Phylogenetic analyses reveal that Schellackia parasites (Apicomplexa) detected in American lizards are closely related to the genus Lankesterella: is the range of Schellackia restricted to the Old World?

Background Species of Schellackia Reichenow, 1919 have been described from the blood of reptiles distributed worldwide. Recently, Schellackia spp. detected in European and Asian lizards have been molecularly characterised. However, parasites detected in American lizard hosts remain uncharacterised. Thus, phylogenetic affinities between the Old and New World parasite species are unknown. Methods In the present study, we characterised morphologically and molecularly the hemococcidian parasites (sporozoites) that infect three lizard hosts from North America and two from South America. Results In total, we generated 12 new 18S rRNA gene sequences of hemococcidian parasites infecting New World lizard hosts. By the microscopic examination of the smears we identified Schellackia golvani Rogier & Landau, 1975 (ex Anolis carolinensis Voigt) and Schellackia occidentalis Bonorris & Ball, 1955 (ex Uta stansburiana Baird & Girard and Sceloporus occidentalis Baird & Girard) in some samples, but the phylogenetic analysis indicated that all 18S rDNA sequences are distant from Schellackia species found in Old World lizards. In fact, the hemococcidian parasites detected in the New World lizards (including S. occidentalis and S. golvani) were closely related to the genus Lankesterella Labbé, 1899. Consequently, we suggest these two species to be included within the genus Lankesterella. Conclusions Life history traits of hemococcidian parasites such as the type of host blood cells infected, host species or number of refractile bodies are not valid diagnostic characteristics to differentiate the parasites between the genera Schellackia and Lankesterella. Indeed, lankesterellid parasites with a different number of refractile bodies had a close phylogenetic origin. Based on the phylogenetic results we provide a systematic revision of the North American hemococcidians. Our recommendation is to include the species formerly described in the genus Schellackia that infect American lizards into Lankesterella (Lankesterellidae) as Lankesterella golvani (Rogier & Landau, 1975) n. comb and L. occidentalis (Bonorris & Ball, 1955) n. comb. Electronic supplementary material The online version of this article (10.1186/s13071-017-2405-0) contains supplementary material, which is available to authorized users.


Background
Lankesterella Labbé, 1899 [1] and Schellackia Reichenow, 1919 [2] are two genera of haemococcidian parasites of independent evolutionary origin that are nested within a paraphyletic Eimeriidae [3]. The evolutionary novelty in the life-cycle of parasites in these genera is the participation of a blood-sucking vector that exerts a mechanical role in the transmission between hosts. Therefore, the infective stages of the protozoans remain dormant in the transmitter without undergoing any development or modification [4]. Thus, at least in lizard hosts, transmission is accomplished by predation on the infected invertebrate (but see [5]). Some authors consider Lainsonia Landau, 1973 [6] as a third genus of hemococcidia that undergoes sporogony within reticuloendothelial host cells of the liver, spleen, lungs, kidney, and capillaries of the brain [4] in South American lizard host species [4,7]. However, other authors prefer to consider Lainsonia as a synonym of Schellackia based on the common characteristic of their life-cycles, i.e. the presence of eight sporozoites in the oocyst [4,8].
The genus Schellackia was originally described in European lizard hosts [2]. So far, species of Schellackia have been described infecting frog and lizard hosts from Europe, America, Asia and Africa [9][10][11][12][13][14]. Schellackia sp. was reported to infect the scincid lizard Egernia stokesii in Australia [15] but the identification of these parasites remains to be molecularly confirmed. The second hemococcidian genus, Lankesterella, was erected for a species infecting frogs [1]. Later, species of this genus were described infecting birds and lizards [3,[16][17][18][19][20][21][22][23]. The main morphological difference used to classify the hemococcidia into Lankesterella or Schellackia has been the characteristics of the oocyst during the endogenous development of the parasite. The presence of oocysts, normally in the lamina propia of the intestine, with eight naked sporozoites surrounded by a soft-walled oocyst, has been considered a diagnostic character for the genus Schellackia. In contrast, the presence of 32 or more sporozoites is the diagnostic characteristic for the genus Lankesterella [4]. However, the latter was true for Lankesterella spp. described in anuran hosts [8,18,21], since no endogenous development of Lankesterella spp. infecting lizard hosts has been described so far. The apicomplexan genera Lankesterella and Schellackia were largely believed to form a monophyletic clade within the family Lankesterellidae [4,8]. However, phylogenetic analyses revealed they have an independent evolutionary origin [3].
No information on the molecular diversity and the phylogenetic affinities of the hemococcidia that infect New World lizard hosts is available. Therefore, we have sampled different populations of New World lizards belonging to the families Phrynosomatidae (genera Uta Baird & Girard and Sceloporus Wiegmann), Iguanidae (genus Dipsosaurus Hallowell), and Liolaemidae (genera Liolaemus Wiegmann and Phymaturus Gravenhorst) to obtain molecular data that allowed the study of the evolutionary affinities between hemococcidia detected in Oldand New World host lizards. In addition, we present morphological data on the sporozoites found infecting the blood of the American lizard hosts studied here.

Sampling methods
We gathered several blood samples with hemococcidian parasites that were collected in different localities in North and South America. From California, blood was obtained from six infected Uta stansburiana hesperis Richardson (Phrynosomatidae) from Corn Springs (South California) and 12 from Los Baños (West California), 19 Sceloporus occidentalis bocourtii Boulenger (Phrynosomatidae) from Santa Cruz (West California), two S. occidentalis biseriatus Hallowell from high elevation (1800 masl) from the Sierra Nevada mountains (CA, USA) and one Dipsosaurus dorsalis Baird & Girard (Iguanidae) from Boyd Canyon, Riverside County (South California). One infected sample of Liolaemus pictus Müller & Hellmich (Liolaemidae) from Huinay (Chile) was also included in the study. All authors captured the reptiles using a noose tied to the end of a fishing pole (e.g. [3,14]). All animals were processed and released on the same spot where they had been captured.
Blood samples were collected at the base of the tail using individual sterile needles (BD Microlance, Huesca, Spain). In the case of male lizards, the hemipenes bulges were always avoided to prevent harming the individuals. The area where the blood sample was obtained was previously cleaned with ethanol 96%. Blood droplets were collected with the help of Na-heparinized microhematocrit tubes (BRAND, Wertheim, Germany). Blood samples were preserved in two different ways. First, thin blood smears were made for microscopic examination of infection. Blood smears were air-dried and fixed for 5 min in absolute methanol. All blood smears were stained with Giemsa stain (1/10 v/v) for 40 min. Slides were examined by the same operator for hemoparasites counting to 15,000 red blood cells (RBC) following Merino & Potti [29]. We made microphotographs of the sporozoites observed under light microscopy and measured the length and the width of the sporozoites using the MBruler 5.0 free software (http://www.markus-bader.de/MB-Ruler/index.php) [3,14]. Thereafter, we performed an ANOVA analysis to compare the size of the sporozoites that we detected in the two known hosts for Schellackia occidentalis: Sceloporus occidentalis and U. stansburiana. Secondly, the remaining blood (less than 20 μl) was preserved on Whatman FTA Classic Cards (GE Healthcare UK Limited, Buckinghamshire, UK). These cards allowed us to perform DNA extraction later [3,14,23].
D' Amico & Aguilar [30] reported the presence of hemococcidian parasites in the blood of Phymaturus payuniae Cei & Castro (Liolaemidae) from the cold desert in Patagonia, Argentina. These authors provided a tissue sample (tail tip) of one infected individual preserved in 70% ethanol for this study. Tail tissue samples allow the proper extraction of the apicomplexan DNA from the host tissue [31]. Similarly, we microscopically analysed the blood smear corresponding to the infected polychrotid lizard, Anolis carolinensis, reported in a previous study [23]. The infected lizard was one individual recently imported from Florida [23] that presented an intensity of infection of 21/15,000 RBC. In the present study, we compared the morphology of parasites found in the American lizards of genera Anolis [23], Sceloporus, Uta, Dipsosaurus, Phymaturus [30] and Liolaemus with previous species described infecting American lizard hosts using the morphological data available in Telford [4].

Molecular methods
We extracted genomic DNA from the blood preserved on FTA cards following the protocol described by Megía-Palma et al. [14]. The DNA was then purified using the NZYGelpure kit (NZYTech, Lda., Lisbon, Portugal). Partial amplification of the 18S ribosomal RNA gene sequence was performed using the primer set BT-F1/hep1600R or BT-F1/EimIsoR1 or BT-F1/EimI-soR3 (Table 1). These primers were previously used to amplify other coccidian species [3,23]. PCR reactions (total volume of 20 μl) contained between 20 and 100 ng of the DNA template. Supreme NZYTaq 2× Green Master Mix (NZYTech, Lda.) and 250 nM of each primer were used. Using a Veriti thermal cycler (Applied Biosystems, Foster City, CA, USA), reactions were run using the following conditions: 95°C for 10 min (polymerase activation), 40 cycles at 95°C for 30 s, annealing temperature at 58°C for 30 s, 72°C for 120 s, and a final extension at 72°C for 10 min. All amplicons were recovered from agarose gels and subjected to direct sequencing using an ABI 3730 XL automated sequencer (Applied Biosystems). Amplicons whose chromatograms yielded double peaks were cloned using NZY-A Speedy PCR cloning kit (NZYTech, Lda.). Plasmids were purified using NZYMiniprep kit (NZYTech, Lda.).
Due to the weak signal obtained by PCR in some samples, we could only sequence 22 of the 41 amplicons that we obtained. The twelve different DNA sequences (18S rRNA gene) obtained in the present study were aligned together with other 98 sequences obtained from GenBank. The alignment was performed using PROBCONS (http:// toolkit.tuebingen.mpg.de/probcons). Poorly aligned positions and divergent regions of the alignment were suppressed using the GBlocks program [32] selecting the following options: minimum number of sequences for a conserved position = 56; minimum number of sequences for a flank position = 56; maximum number of contiguous nonconserved positions = 10; minimum length of a block = 5; and allowed gap positions = with half. The final alignment contained 1441 positions and 110 sequences. The substitution model GTR + I + G was selected using jModeltest 2.1.4 [33] to perform the Bayesian analysis with MrBayes 3.2.6 software [34][35][36]. This analysis consisted of 2 runs of 4 chains each with 2,000,000 generations per run and a sampling interval every 100 generations. After a 'burn-in' of 500,000 generations was applied and 30,000 trees were obtained for consensus tree. The final standard deviation of the split frequencies was lower than 0.01. In addition, the alignment was also analysed using a maximum likelihood inference (PhyML 3.0 software) [37], using the substitution model listed above. The subtree pruning and regrafting (SPR) tree rearrangement option were selected, and a Bayesian-like transformation of aLRT (aBayes) was used to obtain the clade support [38]. Both trees were rooted with the family Sarcocystidae because this taxon is well established as a distinct but closely related family to the Eimeriidae [8,14].

ZooBank registration
To comply with the regulations set out in article 8.

Morphological identification
In total, 99% (207/209) of the observed sporozoites had light blue refractile bodies (RB) (Fig. 1). The presence of RB is a morphological characteristic compatible with the morphology of the sporozoites in the hemococcidian genera Lankesterella and Schellackia [3,4]. The number of RB in each of the parasite haplotypes is shown in Table 2. We observed 5% (11/209) of the sporozoites free in the plasma, whereas the remaining 95% (198/ 209) of the sporozoites were found infecting the cytoplasm of host blood cells. More, specifically 65% of the sporozoites infected red blood cells, and 30% leukocytes. Host blood cells for the sporozoites of each parasite haplotypes are indicated in Table 2. The hemococcidian parasite found in the blood of Anolis carolinensis was morphologically identified as Schellackia golvani. The mean size of the sporozoites was 8.4 × 3.9 μm and all of them were observed in the cytoplasm of polymorphonuclear leukocytes (21/21) with a single RB.

Molecular results
We performed DNA extraction from the 41 samples where we found hemococcidians infecting blood cells in blood smears. We only sequenced 22 of the 41 (53%) amplicons obtained. In total, we found 12 new haplotypes of hemococcidian parasites infecting New World lizard hosts ( Table 3). The new sequence data are available in the GenBank database under the accession numbers MF167544-MF167555 (see also Additional file 1: Table S1). In North American hosts, we found six new haplotypes of the 18S rRNA gene of hemococcidian parasites infecting Phrynosomatidae and four infecting Iguanidae whereas, in South American hosts, we found two haplotypes infecting Liolaemidae lizards. The prevalence of each parasite haplotype in the host species sample is shown in Table 2.
Bayesian inference and maximum likelihood phylogenetic analyses produced trees with an almost identical topology. All new haplotypes grouped together with Lankesterella parasites (Fig. 2). Ten of the 12 new haplotypes, including parasite sequences detected in U. stansburiana and Sceloporus occidentalis, were included with high phylogenetic support (≥ 95%) in the same clade with sequences of Lankesterella parasites (i.e. Lankesterella sp. Ae-Lk and Lankesterella sp. Lank_anocar) detected in the North American host species, Anolis carolinensis, and in the Mediterranean host species, Acanthodactylus erythrurus (clade A). The group (named clade B) formed by L. valsainensis (detected in Eurasian avian hosts) and the type-species of the genus Lankesterella, L. minima (detected in Eurasian amphibian hosts), was a sister group to clade A (support < 80%, Fig. 2). In addition, clade C that grouped cloned haplotypes DD1 and DD4 detected in the desert iguana, D. dorsalis, from California, formed a third monophyletic clade within the family Lankesterellidae. This last clade was the sister group to the clade formed by the groups A and B with a statistical support of 0.78 by Bayesian inference and 95% by maximum likelihood (Fig. 2). Additionally, clade D contained all previously known sequences of Schellackia parasites that were molecularly characterised from blood samples of Asian and European lizard hosts in previous studies [3,14].

Discussion
Three Schellackia species are known to infect American lizard hosts: S. landaue, S. occidentalis, and S. golvani [4]. We detected parasites that are biologically and morphologically compatible with S. occidentalis and S. golvani. On the one hand, the sporozoites observed infecting the blood cells of Anolis carolinensis are morphologically (i.e. mean length and width [4]) compatible with the original description of S. golvani. Besides, the sporozoites were found infecting exclusively polymorphonuclear leukocytes, which is a unique characteristic of S. golvani [4,10]. On the other hand, parasites found in the blood samples of the lizard hosts Sceloporus occidentalis and Uta stansburiana are likely the same parasite formerly described as Schellackia occidentalis.
The size (i.e. mean length and width) of the sporozoites found infecting the blood cells of Sceloporus occidentalis was compatible with that originally described for the sporozoites of Schellackia occidentalis (Table 4) although the number of RB was not always consistent with the original description. However, previous experimental studies demonstrated that RB are dynamic structures within the Eimeriorina that can shift in size and number during the development of the sporozoite and fluctuate in number even in the sporozoites infecting an individual host [4,40,41].
Considering that we included sequences of S. golvani and S. occidentalis in the phylogenetic analyses, it was unexpected that none of them were closely related to the family Schellackiidae (clade D) [23]. To our knowledge, this phylogenetic study is the first to compare hemococcidian parasites from American, European and Asian lizard species. We found strong phylogenetic support to conclude that the hemococcidian haplotypes detected in American lizard hosts are closely related to the genus Lankesterella and distant to the genus Schellackia. Indeed, lankesterellids detected in North and South American lizards were more closely related to Lankesterella sp. Ae-Lk1 and Lankesterella sp. Lank_anocar, previously detected in Old and New World lizards [3,23] (clade A) than to Lankesterella spp. detected in European bird and amphibian hosts (clade B). The phylogenetic affinity of clade C with the remaining clades within the Lankesterellidae showed medium to high phylogenetic support and, hence, it provides molecular evidence for a monophyletic origin of the genus Lankesterella.
Based on the phylogenetic affinities achieved here and the phylogenetic support for the monophyletic origin of the genus Lankesterella, our recommendation is to reclassify Schellackia golvani as Lankesterella golvani (Rogier & Landau, 1975) n. comb., and S. occidentalis as L. occidentalis (Bonorris & Ball, 1955) n. comb. Lankesterella occidentalis n. comb. seems to be composed of a diverse complex of haplotypes from parasites that infect different host species in North America (see [4]). Indeed, we characterised six different 18S RNA haplotypes in only two different host species of the family Phrynosomatidae. However, the specificity at the host genus level of American lankesterellids was high because none of the parasite haplotypes detected in Uta stansburiana was found in Sceloporus occidentalis (and vice versa) despite sampling ranges of the hosts and parasites overlapped (e.g. in California). This result suggests that L. occidentalis n. comb. may be a complex group of cryptic species similar to previous cases of parasitic species complexes [42]. Indeed, the significantly longer and narrower sporozoites of L. occidentalis n. comb. that infect U. stansburiana, in comparison to the sporozoites that infect S. occidentalis hosts, suggest that Schellackia     Table S1. The scale-bar indicates the number of substitutions per site occidentalis was formerly described based on a complex of species. Future description of the endogenous development of the parasite in the walls of the intestine of its host may provide new species designations in order to split L. occidentalis n. comb. We provide evidence that lankesterellids are a molecularly diverse group of parasites in lizard hosts, at least in the Americas. Indeed, we found 12 different haplotypes of these parasites in five host lizard species. The absence of Schellackia haplotypes in this sample of New World lizard hosts suggests that the genus Schellackia might be restricted to Old World lizards. Nevertheless, further studies should provide a wider sampling including other localities and host species in North and South America to provide additional evidence on the absence of Schellackia spp. in the Americas. This might have important conservation implications because the eventual introduction of a novel pathogen into America (i.e. Schellackia) through human activity (e.g. pet trade, see [43]) could have a dramatic negative effect on native lizards.
One of the host lizards included in this study, the desert iguana Dipsosaurus dorsalis from Boyd Deep Canyon (CA), was infected by at least four Lankesterella haplotypes (i.e. DD1-DD4). Two possible scenarios may explain the presence of these four clonal parasite haplotypes in a single host individual. On the one hand, these haplotypes could be copies of the same Lankesterella species. Although all known eukaryotes have several clonal copies of the 18S rRNA gene, some apicomplexan species transcribe different copies of this gene along their life-cycle [43]. This is the case for Plasmodium vivax (Apicomplexa: Haemosporidia) [44], and it has also been suggested for the coccidian genus Hepatozoon [45]. In our sample, the absence of multiple peaks in the chromatograms from the uncloned sequences does not rule out the possibility that multiple copies of the 18S gene were also present. On the other hand, the phylogenetic position of haplotypes D1 and D4 as the sister clade to groups A and B only partially support this assumption. In this sense, the presence of different haplotypes of the 18S rRNA gene in one individual might reflect the presence of a mixed infection with different parasite species. Indeed, many cryptic species may occur within Apicomplexa [42,46] because hematic stages of unicellular parasites may lack enough diagnostic morphological traits to distinguish easily among species when mixed infections occur [47].
Previous studies on vector-parasite evolution suggest that vector diversity and vector-protozoan specificity may be high in avian models [48]. If that also occurs in lizards, the high genetic diversity of hemococcidians detected in the present study might be related to the high diversity of their vectors, indicating high vector-parasite specificity. A recent publication described high genetic diversity of Lutzomyia (Diptera: Psychodidae) sand flies in California [49]. Measures of genetic differentiation indicated genetic differentiation of this vector between sites that were approximately 0.5-3.8 km distant. Lutzomyia sand flies, together with some mite species, have been described as natural vectors for several diseases of lizards caused by protozoans including hemococcidia [4,50,51]. Thus, the presence of a high genetic diversity of transmitters would favour diversification of host-parasite associations and, hence, diversification of hemococcidian parasites in America.

Conclusions
Life history traits of hemococcidian parasites such as the type of host blood cells infected, host species or number of refractile bodies are not valid diagnostic characteristics to differentiate the parasites between the genera Schellackia and Lankesterella. Indeed, Lankesterella spp. with a different number of refractile bodies had a close phylogenetic origin. Based on the phylogenetic results we provide a systematic revision of the North American hemococcidians. Our recommendation is to include the species formerly described in the genus Schellackia (i.e. S. golvani and S. occidentalis) that infect American lizards into Lankesterella (Lankesterellidae).

Additional file
Additional file 1: Table S1. GenBank accession numbers for all sequences included in the phylogenetic analyses. Numbers in column C identify collapsed clades in Fig. 2 (DOCX 24 kb)