Logo of mbioJournal InfoAuthorsReviewersBoard of EditorsJournals ASM.orgmBiomBio Article
mBio. 2012 Sep-Oct; 3(5): e00204-12.
Published online 2012 Sep 4. doi:  10.1128/mBio.00204-12
PMCID: PMC3445966

Comparative Genome Analysis of Three Eukaryotic Parasites with Differing Abilities To Transform Leukocytes Reveals Key Mediators of Theileria-Induced Leukocyte Transformation


We sequenced the genome of Theileria orientalis, a tick-borne apicomplexan protozoan parasite of cattle. The focus of this study was a comparative genome analysis of Torientalis relative to other highly pathogenic Theileria species, Tparva and Tannulata. Tparva and Tannulata induce transformation of infected cells of lymphocyte or macrophage/monocyte lineages; in contrast, Torientalis does not induce uncontrolled proliferation of infected leukocytes and multiplies predominantly within infected erythrocytes. While synteny across homologous chromosomes of the three Theileria species was found to be well conserved overall, subtelomeric structures were found to differ substantially, as Torientalis lacks the large tandemly arrayed subtelomere-encoded variable secreted protein-encoding gene family. Moreover, expansion of particular gene families by gene duplication was found in the genomes of the two transforming Theileria species, most notably, the TashAT/TpHN and Tar/Tpr gene families. Gene families that are present only in Tparva and Tannulata and not in Torientalis, Babesia bovis, or Plasmodium were also identified. Identification of differences between the genome sequences of Theileria species with different abilities to transform and immortalize bovine leukocytes will provide insight into proteins and mechanisms that have evolved to induce and regulate this process. The Torientalis genome database is available at http://totdb.czc.hokudai.ac.jp/.


Cancer-like growth of leukocytes infected with malignant Theileria parasites is a unique cellular event, as it involves the transformation and immortalization of one eukaryotic cell by another. In this study, we sequenced the whole genome of a nontransforming Theileria species, Theileria orientalis, and compared it to the published sequences representative of two malignant, transforming species, Tparva and Tannulata. The genome-wide comparison of these parasite species highlights significant genetic diversity that may be associated with evolution of the mechanism(s) deployed by an intracellular eukaryotic parasite to transform its host cell.


Theileria spp. are tick-borne intracellular parasites that belong to the phylum Apicomplexa and infect domestic and wild ruminants, including cattle, Asian water buffalos, sheep, goats, and African buffalos. Although infection by some Theileria species is asymptomatic or persists as a chronic infection, Theileria parva and Theileria annulata can be highly pathogenic to cattle and Theileria lestoquardi can cause significant disease in sheep. These three species are among the “transforming Theileria” species because of their ability to transform and induce indefinite proliferation of infected host leukocytes (14). The resulting disease syndromes can be described as lymphoproliferative disorders, which often culminate in disorganization and destruction of the host lymphoid system. Although detailed information has been generated for a number of host cell signal transduction pathways that are perturbed during leukocyte transformation, parasite molecules responsible for the initiation or regulation of the host cell transformation event have yet to be identified or fully validated (5, 6).

A comparative analysis of the Tparva and Tannulata genome sequences was reported in 2005 (7, 8). Despite the identification of a number of Theileria genes that could be involved in the transformation process, the selectivity of the approach was compromised by a high number of hypothetical proteins of unknown function and the high number of shared genes that exists across the genomes of these two closely related species. One way in which the discriminatory power of a comparative genomic approach could be increased would be to conduct bi- and trilateral genome comparisons with Theileria and Babesia parasites that lack the ability to transform host leukocytes but otherwise show strong similarity over the rest of their parasitic life cycle (9).

Theileria orientalis, an intraerythrocytic parasite of cattle, is a member of the nontransforming group of Theileria species that proliferate in the bovine host as an intraerythrocytic form and can generate anemia and icterus but rarely cause fatal disease (10). This parasite has frequently been referred to as Tsergenti, but this specific name is now considered invalid (11). Bovine piroplasmosis caused by this species causes enormous economic losses in the livestock industry in Japan (1214). Torientalis is often classified into two major genotypes, the Chitose type and the Ikeda type, which are distinguishable on the basis of diversity in the small-subunit rRNA and major piroplasm surface protein (MPSP) gene sequences (15). The Torientalis Ikeda type is limited to eastern Asian countries, including Japan, South Korea, the northeastern part of China, and Australia (16), and it is present in areas where livestock succumb to severe clinical cases of theileriosis and serious production losses. In contrast, Torientalis Chitose is found throughout the world and is usually associated with benign infection (15, 17). Thus, even though it is believed to be relatively mild compared to the transforming Theileria species, Torientalis can be an important pathogen in its own right and many researchers have been looking forward to the derivation of the genomic sequence to provide an important resource for further studies.

Unlike transforming Theileria species, the macroschizonts of nontransforming Theileria parasites are only transiently found in cells within lymph nodes or the spleen following the invasion of host cells by the infective sporozoite, and no evidence for proliferation of infected cells has been reported in vivo or in vitro. Indeed, in vivo studies indicate that the schizont undergoes continual enlargement over the course of 4 to 8 days before generating multiple merozoites that are released upon host cell destruction. A lack of host proliferation is indicated by a substantial increase in host cell size, but it is unknown whether the parasite manipulates the cell at the molecular level or inhibits an apoptotic response to infection (13, 18). Free merozoites subsequently invade erythrocytes and, unlike the case with transforming species, undergo significant rounds of proliferation in red blood cells, similar to the proliferation observed with Babesia parasites. Clinical signs, when observed, are associated primarily with anemia and icterus. In addition to the schizont stage, the intraerythrocytic stage of Tannulata can also cause anemia.

In this study, we focused primarily on a comparative analysis of the genome of the Torientalis Ikeda type relative to the genomes of the transforming Theileria species Tparva and Tannulata and a closely related hemoparasite species, Babesia bovis. The main goals of this analysis were to provide supportive data on existing candidate genes and/or identify novel candidate genes that enable the transformation of bovine leukocytes upon infection with Tannulata and Tparva.


Structure of the Torientalis genome.

Whole-genome shotgun sequence data on Torientalis (Ikeda strain) were assembled, and physical gaps between scaffolds were manually closed, resulting in the complete sequence of all four chromosomes. The derived sequence has been deposited in the DNA Data Bank of Japan (DDBJ) under project accession numbers AP011946 to AP011951. In addition to the nuclear genome, partial sequences of the apicoplast and mitochondrial genomes were also obtained. The compete genome sequence of the mitochondria has already been published (accession number AB499090) (19).

At 9.0 Mb, the genome size of Torientalis is approximately 8% larger than the reported genome sizes of Tparva, Tannulata, and B. bovis. The number of predicted protein-coding genes identified in Torientalis is, however, almost the same as that found in Tparva (Table 1). The G+C composition of the Torientalis genome (41.6%) is higher than those of Tparva and Tannulata (34.1% and 32.5%, respectively) but similar to that of Bbovis (41.8%). The frequencies of the top 50 InterPro entries (see Table S1 in the supplemental material) are similar for the three Theileria species, suggesting that, in general, the three parasite species possess similar sets of gene families and encoded protein domains. For example, the InterPro domain of DUF529, known as the FAINT (frequently associated in Theileria) domain, described later in detail, is found frequently in all of the Theileria species sequenced to date. In contrast, the PEST motif, associated with rapid degradation of (nuclear) proteins, was found to be encoded by several gene families in the genomes of the two transforming Theileria species but was not identified in Torientalis.

Comparison of genome characteristics of Torientalis, Tparva, Tannulata, and B. bovis

Synteny across all of the chromosomes of all three Theileria species is generally conserved, except for the subtelomeric regions, and several internal inversions were identified for each chromosome (Fig. 1). Most large-scale inversions were found when comparing Torientalis versus Tannulata (Fig. 1, lower half of each panel) and were not present in the Tannulata-versus-Tparva comparison (Fig. 1, upper half of each panel), suggesting that these structural changes occurred following the speciation of Torientalis and a common ancestor of Tannulata/Tparva. However, a large inversion of approximately 113,000 bp in chromosome 3 of Tannulata may have occurred after the speciation of Tannulata and Tparva (Fig. 1, upper right panel, indicated by a double-headed arrow). A striking difference between the genomes is that a number of gene families show evidence of expansion and diversification specific to the genomes of the transforming Theileria species, while few instances of Torientalis-specific gene family expansion were recorded. In addition, several lineage-specific genes were identified at microsynteny breakpoints. Finally, the subtelomeric regions of all four Torientalis chromosomes are markedly different from those of Tannulata and Tparva because they completely lack the largest subtelomeric gene family reported for Tannulata and Tparva, which encodes subtelomere-encoded variable secreted proteins (SVSPs) (see Fig. 2) of unknown function (20).

Genome scale synteny among three species of Theileria chromosomes. Shown is an Artemis Comparison Tool (65) plot of Torientalis (bottom) versus Tannulata (middle) and Tparva (top). Blue bars indicate matching regions in the ...
Genes in the subtelomeric region of each chromosome. Annotation of the subtelomeric regions of Torientalis chromosomes identified mainly ABC transporter family or SfiI-related family genes but not SVSP genes, while subtelomeric regions of T ...

Metabolic pathways.

To reconstruct KEGG metabolic pathways of Torientalis, we assigned 263, 263, 273, and 264 KEGG orthology (KO) identifiers (21) to the predicted proteomes of Torientalis, Tparva, Tannulata, Torientalis, and Bbovis, respectively (see Fig. S1 in the supplemental material). These four species had 255 KOs in common, indicating no significant differences in known metabolic pathways between nontransforming and transforming Theileria and Babesia species, despite the known preference to proliferate in different host cell types (leukocytes versus erythrocytes).

K00626 is the only KO common to Bbovis and Torientalis and not identified in Tparva and Tannulata. It codes for a putative thiolase that catalyzes the conversion of acetyl coenzyme A (acetyl-CoA) into acetoacetyl-CoA. This enzyme is known to function in a variety of metabolic pathways, including fatty acid metabolism, nucleotide metabolism, and amino acid degradation. Phylogenetic analysis has indicated that the two transforming Theileria species diverged from Torientalis after the speciation of Theileria and Babesia (22). Therefore, the acetoacetyl-CoA thiolase might represent an example of a reduction of metabolic capacity due to an increasing host cell metabolite-scavenging ability/dependence of Theileria species.

Gene families.

Expansion of gene families specific to different Theileria species could offer a valuable insight into how these parasites have evolved and adapted to their different host environments, including the acquisition of leukocyte transformation capability. To examine the expansion processes of gene families in the Theileria lineages in detail, we constructed gene families composed of sequences representing the three Theileria species, Bbovis, and two Plasmodium species (Plasmodium falciparum and Pvivax) on the basis of the ortholog clustering framework of OrthoMCL (23), as well as additional computational and manual curations. We assigned 3,419 orthologous groups in which at least one Theileria species was included. While 1,740 of these orthologous groups consisted of single-copy genes across all six species, 223 orthologous groups possessed Theileria paralogs (see Data set S1 in the supplemental material). We focused on several family groups in the Theileria lineage that showed evidence of marked expansion that could be associated with acquisition of the ability to generate the proliferating, transformed, infected leukocyte.

Expansion of gene families in the genomes of transforming Theileria species.

Three gene families showed a striking association with the genomes of the two transforming Theileria species. PiroF0100022 (Tar/Tpr family), PiroF0100037 (SVSP family), and PiroF0100038 (TashAT/TpHN family) are all significantly expanded within or unique to the genomes of the host cell-transforming Theileria lineage and are composed of genes predicted to encode proteins possessing FAINT domains. The TashAT family of Tannulata contains 17 tandemly arrayed genes, some of which have been shown to encode proteins that are translocated to the host nucleus, bind DNA, and alter gene expression and protein profiles of transfected bovine cells (24, 25). An orthologous cluster of 20 genes (TpHN) has also been identified in Tparva (25). In sharp contrast, only a single TashAT/TpHN-like gene, TOT0100571, was identified in the genome of Torientalis. Reciprocal best hits using BLASTP indicate that the Torientalis gene is likely to be the ortholog of Tash-a (TA03110) and TP01_0621 in the transforming Theileria species. Both of these genes are located at the 3′ ends of their respective clusters in the Tannulata and Tparva genomes (Fig. 3A).

Genomic and phylogenetic structures of the TashAT gene family (PiroF0100038). (A) Schematic representation of the TashAT clusters in Tparva and Tannulata and the corresponding locus in Torientalis. Genes in the same ortholog ...

To gain further insight into the species-conserved Tash-a gene relative to the other members of the TashAT cluster, we obtained microarray data to examine whether gene expression of the different TashAT genes is associated with proliferating, macroschizont-infected leukocytes (26). Analysis of the normalized dataset showed that, in general, TashAT family expression is consistently downregulated as the macroschizont undergoes differentiation to the merozoite and host cell proliferation subsides, as demonstrated previously for a number of individual family members (25). In marked contrast, transcripts representing Tash-a were found to be significantly upregulated during the differentiation process (see Fig. S2A in the supplemental material). This result may indicate a requirement for synthesis of the protein during merozoite production. This postulation was supported by an indirect fluorescent-antibody test (IFAT) using serum raised against a Tash-a fusion protein (see Fig. S2B) and colocalization of Tash-a staining with a merozoite rhoptry antigen (see Fig. S2C). We conclude that the Tash-a protein performs a function that is required during or following merozoite production and that the temporal expression and location of the protein are distinct from those of other members of the family. Phylogenetic analysis suggests that Tash-a and its orthologs represent ancestral members of the TashAT and TpHN clusters (see Fig. 3B). In addition, we did not find any obvious TashAT orthologs in Bbovis or two Plasmodium species genomes. We propose that Tash-a diverged after the separation of Theileria from a common ancestor of Theileria and Babesia and that gene duplication and functional diversification of the TashAT and TpHN clusters has then occurred as Theileria species of the transforming lineage evolved. Whether expansion of the cluster coincided with acquisition of a transforming capability is unknown.

Polypeptides encoded by the subtelomeric SVSP gene family (PiroF010037) are a major component of the predicted macroschizont secretome of Tannulata and Tparva, and a number of SVSPs have been predicted to translocate to the nucleus of the infected cell. Most SVSP genes are coexpressed in cultures of macroschizont-infected cells, and the SVSP family shows a high level of amino acid sequence diversity (20). Further work is needed to determine the function of SVSPs, whether they contribute directly to the transformation of the host cell or play a role in subverting the bovine immune response. Some of the SVSPs contain bioinformatically detectable signal peptides, suggesting secretion into the host cell cytoplasm. Though the expression patterns of Tparva SVSPs appear complicated and their involvement in phenotypic changes in host leukocytes remains unclear, the fact that some SVSPs encode functional nuclear localization signals (NLSs) in addition to a predicted signal sequence for secretion suggests that they might be transported to the host nucleus and modulate signaling pathways (20). In this context, the absence of SVSP loci in Torientalis is noteworthy. Thus, like the TashAT/TpHN clusters, SVSP gene expansion in Tannulata/Tparva appears to be associated with species of the transforming Theileria lineage and may provide an as-yet-unknown function that promotes the establishment or maintenance of proliferating macroschizont-infected leukocytes.

In addition to the SVSP and TashAT clusters, the Tar/Tpr (PiroF0100022) family of orthologous genes showed evidence of significant expansion in the transforming Theileria lineages, as only five genes dispersed over the four chromosomes were detected in Torientalis, compared with the 69 dispersed Tar genes in Tannulata. The function of the proteins encoded by Tar/Tpr genes is unknown. They lack a FAINT domain, and the presence of multiple transmembrane domains predicts a membrane location. Transcriptome studies indicate that copies of Tpr genes dispersed throughout the Tparva genome are expressed in the macroschizont stage (27), while those organized in a tandem array of 28 genes are expressed by the intraerythrocytic piroplasm (28).

The CD8 T cell response is considered to play a key role in immunity to Tparva/Tannulata (29). Of the macroschizont antigens that are recognized by CD8 T cells from immune animals (30, 31), one, TA9/TP9 (TA15705/TP02_0895), is encoded by a member of a small orthologous gene family (PiroF0100041) in the genomes of transforming Theileria species. The family consists of five and six members in Tannulata and Tparva, respectively, all of which encode predicted proteins with a signal peptide for secretion by the parasite. Expressed sequence tag (EST) data and microarray data indicate that one of the TA9 family members (TA15705) is expressed in a specific manner by the transforming macroschizont stage (see Fig. S3C in the supplemental material), and it has been reported that the protein can be detected in the host cell cytosol (32). In the Torientalis genome, a single gene (TOT020000921) showing weak homology in the signal peptide region and C-terminal region with the TA9/TP9 family was found in a syntenic region of chromosome 2 (see Fig. S3). The data indicate that the TA9/TP9 gene family has expanded uniquely in the transforming Theileria species. A role for TA9-encoded polypeptides in the transformation of the host cell requires further investigation.

Evolution of the FAINT domain superfamily.

As observed for Tannulata (8) and Tparva (7), a large number of genes whose predicted polypeptides encode DUF529 domains (IPR007480 in InterPro), alternatively called FAINT domains, were found in Torientalis (see Table S1 in the supplemental material). Previous analysis revealed that ~900 copies of FAINT domains are present in the genomes of Tannulata and Tparva (8). With our pipelines for InterPro annotation, 686 FAINT copies were identified in 137 predicted Torientalis proteins, and 913 and 725 copies were identified in 126 Tannulata and 142 Tparva putative proteins, respectively. This suggests that expansion of FAINT domain-containing polypeptides (FAINT superfamily) is likely to have occurred in the common ancestor of the three Theileria species. In addition, ortholog clustering indicated that different FAINT families have been expanded in Torientalis than in Tparva and Tannulata. For example, the FAINT superfamilies of PiroF0001942 and PiroF0001943 are specifically expanded in Torientalis (see Table S2 in the supplemental material). In contrast, the PiroF0100056 orthologous group of SfiI-related genes showed greater expansion in Tparva and Tannulata (see Table S2). A protein of the FAINT superfamily was also found in Tequi (8), which has been considered to be an outlier species in the genus Theileria (33). This indicates that FAINT domain polypeptides were present in early ancestral species of the Theileria genus and have subsequently been subjected to differential expansion or contraction pressures as the different species evolved.

Many of the FAINT superfamily members in Tparva and Tannulata are inferred to be secretory proteins (5). Out of 137 proteins of the FAINT superfamily identified in Torientalis, signal peptides were found in 103, indicating that members of the FAINT superfamily are significantly enriched for proteins with a predicted signal peptide (P = 5.97 × 10−55, Fisher’s exact test). Thus, the differential expansion and diversification of FAINT domain proteins could be associated with the adaptation of different Theileria species to preferential host niches that require specific host-parasite interactions. Comparison of additional genome sequences derived from both nontransforming and transforming Theileria species may be informative.

Candidate genes responsible for Theileria-induced host cell transformation.

Comparative genomic analysis of Torientalis and Tannulata/Tparva provides a tool for identifying candidate genes responsible for Theileria-mediated host cell transformation. This premise is based on the assumption that transformation-related genes are unique to the Tannulata/Tparva lineage, as there is no evidence that Torientalis can transform leukocytes into proliferating infected cells. It can also be predicted that molecules that regulate the transformation event are likely to be secreted or localized to the macroschizont membrane, since Theileria parasites have direct contact with the host cell cytoplasm (34). In the course of ortholog classification analysis, we applied both of these criteria and identified 97 ortholog groups present in the Tparva and Tannulata lineages that were absent from Torientalis, Bbovis, Pfalciparum, and Pvivax. Of these lineage-specific ortholog groups, 29 are predicted to encode polypeptides with an endoplasmic reticulum signal sequence (several of which also contain a GPI anchor motif), indicating potential interaction with the host cell compartment (Table 2). The majority of these genes encode hypothetical proteins and do not show any similarities to known cancer-related genes, although several domains are predicted in the InterPro entries. We propose that genes placed within these 29 groups, plus the TashAT/TpHN family, can be considered candidates for involvement in the transformation process.

Possible candidate transforming genes in Tparva and Tannulata

Identification of candidate genes as host cell phenotype manipulators has been reported previously (5, 35). The predicted proteins have signal sequences, protein kinase properties, phosphatase properties, NLSs, or DNA biding motifs, or they show identity with higher eukaryotic proteins that are involved in neoplasia. We searched for these genes in the genome of Torientalis and found that all of them, with the exception of TashAT and SVSP family genes, are conserved across the three Theileria species (see Table S3 in the supplemental material). However, four Torientalis genes lack the signal sequence or NLS that is predicted in each of the Tannulata/Tparva orthologs. Thus, it is possible that the function or localization of the encoded polypeptides has diverged between Torientalis and the transforming Theileria species, and this may be worthy of further investigation.


This is the first genome sequence of a nontransforming Theileria species that occupies a phylogenetic position close to that of the transforming Theileria species and thus provides an ideal opportunity to analyze unique features of Theileria parasitism from an evolutionary viewpoint. Genome sequencing of the nontransforming Theileria species Torientalis and comparison with the transforming Theileria species Tannulata and Tparva highlighted lineage-specific evolutionary features. Several transforming Theileria lineage-specific gene family expansions were identified, including the SVSP, Tash/TpHN, Tpr/Tar, and TP9/TA9 families, that may have been coincident with development of the ability to transform host leukocytes. Additional genes identified as specific to the genomes of transforming Theileria species can also be considered transformation candidates. This study provides increased understanding of the evolution of transforming Theileria species at the genomic level and has generated a database that will serve as the foundation for future studies on Theileria pathobiology and parasite-host cell interaction.


Parasite samples.

Torientalis (Shintoku stock) was used as the starting genomic material in this study. This stock contains two different genotypes, Ikeda and Chitose. Parasites of a single genotype (Ikeda) were selected following syringe passage of the original isolate through calves and then used to infect an animal for parasite isolation. Blood collected from the infected animal was passed through a leukocyte removal filter (Terumo), and the resulting red cells were washed three times with phosphate-buffered saline (PBS). Erythrocytes were resuspended in an equal volume of PBS and disrupted by nitrogen cavitation, and piroplasms were purified by differential centrifugation as described previously (36). Infection of the cow was conducted in accordance with protocols approved by the National Institute of Animal Health, Japan, Animal Care and Use Committee (approval no. 2000/901). Genomic DNA was purified by proteinase K and SDS treatment, followed by phenol-chloroform extraction. Purified parasite DNA was dissolved in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0). Confirmation that the DNA represented the Ikeda genotype was carried out by PCR targeting genes encoding small-subunit rRNA and the MPSP as previously described (37).

Genome sequencing.

The complete genome sequence of Torientalis was determined by a combination of the whole-genome shotgun method and fosmid end sequencing. Genomic DNA was fragmented for plasmid library construction with an average insert size of 2 to 4 kb using a HydroShear DNA Shearing Device (Genemachines). Plasmid DNA was amplified with a TempliPhi DNA amplification kit (GE Healthcare) from the bacterial culture. The fosmid library was constructed by TaKaRa Bio Inc. using a CopyControl pCC1FOS vector (Epicentre, Madison, WI). Fosmid DNA was extracted with PI-1100 plasmid isolators (Kurabo). Both ends of 40,704 plasmid inserts and 3,840 fosmid clones were sequenced with ABI 3730 sequencers (Applied Biosystems) and MegaBACE 4500 sequencers (GE Healthcare). Contigs were assembled by using 111,945 shotgun reads. Gap closing and resequencing of low-quality regions in the assembled data were performed by shotgun sequencing of fosmid clones that covered the target regions, nested deletion (38), construction of short-insert libraries (39), and primer walking on selected clones and PCR-amplified DNA fragments. The overall accuracy of the finished genome sequence was estimated to have an error rate of less than 1 per 10,000 bases. The sequence is available from DDBJ/GenBank/EMBL under accession numbers AP011946 to AP011951.


Six volumes of Trizol LS was added to 1 volume of parasite-infected erythrocytes and homogenized with a Polytron homogenizer. Total RNA was then isolated according to the manufacturer’s protocol, and full-length cDNA libraries were produced by either the oligocapping or the vector-capping method (40). Random clones were picked from the oligocapped and vector-capped library, and inserts were amplified by PCR from the single colonies sequenced at the 5′ end or both the 5′ end and the 3′ end. Sequences were aligned with available whole-genome sequences by using the est2genome (41) program. These sequences are available from DDBJ/GenBank/EMBL under accession numbers FS557591 to FS578553.

Gene structure prediction and annotation.

All of the repetitive and low-complexity sequences in the Torientalis genome sequence were masked by using RepeatMasker (http://www.repeatmasker.org) with Repbase. rRNA and tRNA genes were detected by using BLAST searches against Rfam (42) and the tRNAscan-SE program (43).

Torientalis genes were first predicted computationally by using Torientalis EST pair gene models and several gene prediction programs and then finally identified by genome-wide manual curation. Torientalis EST sequences, identified from a full-length cDNA library made from parasite-infected erythrocytes, were mapped onto the Torientalis genome. Based on EST-genome alignments using est2genome (41), EST pair gene models were constructed by merging the exon overlap on the same strand of ESTs of the same clone. We identified 544 Torientalis EST pair gene models. Genes were predicted by several gene-finding software packages, including GlimmerHMM (44), GeneMark.hmm (45), GeneWise (46), and JIGSAW (47). GlimmerHMM was trained on two sets of full-length gene sequences. The first set consisted of Torientalis genes (544 EST pair gene models), and the second set consisted of these Torientalis genes and annotated genes of Tparva and Tannulata that were predicted to be longer than 400 amino acids. GeneWise was trained on all of the annotated genes of Tparva and Tannulata. The genes from four sets of results of genome coordinates provided by GlimmerHMM, GeneMark.hmm, GeneWise, and Torientalis (544 EST pair gene models) were summarized by using JIGSAW. JIGSAW was also trained on Torientalis EST pair gene models.

We essentially used annotation procedures described previously (48, 49). For each Torientalis gene product, we conducted InterProScan (50). We then assigned a standardized functional annotation to each gene as illustrated in Fig. S4 in the supplemental material, based on the results of a BLASTX similarity search against the UniProtKB/Swiss-Prot, UniProtKB/TrEMBL, and RefSeq protein databases and InterProScan (48, 51). Finally, to identify the representative Torientalis genes, manual curation was performed by using a custom-made annotation system named TOT-SOUP/G-integra (48, 52). The numbers of manually curated Torientalis genes are summarized in Table S4 in the supplemental material. Signal peptides were inferred by SignalP 3.0 (53).

Ortholog clustering.

Ortholog groups consisted of Torientalis, Tannulata, Tparva, Bbovis, Pfalciparum, and Pvivax proteins derived primarily from gene annotation. Tannulata orthologs were from GeneDB (http://old.genedb.org/genedb/annulata/); Tparva, except for the mitochondrion proteome, and Bbovis orthologs were from RefSeq (http://www.ncbi.nlm.nih.gov/RefSeq/); the Tparva mitochondrial proteome was from UniProt (http://www.uniprot.org/); and Pfalciparum and Pvivax orthologs were obtained from PlasmoDB (http://plasmodb.org/plasmo/). Ortholog groups were generated by OrthoMCL (23) on the basis of sequence similarity by using an all-versus-all NCBI BLASTP search (54) with a bit score cutoff of <60 and default parameters. Because E values from the BLASTP search were applied for a similarity measure, we recomputed the exact E values between closely related proteins if the E value was approximated at 0.0. We integrated the orthologous groups assumed to be duplicated in the Theileria lineage after separation from Babesia into a single group by using both automatic algorithms/software and manual integration as described below. Ortholog groups A and B were merged if any Theileria-Theileria gene pairs in which two genes belonging to A and B, respectively, had higher bit scores than any Theileria-Babesia/Plasmodium gene pairs within single ortholog group A or B. Several ortholog groups were merged by manual curation based on sequence homology and genomic location if they generated tandem arrays on the chromosomes. We also merged nonclustered genes using OrthoMCL into the ortholog groups with the same procedure. Finally, 3,502 ortholog groups were used for the following analyses; PiroF0100001 to PiroF0100062 represent the merged ortholog groups, and PiroF0000001 to PiroF0003675 represent the other ortholog groups. The ortholog clustering left 436, 112, and 293 nonclustered genes in Torientalis, Tannulata, and Tparva, respectively.

KEGG metabolic pathway reconstruction.

Metabolic pathways in Torientalis were analyzed by KEGG metabolic pathway reconstruction. First, BLAST searches were performed for protein sequences in each orthologous cluster against the KEGG GENES database. A KO identifier was then assigned to each cluster according to the most similar hit with a KO annotation; the E value threshold was <1.0e−5.

Molecular phylogenetic analysis.

Amino acid sequences of each ortholog group were multiply aligned with the L-INS-I alignment strategy in MAFFT (55), and gap-rich sequences, such as truncated ones, were removed from the alignments with MaxAlign (56). Ambiguously and/or poorly aligned sites were removed by Gblocks (57), and the rest were subjected to phylogenetic analysis. Phylogenetic trees were inferred by maximum likelihood (ML) (58, 59) with a heuristic ML tree search using RAxML (60) with the WAG-F model (61). Heterogeneity of evolutionary rates among sites was modeled by a discrete gamma distribution, with optimization of gamma shape parameter alpha for each alignment set (62). Bootstrap probability (59) was calculated for each tree node with 1,000 replications.

Generation of recombinant protein and antiserum.

A 1,788-bp fragment of TA03110 was PCR amplified with the C9 (genome) strain of Tannulata as a template. This corresponds to the full-length encoded protein minus the N-terminal signal peptide sequence and spans nucleotide positions 70 to 1,857 relative to the translation start codon. In addition to gene-specific sequences, the PCR primers incorporated attB adaptors to facilitate the use of Gateway Recombination Cloning Technology (Invitrogen); the forward primer was 5′-forward attB adaptor-GAGGACTTGGACCTAAACTCTCC-3′, and the reverse primer was 5′-reverse attB adaptor-AGGATTTTGATCAGTGTTAATATCG-3′. The amplicon was cloned into the pDONR221 shuttle vector and subcloned into the expression vector pDEST17, which has a six-histidine (His6) repeat at the 5′ end of the multiple cloning site. After the transformation of chemically competent Escherichia coli BL21 cells (Invitrogen), expression of the His6-tagged fusion protein was induced by adding l-arabinose to a final concentration of 0.2% in LB liquid medium. Recombinant protein was purified by affinity chromatography on nickel agarose columns under denaturing conditions by using the manufacturer’s protocol (Qiagen). Eluted fractions containing the recombinant protein were assessed by using SDS-PAGE before being pooled. To generate polyclonal anti-TA03110 serum, two rats were immunized a total of four times with 30 µg of recombinant protein per immunization. Immunizations were conducted under a project license issued by the United Kingdom Home Office, i.e., Animals (Scientific Procedures) Act 1986 contract immunization project license PPL 60/3464.

Parasite material and IFAT.

The Tannulata-infected cloned cell line Ankara A2 D7 (26) was used to provide material for the microarray experiment and for the IFAT. To stimulate differentiation from the macroschizont stage to the merozoite stage, cultures were maintained at 41°C by using a previously described protocol (26). Cytospin preparation of Tannulata-infected cells, paraformaldehyde fixation, and the IFAT were performed as described previously (62). The anti-Tash-a serum was used at dilution of 1:500 in cell culture medium, and the anti-His6 tag antibody (sc-65902; Stratagene) was used at 1:200; monoclonal antibodies against a macroschizont surface antigen (1C12), the Tams1 merozoite surface antigen (5E1), and a merozoite rhoptry antigen (1D11) were used as undiluted hybridoma culture medium as previously described (63); anti-rat IgG and anti-mouse IgG secondary antibodies conjugated to Alexa 488 or Alexa 555 (Invitrogen) were used at a 1:200 dilution.

Microarray analyses.

Parasite gene expression was investigated by using a custom-designed tiling microarray (Roche NimbleGen Inc., Madison, WI). Each gene in the TashAT cluster was represented by a set of 45-mer oligonucleotides that were specific to that gene. cDNA was generated from 10 µg total RNA by using an oligo(dT) primer and tagged with 3′-Cy3 dye, after which labeled cDNA was hybridized to the array. Gene expression values were calculated from a robust multiarray average-normalized dataset (64).


Data set S1

Orthologue clustering in piroplasms and Plasmodium genes. Download Data set S1, XLSX file, 0.4 MB.

Figure S1

Venn diagram showing the overlap and number of assigned KOs among Torientalis, Tparva, Tannulata, and Bbovis. Genes of the four species are annotated with KOs found in KEGG metabolic pathways. The total numbers of KOs assigned to Torientalis, Tparva, Tannulata, and Bbovis were 263, 263, 273, and 264, respectively. The large number (255) of KOs common to the four species indicates strong similarity in terms of known metabolic pathways. Download Figure S1, TIF file, 0.5 MB.

Figure S2

(A) Gene expression across the TashAT family in Tannulata through differentiation from the macroschizont to the merozoite stage as determined by microarray analysis. Expression levels are shown as log2 intensity values. (B) Detection of Tash-a antigen following differentiation from the macroschizont stage (day 0) to the merozoite stage in vitro at 41°C (day 9). (a) Detection of macroschizonts by monoclonal antibody (MAb) 1C12 (green) in a day 0 (37°C) culture. (b) Lack of reactivity of anti-Tash-a to macroschizonts. (c) Detection of host and macroschizont nuclei by 4[prime],6-diamidino-2-phenylindole (DAPI) staining (same field as in panel b). (d) Detection of merozoites in a day 9 culture by MAb 5E1 (green). (e) Detection of merozoites by anti-Tash-a antibody (green). (f) Detection of merozoite nuclei by DAPI staining (same field as panel e). (g) Lack of reactivity of MAb 5E1 to macroschizonts. (h) Lack of reactivity of MAb to the His6 tag in a day 9 culture. (i) Detection of merozoite and host nuclei in a day 9 culture by DAPI staining (same field as panel h). Red color is due to Evans blue counterstain. Arrows denote merozoites. Bar = 15 µm. (C) Double labeled IFAT for detection of Tash-a antigen and ID11 rhoptry antigen following differentiation to the merozoite stage in vitro at 41°C (day 7). (a) Detection of Tash-a (green). (b) Detection of rhoptry antigen by MAb ID11. (c) Merged image of panels a and b. (d) Merged image of panel c with DAPI staining for host and parasite nuclei. Bar = 7 µm. Download Figure S2, TIF file, 5.6 MB.

Figure S3

TA9/TP9 family (PiroF0100041) expansion in the genomes of transforming Theileria species. (A) Chromosome synteny at the locus of the TA9/TP9 family (PiroF0100041). In Tannulata and Tparva, five genes were found at this tandemly arrayed gene family locus (red box). Conversely, in Torientalis, only a single gene, TOT020000921, was found at this locus and the similarity of this gene to the TA9/TP9 family was weak (TOT020000921 did not cluster into the PiroF0100041 ortholog group because of very low similarity). (B) Phylogenetic tree of the TA9/TP9 family in transforming Theileria and the corresponding gene in Torientalis. (C) Gene expression across the TA9/TP9 family in Tannulata through differentiation from the macroschizont to the merozoite stage as determined by microarray analysis. Expression levels are shown as log2 intensity values. Download Figure S3, TIF file, 0.7 MB.

Figure S4

Schematic diagram of a standardized pipeline for functional annotation. Download Figure S4, TIF file, 0.4 MB.

Table S1

The top 50 InterPro entries with the highest numbers of protein matches for every species.

Table S2

Expanded gene families specifically in the T. orientalis and transforming Theileria lineages.

Table S3

Transforming candidate genes (previously proposed).

Table S4

Statistics related to functional annotation.


We are grateful to the technical staff of the Sequence Technology Team at RIKEN GSC and the Department of Medical Genome Sciences, Graduate School of Frontier Sciences, at the University of Tokyo for their assistance.

This research was supported by Grants-in-Aid for Scientific Research (http://www.jsps.go.jp/english/e-grants/grants.html) from the Ministries of Education, Culture, Sports, Science, and Technology (17208026, 21248035) to C.S. and the Program of Founding Research Centers for Emerging and Reemerging Infectious Diseases (http://www.crnid.riken.jp/english/index.html) from the Ministries of Education, Culture, Sports, Science, and Technology. Worked performed by B.R.S. and W.W. was funded by a grant from the Wellcome Trust (083488/Z/07/Z).

The funder had no role in study design, data collection and analysis, the decision to publish, or preparation of the manuscript.


Citation Hayashida K, et al. 2012. Comparative genome analysis of three eukaryotic parasites with differing abilities to transform leukocytes reveals key mediators of Theileria-induced leukocyte transformation. mBio 3(5):e00204-12. doi:10.1128/mBio.00204-12.


1. Brown CG, Stagg DA, Purnell RE, Kanhai GK, Payne RC. 1973. Letter: infection and transformation of bovine lymphoid cells in vitro by infective particles of Theileria parva. Nature 245:101–103 [PubMed]
2. Irvin AD, Brown CG, Kanhai GK, Stagg DA. 1975. Comparative growth of bovine lymphosarcoma cells and lymphoid cells infected with Theileria parva in athymic (nude) mice. Nature 255:713–714 [PubMed]
3. Brown CG. 1990. Control of tropical theileriosis (Theileria annulata infection) of cattle. Parassitologia 32:23–31 [PubMed]
4. Hooshmand-Rad P, Hawa NJ. 1973. Malignant theileriosis of sheep and goats. Trop. Anim. Health Prod. 5:97–102
5. Shiels B, et al. 2006. Alteration of host cell phenotype by Theileria annulata and Theileria parva: mining for manipulators in the parasite genomes. Int. J. Parasitol. 36:9–21 [PubMed]
6. Dobbelaere D, Baumgartner M. 2009. Theileria, p 613–632 In Schaible UE, Haas A, editors. , Intracellular niches of microbes: a pathogens guide through the host cell. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany
7. Gardner MJ, et al. 2005. Genome sequence of Theileria parva, a bovine pathogen that transforms lymphocytes. Science 309:134–137 [PubMed]
8. Pain A, et al. 2005. Genome of the host-cell transforming parasite Theileria annulata compared with T. parva. Science 309:131–133 [PubMed]
9. Bishop RP, et al. 2009. Theileria, p. 191–224 In Nene V, Kole C, editors. , Genome mapping and genomics in animal-associated microbes. Springer-Verlag, Berlin, Germany
10. Onuma M, Kakuda T, Sugimoto C. 1998. Theileria parasite infection in East Asia and control of the disease. Comp. Immunol. Microbiol. Infect. Dis. 21:165–177 [PubMed]
11. Uilenberg G. 2011. Theileria sergenti. Vet. Parasitol. 175:386 [PubMed]
12. Shimizu S, Yoshiura N, Mizomoto T, Kondou Y. 1992. Theileria sergenti infection in dairy cattle. J. Vet. Med. Sci. 54:375–377 [PubMed]
13. Uilenberg G, Perié NM, Spanjer AA, Franssen FF. 1985. Theileria orientalis, a cosmopolitan blood parasite of cattle: demonstration of the schizont stage. Res. Vet. Sci. 38:352–360 [PubMed]
14. Fujisaki K, Kawazu S, Kamio T. 1994. The taxonomy of the bovine Theileria spp. Parasitol. Today 10:31–33 [PubMed]
15. Kubota S, Sugimoto C, Onuma M. 1996. Population dynamics of Theileria sergenti in persistently infected cattle and vector ticks analysed by a polymerase chain reaction. Parasitology 112(Pt 5):437–442 [PubMed]
16. Kamau J, et al. 2011. Emergence of new types of Theileria orientalis in Australian cattle and possible cause of theileriosis outbreaks. Parasit. Vectors 4:22 [PMC free article] [PubMed]
17. Kakuda T, et al. 1998. Phylogeny of benign Theileria species from cattle in Thailand, China and the U.S.A. based on the major piroplasm surface protein and small subunit ribosomal RNA genes. Int. J. Parasitol. 28:1261–1267 [PubMed]
18. Sato M, et al. 1993. Histological observations on the schizonts in cattle infected with Japanese Theileria sergenti. J. Vet. Med. Sci. 55:571–574 [PubMed]
19. Hikosaka K, et al. 2010. Divergence of the mitochondrial genome structure in the apicomplexan parasites, Babesia and Theileria. Mol. Biol. Evol. 27:1107–1116 [PubMed]
20. Schmuckli-Maurer J, et al. 2009. Expression analysis of the Theileria parva subtelomere-encoded variable secreted protein gene family. PLoS One 4:e4839 [PMC free article] [PubMed]
21. Kanehisa M, Goto S, Furumichi M, Tanabe M, Hirakawa M. 2010. KEGG for representation and analysis of molecular networks involving diseases and drugs. Nucleic Acids Res. 38:D355–D360 [PMC free article] [PubMed]
22. Chae JS, et al. 1999. A study of the systematics of Theileria spp. Based upon small-subunit ribosomal RNA gene sequences. Parasitol. Res. 85:877–883 [PubMed]
23. Li L, Stoeckert CJ, Jr, Roos DS. 2003. OrthoMCL: identification of ortholog groups for eukaryotic genomes. Genome Res. 13:2178–2189 [PMC free article] [PubMed]
24. Swan DG, et al. 2001. Characterisation of a cluster of genes encoding Theileria annulata AT hook DNA-binding proteins and evidence for localisation to the host cell nucleus. J. Cell Sci. 114:2747–2754 [PubMed]
25. Swan DG, Phillips K, Tait A, Shiels BR. 1999. Evidence for localisation of a Theileria parasite AT hook DNA-binding protein to the nucleus of immortalised bovine host cells. Mol. Biochem. Parasitol. 101:117–129 [PubMed]
26. Shiels B, et al. 1992. Disruption of synchrony between parasite growth and host cell division is a determinant of differentiation to the merozoite in Theileria annulata. J. Cell Sci. 101(Pt 1):99–107 [PubMed]
27. Baylis HA, Sohal SK, Carrington M, Bishop RP, Allsopp BA. 1991. An unusual repetitive gene family in Theileria parva which is stage-specifically transcribed. Mol. Biochem. Parasitol. 49:133–142 [PubMed]
28. Skilton RA, et al. 2000. A 32 kDa surface antigen of Theileria parva: characterization and immunization studies. Parasitology 120(Pt 6):553–564 [PubMed]
29. McKeever DJ, et al. 1994. Adoptive transfer of immunity to Theileria parva in the CD8+ fraction of responding efferent lymph. Proc. Natl. Acad. Sci. U. S. A. 91:1959–1963 [PMC free article] [PubMed]
30. Graham SP, et al. 2006. Theileria parva candidate vaccine antigens recognized by immune bovine cytotoxic T lymphocytes. Proc. Natl. Acad. Sci. U. S. A. 103:3286–3291 [PMC free article] [PubMed]
31. Graham SP, et al. 2008. Characterization of the fine specificity of bovine CD8 T-cell responses to defined antigens from the protozoan parasite Theileria parva. Infect. Immun. 76:685–694 [PMC free article] [PubMed]
32. MacHugh ND, et al. 2011. Extensive polymorphism and evidence of immune selection in a highly dominant antigen recognized by bovine CD8 T cells specific for Theileria annulata. Infect. Immun. 79:2059–2069 [PMC free article] [PubMed]
33. Mehlhorn H, Schein E. 1998. Redescription of Babesia equi Laveran, 1901 as Theileria equi Mehlhorn, Schein 1998. Parasitol. Res. 84:467–475 [PubMed]
34. Shaw MK. 2003. Cell invasion by Theileria sporozoites. Trends Parasitol. 19:2–6 [PubMed]
35. Weir W, et al. 2009. Highly syntenic and yet divergent: a tale of two Theilerias. Infect. Genet. Evol. 9:453–461 [PubMed]
36. Shimizu S, et al. 1988. Isolation of Theileria sergenti piroplasms from infected erythrocytes and development of an enzyme-linked immunosorbent assay for serodiagnosis of Tsergenti infections. Res. Vet. Sci. 45:206–212 [PubMed]
37. Kubota S, Sugimoto C, Onuma M. 1995. A genetic analysis of mixed population in Theileria sergenti stocks and isolates using allele-specific polymerase chain reaction. J. Vet. Med. Sci. 57:279–282 [PubMed]
38. Hattori M, et al. 1997. A novel method for making nested deletions and its application for sequencing of a 300 kb region of human APP locus. Nucleic Acids Res. 25:1802–1808 [PMC free article] [PubMed]
39. McMurray AA, Sulston JE, Quail MA. 1998. Short-insert libraries as a method of problem solving in genome sequencing. Genome Res. 8:562–566 [PMC free article] [PubMed]
40. Maruyama K, Sugano S. 1994. Oligo-capping: a simple method to replace the cap structure of eukaryotic mRNAs with oligoribonucleotides. Gene 138:171–174 [PubMed]
41. Mott R. 1997. EST_GENOME: a program to align spliced DNA sequences to unspliced genomic DNA. Comput. Appl. Biosci. 13:477–478 [PubMed]
42. Gardner PP, et al. 2009. Rfam: updates to the RNA families database. Nucleic Acids Res. 37:D136–D140 [PMC free article] [PubMed]
43. Lowe TM, Eddy SR. 1997. tRNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res. 25:955–964 [PMC free article] [PubMed]
44. Majoros WH, Pertea M, Salzberg SL. 2004. TigrScan and GlimmerHMM: two open source ab initio eukaryotic gene-finders. Bioinformatics 20:2878–2879 [PubMed]
45. Lomsadze A, Ter-Hovhannisyan V, Chernoff YO, Borodovsky M. 2005. Gene identification in novel eukaryotic genomes by self-training algorithm. Nucleic Acids Res. 33:6494–6506 [PMC free article] [PubMed]
46. Birney E, Clamp M, Durbin R. 2004. GeneWise and Genomewise. Genome Res. 14:988–995 [PMC free article] [PubMed]
47. Allen JE, Salzberg SL. 2005. JIGSAW: integration of multiple sources of evidence for gene prediction. Bioinformatics 21:3596–3603 [PubMed]
48. Imanishi T, et al. 2004. Integrative annotation of 21,037 human genes validated by full-length cDNA clones. PLoS Biol. 2:856–875 [PMC free article] [PubMed]
49. Yamasaki C, et al. 2008. The H-invitational database (H-InvDB), a comprehensive annotation resource for human genes and transcripts. Nucleic Acids Res. 36:D793–D799 [PMC free article] [PubMed]
50. Hunter S, et al. 2009. InterPro: the integrative protein signature database. Nucleic Acids Res. 37:D211–D215 [PMC free article] [PubMed]
51. Yamasaki C, et al. 2006. TACT: transcriptome auto-annotation conducting tool of H-InvDB. Nucleic Acids Res. 34:W345–W349 [PMC free article] [PubMed]
52. Yamasaki C, et al. 2005. Investigation of protein functions through data-mining on integrated human transcriptome database, H-Invitational database (H-InvDB). Gene 364:99–107 [PubMed]
53. Emanuelsson O, Brunak S, von Heijne G, Nielsen H. 2007. Locating proteins in the cell using TargetP, SignalP and related tools. Nat. Protoc. 2:953–971 [PubMed]
54. Altschul SF, et al. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389–3402 [PMC free article] [PubMed]
55. Katoh K, Kuma K, Toh H, Miyata T. 2005. MAFFT version 5: improvement in accuracy of multiple sequence alignment. Nucleic Acids Res. 33:511–518 [PMC free article] [PubMed]
56. Gouveia-Oliveira R, Sackett PW, Pedersen AG. 2007. MaxAlign: maximizing usable data in an alignment. BMC Bioinformatics 8:312 [PMC free article] [PubMed]
57. Castresana J. 2000. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol. Biol. Evol. 17:540–552 [PubMed]
58. Kishino H, Miyata T, Hasegawa M. 1990. Maximum-likelihood inference of protein phylogeny and the origin of chloroplasts. J. Mol. Evol. 31:151–160
59. Felsenstein J. 2004. Inferring phylogenies. Sinauer Associates, Inc, Sunderland, MA
60. Stamatakis A. 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22:2688–2690 [PubMed]
61. Whelan S, Goldman N. 2001. A general empirical model of protein evolution derived from multiple protein families using a maximum-likelihood approach. Mol. Biol. Evol. 18:691–699 [PubMed]
62. Yang Z. 1994. Maximum likelihood phylogenetic estimation from DNA sequences with variable rates over sites: approximate methods. J. Mol. Evol. 39:306–314 [PubMed]
63. Shiels BR, et al. 2004. A Theileria annulata DNA binding protein localized to the host cell nucleus alters the phenotype of a bovine macrophage cell line. Eukaryot. Cell 3:495–505 [PMC free article] [PubMed]
64. Irizarry RA, et al. 2003. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4:249–264 [PubMed]
65. Carver TJ, et al. 2005. ACT: the Artemis Comparison Tool. Bioinformatics 21:3422–3423 [PubMed]

Articles from mBio are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...