Molecular and Biochemical Parasitology, 54 (1992) 87-96
87
~c? 1992 Elsevier Science Publishers B.V. All rights reserved. / 0166-6851/92/$05.00 MOLBIO 01782
Ribosomal DNA
s e q u e n c e c o m p a r i s o n o f Babesia and Theileria
John Ellis a', Chris Hefford a, Peter R. Baverstock b, Brian P. Dalrymple ~ and Alan M. Johnson a° ~Department of Microbiology and Infectious Diseases, Flinders University Medical School, Bedford Park, South Australia, Australia: bCentre for Conservation Technology, University of Northern Rivers, Lismore, New South Wales, Australia; and ~CSIRO Division ~ Tropical Animal Production, lndooroopilly, Queensland, Australia (Received 7 November 1991; accepted 20 April 1992)
Previous studies on the taxonomy of Babesia spp. (phylum Apicomplexa) using morphological and life cycle characteristics have resulted in their classification into 3 subgenera, with the genus Theileria being most closely related to them. Using a strategy based on the direct sequence analysis of products derived by asymmetric PCR to determine the nucleotide sequences, we have tested the validity of this classification by sequencing the small subunit ribosomal RNA genes amplified from 2 Babesia species, namely Babesia hovis and Babesia rodhami, and comparing these with previously published sequences of Theileria annulata and Babesia bigemina using Plasmodiumfalciparum as an outgroup. The results of this phylogenetic analysis support the recognition of at least 2 genera in Babesia - - one to include B. bigemma and B. boris, the other to include B.
rodhaini. Key words: Phylogeny; Babesia spp; Theileria annulata; Sarcocystis muris; Polymerase chain reaction; Ribosomal RNA gent
Introduction
Haemogregarina, Gregarina, Isospora, Haemoproteus, Plasmodium, Sarcocystis and Babesia)
The phylum Apicomplexa contains over 4600 named species of parasitic protozoa, many of which are known to cause serious and therefore important diseases of man and livestock. Of the estimated 180 genera known to comprise this phylum only 8 (Eimeria,
contain more than half of the named species [1]. Members of a number of these genera, Eimeria, Plasmodium, Sarco~Tstis and Babesia and the related genera Theileria and Toxoplasma are the focus of intensive study aimed at the production of recombinant vaccines. It is probable that generic approaches to vaccination against members of the different genera in the Apicomplexa will be developed as a result of this work and will be significantly aided by an understanding of the evolutionary relationships between the different genera. The taxonomic grouping of the genera within the phylum Apicomplexa has undergone many reorganisations and is still the subject of much discussion [2-8]. Molecular systematics is currently a flourishing area of research in which the techniques of molecular biology can be used to investigate controver-
"Present address." Department of Microbiology, School of Biological and Biomedical Sciences, University of Technology Sydney, Westbourne Street, Gore Hill, New South Wales 2065, Australia. Correspondence address: A.M. Johnson, Department of Microbiology, School of Biological and Biomedical Sciences, University of Technology Sydney, Westbourne Street, Gore Hill, New South Wales 2065, Australia.
Note." Nucleotide sequence data reported in this paper have been submitted to the GenBank T M data base with the accession numbers M87566 (Babesia boris) and M87565 (Babesia rodhaini).
88
sies that have arisen from classical taxonomic approaches [9,10]. Within the subclass Piroplasmia of the Apicomplexa the classification of some members of the genus Bahesia has not been resolved and it has been suggested that some species should be transferred to another genus (Theileria) within the subclass [11]. Recent analysis of ribosomal RNA (rRNA) gene restriction maps has supported the division of the genus Babesia into at least 2 subgenera, if not 2 distinct genera (Dalrymple, B.P., Dimmock, C., Parrodi, F. and Wright, I.G., submitted tbr publication). In order to test the validity of such a classification we report a phylogenetic analysis using nucleotide sequences derived from the small subunit rRNA (ssrRNA) genes of Babesia boris and Babesia rodhaini by direct sequencing of products derived from asymmetric polymerase chain reactions. The relationships of these species of Babesia to other Apicomplexans is also discussed.
Materials and Methods
The cloning of the rRNA genes of B. boris has been previously described [12]. A 7-kb EcoRI restriction fragment of B. boris which contained the rRNA unit was subcloned into the EcoRI site of p G E M 3 Z f ( + ) . Genomic DNA of B. rodhaini was isolated from merozoites according to standard procedures [12]. The ssrRNA gene was amplified by PCR [13] using primers A and B [14] and the following temperature profiles. The mixture was heated at 95:~C for 5 min, followed by 25 cycles of 55~C (2 min), 72~C (2 min) and 93~C (2 min). This was terminated by a final cycle of 55°C (2 min) and 72°C (5 min). The doublestranded PCR products were analysed by agarose gel electrophoresis. Single-stranded template DNA for sequencing was prepared as follows. A doublestranded DNA template was amplified from plasmid or genomic DNA using conserved primers (50 pmol per reaction each, 100 ltl reaction volume) located throughout the
ssrRNA gene [15]. One microliter of doublestranded template formed in this way and verified by gel electrophoresis was used to initiate asymmetric PCR with a single homologous primer (50 pmol per reaction). The products were desalted using Centricon 100 microconcentrators, precipitated with ethanol and resuspended in distilled water. Nucleotide sequences were determined by the dideoxy chain termination procedure using Sequenase, [35S]dATP, denaturing polyacrylamide gel electrophoresis and autoradiography [16]. Overlapping, contiguous sequences were identified by eye. The nucleotide sequences for the ssrRNA genes of B. boris and B. rodhaini were aligned with previously published sequences for B. higemina (GenBank accession number X59604) [17], T. annulata and Sarcocystis muris (GenBank accession numbers M34845 and M34846) [18] using the program CLUSTAL V (Higgins, D.G.. Bleasby, A.J. and Fuchs, R., submitted for publication) run on a Sun SPARC Server 470 computer. The phylogenetic relationship of these sequences were analysed by parsimony using PAUP [19] with the sequence of the ssrRNA gene of Plasmodium fak'iparum (Genbank Accession Number M19172) [20] used to root the tree. In order to test the rigour of the tree the data were subject to bootstrapping [21]. This procedure randomly samples subsets of the data, and asks whether nodes on the tree supported by the entire data set are also supported by subsets of the data. The data were also subjected to distance analysis. K,,uc is a measure of nucleotide divergence that takes into account the possibility of multiple changes at a single site, and these values were calculated as described [22]. These distance data were analysed phylogenetically by the Fitch-Margoliash method [23] using the FITCH option in PHYLIP version 3.2 [21]. Consensus trees were constructed as described [24,25].
89
Results
The D N A sequences of the ssrRNA genes of B. bovis and B rodhaini were determined from overlapping contiguous sequences. Only a few bases directly adjacent to the terminal primers A and B were not determined. 1586 and 1684 bp of B. boris and B. rodhaini sequences were aligned with the homologous regions of B. bigemina, T. annulata, S. muris and P. falciparum (Fig. 1). Based on these sequence alignments the total number of nucleotide substitution differences were calculated (Table I). The extent of nucleotide divergence varies from approximately 10--15% in the ingroup, and up to 20% in the outgroup. Thus the 18S r R N A gene is an appropriate gene for a phylogenetic analysis of this group because it shows some, but not too much variation [10,26]. The entire data set was analysed by searching for the most parsimonious trees using P A U P [19]. For 6 taxa, there are only 105 possible trees [27] and so it was possible to evaluate all of them. The most parsimonious tree had a length of 917, with the next 6 most parsimonious trees having lengths of 921, 923 (2 trees), 924, 926 and 927 respectively. The 8th most parsimonious tree was 3 steps longer (930), and thus trees with more than 927 steps were excluded from further analyses. The most parsimonious tree rooted using P. Jalciparum as an outgroup is shown in Fig. 2a. On this tree
IABLE 1 S u m m a r y o f the extent of nucleotide differences a m o n g the Babesia species. (Bbi-B. higemim~, Bbo-B. boris, Bro-B. rodhaini), Theileria ITan-7". annulata), Sarco¢3'stis (Smu-S. muris) and fi~r P. falciparum (Pla) for 2046 nucleotide positions
Bbi Bbo Bro Tan Smu Pfa
Bbi
Bbo
Bro
Tan
Smu
Pfa
0.29 0.26 0.23 0.28 I).39
201 0.34 0.31 0.36 0.48
175 247 0.22 0.25 0.36
150 221 144 0.23 0.37
221 277 205 184 0.39
354 405 346 361 422 -
Upper matrix, n u m b e r of pairwise nucleolide differences: lower matrix, K~oc values.
B. bigemina and B. boy& form a monophyletic group, but B. rodhaini falls outside of a clade consisting of B. bigemina, B. boris and T. annulata. The 3 Babesia species along with Theileria form a clade to the exclusion of S. muris. A strict-consensus tree [24] of the 7 most parsimonious trees of length less than 928 is shown in Fig. 2b. An Adams consensus [25] gave exactly the same tree. Thus the only consensus of all 7 trees is the monophyly of B. bigemina and B. boris. In order to test the robustness of the tree, the data were subjected to bootstrapping [21] 20 times. The results (Fig. 2c) showed that in all 20 bootstraps monophyly of B. h~emina and B. hovis was supported. Moreover, the majority of bootstraps (63%) placed B. rodhahfi outside a clade consisting of B. higemina, B. hovis and Theileria. The data were also subjected to distance analysis. Table 1 shows the number of nucleotide substitution differences among the 6 taxa calculated as Knu~ [22]. The best fit tree produced by the Fitch -Margoliash method (Fig. 3) had a standard deviation for goodnessof-fit of 1.28%. This analysis again shows B. rodhaini lying outside a clade containing B. boris, B. higemina and T. annulata. Indeed~ on this tree B. rodhaini lies outside a clade consisting of B. boris, B, h~emina, 7". atmulata and S. mur{s. However, other than the common branch uniting B. hovis and B. higemina, all inter-node distances are small with Kn~c values less than 0.01. The alignments shown in Fig. 1 matched bases that are in similar positions within the secondary structure of the ssrRNA. The 18S r R N A gene in the study group, as in others, is variable in length because of the presence of expansion segments which are frequently hyperwlriable. For example, by comparison to secondary structure models presented for ssrRNAs derived from several eukaryotes including P. Jalciparum, [28,29] the Babesia D N A sequences which probably encode region V4 appear the most variable, since the alignments around this region contained a large number of gaps. The alignment of
90
> Bbi
87
~c? 1992 Elsevier Science Publishers B.V. All rights reserved. / 0166-6851/92/$05.00 MOLBIO 01782
Ribosomal DNA
s e q u e n c e c o m p a r i s o n o f Babesia and Theileria
John Ellis a', Chris Hefford a, Peter R. Baverstock b, Brian P. Dalrymple ~ and Alan M. Johnson a° ~Department of Microbiology and Infectious Diseases, Flinders University Medical School, Bedford Park, South Australia, Australia: bCentre for Conservation Technology, University of Northern Rivers, Lismore, New South Wales, Australia; and ~CSIRO Division ~ Tropical Animal Production, lndooroopilly, Queensland, Australia (Received 7 November 1991; accepted 20 April 1992)
Previous studies on the taxonomy of Babesia spp. (phylum Apicomplexa) using morphological and life cycle characteristics have resulted in their classification into 3 subgenera, with the genus Theileria being most closely related to them. Using a strategy based on the direct sequence analysis of products derived by asymmetric PCR to determine the nucleotide sequences, we have tested the validity of this classification by sequencing the small subunit ribosomal RNA genes amplified from 2 Babesia species, namely Babesia hovis and Babesia rodhami, and comparing these with previously published sequences of Theileria annulata and Babesia bigemina using Plasmodiumfalciparum as an outgroup. The results of this phylogenetic analysis support the recognition of at least 2 genera in Babesia - - one to include B. bigemma and B. boris, the other to include B.
rodhaini. Key words: Phylogeny; Babesia spp; Theileria annulata; Sarcocystis muris; Polymerase chain reaction; Ribosomal RNA gent
Introduction
Haemogregarina, Gregarina, Isospora, Haemoproteus, Plasmodium, Sarcocystis and Babesia)
The phylum Apicomplexa contains over 4600 named species of parasitic protozoa, many of which are known to cause serious and therefore important diseases of man and livestock. Of the estimated 180 genera known to comprise this phylum only 8 (Eimeria,
contain more than half of the named species [1]. Members of a number of these genera, Eimeria, Plasmodium, Sarco~Tstis and Babesia and the related genera Theileria and Toxoplasma are the focus of intensive study aimed at the production of recombinant vaccines. It is probable that generic approaches to vaccination against members of the different genera in the Apicomplexa will be developed as a result of this work and will be significantly aided by an understanding of the evolutionary relationships between the different genera. The taxonomic grouping of the genera within the phylum Apicomplexa has undergone many reorganisations and is still the subject of much discussion [2-8]. Molecular systematics is currently a flourishing area of research in which the techniques of molecular biology can be used to investigate controver-
"Present address." Department of Microbiology, School of Biological and Biomedical Sciences, University of Technology Sydney, Westbourne Street, Gore Hill, New South Wales 2065, Australia. Correspondence address: A.M. Johnson, Department of Microbiology, School of Biological and Biomedical Sciences, University of Technology Sydney, Westbourne Street, Gore Hill, New South Wales 2065, Australia.
Note." Nucleotide sequence data reported in this paper have been submitted to the GenBank T M data base with the accession numbers M87566 (Babesia boris) and M87565 (Babesia rodhaini).
88
sies that have arisen from classical taxonomic approaches [9,10]. Within the subclass Piroplasmia of the Apicomplexa the classification of some members of the genus Bahesia has not been resolved and it has been suggested that some species should be transferred to another genus (Theileria) within the subclass [11]. Recent analysis of ribosomal RNA (rRNA) gene restriction maps has supported the division of the genus Babesia into at least 2 subgenera, if not 2 distinct genera (Dalrymple, B.P., Dimmock, C., Parrodi, F. and Wright, I.G., submitted tbr publication). In order to test the validity of such a classification we report a phylogenetic analysis using nucleotide sequences derived from the small subunit rRNA (ssrRNA) genes of Babesia boris and Babesia rodhaini by direct sequencing of products derived from asymmetric polymerase chain reactions. The relationships of these species of Babesia to other Apicomplexans is also discussed.
Materials and Methods
The cloning of the rRNA genes of B. boris has been previously described [12]. A 7-kb EcoRI restriction fragment of B. boris which contained the rRNA unit was subcloned into the EcoRI site of p G E M 3 Z f ( + ) . Genomic DNA of B. rodhaini was isolated from merozoites according to standard procedures [12]. The ssrRNA gene was amplified by PCR [13] using primers A and B [14] and the following temperature profiles. The mixture was heated at 95:~C for 5 min, followed by 25 cycles of 55~C (2 min), 72~C (2 min) and 93~C (2 min). This was terminated by a final cycle of 55°C (2 min) and 72°C (5 min). The doublestranded PCR products were analysed by agarose gel electrophoresis. Single-stranded template DNA for sequencing was prepared as follows. A doublestranded DNA template was amplified from plasmid or genomic DNA using conserved primers (50 pmol per reaction each, 100 ltl reaction volume) located throughout the
ssrRNA gene [15]. One microliter of doublestranded template formed in this way and verified by gel electrophoresis was used to initiate asymmetric PCR with a single homologous primer (50 pmol per reaction). The products were desalted using Centricon 100 microconcentrators, precipitated with ethanol and resuspended in distilled water. Nucleotide sequences were determined by the dideoxy chain termination procedure using Sequenase, [35S]dATP, denaturing polyacrylamide gel electrophoresis and autoradiography [16]. Overlapping, contiguous sequences were identified by eye. The nucleotide sequences for the ssrRNA genes of B. boris and B. rodhaini were aligned with previously published sequences for B. higemina (GenBank accession number X59604) [17], T. annulata and Sarcocystis muris (GenBank accession numbers M34845 and M34846) [18] using the program CLUSTAL V (Higgins, D.G.. Bleasby, A.J. and Fuchs, R., submitted for publication) run on a Sun SPARC Server 470 computer. The phylogenetic relationship of these sequences were analysed by parsimony using PAUP [19] with the sequence of the ssrRNA gene of Plasmodium fak'iparum (Genbank Accession Number M19172) [20] used to root the tree. In order to test the rigour of the tree the data were subject to bootstrapping [21]. This procedure randomly samples subsets of the data, and asks whether nodes on the tree supported by the entire data set are also supported by subsets of the data. The data were also subjected to distance analysis. K,,uc is a measure of nucleotide divergence that takes into account the possibility of multiple changes at a single site, and these values were calculated as described [22]. These distance data were analysed phylogenetically by the Fitch-Margoliash method [23] using the FITCH option in PHYLIP version 3.2 [21]. Consensus trees were constructed as described [24,25].
89
Results
The D N A sequences of the ssrRNA genes of B. bovis and B rodhaini were determined from overlapping contiguous sequences. Only a few bases directly adjacent to the terminal primers A and B were not determined. 1586 and 1684 bp of B. boris and B. rodhaini sequences were aligned with the homologous regions of B. bigemina, T. annulata, S. muris and P. falciparum (Fig. 1). Based on these sequence alignments the total number of nucleotide substitution differences were calculated (Table I). The extent of nucleotide divergence varies from approximately 10--15% in the ingroup, and up to 20% in the outgroup. Thus the 18S r R N A gene is an appropriate gene for a phylogenetic analysis of this group because it shows some, but not too much variation [10,26]. The entire data set was analysed by searching for the most parsimonious trees using P A U P [19]. For 6 taxa, there are only 105 possible trees [27] and so it was possible to evaluate all of them. The most parsimonious tree had a length of 917, with the next 6 most parsimonious trees having lengths of 921, 923 (2 trees), 924, 926 and 927 respectively. The 8th most parsimonious tree was 3 steps longer (930), and thus trees with more than 927 steps were excluded from further analyses. The most parsimonious tree rooted using P. Jalciparum as an outgroup is shown in Fig. 2a. On this tree
IABLE 1 S u m m a r y o f the extent of nucleotide differences a m o n g the Babesia species. (Bbi-B. higemim~, Bbo-B. boris, Bro-B. rodhaini), Theileria ITan-7". annulata), Sarco¢3'stis (Smu-S. muris) and fi~r P. falciparum (Pla) for 2046 nucleotide positions
Bbi Bbo Bro Tan Smu Pfa
Bbi
Bbo
Bro
Tan
Smu
Pfa
0.29 0.26 0.23 0.28 I).39
201 0.34 0.31 0.36 0.48
175 247 0.22 0.25 0.36
150 221 144 0.23 0.37
221 277 205 184 0.39
354 405 346 361 422 -
Upper matrix, n u m b e r of pairwise nucleolide differences: lower matrix, K~oc values.
B. bigemina and B. boy& form a monophyletic group, but B. rodhaini falls outside of a clade consisting of B. bigemina, B. boris and T. annulata. The 3 Babesia species along with Theileria form a clade to the exclusion of S. muris. A strict-consensus tree [24] of the 7 most parsimonious trees of length less than 928 is shown in Fig. 2b. An Adams consensus [25] gave exactly the same tree. Thus the only consensus of all 7 trees is the monophyly of B. bigemina and B. boris. In order to test the robustness of the tree, the data were subjected to bootstrapping [21] 20 times. The results (Fig. 2c) showed that in all 20 bootstraps monophyly of B. h~emina and B. hovis was supported. Moreover, the majority of bootstraps (63%) placed B. rodhahfi outside a clade consisting of B. higemina, B. hovis and Theileria. The data were also subjected to distance analysis. Table 1 shows the number of nucleotide substitution differences among the 6 taxa calculated as Knu~ [22]. The best fit tree produced by the Fitch -Margoliash method (Fig. 3) had a standard deviation for goodnessof-fit of 1.28%. This analysis again shows B. rodhaini lying outside a clade containing B. boris, B. higemina and T. annulata. Indeed~ on this tree B. rodhaini lies outside a clade consisting of B. boris, B, h~emina, 7". atmulata and S. mur{s. However, other than the common branch uniting B. hovis and B. higemina, all inter-node distances are small with Kn~c values less than 0.01. The alignments shown in Fig. 1 matched bases that are in similar positions within the secondary structure of the ssrRNA. The 18S r R N A gene in the study group, as in others, is variable in length because of the presence of expansion segments which are frequently hyperwlriable. For example, by comparison to secondary structure models presented for ssrRNAs derived from several eukaryotes including P. Jalciparum, [28,29] the Babesia D N A sequences which probably encode region V4 appear the most variable, since the alignments around this region contained a large number of gaps. The alignment of
90
> Bbi
