Polychromophilus spp. (Haemosporida) in Malagasy bats: host specificity and insights on invertebrate vectors

Sampling sites and techniques

Fifty-two sites were visited from February 2012 to March 2013 in different areas of Madagascar to sample bats associated with taxonomic studies [25], ecology and distribution [27], as well as hosted ectoparasite and microparasites [30–33]. Bats were captured using mist nets and harp traps, which were most often placed at cave entrances or across bat flight pathways (Fig. 1). In a few cases, a butterfly net was used to sample individuals from cave and synanthropic day-roost sites. The exception was for Pteropus rufus, a CITES Appendix II species, for which living individuals were purchased in a market and not physically captured by the field research group. This species is considered as bushmeat on Madagascar and exportation of tissue samples for scientific work needed a CITES permits (cf. Ethical approval and consent to participate).

Upon capture, each individual bat, excluding Pteropus rufus as noted above, was placed in a clean cloth bag. Further, Nycteribiidae flies, an obligate parasite of bats [34] and parasitizing members of the families Miniopteridae and Vespertilionidae, were directly collected from individual bats for morphological and molecular identification [30, 33]. Voucher specimens were deposited at the Université d’Antananarivo, Mention Zoologie et Biodiversité Animale (formerly Département de Biologie Animale) (UADBA), Antananarivo, and at the Field Museum of Natural History (FMNH), Chicago.

DNA extraction and PCR screening

A pool of approximately 1 mm³ of kidney, lung and spleen from each individual bat was crushed in DMEM medium using two 3 mm-tungsten beads in a Tissue Lyser II (Qiagen, Valencia, CA). Subsequently, the mixture was centrifuged and the supernatant used for nucleic acids extraction. Total nucleic acids were extracted using EZ1 robot, with the viral mini kit v2.0 according to the manufacturer’s protocol (Qiagen Valencia, CA, USA), which has been routinely used in the PIMIT laboratory for the screening of both RNA and DNA parasites as previously described [32, 35, 36]. Nycteribiidae recovered from bats belonging to the families Miniopteridae and Vespertilionidae were crushed in a Tissue LyserII using two 3 mm tungsten beads, and total nucleic acids were extracted with the Qiagen biorobot with the viral mini kit v2.0 following manufacturer’s protocol (Qiagen Valencia, CA, USA) (see [30] for additional information).

Bat samples and Nycteribiidae flies were screened for the presence of Polychromophilus spp. using a previously described nested PCR protocol targeting the mitochondrial cytochrome b locus (cyt b) of haemosporidian parasites, which has been used to screen bats from Madagascar, Cambodia and Switzerland [8, 17, 18]. Primary PCR were conducted in a 25 µl reaction volume containing 12.5 µl of GoTaq^® Hot start green master mix (Promega, Madison, WI, USA), 1 µl of each Plas1/Plas2 primer at 0.4 µM and 1 µl of total nucleic acid as template. The balance of the reaction volume was supplemented by nuclease free water. The cycling conditions were 40 s at 94 °C, 40 s at 50 °C, 1 mn at 72 °C for 35 cycles. Secondary PCRs were performed using 50 µl reactions containing 25 µl of GoTaq^® Hot start, 1 µl of each Plas3/Plas4 primers, and 1 µl of the primary PCR product completed with nuclease free water. The cycling conditions were 40 s at 94 °C, 40 s at 50 °C, 1 mn at 72 °C, and a final extension at 72 °C for 7 mn. All PCR reactions were preceded by an initial denaturation at 94 °C for 5 mn. PCR products were visualized in an electrophoresis gel and sent to Genoscreen (Lille, France) for direct Sanger sequencing on both strands using forward and reverse primers.

Phylogenetic analyses

Sequences from positive individual of bats and nycteribiid flies were visually edited in Geneious 6.1.4 (http://www.geneious.com/) [37]. For positive individuals, the consensus nucleotide sequences from each were saved and used in the phylogenetic study. For similar consensus nucleotide sequences obtained from bats, the different haplotype sequences were identified using “pegas” package [38] implemented in R software [39] and only one sequence per haplotype was used in the analysis (Additional file 1: Table S1). Selected nucleotide sequences were then combined with those downloaded from GenBank and subsequently aligned using MAFFT implemented in Geneious software [37].

Prior to conducting phylogenetic analyses, a jModelTest version 2.1.3 [40, 41] was performed revealing GTR + G as the best substitution model. Subsequently, Bayesian inference consisting of two independent runs of four incremental Metropolis Coupled Markov Chain Monte Carlo (MC³) iterations starting from a random tree was conducted using MrBayes 3.1.2 [42]. This analysis was run for 5,000,000 generations with trees and associated model parameters sampled every 500 generations. The first 10% of the trees were discarded as a burn-in. New nucleotide sequences produced in the context of this study were deposited in GenBank under accession numbers MH744503 to MH744537 (Additional file 1: Table S1).

Morphological examination of blood parasites

In the field, a thin blood smear was prepared from each captured bat, specifically to examine haemoparasite morphology. Blood smears were fixed with methanol and stained with GIEMSA at room temperature. In the context of this study, blood smears were not used for detecting infection, as PCR analysis is distinctly more sensitive than visual inspection [17, 43–45]. Further, in many cases, morphology of apicomplexan parasites does not provide sufficient characters for species-level identification. Nevertheless, some blood smears of PCR positive samples were examined to document the morphology of associated haemosporidian parasites for illustrative purposes. In such cases, blood smears were examined using a binocular microscope under immersion oil objective with 1000× magnification.

Statistical analyses

Generalized Linear Model (GLM) analyses were conducted to investigate variation in Polychromophilus infection rate between species and sex classes in Miniopterus spp. and a Chi square analysis was conducted to compare infection rates between sex classes in M. goudoti. All statistical computations were conducted using R version 3.0.0 [39].

Phylogenetic analyses and host-parasite relationship

Phylogenetic analysis based on cyt b revealed that Polychromophilus infecting bats in different portion of the world forms a monophyletic clade composed of four clusters. Polychromophilus in Malagasy bats is segregated into two clusters. A first cluster is composed of P. melanipherus (Fig. 2a) lineages infecting all eight Miniopterus spp. from Madagascar screened in the present investigation, as well as a single individual of P. furculus (H1 to H24 in the phylogenetic tree, Fig. 3, Additional file 1: Table S1). Polychromophilus melanipherus identified from Malagasy Miniopteridae clustered with those identified in M. schreibersii from Switzerland, Miniopterus inflatus from Gabon, and Miniopterus villiersi from Guinea (Fig. 3). Hence, excluding P. furculus, P. melanipherus is known from the Miniopteridae of Africa, Europe, and Madagascar. The second cluster is composed of P. murinus (Fig. 2b) recovered from M. goudoti on Madagascar (H1 to H5 in the phylogenetic tree, Fig. 3, Additional file 1: Table S1). Although the Malagasy clade of P. murinus forms a monophyletic group with haemoparasites recovered from Myotis daubentonii and other European Vespertilionidae (Fig. 3), they are not embedded within the clade and show a certain level of genetic divergence. Haemosporidian parasites identified from Kerivoula hardwickii (Vespertilionidae) sampled in Cambodia and from Neoromicia capensis and Pipistrellus aff. grandidieri (Vespertilionidae) sampled in Guinea cluster in two distinct clades labelled Polychromophilus sp. 1 and 2, respectively (Fig. 3).

Molecular screening of Nycteribiidae: insights on their role as potential vectors

In total, 38 individual nycteribiids belonging to three species (17 Penicillidia leptothrinax, 2 Penicillidia sp. (cf. fulvida), and 19 Nycteribia stylidiopsis) were tested via PCR screening for the presence of Polychromophilus. Five bat fly specimens were positive for P. melanipherus and one for P. murinus. Bat flies positive for P. melanipherus via the PCR analyses included four P. leptothrinax (two sampled on Miniopterus aelleni and two on M. manavi s.l.) in addition to a N. stylidiopsis obtained on M. gleni. Of note, this latter nycteribiid was sampled on a PCR negative bat host. Further, one Penicillidia sp. (cf. fulvida) specimen collected on Miniopterus griveaudi was found positive for P. murinus, whereas the associated bat host was infected with P. melanipherus.

Phylogenetic relationship of haemosporidian parasites

Previous phylogenetic studies of haemosporidian parasites infecting Malagasy bats revealed the presence of two different parasite species belonging to the genus Polychromophilus [6, 8, 17]. In this study, the presence of two Polychromophilus sister species infecting two widely distributed bat genera on Madagascar, namely Miniopterus (Miniopteridae) and Myotis (Vespertilionidae) was confirmed. Polychromophilus melanipherus infects all eight Miniopterus spp. sampled in the present study, which are endemic to the Malagasy Region (Madagascar and the Comoros). The haplotypes of haemoparasites infecting Malagasy bats are embedded within the P. melanipherus clade occurring in African and European Miniopterus spp. This topology strongly supports a tight specificity of P. melanipherus within hosts belonging to the Miniopteridae [6, 17, 48]. One individual of P. furculus (Rhinonycteridae) also tested positive and based on the cyt b sequence, it fell within the P. melanipherus clade. Raharimanga and colleagues [28] and Duval and colleagues [8], already reported the presence of one individual of P. furculus with haemosporidian infection based on microscopic examination. This P. furculus specimen might have been accidentally and transiently infected by an infected nycteribiid during a blood meal. The haplotype identified in this P. furculus is unique and different from haplotypes obtained from infected Miniopterus sampled in the same day roost. However, further investigations are needed to elucidate the importance of P. melanipherus infection in P. furculus with a greater sample from different localities.

The second cluster of Polychromophilus only infected M. goudoti and is genetically related to P. murinus. Interestingly, P. murinus from M. goudoti, a bat endemic to Madagascar, was not embedded within the cluster of P. murinus from European bat species and suggesting a different diversification history of these two subgroups. It has been suggested that the lineage in which M. goudoti is placed (Ethiopian clade V) diverged from the other lineages mainly composed of Palearctic and Oriental Myotis spp. (Clade I, II, III, IV) in the Miocene, over 11 Mya ago [49]. It can thus be hypothesized that P. murinus parasites infecting M. goudoti co-diverged with their host and infects a higher diversity of bats within the family Vespertilionidae such as Myotis daubentonii, Eptesicus serotinus, and Nyctalus noctula [8, 18].

No co-infection was found in the present study, which confirms previously published information [6, 17]. This suggests that infection is actually specific at the level of host genus. This pattern is remarkable in that on Madagascar Miniopterus spp. and M. goudoti are syntopic in the same day-roost sites. Such physical contact has been previously reported to favour host switching of pathogenic Leptospira between these two syntopic occurring host genera [32], but this does not seem to be the case for Polychromophilus spp.

Polychromophilus infection in bats from Madagascar

The molecular screening of Malagasy bats revealed a total infection rate of 13.5%, which is congruent with previously reported rates [8, 28]. Based on the analysis of 31 Malagasy bat taxa, haemosporidian infection is limited to the families Miniopteridae, Vespertilionidae, and Rhinonycteridae. Bayesian reconstruction showed the presence of P. melanipherus infecting all eight Miniopterus spp. sampled herein, as well as a single P. furculus specimen, while P. murinus was only detected in M. goudoti. While a significant difference was observed between sexes in P. melanipherus infecting Miniopterus spp., no such pattern was observed in P. murinus infecting M. goudoti. These results can be explained by the roosting ecology and behaviour of these two syntopic species. Both sexes of M. goudoti live together in colonies throughout the year. For Miniopterus spp., there is at least partial sexual segregation in day-roosting sites. For example, M. manavi s.l. was sampled on several occasions in an open rock-shelter cave at Ambohitantely, where only males were present, suggesting that the two sexes are in contact only during the initial stages of reproduction. Further, bats roosting in this shallow cave are presumably more exposed to haematophagous insects than bats of either sex roosting deep in caves. Nevertheless, for species such as M. mahafaliensis sampled in the present study and at different localities, both sexes were present within their day roost site.

Insight into invertebrate vectors

Polychromophilus spp. have been reported to be transmitted by nycteribiid flies [13, 14, 50], wingless Diptera that are obligate blood-sucking parasites of bats [34]. Molecular screening of bat flies sampled in the context of this study was carried out to identify potential candidate vectors. The screening results should be interpreted with some caution, as the presence of Polychromophilus DNA in a nycteribiid might simply be the result from a recent blood meal. Flies sample included P. leptothrinax, specific to Miniopterus spp., and Penicillidia sp. (cf. fulvida) and N. stylidiopsis, parasitizing both Myotis and Miniopterus [30, 33]. In several cases, Penicillidia sp. (cf. fulvida) and N. stylidiopsis were found on the same bat host [30, 33]. Both P. leptothrinax and N. stylidiopsis tested positive for P. melanipherus. Further, P. murinus was detected in a Penicillidia sp. (cf. fulvida) fly that was collected on a Miniopterus griveaudi specimen testing positive for P. melanipherus. Although, cross contamination cannot be excluded during bat sampling as Miniopterus griveaudi and M. goudoti occur in syntopic cave day-roost sites, the result can alternatively be interpreted as additional evidence for Miniopterus spp. being non-permissive to P. murinus infection. This situation may be related to the ecology of Miniopterus and Myotis. In fact, colonies of Miniopterus spp. and M. goudoti often live in syntopy (physical contact). Members of these two genera seem to have ectoparasites with relaxed host preference [30, 33]. However, given their close physical contact, it might be expected to favor parasite co-infection and/or host switching. However, no P. murinus/P. melanipherus co-infection was detected either in bat genera or in their nycteribiid ectoparasites [6, 17, 51], and thus suggesting host-specificity.