INTRODUCTION
Haemogregarines form a group of about 400 species of adeleid blood parasites with a suspected heteroxenous life cycle, species of the genus Haemogregarina being primarily parasites of reptiles and fish (Desser, Reference Desser and Kreier1993). In the past, the majority of new Haemogregarina species were described based on the morphology of their erythrocytic stages – e.g. gamonts and meronts. Moreover, strict host specificity was supposed, so each new studied host species led to a description of a new parasite. Such a practice led to considerable synonymy. In his review, Levine (Reference Levine1988) placed 300 species into the genus Haemogregarina, whereas Siddall (Reference Siddall1995) designated only 19 to represent Haemogregarina sensu stricto. The biology, vectors and transmission routes of haemogregarines are mostly unknown (Desser, Reference Desser and Kreier1993). Complete life cycles, allowing for taxonomic conclusions, have been described for only two Haemogregarina species of turtles: the type species of the genus, the parasite of the European pond turtle (Emys orbicularis) – Haemogregarina stepanowi (see Danilewsky, Reference Danilewsky1885; Reichenow, Reference Reichenow1910); and for the parasite of Nearctic snapping turtles (Chelydra serpentina) – Haemogregarina balli (see Paterson and Desser, Reference Paterson and Desser1976; Siddall and Desser, Reference Siddall and Desser1990, Reference Siddall and Desser1992). In both haemogregarine species, sexual development occurs in the leech hosts, where gametogenesis, the formation of the zygote and the monosporoblastic oocyst occur within intestinal epithelial cells. Subsequent merogony in endothelial cells produces hundreds of merozoites, invading the proboscis of the leech and inoculating the turtle during feeding. Pre-erythrocytic merogony takes place in the lungs, liver and spleen of the turtle. The last part of the life cycle comprising secondary merogony and the formation of gamonts occurs in the erythrocytes of the turtle, allowing for relatively easy detection (Reichenow, Reference Reichenow1910; Paterson and Desser, Reference Paterson and Desser1976; Siddall and Desser, Reference Siddall and Desser1990, Reference Siddall and Desser1992).
The Western Palaearctic is inhabited by four species of hard-shelled freshwater turtles belonging to two families – E. orbicularis (Emydidae), Mauremys caspica, Mauremys rivulata and Mauremys leprosa (Geoemydidae). As well as H. stepanowi from E. orbicularis (see Mihalca et al. Reference Mihalca, Racka, Gherman and Ionescu2008), the occurrence of haemogregarines has been reported in the blood of M. rivulata (see Desser and Yekutiel, Reference Desser and Yekutiel1987; Paperna, Reference Paperna1989), albeit without taxonomic considerations. Additionally, Haemogregarina bagensis (Ducloux, Reference Ducloux1904) was described from M. leprosa from north-western Tunisia. Subsequently, Billet (Reference Billet1904) described two forms of this parasite in M. leprosa from Algeria. The description of this haemogregarine from this same host was completed by Laveran and Pettit (Reference Laveran and Pettit1909). Surprisingly, more than a hundred years after their description, very little is still known about the biodiversity, taxonomy, evolutionary relationships and biogeography of these turtle haemogregarines.
The development of modern molecular-genetic methods has provided new, powerful tools to enable the acceptance or refutation of the validity of numerous previously described species and to study their genealogy and evolution. Surprisingly, these simple methods have still not been widely applied in the taxonomy of haemogregarines (see e.g. Wozniak et al. Reference Wozniak, Telford and McLaughlin1994), which still awaits revision. Even recent descriptions of new species frequently lack molecular-genetic specification and genealogical placement within the evolutionary tree of adeleids. Among Haemogregarina species, only H. balli phylogeny has been studied using molecular-genetic tools (Barta et al. Reference Barta, Ogedengbe, Martin and Smith2012).
By sampling all western Palaearctic hard-shelled freshwater turtle species from a wide geographical range, we set the following objectives for this study: (1) to test the relationships between H. balli and H. stepanowi by comparing their 18S rDNA sequences and to confirm their congeneric status within the genus Haemogregarina; (2) to evaluate the conspecificity of Haemogregarina isolates from all four western Palaearctic turtle species (genera Emys and Mauremys) by comparing the morphology of their blood-stages and by phylogenetic analyses of 18S rDNA sequences; (3) to consider the host specificity of Haemogregarina at the level of their vertebrate intermediate host.
MATERIALS AND METHODS
Sampling
Eighty-one turtles were collected by hand during 2005–2013 in Morocco, north-eastern Algeria, southern Bulgaria, Turkey, western Syria and Iran (Table 1). With a few exceptions, the animals were classified as male, female and juvenile. All turtles were released immediately afterwards at the location of capture. Blood samples were taken by puncture of the dorsal coccygeal vein and fixed in 96% pure ethanol. Thin blood smears were air dried, fixed in absolute methanol for 5 min and stained with Giemsa diluted 1 : 10 in distilled water (pH 7) for 20 min. The smears were examined by light microscopy using a 100× magnification objective lens equipped with immersion oil. Images were captured by Olympus DP 73 digital camera. For each infected turtle, the intensity of parasitaemia was estimated as the percentage of infected erythrocytes present in approximately 104 cells (e.g. 100 fields each with estimated 100 erythrocytes per field).
Studied material; m – males, f – females, juv – juveniles, nd – not detected
DNA extraction, PCR, sequencing
DNA was isolated using the NucleoSpin Tissue kit (Macherey-Nagel, Germany) according to the manufacturer's protocol and eluted in 100 μL of PCR water. Extracted DNA was quantified spectrophotometrically using a Nanodrop ASP-3700 (ACTGene, USA) and then stored at −20 °C. A pair of specific primers (EF: 5′-GAAACTGCGAATGGCTCATT-3′ and ER: 5′-CTTGCGCCTACTAGGCATTC-3′) designed originally for Eimeria by Kvičerová et al. (Reference Kvičerová, Pakandl and Hypša2008) was used to confirm the presence of DNA of apicomplexan parasites, amplifying up to 1500 bp long fragments of nuclear 18S rDNA of Haemogregarina. This gene proved to be sufficiently variable and informative both on the generic and specific levels within Apicomplexa (Barta et al. Reference Barta, Ogedengbe, Martin and Smith2012). PCR reactions were carried out in a 25 μL volume, including 1 μL each of 10 μ m PCR primer, 12·5 μL of Combi PPP Master Mix (Top-Bio, Czech Republic), 8·5 μL of PCR water and 2 μL of extracted DNA.
PCR conditions were as follows: initial denaturation at 95 °C for 4 min, followed by 30 cycles consisting of denaturation at 92 °C for 45 s, annealing at 58 °C for 45 s and extension at 72 °C for 90 s, and a final extension at 72 °C for 10 min. The results of amplification were visualized on 1·2% agarose gel using ethidium bromide under UV light. PCR products were purified using the Gel/PCR DNA Fragments Extraction Kit (Geneaid Biotech Ltd., Taiwan). The concentration of DNA was then measured. DNA sequencing was provided by the service laboratory (Macrogen Inc., the Netherlands) on an automatic 3730XL DNA analyser.
Phylogenetic analyses
Obtained sequences were identified by BLAST analysis, edited using the DNASTAR program package (DNASTAR Inc.) and deposited to the NCBI GenBank database under the accession numbers KF257926, KF257927, KF257928 and KF257929. Additionally, selected sequences from members of genera Adelina, Babesiosoma, Cryptosporidium, Dactylosoma, Haemogregarina and Hepatozoon were obtained from the GenBank (NCBI) (Table 2). Since only a single sequence of the genus Haemogregarina, namely H. balli (Barta et al. Reference Barta, Ogedengbe, Martin and Smith2012), was available in the GenBank, we enlarged the dataset by adding sequences of three Haemogregarina-like parasites collected from three African terrapins of the genus Pelusios available in our lab – P. marani from Gabon (KF257924), P. williamsi from Kenya (KF257923) and P. subniger from Mozambique (KF257925) (Table 2). The sequences were aligned in BioEdit (Hall, Reference Hall1999) using the ClustalW algorithm (Thompson et al. Reference Thompson, Higgins and Gibson1994). MEGA 5.0 (Tamura et al. Reference Tamura, Peterson, Peterson, Stecher, Nei and Kumar2011) was used to calculate genetic distances based on 18S rDNA sequences of Haemogregarina included in the analyses. Bayesian inference analysis (BI) was carried out with Mr Bayes 3.1.2. using a GTR+Γ+I model for 10 million generations (Ronquist and Huelsenbeck, Reference Ronquist and Huelsenbeck2003). Chain convergence and burn-in were estimated according to the indices implemented in the MrBayes program (deviation of split frequencies, potential scale reduction factor – PSRF) and using a Tracer program (Rambaut and Drummond, Reference Rambaut and Drummond2007). The trees were summarized after removing burn-in (600 trees). Maximum likelihood analysis (ML) was performed in PHYML 2.4.4. (Guindon and Gascuel, Reference Guindon and Gascuel2003), with the GTR+Г+I model and parameters estimated from the data; bootstrap values were calculated for 1000 replicates. Resulting trees, including Cryptosporidium serpentis as an outgroup [AF093499], were visualized using TreeView 1.6.6. (Page, Reference Page1996).
The GenBank accession numbers of the sequences included in the phylogenetic analyses
RESULTS
Microscopy
Altogether, 47/81 (58·0%) of turtles were infected with Haemogregarina sp. (Table 3). The highest prevalence occurred in E. orbicularis, where 13/15 turtles (86·7%) were parasitized. Among turtles of the genus Mauremys, 8/17 M. caspica (47·1%), 2/8 M. leprosa (25·0%), and 24/41 M. rivulata (58·5%) were infected with Haemogregarina sp. Parasitaemia in infected individuals reached 0·08–1·36%. We observed four main developmental stages as reported for Haemogregarina in turtle hosts by Telford (Reference Telford2009), namely intraerythrocytic trophozoites, premeronts, meronts and gamonts (see Fig. 1), including their intermediate forms. Gamonts were the most frequently observed, followed by premeronts and trophozoites. Meronts were the rarest (Table 4).
Endogenous life stages of Haemogregarina stepanowi from Emys orbicularis (a–d), Mauremys caspica (e–h, p), M. rivulata (i–l), and M. leprosa (m–o), all in the same scale. Trophozoites contain numerous vacuoles (a, e, i, m), premeronts possess the nucleus in a central position and lack the vacuoles (b, f, j, n), meronts of variable sizes possess various number of nuclei (c, g, k), gamonts are curved in a capsula, with nucleus located at the bend (d, h, l, o), premeronts of different developmental stages in a single erythrocyte of M. caspica (p). Scale bar 10 μm.
Numbers of collected turtles and prevalence of infection; m – males, f – females, juv – juveniles, nd – not detected
Measurements (μm) of Haemogregarina species reported in literature and H. stepanowi isolates from four host turtles in this study; na – data not available
Trophozoites were the smallest life-stages. They were slightly curved in shape with a nucleus located close to one end, and their cytoplasm contained a large number of vacuoles (Fig. 1a, e, i and m). Encapsulated premeronts were more elongated, slightly curved, and, compared with trophozoites, they lacked vacuoles. Their nuclei were usually located in the central position (Fig. 1b, f, j and n). Erythrocytic meronts containing multiple nuclei were rarely found (Fig. 1c, g and k). The prevailing gamonts laid in erythrocytes in a bean-shaped capsule. They were recurved, and with the majority showing a central brightening (Fig. 1d, h, l and o). The stained deep purple nucleus was located near or directly at the bend of the parasite. Young forms of gamonts had nuclei with stranded chromatin, but in mature forms the nucleus shape was dependent on its position. Infected erythrocytes were wider and their nucleus was displaced to the edge of the cell; a few erythrocytes contained two gamonts, two premeronts (Fig. 1p), or a gamont together with a premeront. For each host species, the measurements of all developmental stages are reported in detail in Table 4. The observed morphology of all forms fitted well the morphology of H. stepanowi as described by Danilewsky (Reference Danilewsky1885) and Reichenow (Reference Reichenow1910), thus our isolates were considered to represent this species. Additionally, a considerable overlap within morphological traits further suggested the conspecificity of isolates from all four western Palaearctic hard-shelled freshwater turtle species.
Phylogenetic analyses
PCR analysis offered the same sensitivity as microscopy; all microscopically Haemogregarina-positive samples yielded corresponding PCR products. 18S rDNA sequences were obtained from 19/47 Haemogregarina-positive samples, the lengths of which ranged from 737 and 1421 bp. We used the 13 longest sequences for subsequent analyses. Alignment along the 1085 bp overlapping region was identical for all samples (for p distances between studied taxa, see Table 5). Four sequences of H. stepanowi (one for each host species) were included in the phylogenetic analyses. Resulting phylogenetic trees provided identical topology for both BI and ML analyses (Fig. 2). Analysed haemogregarines formed a monophyletic cluster with three main branches: (i) a clade of sequences of Dactylosoma ranarum as a sister taxon of Babesiosoma stableri; (ii) a large clade of Hepatozoon sp. sequences, divided into two well-separated groups, one being composed of mammalian species of Hepatozoon, and a second formed by Hepatozoon species from reptiles and amphibians; and (iii) a clade consisting of sequences of Haemogregarina. The Nearctic species, H. balli, represented a sister taxon to our four isolates of H. stepanowi, while isolates of Haemogregarina-like parasites from African hinged terrapins formed a sister clade to H. balli and H. stepanowi. All nodes were well-resolved and highly supported. Our analyses revealed the evident conspecificity of all our isolates from E. orbicularis and Mauremys spp. and their affiliation to the genus Haemogregarina together with H. balli and with haemogregarines from African hinged terrapins.
Bayesian inference phylogenetic tree of 18S rDNA sequences (1207 bases) of haemogregarines. The tree is rooted with Cryptosporidium serpentis. Numbers at the nodes show posterior probabilities under BI and bootstrap values for ML, respectively. Maximum posterior probabilities and bootstrap supports 1·0 or 100%, respectively, are marked with asterisk (*).
P distances based on total of 1207 positions in the final dataset. The numbers of base differences per site from estimation of net average between groups of sequences are shown. Standard error estimate(s) are shown above the diagonal. The analysis involved 17 nucleotide sequences. Codon positions included were 1st+2nd+3rd+Noncoding. All ambiguous positions were removed for each sequence pair
DISCUSSION
Apicomplexans are among the most tangled organisms in respect to their phylogenetic reconstructions. Their perceived evolutionary relationships can be distorted in some instances by the lack of distinct morphological features at the light microscopy level. Our knowledge of the phylogeny of Apicomplexa has considerably increased with the expansion of molecular methods based on a variety of markers. However, though medically and veterinary important groups of Apicomplexa have become frequently studied via these methods, others have been largely overlooked. With the exception of the genus Hepatozoon, such neglect applies to haemogregarines of the suborder Adeleorina.
The type species of the genus Haemogregarina, H. stepanowi, has been frequently studied, however, without emphasis on its proper generic allocation, genealogy, and phylogeny with respect to other haemogregarines. Moreover, it has also been formerly reported from various Nearctic turtles, such as Emydoidea blandingii, Chrysemys picta, and C. serpentina (Hahn, Reference Hahn1909; Roudabush and Coatney, Reference Roudabush and Coatney1937). The latter is the host of H. balli – the best studied of the Haemogregarina. Our phylogenetic analyses based on partial 18S rDNA sequences confirmed the close relationships between H. stepanowi and H. balli, however, always as two individual species. Both applied methods – Bayesian inference and maximum likelihood – confirmed their affiliation to the genus Haemogregarina together with three studied isolates of haemogregarines from African hinged terrapins.
The size and other morphological traits of developmental stages of Haemogregarina available in the peripheral blood of all four studied Palaearctic turtles considerably overlap, possessing low variability. They correspond well with the morphology of the appropriate stages of H. stepanowi as described by previous authors (Danilewsky, Reference Danilewsky1885; Reichenow, Reference Reichenow1910; Telford, Reference Telford2009). However, slight morphological differences are present. In particular, haemogregarines found in M. leprosa resembled small ‘endoglobular’ (cylindrical with round edges, possessing a nucleus composed of a large amount of chromatin) and ‘vermicular’ (recurved with a nucleus located at the bend) forms of H. bagensis as reported by Billet (Reference Billet1904), Ducloux (Reference Ducloux1904) and Laveran and Pettit (Reference Laveran and Pettit1909). This necessitates comparison of H. stepanowi with H. bagensis, since the latter was considered by Siddall (Reference Siddall1995) as a valid species. The morphology of the individual stages is very similar and their mutual comparison is also made difficult by the uneven quality of old descriptions. Only Laveran and Pettit (Reference Laveran and Pettit1909) provided dimensions of premeronts, gamonts and erythrocytic meronts of H. bagensis comparable to those of H. stepanowi (Table 4). The size of premeronts of the two haemogregarines is slightly different, H. bagensis being more slender. A complication arises in gamonts, because Laveran and Pettit (Reference Laveran and Pettit1909) did not consider the recurving of the parasite within the erythrocyte. Nevertheless, taking this into account, their size is similar. The dimensions of erythrocytic meronts vary depending on the number of nuclei; 4–16 nuclei being reported in H. bagensis and 2–24 nuclei being reported in H. stepanowi. Therefore, the size of meronts is not a reliable tool for comparison among the species of haemogregarines. We suppose that slight differences in morphology of blood stages (Table 4) might be attributed to their intraspecific variability, to the various phases of infection, to the influence of different host species, and to the different measuring methods used by former authors.
The prevalence of individual stages in the bloodstream depends on the duration of infection, when long-term infection can be characterized by a higher number of gametogonic and merogonic forms, and by fewer trophozoites (Mihalca et al. Reference Mihalca, Achelaritei and Popescu2002). The old infections are characterized by prevailing gamonts in the peripheral blood. This trait could explain the absence of some stages in the blood of M. leprosa. Both haemogregarines were reported from species of aquatic turtles with partly sympatric distribution (Iverson, Reference Iverson1992; Fritz and Havaš, Reference Fritz and Havaš2007). Moreover, the developmental stages of H. bagensis were also found in the leech Placobdella costata (reported also as Placobdella catenigera or Haementeria costata), the same definitive host and vector for H. stepanowi (Brumpt, Reference Brumpt1904; Bielecki et al. Reference Bielecki, Cichocka, Jabłoński, Jeleń, Ropelewska, Biedunkiewicz, Terlecki, Nowakowski, Pakulnicka and Szlachciak2012). It would also be interesting to compare the developmental stages in the organs and tissues of the leech vectors, which, however, were not available for this study.
We revealed consistency in the comparison of the 18S rDNA sequences of our isolates, when homologous sequences were identical in haemogregarines from all four turtle host species. Based on our analyses, we assume that a haemogregarine common for E. orbicularis, M. caspica, M. leprosa and M. rivulata is conspecific with H. stepanowi. Nevertheless, we cannot exclude the possibility that each turtle species could be parasitized by more than one haemogregarine species that were not detected in this study (e.g. M. leprosa by H. bagensis). Apicomplexans had already emerged in the Precambrian era and all currently known species are obligatory intracellular parasites (Levine, Reference Levine1988; Perkins et al. Reference Perkins, Barta, Clopton, Pierce, Upton, Lee, Leedale and Bradbury2000). It is believed that their ancestors evolved in invertebrate hosts. Phylogenetic studies suggest that extant species possessing heteroxenous life cycles evolved from monoxenous ancestors parasitizing invertebrates (Barta, Reference Barta1989; Kopečná et al. Reference Kopečná, Jirků, Oborník, Tokarev, Lukeš and Modrý2006). A reverse scenario of secondary simplification of complex development is less likely (Landau, Reference Landau1974). Notably, the vector-transmitted apicomplexans retain their sexual development in the invertebrate host. This trait argues for a low host specificity of haemogregarines regarding their vertebrate host – turtles. We conclude that the presence of this parasite is likely to be strictly linked to the vector and definitive host – the leech.
ACKNOWLEDGEMENTS
We would like to thank M. Hostovský, D. Jandzik, L. Kadlčík, M. Kautman, J. Kopečná, P. Mikulíček and D. Modrý for assistance in the field. We also thank Christopher Mark Steer for language correction, and to both reviewers for their effort and improvement of our manuscript.
FINANCIAL SUPPORT
This work was supported by the Czech Science Foundation (project P506/11/1738), by the project ‘CEITEC – Central European Institute of Technology’ (CZ.1.05/1.1.00/02.0068) provided by the European Regional Development Fund, and by the IGA VFU grant (project No. 11/2012/FVHE). IP was also supported by the Operational Programme ‘Education for Competitiveness’ project CZ.1.07/2.3.00/30.0014 from the European Social Fund.
