Transfection of the Primate Malaria Parasite Plasmodium knowlesi Using Entirely Heterologous Constructs

Plasmid pDT.Tg23 and pchD5.1/C3 have previously been described (5, 6). Plasmid pD.D_B.D. contains the same elements as pMD204 (3) except that the selection cassette was cloned into pUC-19 for pD.D_B.D. instead of pBluescript, and the elements (upstream, ORF and downstream) were engineered so that the ORF is readily replaceable through excision with BamHI. The pyrimethamine-resistant M2M3 mutant form of the T. gondii dhfr-ts gene (7) was amplified from pDT.Tg23 by PCR, using primers 5′-CGTGATCAATGCATAAAACCGGTGTGTC-3′ (TOX3) and 5′-CGTGATCAAAGCTTCTGTATTTCCGC-3′ (TOX4). PCR with pfu polymerase (Stratagene Inc., La Jolla, CA) yielded an amplified product that was kinated, gel-purified, and cloned into the blunted BamHI site of plasmid pD.D_B.D. (replacing the P. berghei dhfr-ts) to yield pD.D_T.D. Plasmids were purified using Plasmid Mega columns (Qiagen, Chatsworth, CA).

A P. knowlesi (Nuri strain) (15) infection was initiated in a female rhesus monkey (Macaca mulatta) by intravenous injection of 1 × 10⁵ parasites. Parasitemia was monitored daily on blood obtained from finger pricks. When 35% of erythrocytes were infected with mature schizonts, blood was collected by cardiac puncture. After centrifugation (450 g, 10 min, room temperature) the top brown layer of the erythrocyte pellet, containing >90% schizonts, was collected. Leukocytes were removed from this material using PlasmodiPur filters (Eurodiagnostica, Apeldoorn, the Netherlands) (16). Schizonts were then suspended in either PBS (3) or incomplete Cytomix (120 mM KCl, 0.15 mM CaCl₂, 2 mM EGTA, 5 mM MgCl₂, 10 mM K₂HPO₄/ KH₂PO₄, 25 mM Hepes, pH 7.6) (2) at a concentration of 5 × 10⁹ schizonts/ml.

Two different plasmid mixtures were prepared: Mix 1 consisted of pD.D_B.D. and pD.D_T.D. (mixed 1:1 wt/wt) and Mix 2 consisted of pDT.Tg23 and pchD5.1/C3 (mixed 1:1 wt/wt). For each electroporation, a total of 100 μg plasmid of Mix 1 or Mix 2 was added to 5 × 10⁸ schizonts in a 0.4-cm electroporation cuvette and electroporated using a Bio-Rad Gene Pulser using the following conditions (previously established for P. falciparum and P. berghei [2, 3]). DNA dissolved into 85 μl TNE (10 mM Tris-HCl, 100 mM NaCl, 5 mM EDTA, pH 7.5) and 115 μl PBS was mixed with schizonts in the same buffer and electroporated at either 600, 800, or 1,200 V, at a capacitance of 25 μF (time constants ranged between 1.1–1.5 ms). DNA dissolved into 700 μl incomplete Cytomix was mixed with schizonts in Cytomix and subjected to a pulse of either 1,500, 2,000, or 2,500 V, at a capacitance of 25 μF and a resistance of 200 \xbd (time constants ranged between 0.7–0.8 ms). Samples electroporated under these conditions were pooled, placed on ice for 5–8 min and injected intravenously into two non-splenectomized rhesus monkeys. Monkey R3106 received pooled electroporated samples of Mix 1 and monkey R3126 received pooled electroporated samples of Mix 2. Starting 40 h after injection of schizonts both monkeys orally received 2 mg/kg pyrimethamine per day, supplemented once a week with 3.5 mg folinic acid to counteract the bone marrow suppression caused by pyrimethamine (17). The parasitemia of the two monkeys was monitored daily. After 11 d of pyrimethamine pressure, blood was collected by cardiac puncture and leukocytes were removed by PlasmodiPur filtration. Parasite DNA was isolated and analyzed according to standard protocols.

Based on the successful use of mutated dhfr-ts genes conferring pyrimethamine resistance as selectable markers in transfection systems of P. berghei and P. falciparum, we transfected P. knowlesi with plasmid constructs containing resistant forms of the dhfr-ts gene of P. berghei or T. gondii (Table 1). 36 h after inoculation of transfected schizonts in monkey R3106 and R3126, newly invaded parasites were readily detectable in thin smears. After pyrimethamine treatment was initiated, parasitemias rapidly dropped to levels undetectable by thickfilm analysis, confirming the sensitivity of P. knowlesi to this drug. However, under continuous pyrimethamine administration, parasitemias in both monkeys rose to detectable levels by day 8. On day 12, when >1% of the red blood cells were infected, the pyrimethamine-resistant parasites were collected for further analyses (Table 1).

To assess which genes and control regions are active in P. knowlesi, we separately introduced two different plasmid combinations into P. knowlesi. Monkey R3106 received schizonts transfected with Mix 1, containing plasmid constructs with P. berghei dhfr-ts or T. gondii dhfr-ts flanked in both cases by control regions of P. berghei dhfr-ts. Monkey R3126 received schizonts transfected with Mix 2 containing a plasmid construct with T. gondii dhfr-ts flanked by P. falciparum control regions and a non-selectable plasmid to control for the possibility of continued presence of plasmid which does not confer pyrimethamine resistance.

DNA was isolated from the parasites of monkeys R3106 and R3126 and hybridized with probes against P. berghei dhfr-ts, T. gondii dhfr-ts or P. falciparum cam (Fig. 1 A). Hybridization of DNA of parasites from both monkeys was evident with the T. gondii dhfr-ts probe, indicating that, consistent with findings in the P. falciparum and P. berghei systems (5, our unpublished observation), T. gondii dhfr-ts was effective in conferring pyrimethamine resistance to P. knowlesi. The plasmid containing P. berghei dhfr-ts was not detected by this analysis. However, PCR analysis of DNA from parasites of monkey R3106 was positive for P. berghei dhfr-ts (data not shown). The presence of only minor amounts of pD.D_B.D. compared with pD.D_T.D. might suggest that parasites containing P. berghei dhfr-ts are overgrown by parasites containing T. gondii dhfr-ts. In the P. berghei system T. gondii dhfr-ts was found to confer a 10–100-fold higher pyrimethamine resistance to the parasites than P. berghei dhfr-ts (Janse, C.J., unpublished data). This may result in a selective advantage for parasites containing the T. gondii dhfr-ts under drug pressure, although other effects on growth kinetics, for example direct effects of the expression product of both plasmids cannot be ruled out.

The presence of plasmid pD.D_T.D. and plasmid pDT.Tg23 in resistant parasites was confirmed by rescue experiments through transformation of E. coli with parasite DNA. Unrearranged plasmids were recovered, as was shown by restriction analysis (Fig 1 B). We failed to rescue plasmid pD.D_B.D. and pchD5.1/C3.

To confirm that plasmids had been replicated in a eukaryotic environment, susceptibility to cleavage by MboI was evaluated (18). Plasmids isolated from E. coli were not susceptible to MboI digestion, but were susceptible to DpnI, an isoschizomer that is active when the adenine in the recognition site is methylated, as occurs in prokaryotic systems (5, 18). Plasmids isolated from parasites from both monkeys were susceptible to MboI digestion, demonstrating their eukaryotic replication (Fig 1 C). Transcription of T. gondii dhfr-ts in P. knowlesi was shown by hybridization with a Northern blot containing RNA isolated from parasites of monkey R3106. A transcript with a size of 2.4 kb was detected (not shown), comparable to the size of the transcript produced in P. berghei by plasmid pD.D_B.D. (Tomás, A.M., unpublished data).

Considering the phylogenetic distance between P. knowlesi, P. berghei, and P. falciparum (19, 20) and the large differences in the GC-contents of the genomes of these species (18% for both P. berghei and P. falciparum and 30% for P. knowlesi [21]), it is of interest that control regions of both P. falciparum and P. berghei were functional (were able to drive expression of the dhfr-ts genes) in P. knowlesi. This finding may be related to specific characteristics of the genome composition of P. knowlesi. In a study using CsCl-density centrifugation it has been shown that the genomes of the closely related species P. vivax and P. cynomolgi, separate into highdensity (GC-rich) and low-density (AT-rich) components. Although the genome of P. knowlesi contains only a highdensity component (21), specific sequences of the P. knowlesi genome do hybridize with the low-density component of P. cynomolgi, indicating that AT-rich sequences are present in the P. knowlesi genome. Additionally, in a comparison of introns of different Plasmodium species, introns in both P. vivax and P. knowlesi were found to have either a GC-rich or an AT-rich composition (22). Introns are likely to reflect their genomic environment, and therefore, it was suggested that the differences in the introns of these genes reflect their maintenance in distinct isochores (very long DNA segments with fairly homogeneous base compositions [23]). This suggests that the genome organization of P. knowlesi may share the characteristics of both P. vivax and P. cynomolgi. If this is the case, transcription machinery in P. knowlesi may be able to respond to control regions of varied base composition, and this would explain why in P. knowlesi, gene control regions derived from phylogenetically distinct parasites with a different overall genomic GC-content are functionally active.

In summary, we have shown that P. knowlesi can be stably transfected using entirely heterologous constructs, offering a model system which allows investigations into parasite-host interactions, in hosts closely related to humans. The successful use of heterologous constructs in this parasite will facilitate the creation of transgenic and knockout parasites through integration-dependent transfection. The phylogenetic distance over which the control elements have been shown to be effective in this study suggests that similar constructs may also be effective in P. vivax and in other important non-human primate malaria parasites such as P. cynomolgi and P. fragile.

We are grateful to Dr. Tom Wellems for providing vector pDT.Tg23 and Dr. Alan Cowman for providing vector pchD5.1/C3. We thank Eugène Uytendaal and Kristel de Brouwer for excellent technical assistance, members of the Animal Science Department of the Biomedical Primate Research Centre for expert animal care and Milly van Dijk for helpful advice. The research protocol was approved by an independent animal care and use committee and performed according to Dutch and European laws.

This work was supported by European Commission, Directorate General XII (INCO-DC programme) contracts CT95-0022 and CT94-0275. Dr. A.M. Tomás was cofinanced by a post-doctoral grant from Programa Praxis (JNICT, Portugal) and by Instituto de Ciências Biomédicas de Abel Salazar, Porto University.