The preliminary sequence data of T. parva (Muguga) chromosome 1 13 was run through the gene finding programs GlimmerM 14 and Phat 15, and the proteins encoded by the predicted T. parva genes subjected to a variety of analyses to identify secreted or membrane-anchored proteins. Fifty-five ORFs encoding candidate antigens were selected for screening, of which 36 were successfully cloned into pTargeT 16. Each of the 36 ORFs were transiently transfected into autologous immortalized skin fibroblasts (iSF) and screened with a panel of schizont specific polyclonal CTL lines or CTL clones generated from 13 T. parva immune cattle representing a spectrum of bovine leukocyte antigen (BoLA) class I genotypes. CTL from four animals (BW002, BW013, BW014 and D409) exhibiting high lytic activity against autologous schizont-infected lymphoblasts (Fig. 1A), all showed significant IFN-γ ELISpot responses to iSF transfected with cDNA Tp01_56 (Fig. 1B; p < 0.01). CTL lines from the other nine cattle did not recognise iSF transfected with any of the selected ORF's and no false positives were detected during the screening (data not shown). The ability of CTL to lyse iSF transfected with cDNA Tp01_56 was assessed using a ⁵¹Chromium release assay (Fig. 1C). CTL exhibited specific lysis against cDNA Tp01_56 transfected cells and not against cells transfected with an irrelevant cDNA cloned into the same vector. Lysis of cDNA Tp01_56 transfected cells by CTL was inhibited by pre-incubation of target cells with a blocking mAb against MHC class I (IL-A88; ILRI, Nairobi, Kenya). This observation together with the inability of CTL to lyse cDNA Tp01_56 transfected allogeneic cells suggested that the recognition and lysis was MHC class I restricted. The gene encoding cDNA Tp01_56 was assigned the name Tp2 (GenBank accession number XM_760490).

The Tp2 gene encodes 174 amino acid residues, has a calculated mass of 19,141Da and an N-terminal signal peptide with a predicted cleavage site between residues 23 and 24. The predicted mature protein contains twelve cysteine amino acid residues suggesting that the tertiary structure may be dependent on di-sulfide linkages. The nucleotide and deduced amino acid sequences of the Tp2 gene are shown in Fig. 2A. Reverse transcriptase(RT)-PCR revealed transcription of Tp2 in sporozoite, schizont and piroplasm life-cycle stages of T. parva (Fig. 2B) and Northern blots detected the presence of ~1 kb transcript in schizont and piroplasm RNA (Fig. 2C). Transcriptome analysis of T. parva using massively parallel signature sequencing indicated that Tp2 is expressed at a relatively high level (1069 tpm) in the schizont stage 17. The cDNA clone was isolated from the schizont cDNA library contains a 1026 bp insert with 105 and 396 bp 5' and 3' UTR's, respectively. Searches of protein and DNA databases using BLAST identified a homolog of Tp2 in Theileria annulata, a molecule described as a membrane-associated antigen, TaD 18.

The CTL epitopes on Tp2 were mapped using a library of 12-mer overlapping synthetic peptides covering the full length of the Tp2 protein. Recognition of Tp2 peptides was assessed by IFN-γ ELISpot using autologous iSF as antigen-presenting cells. Fig. 3A shows the results of screening the Tp2 peptide library with CD8⁺ polyclonal CTL lines from animal BW002, BW013, BW014 and CTL clone T29.10 from animal D409. This allowed the identification of four epitope-containing regions. Minimal length antigenic peptides were defined by screening all possible 8-, 9-, 10- and 11-mers covering the regions identified above. The four minimal length antigenic peptides recognized by CTL were 11-mers, SHEELKKLGML (BW002) and KSSHGMGKVGK (BW013) and overlapping 9-mers, FAQSLVCVL (D409) and QSLVCVLMK (BW014) (data not shown).

As a prelude to directly testing the vaccine potential of the Tp2 antigen in cattle, an experiment was conducted to look for a Tp2-specific component within the protective CD8⁺ T cell response following the challenge of three ECF immune cattle with a lethal dose of sporozoites. All animals were solidly resistant to challenge, exhibiting only a transiently detectable parasitosis and no fever (data not shown). CD8⁺ T cells enriched from peripheral blood responded to Tp2 antigenic peptides as measured by IFN-γ ELISpot assay (Fig 4). Responses to the Tp2 antigenic peptides were significantly greater than those to peptides from other regions of Tp2 from day 9–11 post-challenge (Fig. 4A–C; p < 0.05). Both the kinetics and magnitude of the response correlate well with those previously described for schizont-specific CTL in efferent lymph and blood following challenge of immune animals 5. Attempts to detect schizont and Tp2 specific cytotoxic activity directly in the blood of cattle post-challenge failed to demonstrate any significant lysis (data not shown). However, 14 days post-challenge, BW002 PBMC stimulated in vitro with autologous irradiated schizont- infected cells and tested for cytotoxicity 7 days later exhibited significant cytolytic activity against both schizont-infected cells and iSF pulsed with the Tp2 antigenic peptide SHEELKKLGML (Fig. 4D; p < 0.001). This finding is consistent with the hypothesis that the CTL precursor frequency in peripheral blood was too low to be measured directly but could be increased to detectable levels following a single re-stimulation in vitro.

Preliminary contigs of the T. parva genome were searched against a non-redundant database of proteins extracted from GenBank. The search results were used to produce a set of putative T. parva genes that encoded highly conserved eukaryotic genes. This set of genes was complemented with a number of full length genes isolated from a cDNA library made from purified schizonts and were used to train the gene finding programs GlimmerM 14 and Phat 15, which were subsequently run against all of the preliminary contigs to produce gene models for the entire preliminary genome sequence. The proteins encoded by the predicted T. parva genes on chromosome 1 were searched against a non-redundant protein database using WU-BLAST 22, and predictions of signal peptides and signal anchors 23 and transmembrane domains 24 analysed. The results were reviewed and a set of 55 genes encoding candidate antigens was selected for cloning and screening 16.

Open reading frames (ORFs) less than 3.1 kb were successfully amplified by OneStep RT-PCR kit (QIAGEN Ltd., Crawley, UK) using RNA purified from T. parva (Muguga) schizont-infected lymphoblasts. Thermal cycles were: Genes up to 1 kb: 50°C 30 minutes; 95°C 15 minutes; 94°C 30 seconds; 55°C 30 seconds; 72°C 1 minute; 35 times from 94°C cycle and finally 72°C 10 minutes. Genes between 1 and 2 kb: 50°C 30 minutes; 95°C 15 minutes; 94°C 1 minute; 55°C 1 minute; 72°C 1 minute; 35 times from 94°C cycle and finally 72°C 10 minutes. Genes between 2 and 3 kb: 45°C 30 minutes; 95°C 15 minutes; 94°C 10 seconds; 55°C 1 minute; 68°C 3 minutes; 35 times from 94°C cycle and finally 68°C 10 minutes. Amplified genes were purified from agarose gels by QIAquick Gel Extraction kit (QIAGEN) and cloned into eukaryotic expression T-vector pTargeT (Promega, Madison, WI, USA). Ligated samples were electroporated into JM109, and colonies were screened by PCR using gene specific internal forward primers and vector specific reverse primer (5'GAGCGGATAACATCACACAGG3'). Positive colonies were cultured in 2 ml and plasmids purified by QIAprep Spin Miniprep kit (QIAGEN) and sequenced by ABI PRISM 377 DNA sequencer (Applied Biosystems, Foster City, CA, USA). Although 51 of 55 selected genes were amplified, 36 were cloned successfully in pTargeT 16. Some of the genes were consistently cloned in wrong orientation suggesting the gene products were toxic to E. coli. Some genes were amplified by the reverse primer only and missed the 5' end.

Total RNA from sporozoites, schizonts and piroplasms was reverse-transcribed into ss-cDNA by reverse transcriptase using oligo-(dT) as primer following the recommendation of the supplier (Invitrogen). After removal of RNA complementary to the cDNA by treatment with RNase H and purification with phenol-chloroform, the ss-cDNA was PCR-amplified in the presence of specific primers used to generate the Tp2 ORF (5'-ATGAAATTGGCCGCCAGA-3' and 5'-CTATGAAGTGCCGGAGGCTTCTCC-3'). Northern blots of RNA from the schizont and piroplasm stages of T. parva were probed with the Tp2 ORF DNA fragment. This probe was also used to screen a schizont cDNA library and restriction enzyme analysis was performed by digestion with Eco RI and Bam HI.

All animal experimentation was reviewed and approved by the ILRI Institutional Animal Care and Use Committee. Four Boran (Bos indicus), 1 Jersey (Bos taurus), 1 Holstein-Friesian (Bos taurus) and 4 crossbreds, selected on the basis of BoLA type, were immunized against T. parva Muguga stock, challenged after 3 months and schizont-specific polyclonal CTL lines and clones established as described 16.

Immortalized skin fibroblasts (iSF) were transfected in 96-well plates with cDNA clones (100 ng/well) using Fugene-6 (Roche Diagnostics GmbH, Mannheim, Germany) and cultured for 24 hours.

Transfectants were co-cultured with schizont-specific CTL and recognition assessed using an IFN-γ ELISpot assay 16. Transfectants and schizont-infected cells were labelled with ⁵¹Chromium (Amersham Biosciences Europe GmbH, Freiburg, Germany) and lysis by schizont-specific CTL lines assessed 25.

A peptide library of Tp2 was generated that contained every 12 mer, 11 mer, 10 mer and 9 mer offset by 2 amino acids from the protein sequence (Cleaved PepSet; Mimotopes, Clayton, Australia). Peptides were prepared by truncations of 12 mers at the N-terminus and supplied lyophilized with each tube containing a nominal 12 mer and the 9, 10, 11 mer truncations with the same C-terminus. Peptides were dissolved in 400 μl 50% (v/v) DNA synthesis grade acetonitrile/water (Applied Biosystems). Peptides were further diluted to 10 μg/ml in complete RPMI-1640 and 10 μl added to triplicate wells of an ELISpot plate. Autologous iSF were adjusted to a density of 4 × 10⁵/ml and 50 μl added to wells containing peptides. The plates were incubated at 37°C for 1 hour before 50 μl of CTL at a density of 2 × 10⁵/ml were added to each well. Plates were incubated and developed as described above. Based on the results of screening the peptide library, individual 8, 9, 10 or 11 mer peptides were synthesized and tested as described above. Peptide-pulsed iSF were prepared as targets for ⁵¹Chromium release assays by overnight incubation with peptide at 1 μg/ml in complete DMEM. Cells were harvested, labelled and assayed as described above.

Immune cattle, BW002, BW013 and BW014, whose schizont-specific CTL had recognized Tp2, were challenged with a lethal dose of T. parva (Muguga) sporozoites and bled daily from day 6 to 13 post-challenge. CD8⁺ T cells and CD14⁺ monocytes were purified from PBMC by MACS magnetic cell sorting according to the manufacturer's instructions (Miltenyi Biotec, Gergisch Gladbach, Germany). CD8⁺ T cells were sorted indirectly using a monoclonal antibody specific for bovine CD8 (IL-A105; ILRI, Nairobi, Kenya) followed by incubation with goat anti-mouse IgG microbeads (Miltenyi Biotec). CD14⁺ monocytes were sorted directly by incubation with CD14 microbeads (Miltenyi Biotec). CD8⁺ T cells (2.5 × 10⁵/well) and CD14⁺ monocytes (2.5 × 10⁴/well) were added to IFN-γ ELISpot plates containing Tp2 peptides (1 μg/ml final concentration) and were incubated and developed as described 21. To recall Tp2 specific CTL responses, PBMC were stimulated with autologous infected lymphoblasts 14 days post-challenge, viable cells were harvested 7 days post-stimulation and lytic activity against infected lymphoblasts and Tp2 peptide pulsed iSF assessed as described above.

Analysis of variance (ANOVA) was used for the analysis of fixed effects on different traits using SAS Release 8.2 (SAS Institute Inc., Cary, USA).