Ehrlichia spp. and Anaplasma spp. in Xenarthra mammals from Brazil, with evidence of novel ‘Candidatus Anaplasma spp.’

Anaplasmataceae agents are obligatory intracellular Gram-negative α-proteobacteria that are transmitted mostly by arthropod vectors. Although mammals of the Superorder Xenarthra (sloths, anteaters, and armadillos) have been implicated as reservoirs for several zoonotic agents, only few studies have sought to detect Anaplasmataceae agents in this group of mammals. This study aimed to investigate the occurrence and genetic diversity of Anaplasma spp. and Ehrlichia spp. in blood and spleen samples of free-living Xenarthra from four different states in Brazil (São Paulo, Mato Grosso do Sul, Rondônia, and Pará). Nested and conventional PCR screening assays were performed to detect the rrs and dsb genes of Anaplasma spp. and Ehrlichia spp., respectively. The assays were positive in 27.57% (91/330) of the Anaplasma spp. and 24.54% (81/330) of the Ehrlichia spp. Of the 91 positive Anaplasma spp. samples, 56.04% were positive in a conventional PCR assay targeting the 23S–5S intergenic region. Phylogenetic and distance analyses based on the rrs gene allocated Anaplasma sequences from sloths captured in Rondônia and Pará states in a single clade, which was closely related to the A. marginale, A. ovis, and A. capra clades. The sequences detected in southern anteaters from São Paulo were allocated in a clade closely related to sequences of Anaplasma spp. detected in Nasua nasua, Leopardus pardalis, and Cerdocyon thous in Brazil. These sequences were positioned close to A. odocoilei sequences. Genotype analysis corroborated previous findings and demonstrated the circulation of two distinct Anaplasma genotypes in animals from north and southeast Brazil. The first genotype was new. The second was previously detected in N. nasua in Mato Grosso do Sul state. The intergenic region analyses also demonstrated two distinct genotypes of Anaplasma. The sequences detected in Xenarthra from Pará and Rondônia states were closely related to those in A. marginale, A. ovis, and A. capra. Anaplasma spp. sequences detected in Xenarthra from São Paulo and were allocated close to those in A. phagocytophilum. The analyses based on the dsb gene grouped the Ehrlichia spp. sequences with sequences of E. canis (São Paulo, Mato Grosso do Sul, and Pará) and E. minasensis (Rondônia and Pará). The data indicate the occurrence of E. canis and E. minasensis and two possible new Candidatus species of Anaplasma spp. in free-living mammals of the Superorder Xenarthra in Brazil.

Reaction conditions. All qPCR assays were performed with a final volume of 10 μL containing 1 μL of DNA sample (concentration mean: 162.5 and 50.55 ng/µL for spleen and blood samples, respectively), 0.2 μM of each primer and hydrolysis probe, 5 μL GoTaq Probe qPCR Master Mix (Promega Corporation, Madison WI, USA),  www.nature.com/scientificreports/ and sterilized ultrapure water (Nuclease-Free Water; Promega Corporation) q.s. 9 μL. PCR amplifications were performed in low-profile multiplate unskirted PCR plates (Bio-Rad, Hercules, CA USA) using a CFX96 Thermal Cycler (Bio-Rad). Quantification of the number of copies of target DNA/μL was performed using IDT psmart plasmids (Integrated DNA Technologies, Coralville, IA, USA) containing the target sequences. Serial dilutions were performed to construct standard curves with different plasmid DNA concentrations (2.0 × 10 7 to 2.0 × 10 0 copies/μL). The number of plasmid copies/µL of the amount (g/µL) of DNA/plasmid (bp) was determined by multiplying by 6.022 × 10 23 . Each qPCR assay was performed in duplicate for each DNA sample. All duplicates showing cycle quantification (Cq) values differing by > 0.5 were re-tested. Amplification efficiency (E) was calculated from the slope of the standard curve in each run (E = 10 -1/slope ). The reactions followed the standards established by the Minimum Information for Publication of Quantitative real-time PCR experiments 39 . All the cPCR assays were performed using 5 μL of the DNA samples (concentration mean: 162.5 and 50.55 ng/ µL for spleen and blood samples, respectively) in a mixture containing 1.25 U Platinum Taq DNA Polymerase Table 1. Description of primers, hydrolysis probes and thermal sequences used in qPCR assays for Ehrlichia spp. and Anaplasma spp. a qPCR multiplex.
The consensus sequences obtained in this study and those retrieved from GenBank were aligned using ClustalW software version 7 46 using Bioedit v. 7.0.5.3 47 . The best evolutionary model was chosen using the jModelTest2 software (version 2.1.6) on XSEDE 48 via the CIPRES Science Gateway 49 . The phylogenetic analyses were based on Bayesian inference (BI) and maximum likelihood (ML) methods. The BI analyses were performed using MrBayes 3.1.2 50 via the CIPRES Science Gateway. Markov chain Monte Carlo simulations were run for 10 6 generations with a sampling frequency of every 100 generations and a 25% burn-in. ML analyses were performed using the Blackbox RaxML cluster 51 using 1,000 bootstrapping replicates 52 . The phylogenetic trees were edited using TreeGraph 2.0.56-381 beta software 53 .
Genetic diversity and genealogies. The genetic diversity analyses for the rrs gene and 23S-5S intergenic region of Anaplasma spp. and for the dsb gene of Ehrlichia spp. were performed with the sequences obtained in this study aligned to phylogenetically closer sequences of A. phagocytophilum, A. marginale, A. ovis, A. odocoilei, Anaplasma spp., E. canis, E. minasensis, and Ehrlichia spp. retrieved from GenBank. Clustal/W software 46 via Bioedit v. 7.0.5.3 47 was used for the alignment. The sequences used were at least 420 bp, 310 bp, and 310 bp for the rrs gene, 23S-5S intergenic region, and dsb gene, respectively. Sequences that were smaller in size were excluded from the phylogenetic analysis. These alignments were used to calculate the nucleotide diversity (π), polymorphism level (diversity of haplotypes [Dh], number of haplotypes [h], and the average number of nucleotide differences [K]), using DnaSP v5 software 54 . The sequences were submitted to the TCS Network 55 and distance analysis based on the split-network was inferred using the programs Population Analysis with Reticulate Trees (popART) 56 and Splitstree v 4.14.6 57 , respectively.

Results
occurrence of Anaplasmataceae agents in Xenarthra mammals. The mean concentration of the extracted DNA was 162.5 ± 37.8 and 50.55 ± 12.1 ng/µL for spleen and blood samples, respectively. All 330 DNA samples were positive in the cPCR assay based on the endogenous gapdh gene. A total of 147 (44.54%) animals were positive for at least one agent.
Financial restraints limited the selection of samples for sequencing to only a few among the large number of positive samples. The selection was based on two steps. First, we selected samples that presented high intensity amplicons in agarose gel electrophoresis. These samples were then separated according to animal species and region of origin. A random selection was performed to obtain at least one representative of each positive species and of each location.  (Table 3). No sample was positive in the qPCR for Anaplasma spp. based on the groEL gene. Of the 91 samples positive in the screening assays for Anaplasma spp., 51 (56.04%) samples were positive for the 23S-5S intergenic region of Anaplasma spp., and 7/91 (7.7%) samples were positive in the qPCR assay for A. phagocytophilum based on the msp-2 gene. The latter positive samples were not quantified because of the low amount of DNA of the agent in the tested samples (Monte Carlo effect) 42 (Table 3). All samples were negative in the qPCR assays for A. marginale (msp1-β gene) (Supplementary Table 1 Table 4). The phylogenetic analysis inferred by the BI method ( Fig. 3a) positioned all the sequences obtained in Rondônia and Pará states in a single clade that was phylogenetically closer to Anaplasma spp. genotypes detected in rodents in Brazil (KP757841 and KY391803), with 89% branch support. Despite forming a single clade, the obtained sequences showed greater proximity to the clade of Anaplasma spp. found in ruminants (A. capra, A. ovis, and A. marginale). On the other hand, the sequences obtained from anteaters of São Paulo state were allocated to a clade closer to the sequences of Anaplasma spp. detected in ocelots (Leopardus pardalis), coatis (Nasua nasua), and crab-eating foxes (Cerdocyon thous) from the Pantanal natural region in southern Brazil, with 87% branch support. This clade was sister to the clade of A. odocoilei, with 90% branch support. Results of the ML analysis ( Fig. 3b) concurred (or agreed) partially with the BI analysis. While the sequences obtained in Rondônia and Pará states formed a single clade close to the clade of Anaplasma spp. found in ruminants, the Anaplasma sequences obtained from Xenarthra of São Paulo were subdivided into different clades, allocated close to the clade of A. odocoilei.
Additionally, four (BM31, BM32, BM128, and BM129) of the seven positive samples for A. phagocytophilum, based on the msp2 gene, were sequenced for the rrs gene. Interestingly, three sequences had a high identity with A. phagocytophilum by the BLASTn analysis. However, in the phylogenetic analysis, they were positioned together with rrs sequences of other Anaplasma spp. obtained in Rondônia and Pará states. BLASTn directly presented a greater identity of one sequence (BM31) with Anaplasma sp. of Rattus rattus.
Genotype analysis based on 38 rrs Anaplasma sequences, including 18 sequences obtained in this study and sequences of A. marginale, A. ovis, A. phagocytophilum, A. odocoilei, and Anaplasma spp., indicated the presence of 12 genotypes, with π = 0.01719, hd = 0.808, and K = 7.16643 (Table 5, Fig. 4a). Eight genotypes comprised more than one sequence. Genotype #3 comprised two sequences of A. phagocytophilum detected in South Korea. Genotype #4 comprised 16 sequences detected in sloths from Rondônia and Pará (this study). Genotype #5 comprised three sequences of A. phagocytophilum detected in the United States and Norway. Genotype #6 grouped two sequences of A. marginale detected in cattle from the Philippines and Uganda. Genotype #7 grouped two sequences of A. ovis detected in sheep and deer in China. Genotype #8 grouped the two sequences detected in anteaters from São Paulo state (this study) and one sequence detected in a coati (Nasua nasua) from MS. Genotype #11 comprised three sequences of A. odocoilei sequences detected in cervids and a fly (Lipoptena depressa) in the United States. Finally, genotype #12 grouped sequences of Anaplasma sp. detected in ocelot (Leopardus pardalis), crab-eating fox (Cerdocyon thous), and coati from Brazil.
Regarding the genotype network (Fig. 4a), it could be inferred that genotype #4, which comprises the rrs sequences of Anaplasma detected in sloths from Rondônia and Pará states, was derived from genotype #1, which comprises a sequence found in Rattus in Brazil (KY391803), upon a mutational event. Both genotypes originated from a median vector (genotype not contemplated in the presented tree). On the other hand, genotype #8, which covers the two sequences obtained in this study in anteaters from São Paulo state and a coati sequence obtained in the state of MS, originated from a median vector upon a mutational event. Moreover, the sequences of A. odocoilei also originated from the same median vector. Additionally, genotype #8 gave rise to genotype #12, composed of sequences obtained from different wild animals sampled in the state of MS.
Split-network analysis based on the rrs gene (Fig. 4b) corroborated the genotype network, since the Rondônia and Pará sequences were all positioned together and closer to the Anaplasma spp. sequence previously detected    www.nature.com/scientificreports/ Phylogenetic analyses based on the 23S-5S intergenic region of Anaplasma spp. positioned the sequences detected in Xenarthra in two distinct clades, composed only of sequences found in this study. The first clade was composed of the sequences detected in sloths from the Rondônia and Pará states, which was in close proximity to the clade of A. marginale and A. ovis. The second clade was composed of sequences obtained from anteaters from SP, which were in close proximity to the clade of A. phagocytophilum. Both the BI (Fig. 5a) and ML analyses (Fig. 5b) presented the same topology, although ML presented a better definition of the clades. The index of clade support was 100 and 99% for Rondônia and Pará clade 1 and 89 and 100% for SP clade 2, in the BI and ML analyses, respectively. Of the seven (7.7%) positive samples in the qPCR for the msp2 gene of A. phagocytophilum, only one (14.28%) was positive in the PCR based on the 23S-5S region. However, it was closer to A. marginale in both the BLASTn and phylogenetic analyses.
Genotype analysis based on 23 Anaplasma 23S-5S sequences, which included 14 sequences obtained from the Xenarthra sampled in this study as well as A. marginale, A. ovis, and A. phagocytophilum sequences, obtained eight different genotypes, with π = 0.08454, hd = 0.719, and K = 24.09486 (Table 5, Fig. 6a). Four genotypes comprised more than one sequence. Genotype #1 comprised all 12 Anaplasma 23S-5S sequences from sloths sampled in Rondônia and Pará states. Genotype #8 comprised two sequences detected in giant anteaters from São Paulo state. Genotype #2 comprised three sequences of A. marginale previously detected in Brazil and Mexico. Genotype #5 comprised two sequences of A. phagocytophilum detected in the United States and Norway. Based on the network of genotypes (Fig. 6a), genotype #1 originated from a median vector and several mutational events. The same median vector also originated from genotypes #2 and #4, corresponding to sequences of A. marginale. Genotype #8 also originated from a median vector, which in turn originated from genotype #7, and gave rise to genotypes #5 and 6, comprised of A. phagocytophilum sequences.  BLASTn analysis showed that 10 sequences were identical to E. canis detected in a dog from Colombia (MK783026). The remaining nine sequences showed identities ranging from 98.96 to 100% with E. minasensis detected in Rhipicephalus microplus from Brazil (JX629808) and in bovine from Australia (MH500007) ( Table 4).
Phylogenetic analyses based on the dsb gene of Ehrlichia spp. positioned the sequences obtained in this study in the E. canis and E. minasensis clades. The sequences obtained from São Paulo and Mato Grosso do Sul, and three from Pará were grouped in the clade of E. canis, and four sequences from Pará and all from Rondônia were grouped in the clade of E. minasensis. The topologies of phylogenetic trees obtained by both BI (Fig. 7a) and ML (Fig. 7b) methods corroborated the findings. Additionally, ML analysis inferred a subdivision within the E. canis clade. Five sequences found in Xenarthra mammals formed a minor clade close to a sequence detected in a dog from Colombia. Four sequences formed a second clade with the other sequences of E. canis analyzed. Finally, a sequence obtained from a M. tridactyla (83) from São Paulo was positioned separately from the others, with 98% branch support.
Genotypes based on 29 Ehrlichia sequences were analyzed. These included 17 sequences obtained in this study as well as E. canis and E. minasensis sequences detected in different countries. A total of four genotypes were found, with π = 0.03570, hd = 0.569, and K = 9.28079 (Table 5). Genotype #1 comprised all sequences of E. canis retrieved from GenBank as well as the sequences detected in Xenarthra sampled in this study in São Paulo and Mato Grosso do Sul, and three sequences from Pará (BM24, BM51, and BM61). Genotype #2 comprised all E. minasensis sequences retrieved from GenBank, three Xenarthra sequences sampled in Rondônia, and four from Pará (BM16, BM177, BM180, PV14). Genotypes #3 and #4, on the other hand, comprised unique sequences detected in specimens of B. variegatus from Rondônia, PV337, and PV41, respectively. Based on genotype network analysis (Fig. 8a), genotype #1 (E. canis) originated from genotype #3 through several mutational events. The latter seems to have originated from genotype #2 (E. minasensis), which, in turn, originated from genotype #4, both from a mutational event. The split-network analysis corroborated the main findings described by the genotype network analysis (Fig. 8b).

Discussion
The present study revealed a high rate of positivity for Anaplasma spp. and Ehrlichia spp. in Xenarthra mammals sampled in four different Brazilian states. Of the 330 animals, 147 (44.54%) were positive for at least one of the agents. Of these, 25 (17%) were positive for both Ehrlichia spp. and Anaplasma spp. Until this study, molecular data concerning Anaplasmataceae agents in biological samples of the Superorder Xenarthra have been scant. www.nature.com/scientificreports/ Guillemi et al. 28 detected the presence of A. marginale morulae in blood smears from a giant anteater in Argentina, which was confirmed by PCR assays based on msp-5 and msp1-α genes. In Brazil, Soares et al. 20 detected a new dsb genotype of Ehrlichia sp. in a three-toed sloth (Bradypus tridactylus) obtained in the state of Pará. The sequence was allocated a separate clade that was sister to the clade of E. ruminantium, and close to the sequences of Ehrlichia spp. detected in a horse and a fox in Brazil. The same animal was positive for the rrs gene, whose sequence was allocated in a clade close to A. phagocytophilum. The present study reports the occurrence of two possible species of Anaplasma spp. in mammals of the Superorder Xenarthra from Brazil, since the two genes analyzed showed low identity values obtained by BLASTn and the phylogenetic findings positioned the sequences obtained in this study in single clades that were separate from the others.
BLAST analyses performed for Anaplasma spp. showed that all rrs sequences detected in sloths from the states of Rondônia and Pará showed identity values lower < 99% (not exceeding 98.5%) with sequences of A. ovis, A. marginale, and A. centrale. Additionally, the sequences detected in anteaters from São Paulo also showed an identity < 99% with sequences of A. phagocytophilum and A. odocoilei. However, these same sequences showed an identity > 99% with an Anaplasma sequence previously detected in a coati from MS (KY4999186). Previous  www.nature.com/scientificreports/ studies have defined rrs sequences as having at least 95% identity to be identified at the genus level and 99% to be identified at the species level [58][59][60] . In view of this, we have proposed two new species circulating in these animals, and the species detected in São Paulo's anteaters is probably the same as that found in the Mato Grosso do Sul coatis. The analyses performed in the 23S-5S region intergenic sequences corroborated with the rrs gene. The identities obtained were quite low, not exceeding 90%, strengthening the hypothesis of two new species. The phylogenetic analyses corroborated the BLASTn results. The analysis of the rrs sequences of Anaplasma spp. obtained from sloths of Rondônia and Pará were allocated close to two Anaplasma sequences previously detected in rodents (R. rattus and H. megacephalus) from Brazil. In addition, the clades formed by these two sequences and those found in the present study had a sister clade formed by A. marginale, A. ovis, and A. capra. The Anaplasma sequences detected in anteaters from São Paulo were allocated in a clade close to a new genotype of Anaplasma spp. previously detected in wild mammals from the Pantanal Sul-matogrossense.
A similar topology was observed in the phylogeny based on the 23S-5S intergenic region, in which the clade formed by the Anaplasma sequences from Xenarthra sampled in Rondônia and Pará was a sister clade that was formed by Anaplasma species detected in ruminants. The two Anaplasma sequences obtained from anteaters in the state of São Paulo were in a clade completely separated from the other sequences, although they were closer to the A. phagocytophilum clade, raising questions about the possible influence of the geographical location or host species on the occurrence of Anaplasma species that affect these animals.
Out of the seven samples positive in the qPCR for the msp2 gene of A. phagocytophilum, only one (14.2%) was positive in the PCR based on the 23S-5S region. Despite this, it was phylogenetically related to the clade formed by A. marginale and A. ovis. Similarly, four samples that were also positive for the msp2 gene were positioned close to the clade of A. marginale, A. ovis, and A. capra in the phylogenetic analysis based on the rrs gene. This potentially indicates the possibility of the aforementioned qPCR protocol to amplify msp-2 gene fragments from Anaplasma species phylogenetically related to A. phagocytophilum. Alternatively, the animals may have been co-infected with A. phagocytophilum and the new Candidatus species. MSP2, an external membrane protein present in all Anaplasma species, is encoded by several polymorphic genes in A. marginale, A. centrale, and A. ovis, and by only one gene in A. phagocytophilum 61 .
To better understand the results, genetic diversity analysis and distance genealogies were performed. The results of both corroborated the previously presented phylogenetic positioning. For both the rrs gene and the 23S-5S intergenic region, the genotype analyses showed that the Anaplasma sequences obtained in Rondônia and Pará states formed new genotypes (#4 and #1, for rrs and intergenic regions, respectively), whereas the sequences obtained from anteaters in the SP state comprised one genotype (#8 for rrs and intergenic region).
In addition, the genotype network based on the rrs gene suggests that the genotype circulating in sloths in Rondônia and Pará states might have originated through three mutational events from genotype #1, which was detected in a R. rattus from Brazil, explaining the proximity of both in phylogenetic analysis. The genotype circulating in São Paulo anteaters is the same genotype previously found in a coati in Mato Grosso do Sul, which might have originated through a mutational event from a median vector. Additionally, the genotype network based on the intergenic region suggests that both genotypes (#1 and #8) found in this study might have originated from different median vectors through numerous mutational events. The distance analysis for both genes showed that Anaplasma sequences obtained from Xenarthra sampled in the northern region of the country were positioned apart from the others, but were closer to A. marginale and A. ovis. On the other hand, the sequences obtained from Xenarthra sampled in São Paulo were positioned apart, albeit closer to A. odocoilei and A. phagocytophilum, based on the rrs gene and intergenic region, respectively.
All analyses corroborated and provided strong evidence of the circulation of two new species of Anaplasma in Xenarhtra in Brazil, related to the region inhabited by these animals. We propose naming the species circulating in Rondônia and Pará states as 'Candidatus Anaplasma amazonensis' and the species detected in São Paulo as 'Candidatus Anaplasma brasiliensis' . Further studies are needed to validate these species as well as to determine the vectors responsible for their transmission.
Regarding the findings for Ehrlichia spp., all analyses including BLASTn, corroborated and grouped the sequences as E. canis or E. minasensis. In addition, two sequences detected in sloths phylogenetically close to E. minasensis formed two new and distinct genotypes (#3 and #4).
Interestingly, the dsb sequences of E. canis obtained from Xenarthra mammals comprised animals sampled in the states of São Paulo, Mato Grosso do Sul, and portions of Pará. The dsb sequences of E. minasensis were obtained from animals from Rondônia and Pará states. Similar to the analysis of rrs and 23S-5S Anaplasma sequences, these findings again raise questions about the possible regionalization of the Ehrlichia and Anaplasma species found in Xenarthra from Brazil.
The clear dichotomy between the Ehrlichia and Anaplasma species that infect Xenarthra from different regions of the country may be related to the distribution and abundance of tick species that function as Anaplasmataceae vectors 78 . However, previous studies performed in different Brazilian states have shown that the different tick species that parasitize Xenarthra mammals are more correlated to the host species than to the geographic region. For instance, sloths are mainly parasitized by Amblyomma varium and A. geayi, with the latter found Scientific RepoRtS | (2020) 10:12615 | https://doi.org/10.1038/s41598-020-69263-w www.nature.com/scientificreports/ more in the state of Pará. While giant anteaters and southern anteaters are usually parasitized by A. nodosum, A. sculptum, and A. calcaratum, armadillos are frequently parasitized by A. pseudoconcolor, A. auricularium, and A. sculptum 27,[79][80][81][82] . These studies show the high parasitic specificity of some species of ticks that infest these animals. For instance, the ticks collected from anteaters from the state of São Paulo in the present study were all identified as A. nodosum, which was the most frequently found species in this group of animals (data not shown). The vectorial capacity of the tick species frequently found in Xenarthra mammals for Anaplasmataceae agents is still unknown. Although the majority of ticks that parasitize this group of mammals belong to the genus Amblyomma spp., the main vectors of Ehrlichia spp. in Brazil belong to the genus Rhipicephalus spp. Further studies should be conducted to assess the vectorial capacity of these ticks and to better understand the genetic diversity of Anaplasmataceae agents that infect mammals of the Xenarhra superorder in Latin America as well as the possible role of these animals in the epidemiological cycle of these agents.

conclusion
The present study showed the high occurrence of Anaplasma spp. (27.57%) and Ehrlichia spp. (24.54%) in freeliving Xenarthra mammals sampled from four states in Brazil. In addition, the study provides the first description of the occurrence of E. canis and E. minasensis in this group of mammals. The analysis of two genetic regions of Anaplasma spp., one conserved (rrs), and another one more diverse (intergenic region 23S-5S) revealed similar results of the low identity in BLASTn analysis, phylogenetic positioning in two different clades that were separate from the others containing known species, and formation of two different genotypes from those comprising known Anaplasma species by the diversity analysis. Based on these findings, we propose two new Candidatus species-'Candidatus Anaplasma amazonensis' and 'Candidatus Anaplasma brasiliensis'-in Xenarthra from Brazil.