Introduction

Piroplasm species are tick-borne parasitic protozoa which are differentiated into the genera Theileria and Babesia. A number of these parasites are highly pathogenic for cattle, sheep and goats; and the diseases emerging from these infections are referred to as theileriosis and babesiosis, respectively. The economic losses due to theileriosis and babesiosis are enormous in tropical and subtropical areas (Mehlhorn and Schein 1984; Mehlhorn et al. 1994).

18S RNA genes have been successfully applied to identify and classify several previously unknown Theileria and Babesia parasites (Persing et al. 1993, 1995; Quick et al. 1993; Thomford et al. 1994; Herwaldt et al. 1997; Katzer et al. 1998; Gubbels et al. 2000; Schnittger et al. 2000a). Furthermore, the phylogenetic classification of cattle-infecting piroplasms by the analysis and comparison of 18S RNA genes has been shown to correspond with traditional taxonomy and provides additional refined information on their evolutionary relationship.

It was shown by molecular phylogeny of 18S RNA genes that the highly pathogenic cattle parasites T. annulata and T. parva, which share the ability to transform their host cells, are evolutionarily closely related and that parasites of the T. buffeli group (including T. orientalis, T. buffeli, T. sergenti) have a monophyletic origin and cannot presently be distinguished into different species (Allsopp et al. 1994; Chae et al. 1999; Gubbels et al. 2000, 2002). In contrast, cattle Theileria species (T. mutans, T. velifera) and Babesia species (B. bovis, B. bigemina, B. divergens) have been shown to be of individual origin and show no close relation to each or to other cattle piroplasms (Allsopp et al. 1994; Chae et al. 1999; Gubbels et al. 2000; Homer et al. 2000).

In comparison to cattle piroplasms, piroplasm species infecting sheep and goats have been somewhat neglected but increasingly attract attention due to their economic importance being higher than cattle-infecting piroplasms in some countries (e.g. China, Iran; Luo and Yin 1997; Schnittger et al. 2000a; Ahmed et al. 2002; Bai et al. 2002; Yin et al. 2002, 2003).

The objective of the current study was to elucidate the evolutionary relationship of sheep/goat piroplasm species on the basis of their 18S RNA gene sequences. The sequences of the 18S RNA genes of a panel of isolates of the species T. lestoquardi, T. separata, T. ovis, B. motasi, B. crassa, B. ovis and two recently reported Theileria species originating from China were deduced. Establishment of a phylogenetic tree allowed integration and definition of sheep/goat piroplasm species in comparison with cattle piroplasms and piroplasms infecting other hosts.

Materials and methods

Parasite isolates

Information on parasite stocks used in this study and their vertebrate host, geographic origin and accession number under which the corresponding 18S RNA gene sequence was submitted to the GenBank data base is listed in Table 1. Parasite stocks isolated from sheep include Theileria ovis (Sudan), T. ovis (Turkey), Babesia motasi (Ameland), B. crassa (Iran), Babesia sp. (Turkey) and B. motasi (Turkey). Parasite stocks Theileria sp. G4 (Tanzania) and Theileria sp. G6 (Tanzania) were isolated from goats. Parasite stocks isolated after experimental infection of sheep with ticks collected in the field include Theileria sp. (Lintan 2), Theileria sp. (Lintan 3), Theileria sp. (Zhangjiachuan), Theileria sp. (Longde), Theileria sp. (Qinghai), Theileria sp. (Madang) and Theileria sp. (Ningxian), which all originate from Haemaphysalis qinghaiensis ticks, T. separata (South Africa) which originates from Rhipicephalus evertsi ticks, B. motasi (Texel) which originates from H. punctata ticks and B. ovis (Turkey) which originates from R. bursa ticks. All parasite isolates were stored as EDTA–blood stabilates in liquid nitrogen.

Table 1. Isolate, vertebrate host, origin and accession number of 18S RNA genes used for phylogenetic tree constuction. Isolates isolated from sheep or goat are indicated by Sheep or Goat under Host. 18S RNA genes of sheep/goat piroplasms sequenced in this work are given in the Materials and methods. Other 18S RNA gene sequences were obtained from the GenBank database. All 18S RNA gene sequences of listed isolates were used to construct the phylogenetic tree shown in Fig. 1. Sequences which consistently behaved as a single operational unit were combined in clusters/subclusters, as listed here
Full size table

DNA isolation and PCR amplification

Genomic DNA was prepared from blood, using a genomic DNA extraction kit (Qiagen) and following the protocol of the manufacturer; and the amount of isolated DNA was assessed by photometry. The PCR amplification protocol performed for the determination of 18S RNA gene sequences was in brief: 100 ng of genomic DNA was added to a reaction mixture (final volume of 30 µl) containing 1 µM primer A and 1 μM primer B which cover the eukaryotic 18S RNA gene sequence (Medlin et al. 1988), PCR buffer (10 mM Tris-HCl, pH 8.0, 50 mM KCl, 0.01% gelatine), 200 µM each deoxynucleotide triphosphate, 1.5 mM MgCl2 and 1.5 units of Taq polymerase; and the amplification was done in an automatic DNA thermocycler (Eppendorf). The reaction was incubated at 94 °C for 3 min to denature genomic DNA and the thermal cycle reaction programmed for 40 cycles of 1 min at 94 °C, 90 s at 40 °C and 90 s at 72 °C, with a final extension step of 5 min at 72 °C. Samples were held at 4 °C until analysed.

Sequence analysis of 18S RNA genes

After amplification of the complete 18S RNA gene with primers A and B as described above, the generated DNA fragments were either sequenced directly or ligated into pCR 2.1 and transformed into One Shot-competent Escherichia coli, using the TA cloning kit (Invitrogen) according to the manufacturer′s instructions. Transformed bacteria were plated and qualified clones selected. A restriction digest was performed (with EcoRI), run on a 1% agarose gel and visualised using an ultra-violet transilluminator to recognise clones containing the desired 18S RNA gene fragment of approximately 1,750 bp. The sequences of coding and non-coding strands of four clones obtained in at least three independent PCR reactions were determined on an automatic sequencer according to standard procedures (Applied Biosystems). The sequencing output was analysed using the accompanying DNA Sequence Analyzer software (Applied Biosystems). Compilation, editing and assembly of multiple sequences generated from each template were performed with the EditSeq and SeqMan analysis program components of the Lasergene software package for Windows (DNASTAR, Madison, Wis.). The nucleotide sequence data for the analysed 18S RNA gene sequences of sheep and goat piroplasm isolates are available in the GenBank database under accession numbers AY262115, AY262119, AY262117, AY262118, AY260174, AY260173, AY260171, AY260172, AY260175, AY262116, AY262121, AY262120, AY260179, AY260180, AY260176, AY260177 and AY260178.

Sequence alignment and phylogenetic analysis

For sequence alignment and phylogenetic analysis, the MegAlign component of the Lasergene program ver. 4.01 (DNASTAR) was used. The MegAlign component performs multiple sequence alignments by the program algorithm of the Clustal method (Thompson et al. 1994) and phylogenetic analysis by the neighbour-joining method (Saitou and Nei 1987; Atteson et al. 1999). The constructed phylogenetic tree (cladogram; Fig. 1) is based on Theileria and Babesia 18S RNA gene sequences determined in our laboratory or obtained from the GenBank database (Table 1). To estimate the reproducibility of the tree, bootstrapping was done by tree reconstruction with random draws of sub-samples. While there was a minor change in the order of clades/clusters observed in the Theileria branch of the tree, the grouping of sequences into individual clusters/clades was in all cases identical, verifying that each cluster/clade represents an operational taxonomic unit (OTU). The phylogeny was established, assuming a biological clock; and the length of each pair of branches in Fig. 1 represents the distance between sequence pairs. The units on the horizontal axis indicate the number of substitution events. The matrix of pairwise sequence divergence and identity is not shown but is referred to in the Results and the Discussion.

Fig. 1.
figure 1

Phylogenetic tree inferred from Theileria and Babesia 18S RNA gene sequences given in Table 1. Sequences combined in clusters/subclusters behave consistently as operational taxonomic units. For reasons of clarity only representative 18S RNA gene sequences of a given cluster/subcluster are shown (see footnote in Table 1). Sheep and goat piroplasm isolates are printed in bold. The scale gives the number of nucleotide exchanges. The line-break (//) marks the major division, as characterised by the highest phylogenetic diversity of 18S RNA genes between Theileria and Babesia parasites (with regard to this definition, Babesia sensu latu parasites B. microti and B. rodhaini belong to the genus Theileria). Clusters are designated by numbers and subclusters by letters

Full size image

Results

18S RNA gene sequences

Within the scope of this work, complete 18S RNA gene sequences were determined for isolates Theileria separata (South Africa), T. ovis (Sudan), T. ovis (Turkey), Theileria sp. G4 (Tanzania), Theileria sp. G6 (Tanzania), Theileria sp. (Lintan 2), Theileria sp. (Lintan 3), Theileria sp. (Zhangjiachuan), Theileria sp. (Longde), Theileria sp. (Qinghai), Theileria sp. (Madang), Theileria sp. (Ningxian), Babesia motasi (Ameland), B. crassa (Iran), Babesia sp. (Turkey), B. motasi (Turkey) and B. ovis (Turkey). Table 1 gives the parasite isolate, vertebrate host, geographic origin and the accession number under which the corresponding 18S RNA gene sequence was submitted to the GenBank database. Additionally, all other 18S RNA gene sequences from Theileria and Babesia parasites currently available in the GenBank database are registered in Table 1.

Phylogenetic tree construction

To integrate the small ruminant-infecting piroplasm parasites into the phylogenetic system of other Theileria and Babesia parasites, a phylogenetic tree was inferred from all 18S RNA sequences by neighbour-joining analysis (Table 1, Fig. 1). The robustness of the tree was tested by performing tree reconstructions with random draws of sub-samples. Some minor changes were observed in the ordering of clusters/clades within the Theileria branch of generated trees. Given the relatively short branch length in this part of the tree, this is not surprising and was also observed by Gubbels et al. (2002). The observed change of cluster ordering is disregarded in the following because it is of no further relevance for the presented results. In contrast to Theileria sequences, the branching order of Babesia sequences stayed identical in all trials. In accordance with the high robustness of the Babesia part of the tree, the branch lengths were significantly enhanced in segregating Babesia, compared with Theileria 18S RNA sequences (Fig. 1). It was however essential that, in all generated trees, the grouping of sequences into individual clusters/clades stayed identical, verifying that each cluster/clade represents an OTU. Accordingly, the Theileria and Babesia 18S RNA sequence clusters, as presented, behaved as single OTUs and corresponded to clusters/clades described by other authors for Theileria and Babesia 18S RNA sequences (Allsopp et al. 1994; Chae et al. 1999; Gubbels et al. 2000, 2002; Homer et al. 2000). For reasons of clarity, only representative selected 18S RNA gene sequences (see footnote in Table 1) are shown in several clusters/clades/OTUs of the phylogenetic tree in Fig. 1.

Molecular phylogeny of sheep/goat Theileria and Babesia species

The affiliation of an isolate to a piroplasm species based on the segregation of its 18S RNA gene sequence within a phylogenetic tree is somewhat subjective (e.g. cluster 5a contains four different species, while cluster 6 contains one species only). To allow a more objective decision whether 18S RNA genes of two different isolates belong to the same or to separate species, a decision procedure was applied, based on the calculated matrix of pairwise sequence identity (see Materials and methods). The identity between 18S RNA genes is an estimate of the phylogenetic distance between corresponding isolates. Therefore, it is possible to determine an upper/lower identity limit of 18S RNA gene sequences above/below which corresponding isolates would be affiliated to the same or different species. These identity limits have to be established separately for the genera Theileria and Babesia, because each has a different evolutionary history, as apparent from the different branch lengths of the phylogenetic tree.

The highest identity values of a pair of 18S RNA genes belonging to different species were observed between T. annulata vs T. lestoquardi (99.7%) and B. divergens vs B. canis (97.9%), respectively. Accordingly, isolates were considered to belong to a single species if their identity was found to be higher than 99.7%/97.9% for Theileria/Babesia 18S RNA genes. In contrast, the lowest identity values between 18S RNA gene sequences of isolates of a single species existed between Theileria sp. (Lintan 1) vs Theileria sp. (Ningxian) at 99.3% (both representing Theileria sp. 1 (China)] and between B. bovis (Australia) vs B. bovis (USA) at 96.6%. Therefore, isolates were considered to belong to different species if the identity was found to be lower than 99.3%/96.6% for Theileria/Babesia 18S RNA gene sequences, respectively. On the basis of these limits, it was possible to unequivocally assign all isolates to piroplasm species by applying the presented decision procedure (Table 2). In each case, the biological characteristics of assigned isolates were in accordance with that accepted and published by the known classic taxonomy of piroplasm species.

Table 2. Isolates assigned to piroplasm species by phylogenetic classification. Goat and sheep piroplasm isolates were assigned to different species by systematic classification of the 18S RNA gene, as described in the Materials and methods. Compare Cluster/subcluster data with Fig. 1. n.k. Not known
Full size table

Five different Theileria species can be defined by the outlined phylogenetic classification of 18S RNA genes, each segregating in a single cluster/subcluster (Fig. 1, Table 2). Theileria sp. 1 (China) and Theileria sp. 2 (China) display identical biological characteristics in regard to tick and vertebrate host specificity, morphology and pathogenicity and can at present only be distinguished by molecular approaches. The similar aggregation of the 18S RNA sequences of four T. ovis isolates originating from different geographic regions into one cluster supports the view that T. ovis represents a single entity and does not represent a complex of different parasite species, as some authors suggested (Uilenberg 1981; Leemans et al. 1997).

Three different sheep Babesia species could be distinguished by phylogenetic classification: B. motasi, B. ovis and B. crassa. Furthermore, it could be shown that B. crassa is distributed not only within Iran but also in Turkey: B. crassa (Turkey). In Table 2, all investigated isolates are attached to the different defined species.

Discussion

The present study reports on the comparison and classification of sheep- and goat-infecting Theileria and Babesia piroplasms, based on the structure of their 18S RNA genes. The established phylogenetic tree can be divided in two major divisions, each having a single ancestor. One division represents the Theileria parasites, including the parasites T. equi, B. rodhaini and B. microti (Babesia sensu latu) and the other constitutes the true Babesia parasites (Babesia sensu strictu). Apparently, this division corresponds with biological characteristics: Theileria parasites can be distinguished from Babesia parasites by the presence of a schizont parasite stage and by their trans-stadial transmission by ticks. Both these features must have been a characteristic of a common Theileria ancestor. As trans-stadial transmission is a trait also displayed by some Babesia parasites, it is likely to be inherited by the genus Theileria.

As it is evident from the structure of the generated phylogenetic tree, the genera Theileria and Babesia have different evolutionary histories. In the Theileria parasites, the branch lengths are shortened and accordingly the identity between different species is much lower than in the Babesia parasites. Under the supposition that the evolutionary rate of 18S RNA genes is similar within both groups, this suggests that the genus Theileria developed later and has a shorter evolutionary history than Babesia.

The generally observed higher sequence identity between Theileria 18S RNA sequences compared with those of Babesia may be best explained by an evolutionary process of adaptive radiation. That is the relatively recent and fast spreading of members of this group to different hosts, not allowing the accumulation of species-specific mutations within slowly evolving RNA genes (Tibayrenc 1998). This would furthermore explain why some Theileria parasites, as for example parasites belonging to the T. buffeli group, cannot be clearly separated into different species: they may presently be in a dynamic evolutionary process of diversification in which clear separation of species has not yet been completed (Chae et al. 1999; Gubbels et al. 2000, 2002). This assumption would also be consistent with the finding that the species Theileria sp. 2 (China) with sheep as the vertebrate host and Theileria sp. Serow (Japan) with serow as the vertebrate host show an overlapping sequence diversity of their 18S RNA genes and that T. annulata and T. lestoquardi have speciated evidently very recently, as outlined by some authors (Katzer et al. 1998; Schnittger et al. 2000a, 2000b).

Unfortunately, there is no general consensus on the extent of sequence variation in 18S RNA genes which must be present to consider organisms as different species. However, it would be somewhat subjective to assign isolates to species by mere visual inspection of a phylogenetic tree. In order to objectify the assignment of isolates to species, we decided to establish identity values as decision thresholds by an empirical method (see Results). Due to the dissimilar evolution of Theileria and Babesia, different threshold values were determined for each genus. When the identity of two 18S RNA genes exceeded the established maximum threshold value, the respective isolates were considered as separate species. In contrast, when the identity of two 18S RNA genes was lower than the established minimum threshold value, the respective isolates were considered as similar species. The assignment of isolates to species on the basis of sequence identities was revised with their biological characteristics and compared with the known classic taxonomy. In each case, the sequence data and biological characteristics of the isolates could be reconciled with classic taxonomy, verifying the validity of the applied decision procedure.

The 18S RNA sequence of Theileria sp. 1 (China) isolates [Theileria sp. (Lintan 1), Theileria sp. (Lintan 2), Theileria sp. (Qinghai) Theileria sp. (Madang), Theileria sp. (Ningxian)] all segregate in cluster 1a, together with sequences originating from Theileria isolates of Japanese serow: Theileria sp. Serow (Japan) (Schnittger et. al. 2000a; Gubbels et al. 2002). Thus, sequences of Theileria sp. 1 (China) isolates show an overlapping diversity to that of Theileria sp. Serow (Japan). Both probably descended from a single direct ancestor species; and it seems likely that the parasite was transferred very recently by a compatible tick from the wild Japanese serow to domestic sheep, not yet allowing complete separation into different species. Unfortunately, biological data to support this hypothesis as the species of the transmitting tick have not been reported from Theileria sp. Serow. The very close relationship between both parasites has been already supposed by Gubbels et al. (2002).

The phylogenetic tree placed T. ovis 18S RNA sequences of isolates Theileria sp. G4 (Tanzania), Theileria sp. G6 (Tanzania), T. ovis (Sudan) and T. ovis (Turkey) together in cluster 2a. Until recently, T. ovis was identified mainly by its non-pathogenicity for sheep and goat and it was therefore supposed this species may constitute a complex of different species (Leemans et al. 1997; Uilenberg 1981). The observation that T. ovis isolates originating from three entirely different geographic regions cluster together suggests that T. ovis does represent a single species. Most closely related to T. ovis is the sequence of a Theileria species isolated from sika deer [Theileria sp. CNY1B Sika deer (Japan) in cluster 2b] and both may share a direct ancestor. Interestingly, there is another Theileria parasite which was also isolated from sika deer [Theileria sp. CNY2A Sika deer (Japan), but which clearly represents a different species (cluster 1b) more closely related to Theileria sp. CC3A Serow (Japan) and Theileria sp. 1 (China) in cluster 1a.

The sequence of the sheep piroplasm T. separata is most closely related to that of Theileria sp. Sable (cluster 4b). The finding that T. separata and Theileria sp. Sable are transmitted by the same tick species would provide support for the hypothesis that the parasite was passed from wild sable antelope to domestic sheep. However, the transmitting tick species of Theileria sp. Sable is not known (Thomas et al. 1982).

Apart from T. taurotragi, piroplasm species of cluster 5a are of great economic importance because of their pronounced pathogenicity for domestic animals. The cattle-infecting T. annulata, T. parva and the sheep-infecting T. lestoquardi parasites share the evolutionarily unique ability of host-cell transformation, which must have been developed by a Theileria parasite ancestral to these species, supporting the close relationship of these parasites as shown by the clustering of their sequences. T. lestoquardi seems to be closely related to T. annulata, as suggested by the high identity of their sequences, high antigenic cross-reaction, overlapping distribution patterns and the same transmitting tick species, Hyalomma sp. (Hooshmand-Rad and Hawa 1973; Leemans et al. 1997; Katzer et al. 1998; Schnittger et al. 2000a, 2000b). Interestingly, T. lestoquardi seems to be the only sheep/goat piroplasm which has likely evolved from a cattle-infecting piroplasm species. It is furthermore noteworthy that, except cluster 8, cluster 5a is the only cluster of the genus Theileria harbouring Theileria isolates which have been clearly defined as different species.

Besides Theileria sp. 1 (China), a second parasite pathogenic for sheep and goat has been identified (Yin et al. 2003). The species is designated Theileria sp. 2 (China); and 18S RNA gene sequences of isolates Theileria sp. (Lintan 2), Theileria sp. (Zhangjiachuan) and Theileria sp. (Longde) segregated in cluster 6. Theileria sp. 2 (China) is relatively isolated in the phylogenetic tree and shares with the most closely related Theileria sp. CNY1B Sika deer (Japan) sequence (cluster 2b) an identity of 96.1% only. It is remarkable that Theileria sp. 2 (China) has a similar morphology and is transmitted by the same tick species (Haemaphysalis qinghaiensis) as Theileria sp. 1 (China). Moreover, differentiation of both species is complicated by overlapping endemic regions and the occurrence of mixed infections. Possibly for these reasons, the two were not distinguished until molecular classification methods were applied (Yin et al. 2003).

The parasites B. motasi (Ameland) and B. motasi (Texel) were isolated in the Netherlands and, accordingly, both segregate into cluster 9a. The species B. crassa is represented by two geographic variants having highly related 18S RNA gene sequences: Babesia sp. (Turkey) and B. crassa (Iran) in cluster 9c. So far B. crassa has been identified in Iran only. However, the present report of B. crassa in Turkey may not be surprising, since Turkey and Iran share borders.

The species B. ovis (Turkey), like Theileria sp. 2 (China), is also somewhat isolated in the phylogenetic tree (cluster 11). It represents a small Babesia species and is based on 18S RNA gene identity most closely related to B. crassa (93%).

All sheep/goat piroplasm species investigated in this work developed independently by polyphyletic evolution. Each of the following pairs of parasites descended most likely from a single direct ancestor: Theileria sp. 1 (China)/Theileria sp. Serow (Japan), T. ovis/Theileria sp. CNY2A Sika deer (Japan), T. separata/Theileria sp. Sable (South Africa) and T. lestoquardi/T. annulata. Possibly, the parasite has been transferred from the wild serow, sika deer and/or sable antelope to imported domestic sheep by transfer via compatible tick species. While the known transmission of T. annulata and T. lestoquardi by the same tick species is consistent with this assumption, the tick species transmitting Theileria sp. Serow (Japan), Theileria sp. CNY2A Sika deer (Japan) and Theileria sp. Sable (South Africa) are unfortunately not known at present (Thomas et al. 1982; Stoltsz and Dunsterville 1992; Takahashi et al. 1992).

In contrast, there are no close relatives to the species Theileria sp. 2 (China), B. crassa and B. ovis; and all three are placed relatively isolated in the tree as exemplified by the low identity of 18S RNA genes with their closest relatives: Theileria sp. CNY1B Sika deer (Japan) (97.5%), B. ovata (96.2%) and B. bigemina (Mexico) (91.8%), respectively.

In summary, five sheep/goat Theileria and three sheep/goat Babesia piroplasm species have been recognised in this study: Theileria sp. 1 (China), T. ovis, T. separata, T. lestoquardi, Theileria sp. 2 (China), B. motasi, B. crassa and B. ovis. Of these, the species Theileria sp. 1 (China) and Theileria sp. 2 (China) have been reported very recently (Schnittger et al. 2000a, Yin et al. 2003).

Two additional sheep/goat piroplasm species, T. recondita and a very recently reported unknown Babesia parasite from China [Babesia sp. (China)] have not been included in this study (Alani and Herbert 1988; Bai et al. 2002). T. recondita, in contrast to Theileria sp. 1 (China), T. lestoquardi and Theileria sp. 2 (China), is non-pathogenic and certainly not identical with the species T. ovis and T. separata (Uilenberg and Andreasen 1974; Uilenberg 1981). Hence, T. recondita is most likely an additional species and it must therefore be concluded that at least six sheep/goat Theileria species do exist. Babesia sp. (China) seems to be most closely related to B. motasi. However, in contrast to B. motasi, it is transmitted by the tick species H. longicornis, is highly pathogenic, infects both sheep and goats and exhibits a smaller morphology (Bai et al. 2002). Due to these attributes, it certainly represents an additional fourth sheep/goat Babesia parasite species.