Novel genotypes of Hepatozoon spp. in small mammals, Brazil

Background Small mammals (rodents and marsupials) have been poorly explored for the occurrence of apicomplexan (genus Hepatozoon and genera of the order Piroplasmorida) and Anaplasmataceae agents in Brazil. Thus, this study investigated the occurrence of Hepatozoon spp., Piroplasmorida, and Anaplasmataceae agents in small mammals in seven forest fragments in Brazil. Methods During 2015–2018, small mammals were captured in six forest fragments in the State of São Paulo (Cerrado and Atlantic Forest biomes) and one fragment in the State of Mato Grosso do Sul (Pantanal biome). Mammal blood, liver, spleen, and lung samples were tested molecularly for the presence of DNA of Hepatozoon, Piroplasmorida, and Anaplasmataceae agents. Results A total of 524 mammals were captured, comprising seven species of marsupials, 14 rodents, two carnivores, and one Cingulata. Four novel haplotypes (1, 2, 3, 4) of Hepatozoon spp. were detected in small mammals from different biomes. In São Paulo state, haplotype 1 was detected in rodents from Cerrado and a transition area of Cerrado and Atlantic Forest biomes, whereas haplotype 2 was detected in rodents from the Atlantic Forest biome. On the other hand, haplotypes 3 and 4 were restricted to rodents and marsupials, respectively, from the Pantanal biome of Mato Grosso do Sul. No host species shared more than one haplotype. Despite these distinct geographical and host associations, our phylogenetic analyses indicated that the four Hepatozoon haplotypes belonged to the same clade that contained nearly all haplotypes previously reported on rodents and marsupials, in addition to several reptile-associated haplotypes from different parts of the world. No mammal samples yielded detectable DNA of Piroplasmorida agents. On the other hand, the Anaplasmataceae-targeted polymerase chain reaction (PCR) assay amplified a sequence 100% identical to the Wolbachia pipientis endosymbiont of the rodent filarid Litomosoides galizai. Conclusions We report a variety of Hepatozoon haplotypes associated with small mammals in three Brazilian biomes: Cerrado, Atlantic Forest, and Pantanal. Through phylogenetic analyses, the Hepatozoon agents grouped in the rodent-marsupial-reptile large clade of Hepatozoon spp. from the world. The detection of a W. pipientis associated with the rodent filarid L. galizai indicates that the rodent was infected by filarial nematodes. Graphical Abstract


Background
Tick-borne protozoans of the genus Hepatozoon and order Piroplasmorida (genera Babesia, Cytauxzoon and Theileria) have been associated with infections and diseases of domestic and wild mammals in Brazil [1][2][3][4][5]. Hepatozoon spp. are apicomplexan parasites characterized by a heteroxenous life-cycle, in which the intermediate hosts (vertebrates) become infected primarily through the ingestion of hematophagous arthropods (definite hosts such as ticks and mosquitoes) containing mature oocysts [6]. Alternative routes of transmission, such as the predation of infected vertebrates containing Hepatozoon cysts in their tissues, have been described and seem to be an important infection route for carnivore hosts [7,8]. The order Piroplasmorida comprises tick-borne agents that infect mammalian blood cells (mostly erythrocytes), and have a major impact on farm and pet health-associated costs worldwide, but can also occur in wildlife [9,10].

Parasites & Vectors
The Anaplasmataceae family includes bacteria of the genera Anaplasma, Ehrlichia, Neorickettsia, and Wolbachia, from which the first two encompass tick-borne agents of veterinary and public health significance world widely [11]. All genera except Wolbachia are known to infect vertebrate cells (mammals or birds). Through the tick bite, the bacterium enters the bloodstream and infects specific host cell types, such as neutrophils, monocytes and macrophages, platelets, erythrocytes, or endothelial cells depending on the agent [5,12].
Small mammals such as wild rodents and marsupials are hosts of numerous species of ticks at some stage (larva, nymph, or/and adult) of their life-cycle [13]. In Brazil, several species of ticks have been found parasitizing small mammals [14][15][16]. Furthermore, these vertebrates have frugivorous-omnivorous habits, including the consumption of small vertebrates, invertebrates, and fruits [17]. Indeed, these habits are likely to increase the acquisition of these pathogens interspecies, especially in the case of Hepatozoon spp.
Various studies have reported the occurrence of apicomplexan and Anaplasmataceae agents in wildlife in Brazil during this century [3,[18][19][20][21][22][23][24][25][26], but the focus on small mammals has been much lower, which suggests that the diversity of these protozoa and bacteria might be by far underestimated, especially among marsupials. In this context, the aim of the present study was to investigate the occurrence of Hepatozoon spp., Piroplasmorida, and Anaplasmataceae agents in small mammals in seven forest fragments in Brazil.

Study areas and sampling procedures
Seven forest fragments, six of them located in the State of São Paulo and one in the State of Mato Grosso do Sul, were prospected (Fig. 1). Areas A1, A2, A3, and A4 are located in transition areas of the Cerrado and Atlantic Forest biomes; A5 belongs to the Cerrado biome; A6 is within the Atlantic Forest biome, whereas A7 is in the Pantanal biome of the state of Mato Grosso do Sul. Fieldwork was performed during the years 2015-2018 in the dry (summer) and wet (winter) seasons of each year. Small mammals were captured by Tomahawk and Sherman-like traps. General characteristics of the study sites, details on field study, and the protocols for handling the animals, anesthetic doses, storage, and identification have been published recently [16]. Briefly, blood samples were collected from all captured animals, whereas fragments of internal organs (liver, lung, and spleen) were collected from those animals that were euthanized. Collected samples were kept frozen at −20 °C until molecular analyses.

Molecular analyses
DNA extractions from blood, liver, lung, and spleen samples were carried out using the DNeasy Tissue & Blood Kit (Qiagen, Chatsworth, CA, USA) according to the manufacturer's instructions. In order to verify the success of extraction, an initial polymerase chain reaction (PCR) targeting the mammalian mitochondrial cytochrome b gene (cytb) was performed as described [27]. Subsequently, DNA samples were individually tested by a battery of PCR assays targeting protozoa of the genus Hepatozoon, protozoa of the order Piroplasmorida (genera Babesia, Cytauxzoon, and Theileria), and bacteria of the family Anaplasmataceae (Table 1). Samples yielding amplicons by the initial Hepatozoon genus-specific PCR were submitted to PCR protocols to amplify a larger fragment of the 18S ribosomal RNA (rRNA) gene of this protozoan genus. PCR assays were performed in a total volume of 25 μl, using DreamTaq Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA). Primers, PCR conditions, and positive controls are provided in Table 1. Negative controls consisted of ultrapure water. Products were resolved in 1.5% agarose gels and amplicons with the expected size purified and prepared for sequencing with the BigDye kit (Applied Biosystems, Foster City, CA, USA). An ABI PRISM 3500 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) was employed for sequencing using the same primers to perform PCRs. Obtained sequences were subjected to Basic Local Alignment Search Tool nucleotide (BLASTn) analyses to check identities with congeneric organisms available in GenBank [28].

Phylogenetic analyses
Partial 18S rDNA sequences of Hepatozoon generated in the present study were aligned with Clustal X [29] and manually refined by GeneDoc [30] with corresponding Hepatozoon sequences available in GenBank. Two phylogenetic trees, one for a small fragment (≈550 bp) of the 18S rRNA gene (generated from the initial PCR assays), and one for a larger fragment (≈1700 bp) of the 18S  rRNA gene were constructed. Both trees were inferred by using maximum parsimony (MP), as implemented in PAUP version 4.0b10 [31] with 1000 bootstrap replicates, and by Bayesian analysis (BA) using MrBayes v3.1.2 [32] with 1,000,000 replicates. The first 25% of the trees represented burn-in, and the remaining trees were used to calculate Bayesian posterior probabilities.

Results
A total of 524 mammals were captured, comprising seven species of marsupials, 14 rodents, two carnivores, and one Cingulata ( Table 2). Blood samples were collected in all cases. Among the 524 animals, 277 were euthanized and samples of liver, lung, and spleen collected. We tested a total of 1355 samples (524 blood, 277 liver, 277 lung, and 277 spleen). Expected size amplicons for cytb gene were obtained in all 1355 samples, thus confirming successful DNA extractions. Among the 524 sampled animals, 21 (4.0%) yielded amplicons in the initial Hepatozoon genus-specific PCR targeting a 574-bp fragment of the 18S rRNA gene. These 21 animals comprised 11 individual rodents from the state of São Paulo (one from A4, one from A5, and nine from A6), and six individual marsupials and four rodents from the state of Mato Grosso do Sul (A7) ( Table 3). Only one of the 21 Hepatozoon-infected animals yielded amplicons from blood. The remaining 20 infected mammals yielded amplicons from the lung only (7 animals), liver only (3), spleen only (1), lung and spleen (4), and lung, liver, and spleen (5) Based on the above samples that generated four distinct haplotypes, we attempted to generate a larger sequence (≈1700 bp) of the Hepatozoon 18S rRNA gene from at least one small mammal of each haplotype. Although no larger fragment was generated from haplotype 1-representatives, we did generate it from four haplotype 2-representatives, from which their identical 18S rRNA Table 3 Data of the 21 small mammals that were shown to contain Hepatozoon DNA by two PCR protocols, one targeting a small fragment (≈540 bp) and the other targeting a large fragment (≈1700 bp) of the Hepatozoon 18S RNA gene a Areas A4, A5, A6, and A7 are indicated in Fig. 1 Table 3).
In the phylogenetic tree inferred from short partial sequences (≈540 bp) of the 18S rRNA gene of Hepatozoon spp., haplotypes 1, 2, 3, and 4 branched within a large clade composed by many Hepatozoon haplotypes associated with rodents and reptiles from different parts of the world. This large clade was sister to another large clade that contained Hepatozoon haplotypes associated with canids (Hepatozoon canis) and felids (Hepatozoon felis) under 66 (MP) and 0.8 (BA) bootstrap supports (Fig. 2). This topology was somewhat similar in the phylogenetic tree inferred from long partial sequences (≈1700 bp) of the 18S rRNA gene of Hepatozoon spp., in which the sequences of the present study branched within a large clade containing many haplotypes associated with rodents and a few reptiles under 84 (MP) and 1.0 (BA) bootstrap supports (Fig. 3).
None of the 1355 samples yielded amplicons by the Piroplasmorida-PCR protocol. One blood sample from an Oligoryzomys sp. from Piracicaba (A1) yielded amplicons by the Anaplasmataceae-PCR protocol. These amplicons generated a DNA sequence 100% (306/306 bp) identical to the Wolbachia pipientis endosymbiont of the rodent filarid Litomosoides galizai (AJ548800).

Discussion
In the present study, four different haplotypes (1,2,3,4) of Hepatozoon spp. were detected in small mammals from different biomes among two Brazilian states, São Paulo and Mato Grosso do Sul. In São Paulo state, haplotype 1 was detected in rodents from Cerrado and a transition area of Cerrado and Atlantic Forest biomes, whereas haplotype 2 was detected in rodents from the Atlantic Forest biome. On the other hand, haplotypes 3 and 4 were restricted to rodents and marsupials, respectively, from the Pantanal biome of Mato Grosso do Sul. No host species shared more than one haplotype. Despite these distinct geographical and host associations, our phylogenetic analyses indicated that the four Hepatozoon haplotypes belonged to the same clade that contained nearly all haplotypes previously reported on rodents and marsupials, in addition to several reptile-associated haplotypes from different parts of the world. This general topology was confirmed by our analysis inferred from nearly complete sequences (≈1700 bp) of the 18S rRNA gene of Hepatozoon spp.
Previous molecular detections of Hepatozoon spp. on small rodents and marsupials in Brazil are restricted to six studies [20,25,[33][34][35][36]. Indeed, the vast majority of these records were on small rodents (at least 12 rodent species), in contrast to only two records on marsupials: Hepatozoon sp. PCR165 on one Thylamys macrurus [35], and H. canis on two Didelphis albiventris [34]. Except for this later record of H. canis, all previous records on small rodents and marsupials in Brazil demonstrated that the Hepatozoon spp. haplotypes belonged to the large clade composed of haplotypes associated with rodents, marsupials, and reptiles, in agreement with the results of the present study. Unfortunately, many of the sequences from previous studies were from regions of the 18S rRNA gene different from the present study (in the case of our Fig. 2) or much shorter than 1700 bp (in the case of our Fig. 3); for this reason, they were not included in our phylogenetic analysis.
We retrieved rodent-associated-Hepatozoon haplotypes (only those > 500 nucleotides) from previous Brazilian studies [20,25,[33][34][35][36] from GenBank and aligned them with the 18S rRNA long haplotypes (≈1700-haplotypes) of the present study (data not shown), in order to determine whether any of them matched 100% to any part of the long haplotypes. As shown in Table 4, haplotypes HF005, HF009, HF014, and HF015 (all from the rodents Akodon sp. and E. russatus of the state of São Paulo) had internal regions that were 100% identical to 516-1007-bp haplotypes of Hepatozoon sp., previously reported on the rodents Akodon montensis, Akodon cursor, and Galea spixii from three Brazilian states, including São Paulo [25,33]. Haplotype PS085 (from the rodent O. mamorae from the Pantanal biome of Mato Grosso do Sul) had regions that were 100% identical to 573-625-bp haplotypes of Hepatozoon sp., previously reported on this same rodent species and the same Brazilian biome [35], indicating that PS085 might be primarily related to O. mamorae in the Pantanal. On the other hand, no Hepatozoon sequence > 500 bp from GenBank matched 100% to haplotype PS001, which was detected in the marsupial M. domestica in the Pantanal biome. This result points PS001 as a novel Hepatozoon haplotype associated with marsupials in the Pantanal biome of Mato Grosso do Sul state, where other closely related haplotypes seem to be associated primarily with rodents and reptiles.
Vertebrate hosts acquire Hepatozoon infection through the ingestion of a hematophagous arthropod containing mature oocysts. Alternatively, infection can be acquired by intrauterine transmission or by predation of infected vertebrates containing merozoites (intermediate hosts) or cystozoites (paratenic host) [37]. Previous studies with wild rodents have detected Hepatozoon meronts in the liver or lungs (reviewed by [33]). Regarding the infection by Hepatozoon milleri in the rodent A. montensis in Brazil, Demoner et al. [33]  detected meronts only in the liver, whereas cystozoites were detected in the spleen, lungs, and kidneys. Since only one of the 21 infected hosts yielded amplicon in the blood (i.e., gamont detection), we infer that the small mammals of the present study contained meronts and/or cystozoites, hence, they might be acting as intermediate or paratenic hosts to Hepatozoon spp. in the studied areas. As previously stated [35,36,38], rodents might play a role in the epidemiological cycle of reptile-associated Hepatozoon spp. rather than the Carnivora-associated species (e.g., H. canis, H. felis), as indicated by ours and previous phylogenetic analyses. Among the seven sampled areas of the present study, Hepatozoon spp. infection was detected only in areas A4, A5, A6, and A7. According to previous analyses [16], these four areas presented significantly higher diversity of small mammals than areas A1, A2, and A3. Higher biodiversity could favor the transmission of Hepatozoon by interspecies predation, a likely important transmission route of this protozoan [35]. In addition, higher diversity of ticks could also facilitate Hepatozoon transmission. Again, it is noteworthy that areas A4, A5, A6, and A7 presented higher diversity of ticks than the remaining areas [16], suggesting that the life-cycle of small mammal-associated Hepatozoon spp. is under a complex interaction of vertebrate and invertebrate hosts.
Regarding the other PCR assays of the present study, none of the samples yielded detectable DNA to Piroplasmorida agents. This result contrasts with previous studies that reported different Piroplasmorida haplotypes among small mammals from other regions of Brazil, including studies that employed the same PCR protocol of the present study [3,20,24,39]. Therefore, these contrasting results could be related to uneven distribution of Piroplasmorida agents among the populations of small mammals along different geographical regions of Brazil, due to reasons yet to be determined.
The PCR assay targeting Anaplasmataceae bacteria yielded amplicons from the blood of an Oligoryzomys sp.: a 16S rRNA partial sequence 100% identical to W. pipientis associated with the rodent filarid L. galizai. This is indirect evidence that this rodent was infected by filarial nematodes. In previous studies, similar PCR assays amplified Wolbachia fragments from wild mammals (monkeys and opossums) and domestic dogs, leading the authors to conclude that these animals were infected by filarial parasites [22,40].

Conclusions
We report a variety of Hepatozoon haplotypes associated with small mammals in three Brazilian biomes: Cerrado, Atlantic Forest, and Pantanal. Through phylogenetic analyses, the Hepatozoon agents grouped in the rodent-marsupial-reptile large clade of Hepatozoon spp. from the world. Our phylogenetic analyses suggest that small mammals might play a role in the epidemiology cycle or reptile-associated Hepatozoon spp. rather than the Carnivora-associated Hepatozoon species. Since the Hepatozoon detections were restricted to areas with higher diversity of small mammals and ticks, this condition could facilitate Hepatozoon transmission. Finally, the detection of a W. pipientis associated with the rodent filarid L. galizai indicates that the rodent was infected by filarial nematodes. Table 4 List of short haplotypes (at least 500 bp long) previously reported in Brazil that were 100% identical to parts of the 18S rRNA long haplotypes (≈1700-haplotypes) of the present study a None of these sequences was included in the phylogenetic trees of the present study because they were from regions of the 18S rRNA gene different from the alignment used in Fig. 2 or they were much shorter than the ≈1700-bp alignment of Fig. 3 Long haplotypes of the present study (Brazilian state) Hepatozoon short haplotypes from GenBank that were 100% identical to regions of the long ≈1700 bp Hepatozoon haplotypes of the present study