Anaplasma phagocytophilum, Ehrlichia canis, and Ehrlichia chaffeensis are tick-borne pathogens of a wide range of vertebrate hosts (McQuiston et al., 2003; Yabsley, 2010; Atif, 2015). Their ecology involves hard ticks (Ixodidae) as vectors and wildlife and domestic animals or human-associated fauna (D. virginiana, rodents) as hosts (McQuiston et al., 2003; Yabsley, 2010; Atif, 2015) (Figure 1). The emerging zoonotic pathogens E. chaffeensis and A. phagocytophilum are increasingly implicated as a human pathogen worldwide. The diseases they cause are human monocytic ehrlichiosis (HME) and human granulocytic anaplasmosis (HGA), respectively (Yabsley, 2010; Atif, 2015). Both pathogens cause similar clinical illnesses that involve malaise, gastrointestinal disorders, high fever, and severe headache (Dahlgren et al., 2011; Atif, 2015). During 2000–2007, the cases-fatality rate in the USA was 1.9 and 0.6%, respectively (Dahlgren et al., 2011). E. canis is not common in humans and his reservoir hosts are canids worldwide, especially domestic dogs (Sainz et al., 2015). It is a pathogen of high veterinary importance for dog health.

Despite the concerns for human health, the information about the presence of those pathogens in Mexico is scarce, and most reports have been focused on E. canis, domestic dogs as well as performed with preliminary serological tests (Pat-Nah et al., 2015; Movilla et al., 2016). Only recently the first Human fatality by HME in Mexico has been reported, suggesting the emergence of E. chaffeensis in Mexico (Sosa-Gutiérrez et al., 2016). Another recent study, reported the presence of these three pathogens in wild rodents from central Mexico, highlighting the potential risk to humans and wildlife exposed to these tick-borne rickettsiae (Sosa-Gutiérrez et al., 2014). Only domestic dogs have been reported as a host for E. canis (Pat-Nah et al., 2015), but no other definitive report has been made on wild vertebrates or humans in the study area. The domestic-wildlife interface is important in the transmission of zoonotic pathogens and represents an increasing and a significant threat to global health. Generally, emerging infectious diseases (EIDs) events are dominated by zoonosis (60.3% of EIDs) and the majority of these (71.8%) originate in wildlife and are increasing significantly over time (Jones et al., 2008). Similar conclusions are shown in Daszak et al. (2000), Morens et al. (2004), Weiss and McMichael (2004), and King et al. (2006). For this specific system, several studies identify the importance of the domestic-wildlife interface in the transmission of these three pathogens, associating factors to the risk of infection to A. phagocytophilum and E. chaffeensis. Such as the presence and abundance of vertebrate host species and vectors, forest cover, and high-risk outdoor activities (Paddock and Yabsley, 2007; Dahlgren et al., 2011; Bayles and Allan, 2014; Atif, 2015). Even though E. canis appears to be host-specific, one human case has been reported (Perez et al., 1996) and other wildlife canids, such as foxes, could be involved in the transmission of this pathogen (Guerrero-Sánchez and Weber, 2012). Thus, monitoring the presence of those pathogens in the most common and abundant host of this interface is relevant for understanding the disease dynamics, taking protection measures and preventing epidemiological outbreaks (Gompper, 2014). In this sense, species that are common and abundant in this interface, such as opossums (both Didelphis virginiana and D. marsupialis) and dogs, which in turn could be potential hosts for those pathogens (Figure 1), can be elements that facilitated zoonotic and anthropomorphic processes. It has been reported that only D. marsupialis, represent 40% of the total biomass of small mammals (Colchero et al., 2014), and human: dog ratio in rural towns of the study area can be as high as 1.7:1 (Ortega-Pacheco et al., 2007). Therefore, the objective of our base line study was to identify the role of opossums (D. virginiana and D. marsupialis) and owned, free-ranging dogs in the epidemiology of A. phagocytophilum, E. canis, and E. chaffeensis in a rural area near the city of Campeche, Mexico. To accomplish this objective, first, we determined the prevalence of those three pathogens in opossums and dogs through a polymerase chain reaction (PCR) diagnosis method, as well as assessing the abundance of Ixodidae ticks in both hosts. Then we identify factors associated with the abundance of ticks and the prevalence of these pathogens within the hosts.

High densities of dogs inhabiting the human-wildlife interface may facilitate their potential reservoir role for multihost pathogens (Gompper, 2014). The positive results we found support previous reports of dogs as a host for A. phagocytophilum (Woldehiwet, 2010; Atif, 2015; Çetinkaya et al., 2016). The previous report of A. phagocytophilum in dogs by PCR worldwide (serological studies excluded) vary from 0.5 to 6.3% (Çetinkaya et al., 2016), which is much less than the 27% prevalence we found. This difference could be due to: (1) A. phagocytophilum strain, (2) the presence of this pathogen in the rest of the vertebrate host community in the study area, and (3) previous studies have found that dogs with access to veterinary care might have low prevalence values, opposite to our study area (Jensen et al., 2007; Bowman et al., 2009; Çetinkaya et al., 2016; Movilla et al., 2016). We need further investigation to support the first two hypotheses and to discard that the high prevalence are due to the low sample size. The high density of free-ranging dogs and the unusually high prevalence we found suggest their role as a major reservoir host in the area. However, it is difficult to assess this conclusion, since there wasn't an exhaustive survey on dog population and on other potential hosts, such as squirrels and racoons (Keesing et al., 2010). We found that both species of opossums are potential hosts for A. phagocytophilum in southeastern Mexico. D. virginiana has already been reported as a competent host for A. phagocytophilum in the USA with a higher prevalence (36%) than the one we found (Keesing et al., 2012). This difference could be explained because opossums were captured in an area considered as of high risk of infection by this pathogen due to the population growth of his main reservoir host and vector (Dugan et al., 2006; Paddock and Yabsley, 2007; Dahlgren et al., 2011). To the best of our knowledge, this is the first study screening D. marsupialis for the presence of the three pathogens using molecular tools (PCR analysis) in southeastern Mexico.

The absence of E. canis and E. chaffeensis in opossums in the present study is an argument in favor that these pathogens cannot infect them. E. canis has never been detected in opossums. E. chaffeensis, has only been detected by serological methods in D. virginiana in two different studies, which may have cross-reacted with a related Ehrlichial agent (Yabsley, 2010). Furthermore, we need experimental studies to determine if both species of Ehrlichia can infect the Didelphidae. The prevalence of E. canis that we found in dogs is lower than the 36% reported previously in a city near the study area and despite they are considered as reservoir hosts for this pathogen (Pat-Nah et al., 2015). However, they did the study with dogs from animal shelters from the city, and this type of provenance is considered a risk factor of infection by E. canis de facto (Pat-Nah et al., 2015). Furthermore, infections by tick-borne pathogens may appear aggregated among hosts collected from closely spaced geographical sites (Paddock and Childs, 2003; Estrada-Peña and de la Fuente, 2014). Further research is needed in order to infer the low prevalence of E. canis in dogs and the absence of E. chaffeensis in the dog population from the study area.

Opossums from Bobolá had lower tick abundance than the rest of the localities. In part, this can be explained by the environmental history of capture sites. The proximity of the capture sites in this locality with the agricultural fields could affect tick abundance, since this activity is negatively associated with questing tick and tick abundance in the host because it excludes the hosts of ticks and the loss of habitat for these (Pfäffle et al., 2013). In contrast, cattle practices in Xcampeu and high dog densities in Chiná, may have favored the abundance of ticks in opossums. Nevertheless, it is important to mention that the environmental history is not sufficient to explain the observed pattern, since host-tick relationships are complex and depend on many local factors (microclimate, structure, and composition of vegetation and host community) (Pfäffle et al., 2013) and factors that occur at the landscape level (proportion of agricultural, livestock and vegetation areas, vegetation patch size, and connectivity with other patches) (Estrada-Peña and de la Fuente, 2014). The sex also influenced the abundance of ticks in opossums, being females the ones with the lower tick abundance. Even if this pattern hasn't been observed in opossums, it is common in other vertebrates such as deer and rodents (Brunner and Ostfeld, 2008; Ruiz-Fons et al., 2013). One of the main hypotheses to explain this, is the difference in behavior between males and females: for the case of opossums, it has been reported that males travel more distances during the day, even covering the territory of several females (Sunquist et al., 1987; Zarza and Medellín, 2014), which makes them more vulnerable to tick infestations. On the other hand, opossums had a lower tick abundance than dogs. This can be explained because the dogs had a higher body surface than opossums. Furthermore, dogs are present in high population densities in the study area, and it is well-known that tick infestations are density dependent. The habitat use of owned free-ranging dogs in the study area may facilitate tick infestation because they usually frequent open areas of the towns with high probability to be infested by the tick Rhipicephalus sanguineus, the domestic dog's main tick (Pat-Nah et al., 2015). In addition, the little veterinary health care that dogs receive here, makes them vulnerable to contracting high tick infestations as we found.

The only risk factors associated with infections by A. phagocytophilum and E. canis in the present study were the kind of host and locality. Opossums had a lower risk of infection by both pathogens than dogs. This result was to be expected because: (1) dogs are considered a common reservoir host for these two pathogens, (2) because of the greater abundance of ticks in dogs than opossums and; (3) the little (or none) veterinary health care they receive (Beall et al., 2012; Atif, 2015). There was also a locality effect: Chiná had the highest dog abundance, and this could have caused the infections with A. phagocytophilum to be concentrated there because dogs are considered the common host of this pathogen (Atif, 2015). Instead, in the rest of the localities, where the abundance and density of dogs are lower, so were the prevalence values found. In addition, the distribution of the three positive opossums, suggests that the localities with a greater abundance of dogs, also have an increased risk of infection by A. phagocytophilum.

Even though the small prevalence of A. phagocytophilum found in both opossums in the study area, it is evident that they can be infected. Those small frequencies of infection could be difficult to interpret, as the abundance that opossums can achieve in some regions is very high, therefore potentially increasing the chances of higher infection rates regionally. Because of the extremely high tick infestation rates we found, our data suggest that both opossums and dogs may favor tick populations regionally (Table 1, Figure 4). From 102 opossums, 25 had tick infestation rates greater than 25 ticks/individual. From this group, 9 were hosting 100 ticks or more (up to 215); that is an extremely high infestation rate considering the opossum's body size. This finding makes us wonder about their role in tick population dynamics and therefore their role in the ecology of tick-borne pathogens. In this sense, our data contradicts the assumption that “Virginia opossum kill the vast majority of ticks that attempt to feed on them” (Keesing et al., 2010).

The large dog abundance in the study area and the little health care they receive, place them as risk species in the transmission of the tick-borne pathogens for which they may be hosts. Due to the unusually high prevalence of A. phagocytophilum in dogs, they could even be considered reservoir hosts for this Ricketsial. In this context, the role of owned free-ranging dogs in the transmission of A. phagocytophilum becomes important because it is a pathogen that can infect a wide variety of vertebrates, including humans, and because they are an active species at the domestic-wildlife interface (Gompper, 2014). Furthermore, the study area is situated on The Yucatan Peninsula, which is a major stopover site for migratory birds, potentially connecting ticks and pathogen populations from North and Central America (Viana et al., 2016). Within this context, the ecology of A. phagocytophilum as a zoonotic pathogen, involve the diversity of local vertebrate hosts and human health through North and Central America. Further research is needed to determine which vertebrate species are involved in such dynamics; their host-competence; and which tick species from the high extent diversity are competent vectors of these and other tick-borne pathogens.

Understanding the dynamics of disease transmission in the human-domestic-wildlife interface has become a major priority in current ecological and epidemiological research. Under this approach, we found evidence of the presence of A. phagocytophilum in opossums and dogs in a rural area of The Yucatan Peninsula. Despite of the low prevalence of this pathogen in opossums, or null (for E. canis and E. chaffeensis), they could be indirectly involved in the dynamic transmission of tick-borne pathogens due to the high abundance of ticks that they presented. Males may have a major role in the transmission of those pathogens due to the greater average abundance of ticks in relation to the females. Other regional factors also appear to affect the abundance of ticks, such as livestock practices, which may favor it, or agriculture that may decrease it; and that eventually translates into a greater or lesser risk of infection. Regarding the high prevalence of A. phagocytophilum in dogs we found as well as previous reports in Mexico and in the world, they could be considered as a reservoir host in this region. This is a finding that cannot be overlooked in terms of public human and animal health, nor for future research with rickettsiosis in this and other Neotropical socio-ecosystems.

ER-V and MW: study design, field work, and data analysis; ER-V and GG-P: laboratory analysis; ER-V, GG-P, and MW: funding and manuscript writing.