A Cross-Sectional Study on the Occurrence of the Intestinal Protist, Dientamoeba fragilis, in the Gut-Healthy Volunteers and Their Animals

Dientamoeba fragilis is a cosmopolitan intestinal protist colonizing the human gut with varying prevalence depending on the cohort studied and the diagnostic methods used. Its role in human health remains unclear mainly due to the very sporadic number of cross-sectional studies in gut-healthy populations. The main objective of this study was to expand knowledge of the epidemiology of D. fragilis in gut-healthy humans and their animals. A total of 296 stool samples from humans and 135 samples from 18 animal species were analyzed. Using qPCR, a prevalence of 24% was found in humans in contrast to conventional PCR (7%). In humans, several factors were found to influence the prevalence of D. fragilis. A more frequent occurrence of D. fragilis was associated with living in a village, traveling outside Europe and contact with farm animals. In addition, co-infection with Blastocystis spp. was observed in nearly half of the colonized humans. In animals, D. fragilis was detected in 13% of samples from eight species using qPCR. Our molecular phylogenies demonstrate a more frequent occurrence of Genotype 1 in gut-healthy humans and also revealed a likely a new protist species/lineage in rabbits related to D. fragilis and other related organisms.


Introduction
Dientamoeba fragilis is a human intestinal protist with cosmopolitan distribution and prevalence ranging from 0.2% to 82% (reviewed in [1][2][3]). Most studies addressing its occurrence are from high-income countries and surprisingly, less is known from lowincome countries [1,2,4]. This protist was first described more than a century ago and was originally classified as a binucleate intestinal amoeba due to the absence of flagella (summarized in [5]). However, later morphological and molecular-phylogenetic analyzes revealed that D. fragilis is related to the flagellating trichomonads [2,5]. Nevertheless, some aspects of the life cycle of D. fragilis are still unclear. While colonization of the host hindgut by the trophozoite stage has been described in detail [2], the question of the mode of human-to-human transmission remains open. One hypothesis assumes involvement of the helminth vector with Enterobius vermicularis being the main candidate [6,7]. This is supported by observations such as (i) the colonization of a volunteer with D. fragilis after ingestion of E. vermicularis eggs derived from a donor colonized with both parasites [8], (ii) the molecular detection of D. fragilis DNA in E. vermicularis eggs [6,7], and (iii) also and gender. In addition, using the data from our previous study, Lhotská et al. [47], we were able to investigate the frequency of co-infection of D. fragilis and Blastocystis spp.

Prevalence of Dientamoeba fragilis Based on Two Molecular Approaches
Conventional PCR (cPCR) detected D. fragilis in 32 samples (31 from humans and one from animal) out of total 431 samples. However, Sanger sequencing confirmed its presence in only 23 samples, 22 of which were from humans and one from the animal host (rabbit). Only these verified samples were considered positive based on cPCR (Table 1). According to this, the overall prevalence was 7% (22/296) in humans and 0.7% in nonhuman hosts (1/135). We were unable to obtain sequences from the remaining nine samples, so they were classified as negative.
Real-time PCR (qPCR) detected D. fragilis in 72 human samples with a prevalence of 24% (72/296) and in 15 samples from animals (11%; 15/135). Positive animals included dogs (Canis lupus familiaris), cat (Felis silvestris catus), guinea pig (Cavia aperea porcellus), rabbits (Oryctolagus cuniculus domesticus), goat (Capra aegagrus hircus), sheep (Ovis aries), and horses (Equus caballus) (see Table 1 for details). All 23 positive samples detected by cPCR (confirmed by Sanger sequencing) were also positive in qPCR (see Table 3 for details). In addition, we verified the presence of D. fragilis in the qPCR amplicons of all qPCR-positive samples by Sanger sequencing. Because we did not have a culture of D. fragilis to generate a quantification curve in qPCR, only Ct values are shown for each positive sample (Table 3).

Comparison of the Sensitivity of cPCR and qPCR for Dientamoeba fragilis Detection
We compared the sensitivity of these two methods on a set of human samples only (N = 296). The McNemar test showed that qPCR was a significantly more sensitive method for diagnostics of D. fragilis (p < 0.001; χ 2 = 49; Supplementary Data S1). In addition, qPCR was also more successful in detecting D. fragilis in animal samples (Table 3), but these were not included in the statistical analyses.

Influence of the Specific Factors on the Occurrence of Dientamoeba fragilis
An effect of the specific factors on the distribution on D. fragilis was tested on the set of human samples only (N = 296) and statistically evaluated (for details see Section 4.7).

Lifestyle
All 296 human samples were tested for the effect of lifestyle on D. fragilis colonization. Two factors were selected for this investigation: (i) urban life (town/city) versus village life, and (ii) impact of traveling (Table 4). For the latter factor, we distinguished three categories of individuals: (a) travelers outside Europe (104/296), (b) travelers only in Europe (104/296), and (c) non-travelers (52/296; Table 4). Regarding the living locality, a higher prevalence was observed in individuals living in village areas (29%, 24/83), compared to individuals from urban areas (23%, 48/213; Table 4), but this difference was not statistically significant. We observed a higher prevalence of D. fragilis in travelers. The prevalence among travelers within Europe was 24% (34/140), while 29% (30/104) was in travelers outside Europe (Table 4). However, the prevalence in non-travelers was significantly lower (15%, 8/52; p = 0.0324; Table 4).  The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm village 83/296 29% The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm

Contact with Animals
The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm

Contact with Animals
The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm

Contact with Animals
The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm

Contact with Animals
The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm

Contact with Animals
The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm in contact with pets

Contact with Animals
The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm

Contact with Animals
The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm

Contact with Animals
The overall prevalence of D. fragilis in the group of volunteers who reported contact with animals was 25% (62/252; Table 4). Among those without animal contact, the prevalence was 23% (10/44, Table 4). Based on the information from questionnaires, we divided the group of subjects with animal contact into (i) those with contact with pets (180/252) and (ii) those with contact with farm animals (72/252; Table 4). We found a significant difference in the incidence of D. fragilis between these two groups (p = 0.004; Table 4), specifically between 36% prevalence (26/72) in the group with contact with farm animals and 20% prevalence (36/180) in the group with contact with pets.

Occurrence of Dientamoeba fragilis in the Family Environment
We obtained samples from several family members from a total of 69 families. Colonization of D. fragilis was detected in 34 families (i.e., 49%; Table 5). In addition, D. fragilis was detected in some families whose members were in close contact with a pet or farm animal that was positive for D. fragilis in qPCR (i.e., rabbits, cat, horses, dogs, and small ruminants; Table 5). Table 5. Overview of families including individuals colonized by Dientamoeba fragilis as well as D. fragilis-positive animals with which these families were in close or frequent contact (N family members-total number of family members; N positive members-number of family members colonized with D. fragilis; contact animal-animals colonized by D. fragilis which were in close contact with family members).

Genetic Diversity of Dientamoeba fragilis Based on Molecular Phylogenies
For the phylogenetic analyses, a total of twenty partial sequences of the SSU rDNA of D. fragilis were examined, specifically 19 sequences from colonized humans and one from a rabbit (B174). In addition, the Trichomonas sp. sequence (B131) derived from a chicken was included in the final dataset. This dataset included ten contigs (B9, B13, B126, B224, B238, B288, B329, B338, B174, B131), eight forward sequences (B113, B326, B328, B341, B371, B393, B413, B418), and three reverse sequences (B24, B82, B345). All our sequences were cloned to avoid unclear multiple signals or chimeras. The DNA sequences obtained in this study were deposited in the GenBank TM under accession numbers OP375680-OP375701.
The maximum likelihood tree (TVMe+G4+F, IQ-TREE) ( Figure 1) shows a large clade containing D. fragilis-, H. meleagridisand P. wenrichi-subclades, and then two trichomonad clades, one of which contains our sequence of Trichomonas sp. from chicken (B131/OP375701). The D. fragilis subclade includes our human sequences (OP375679-99), sequences defined as Genotype 1 from GenBank TM , and also contains only one Genotype 2 sequence (U37461) that is currently available in the GenBank database and it is represented by a separate branch (Figure 1). All our sequences from humans were unambiguously assigned to the Genotype 1 group, which has some internal polytomy (identical sequences) and three stable subgroups (one containing sequences OP375681, OP375694; the second containing sequences OP375686, OP375688; and the third with OP375685, OP375699). A very interesting result is the phylogenetic position of the sequence from rabbit (B174/OP375700). Although this sequence has a high identity with D. fragilis and H. meleagridis in the BlastNt database (concordance with D. fragilis: query coverage-70-76%, e-value-0.0, percent identity-89-92%; concordance with H. meleagridis: query coverage-75%, e-value-0.0, percent identity-89%), it forms a separate, strongly supported branch representing a new lineage of organisms closely related to D. fragilis, H. meleagridis and P. wenrichi. 2 sequence (U37461) that is currently available in the GenBank database and it is represented by a separate branch (Figure 1). All our sequences from humans were unambiguously assigned to the Genotype 1 group, which has some internal polytomy (identical sequences) and three stable subgroups (one containing sequences OP375681, OP375694; the second containing sequences OP375686, OP375688; and the third with OP375685, OP375699). A very interesting result is the phylogenetic position of the sequence from rabbit (B174/OP375700). Although this sequence has a high identity with D. fragilis and H. meleagridis in the BlastNt database (concordance with D. fragilis: query coverage-70-76%, e-value-0.0, percent identity-89-92%; concordance with H. meleagridis: query coverage -75%, e-value-0.0, percent identity-89%), it forms a separate, strongly supported branch representing a new lineage of organisms closely related to D. fragilis, H. meleagridis and P. wenrichi.

Discussion
To our knowledge, this is the first cross-sectional study addressing the occurrence of Dientamoeba fragilis in the Czech Republic in a set of stool and fecal samples obtained from gut-healthy volunteers and their animals of different species (see Table 1 for details). In humans, using real-time diagnostics (qPCR), we found an overall prevalence of 24% in a group of 296 humans. We also investigated the difference in sensitivity for detection of D. fragilis between qPCR and conventional PCR (cPCR) in human samples. Real-time PCR identified 50 more positive samples than cPCR, making it more sensitive for the detection of D. fragilis. The prevalence based on cPCR was 7%. Our observation is consistent with other studies comparing the sensitivity of molecular methods for the detection of other intestinal protists (e.g., [48]). In general, the diagnostics of D. fragilis are challenging and the sensitivity of different diagnostic methods can vary widely [49]. For example, Sharzhanov et al. [50] showed that the sensitivity of traditional microscopic methods for the detection of D. fragilis is lower compared to qPCR. Thus, qPCR appears to be the gold standard for D. fragilis diagnostics, and some recent studies already use this method routinely (reviewed in [1]).
Current data on the occurrence and prevalence of D. fragilis in asymptomatic individuals (i.e., subjects without gastrointestinal symptoms or disease) are generally very sporadic and come primarily from control groups of healthy subjects included in studies that focused on patients with various gastrointestinal problems (acute or chronic diarrhea, Inflammatory Bowel disease-IBD, or Inflammatory Bowel Syndrome-IBS). Because epidemiologic studies of D. fragilis use a variety of diagnostic approaches that may differ in specificity and sensitivity [1], we compared our data only with studies using qPCR.
The available literature shows that the prevalence of D. fragilis in control groups of healthy individuals ranges from 36% to 71% [25,27,29,30,51,52]. For example, Heusinkveld et al. [52] conducted a large-scale cross-sectional survey to investigate the association between the occurrence of a number of enteropathogens (including selected protists) and acute gastroenteritis, and included more than 8500 pairs of a preschool-aged child and a parent living in the same household. They found that the prevalence of D. fragilis ranged from 36% to 39% in individuals without acute intestinal problems. Our results show a prevalence of 24% in gut-healthy people in different age groups, which is slightly lower than in the above-mentioned studies. Interestingly, a ten-year study conducted at the University Hospital of Parma demonstrated an almost comparable prevalence of D. fragilis (i.e., 20%) in patients with gastrointestinal symptoms [3]. Following data from patient cohorts and healthy subjects, the incidence of D. fragilis seems to be more or less comparable. Therefore, it is not possible to decide whether it is rather a pathogenic or a commensal organism. Determination of the so-called fecal protist load [48] could shed more light on deciphering the role of D. fragilis in human health. Such quantitative data could subsequently be compared between asymptomatic and symptomatic individuals, and in patients with IBD. Unfortunately, we were not able to quantify the fecal Dientamoeba load in our study because we did not have a monoxenic culture of D. fragilis to construct the quantification curve as was done in case of Blastocystis spp. in the study by Šloufová et al. [48].
It is important to mention that the prevalence of D. fragilis may be influenced by various factors, for example, age group selection, geographic location, or socioeconomic status [1,2]. The prevalence observed in our study may be affected by the wide age range in the group of volunteers because it ranged from 0 months to >60 years.
Our set of human samples allowed us to observe the frequency of co-infections of D. fragilis with another intestinal protist Blastocystis spp. which was the subject of our previous study [47]. Based on the results of both studies, we found that almost half of the samples (i.e., 46%) from individuals colonized with D. fragilis had co-infection with Blastocystis spp. Interestingly, these cases of co-infections were associated with factors such as frequent contact with animals and traveling. Co-infection with these protists was also recorded in other studies that focused primarily on patients with gastrointestinal disease or symptoms, and the rate of co-infection rate varied widely. While several studies found co-infection rates ranging from 17.6% to 33.6% in patients with gastrointestinal symptoms [17,21,50,53], Yakoob et al. [20] detected co-infections in only 3% of IBS patients but found nothing comparable in a group of healthy subjects. In contrast, our results show a higher rate of co-infection in gut-healthy volunteers. Similarly, a higher co-infection rate of D. fragilis with Blastocystis spp. (i.e., 66.2%) was demonstrated in mostly asymptomatic children [54]. The authors of the latter study hypothesize that co-infection of D. fragilis and Blastocystis spp. is associated with healthy gut microbiome, as was the case in the study by de Boer et al. [26]. In addition, our data are unique because we were able to assess the presence D. fragilis and Blastocystis subtypes. Dientamoeba was most frequently detected in co-infection with ST1, ST2, ST3, and ST4 subtypes.
As far as we know, only one larger epidemiological study has been conducted for D. fragilis, covering a wider range of animal species [38]. However, D. fragilis was detected in only two fecal samples out of 420 samples and in two animal species (dog and cat) out of a total of 37 species. In our study, we found 13% of animals in close contact with human volunteers (135 samples obtained from 18 species) that were positive for D. fragilis using qPCR method. Dientamoeba was detected in eight animal species, several of which have already been described as potential hosts, specifically dogs (4/55), cat (1/19), and sheep (1/6). In dogs and cats, D. fragilis has been previously detected by PCR, whereas in sheep it was detected by traditional coproscopic methods [38,40,55]. In addition, we found D. fragilis in rabbits (5/13), horses (2/15), goat (1/4), and guinea pig (1/2). Surprisingly, we found no evidence of D. fragilis presence in domestic pigs, although a high prevalence (44-63%) has been recorded in the past [32,33].
Regarding the specific factors that might affect the distribution of D. fragilis in the gut-healthy human population, we examined the influence of living locality (urban versus rural), contact with animals, traveling (within and outside Europe, no traveling), age, and gender. Our results showed that prevalence was higher in people living in villages, and in travelers, especially outside Europe. Traveling is also considered an important predisposing factor for D. fragilis colonization in several other studies that focused on patients with diarrhea [15,16] or in a cross-sectional study in children under six years of age [29]. Interestingly, in a study by Norberg et al. [15], there was recorded a trend toward a higher incidence of D. fragilis in more than half of the patients with a history of traveling, and almost half of them had traveled outside Europe.
Not much is known about the effect of living locality; only Heusinkveld et al. [52] found a correlation between more frequent occurrence of D. fragilis and less urbanized areas. They also described a correlation between the occurrence of D. fragilis and the ownership of dogs. In contrast, Jokelainen et al. [29] concluded that contact with animals, especially those species considered potential hosts, had no effect on the spread of D. fragilis. Similarly, in our study, we found no significant effect of general contact with animals on the spread of D. fragilis in humans, but we showed a possible effect of keeping some farm animals, as more than one third of their breeders were colonized. Unfortunately, our data do not allow us to assess zoonotic transmission because we confirmed the presence of D. fragilis in animals only by qPCR without any subsequent sequence information (detailed information further). So far, the only case of possible zoonotic transmission was described in the study by Cacciò et al. [32], in which Genotype 1 was confirmed in pigs and their keepers based on molecular analyses of partial SSU-and 5.8S-rDNA sequences within one farm.
Age is generally assumed to have a significant impact on the distribution of D. fragilis in humans [1], but this has not always been confirmed [4,30]. In general, a higher prevalence is thought to occur in children up to about 15 years of age [29,51,56]. Although our results showed a more frequent occurrence of D. fragilis in the group of children younger than 12 years, these differences were not statistically significant, as in other studies [17,49,57]. Here, this trend could be due to an imbalanced number of samples collected from children under 18 years of age. The occurrence of D. fragilis in the other age groups was very similar. Regarding gender, we did not observe a difference in the incidence of D. fragilis between men and women, as was the case in the recent study by Sarzhanov et al. [50].
Nevertheless, findings of higher D. fragilis prevalence in women of childbearing age are frequently described [4,58].
A total of 69 families provided us with samples from all or several of their members so that we could track the occurrence of D. fragilis within the family. In almost half of the families (49%), the presence of D. fragilis was detected in at least one member, but usually in several members. Our results suggest that D. fragilis appears to circulate among family members or that positive family members may have been colonized from the same source. In six families, we also detected the presence of D. fragilis in their animals. However, based on our data, we were unable confirm zoonotic transmission.
The genetic diversity of D. fragilis appears to be low, and so far, only two genotypes have been proposed based on the variability of three genes (SSU rRNA, EF1, actin) [16,43,59,60]. However, most studies dealing with molecular or phylogenetic identification of D. fragilis isolates describe a predominant occurrence of Genotype 1 in different geographical areas, in different human cohorts (symptomatic versus asymptomatic individuals) and in animals [16,24,40,41]. Our nineteen SSU rDNA sequences of D. fragilis obtained from human samples also belong to Genotype 1 according to our phylogenetic analyses, and the sequences appeared in the phylogenetic tree together with Genotype 1 sequences from GenBank. Nevertheless, this Genotype 1 group exhibits some small genetic variability. In addition to the identical sequences forming a polytomy, we identified three highly supported subgroups. Analysis of the entire SSU rDNA sequence or another gene-as suggested by Simone M. Cacciò [1]-would perhaps provide more insight into the genetic diversity of D. fragilis.
In contrast to the human isolates, we were not able to obtain SSU rDNA sequences of D. fragilis by cPCR from most of the positive animals (with the exception of one sequence from a rabbit), apparently due to the low intensity of colonization, considering the Ct values obtained in qPCR (for details, see Table 3). Therefore, the presence of D. fragilis in these samples was confirmed only by qPCR followed by sequencing of very short qPCR amplicons (∼97 bp). Therefore, we were not able to determine whether it was indeed D. fragilis. To determine the genotypes of D. fragilis in animals, the partial SSU rDNA sequences obtained by cPCR (∼850 bp) are required. So far, only the occurrence of Genotype 1 has been described in pigs, cattle and budgerigars [40,41,44].
The fact that these 850 bp-SSU rDNA sequences are absolutely essential for the definitive genetic identification of D. fragilis in animals is also underlined by our finding of an apparently new protist in a rabbit, that appears to be closely related to D. fragilis. This SSU rDNA sequence was assigned to D. fragilis using the BlastNt database, but our phylogenetic analyzes indicated that it is the strongly supported separate branch next to D. fragilis, Histomonas meleagridis, and Parahistomonas wenrichi within one clade. These results suggest that it may be a new lineage or genus of a related protist that has not been described previously. However, further studies focusing on detailed morphological and phylogeneticmolecular identification is required for a taxonomic description. Based on all these facts, there is also a need to optimize or develop a new diagnostic qPCR protocol that is highly specific for D. fragilis.

Sample Collection and Ethical Approval
The present study was conducted between 2017 and 2019 in the Czech Republic and follows the previous study on Blastocystis spp. [47] with identical sampling strategy and sample dataset. Briefly, stool samples were collected from healthy volunteers, none of whom had gastrointestinal symptoms or intestinal inflammatory diseases (inflammatory Bowel Disease-IBD, or Inflammatory Bowel Syndrome-IBS) at the time of sampling. To investigate possible zoonotic transmission, fecal samples were also collected from animals, with which the volunteers were in a close contact. All participants completed a questionnaire that included information on living locality (city/village), traveling (nontravel/travel in Europe/travel outside Europe), contact with animals (no contact/contact with pets only/contact with pets and livestock), gender and age (age categories: 0-3, 4-6, 7-12, 13-17, 18-30, 31-49, 50-60, >60) (for more details see [47]).
Each participant signed an informed consent declaration to participate in the study. The procedure, conditions, and ethical rules of this study are in accordance with the Declaration of Helsinki 2013 (World Medical Association). All data were strictly anonymized and processed in accordance with applicable laws of the Czech Republic (e.g., Act No 101/2000 Coll and subsequent regulations). The study was approved by the Ethics Committee of the Biology Centre of the Czech Academy of Sciences (reference number: 1/2017).

DNA Extraction and Samples Processing
Total DNA was extracted directly from stool/fecal samples using the commercial kit PSP Spin Stool DNA Kit (Stratec, Germany) according to the manufacturer's protocol. All collected samples were subjected to two molecular diagnostic procedures such as conventional PCR (cPCR) [16] and real-time PCR (qPCR) [61]. As a positive control, we used DNA from the xenic culture of D. fragilis. This DNA was provided to us by our colleague Prof. J. Kulda from Charles University (Prague, Czech Republic).

Cloning and Sequencing of Positive Samples
Sequences were always obtained from positive samples using cloning using pGEM ® -T Easy Vector System I (Promega, Madison, WI, USA) followed by Sanger sequencing by a commercial company (Eurofins GATC Biotech, Konstanz, Germany). PCR amplicons of the appropriate size were purified using the GenElute TM Gel Extraction Kit (Sigma-Aldrich, St. Louis, MO, USA) and sequenced in both directions using PCR primers. All sequences were then processed using the Geneious Prime 2019.0.4. software and subsequently compared with sequences in the GenBank TM database (National Centre for Biotechnology Information) using BLASTn.

Phylogenetic Analysis
The newly obtained SSU rRNA gene sequences were aligned to the available homologs from GenBank using the Clustal Omega algorithm in SeaView [62]. Alignment was manually edited in BioEdit [63] removing gaps, ambiguously aligned and hypervariable positions. Maximum likelihood trees were constructed using a TVMe+G4+F evolutionary model (selected by ModelFinder [64]) implemented in IQ-TREE [65]. Branch support was estimated from 10,000 UFBoot (ultrafast bootstrap approximation) [66] using the above software and conditions.

Statistical Analysis
Difference in sensitivity of cPCR and qPCR for detection of D. fragilis was evaluated using the McNemar test with Yates's correction. We employed Fisher's Exact Test to examine differences between age categories in the incidence of D. fragilis in humans detected by qPCR. In addition, we used a generalized linear model to examine the effects of age as a continuous variable. The chi-square test was used to statistically evaluate the significance of the difference in D. fragilis prevalence between groups of humans within a single observed factor, such as lifestyle (urban versus rural), traveling (within and out of Europe), contact with animals (with and without), and gender. These analyzes were performed using GraphPad Prism 5.0 software and R Studio version 3.6.2. A p-value of less than 0.05 was considered to indicate statistical significance.

Conclusions
The prevalence of D. fragilis found in our study in the gut-healthy individuals suggests that the mere presence of this protist should not generally lead to efforts to eliminate it from the intestine, especially in asymptomatic individuals. This fact has been also considered for other intestinal protists, for example Blastocystis spp. [67]. Moreover, it is very difficult to decide whether D. fragilis is more of a commensal or a pathogen, since the frequency of its occurrence in patients with gastrointestinal symptoms is more or less comparable to healthy cohorts. It seems that the information about the presence of D. fragilis is not sufficient to assess its role in human health and that a quantitative indicator, the so-called fecal protist load, should be introduced into its diagnostics [48]. Such quantitative data on D. fragilis colonization and their comparison between different cohorts could facilitate understanding of its role in human health and disease. In addition, knowledge about D. fragilis in gut healthy individuals in high-income countries and in the human populations with traditional lifestyles needs to be expanded, including also knowledge about correlations with intestinal bacterial diversity. Worth mentioning is recent evidence of regulatory capabilities of some intestinal eukaryotes, including protists, which may be critical to the host intestinal ecosystem and immune system [68][69][70]. Dientamoeba fragilis may have a similar regulatory function, either alone or in co-infection with Blastocystis spp. but this remains to be tested experimentally. Finally, it is important to emphasize that to understand the zoonotic transmission of D. fragilis, it is necessary to monitor its occurrence in a range of animal species and also to decipher its genetic diversity.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
Publicly available datasets were analyzed in this study. This data can be found in the GenBank TM database under these accession numbers [OP375679-OP375701].