Diversity of Anaplasma species and importance of mixed infections in roe deer from Spain

Received: 30 April 2021 Revised: 24 August 2021 Accepted: 4 September 2021 DOI: 10.1111/tbed.14319 ORIGINAL ARTICLE Diversity of Anaplasma species and importance of mixed infections in roe deer from Spain Susana Remesar Alberto Prieto David García-Dios Gonzalo López-Lorenzo Néstor Martínez-Calabuig José Manuel Díaz-Cao Rosario Panadero Ceferino Manuel López Gonzalo Fernández Pablo Díez-Baños Patrocinio Morrondo Pablo Díaz Investigación en Sanidad Animal: Galicia (Grupo INVESAGA), Facultade de Veterinaria, Universidade de Santiago de Compostela, Lugo, Spain Correspondence AlbertoPrieto,FacultaddeVeterinaria, PabellónI,CampusUniversitarios/n. 27002, Lugo, Spain. Email:[email protected] Funding information SpanishRoeDeerAssociation,Grant/Award Number:2016-CL018;AsociacióndelCorzo Español(ACE),Spain);ProgrammeforConsolidatingandStructuringCompetitiveResearch Groups,Grant/AwardNumber:GRC2019/04; XuntadeGalicia,Spain Abstract Although wildlife can act as reservoirs of some Anaplasma species, studies on the presence and distribution of Anaplasma spp. in wild cervids are mainly limited and focusedonzoonoticspecies.InordertoidentifytheAnaplasma speciesinroedeerfrom Spain and to detect co-infections, 224 spleen samples were tested for Anaplasma spp. using a commercial qPCR; positive samples were further characterized using generic 16S rRNA primers and species-specific primers targeting the msp2 and groEL genes. Anaplasma DNA was detected in the 50.9% of samples, and four Anaplasma species were identified. Anaplasma phagocytophilum (43.8%) was predominant, followed by Anaplasma bovis (13.8%), Anaplasma capra (5.8%) and Anaplasma ovis (2.2%). In addition, strains similar to Anaplasma platys were found in nine animals. Most positive roe deer (71.9%) were infected with a single Anaplasma species, whereas co-infections with two (19.3%) or three (8.8%) Anaplasma species were also found. This study confirms the widespread occurrence of Anaplasma spp. in roe deer from Spain, being the first report of A. platys-like strains and A. capra in this cervid; it is also the first report of A. capra in Spain. The detection of Anaplasma species pathogenic for humans and/or domestic animals in roe deer suggests that this cervid may play a role in the sylvatic cycle of these bacteria contributing to the appearance of clinical anaplasmosis cases. In addition, co-infections are common in roe deer revealing that Anaplasma species specific PCR assays are essential for a reliable identification as well as for determining their real prevalence. KEYWORDS Anaplasma bovis,Anaplasma capra,Anaplasma ovis,Anaplasma phagocytophilum,Anaplasma platys, roe deer This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2021 The Authors. Transboundary and Emerging Diseases published by Wiley-VCH GmbH e374 wileyonlinelibrary.com/journal/tbed Transbound Emerg Dis. 2022;69:e374–e385. REMESAR ET AL.e375 1INTRODUCTION ObligateintracellularbacteriaofthefamilyAnaplasmataceae havebeen reported worldwide in wildlife (García-Pérez et al., 2016). Among them, several species belonging to the genus Anaplasma are considered aetiological agents of a number of tick-borne diseases in mammalian hosts (Dumler et al., 2001) including bovine (Anaplasma marginale, Anaplasma bovis and Anaplasma centrale) and ovine anaplasmosis (Anaplasma ovis), human granulocytic anaplasmosis and tickborne fever of ruminants (Anaplasma phagocytophilum) and canine anaplasmosis (Anaplasma platys) (Battilani et al., 2017;delaFuente et al., 2005). Up to now, only A. phagocytophilum, A. capra, A. ovis and A. platys are considered zoonotic (Arraga-Alvarado et al., 2014;Breitschwerdt et al., 2014; H. Li, Zheng, et al., 2015). The distribution of each Anaplasma species depends on several factors, mainly the presence of proper tick vector species as well as suitable hosts and competent reservoirs (Estrada-Peña & de la Fuente, 2014). In this regard, it has been reported that wildlife, especially cervids, can act as reservoirs or asymptomatic carriers of some Anaplasma species (Atif, 2016; Ben Said et al., 2015; de la Fuente et al., 2008; Overzier et al., 2013; Renneker et al., 2013). In Spain, Ixodes ricinus, Rhipicephalus sanguineus s.l., Rhipicephalus bursa,Haemaphysalis punctata and Dermacentor reticulatus are considered the main vectors of Anaplasma spp. (Dantas-Torres, 2010; Koenen et al. 2013; Portillo et al., 2011; Palomar et al., 2015). However, no transovarial transmission of the pathogen has been demonstrated in their vectors (Rikihisa, 2011), and thus wild animals may play an important role in maintaining the ecological cycle of these bacteria in nature (Di Domenico et al., 2016; Woldehiwet, 2010). In this regard, wild cervids such as roe deer (Capreolus capreolus) and red deer (Cervus elaphus), together with otherdomesticruminants,areconsidered the main reservoirsof A. ovis, showingprevalencesusuallyhigher than 50% using molecular methods (de la Fuente et al., 2008; García-Pérez et al., 2016; Y. Q. Li, Yang, et al., 2015; Renneker et al., 2013). Similarly, A. phagocytophilum has been molecularly identified (6%–98%) in red deer, roe deer and fallow deer (Dama dama) (Hulínská et al., 2004; Hapunik et al., 2011; Overzier et al., 2013; Scharf et al., 2011; Teodorowski et al., 2020; Zeman & Pecha, 2008). A. bovis is a species less frequently found (4%–15%) in roe deer, white-tailed deer (Odocoileus virginianus), red deer or sika deer (Cervus nippon) (García-Pérez et al., 2016; Jilintai et al., 2009; Kawahara et al., 2006; Y. Q. Li, Yang, et al., 2015). In addition, it has been reported that deer can develop a persistent infection with A. marginale or A. centrale; the high seroprevalences detected reveal that these wild ungulates are frequently in contact with these pathogens (Atif, 2016). In recentyears, new Anaplasma species have been proposed; A. capra, first reported in goats from China (H. Li, Zheng, et al., 2015), has also been detected in red deer from France (Jouglin et al., 2019). Finally, a molecular investigation performed in China also identified organisms similar to A. platys in red deer (9%) and sika deer (15%) (Y. Q. Li, Yang, et al., 2015). Unravelling the role of wild ruminants as reservoirs of different Anaplasma species is of major importance for a proper understanding of the epidemiology of these bacteria. Nevertheless, most Anaplasma investigations in wildlife are focused on few Anaplasma species, mainly A. phagocytophilum. Molecular studies on the distribution of Anaplasma spp. in wild cervids from Spain are scarce and restricted to some areas (de la Fuente et al., 2008; García-Pérez et al., 2016; Portillo et al., 2011). Four Anaplasma species were previously detected in roe deer from Spain. In northern areas, A. phagocytophilum was the most prevalent (61%), followed by A. ovis (53%), A. bovis (3.81%) and A. centrale (0.95%) (García- Pérez et al., 2016). However, in southern areas, only A. ovis (53%) and A. phagocytophilum (18%) were detected (de la Fuente et al., 2008). Therefore, the objective of the present study was to determine the prevalence of Anaplasma spp. in roe deer hunted in four ecological areas covering the geographical distribution of this wild cervid in Spain. In addition, the molecular identification of these species, their distribution throughout the country and the presence of co-infections were assessed. Finally, the possible influence of the age and sex of the animals on the prevalence of these pathogens was studied. 2MATERIAL AND METHODS 2.1 Sample collection and preservation Amongst the wild cervids in Spain, roe deer is the second most abundant and hunted species, with a population around 200,000 specimens (Escudero et al., 2020); available official sources showed that 66,737 roe deer were hunted in 2018 in the country (MAPA, 2021). Between 2013 and 2020, the whole spleen of 224 roe deer from all the distribution areas of this wild cervid in Spain (Figure 1a) was collected during field evisceration by the hunters of the Spanish Roe Deer Association (Asociación del Corzo Español). Most of these samples (n=212) were included in a previous study on the molecular characterization of A. phagocytophilum in roe deer (Remesar et al., 2020). All samples were collected, classified and preserved as previously reported (Remesar et al., 2020). The location, age and sex of each deer were recorded. The age was estimated on the basis of teeth analysis (Høye, 2006). Regarding the location, four ecological areas (continental, Mediterranean, mountainous and oceanic) were established as previously described (Morrondo et al., 2017). The approval of the Ethics Committee/Welfare Authority was not required since all samples were collected post-mortem. 2.2 DNA extraction, detection and identification of Anaplasma species DNA extraction from splenic tissue was performed as previously described (Remesar et al., 2020). Detection of Anaplasma spp. DNA was performed in those samples that tested negative for A. phagocytophilum (124/212) in the previous study (Remesar et al., 2020)aswell as in 12 new samples using a commercial qPCR targeting the msp4 gene (EXOone Anaplasma spp., Exopol, Zaragoza, Spain). qPCR was e376 REMESAR ET AL. FIGURE 1 Maps showing roe deer distribution in Spain (a) and the four ecological areas (b-f). Dots represent the presence of Anaplasma spp. (a), Anaplasma phagocytophilum (b), Anaplasma bovis (c), Anaplasma capra (d), Anaplasma ovis (e) and Anaplasma platys-like (f) performed in an Applied Biosystems 7500 Fast Real-Time PCR System (Thermo Fisher Scientific, Massachusetts, USA) following the manufacturer’s instructions. All Anaplasma spp. qPCR positive samples, and those previously reported as positive to A. phagocytophilum (Remesar et al., 2020), were further tested using a PCR targeting the 16S rRNA gene of Anaplasma spp. (Table 1). A previously sequenced A. phagocytophilum sample and distilled water were included in each amplification reaction as positive and negative controls, respectively. PCR products were separated by electrophoresis on 1.5% agarose gels stained with RedSafe (iNtRON Biotechnology, South Korea) and then visualized using a Fluor-S MultiImager (Bio-Rad Laboratories, California, USA). The obtained 16S rRNA products were purified and sequenced in both senses on an ABI 3730xl (Applied Biosystems, Foster City, California, USA) using a BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) at the Sequencing and Fragment Analysis Unit of the Santiago de Compostela University (Spain). Sequences were aligned and edited using ChromasPro (Technelysium, Brisbane, Australia) and consensus sequences were compared with sequences available from REMESAR ET AL.e377 TABLE 1 Primers and protocols used for detection and identification of Anaplasma spp Gene target Primer name Primer sequence 5′-3′Fragment size Reference 16sRNA of Anaplasma spp. AnaplsppF AGA AGA AGT CCC GGC AAA CT 518 bp (Zobba et al., 2014) AnaplR3 GAG ACG ACT TTT ACG GAT TAG CTC msp2 of Anaplasma phagocytophilum msp2-3F CCA GCG TTT AGC AAG ATA AGA G 334 bp (Zeidner et al., 2000) msp2-3R GCC CAG TAA CAA CAT CAT AAG C groEL of Anaplasma bovis Ab groELF1 GTTCGCAGTATTTTGCCAGT ≈500 pb (Guo et al., 2019) Ab groELR CTGCRTTCAGAGTCATAAATAC Ab groELF2 ATCTGGAAGRCCACTATTGAT Ab groELR CTGCRTTCAGAGTCATAAATAC groEL of Anaplasma ovis Ao groELF AGCAAAATAGCGCAATGCGTC 722 bp (Belkahia et al., 2019) Ao groELR TCAACTCTATCCTTAAGCTC groEL of Anaplasma capra Ac groELF1 GCGAGGCGTTAGACAAGTCCATT 1264/1087 bp (Jouglin et al., 2019) Ac groELR3 TCCAGAGATGCGAGCGTGTATAG Ac groELF2 TGCACTGCTGGTCCAAAGGGGCT Ac groELR2 CAACTTCGCTAGAGCCGCCAACC groEL of Anaplasma platys Ap groELF ATGGTATGCAGTTTGATCGC 624/515 bp (Belkahia et al., 2019) Ap groELR1 TCTACTCTGTCTTTGCGTTC Ap groELF ATGGTATGCAGTTTGATCGC Ap groELR2 CATAGTCTGAAGTGGAGGAC the GenBank database using the Basic Local Alignment Search Tool (BLAST; http://blast.ncbi.nlm.nih.gov/Blast.cgi). Five different species-specific PCR tests targeting the msp2 gene of A. phagocytophilum and the groEL gene of A. bovis,A. capra,A. platys and A. ovis were also performed in all qPCR positive samples, including those previously positive to A. phagocytophilum, in order to determine the presence of co-infections. All PCR protocols were performed using previously reported protocols (Table 1), including positive and negative controls in each run assay. Unique partial sequences identified in this study were deposited in GenBank under accession numbers MW759445-MW759459. 2.3 Statistical analysis The possible influence of the ecological area as well as both the age and sex of roe deer on the prevalence of A. phagocytophilum and A. bovis was analyzed using a logistic regression; risk analysis could not be performed on other Anaplasma species due to their low prevalence. The number of samples of each category is summarized in Table 2; information from six roe deer was incomplete, so they were not age-classified and one animal could not be sexed. Factors were eliminated from the initial model using a backward and forward conditional method based on Akaike information criterion (AIC) value until the best model was built. All pairwise interactions were evaluated. Odds ratio (OR) were computed by raising ‘e’ to the power of the logistic coefficient over the first category of each factor (reference category). The logistic analyses and the AIC selection were performed with glm() and step() functions in the R software (R Core Team, 2020). Level of significance was set at p-values <.05. Phylogenetic analyses were carried out using MrBayes 3.2.7 software (Ronquist et al., 2012) by Bayesian approach with Markov Chain Monte Carlo sampling (10,000,000 generations sampling every 1000 steps). A Hasegawa-Kishino-Yano (HKY+G) and a General Time Reversible substitution model (GTR+G), both with gamma-distributed rate variation across sites were used for the analysis of Anaplasma 16s rRNA and A. bovis groEL sequences, respectively. Both models were selected based on AIC value using the free software jModel- Test v.2.1.10 (Darriba et al., 2012; Guindon & Gascuel, 2003). Trees were visualized and edited in FigTree 1.4.3 (http://tree.bio.ed.ac.uk/ software/figtree/). 3RESULTS Most of the 224 analyzed roe deer were males (n=157) and adults (n=167). Regarding ecological areas, the highest number of samples originated from the oceanic area (n=77), followed by Mediterranean (n=55) and both mountainous and continental areas (n=46 from each area) (Table 2). First, qPCR results showed that 114 out of 224 (50.9%) spleen samples were positive to Anaplasma spp. Subsequently, amplification at the 16S rRNA gene was detected in 103 out of 114 Anaplasma spp. positive samples (90.4%). Sequence analysis at this gene allowed the e378 REMESAR ET AL. TABLE 2 Prevalence of Anaplasma species in roe deer from Spain when considering the hunting location, the age and sex of the roe deer Hunting location Sex Age Oceanic (n=77) (95% CI) Mountain (n=46) (95% CI) Continental (n=46) (95% CI) Mediterranean (n=55) (95% CI) Female (n=66) (95% CI) Male (n=157) (95% CI) Unknown (n=1) (95% CI) Young (n=51) (95% CI) Adult (n=167) (95% CI) Unknown (n=6) (95% CI) Prevalence of each Anaplasma species detected on the total number of animals A. phagocytophilum 61.04% (49.22–71.74) 17.39% (8.32–31.95) 28.26% (16.45–43.68) 54.55% (40.66–67.80) 37.88% (26.47–50.70) 46.50% (38.56–54.60) 0.00% (0.00–94.53) 37.88% (34.95–63.23) 46.50% (36.13–51.59) 0.00% (0.00–48.32) A. bovis 10.39% (4.91–19.97) 8.70% (2.82–21.69) 10.87% (4.07–24.36) 25.46% (15.09–39.27) 10.61% (4.73–21.23) 15.29% (10.23–22.09) 0.00% (0.00–94.53) 10.61% (3.67–22.19) 15.29% (10.10–21.50) 16.67% (0.88–63.62) A. ovis 0.00% (0.00–5.92) 0.00% (0.00–21.69) 6.52% (17.00–48.93) 3.64% (0.63–13.60) 4.55% (1.18–13.56) 1.27% (0.22–5.00) 0.00% (0.00–94.53) 4.55% (1.53–17.23) 1.27% (00.21–4.71) 0.00% (0.00–48.32) A. capra 1.30% (0.00–8.01) 0.00% (0.00–21.69) 6.52% (17.00–48.93) 16.36% (8.20–29.30) 0.00% (0.00–6.85) 8.28% (4.66–14.03) 0.00% (0.00–94.53) 0.00% (0.10–11.79) 8.28% (3.39–12.50) 0.00% (0.00–48.32) A. platys-like 0.00% (0.00–5.92) 0.00% (0.00–21.69) 2.17% (0.11–12.97) 14.55% (6.93–27.22) 0.00% (0.00–6.85) 5.73% (2.82–10.93) 0.00% (0.00–94.53) 0.00% (0.10–11.79) 5.73% (2.24–9.55) 0.00% (0.00–48.32) Prevalence of Anaplasma single species and coinfections detected on the total number of animals A. phagocytophilum 54.55% (42.84–65.79) 13.04% (5.42–26.95) 15.22% (6.84–29.48) 21.82% (12.25–35.36) 30.30% (19.91–43.00) 29.94% (23.03–37.84) 0.00% (0.00–94.54) 35.29% (22.80–50.00) 29.34% (22.69–36.96) 0.00% (0.00–48.32) A. bovis 3.90% (1.01–11.73) 4.35% (0.76–16.04) 6.52% (1.70–18.93) 1.82% (0.00–10.99) 7.58% (2.82–17.50) 2.55% (0.82–6.81) 0.00% (0.00–94.54) 3.92% (0.68–14.59) 3.59% (1.47–8.01) 16.67% (0.88–63.52) A. capra 0.00% (0.00–5.92) 0.00% (0.00–9.60) 6.52% (1.70–18.93) 5.45% (1.12–16.07) 0.00% (0.00–6.69) 3.82% (1.56–8.50) 0.00% (0.00–94.54) 1.96% (0.10–11.79) 2.99% (1.11–7.22) 0.00% (0.00–48.32) A. phagocytophilum +A. bovis 5.19% (1.68–13.47) 4.35% (0.76–16.04) 4.35% (0.76–16.04) 7.27% (2.36–18.43) 3.03% (0.53–11.48) 6.37% (3.27–11.72) 0.00% (0.00–94.54) 5.88% (1.53–17.23) 5.39% (2.66–10.30) 0.00% (0.00–48.32) A. phagocytophilum +A. ovis 0.00% (0.00–5.92) 0.00% (0.00–9.60) 6.52% (1.70–18.93) 0.00% (0.00–8.13) 4.55% (1.18–13.56) 0.00% (0.00–2.98) 0.00% (0.00–94.54) 5.88% (1.53–17.23) 0.00% (0.00–2.80) 0.00% (0.00–48.32) A. phagocytophilum +A. platys-like 0.00% (0.00–5.92) 0.00% (0.00–9.60) 2.17% (0.11–12.97) 9.09% (3.40–20.71) 0.00% (0.00–6.69) 3.82% (1.56–8.50) 0.00% (0.00–94.54) 1.96% (0.10–11.79) 2.99% (1.11–7.22) 0.00% (0.00–48.32) A. bovis +A. capra 0.00% (0.00–5.92) 0.00% (0.00–9.60) 0.00% (0.00–9.60) 1.82% (0.09–10.99) 0.00% (0.00–6.69) 0.64% (0.03–4.03) 0.00% (0.00–94.54) 0.00% (0.00–8.73) 0.60% (0.03–3.80) 0.00% (0.00–48.32) A. phagocytophilum +A. bovis +A. capra 1.30% (0.07–8.01) 0.00% (0.00–9.60) 0.00% (0.00–9.60) 7.27% (2.36–18.43) 0.00% (0.00–6.69) 3.18% (1.18–7.66) 0.00% (0.00–94.54) 0.00% (0.00–8.73) 2.99% (1.11–7.22) 0.00% (0.00–48.32) A. phagocytophilum +A. bovis +A. platys-like 0.00% (0.00–5.92) 0.00% (0.00–9.60) 0.00% (0.00–9.60) 5.45% (1.12–16.07) 0.00% (0.00–6.69) 1.91% (0.49–5.92) 0.00% (0.00–94.54) 0.00% (0.00–8.73) 1.80% (0.47–5.58) 0.00% (0.00–48.32) A. phagocytophilum +A. ovis +A. capra 0.00% (0.00–5.92) 0.00% (0.00–9.60) 0.00% (0.00–9.60) 1.82% (0.09–10.99) 0.00% (0.00–6.69) 0.64% (0.03–4.03) 0.00% (0.00–94.54) 0.00% (0.00–8.73) 0.60% (0.03–3.80) 0.00% (0.00–48.32) A. phagocytophilum +A. bovis +A. ovis 0.00% (0.00–5.92) 0.00% (0.00–9.60) 0.00% (0.00–9.60) 1.82% (0.09–10.99) 0.00% (0.00–6.69) 0.64% (0.03–4.03) 0.00% (0.00–94.54) 0.00% (0.00–8.73) 0.60% (0.03–3.80) 0.00% (0.00–48.32) Total 64.94% (53.14–75.23) 21.74% (11.45–36.76) 41.30% (27.34–56.71) 63.64% (49.51–75.86) 45.45% (33.32–58.11) 53.50% (45.40–61.44) 0.00% (0.00–94.54) 54.90% (40.45–68.61) 50.90% (43.09–58.67) 16.67% (0.88–63.52) Abbreviation: CI, confidence interval. REMESAR ET AL.e379 FIGURE 2 Phylogenetic tree clustering of the partial 16S rRNA gene of Anaplasma spp. The tree was obtained using a Hasegawa-Kishino-Yano with gamma-distributed rate variation across sites method (HKY+G) with the software MrBayes 3.2.7 (Ronquist et al., 2012)byBayesian approach with Markov Chain Monte Carlo sampling (10,000,000 generations sampling every 1000 steps). The nucleotide sequence of Ehrlichia canis was used as an outgroup. Isolates identified in this study (*) identification of five Anaplasma species: A. phagocytophilum (68/103), A. bovis (11/103), A. capra (11/103), A. ovis (4/103) and A. platyslike (9/103). Finally, and after performing the species-specific PCRs, 98 samples tested positive to the msp2 gene of A. phagocytophilum (98/114) and 31 amplified through the study of the partial groEL gene of A. bovis (31/114). A low number of samples were positive to A. capra (13/114) and A. ovis (5/114) groEL gene specific PCRs. All A. platys-like samples were negative to the PCR assay targeting the specific groEL gene of A. platys. Sequence analysis of the species-specific PCRs at the msp2 and groEL genes confirmed species identification in all 16S rRNA- positive samples except those identified as A. platys-like. All A. phagocytophilum, A. bovis, A. capra and A. ovis 16S rRNA sequences showed a homology higher than 99.5% when compared to reference sequences (Supporting Information Material 1)andwere clearly separated in four clades in the phylogenetic analysis of this gene (Figure 2). In addition, A. platys-like sequences were grouped in a clade including A. platys and Candidatus Anaplasma camelii reference sequences (Figure 2). Most of them (5/9) presented a homology higher than 99.5% with the A. platys sequence KX987336 obtained from ticks in China (Lu et al., 2017). The other four A. platys-like samples showed a homology higher than 99.4% when compared to sequences identified as Candidatus Anaplasma camelii (MT510533) and A. platys (MN266939) obtained from camels and cattle in Kenya, respectively (Kidambasi et al., 2020; Sang et al., 2006). A. phagocytophilum sequences obtained through the study of msp2 gene were identical to those deposited in GenBank (Supporting Information Material 1). However, A. bovis groEL obtained sequences exhibited a 93%–97% homology when compared to the deposited A. bovis sequences MH255909, MK340768, MK340781 MK340800 and MK340803 detected in goats and ticks from China (Guo et al., 2018). Finally, all A. capra (5.8%; 13/224) and A. ovis (2.2%; 5/224) groEL sequences showed a similarity higher than 99.3% to A. capra (MH084718) and A. ovis (MG869402) sequences obtained from a red deer in France (Jouglin et al., 2019) and from a goat in China (Guo et al., 2018), respectively. When considering the overall prevalences for each Anaplasma species, the most frequent species was A. phagocytophilum (43.8%; 98/224), followed by A. bovis (13.8%; 31/224), A. capra (5.8%; 13/224) and A. ovis (2.2%; 5/224); Anaplasma platys-like microorganisms were identified in 4% of samples (9/224). Infections with a single Anaplasma species were predominant, with A. phagocytophilum infections the most frequent, followed by A. bovis and A. capra infections (Table 2). The remaining Anaplasma-positive roe deer showed co-infections with two or three different Anaplasma species. The combination of e380 REMESAR ET AL. TABLE 3 Logistic regression model for the prevalence of Anaplasma spp.; A.phagocytophilum; A. bovis and Anaplasma spp. co-infections. Factors were removed following the Akaike information criterion value until the best model was built Estimate z-Value p-Value OR CI 95% Anaplasma spp. (Intercept) 0.89475 2.220 .0265 2.4467287 1.12977872–5.5386893 Oceanicarea ----- Mountainous area −1.86743 −4.269 1.96e-05 0.1545197 0.06289031–0.3532744 Continental area −0.88340 −2.231 .0257 0.4133771 0.18787941–0.8924070 Mediterranean area - - - - - Mountain area −1.90887 −3.884 .000103 0.1482478 0.05439368–0.3771746 Continental area −0.92483 −2.105 .035292 0.3965983 0.16509069–0.9298878 Anaplasma phagocytophilum (Intercept) 0.7141 0.4013 .07519 2.0422583 0.94101393–4.5823094 Oceanicarea ----- Mountainous area −1.9753 −4.289 1.79e-05 0.1387200 0.05310513–0.3283898 Continental area −1.2296 −2.996 .00274 0.2923980 0.12759746–0.6427744 Mediterranean area - - - - - Mountain area −1.7996 −3.523 .000426 0.1653665 0.05773717–0.4337461 Continental area −1.7996 −2.334 .019583 0.3485644 0.14046084–0.8312083 Anaplasma bovis (Intercept) −2.30777 −3.657 .000255 0.09948325 0.02514075–0.3075346 Oceanicarea ----- Mediterranean area 1.04792 2.007 .044790 2.85170011 1.05189758–8.3249287 Anaplasma spp. coinfections (Intercept) −0.23295 −0.310 .756305 0.7921971 0.17178867–3.3844292 Mediterraneanarea----- Oceanic area −2.13901 −3.617 .000298 0.1177710 0.03319924–0.3488534 Mountain area −2.51042 −3.017 .002554 0.0812342 0.01145065–0.3436545 Continental area −1.20123 −2.090 .036607 0.3008250 0.08979144–0.8790248 Abbreviations: CI, confidence interval; OR, odds ratio. A. phagocytophilum/A. bovis and A. phagocytophilum/A. platys-like was the most prevalent dual co-infections. Finally, the most common triple co-infection was the association of A. phagocytophilum/A. bovis/A. capra (Table 2). Using logistic regression, significant differences in the prevalence of A. phagocytophilum and A. bovis were only found when considering the ecological area (Table 3). Thus, roe deer from oceanic areas showed a lower probability of being positive to A. phagocytophilum than those from continental (OR =0.29) and mountainous (OR =0.14) areas;the risk of being positivewas also lower inanimalsfrom Mediterranean areas than in those from continental (OR =0.35) and mountainous areas (OR =0.17) (Table 3). In addition, logistic regression showed that roe deer from Mediterranean areas presented a probability to be positive to A. bovis 2.9-fold higher than those from oceanic areas (Table 3). Due to the low number of A. capra, A. ovis and A. platys-like positive animals, risk analysis was not performed for these species. 4DISCUSSION It has been demonstrated that wild ungulates can play an important role in the epidemiology of some tick-borne pathogens, mostly acting as carriers (Atif, 2016). In the past years, changes in land use and urbanization have led to an increased interaction between wildlife and humans and domestic animals, increasing the risk of transmission of these pathogens (Mackenstedt et al., 2015). Detection and identification of Anaplasma species present in wild ungulates is a major goal for assessing the role of these animals on their epidemiology. Our data revealed that Anaplasma infections are very prevalent in roe deer from Spain, agreeing with previous molecular investigations performed in roe deer (65.7%–70.6%), red deer (50%) and fallow deer (50%) from northern areas of the country (de la Fuente et al., 2008; García-Pérezet al., 2016). Noticeableprevalencevalueshave also been reported in other wild cervid species from other countries such as sika deer from Japan (39.7%; Kawahara et al., 2006) and China (50%; Y. Q. REMESAR ET AL.e381 Li, Yang, et al., 2015). All these data demonstrate that Anaplasma infections are very frequent in cervids, suggesting that these wild animals may act as reservoirs of this pathogen. The results of this study show that wild cervids can be infected by a wide diversity of Anaplasma species. In this regard, three of the four Anaplasma species detected (A. phagocytophilum, A. bovis and A. ovis) have been previously reported in roe deer from Spain (de la Fuente etal., 2008; García-Pérez etal., 2016). Inaddition, this is thefirst report of A. capra and A. platys-like strains in European roe deer. A. phagocytophilum has been reported in a wide range of animal speciesaswellasinhumans(Atif,2016; Stuen et al., 2013); nevertheless, not all A. phagocytophilum strains are zoonotic since different variants adapted to particular geographical areas and hosts have been identified (Jahfari et al., 2014). Our results are consistent with previous investigations demonstrating that A. phagocytophilum is very prevalent in roe deer from different European countries (Atif, 2016; Teodorowski et al., 2020). Previous data from roe deer in Spain showed a high prevalence and variability of A. phagocytophilum strains, demonstrating that this wild ungulate is a reservoir for their own strains as well as some pathogenic A. phagocytophilum variants for humans and domestic animals (Remesar et al., 2020). It is worth noting that most molecular investigations on the presenceof Anaplasma spp. in wild ungulates were only focused on A. phagocytophilum because of its zoonotic potential, so data on the prevalence of other Anaplasma species is still limited. Our results revealed that A. bovis was the second most prevalent Anaplasma species in the sampled animals. This species is common in several domestic animals, being more prevalent in sheep and goats (16%–43%) than in cattle (≈4%) (Ben Said et al., 2015; Belkahia et al., 2015; Ceci et al., 2014; Liuetal.,2012; Nair et al., 2013; Ooshiro et al., 2008; Yang et al., 2015). It was also identified in a low-to-moderate percentage (4%– 15%) of wild ruminants such as roe deer, white-tailed deer, red deer and sika deer (García-Pérez et al., 2016; Kawahara et al., 2006;Y.Q. Li, Yang, et al., 2015), agreeing with our results. Previous studies analyzing the groEL gene of A. bovis demonstrated a high genetic intraspecies diversity suggesting the existence of different lineages (Guo et al., 2018), as observed in a phylogenetic tree (Supporting Information Material 2); this fact may explain the percentages of identity between our sequences and other deposited A. bovis sequences (Supporting Information Material 1). A. capra and A. ovis were detected in a low percentage of roe deer from Spain. A. capra was firstly reported in goats from China (H. Li, Zheng, et al., 2015) and since then it has been largely detected in Asian countries; thus, it has been found in sheep, cattle, Siberian roe deer, dogs, ticks and even humans with prevalences ranging from 6% to 12% (Peng et al., 2018; Shi et al., 2019; Seo et al., 2020; Yang et al., 2016). Nevertheless, a recent investigation also identified this species in some deer species such as red deer (3.4%) and swamp deer (Rucervus duvaucelii) (14.3%) from France, being the first report of A. capra outside Asia (Jouglin et al., 2019); these prevalence values agree with that detected in the present study. Thus, this is the first report of A. capra in Spain and the second report in Europe. It is worth noting the similarity of A. capra and A. centrale sequences at the 16S rRNA gene, as can be observed in Figure 2, that may lead to their misidentification. In fact, A. centrale isolates detected in deer from Japan by Kawahara et al. (2006) showed 16S rRNA sequences identical to the A. capra sequence MH762077 (Guo et al., 2018), suggesting that molecular analysis at more than one gene is required in order to achieve a reliable identification of both species. In addition, A. ovis was identified in a limited number of animals; this species has been detected in sheep and goats from Southern Europe (Italy and Portugal), Asia (Turkey, Iran, Iraq, Pakistan and China) and Africa (Kenya) with prevalence values ranging from 37% to 87% (Ahmadi-hamedani et al., 2012; Khan et al., 2015; Renneker et al., 2013; Torina et al., 2010; Yang et al., 2015). In contrast, data on the prevalence of A. ovis in deer is limited and restricted to a report in sika deer (20%) and red deer (32%) from China (Y. Q. Li, Yang, et al., 2015). It has been also found in a high percentage (53%) of roe deer from Spain (de la Fuente et al., 2008); the noticeable differences observed when compared to our data may be related to a decline in the number of sheep extensive farms in Spain in the last decade (Atif, 2016). In fact, a 22.4% reduction in the Spanish sheep population was reported in the last 10 years (Escudero et al, 2020). In the past years, strains closely related to A. platys have been identified in several domestic animals such as cattle, goats, sheep, cats and camels(AitLbachaetal.,2017; Belkahia et al., 2015; Dahmani et al., 2015; H. Li, Zheng, et al., 2015; Y. Li, Chen, et al., 2015; Selmi et al., 2019; Wei et al., 2020; Zobba et al., 2014) as well as red deer and sika deer from China (Y. Q. Li, Yang, et al., 2015). Molecular data revealed percentages of identity ranging from 92% to 99% when compared to canine A. platys sequences at 16S rRNA and groEL genes (Belkahia et al., 2015). In this study, A. platys-like strains were detected in nine animals, most of them from the Mediterranean area (Figure 1f), being its first report in roe deer. It is also worth noting that all positive A. platys-like roe deer showed co-infections with A. phagocytophilum. It must be considered that only the 16S rRNA gene was studied since the results of all samples were negative when analyzed using an A. platys specific PCR targetingthe groEL gene.These resultsmight indicate that thesestrains may be more related to the strains previously detected in camels than to A. platys reported in dogs (Belkahia et al., 2015). Riskanalysis showed that theprevalenceof someAnaplasma species was only influenced by the ecological area (Figure 1). A. phagocytophilum and A. bovis showed a wide geographical distribution since they were the only species detected in all the studied areas (Figure 1b,c), although it could be the result of their high prevalences. The significant differences observed in the prevalences of these species are probably related to the distribution of their major vectors (Remesar et al., 2020). The prevalence of A. phagocytophilum in roe deer (Figure 1b) was significantly highest in areas with oceanic climate located in the north of the country, where its main vector, Ixodes ricinus,ismore abundant since it needs narrow diurnal temperature variations and high humidity (Estrada-Peña, 2017; Pérez-Latorre et al., 1999); in fact, I. ricinus is the most frequently reported tick species parasitizing deer from northern oceanic areas (Vázquez et al., 2011). In contrast, ticks of the genus Rhipicephalus and Haemaphysalis, which are considered the main vectors of A. bovis (Atif, 2016), are ecologically very adaptable and tolerate different climatic environments (Estrada-Peña et al., 2017). e382 REMESAR ET AL. In addition, previous studies showed that ticks from the genus Rhipicephalus are very common in ungulates from Mediterranean areas of Spain(Contrerasetal.,2020; Ruíz-Fons et al., 2006).Inthesameway,A. capra, A. ovis and A. platys-like have also been reported in some species of the genus Haemaphysalis and Rhipicephalus (de la Fuente et al., 2007; Guo et al., 2018; Yang et al., 2016),whichmayexplaintheirhighest prevalences in continental and Mediterranean areas. Information about Anaplasma co-infections in deer is currently limited. In this regard, it has been suggested that infections with a particularAnaplasma species mayexcludeinfectionswithotherAnaplasma spp. (de la Fuente et al., 2002; Stuen et al., 2005). Nevertheless, the use of species-specific primers in the present study allowed determining that Anaplasma co-infections were very frequent in roe deer since about 30% of animals were infected with two or three Anaplasma species. These results are consistent with the fact that deer is usually highly infested with ticks, showing high prevalences of tick-borne pathogens (Portillo et al., 2011; Remesar et al., 2020; Vázquez et al., 2011). The little information available on Anaplasma co-infections may be related to the frequent use of generic PCR protocols in Anaplasma molecular investigations, allowing the selective amplification of the dominant species (García-Pérez et al., 2016; Yu et al., 2020) and thus masking coinfections. Our results reveal that performing species-specific PCRs is strongly needed for detecting co-infections as well as for achieving a reliableidentification of Anaplasma species, especially thoselessprevalent, and estimating their real prevalence. 5CONCLUSIONS This study confirms that Anaplasma spp. is very prevalent in roe deer from Spain, being widespread in the distribution area of this cervid. In addition, our data support that roe deer can host a wide diversity of Anaplasma species. This study also represents the first report of A. capra and A. platys-like strains in Spain. Since some Anaplasma species detected are considered pathogenic for domestic animals (A. phagocytophilum,A.bovis,A. ovis and A. capra) or even zoonotic (A. capra and A. phagocytophilum), our results suggest that roe deer may play an important role in the sylvatic cycle of these pathogens contributing to the appearance of clinical anaplasmosis cases in both domestic animals and humans. The major factor influencing the presence of A. phagocytophilum and A. bovis was the geographical location, which could be related to the presence of their main vectors. Finally, co-infections with two or three different Anaplasma species are frequent in roe deer and more common than expected; therefore, using Anaplasma species specific PCR protocols is essential for a reliable identification as well as for determining their real prevalence. ACKNOWLEDGEMENTS We would like to thank ACE hunters and “O Veral” Recovery Centre of Wild Animals staff for their inestimable collaboration in collecting samples. This research was supported by a project grant awarded by the Spanish Roe Deer Association (2016-CL018; Asociación del Corzo Español (ACE), Spain), the Programme for Consolidating and Structuring Competitive Research Groups (GRC2019/04; Xunta de Galicia, Spain). CONFLICT OF INTEREST The authors declare no conflict of interest. ETHICS STATEMENT No animals were culled for the purpose of this study. All samples were obtained post-mortem and kindly provided by the Spanish Roe Deer Association (ACE). Animals were hunted in accordance with the Spanish Hunter Code: (https://www.boe.es/biblioteca_juridica/ codigos/abrir_pdf.php?fich=095_Codigo_de_Caza.pdf). DATA AVAILABILITY STATEMENT The data that supports the findings of this study are available in the supplementary material of this article. ORCID Alberto Prieto https://orcid.org/0000-0002-3211-3494 José Manuel Díaz-Cao https://orcid.org/0000-0002-8119-7057 Pablo Díez-Baños https://orcid.org/0000-0003-2445-1095 REFERENCES Ahmadi-hamedani, M., Khaki, Z., Rahbari, S., & Ahmadi-hamedani, M. A. (2012). Hematological profiles of goats naturally infected with Anaplasma ovis in north and northeast Iran. Comparative Clinical Pathology,21, 1179–1182. https://doi.org/10.1007/s00580-011-1257-9 Ait Lbacha, H., Alali, S., Zouagui, Z., El Mamoun, L., Rhalem, A., Petit, E., Haddad,N.,Gandoin, C.,Boulouis, H.J.,&Maillard,R. (2017).High prevalence of Anaplasma spp. in small ruminants in Morocco. Transboundary and Emerging Diseases,64, 250–263. https://doi.org/10.1111/tbed.12366 Arraga-Alvarado, C. M., Qurollo, B. A., Parra, O. C., Berrueta, M. A., Hegarty, B. C., & Breitschwerdt, E. B. (2014). Case report: Molecular evidence of Anaplasma platys infection in two women from Venezuela. The American Journal of Tropical Medicine and Hygiene,91(6), 1161–1165. https://doi. org/10.4269/ajtmh.14-0372 Atif, F. A. (2016). Alpha proteobacteria of genus Anaplasma (Rickettsiales: Anaplasmataceae): Epidemiology and characteristics of Anaplasma species related to veterinary and public health importance. Parasitology, 143(6), 659–685. https://doi.org/10.1017/S0031182016000238 Battilani, M., De Arcangeli, S., Balboni, A., & Dondi, F. (2017). Genetic diversity and molecular epidemiology of Anaplasma.Infection, Genetics and Evolution: Journal of Molecular Epidemiology and Evolutionary Genetics in Infectious Diseases,49, 195–211. https://doi.org/10.1016/j.meegid.2017. 01.021 Belkahia, H., Ben Said, M., Sayahi, L., Alberti, A., & Messadi, L. (2015). Detection of novel strains genetically related to Anaplasma platys in Tunisian one-humped camels (Camelus dromedarius). Journal of Infection in Developing Countries,9(10), 1117–1125. https://doi.org/10.3855/jidc.6950 Belkahia, H., Ben Said, M., Ghribi, R., Selmi, R., Ben Asker, A., Yahiaoui, M., Bousrih, M., Daaloul-Jedidi, M., & Messadi, L. (2019). Molecular detection, genotyping and phylogeny of Anaplasma spp. in Rhipicephalus ticks from Tunisia. Acta tropica,191, 38–49. https://doi.org/10.1016/j. actatropica.2018.12.034 Ben Said, M., Belkahia, H., Karaoud, M., Bousrih, M., Yahiaoui, M., Daaloul- Jedidi,M., &Messadi, L.(2015).First molecularsurveyofAnaplasma bovis in small ruminants from Tunisia. Veterinary Microbiology,179(3–4), 322– 326. https://doi.org/10.1016/j.vetmic.2015.05.022 Breitschwerdt, E. B., Hegarty, B. C., Qurollo, B. A., Saito, T. B., Maggi, R. G., Blanton, L. S., & Bouyer, D. H. (2014). Intravascular persistence of