High efficacy in vitro selection procedure for generating transgenic parasites of Plasmodium berghei using an antibiotic toxic to rodent hosts

The malaria parasite Plasmodium berghei is one of the main rodent malaria models. A shortcoming of this model parasite is its low flexibility in genetic manipulation. As this parasite cannot be continuously propagated in cell cultures, in vivo drug selection procedures are necessary to isolate genetic mutants. Drugs harmful to rodents therefore cannot be used for drug selection, which restricts the range of genetic manipulation. In this study, we addressed this problem by establishing a novel in vitro culture drug selection method, which we used in combination with other established methods to successfully isolate genetically manipulated parasites. The target mutants were enriched to the desired level within two weeks. We show that our system can also be used for sequential genetic manipulation of parasites carrying the traditionally used selection markers, demonstrate the procedure’s versatility, and show its use in isolating specific genetically manipulated parasites. This novel in vitro selection method increases the number of available selection markers, allowing more extensive genetic manipulation in malaria parasite research.

thus be used as a secondary marker in sequential manipulation experiments. However, sequential manipulation is challenging because hdhfr only confers slight resistance to WR99210 6 , and only two reports describe sequential gene disruptions using these markers 2,9,10 . A marker recycling method using negative selection was developed to address this issue 7,11,12 , but did not prove to be an effective solution.
In this study, we developed a novel selection method to overcome the genetic manipulation problems in P. berghei models. We focused on the puromycin-N-acetyltransferase (pac)-puromycin system, which is commonly used in mammalian cells 13,14 . pac can be isolated from Streptomyces alboniger and shows resistance to the aminonucleoside antibiotic, puromycin 15 . A previous study reported that pac might be used as a selection marker for blood stage long-term in vitro cultures of P. falciparum, the causative agent of tropical malaria in humans 15 . As puromycin is severely toxic to rodents and thus cannot be used for P. berghei in vivo, we developed a pac-puromycin system based on a novel short-term in vitro culture method.
We identified a suitable puromycin concentration for drug selection by selecting pXL/ hdhfr-pac-egfp-transfected parasites at various puromycin concentrations. The empirical optimal concentration was 1.0 µg/ml (1.8 µM) and the target mutant was not obtained at concentrations above 1.5 µg/ml. Generation of transgenic parasites using pac markers through piggyBac transposon. To determine whether pac can be used as a positive selection marker for genetically manipulated parasites, we used it to select pXL/hdhfr-pac-egfp-transfected parasites (Fig. 2a). Three selections were performed after transfection: the first after two days, the second after 8-9 days, and the third after 13-14 days. After the second selection, over 95% of parasites were expressing enhanced green fluorescent protein (eGFP) (Fig. 2b,c). The target mutant ratio increased significantly from first to second selections. There was no significant difference between second and third selections (Fig. 2b). A typical transfectant line was also analyzed using flow cytometry. Over 95% of parasites were expressing eGFP after the third selection (Fig. 2f). Three clones were obtained and analyzed using Southern blot analysis (Fig. 2d) to confirm genomic integration. The fragments of integrated pac cassettes (4.8, 5.7, 6.2 kbp) were detected with a pac probe. We did not detect a signal corresponding to plasmid size. Plasmid rescues on parasite DNA extracted from these clones confirmed the absence of episomally maintained plasmids. Sequence analyses of inverse PCR products showed that these three clones had single copy insertions in unique loci (Table 1). To determine the inhibitory effect of Pac expression on parasite growth, the growth of the three pac clones was analyzed. Clones 1 and 2 showed no significant difference from wild type parasites, but clone 3 exhibited significantly higher parasitemia at seven and eight days post infection (Fig. 2e). These results demonstrate that pac represents a valid selection marker in P. berghei.
Generation of transgenic parasites using pac through double homologous recombination. We investigated whether pac can be used for gene targeting by assembling constructs that contained pac under the control of an hsp70 promoter and terminator (Fig. 3a). We used a GFP-expressing parasite 16 under the control of an hsp70 promoter and terminator as a parental strain for transfection. As shown in Fig. 3a, the gfp expression cassette would be replaced by the pac expression cassette after the homologous recombination event, and GFP expression would be lost.
Three selections were performed after transfection: the first after 2 days, the second after 8 days, and the third after 12 days. After the second selection, more than 99% of parasites were GFP-negative (Fig. 3b,d). The mutant ratio increased significantly from first to second selections, but there was no significant difference between second and third selections (Fig. 3b). A typical transfectant line was also analyzed using flow cytometry. More than 99%  of parasites were GFP-negative after the second selection (Fig. 3e). From two of these experiments we obtained two independent clones, which were analyzed using Southern blot analyses (Fig. 3c). The 8.9-kbp fragment containing the gfp cassette was only detected in the parental GFP-expressing parasite with the gfp probe. The identical 8.9-kbp signals were detected in both clones, but not in the wild type, with the pac probe and the 5′hsp70 probe. Based on these results, the in vitro pac-puromycin system is suitable for gene targeting as well as random insertion by transposon.
Sequential genetic manipulation of pac to pbdhfr-ts parasites. We demonstrated sequential genetic manipulation using pac and pbdhfr-ts markers for gene targeting. To this end, we assembled a targeting vector composed of a pac cassette flanked by sporozoite microneme proteins essential for the cell traversal 2 (spect2, PBANKA_1006300) 17 gene sequence (Fig. 4a). A GFP-expressing parasite 16 containing pbdhfr-ts marker was used as the parental strain for transfection. The parasite was transfected with the targeting construct, after which puromycin selection was applied. The ratio of the target mutant was determined using genomic Southern blot analysis (Fig. 4b). KO locus signals (3.7 kbp) appeared after the first selection but became dominant after the second selection. We confirmed that GFP was continuously expressed from before transfection to after third selection (Fig. 4c). The KO of spect2 was confirmed by PCR (Fig. 4d). These results show that the in vitro pac-puromycin system can be used in combination with traditional pbdhfr-ts markers to establish double mutant parasites.
Generation of mutants using the pac::egfp fusion marker. We investigated whether the pac::egfp fusion gene can be used as a marker by assembling a piggyBac transposon vector that contained the fusion gene under the control of an hsp70 promoter and terminator (Fig. 5a). To confirm that pac::egfp functionally expressed eGFP, in vivo pyrimethamine selection of pXL/hdhfr-pac::egfp transfected parasites was performed prior to in vitro puromycin selection. The expression of eGFP was confirmed by fluorescence microscopy.
After transfection of pXL/hdhfr-pac::egfp, in vitro puromycin selection was conducted three times. Fluorescence analyses showed that more than 60% of parasites expressed egfp after the second selection (Fig. 5b). The mutant ratio increased significantly between first and second selections, but there was no significant difference between second and third selections (Fig. 5b). Fluorescence microscopy confirmed the expression of the eGFP signal (Fig. 5c). Based on these results, pac::egfp is suitable as a dual marker for fluorescence and drug selection.
Application of the in vitro selection method to the hdhfr-WR99210 system. To examine the universality of our in vitro selection method, we used the hdhfr-WR99210 system as a model. Preliminary experiments showed that the optimal WR99210 concentration for selection was 6.25 ng/ml (1.6 × 10 −2 µM). A wild type parasite was transfected with pXL/hdhfr-pac-egfp, after which WR99210 selection was performed. Proportions of eGFP-positive parasites increased significantly up to the fourth and fifth selections, and the mutant ratio reached more than 90% after the fifth selection ( Fig. 6a,b). A typical transfectant line was also analyzed using flow cytometry. Over 90% of parasites were expressing eGFP after the fifth selection (Fig. 6c). These results show that the employed in vitro selection method may be suitable for application in drug selection systems other than the pac-puromycin system.

Discussion
Reverse genetics has been a powerful approach in studying the biology of malaria parasites 3,4 . However, the limited number of selection markers has hampered the development of the method 6,7 . To provide further selection markers, we developed a pac marker in P. berghei based on a novel in vitro selection method.
We established this method by combining repeated short-term in vitro cultures with parasite recovery in vivo. The pac marker conferred sufficient resistance to the P. berghei parasite for puromycin selection. Its resistance made it possible to isolate the mutant parasites using piggyBac transposon mutagenesis and gene targeting by homologous recombination. We showed that pac::egfp was also an effective drug selection marker and might constitute a useful tool for imaging experiments. Blood stage parasite growth analyses of Pac-expressing clones indicated that our pac marker cassette did not inhibit parasite growth. While currently available positive selection markers for this parasite only confer resistance to antifolates 2 , pac is independent of antifolate pathways 15 . Thus, the pac-puromycin system can be used in conjunction with the traditional antifolate resistant markers (hdhfr and pbdhfr-ts). The small size of pac (600 bp) is one of its advantages; in Plasmodium, assembling targeting constructs causes a transfection bottleneck because large-sized constructs often result in instability in E. coli caused by the AT-rich Plasmodium genome.
We confirmed the universality of this in vitro selection method by using the traditional hdhfr-WR99210 system as a model. The results indicate that our in vitro method has potential for use in other drug selection systems already established for P. falciparum 18 . In addition, this system could also be used for negative selection in vitro, to recycle markers. A further benefit of our method is its relative cheapness, as WR99210 is approximately 100 times more expensive than pyrimethamine (based on Sigma Aldrich prices). The new system significantly reduces the amount of WR99210 required (0.001%) relative to the conventional in vivo method. Our method can enrich mutant parasites by more than 90% within 2 weeks after transfection, using 2 mice. This level of efficacy has great advantages for cloning mutant parasites, and may reduce the number of mice required to obtain the desired production. Traditional cloning procedures involve using the in vivo limiting dilution method 1 . In our experience, the percentage of target KO parasites obtained when using the conventional Confirmation of the predicted recombination events was achieved with primer combinations specific for 5′ (P3 + P4) or 3′ integration (P5 + P6). An additional primer combination (P1 + P2) was used to assess integration of the pac cassette. The primer sequences can be found as Supplementary Table S1. P: primer.
Scientific RepoRts | 7: 4001 | DOI:10.1038/s41598-017-04244-0 selection method is around 10-20%, so double limited dilution procedures and at least 20 mice are required to isolate a single cloned parasite. Our in vitro pac-puromycin system, however, allows us to isolate a clone using fewer than five mice. This reduction may save significant expense, time, and labor in rearing mice, and has obvious animal welfare benefits. Furthermore, the method allows the ready generation of several mutants in parallel. In some cases, the drug-selected parasites can be used directly without cloning or flow-sorting steps 2 .
A further advantage of this system is its flexible application for sequential manipulation to mutant parasite generated using traditional antifolate-resistant markers. While various strategies have successfully used negative selection in the P. berghei model 7, 12 , the disadvantage of the traditional marker recycling method is its complexity and the time requirement of several months for the isolation of a double mutant parasite 7,19 . The "gene out marker out" (GOMO) strategy was developed to address this problem 12 . However, this method is inapplicable  to parasites already carrying antifolate resistant markers as it requires positive selection using antifolates. Our in vitro pac-puromycin system can readily be used in sequential manipulation experiments (for example when employing gene complementation methods) with the mutant parasites established in past malaria research, using any type of selection method including traditional antifolate markers. In our experiment, the target cloned parasite was obtained within 3 weeks using less than 10 mice.
In summary, we succeeded in establishing a novel in vitro selection method using the pac marker in P. berghei. This in vitro pac-puromycin selection system exhibited high selection efficacy relative to the conventional in vivo method. The method should allow the generation of multiple mutants with greater flexibility and help develop our in-depth understanding of malaria parasite biology.

Methods
Ethics statement. This study was carried out in strict accordance with the recommendations in the Guide for Laboratory Animals of the Obihiro University of Agriculture and Veterinary Medicine. The protocol was approved by the Committee of Animal Experiments of the Obihiro University of Agriculture and Veterinary Medicine (permit number 28-91).
Experimental animals and parasites. ICR and BALB/c (five-week old) mice were obtained from CLEA (Tokyo, Japan). We used the P. berghei ANKA strain (obtained from Dr M. Torii, Ehime University) and a strain constitutively expressing GFP 16 (obtained from Dr M. Yuda, Mie University).
The transposase expression vector EGF-pgT contained a transposase gene excised from pHTH (obtained from the BEI Resources Repository, MRA-912, deposited by John Adams), under the control of an ef-1α promoter and hsp70 terminator, cloned into a pBluescript (pBS) vector 20 .
Elements of pBS/spect2 KO were generated into a pBS backbone. The pac expression cassette was flanked by 5′ and 3′ sequences of spect2 (pBS/spect2 KO) amplified by PCR (Table S1).

Parasite transfection and drug selection in vitro.
Parasite transfection experiments were run following standard protocols 23 . Briefly, about 1 ml of infected blood (1.0-3.0% parasitemia) collected from an ICR mouse by heart puncture was cultured in 50 ml of RPMI 1640 medium (cat no. 23400-021, Thermo Fisher Scientific, MA, USA,) supplemented with NaHCO 3 (0.85 g/l), neomycin sulfate (50 mg/l), and 20% fetal bovine serum for schizonts collection. Schizonts were purified by Nycodenz density gradient centrifugation 24 . Obtained schizonts were sufficient for at least 5 independent transfections. Transfection was performed with a Nucleofector 2b device (Lonza, Basel, Switzerland) using a T Cell Nuclofector Kit (Lonza) under the U-33 program. All in vitro selection procedures were performed separately from the in vivo drug selection using pyrimethamine or WR99210. All in vitro culture and drug selection procedures were performed without shaking, and started at a time between 13:00 h and 15:00 h.
During the piggyBac experiment, 5 μg of target plasmid was co-transfected with 5 μg of transposase plasmid EGF-pgT. In the double crossover homologous recombination experiment, 10 μg of linearized plasmid was used in each transfection. After transfection, the parasites were immediately injected intravenously into a naive ICR mouse. The first drug selection was performed about 1-2 days post-injection when parasitemia reached 0.5-3.0%.
In piggyBac experiments, about 7 μl of infected tail blood were collected in 0.5 ml of culture medium (described above) with 2 μl of heparin solution (143 units/ml), and centrifuged at 500 × g for 5 min at room temperature (RT). The supernatant was discarded and the parasites were resuspended in 0.5 ml of culture medium. The suspension (450 µl) was put into a 24-well plate (Thermo Fisher Scientific), and either 50 µl of either puromycin solution (10 µg/ml, diluted by culture medium; Wako, Osaka, Japan; stock: 50 mg/ml in distilled water) to a final concentration of 1.0 µg/ml (1.8 µM), or WR99210 solution (62.5 ng/ml, diluted by culture medium; Jacobus Pharmaceuticals; stock: 5 mg/ml in DMSO) to a final concentration of 6.25 ng/ml (1.6 × 10 -2 µM) was added. The parasites were incubated for 20 h (36.5 °C, 5% CO 2 , 5% O 2 , 90% N 2 ). After incubation, the parasites were separated by centrifuge at 500 × g for 5 min at RT, and resuspended in 100 μl of PBS. They were then injected intravenously into a naive ICR mouse. When parasitemia reached 0.5-3.0% (about 5 days post-injection), the same selection procedure was repeated for the 2 nd and 3 rd selection using about 7 μl of tail blood. Cloning of parasites was done by in vivo limiting dilution method using male BALB/c mice.
In the first selection of double crossover homologous recombination experiments, 200 μl of infected blood was collected in 5 ml of culture medium by heart puncture under anesthesia, using a syringe containing heparin solution. After centrifugation at 500 × g for 8 min at RT, the supernatant was discarded. The blood was resuspended in 10.8 ml of culture medium. The suspension was put into a 25 cm 2 flask (Thermo Fisher Scientific) with 1.2 ml of puromycin (10 µg/ml, diluted by culture medium) solution to make a final concentration of 1.0 µg/ml and a total volume of 12 ml. This suspension was then incubated for 20 h. After drug selection, the sample was centrifuged at 500 × g for 8 min at RT and resuspended in 100 μl of PBS. The suspension was injected intravenously into a naive ICR mouse. When parasitemia reached 0.5-3.0% (about 5 days post-injection), the selection procedure was repeated. In the 2 nd and 3 rd selection, about 7 μl of tail blood in 0.5 ml culture medium was used as described above for the piggyBac experiment.
A schematic diagram of pac-puromycin in vitro selection procedures is shown in Supplementary Figure  Fluorescence analysis. Once parasite nuclei had been stained with Hoechst 33342, fluorescent images were obtained using a BZ-9000 fluorescence microscope (Keyence, Osaka, Japan) and were analyzed using a BZ-II Analyzer (Keyence).
Drug sensitivity test. Wild type P. berghei were transfected with pXL/hdhfr-pac-egfp and EGF-pgT, selected using pyrimethamine in vivo 23 , and named HSP70-PAC. Drug sensitivity tests were performed as previously described 6,26 . Briefly, infected blood (parasitemia 1.0-3.0%) was resuspended in a culture medium to a final cell concentration of 2%. A 500-µl cell suspension was cultured with various concentrations of puromycin (0-15 µM) for 20 h in a 24-well plate. The parasites were harvested by centrifugation and analyzed using Giemsa-stained thin blood smear films. The number of mature schizonts per 300 parasites was counted and normalized to the control. Three experiments were performed in triplicate. IC 50 values were calculated using program GraphPad Prism ver.5 (GraphPad, San Diego, CA, USA).

Southern blot analysis.
Two micrograms of extracted DNA were digested using Eco RI or Eco RV, separated on agarose gel, transferred onto a Hybond N + membrane (GE Healthcare, Chalfont St. Giles, UK), and then hybridized with probes labeled using an AlkPhos Direct Kit (GE Healthcare). The signal was detected using a CDP-star detection reagent (GE Healthcare).

Identification of piggyBac insertion site.
In summary, genomic DNA of mutant parasites was digested using Nde I and then ligated with T4 DNA ligase. Inverse PCRs were performed using the primer F: ATGTCCAGGAGGAGAAAGGC, R: GCCCCCAAATAAAAACTTCC. The sequences of these products were obtained using an ABI3100 Analyzer, and the insertion sites were identified using the PlasmoDB database.

Statistical analysis.
Statistical analyses comparing each pac-integrated parasite clone against the wild type parasite were performed using two-tailed unpaired t-tests. The mutant ratio of each selection was compared using paired t-tests. IC 50 values were calculated using program GraphPad Prism ver.5 (GraphPad).