Cryptosporidium proliferans n. sp. (Apicomplexa: Cryptosporidiidae): Molecular and Biological Evidence of Cryptic Species within Gastric Cryptosporidium of Mammals.

The morphological, biological, and molecular characteristics of Cryptosporidium muris strain TS03 are described, and the species name Cryptosporidium proliferans n. sp. is proposed. Cryptosporidium proliferans obtained from a naturally infected East African mole rat (Tachyoryctes splendens) in Kenya was propagated under laboratory conditions in rodents (SCID mice and southern multimammate mice, Mastomys coucha) and used in experiments to examine oocyst morphology and transmission. DNA from the propagated C. proliferans isolate, and C. proliferans DNA isolated from the feces of an African buffalo (Syncerus caffer) in Central African Republic, a donkey (Equus africanus) in Algeria, and a domestic horse (Equus caballus) in the Czech Republic were used for phylogenetic analyses. Oocysts of C. proliferans are morphologically distinguishable from C. parvum and C. muris HZ206, measuring 6.8-8.8 (mean = 7.7 μm) × 4.8-6.2 μm (mean = 5.3) with a length to width ratio of 1.48 (n = 100). Experimental studies using an isolate originated from T. splendens have shown that the course of C. proliferans infection in rodent hosts differs from that of C. muris and C. andersoni. The prepatent period of 18-21 days post infection (DPI) for C. proliferans in southern multimammate mice (Mastomys coucha) was similar to that of C. andersoni and longer than the 6-8 DPI prepatent period for C. muris RN66 and HZ206 in the same host. Histopatologicaly, stomach glands of southern multimammate mice infected with C. proliferans were markedly dilated and filled with necrotic material, mucus, and numerous Cryptosporidium developmental stages. Epithelial cells of infected glands were atrophic, exhibited cuboidal or squamous metaplasia, and significantly proliferated into the lumen of the stomach, forming papillary structures. The epithelial height and stomach weight were six-fold greater than in non-infected controls. Phylogenetic analyses based on small subunit rRNA, Cryptosporidium oocyst wall protein, thrombospondin-related adhesive protein of Cryptosporidium-1, heat shock protein 70, actin, heat shock protein 90 (MS2), MS1, MS3, and M16 gene sequences revealed that C. proliferans is genetically distinct from C. muris and other previously described Cryptosporidium species.

Introduction

Apicomplexan parasites of the genus Cryptosporidium infect the gastrointestinal tract of most vertebrates, including humans [1]. The organ specificity (localization of endogenous development in the host) of Cryptosporidium species and genotypes can vary, and two major groups are recognized: the larger intestinal group, which also includes species and genotypes with affinity for the lungs and bursa of Fabricius, and the smaller gastric group which has affinity for the glands of the glandular stomach [2]. Within the gastric group, two species, C. muris and C. andersoni, are specific for mammals, and a number of different strains of these species have been identified worldwide (Tables 1 and 2) [3, 4]. Cryptosporidium muris was described in laboratory mice [3]. Although it is predominantly a rodent species, it has been detected in, or experimentally transmitted to, various mammalian hosts, including members of Hyracoidea, Carnivora, Lagomorpha, Arctiodactyla, Perissodactyla, and primates (both human and non-human) (Table 1). Cryptosporidium muris also has been detected in the feces of snakes, lizards, frogs, and birds of prey; however, these cases were probably due to the mechanical passage of oocysts following ingestion of infected rodents rather than an active infection [5-12]. Similarly, the detection of C. muris in pig feces and slurry from pig farms [13-15] may have been due to rodents rather than active infections, as pigs are not susceptible to C. muris under experimental conditions [16]. Cryptosporidium andersoni, previously known as C. muris-like, was distinguished from C. muris based on molecular and biological differences [4]. Cattle (Bos taurus) are the typical host for C. andersoni, but it has been found in other ruminants, various rodents, and humans (Table 1).

Table 1

List of identified hosts for Cryptosporidium proliferans (CP), Cryptosporidium muris (CM), Cryptosporidium andersoni (CA), Japan field mouse genotype (JG), and caribou genotype (CG).

Mark indicates susceptibility to infection.

Order	Host (common name)	CP	CM	CA	JG	CG	Reference
Rodentia	Acomys cahirinus (Spiny mouse)	•	•				[24]
	Apodemus speciosus (Large Japanese field mouse)				•		[17]
	Apodemus sylvaticus (Wood mouse)		•				[26]
	Clethrionomys glareolus (Bank vole)		•				[26]
	Cavia porcellus (Guinea pig)		•				[27]
	Dolichotis patagonum (Patagonian mara)		•				[10]
	Eutamias sibiricus (Siberian chipmunk)		•				[28]
	Gerbilus gerbilus (Lesser gerbil)	•		•			[29]
	Meriones tristrami (Tristram's jird)	•		•			[29]
	Meriones unguiculatus (Mongolian gerbil)	•		•			[21, 30]
	Sekeetamys calurus (Bushy-tailed jird)	•		•			[29]
	Marmota bobak (Bobak marmot)			•			[31]
	Mastomys natalensis (Natal multimammate mouse)	•	•	•			[32]
	Mastomys coucha (Southern multimammate mouse)	•	•	•			This study
	Mesocricetus auratus (Golden hamster)		•	•			[33, 34]
	Microtus brandti (Brandt's voles)	•	•				[24]
	Mus spp. (House mouse)	•	•	•	•		[6, 17, 35]
	Phodopus campbelli (Campbell's dwarf hamster)		•	•			[33]
	Phodopus sungorus (Djungarian hamster)		•	•			[33]
	Phodopus roborovskii (Roborovski hamster)		•				[36]
	Rattus norvegicus (Brown rat)		•				[37]
	Sciurus carolinensis (Eastern gray squirrel)	•					[38]
	Tachyoryctes splendens (East African mole-rat)	•					[22]
Peramelemorphia	Macrotis lagotis (Greater bilby)		•				[39]
Hyracoidea	Procavia capensis (Rock hyrax)		•				[40]
Carnivora	Canis familiaris (Dog)		•				[27]
	Canis latrans (Coyote)		•				[41]
	Felis catus (Cat)		•				[37]
	Phoca hispida (Ringed seal)		•				[42]
Lagomorpha	Oryctolagus cuniculus (European rabbit)		•				[27]
Arctiodactyla	Bos grunniens (Yak)			•			[43]
	Bos taurus (Cattle)			•			[4]
	Bison bonasus (European bison)			•			[31]
	Capra hircus (Goat)		•				[22, 44]
	Ovis aries (Sheep)		•	•			[22, 44, 45]
	Camelus bactrianus (Bactrian camel)			•			[46]
	Gazella cuvieri (Cuvier's gazelle)		•				[47]
	Giraffa camelopardalis reticulate (Reticulated giraffe)		•				[48]
	Odocoileus hemionus (Mule deer)						[49]
	Oreamnos americanus (Mountain goat)			•			[50]
	Rangifer tarandus (Reindeer)					•	[51]
	Sus scrofa (Pig)		•				[15]
	Syncerus caffer (African buffalo)	•					[52]
Perissodactyla	Equus africanus (Donkey)	•					[53]
	Equus ferus caballus (Horse)	•	•				This study; [54]
Primates	Homo sapiens (Human)		•	•			[55, 56]
	Macaca fascicularis (Crab-eating macaque)		•				[57]
	Nycticebus coucang (Sunda slow loris)		•				[58]
Struthioniformes	Struthio camelus (Ostrich)		•				[59]
Caprimulgiformes	Podargus strigoides (Tawny frogmouth)		•				[9]
Galliformes	Rollulus roulouli (Crested partridge)			•			[9]
Squamata	Elaphe obsolete (Western rat snake)		•				[60]
	Oxyuranus scutellatus (Coastal taipan)		•				[6]
	Python regius (Python regius)		•				[61]
	Spilotes pullatus (Spilotes pullatus)		•				[19]
	Varanus salvadorii (Varanus salvadorii)		•				[31]
Anura	Ceratophrys ornate (Argentine horned frog)		•				[12]

Table 2

Prepatent and patent period for Cryptosporidium proliferans and different Cryptosporidium muris and Cryptosporidium andersoni strains and isolates in various hosts.

Species / strain	Original host	Recipient host (age in days)	Prepatent period (day)	Patent period (day)	Infectious dose	Reference
C. muris MCR	Mus musculus	Mus musculus ICR (21)	4–10	36–73	2×106	[34, 62]
C. muris RN66	Rattus norvegicus	Mus musculus ICR (21)	6	> 30	2.4×106	[63]
C. muris CB03	Camelus bactrianus	Mus musculus BALB/c (56)	10	15–18	1×106	[25]
C. andersoni LI03	Bos taurus	Mus musculus BALB/c (56)	NI	NI	1×106	[22]
C. proliferans	Tachyoryctes splendens	Mus musculus BALB/c (56)	10–12	23–28	1×106	[20]
C. muris MCR	Mus musculus	Ovis aries (1–20)	28–35	16–38	2×107	[44]
C. muris RN66	Rattus norvegicus	Ovis aries (21)	NI	NI	1×107	[22]
C. muris CB03	Camelus bactrianus	Ovis aries (21)	24	< 10	1×107	[22]
C. andersoni LI03	Bos taurus	Ovis aries (21)	NI	NI	1×107	[22]
C. proliferans	Tachyoryctes splendens	Ovis aries (21)	NI	NI	1×107	[22]
C. muris MCR	Mus musculus	Capra hircus (1–20)	19–35	34–85	2×107	[44]
C. muris RN66	Rattus norvegicus	Capra hircus (21)	NI	NI	1×107	[22]
C. muris CB03	Camelus bactrianus	Capra hircus (21)	28	20–60	1×107	[22]
C. andersoni LI03	Bos taurus	Capra hircus (21)	NI	NI	1×107	[22]
C. proliferans	Tachyoryctes splendens	Capra hircus (21)	NI	NI	1×107	[22]
C. muris RN66	Rattus norvegicus	Mus musculus SCID (28)	6	> 28	1×106	[35]
C. muris CB03	Camelus bactrianus	Mus musculus SCID (56)	7	> 60	1×106	Unpublished data
C. andersoni LI03	Bos taurus	Mus musculus SCID (56)	NI	NI	1×106	[22]
C. andersoni Kawatabi	Bos taurus	Mus musculus SCID (28)	14	> 28	1×106	[35]
C. proliferans	Tachyoryctes splendens	Mus musculus SCID (56)	12–18	> 60	1×106	[20]
C. muris CB03	Camelus bactrianus	Lasiopodomys brandtii	14	14–32	1×106	[24]
C. andersoni LI03	Bos taurus	Lasiopodomys brandtii	NI	NI	1×106	[24]
C. proliferans	Tachyoryctes splendens	Lasiopodomys brandtii	14	> 40	1×106	[24]
C. muris RN66	Rattus norvegicus	Mastomys coucha (56)	7–8	38–56	1×106	Unpublished data
C. muris CB03	Camelus bactrianus	Mastomys coucha (56)	8–10	51–76	1×106	Unpublished data
C. andersoni LI03	Bos taurus	Mastomys coucha (56)	20	46–59	1×106	[64]
C. proliferans	Tachyoryctes splendens	Mastomys coucha (56)	15–20	> 140	1×106	This study
C. muris HZ206	Mus musculus	Mastomys coucha (56)	6–8	48–77	1×106	This study
C. andersoni	Bos taurus	Meriones unguiculatus (56)	15–19	18–65	1×106	[30]
C. proliferans	Tachyoryctes splendens	Meriones unguiculatus (56)	18–22	> 90	1×106	[21]

NI–non infectious

A number of variant strains of C. muris and C. andersoni have been described, based on polymorphisms in the small ribosomal subunit (SSU) gene, and differences in host specificity, pathogenicity, and course of infection. For example, C. muris Japan field genotype (also known as C. muris Kawatabi strain) differs from C. muris RN66 (reference strain) [17, 18]. Similarly, among C. andersoni strains, only C. andersoni Kawatabi, is infectious for the domestic mouse. Previous studies suggest that C. muris and C. andersoni represent a complex of cryptic species (Tables 1 and 2), but phylogenetic and biological data to support separate species are mostly lacking.

We undertook this study to examine the host specificity, course of infection, pathogenicity, oocyst morphology, and molecular characteristics of C. muris strain TS03. Based on the collective data from this and other studies [19-25], which show that C. muris strain TS03 is genetically distinct from C. muris and other known Cryptosporidium species, we propose the species name Cryptosporidium proliferans n. sp.

Materials and Methods

Source of oocysts and DNA for studies

The isolate of C. proliferans (previously known as C. muris TS03) used to determine experimental infectivity and oocyst morphology originated from a naturally infected East African mole rat (Tachyoryctes splendens) trapped in Kakamega, Kenya in 2003, and was maintained in susceptible laboratory rodents (SCID and southern multimammate mice Mastomys coucha) at two laboratories: Institute of Parasitology, Biology Centre of the Academy of Sciences of the Czech Republic and University of Veterinary and Pharmaceutical Sciences Brno, Czech Republic. DNA obtained from the laboratory-propagated C. proliferans isolate, and C. proliferans DNA isolated from the feces of an African buffalo (Syncerus caffer) in Central African Republic [52], a donkey (Equus africanus) in Algeria [53], and a domestic horse (Equus caballus) in the Czech Republic (unpublished) were used for phylogenetic analyses.

Oocysts of C. muris HZ206, originally isolated from a naturally infected domestic mouse (Mus musculus domesticus; Mmd) in Germany in 2012, were used as a C. muris control. Cryptosporidium muris HZ206 has been maintained at the Institute of Parasitology, Biology Centre of the Academy of Sciences of the Czech Republic in a wild-derived Mmd strain from Schweben, central Germany (10th generation of brother-sister mating; kept under the name SCHEST at the Institute of Vertebrate Biology, Brno, Czech Republic). For comparison of oocyst morphology a C. parvum isolate originating from a naturally infected 23-day-old Holstein calf was used.

Parasitological examination and oocyst preparation

Animal feces were screened for Cryptosporidium oocysts using fecal smears stained with aniline-carbol-methyl violet (ACMV) [65]. Fecal specimens were collected daily and stored in a 2.5% potassium dichromate solution at 4–8°C. Cryptosporidium oocysts were purified for morphometry, phylogeny, and infectivity analyses using sucrose gradient [66] and cesium chloride gradient centrifugation [67]. Purified oocysts were stored for up to 4 weeks at 4–8°C in PBS with antimycotics and antibiotics (100 UI penicillin, 10 μg/ml streptomycin, 0.25/ml μg amphotericin and 30 μg/ml gentamicin). The identity of the parasite was confirmed by sequence analysis of the SSU gene, using the method described below. The number of oocysts administered to animals was determined by hemocytometer counting. The viability of oocysts was examined by propidium iodide (PI) staining using a modification of a previously described assay [68]. Examined oocysts were washed in distilled water (DW; 105 oocysts in 100 μl) and mixed with 10 μl of PI (1% solution, SIGMA). After 30 min of incubation at room temperature in the dark, the oocysts were washed twice with DW. Oocyst viability was examined using fluorescence microscopy (filter 420 nm, Olympus IX70). Oocysts with red fluorescence were considered to be dead, and those without fluorescence were considered viable. A total of 500 oocysts were counted.

Oocyst morphology

Oocysts were examined using differential interference contrast (DIC) microscopy following ACMV and Auramine Phenol (AP) staining [69], or fluorescence microscopy following labeling with genus-specific FITC-conjugated antibodies (Cryptosporidium IF Test, Crypto cel, Medac) (Olympus IX70 microscope; Olympus CZ, Czech Republic). Morphology and morphometry were determined using digital analysis of images (M.I.C. Quick Photo Pro v.3.0 software; Optical Service, Czech Republic) collected using an Olympus Digital Colour Camera DP73 (17.29 megapixels). A 20-μl aliquot containing 105 purified oocysts was examined for each measurement. Length and width of oocysts (n = 100) were measured under DIC at 1000× magnification, and these were used to calculate the shape index and length-to-width ratio of each oocyst. As a control, the morphometry of C. parvum (n = 100) from a naturally infected 23-day-old Holstein calf, and C. muris HZ206 (n = 100) were measured. Photomicrographs of C. proliferans oocysts observed by DIC, ACMV, AP and IFA were deposited as a phototype at the Institute of Parasitology, Biology Centre of the Academy of Sciences of the Czech Republic.

DNA extraction and molecular analyses

Total DNA was extracted from 200 mg of feces, 105 purified oocysts, or 200 mg of tissue by bead disruption for 60 s at 5.5 m/s using 0.5 mm glass beads in a FastPrep®24 Instrument (MP Biomedicals, CA, USA) followed by isolation/purification using a commercially available kit in accordance with the manufacturer’s instructions (QIAamp® DNA Stool Mini Kit or DNeasy® Blood & Tissue Kit, Qiagen, Hilden, Germany). Purified DNA was stored at −20°C prior to being used for PCR. A nested PCR approach was used to amplify ∼830 bp of the small ribosomal subunit (SSU) gene [70, 71], ∼1066 bp of the actin gene [72], and four previously described minisatellite genes—MS1 (encoding a hypothetical protein), MS2 (encoding a 90-kDa heat shock protein), MS3 (encoding a hypothetical protein), and MS16 (encoding a leucine-rich repeat family protein) [19]. Both primary and secondary PCR reactions were carried out in a volume of 50 μl; the primary reaction contained 2 μl of genomic DNA (or water as a negative control) and the secondary reaction contained 2 μl of the primary reaction as template. In addition, primers for nested PCR to amplify Thrombospondin-Related Adhesive Protein of Cryptosporidium-1 (TRAP-C1; ~955 bp), Cryptosporidium Oocyst Wall Protein (~ 400 bp), and Heat shock protein (HSP70; ~515 bp) were designed for this study using PrimerQuest online software (IDT, http://www.idtdna.com/) and tested by software Serial Cloner v 2.6.1. (http://serialbasics.free.fr/Serial_Cloner.html). Primers amplifying TRAP-C1 were designed using the C. muris sequence in GenBank (CMU_020100). The primers for primary reactions were TRAP-G-F1 (GGA GAT CCT TTA TGT GTT G) and TRAP-G-R1 (CCT GTA CAA ATT CTT CTG AT) and secondary reaction TRAP-G-F2 (GCT CAG AAG ATC CAA GTA) and TRAP-G-R2 (GAT TGC TCT GAA CTA GGA). Primers amplifying HSP70 gene were designed using the C. muris sequence in GenBank (CMU_009950). The primers for primary reactions were HSPAvA1-F (GCT CGT GGT CCT AAA GAT AA) and HSPAvA1-R (ACG GGT TGA ACC ACC TAC TAA T) and secondary reaction HSPAvA2-F (ACA GTT CCT GCC TAT TTC A) and HSPAvA2-R (GCT AAT GTA CCA CGG AAA TAA). Primers amplifying the COWP gene were designed using the consensus of sequences in GenBank (B471649, AB471650, KF747672, DQ989571, DQ989570, DQ060431, KF419210, AF266275, AF161580, DQ060430, AF266262, AY282693, AF161579, AF266264, AB514044, AB514043, AY643491, AB089289, and AB089287). The primers for primary reactions were COWP-torto-F1 (GAA TGT CCT CCT GGG ACT GTA) and COWP-torto-R1 (AGT TCC TGG TGG ACA TAC TG) and secondary reaction COWP-torto-F2 (TCC TCC TGG GAC TGT ATT GGA) and COWP-torto-R2 (GGT GGA CAT ACT GGT TGT GTT G). The primary PCR reactions, for TRAP-C1, HSP70, and COWP genes, were carried out in a volume of 50 μl containing 2 μl of DNA template (or water as a negative control), 1×PCR buffer, 3 mM MgCl2, 200 μM dNTPs, 0.2 μM of each primer, and 1.5 U of Taq DNA polymerase (Top Bio, Czech Republic). Secondary reactions were carried out under similar conditions with 2 μl of primary product used as template. PCRs were run in a thermo cycler with an initial denaturation of 94°C for 5 min, followed by 35 cycles of 94°C for 45 s, 55°C (TRAP-C1), 50°C (HSP70 and COWP) for 45 s, 72°C for 1 min. A final elongation step of 72°C for 10 min was included to ensure complete extension of amplified products. Conditions were the same for both primary and secondary reactions. DNA of C. andersoni was used as positive control. Secondary PCR products were detected by agarose gel (2.0%) electrophoresis, visualized by ethidium bromide staining (0.2 μg/ml) and extracted using QIAquick® Gel Extraction Kit (Qiagen). Purified secondary products were sequenced in both directions with an ABI 3130 genetic analyser (Applied Biosystems, Foster City, CA) using the secondary PCR primers and the BigDye1 Terminator V3.1 cycle sequencing kit (Applied Biosystems, Foster City, California) in 10 μl reactions. Amplification and sequencing of each locus was repeated three times.

Phylogenetic analyses

The nucleotide sequences of each gene obtained in this study were edited using the ChromasPro 1.7.5 software (Technelysium, Pty, Ltd.), manually edited, and aligned with each other and with reference sequences from GenBank using MAFFT version 7 online server with automatic selection of alignment mode (http://mafft.cbrc.jp/alignment/software/). Phylogenetic analyses were performed and best DNA/Protein phylogeny models were selected using the MEGA6 software [73, 74]. Phylogenetic trees were inferred by the i) neighbor-joining (NJ), ii) maximum likelihood (ML), and iii) maximum parsimony (MP) method. Bootstrap support for branching was based on 1000 replications. Neighbor-joining phylograms were edited for style using CorelDrawX7. Sequences have been deposited in GenBank under the accession numbers KR090615-KR090632 and KT731193-KT731212.

Transmission studies

Animals

The infectivity and pathogenicity of C. proliferans for the eight-week-old southern multimammate mice (Mastomys coucha), and adult budgerigars (Melopsittacus undulatus) (Institute of Parasitology, Biology Centre of the Academy of Sciences of the Czech Republic, Czech Republic) was determined experimentally in this study. The infectivity of C. proliferans for SCID and BALB/c mice, gerbils, calves, kids, and lambs to C. proliferans (previously named C. muris strain TS03) was determined previously (Table 2) [21, 22, 25].

Experimental design

To prevent environmental contamination with oocysts, southern multimammate mice were housed in plastic cages with sterilized wood-chip bedding situated in flexible film isolators (BEM, Znojmo, Czech Republic) with high-efficiency particulate air filters. Birds were kept in cages placed in a room separated from other animals. The southern multimammate mice and budgerigars were supplied with a standard sterilized diet for rodents and birds, respectively, and sterilized water ad libitum. For three week prior to infection, fecal samples from all animals were screened daily for the presence of Cryptosporidium spp. using parasitological and molecular tools as described in previous sections. Each animal was inoculated orally by stomach tube with 106 purified viable oocysts of each species (C. proliferans and C. muris HZ206) suspended in 200 μl of distilled water. Each animal used as negative control was inoculated with 200 μl of distilled water only. A total of nine budgerigars were used: three as negative controls, three infected with C. proliferans, and three infected with C. muris HZ206. Fecal samples from all experimental birds were collected daily and experiments were terminated 30 days post infection (DPI). A total of 94 southern multimammate mice were divided into three groups: i) control group (n = 34), ii) group infected with C. muris HZ206 (n = 30), and iii) group infected with C. proliferans (n = 30). All animals were weighed before the start of the experiment, with a precision of 0.1 g. Each animal was kept in separate cages. Fecal samples from all experimental southern multimammate mice were taken daily for the first 28 days, then every 7 days. Experiments were terminated 140 DPI. Animals were euthanized via cervical dislocation (according Law of the Czech Republic No. 419/2012 Sb.). Fecal samples from all experimental animals were stained by ACMV and the presence of Cryptosporidium specific DNA was confirmed using nested PCR targeting the SSU gene every 7 days. Every 28 days, 4 southern multimammate mice were sacrificed from each group, and each was examined for body weight, stomach size and weight, and the surface ratio of glandular to non-glandular parts of the stomach. Histopathological changes of gastric mucosa due to infection of C. proliferans and C. muris were evaluated using histological methods. Results were compared to uninfected animals, which were tested using the same procedures. Course of infection indicators, including fecal consistency, fecal color and infection intensity, were examined. Infection intensity was reported as the number of oocysts per gram (OPG) of feces as previously described [64].

Clinical and histopathological examinations

A complete examination of all gastrointestinal organs was conducted at necropsy. Tissue samples from the stomach, small intestine, and large intestine (the entire tract was divided into 1 cm sections) were processed for histology [75] and for PCR analyses (see Section 2.4.). Histology sections were stained with hematoxylin and eosin (HE), Wolbach’s modified Giemsa, Periodic Acid Schiff (PAS) stain, and genus-specific FITC conjugated monoclonal antibodies targeting Cryptosporidium oocyst wall antigens (Cryptosporidium IF Test, Crypto Cel, Medac).

Animal care

Animal caretakers wore disposable coveralls, shoe covers, and gloves every time they entered the facility rooms. All wood-chip bedding, feces, and disposable protective clothing were sealed in plastic bags, removed from the buildings and incinerated. All housing, feeding, and experimental procedures were conducted under protocols approved by the Institute of Parasitology, Biology Centre of the Academy of Sciences of the Czech Republic and Institute and National Committees (Protocols No. 52/2014).

Statistical analyses

The hypothesis tested in the analysis of oocyst morphometry and size of stomach was that two-dimensional mean vectors of measurement are the same in the two populations being compared. Hotelling’s T2 test was used to test the null hypothesis. The Bartlett test was used to test homoscedasticity of differences in the prepatent and patent periods of different infections.

Results

Oocysts of C. proliferans measuring 6.8–8.8 (mean = 7.7 μm) × 4.8–6.2 μm (mean = 5.3) with a length to width ratio of 1.48 (n = 100) were significantly longer and narrower (p <0.05) than C. muris HZ206 oocysts, which measured 6.3–8.1 (mean = 7.5 μm) × 5.0–6.6 (mean = 5.7 μm) with length to width ratio of 1.35 (n = 100). Cryptosporidium parvum oocysts were significantly smaller (p<0.05) than C. proliferans and C. muris HZ206, measuring 5.1–5.5 (mean = 5.3 μm) × 4.6–5.2 (mean = 4.7 μm) with length to width ratio of 1.12 (1.07–1. 32) (n = 100) (Fig 1A). Oocysts of C. proliferans recovered from experimentally infected mice were morphologically similar to those used for infection. Oocysts in fecal smears showed typical Cryptosporidium ACMV and AP staining characteristics (Fig 1B and 1C). Fixed C. proliferans oocysts labeled with FITC conjugated anti-Cryptosporidium oocyst wall antibody and examined by epifluorescence microscopy displayed typical apple green, halo-like fluorescence (Fig 1D).

Fig 1

Cryptosporidium proliferans, Cryptosporidium muris HZ206, and Cryptosporidium parvum oocysts in (A) differential interference contrast microscopy and stained by (B) aniline–carbol–methyl violet (C) Auramine Phenol and (D) anti-Cryptosporidium FITC-conjugated antibody.

Bar = 10 μm.

Molecular characterization

At the SSU locus, C. proliferans isolates from an East African mole rat, African buffalo, donkey, and domestic horse shared 100% identity with each other and with an isolate (EU096237) from an Eastern gray squirrel in the USA (Fig 2A). At the TRAP-C1 locus, isolates of C. proliferans from the four different hosts shared 100% identity with each other and differed from C. muris RN66 by five SNPs, two of which were non-synonymous (Fig 2B).

Fig 2

Phylogenetic relationships between Cryptosporidium proliferans (highlighted) and other Cryptosporidium spp. as inferred by a neighbor-joining analysis (NJ)/maximum parsimony(MP)/maximum likelihood (ML) of (A) the SSU (706 base positions in the final dataset; ML = log -2886.67) and (B) TRAP-C1 (531 base positions in the final dataset, ML = log -1929.25).

The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates). Numbers at the nodes represent bootstrap values for the nodes gaining more than 50% support. Scale bar included in each tree.

At the HSP70 locus, C. proliferans isolates from the four different hosts shared 100% identity with each other and with the C. muris Kawatabi isolate (AY643490) from Apodemus speciosus in Japan (Fig 3A), but differed from C. muris RN66 by a synonymous SNP.

Fig 3

Phylogenetic relationships between Cryptosporidium proliferans (highlighted) and other Cryptosporidium spp. as inferred by a neighbor-joining analysis (NJ)/maximum parsimony(MP)/maximum likelihood (ML) of (A) HSP70 (211 base positions in the final dataset, ML = log -1745.42) and (B) COWP (369 base positions in the final dataset, ML = log -532.78).

The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates). Numbers at the nodes represent bootstrap values for the nodes gaining more than 50% support. Scale bar included in each tree.

Isolates of C. proliferans from the East African mole rat and African Buffalo shared 100% identity at the COWP locus, and differed from C. muris RN66 (CMU_007770) by a synonymous SNP (C/T) at position 597, using the C. parvum Iowa isolate as a reference sequence (Cgd6_2090) (Fig 3B). Cryptosporidium proliferans COWP sequences were not obtained from donkey or domestic horse isolates.

At the actin locus, isolates from the East African mole rat, African buffalo, and donkey shared 100% identity with each other and differed from C. muris by four synonymous SNPs (Fig 4A). An actin sequence was not obtained from the domestic horse isolate.

Fig 4

Phylogenetic relationships between Cryptosporidium proliferans (highlighted) and other Cryptosporidium spp. as inferred by a neighbor-joining analysis (NJ)/maximum parsimony (MP)/maximum likelihood (ML) of (A) actin (728 base positions in the final dataset, ML = log = -5522.35) and (B) MS1 (436 base positions in the final dataset, ML = log -886.75). The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates).

Numbers at the nodes represent bootstrap values for the nodes gaining more than 50% support. Interrupted branches have been shortened five-fold. Scale bar included in each tree.

Cryptosporidium proliferans isolates clustered separately from C. muris and C. andersoni at each of the four microsatellite loci examined. Three C. proliferans subtypes formed a single cluster at the MS1 locus: MS1-P1 was detected in isolates from the East African mole rat and African Buffalo; MS1-P2 and MS1-P3 were detected in the donkey and domestic horse, respectively (Fig 4B). A single C. proliferans subtype was detected at the MS2 (MS2-P1) and MS3 loci (MS3-P1) (Fig 5A and 5B). Similar to MS1, three C. proliferans subtypes formed a single cluster at the MS16 locus: MS16-P1 was detected in isolates from the East African mole rat and African Buffalo; MS16-P2 and MS16-P3 were detected in the domestic horse and donkey, respectively (Fig 5C).

Fig 5

Phylogenetic relationships between Cryptosporidium proliferans (highlighted) and other Cryptosporidium spp. as inferred by a neighbor-joining analysis (NJ)/maximum parsimony (MP)/maximum likelihood (ML) of (A) MS2 (442 base positions in the final dataset, ML = log -792.33), (B) MS3 (485 base positions in the final dataset, ML = log = -832.45), and (C) MS16 (580 base positions in the final dataset, ML = log = -956.78).

The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates). Numbers at the nodes represent bootstrap values for the nodes gaining more than 50% support. Scale bar included in each tree.

In a tree constructed from concatenated sequences of SSU, actin, HSP70, and TRAP-C1, C. proliferans clustered separately from C. muris RN66, C. muris HZ206, C. parvum, C. hominis, and C. meleagridis (Fig 6).

Fig 6

Phylogenetic relationships between Cryptosporidium proliferans and selected Cryptosporidium spp. as inferred by a neighbor-joining analysis (NJ)/maximum parsimony(MP)/maximum likelihood (ML) analysis of a concatenated sequence constructed from partial DNA sequences of SSU, actin, COWP, HSP70, and TRAP-C1 genes (1991 base positions in the final dataset, ML = log = -4368.98).

The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates). Numbers at the nodes represent bootstrap values for the nodes gaining more than 50% support. Scale bar included in tree.

Experimental transmission studies

Oocysts used for experimental infections had >95% viability, determined by PI staining. Experimentally inoculated budgerigars did not produce microscopically detectable infection. Histological and molecular examination of gastrointestinal tract tissue from budgerigars did not reveal the presence of Cryptosporidium developmental stages.

Both C. proliferans and C. muris HZ206 were infectious for southern multimammate mice. While southern multimammate mice began shedding C. muris HZ206 oocysts 6–8 DPI, C. proliferans shedding began significantly later at 18–21 DPI (p<0.001). PCR amplification of the Cryptosporidium SSU gene was unsuccessful from 2 to 5 DPI and 2 to 17 DPI in mice infected with C. muris HZ206 and C. proliferans, respectively.

Cryptosporidium muris HZ206 infection intensity ranged from 2.0×103 to 1.4×105 OPG, with maximum shedding at 42 DPI. Oocyst shedding intensity decreased from 42 DPI and microscopic detection was intermittent by 77 DPI, although specific DNA was present in feces throughout the patent period (Fig 7). Cryptosporidium proliferans infection intensity ranged from 1.0×105 to 9.5×106 OPG with maximum shedding at 126 DPI. In contrast to C. muris HZ206, the infection intensity continued to increase throughout the experiment (Fig 7). Southern multimammate mice experimentally infected with C. proliferans mostly developed lifelong infection (data not shown).

Fig 7

Course of infection of Cryptosporidium proliferans and Cryptosporidium muris HZ206 in Mastomys coucha based on coprological examination of feces.

No clinical signs of cryptosporidiosis were detected in southern multimammate mice during the first 84 DPI with C. proliferans; however, these mice subsequently began to lose weight compared to time-matched mice in the C. muris HZ206 infection and uninfected control groups (p<0.05).

Histological examination of the glandular and non-glandular parts of the stomach in uninfected control mice showed no evidence of Cryptosporidium developmental stages, pathological alterations, or activation of glands. Also, mucus production was normal. Developmental stages were found only in the glandular part of the stomach of experimentally infected groups and high numbers were typically associated with high oocyst shedding. In C. proliferans positive southern multimammate mice, infected glands were markedly dilated, and filled with necrotic material, mucus, and numerous development stages. Epithelial cells of infected glands were atrophic and exhibited cuboidal or squamous metaplasia. The epithelium also was significantly proliferated into the lumen of the stomach and formed papillary structures. Such proliferation was not observed in uninfected control and C. muris HZ206 infected southern multimammate mice (Fig 8).

Fig 8

Height of mucosa of Mastomys coucha stomach of A) control group, B) Cryptosporidium muris HZ206 infection, and C) Cryptosporidium proliferans infection at 28 day post infection; D) stomach mucosa proliferation in Mastomys coucha with Cryptosporidium proliferans at 140 DPI.

Haematoxylin and eosin. Bar = 150 μm.

Macroscopically, the gastric mucosa of southern multimammate mice infected with C. proliferans had confluent cauliflower-shaped lesions (Fig 8D). The lamina propria did not contain any inflammatory infiltrates. The gastric mucosa of C. muris HZ206 infected southern multimammate mice was less hyperplastic without significant gain. Mucus production was similar to that in the uninfected control group and was significantly less than in C. proliferans infected southern multimammate mice. A gradual retreat of infection from the lower layers of the epithelium was observed in southern multimammate mice infected with C. muris from 56 DPI, and no developmental stages of C. muris HZ206 were detected from 112 DPI using histological methods. The total stomach weight in infected southern multimammate mice was increased compared to the negative control group. While the stomach weight of southern multimammate mice infected with C. muris HZ206 increased up to 1.5 fold, the stomach weight of mice infected with C. proliferans significantly increased up to 5.7 fold (p<0.001; Fig 9). The proliferating mucosa was the major contributor to the increase in stomach weight, while the submucosa, muscularis, and serosa did not significantly change in any group (p>0.05). The mucosa from C. proliferans and C. muris infections increased 5.6 and 1.6 fold, respectively compared to the uninfected control group (p<0.001). In addition, the ratio of glandular to non-glandular surfaces changed from 55:45 to 80:20 in southern multimammate mice infected with C. proliferans. This change was not observed in uninfected control and C. muris HZ206 infected southern multimammate mice (data not shown).

Fig 9

Change of stomach weight of Mastomys coucha during experimental infection with Cryptosporidium proliferans and Cryptosporidium muris HZ206 compared to the control group.

Taxonomic summary of Cryptosporidium proliferans n. sp.

Diagnosis

Oocysts are shed fully sporulated. Sporulated oocysts 6.8–8.8 (mean = 7.7 μm) × 4.8–6.2 μm (mean = 5.3) with a length to width ratio of 1.48. Four sporozoites and residual body are present in each oocyst. The lifecycle of C. proliferans, including the description of endogenous stages, was described previously [23].

Type host

Tachyoryctes splendens (Rüppell, 1835).

Other natural hosts

Equus africanus (Fitzinger, 1857), donkey [53]; Sciurus carolinensis Gmelin, 1788, eastern gray squirrel [38]; Syncerus caffer (Sparrman, 1779), African buffalo [52]; Equus caballus Linnaeus, 1758, horse (unpublished, and this study).

Experimental host

Mastomys coucha (Smith, 1836), southern multimammate mouse; Mastomys natalensis (Smith, 1834), natal multimammate mouse; Meriones unguiculatus (Milne-Edwards, 1867), Mongolian gerbil; Lasiopodomys brandtii (Radde, 1861), Brandt's vole; Mus musculus Linnaeus, 1758, house mouse (strain BALB/c and SCID) [15, 22, 24, 25].

Prepatent and patent period of C. proliferans in model hosts are presented in Table 3.

Table 3

Prepatent and patent period of Cryptosporidium proliferans in various hosts.

Host	Prepatent period (DPI)	Patent period (DPI)	Reference
southern multimammate mouse	15–21	> 140	This study; [24]
laboratory mouse	6–8	up to 30	[25]
gerbils	18–22	> 90	[21]
voles	14	> 40	[24]

Type locality

Kakamega, Kenya

Site of infection

stomach, specifically the glandular part

Material deposited

A phototype, description of oocysts, and DNA are deposited at the Institute of Parasitology, Biology Centre of the Academy of Sciences of the Czech Republic.

DNA sequences

Partial sequences of SSU, actin, COWP, HSP70, TRAP-C1, MS1, MS2, MS3, and MS16 genes were submitted to GenBank under the accession numbers KR090615-KR090623 and KT731193-KT731212.

Etymology

The species name proliferans is derived from the Latin substantive “proliferatio” (meaning a proliferation) according to ICZN Article 11.9.1–3, as it appears to cause a proliferation of mucosa in infected stomach.

Discussion

The gastric species C. muris and intestinal species C. parvum were the first described Cryptosporidium species [3, 76]. Until the late 1990s, many intestinal Cryptosporidium spp. were regarded as subtypes of C. parvum; for example, C. hominis was known as C. parvum genotype I and C. canis as C. parvum canine genotype [77, 78]. Recognition of the subtypes as separate species was made possible by studies showing molecular and biological differences from C. parvum. In cases where only molecular differences are described, isolates are regarded as distinct genotypes rather than subtypes of C. parvum; for example, Cryptosporidium rat genotype I-IV [79-82]. Also, C. andersoni (previously known as C. muris-like) was separated from C. muris senso lato based on molecular and biological differences [4]. A number of reports over the past 15 years indicate the presence of cryptic species within the mammalian gastric Cryptosporidium group, but evidence to support the naming of a new species has thus far been lacking (see Tables 1 and 2 for references).

Although C. muris and C. andersoni are primarily rodent- and ruminant specific, respectively, there have been reports of C. muris in ruminants and C. andersoni in rodents (Table 1). Similarly, C. proliferans has been found in hosts belonging to the Rodentia, Arctiodactyla, and Perissodactyla. With the exception of the occurrence of C. proliferans in squirrels in the US, all the other isolates have been detected in animals in Africa [22, 38, 52] and Europe (this study).

The present study and a previous study [83] have shown that Cryptosporidium proliferans oocysts (7.7 × 5.3 μm) are longer and narrower than those of C. muris HZ206 (7.5 × 5.7 μm) and C. andersoni (7.6 × 5.5 μm). We have seen no change in the size of the oocysts during more than 10 years of oocyst passage through different hosts (data not shown). However, because the reported size of C. andersoni and C. muris oocysts is quite variable [4, 26, 28], oocyst morphology cannot be used to reliably distinguish these species from C. proliferans. The shape and size of C. proliferans is also significantly different from oocysts of intestinal species such C. xiaoi (3.94 × 3.44 μm), C. parvum (5.3 × 4.7 μm) or C. suis (6.2 × 5.5 μm) [84, 85].

Phylogenetic analyses based on SSU, COWP, TRAP-C1, HSP70, actin, MS1, MS2, MS3, and MS16 gene sequences have shown that C. proliferans is genetically distinct from known Cryptosporidium species. At the SSU locus, C. proliferans is 99.4% and 98.3%, similar to C. muris and C. andersoni, respectively. This is comparable to the similarities between C. hominis and C. cuniculus (98.9%), C. parvum and C. erinacei (99.5%), and C. bovis and C. xiaoi (99.5%). At the COWP locus, C. proliferans is 99.7% and 97.8% similar to C. muris RN66 and C. andersoni, respectively. Other closely related Cryptosporidium species, including C. erinacei and C. parvum and C. cuniculus and C. hominis, are identical at the COWP locus [86, 87]. At the TRAP-C1 locus, C. proliferans is 99.0% and 95.6% similar to C. muris RN66 and C. andersoni, respectively. In comparison, C. parvum and C. erinacei are 99.8% similar at the TRAP-C1 locus [86, 88]. At the HSP70 locus, C. proliferans is 99.5% and 97.1% similar to C. muris RN66 and C. andersoni, respectively. In comparison, C. parvum and C. erinacei are 99.0% similar at this locus. Cryptosporidium proliferans is identical to the C. muris Kawatabi isolate (AY643490) at the HSP70 locus [89]. The C. muris Kawatabi isolate also clusters with C. proliferans at the SSU locus, suggesting that these could be the same species. Further genetic and biological characterization of the Kawatabi isolate would be required to test this. At the actin locus, C. proliferans is 99.4% similar to C. muris RN66. In comparison, C. parvum and C. erinacei share 99.5% similarity at this locus. Cryptosporidium proliferans also is distinguishable from C. muris and C. andersoni at four minisatellite loci examined in this study. It was previously found that C. muris isolates RN66 and CB03 were identical to C. proliferans (C. muris TS03 in their study) at the SSU and minisatellite loci [19]. RN66, a commercially supplied reference strain (Waterborne Inc, LA), has been well characterized and the whole genome sequence is known. Multiple studies have shown that the SSU sequence of this strain differs from that of C. proliferans (TS03). Similarly, previous studies have shown that CB03 is identical to RN66, and different from TS03, at the SSU locus [22, 25]. Therefore, Feng et al. acknowledged that their finding was likely the result of cross contamination during isolate handling [19].

The prepatent period for C. proliferans infection in SCID mice, which lack T- and B-cell immunity, ranged from 12–18 DPI with a mean of 14 DPI [20]. This is similar to the prepatent period for C. andersoni Kawatabi (14 DPI) and longer than the prepatent periods for C. muris RN66 (6 DPI) and C. muris CB03 (7 DPI; unpublished) in SCID mice [35]. In a study on C. muris RN66 infection in nude mice, which specifically lack T-cell immunity [90], a 10 DPI prepatent period was reported following a dose of 1 million oocysts, the dose used in the present study. Longer prepatent periods were observed only with lower doses of 20,000 oocysts (15–18 DPI) and 400 oocysts (16 DPI). Similarly, a prepatent period of 10 DPI was observed in both outbred nude and SCID mice infected with 500,000 oocysts of C. muris RN66 [91]. Consistent with observations in immunocompromised mice, the prepatent period for C. proliferans (15–20 DPI) in immunocompetent southern multimammate mice is similar to that for C. andersoni (20 DPI), and longer than that for C. muris (6–10 DPI). Collectively, these data support the conclusion that C. proliferans has a longer prepatent period than C. muris in mice.

Cryptosporidium proliferans develops exclusively in the glandular part of the stomach, similar to C. muris and C. andersoni [4, 47, 75, 92], and its lifecycle [23] is similar to that of C. muris [93].

The clinical course of C. proliferans infection in immunocompetent southern multimammate mice is considerably different to that of C. muris. These mice shed oocysts of C. proliferans for much longer and at a greater intensity than oocysts of C. muris, and only the C. proliferans infection causes significant clinical and pathological changes, including weight loss and massive proliferation of the gastric mucosa that is associated with an almost 6-fold increase in stomach weight. Although gastric cryptosporidia rarely induce clinical symptoms in mammals [37, 75, 92, 94], C. andersoni infection in cattle has caused up to a three-fold increase in abomasum weight, decreased milk production, and loss of body condition [40, 95]. Cryptosporidium andersoni also causes a lifelong infection in cattle and mice [4, 22, 96], similar to C. proliferans in multimammate mice.

Consistent with most previous studies on C. muris and C. andersoni, infiltrates in propria of the mucosa are absent in animals infected with C. proliferans [22, 29, 37, 75, 92–94].

In conclusion, molecular and biological data support the establishment of C. muris TS03 as a new species and we propose the name Cryptosporidium proliferans.