Introduction

Cryptosporidium species are intracellular protozoan organisms that infect the gastrointestinal epithelial cells of a wide range of vertebrates, including humans (Dubey et al. 1990; Fayer et al. 2000). Cryptosporidiosis in humans has been reported globally and occurs sporadically or as outbreaks following zoonotic transmission from farm animals, person-to-person spread, or contamination of water supplies (O’Donoghue 1995). Since large numbers of oocysts are released from animals to surface waters, animal husbandry is a potential source of human infection (de Graaf et al. 1999). Wild mammals, particularly rodents, have been identified as reservoirs of Cryptosporidium (Chalmers et al. 1997; Morgan et al. 1999). In Japan, the presence of Cryptosporidium has been extensively studied in domestic animals raised on farms and slaughter houses (Kaneta and Nakai 1998; Koyama et al. 2005); however, only a few wild animals have been investigated (Iseki 1998). We studied the presence of Cryptosporidium oocysts in wild animals and zoo housed animals in Japan and detected oocysts in large Japanese field mice, Apodemus speciosus, which were captured on the beef farm where the C. andersoni Kawatabi strain was isolated (Satoh et al. 2003, Nakai et al. 2004). Cryptosporidium has been classified on the basis of oocyst size, host specificity, and parasitized organs. Following elucidation of the DNA sequences of the genes for 18S rRNA (Morgan et al. 1999), heat shock protein (HSP) (Sulaiman et al. 2000), and cryptosporidian oocyst wall protein (COWP) (Xiao et al. 2000) of various Cryptosporidium spp., differences in the DNA sequences have become a new and additional criterion for the classification of Cryptosporidium.

In this study, we used morphological, biological, and genetic analysis to characterize an isolate from the large Japanese field mice, A. speciosus.

Materials and methods

Animals

From April end to mid November 2000 and from mid April to early September 2001, Sherman traps were used to capture small animals on the Kawatabi farm of Tohoku University. This was the farm where C. andersoni had been constantly detected in several beef cattle (Nakai et al. 2004). Kawatabi farm is located in the northern part of the main island of Japan (38°45′N, 140°45′E), and the altitude ranges from 200 to 630 m. Live caught animals were housed individually in cages with a wire mesh floor for 1–4 weeks prior to the experiments. Animals were fed on standard laboratory feed pellets for experimental mice and water ad libitum. Feces were collected, and the sugar flotation method was used to detect Cryptosporidium oocysts in them. Captured large Japanese field mice were mated, and their infants were housed in separate cages. All mice, including the laboratory mice, were kept according to the guidelines for animal experiments of Tohoku University.

Oocyst size

Feces that tested positive for Cryptosporidium oocysts were collected for 3 weeks and were preserved in 2.5% (w/v) potassium dichromate solution at 4°C. Oocysts were purified from feces by the sugar centrifugal flotation method and were preserved at 4°C in PBS; they were used for experiments within 1 week of their collection. Using an optical microscope, at 1,000 × magnification, 50 oocysts were measured from each of the following strains: C. muris RN66, which was originally isolated from a house rat (Iseki 1986); C. andersoni Kawatabi strain, which was originally isolated from grazing cattle in the Kawatabi farm (Satoh et al. 2003); and C. parvum HNJ-1, which was genotype 2 and was originally isolated from a Japanese woman (Masuda et al. 1991). These strains had been serially passaged in SCID mice (Charles River Co., Ltd. Japan) in our laboratory. The lengths, widths, and shape indices of these oocysts were compared using non-parametric Kruskal–Wallis analysis of variance between groups because their variances were different. Individual differences were determined by multiple comparisons of average rank.

Experimental transmission

The transmission experiment was carried out using large Japanese field mice, 1–4 weeks after their capture from the field, 2-week-old infant large Japanese field mice, 4-week-old female SCID mice, and 4-week-old male Crj:CD-1 (ICR) mice (Charles River Co., Ltd.). The absence of Cryptosporidium oocysts in these mice was confirmed by the sugar flotation method prior to the start of experiments. This method was also used to monitor the discharge of oocysts for 30 days post-inoculation.

Histological analysis

Samples were obtained from the liver, heart, kidney, lung, stomach, duodenum, jejunum, intestinum ileum, cecum, colon, and pancreas of a large Japanese field mouse and a SCID mouse that were infected with a Cryptosporidium field isolate; these samples were fixed with 10% formalin in PBS and embedded in paraffin. Sections were stained with hematoxylin and eosin (HE).

Genetic analysis

DNA from oocysts of a Cryptosporidium field isolate was extracted using the Mag Extractor Genome (Toyobo, Osaka, Japan). A primer set was used to amplify fragments of genes, namely, the 18S ribosomal RNA (rRNA) (5′-AACCTGGTTGATCC-3′ and 5′-GAATGATCCTTCCGCAGGTTC-3′). We designed the primer set using OLIGO 5.0 (National BioScience Inc., Plymouth, Minn.), and its sequence was based on the sequence of a rock hyrax strain of C. muris (GenBank accession no. AF093496). PCR amplification was performed in a mixture that had the following final composition: 1× PCR buffer, 1 mM MgSO4, 0.2 mM each dNTP, 0.3 μM each primer, and 1.0 U KOD polymerase (Toyobo). The reactions were preheated at 94°C for 2 min and cycled 40 times at 94°C for 15 s, 58°C for 30 s, and 68°C for 2 min and then at 68°C for 10 min. The PCR products were electrophoresed on 1.0% agarose and visualized by ethidium bromide staining. The PCR fragment of 18S rRNA was cloned using the pT7Blue Perfectly Blunt Cloning Kit (Novagen, Darmstadt, Germany). The PCR products were purified using the Mag Extractor-PCR & Gel Clean up-kit (Toyobo). At least three clones of each purified PCR product were sequenced using the ABI PRISM BigDye Terminator Cycle Sequencing Kit (Applied Biosystems) on an ABI PRISM 310 Genetic Analyzer (Applied Biosystems) according to the manufacturer’s instruction.

Phylogenetic analysis

Nucleotide sequences obtained from our isolate and its homologues were identified using the Blast database search programs (Altschul et al. 1997). Homologous sequences were obtained from GenBank. The following additional Cryptosporidium 18S rRNA sequences were obtained from GenBank: C. muris rock hyrax (AF093498), C. muris bactrian camel (AF093497), C. muris RN66 (AB089284), C. andersoni Kawatabi strain (AB089285), C. serpentis (AF093502), C. baileyi (AF093495), C. felis (AF108862), C. canis (AF112576), C. mursupial genotype (AF108860), C. suis (AF115377), C. meleagridis (AF112574), C. wrairi (AF115378), C. parvum mouse isolate (AF112571), C. hominis (AF093489), C. parvum deer isolate (AF093494), and C. parvum (AF093490). The sequence of our isolate was aligned against those of the others using the ClustalW multiple alignment program (Thompson et al. 1994). The neighbor joining tree was constructed using the MEGA version 2.1 program (Kumar et al. 2001), and the genetic distance was calculated using Kimura’s 2-parameter model. The maximum parsimony analysis was also conducted with the same alignments using the DAMBE program (Xia and Xie 2001). The maximum likelihood tree was also constructed using the DAMBE program. The substitution model was set to an algorithm in which the transition to transversion ratio was 2. Bootstrap proportions computed for the maximum likelihood using the resampling were estimated by the log likelihood (RELL) method. In the construction of both the neighbor joining and maximum likelihood trees, a sequence of Eimeria tenella (AF026388) was used as the outgroup for the 18S rRNA analysis as reported previously (Xiao et al. 1999, 2002).

Results

Survey of Cryptosporidium oocysts from wild animals

Twenty-five large Japanese field mice, A. speciosus; five Norway rats, Rattus norvegicus; two house mice, Mus musculus; and a Japanese field vole, Microtus montebelli were captured on the Kawatabi farm (Table 1). Cryptosporidium oocysts were detected only from two large Japanese field mice (Fig. 1).

Table 1 Prevalence of Cryptosporidium species in wild rodents caught in the Kawatabi Farm
Full size table
Fig. 1
figure 1

Nomarski interference contrast photomicrographs of oocysts from a large Japanese field mouse. Magnification is 1,000×. Bar = 5 μm

Full size image

Oocyst size

Oocysts isolated from the large Japanese field mouse were 5.70±0.59×7.66±0.64 μm in size and had a length/width ratio of 1.36±0.17 (Table 2). The sizes were significantly larger than those of C. parvum HNJ-1 (P < 0.01), while the lengths were significantly smaller than those of C. muris RN66 (P < 0.05).

Table 2 Sizes of Cryptosporidium oocysts (in μm) a
Full size table

Experimental infection

The Cryptosporidium field isolate was infective to adult and infant large Japanese field mice, adult ICR mice, and SCID mice (Table 3). The prepatent periods for the Cryptosporidium field isolate were 17, 19, 11, and 11 days in these mice, respectively (data were not shown). Although C. muris was infective to adult large Japanese field mice at a high dose of oocysts, it was not infective to adult and infant Japanese field mice and adult ICR mice at a medium dose of oocysts. C. andersoni was not infective to adult and infant large Japanese field mice and ICR mice, even at a high dose. Clinical signs were not observed in any animals infected with the Cryptosporidium field isolate.

Table 3 Experimental infection of Cryptosporidium spp. to mice
Full size table

Histological observation

On day 24 post-inoculation of oocysts of the Cryptosporidium field isolate, a large number of developing stages of Cryptosporidium were detected in the gastric gland of the stomach of the large Japanese field mouse. However, no stages were detected from the liver, heart, kidney, lung, stomach, duodenum, jejunum, ileum, cecum, colon, and pancreas. The majority of the gastric glands were dilated, hypertrophied, and filled with numerous parasites (Fig. 2). The result was similar to that observed in the infected SCID mouse.

Fig. 2
figure 2

Hematoxylin and eosin staining of the stomach of a large Japanese mouse infected with the Cryptosporidium field isolate. The gastric glands are filled with cryptosporidial developing stages (arrowhead). Scale bar = 20 μm

Full size image

Genomic analysis of the Cryptosporidium field isolate

The 18S rRNA gene (1,673 bp) was partially sequenced, and the sequence of the field isolate submitted to GenBank had the accession number AY642591. In the 18S rRNA gene, the Cryptosporidium field isolate had 6 bp substitutions and 2 bp deletions compared with C. muris RN66 and 13 bp substitutions and 2 bp deletions compared with the C. andersoni Kawatabi strain (data not shown).

Phylogenetic analysis of 18S rRNA

The neighbor joining phylogenic tree and the maximum likelihood tree were constructed for the 18S rRNA gene sequences (Figs. 3, 4). The lengths of the sequences analyzed in this study ranged from 1,673 to 1,728 bp. The tree resulting from the neighbor joining analysis was identical in overall topology to the tree obtained from the maximum likelihood analysis. Cryptosporidium species examined in this study formed two clades. This result was in agreement with that reported in a previous study (Xiao et al. 1999). One clade contained the gastric Cryptosporidium parasites, while the other contained C. baileyi and intestinal parasites. In phylogenetic analysis of the 18S rRNA gene, the Cryptosporidium field isolate belonged to the gastric clade and clustered with parasites isolated from a bactrian camel, a rock hyrax, and a house rat.

Fig. 3
figure 3

Phylogenetic relationships (neighbor joining tree based on Kimura’s 2-parameter analysis) between the Cryptosporidium field isolate and the Cryptosporidium species inferred from the 18S rRNA sequences. The tree was rooted with the 18S rRNA sequence from Eimeria tenella (AF26388), and the root was removed. Bootstrap values (in percentage) above 50 from 1,000 pseudo-replicates are shown for both the neighbor joining (the first value) and maximum parsimony analyses (the second value)

Full size image
Fig. 4
figure 4

Phylogenetic relationships (maximum likelihood tree) between the Cryptosporidium field isolate and the Cryptosporidium species inferred from the 18S rRNA sequences. The tree was rooted with the 18S rRNA sequence from Eimeria tenella (AF26388), and the root was removed. Bootstrap proportions (in percentage) at each node were estimated using the RELL method

Full size image

Discussion

We isolated Cryptosporidium oocysts from the large Japanese field mouse, A. speciosus. The large Japanese field mice were captured on the farm where C. andersoni had been constantly detected from several beef cattle; however, the infectivity to some mice and the DNA sequence of the 18S rRNA gene of the field isolate markedly differed from those of the C. andersoni Kawatabi isolate. Further, the oocyst size and the infectivity of the field isolate to ICR mice slightly differed from those of C. muris RN66.

A large number of developing stages of field isolates was observed only in the gastric gland of the stomach of large Japanese field mice and SCID mice. Recently, C. muris, C. andersoni, and C. serpentis have been identified as gastric Cryptosporidium parasites. The infectivity of C. muris RN66 (Iseki 1986) and C. andersoni Kawatabi isolate (Satoh et al. 2003) to SCID mice has been confirmed. Although the infectivity of the field isolate to ICR mice and large Japanese field mice differed slightly from that of C. muris and C. andersoni in the experimental infection, the differences may not be significant. Iseki et al. (1989) have reported that C. muris RN66 are infective to 3-week-old ICR mice inoculated with 1×106 oocysts. We consider the difference in the results between the previous study and the present study to be due to the number of inoculation oocysts and the age of the mice. Oocysts isolated from cattle in the US, which were previously referred to as C. muris-like oocysts and which may be C. andersoni, did not infect mice or even cattle (Lindsay et al. 2000). Oocysts similar to those obtained from cattle in the US were not transmissible to neonatal or adult BALB/c mice and SCID mice (Koudela et al. 1998). However, several successful transmissions of bovine-derived large-type oocysts, referred to as C. andersoni, to SCID mice have been reported in Japan (Kaneta and Nakai 1998; Satoh et al. 2003; Matsubayashi et al. 2004a; Koyama et al. 2005). In the present study, we also confirmed the ability of the C. andersoni Kawatabi strain to infect SCID mice.

In phylogenetic analysis of the 18S rRNA gene, the Cryptosporidium field isolate belonged to the gastric clade and was located in the C. muris cluster. At the 18S rRNA locus, the similarity between this isolate and C. muris RN66 and between this isolate and the C. andersoni Kawatabi isolate was 99.5 and 99.1%, respectively. A previous study showed that the genetic similarity between C. hominis and C. parvum, which are considered to be distinct species, is 99.7% (Morgan et al. 2002). Toxoplasma gondii and Neospora caninum are classified into different genera, but there is 99.8% similarity at this locus. Furthermore, the similarity between C. muris and C. andersoni is 99.1% at the 18S rRNA locus (Lindsay et al. 2000). However, they are considered to be valid species based on the differences in their infectivity to laboratory mice and differences in their DNA sequences. The percentage similarity between this isolate and C. muris at this locus suggests that the field isolate and C. muris were different species. Further, high bootstrap values (98 and 95%) obtained by parsimony and neighbor joining analysis suggested that the sequences of C. muris and the field isolate clearly branched.

Based on the characteristics of the oocysts, the parasitized organ, the host specificity, and the DNA sequence in the 18S rRNA gene, the field isolate from large Japanese field mice was considered to be a distinct species from C. muris. However, biological differences between this isolate and C. muris were unclear, and this isolate clustered with C. muris in the phylogenetic analysis. Therefore, evidence for classifying this isolate as a novel species was insufficient. Thus, we proposed that this isolate is a novel genotype of C. muris and denoted it as C. muris Japanese field mouse genotype.

Recently, there have been many reports on new species and genotypes of Cryptosporidium on the base of molecular studies. C. parvum has several genotypes, for instance, mouse, rabbit, monkey, and bovine. C. canis has three genotypes, namely, coyote, dog, and fox (Xiao et al. 2002). In contrast to the small oocyst types of Cryptosporidium, there are few reports on the genotypes present in the large oocyst type of Cryptosporidium (Ryan et al. 2003). Our study is the first report of a novel genotype of C. muris.

Phylogenetic analysis has been used to study the diversity of Cryptosporidium spp. from humans, and wild and domestic animals (Hajdušek et al. 2004). C. parvum was found in a raccoon dog in Japan (Matsubayashi et al. 2004b). In this study, a novel Cryptosporidium genotype was isolated from large Japanese field mice, the most common wild mouse found in forests in Japan. Different Cryptosporidium species and genotypes may be distributed in various wild animals. Since human infection of C. muris has been reported (Katsumata et al. 2000; Gatei et al. 2002) and it is a potential zoonotic parasite, further epidemiological investigation and molecular and biological analysis of the C. muris group should be performed.

The taxonomic status of this isolate and C. muris should be clarified by finding closely related species of C. muris and accumulation of biological and genetic data of these strains.