more than 300–500 people/km2, BINP has little possibility of a buffer zone with the surrounding agricultural landscape.

Sample collection

A total of 203 faecal deposits from Gorilla beringei beringei (mountain gorilla; n = 68), Bos taurus (cattle; n = 45), Capri hircus (goat; n = 57), and Homo sapiens (human; n = 33) were collected from locations in and around BINP during May to June 2015 (see Table 1). All faecal samples were transported to the Conservation through Public Health Laboratory (BINP, Uganda) immediately after collection and fixed with 96% ethanol in a 2:1 ratio of ethanol to faeces. All samples were sent to the Royal Veterinary College for molecular analysis.

Table 1 The total numbers of each host species sampled at each site, together with the seven Gorilla beringei beringei Groups sampled from three separate regions of BINP, and geographical coordinates
Full size table

From seven habituated gorilla groups from three different sectors of the park, samples were collected from night-nests each morning and were less than eight hours old. Duplicated samples were avoided by sampling each group only once and taking just one sample per nest. Livestock faecal deposits were collected on privately owned farms bordering BINP. Samples were collected either directly from the rectum or the ground; care was taken to collect only those parts of the faeces not in contact with the environment. Humans that inhabited villages surrounding BINP and had frequent interactions with free-ranging gorillas provided faecal samples.

Isolation of genomic DNA

Total genomic DNA was isolated from each faecal sample using a QIAamp Fast DNA Stool Mini Kit (Qiagen, Hilden, Germany). In brief, a 500–800 μl aliquot of each sample was centrifuged at 10,000×g/1 m, the supernatant removed, 1 ml of phosphate buffered saline added to the pellet (0.2–0.3 g) and the samples mixed by a vortex mixer for 10 s. Following further centrifugation at 10,000×g/1 m, the supernatant was removed, glass beads (0.4–0.6 mm diameter) (Sigma-Aldrich, St Louis, USA) to the equivalent of 0.5 volume of the faecal pellet added, and the sample homogenised using a BeadBeater (30,000× oscillations/min for 30 s). Total genomic DNA was then extracted as per the manufacturer’s instructions and stored at -20 °C.

PCR amplification, gel electrophoresis, sequencing, and phylogenetic analysis

Nested PCR was used to amplify total genomic DNA. For the specific identification of Cyclospora, Cryptosporidium and Entamoeba, a portion of the small subunit of the ribosomal DNA (SSU) was amplified (~500 bp, ~240 bp, and 382–429 bp, respectively). The genotypic and subgenotypic classification of Cryptosporidium was achieved using part of the 60 kDa glycoprotein gene (gp60; 250–350 bp). For the genetic characterisation of Giardia (to the level of assemblage), portions of the triosephosphate isomerase (tpi; ~ 530 bp), glutamate dehydrogenase (gdh; ~530 bp), beta-giardin (bg; ~511 bp), and SSU (~292 bp) genes were amplified. PCR was carried out in a volume of 50 μl containing ~200 ng of DNA template, 50 pmol of each primer, 25 μl of 2× MyTaq™ Mix (Bioline, Taunton, USA), and made up to 50 μl with DNase/RNase free H2O (ThermoFisher Scientific, Hemel Hempstead, UK). Table 2 lists the primers and cycling protocol used to amplify each gene.

Table 2 PCR primers and cycling protocols to amplify target sequences from Cryptosporidium, Cyclospora, Giardia and Entamoeba
Full size table

Visualisation of PCR amplicons was achieved on 1.5% w/v agarose in 1× TBE (Tris, boric acid, ethylenediaminetetraacetic acid [EDTA] buffer) gel stained with SafeView Nucleic Acid Stain (Novel Biological Solutions, Huntingdon, UK). In brief, 5 μl of each amplicon was mixed with 1 μl of 6× DNA Loading Dye (ThermoFisher Scientific) and then subjected to electrophoresis at 50 V/45 min using TBE buffer (0.89 M Tris base, 0.89 M boric acid, 0.5 M EDTA; Sigma-Aldrich). A GeneRuler Low Range DNA Ladder (ThermoFisher Scientific) was included on each gel for size comparison purposes. All PCR amplicons were purified using a QIAquick® PCR Purification Kit (Qiagen), per the manufacturer’s instructions. Purified amplicons were subjected to cycle sequencing reactions using ABI Ready Reaction Mix (BigDye® Terminator v3.1 chemistry; Applied Biosystems, Foster City, USA) and the same primers employed for PCR (separately), followed by direct automated sequencing at GATC Biotech, Cologne, Germany. Comparison with corresponding electropherograms verified sequence quality and consensus sequences were constructed using the software package CLC Main Workbench v.6.9.1 (CLC bio, Aarhus, Denmark).

Basic Local Alignment Search Tool analyses (BLAST®: http://blast.ncbi.nlm.nih.gov/Blast.cgi) determined the sequence similarity of genetic data determined herein (GenBank Accession nos. KY658103–KY658190; Additional file 1: Table S1). Phylogenetic analysis was used to visualise relationships among Entamoeba sequence types defined here and those of 17 recognised species and 11 published ribosomal lineages, because of criteria defined by Jacob et al. [13]. Sequences were aligned using the software MUSCLE version 3.7 [15, 16] with ClustalW sequence weighting and UPGMA clustering for iterations 1 and 2. The resultant alignment was adjusted manually using the software BioEdit [17]. Phylogenetic analysis was conducted by Bayesian inference (BI) using Monte Carlo Markov Chain analysis in MrBayes 3.1.2 [18, 19]. The likelihood parameters set for BI analysis were based on the Akaike Information Criteria corrected for small sample sizes (AICc) [20] in jModelTest2 [21]. For the SSU data, we employed the general time-reversible model of evolution with a gamma distribution (GTR + Γ). Posterior probabilities (pp) were calculated via 1000,000 generations, utilising four simultaneous tree-building chains, with every 100th tree saved. At this point, the standard deviation of split frequencies was < 0.01, and the potential scale reduction factor approached 1. A 50% majority rule consensus tree was constructed based on the final 75% of trees generated by BI.

Results

Cryptosporidium, Giardia and Entamoeba were detected in individual faecal samples from mountain gorillas, humans, and livestock from in and around BINP, while Cyclospora was not detected in any of the samples analysed (see Table 3). PCR detected three concurrent infections: one cattle from Kanyamisinga had Giardia (KY658189) and Entamoeba (KY658126); one goat from Mukono Church of Uganda had Cryptosporidium (KY658104) and Entamoeba (KY658147); and, a gorilla from South, Group Nkuringo had Cryptosporidium (KY658103) and Entamoeba (KY658168).

Table 3 The numbers of hosts test-positive for species of Cryptosporidium, Giardia, Entamoeba and Cyclospora, and the number of mixed infections
Full size table

Appraisal of sequence data, parasite identity and prevalence of infection

Cryptosporidium

We conducted sequence analyses of all gp60 and SSU amplicons (n = 20 and 1, respectively) following PCR. These analyses identified amplicons from two of 20 (10%) faecal DNA samples to represent species and genotypes of Cryptosporidium. One sample was characterised at gp60 (KY658103) and the other at SSU (KY658104); no sample was successfully characterised at both genes. The remaining 18 amplicons were identified as false positives (various bacteria) and are not considered further.

Comparison of the unique gp60 and SSU sequence types determined herein with information available in the GenBank database inferred Cryptosporidium parvum from 1.5% of 68 gorillas (based on gp60) and 1.7% of 57 goats (based on SSU) tested. For the single individual faecal sample test positive in PCR for gp60, we characterised this isolate as genotype IId, subgenotype IIdA23G2, using the system of nomenclature proposed previously [22, 23].

Giardia

Sequencing of all tpi (n = 47), gdh (n = 24), bg (n = 3), and SSU (n = 5) test-positive PCR amplicons from 47 samples (nine from gorillas, 15 from humans and 23 from livestock) identified five samples to contain Giardia isolates representing a single genetic assemblage (A, B or E) of G. duodenalis. No sample represented mixed assemblage populations based on direct sequence comparisons. The remaining 42 samples were identified as false positive ‘bacteria’ or failed to sequence. For the five mono-assemblage samples, our analyses defined two distinct genotypes for tpi (represented by KY658189 and KY658190) and SSU (KY658186 and KY658187, and KY658188), and three genotypes for gdh and bg (KY658183–KY658185 and KY658180–KY658182, respectively). Comparison of these ten sequence types with information available in GenBank inferred G. duodenalis assemblage A in one of three (33%) individual faecal samples from humans from Buhoma and one of 15 (6.6%) cattle from Kanyamisinga; G. duodenalis assemblage B was inferred from two of 27 (7.4%) individual faecal samples from humans from Bujengwe Parish; and, G. duodenalis assemblage E was inferred from one of 12 (8.3%) individual faecal samples from cattle from Mukono, Church of Uganda. Giardia was not detected in samples from mountain gorillas or goats by PCR.

Entamoeba

Sequence analyses of all SSU amplicons (n = 111) identified the same number of faecal DNA samples to represent species/genotypes of Entamoeba. Comparison of the 74 unique SSU sequence types determined herein with information available in the GenBank database inferred the sequences with GenBank accession numbers KY658141 and KY658143 were identical to FN666250 and FN666252, respectively, for Entamoeba bovis, from 7.0% of 57 goats tested.

A further 72 new sequence types (GenBank accession nos. KY658105–KY658140, KY658142, KY658144–KY658179) were recorded during the present investigation. Two sequence types (KY658179 and KY658178), which differed by one and three nucleotides (1% over 630 bp) from a sequence of E. coli (FR686364), were each recorded in 3.0% of 33 human samples tested. A third sequence type (KY658177), which was 26 nucleotides different (4% over 631 bp) from a second E. coli sequence (AB4444953), was also recorded from 3.0% of humans tested. This same sequence type (KY658157) was detected in 1.5% of 68 gorillas tested, and is the only instance where a sequence type was shared between/among species. In the remaining 37 gorilla samples that tested positive for Entamoeba, four sequence types (KY658172, KY658155–KY658157) from five individuals, which are 20–27 nucleotides different (3–5% over 580 bp) from a sequence of E. coli (AB4444953), were recorded in 7.4% of 68 gorillas tested. Among these four sequence types, there are 2–7 nucleotide differences. The remaining 32 gorilla faecal samples contained Entamoeba samples with 19 different sequence types (KY658154, KY658158–KY658171, KY658173–KY658176) that differed from each other by 1–18 nucleotide differences (0.2–3.3% over 539 bp), and are 6–16 nucleotides different (1–3% over 539 bp) from a sequence of E. hartmanni (AF149907), originally reported from humans. From cattle, 29 sequence types (KY658105–KY658133) were defined from 36 individual faecal samples. These sequences differed by 1–51 nucleotides (0.2–9.4% over 545 bp) from each other, and by 15–27, 1–17, and 4–13 nucleotides (3–5%, 1–3%, and 1–2% over 545 bp) from three sequences of E. bovis (FN666249–FN666251, respectively), originally reported from cattle (FN666249) and sheep (FN666250 and FN666251). Finally, from goats, 18 new sequence types (KY658134–KY658153) were defined from 30 individual faecal samples, which differed by 1–53 nucleotides (0.2–9.7% over 546 bp) from each other. These sequence data are 4–7, 14–34, and 2–6 nucleotides different (1%, 3–6%, and 1% over 546 bp) from three sequences of E. bovis (FN666250–FN666252, respectively), originally reported from sheep (FN666250 and FN666251) and caribou (FN666252).

The novel Entamoeba sequences defined here were aligned with 39 reference sequences obtained from GenBank. These reference sequences represented 17 recognised species and 11 published ribosomal lineages. All sequences were aligned across 611 positions. Phylogenetic analyses consistently grouped all 74 SSU sequence types with publicly available reference sequences representing E. coli, E. bovis and E. hartmanni (see Fig. 1), with strong support (pp = 0.95–1.00), to the exclusion of all sequences representing other recognised Entamoeba species, their subtypes, and ribosomal lineages.

Fig. 1
figure 1

The relationships between species and ribosomal/conditional lineages of Entamoeba inferred from SSU sequence data following analysis by Bayesian inference (BI). Posterior probabilities are indicated at all major nodes. New sequences generated here shown in bold

Full size image

Cyclospora

Sequencing of all amplicons (six and 30 from humans and mountain gorilla samples, respectively) identified that none of the samples tested contained Cyclospora DNA detectable by PCR. Based on sequence comparisons with data available on GenBank, genetic data determined herein indicated that amplicons were the result of the amplification of SSU from passerine Eimeria species (data not shown).

Discussion

The present study genetically classified Cryptosporidium, Giardia, and Entamoeba in individual faecal samples from three potential host groups in and around BINP. Our systematic molecular analysis categorised all samples by comparison with reference data available in the GenBank database. The data provide no clear evidence for multiple inter-species transmission cycles (i.e. protists with the same sequence types shared among mountain gorillas, humans, or livestock). A fourth, Cyclospora, appears absent, or was below the limit of detection, from the 203 samples tested. The only sequence derived from multiple host groups was amplified from DNA extracted from an individual faecal sample from a human (fr. Buhoma) and a gorilla (fr. South, Group Bushaho) (c.10 km apart based on geographical coordinates; see Table 1). Comparison of this sequence type with public data indicated it was from an Entamoeba sample, which was 26 nucleotides or 4% different, over 631 bp, from an E. coli sequence (AB4444953). In contrast to these findings, previous epidemiological studies, on a broad range of pathogens, have frequently inferred cross-species transmission to be likely in the case of Cryptosporidium [24], Giardia [25], Escherichia coli [26] and Encephalitozoon intestinalis [27] in Uganda and Tanzania. However, in most cases, the prevalence of infection was higher than the 1.4%, 2.4%, and 54.7% for Cryptosporidium, Giardia and Entamoeba, respectively, determined here. Differences in prevalence of infection may be associated with the times of year samples were collected or differences in local habitat. Alternatively, these findings may reflect changes in Park management practices on faecal contamination of the region by local inhabitants, researchers, tourists, and Park Wardens, and proximity to agricultural land.

Using our PCR-based approach, we genetically characterised seven samples, which were assigned to Cryptosporidium parvum, and to the genetic assemblages A, B and E of G. duodenalis. Cryptosporidium parvum has an extremely broad host distribution making it the greatest zoonotic risk. It is also a cause of economic losses associated with bovine cryptosporidiosis [28]. Despite detecting C. parvum here, it was not found in humans or cattle, but from a gorilla and a goat. The low prevalence (1.4 and 1.7%, respectively, or 1% in all livestock) detected here may suggest a relatively low risk of transmission to humans and other mammals in/around BINP. Nonetheless, previous studies in this area have detected C. parvum from 11% of 100 gorillas [29], 38% of 50 cattle [30] and 8% of 62 humans [31] tested. In addition, in regions < 200 km from BINP, Salyer et al. [24] found the prevalence of infection in humans, non-human primates (NHP) and livestock as high as 32.4%, 11.1%, and 2.2%, respectively. These discrepancies highlight the need for further investigations of the presence and distribution of Cryptosporidium genotypes. Not only are these investigations essential to determine the potential significance of different host groups as sources, reservoirs, and amplifiers of Cryptosporidium, but also to establish which C. parvum transmission cycles occur naturally (e.g. human-to-human, animal-to-human and vice versa, and animal-to-animal) [32]. Understanding the underlying forces behind host-parasite relationships is important, particularly in areas such as BINP that are surrounded by extreme ecological imbalances (i.e. high human and livestock densities, both with low-quality health services), as a reduction in habitat can lead to changes in host density that result in alterations to host-parasite dynamics.

Of the two Cryptosporidium sequence types described here, only one has been reported previously. The single SSU sequence type (KY658104) appears to be distributed globally, having been recorded previously in animal hosts other than goats globally (e.g. [33,34,35]). While this sequence type has not been reported from humans, it has been characterised from cattle [34] and buffalo (Maurya PS. et al., unpublished), both of which are important livestock species. Given (i) the impoverished economic status of the human population surrounding BINP and the reliance on livestock animals to reduce chronic malnutrition, increase food security, and generate an alternate source of income, (ii) the economic losses that may result from bovine cryptosporidiosis, and (iii) the limited genetic information for parasites from herds in the region, ‘tracking’ Cryptosporidium spatially and temporally in Uganda is necessary for the future prevention and control of disease.

Based on gp60 data, a single sequence was characterised as genotype IId of C. parvum, subgenotype IIdA23G2 (KY658103). This subgenotype and sequence type are, to the best of our knowledge, both new records. The genotype is also a new record for Uganda. The IId genotype has been reported from humans (e.g. [36,37,38,39,40]) and livestock animals (e.g. [41,42,43,44,45,46,47,48,49,50]), globally. Recently, genotype IId has been recorded in NHP from China [51], albeit with a different subgenotype (IIdA15G2R1). Consequently, genotype IId, along with IIa, is considered one of the two zoonotic subtype groups of C. parvum [52]. The presence of this genotype highlights the potential occurrence of zoonotic transmission in the region; however, further studies are needed to confirm transmission patterns of this genotype in the region.

From the five samples that tested positive in PCR for G. duodenalis, we did not find polymorphic nucleotide positions in tpi, gdh, or bg sequence types as has been reported previously (e.g. [53, 54]). Also in contrast to previous multilocus genotyping studies (e.g. [55, 56]), there was no discrepancy in assemblage assignment among the four genes for the five samples examined herein, albeit variable amplification success among genes prohibited phylogenetic analysis of a concatenated dataset. The sample characterised and assigned to assemblage A, sub-assemblage AII, based on sequence data from gdh, bg and SSU (KY658185, KY658181 and KY658188, respectively), and recorded here from a human, grouped with G. duodenalis genotypes reported from humans (e.g. [57,58,59,60,61,62,63,64]), wild and domestic animals (including cattle) (e.g. [65,66,67,68]) and water samples (e.g. [69, 70]) from around the world. Samples from a human assigned here to assemblage B, subassemblage BIV, based on sequence data from gdh and SSU (KY658183 and KY658187, respectively), grouped with G. duodenalis genotypes reported from humans from Australia, Brazil, the Netherlands and USA [71,72,73], and wild and domestic animals (including NHP) [74,75,76,77], globally. The second sample assigned to assemblage B, subassemblage BIV, based on phylogenetic analysis of sequence data from bg and SSU (KY658180 and KY658186, respectively), grouped with G. duodenalis SSU genotypes reported from humans from Australia, the Netherlands and USA [73, 78] and wild and domestic animals (including NHP) from Spain and Colombia [76, 77], globally. The corresponding bg sequence is new. Samples from cattle assigned to assemblage E on the basis of genetic data from tpi, gdh and bg (KY658190, KY658184 and KY658182, respectively), grouped with tpi and gdh genotypes reported previously from cattle in Bangladesh, China and USA [67, 79, 80], and NHP from China [51]. The corresponding bg sequence is new, as is the tpi AII sequence type (KY658189) recorded here from cattle.

Giardia duodenalis responsible for the human disease is most commonly linked to assemblages A and B [81, 82], while samples from livestock are typically linked to E. Of the subassemblages/genotypes described, AII, BIII, and BIV were considered specific to humans (see [83]). Nonetheless, multiple, related genotypes within each of these two assemblages have been detected in a range of animals (see [80, 84,85,86]). Our findings are consistent with this. Although it is not possible to illustrate the direct sharing of protist species among host groups using identical sequences, our finding of the G. duodenalis sub-assemblage AII in both a human and cattle indicates the potential for human-to-livestock transmission (based on published data; i.e. [87]).

The remaining samples positive in PCR were phylogenetically closest to Entamoeba. Our results showed an overall prevalence of 54.7% for Entamoeba sequence types from the 203 faecal samples, and 9.1%, 55.9%, 80.0% and 60.0% in human, mountain gorilla, cattle and goat faecal samples (or 68.6% of livestock faecal samples), respectively. The overall prevalence in mountain gorillas is comparable (55.9% in 68 samples vs 48.5% in 70 samples) to figures previously reported from Rwanda [88], despite the authors using microscopic examination and centrifugal flotation techniques. Although not the preferred method of identification [13], Sleeman et al. [88] also characterised Entamoeba samples to the level of species based on cyst and trophozoite morphology, detecting E. coli, E. hartmanni and E. histolytica in 20%, 27% and 1.4% of samples tested. Here, despite using what is the more sensitive/specific technique for the detection and classification of protists in faecal samples, we are ‘technically’ unable to classify all 74 sequence types to the level of species based on the criteria of Jacob et al. [13], having amplified < 80% of SSU gene (i.e. 539–580 of c.1400 nt). However, based on initial sequence comparisons, and our phylogenetic analysis (see Fig. 1), which includes reference sequences representing 17 recognised species and 11 published ribosomal lineages, we interpret 72 new sequence types with caution and classify them as variants (rather than species/subtypes/ribosomal lineages/conditional lineages) of E. bovis, E. coli or E. hartmanni. The two remaining sequence types were detected from faecal samples from goats (KY658141 and KY658143) from Kanyamisinga and Aidah-Rugira (respectively) and were 100% identical to two sequences for E. bovis (FN666250 and FN666252 from sheep and caribou, respectively) [89] over 546 nucleotides. Again, we use caution in this interpretation. Regardless, just two sequence types were shared among host groups, and (typically) these species are not considered pathogenic.

In addition to detecting the target parasites, many samples positive in PCR and sent for sequencing returned results linking amplicons to bacteria or passerine Eimeria (data not shown). This finding highlights the advantages of using phenetic-based approaches, i.e. RFLP, single-strand conformation polymorphism (SSCP), or restriction endonuclease fingerprinting-SSCP, coupled with sequencing to screen large numbers of samples by detecting point mutations, group samples by profile and only sequence representative amplicons. Mutation scanning can, therefore, be a sensitive and powerful tool for the direct analysis of subtle genetic variation within and among populations of protists isolated from animals and the environment.

Conclusions

The present study has provided a snapshot of the occurrence and genetic make-up of Cryptosporidium, Giardia, Entamoeba and Cyclospora in mammals in BINP. The genetic analyses indicated that 54.6% of the 203 samples contained Cryptosporidium, Giardia or Entamoeba that matched species, genotypes or assemblages with the potential to infect humans, mountain gorillas, and livestock species. In addition, 76 new sequence records were identified. As nothing is known about the zoonotic/zooanthroponotic potential of protist samples with these sequences, future work should focus on wider epidemiological investigations of these genetic types together with continued surveillance of protists in humans, other mammals, the environment, and water in this highly impoverished area.