Microsporidiosis is of increasing concern because Microsporidia can infect virtually all animals (Santín and Fayer, 2011; Abu-Akkada et al., 2015; Santin and Fayer, 2015; Zhang et al., 2016a; Zhao et al., 2016). Fecal-oral routes, such as ingestion of contaminated water and food (Zhao et al., 2014a) are the major route to transmit Microsporidia which consist of 1300 named species (Karim et al., 2014c). Of these, Enterocytozoon bieneusi is considered as the most important species and responsible for more than 90% of human microsporidiosis (Desportes et al., 1985). Although E. bieneusi was firstly detected in HIV patients in 1985 (Desportes et al., 1985), as research continues, more and more animals were also considered as susceptible hosts for E. bieneusi. To date, more than 200 distinct genotypes have been reported on the basis of sequences of the internal transcribed spacer (ITS) region (Karim et al., 2014b; Tian et al., 2015). These genotypes were divided into several groups: the zoonotic groups (syn. Group 1), host-adapted groups (groups 2–5 and an outlier genotypes in dogs), and some other small groups (groups 6–9). However, surprisingly, some of the genotypes in Group 2 (the so-called host-adapted group) were found in both animals and humans, which should also be considered as a zoonotic agent (Hu et al., 2014; Karim et al., 2014a; Ma et al., 2015a,b).

To our knowledge, limited data has been published on genotypes of E. bieneusi from China (Tian et al., 2015; Zhang et al., 2015, 2016a,b; Qi et al., 2016) and no data is available for donkeys. The objectives of the present study were to investigate the E. bieneusi prevalence and identify their genotypes in donkeys in Jilin, Liaoning and Shandong Provinces, eastern and northeastern China.

A total of 16 (5.3%) out of 301 donkeys were PCR-positive for E. bieneusi (Table 1). Donkeys in Shandong Province (11/224, 4.9%) showed the higher E. bieneusi prevalence (Table 1). E. bieneusi prevalence in different season groups ranged from 4.2 to 6.5% (Table 1). Prevalence of E. bieneusi in young donkeys and adult donkeys was 4.2 and 5.5%, respectively (Table 1). Moreover, E. bieneusi prevalence in different farm groups varied from 4.2 to 6.9% (Table 2).

In this study, the overall prevalence was 5.3% (16/301) which was lower than that in horses from the Czech Republic (6.9%) (Wagnerová et al., 2012), Colombia (10.8%) (Santín et al., 2010), and Algeria (6.8%). However, it is higher than that in donkeys in Algeria (1.6%) (Laatamna et al., 2015), and Spain (0%) (Lores et al., 2002). Moreover, it is also considerably higher than that in horses in Spain (0%) (Lores et al., 2002). The different infection rates may be due to the different susceptibility between donkeys and horses. Moreover, animal welfare, sample sizes, age distribution of the samples and geo-ecological conditions can also influence the E. bieneusi prevalence.

In the present study, although donkeys from Liaoning Province has relatively high rates of E. bieneusi infection compared with donkeys from Jilin Province and Shandong Province, the difference was not statistically significant (χ² = 0.30, df = 2, P = 0.86) (Table 1). This may be due to the similar climates during sampling times in Jilin, Liaoning and Shandong Provinces, and also relate to the same management mode in these farms. Moreover, the present study also showed that E. bieneusi prevalence in donkeys increased gradually with age, which is consistent with previous reports in that E. bieneusi may accumulate throughout the life time (Thellier and Breton, 2008; Zhao et al., 2014a), but the difference was not statistically significant (χ² = 0.15, df = 1, P = 0.70) (Table 1).

To our knowledge, although two studies of E. bieneusi prevalence in donkeys have been reported previously, these isolates were not genotyped successfully (Lores et al., 2002; Laatamna et al., 2015). Previous studies have identified sixteen E. bieneusi ITS genotypes, namely Horse 1 – Horse 11, G, WL15, D, EpbA and CZ3 in horses who is a close relative to donkeys (Santín et al., 2010; Wagnerová et al., 2012; Laatamna et al., 2015). But probably due to the smaller sample sizes and different collection times, only four E. bieneusi genotypes were observed in donkeys in the present study, including two known genotypes J and D, and two novel genotypes NCD-1 and NCD-2 (Figure 1). Thus, four E. bieneusi genotypes were endemic in donkeys in eastern and northeastern China. This study also indicated that genotype J was the most prevalent in donkeys, which were different from the results reported in horses in Colombia (Santín et al., 2010) and Czech Republic (Wagnerová et al., 2012), where the predominance of genotype D was found, and in horses in Algeria, where predominance of horse1 was identified (Laatamna et al., 2015). Furthermore, a total of 10 polymorphic sites were observed among these genotypes (Table 3), which suggested the high genetic diversity of E. bieneusi in the investigated donkeys.

Genotypes J and D have been widely identified in different hosts in northern and eastern China. For example, genotype D was identified in golden takins (Budorcas taxicolor bedfordi) in Shannxi (Karim et al., 2014a), non-human primates in Henan (Wang et al., 2013; Karim et al., 2014b), dairy cattle, sheep, goats, pig, raccoon dog, cats, blue foxes and dogs in Heilongjiang (Zhao et al., 2014a,b, 2015a,b; Li et al., 2015), and blue foxes in Jilin (Zhao et al., 2015c); genotype J was found in cattle in Heilongjiang, Jilin, Shandong, Henan, captive wildlife in Henan, yaks in Qinghai. In addition, genotypes D was also identified in HIV patients in Henan (Wang et al., 2013; Karim et al., 2014b) and J was found in a child in Jilin (Zhang et al., 2011). Moreover, two of the ITS sequences (accession numbers: KU557671 and KU557672) of the identified E. bieneusi isolates were identical to that of genotypes J (GenBank accession no. KP735178, from a Homo sapiens in Iran) and D (GenBank accession no. KP262379, from a goat in China) sequences available in GenBank, respectively. These results indicate cross-species transmission of these E. bieneusi genotypes in northern China.

Phylogenetic analysis indicated that genotypes D, NCD-1 and NCD-2 belonged to group 1, the most important zoonotic groups; J was grouped into group 2, a cattle-specific groups, but genotype J has also been identified in humans. These results suggest that donkeys are potential source of animal and human microsporidiosis.

This is the first study of E. bieneusi prevalence (5.3%, 16/301) in donkeys in China. Moreover, two known genotypes (genotypes D and J), and two novel genotypes (NCD-1 and NCD-2) were detected for the first time in donkeys. Donkeys should be considered as an important potential source of human microsporidiosis, and effective strategies should be performed to control E. bieneusi infection in donkeys, other animals and humans.

Representative nucleotide sequences were deposited in GenBank with the following accession numbers: KU557670-KU557672 and KY250434.

X-QZ and X-XZ conceived and designed the study, and critically revised the manuscript. D-MY, J-GM, and W-BZ performed the experiments. X-XZ and D-MY analyzed the data. D-MY and J-GM drafted the manuscript. F-CL, J-LH and QZ helped in study design, study implementation and manuscript preparation. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.