Improvement of an in vitro drug selection method for generating transgenic Plasmodium berghei parasites

Background Reverse genetics approaches have become powerful tools to dissect the biology of malaria parasites. In a previous study, development of an in vitro drug selection method for generating transgenic parasite of Plasmodium berghei was reported. Using this method, two novel and independent selection markers using the P. berghei heat shock protein 70 promoter was previously established. While the approach permits the easy and flexible genetic manipulation of P. berghei, shortcomings include a low variety in promoter options to drive marker gene expression and increased complexity of the selection procedure. In this study, addressing these issues was attempted. Methods To secure a variety of promoters, the use of a P. berghei elongation factor-1α promoter for marker gene expression was attempted. To simplify the procedure of in vitro selection, the establishment of a two cell-cycle culture method and its application for drug selection were attempted. Results The P. berghei elongation factor-1α (pbef-1α) promoter, which is commonly used to drive marker gene expression, was successfully applied as an alternative promoter model for marker gene expression, using the parasite’s codon-optimized marker sequence. To simplify the in vitro selection method, a two cell-cycle culture method in which the merozoite was released by filtration of the culture containing matured schizont-infected erythrocytes was also developed and successfully applied for drug selection. Conclusion The pbef-1α promoter was successfully applied in an in vitro selection system. The in vitro selection procedure also could be simplified for practical use using a two cell-cycle culture method. These improvements provide a more versatile platform for the genetic manipulation of P. berghei.


Background
Plasmodium berghei is a useful in vivo rodent model of malaria. Its lifecycle can be reproduced in the laboratory and reverse genetics approaches have been established that have clarified much of its molecular biology [1][2][3]. The transfection efficacy of this parasite is low (10 −3 -10 −4 ), therefore transfected parasites are usually selected by applying drug pressure [4]. As the parasite can't be propagated in vitro [3,4], in vivo drug selection systems have been used for generating transgenic parasites of P. berghei. As this has prevented the use of toxic drug-based selection systems for rodent hosts, only antifolate selection systems have been reported [3]. The limited number of available selection systems has hampered progressive genetic experiments such as multiple gene knockouts (KO) and KO rescues in a parasite [5][6][7]. To solve this problem, an in vitro drug selection method for P. berghei based on repeated short-term in vitro culture with a drug followed by parasite recovery in vivo have been developed [8]. This method enabled us to use toxic drug-based selection systems in rodent hosts. Using this in vitro selection method, puromycin-N-acetyltransferase (pac)-puromycin and blasticidin S

Open Access
Malaria Journal *Correspondence: fukumoto@obihiro.ac.jp National Research Center for Protozoan Diseases, Obihiro University of Agriculture and Veterinary Medicine, Inada-cho, Obihiro, Hokkaido 080-8555, Japan deaminase (bsd)-blasticidin S (blasticidin) selection systems were successfully established for P. berghei [8,9]. The triple sequential gene manipulation was also successfully demonstrated using these systems with a traditional pyrimethamine resistant P. berghei dihydrofolate reductase-thymidylate synthase (pbdhfr-ts)-pyrimethamine selection system [9]. Furthermore, in these in vitro selection systems, the target mutant parasite can be enriched to over 90% within 2 weeks [8,9]. These results facilitated the isolation of the target gene-mutated parasite clones using fewer mice, shorter times, and lower costs in vivo [8,9]. While the in vitro selection system has several advantages, there is still room for improvements in that approach as well.
Flexibility in the selection of a promoter sequence is important in Plasmodium research. However, in the in vitro selection method, only a P. berghei heat shock protein 70 (pbhsp70) promoter was used to drive the expression of marker genes [8,9]. The pbhsp70 promoter is large (~ 1400 bp) and commonly used to drive strong expression of the target gene, such as a fluorescence marker, throughout the life cycle [10][11][12]. In some cases, over-expression of the marker gene causes unexpected effects for the parasite phenotype [13]. Therefore, it is desirable that an optimal choice of promoter should be possible for each assay. In some cases, the size of the targeting vector results in low recombination efficacy during gene manipulation. The choice of a most commonly used promoter P. berghei elongation factor-1α (pbef-1α, around 600 bp) for maker expression, can reduce the size of the targeting vector. Application of many kinds of available promoters for marker gene expression may facilitate useful future Plasmodium research.
The previous in vitro selection method for mutant parasite produced by homologous recombination method or piggyBac transposon-transposes recombination method had made it necessary to repeat the initial selection procedure, which was based on a single cell cycle of in vitro culture and in vivo parasite recovery, at least twice [8,9]. Followed by drug selection, desired mutant parasite clone could be obtained by one round single cell cloning procedure. This repetition is however not optimal for practical use. With the present approach, it is possible to enrich the mutant parasite population to > 90% without repeating the selection procedure if two cell cycles of in vitro drug selection can be applied. In general, continuous in vitro culture of P. berghei is often difficult, for a variety of reasons. The organism cannot propagate in vitro because the merozoites cannot egress from the schizont without additional mechanical shear stress [3]. The low stability of the mouse RBC in vitro and a restricted ability to invade mature RBC is also a bottleneck in the longterm cultivation of P. berghei [14].
The objective of this study is to solve the problems described above for practical use of in vitro selection method. Firstly, to secure a variety of promoters, the application of the pbef-1α promoter, which is most commonly used to drive marker gene expression, was attempted in P. berghei as a model. Secondly, to simplify the procedure of drug selection, the establishment of a two cell-cycle culture method and its application for drug selection were attempted.

Experimental animals and parasites
ICR and BALB/c (5-week old) mice were obtained from CLEA Japan (Tokyo, Japan). The BALB/c mice were used for parasite cloning and the ICR mice were used for other experiments. P. berghei ANKA strain (obtained from Dr. M. Torii, Ehime University, Japan) was used [8].

Codon optimization of marker genes
Codons of pac and bsd were optimized based on the P. berghei codon usage frequency from the Codon Usage Database (http://www.kazus a.or.jp/codon /), and the codons were changed to the most preferred codon without changing amino acid sequences. The codonoptimized markers (pacPB and bsdPB) were chemically synthesized by GenScript (Piscataway, NJ, USA).
The correct sequences of all plasmid inserts were confirmed by DNA sequencing using an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems, CA, USA).

Parasite transfection and drug selection
All mutant parasites were generated using piggyBac transposon-transposase recombination methods. Parasite transfection experiments followed standard protocols basically as previously described [4]. Briefly, the schizonts purified by Nycodenz density gradient centrifugation were co-transfected with piggyBac vectors and transposase expression vector EGF-pgT [8] using Nucleofector 2b (Lonza, Basel, Switzerland) under the U-33 program. In vitro selection utilized puromycin and blasticidin except for the two cell-cycle culture method, which was performed as previously described [8,9]. To generate mutant parasite clone containing pac driven by pbef-1α promoter (pbef-1α-pac) was performed using traditional in vivo pyrimethamine selection [4].

In vitro drug selection using two cell-cycle culture
On day 1, 300 µl of infected P. berghei blood (parasitaemia 0.5-3.0%) was collected and placed in 5 ml of culture medium, centrifuged at 500×g for 8 min at room temperature (RT), and the supernatant was discarded. The blood was resuspended in 14.94 ml of culture medium. The suspension was placed into a 25 cm 2 flask with 1.66 ml of puromycin stock solution (10 mg/ml in distilled water) to form a final concentration of 1.0 µg/ml and a total volume of 16.6 ml. This suspension was incubated for 20 h. On day 2, culture medium was removed after centrifugation of the culture at 500×g for 8 min. The pellet was resuspended in 14.94 ml of culture medium. The resuspended cultures were filtered using a filter unit with a 1.2 µm pore size and 32 mm diameter (Pall Corporation, Port Washington, NY, USA), attached to a 20 ml syringe. The filtrate containing free merozoite was put into a 25 cm 2 flask, and quickly mixed with 300 μl of phenylhydrazine-treated non-infected mouse blood prepared by two washes in culture medium and centrifugation at 500×g for 8 min. The flask was placed on a shaker (NA-M101; Nissin, Tokyo, Japan) and incubated at about 40 rpm for approximately 4 h under the culture conditions described above. After incubation, 1.66 ml of puromycin solution were added into the flask to form a final concentration of 1.0 µg/ml and a total volume of 16.6 ml, which was then incubated for around 20 h. On day 3, the parasites were collected by centrifuging at 500×g for 5 min at RT, resuspended in about 100 μl of PBS, and injected intravenously into a naïve mouse.

Mutant ratio calculation by microscopy and flow cytometry analysis
Parasites were identified with the Hoechst 33342 (Hoechst) stain and quantified using fluorescent microscopy as described previously [8]. The Hoechst-stained cells were analysed with a NovoCyte flow cytometer (ACEA Biosciences, San Diego, CA, USA) equipped with a 488 nm laser for eGFP and 405 nm laser for Hoechst. Parasites were gated using logarithmic forward/side scatter dot plots. The mutant ratio was analysed as previously described [8,9]. At least 20,000 Hoechst-positive parasites were analysed. Data analysis was performed using the Novo Express program ver. 1.0.3 (ACEA Biosciences).

Southern blot analysis
Southern blot analysis was performed as previously described [8]. The extracted genomic DNA (1.5 µg) were digested using NdeI. Digital chemiluminescence images were taken using an Ez-Capture MG (Atto, Tokyo, Japan).

Real-time quantitative RT-PCR assays
Total RNA was isolated using the TRIzol reagent (Thermo Fisher Scientific), treated with RQ1 RNase-Free DNase (Promega, Fitchburg, WI, USA). cDNA was synthesized from 1.0 µg RNA using ReverTra Ace qPCR RT Kit (Toyobo, Osaka, Japan). The pac, pacPB, and pbhsp70 standard curves were generated by tenfold serial dilution of pXL/hdhfr-pac-egfp [8], pXL/pbef-1α-pacPB-egfp, and pBS/pbhsp70 plasmid, respectively. Amplification and qPCR measurements were performed using the Ste-pOne Real-Time PCR System, v 2.1 software (Applied Biosystems). All reactions were carried out using Fast SYBR Green Master Mix (Thermo Fisher Scientific). The qPCR cycle protocol was as follows: 20 s at 95 °C followed by 40 cycles of 3 s at 95 °C and 30 s at 60 °C. All samples were run in triplicate. pbhsp70 was used as an endogenous control to normalize mRNA levels.

Statistical analysis
Statistical analyses comparing each pacPB-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. Statistical analyses comparing IC 50 values were performed using Dunnett's multiple comparisons tests. All statistical analyses were performed using the GraphPad Prism program ver.5 (GraphPad, San Diego, CA, USA).

pbef-1α promoter can be applied for in vitro puromycin selection using codon-optimized markers
To investigate whether pbef-1α promoter can be used for mutant selection, wild type parasite was transfected with pXL/hdhfr-pbef-1α-pac-egfp containing pac driven by pbef-1α promoter, and then applied for puromycin selection. However, the desired eGFP-expressing mutants could not be detected using fluorescence microscopic analysis after the second selection. This result indicated that promoter activity might not have been sufficient for selection. To improve expression levels, codon usage for the pacPB sequence was optimized for the P. berghei genome (Fig. 1a), and a pacPB expression vector (pXL/ pbef-1α-pacPB-egfp) was constructed (Fig. 1b). A wild type parasite was transfected with pXL/pbef-1α-pacPBegfp. The in vitro puromycin selection was repeated twice as previously described [8] and the egfp-expressing mutant ratio was monitored before and after each selection. The ratios were 0.01 ± 0.01, 15.13 ± 12.51, and 97.65 ± 1.70% (mean ± SD) for each respective selection (Fig. 1c, d). A typical transfection line was also analysed using flow cytometry. More than 95% of the parasites expressed eGFP after the second selection (Fig. 1e). Three clones were isolated from the mutant parasite population selected in Fig. 1c by an in vivo method of limiting dilutions to confirm integration of pacPB into the genome. Southern blot analysis confirmed that pacPB was integrated into the genome. Two copies of pacPB were integrated into clone 1, and one copy of pacPB was integrated into clones. 2 and 3 (Fig. 1f ). In clone 1 and 3, insertion position was on non-coding region. In clone 2, insertion position was on coding region of PBANKA_0305500. The growth of clones 1 and 2 was compared with that of wild type. The two clones grew equally as well as wild type parasites, which suggested that the integration of pacPB did not affect parasite growth (Fig. 1g). Then the effect of codon optimization was analysed. The IC 50 values of three mutant clones containing pac driven by the pbhsp70 promoter (pbhsp70-pac) [8] or ef-1α promoter (pbef-1α-pac), as well as pacPB driven by the pbef-1α promoter (pbef-1α-pacPB) were compared. pbef-1α-pac was generated by pyrimethamine selection. The IC 50 values were 2.43 ± 0.29 µg/ml for pbhsp70-pac, 1.21 ± 0.19 µg/ ml for pbef-1α-pac, and 2.12 ± 0.38 µg/ml for pbef-1α-pacPB (Fig. 1h). There was no significant difference between pbhsp70-pac and pbef-1α-pacPB, while the IC 50 value of pbef-1α-pacPB was significantly increased compared to pbef-1α-pac. Real-time quantitative RT-PCR analysis also confirmed a significant increase in pbef-1α-pacPB expression (Fig. 1i). These results suggest that codon-optimization results in the improvement of pac expression. Thus, pacPB under the control of the pbef-1α promoter can be used as a positive selection marker in P. berghei.

pbef-1α promoter can also be applied for in vitro blasticidin selection using codon-optimized marker
To examine whether the pbef-1α promoter can be applied as a blasticidin selection system, a bsdPB sequence optimized for the codon usage of P. berghei was designed (Fig. 2a) and constructed the bsdPB expression vector (pXL/pbef-1α-bsdPB-egfp) (Fig. 2b). The in vitro selection procedure for the pXL/pbef-1α-bsdPB-egfptransfected parasite was applied twice as previously described [9] and the egfp-expressing mutant ratio was monitored before and after each selection. The ratio was 0.01 ± 0.01, 2.81 ± 1.49, and 94.16 ± 2.15% (mean ± S.D.) after each selection (Fig. 2c, d). A typical transfection line was also analysed using flow cytometry. More than 94% of the parasites expressed eGFP after the second selection (Fig. 2e). Thus, bsdPB driven by the pbef-1α promoter can also be used as a positive selection marker in P. berghei.

Development of improved in vitro selection method using two cell-cycle culture method
To simplify the in vitro selection procedure, a method using two cell cycles in culture was examined whether the method could be applicable. A two cell-cycle in vitro culture procedure utilizing a filtration method was (See figure on next page.) Fig. 1 Generation of mutants using pacPB driven by pbef-1α promoter. a Nucleotide sequences of original (pac) and codon-optimized pac (pacPB). In the pacPB sequence, modified nucleotides are highlighted. b Schematic diagram of the piggyBac transposon vector containing pacPB and egfp expression cassettes (pXL/pbef-1α-pacPB-egfp). ITR: inverted terminal repeat. c The eGFP-positive parasite ratio after each puromycin selection. The ratio was analysed using fluorescence microscopy. Each bar represents the mean ± SD of four independent experiments. **p < 0.01, n.s.: not significant (paired t-tests). d Fluorescence images of parasites after each selection. Parasites were stained with Hoechst 33342. The scale bar represents 10 µm. e Flow cytometry analysis of the typical transfection line of eGFP positive mutant ratio after each puromycin selection. Numbers above the bracketed lines indicate the percentage of parasites with eGFP expression. f Southern blot analysis of three pacPB integrated mutant clones. Genomic DNA was digested using NdeI and hybridized with a pacPB probe. Wild wild type, cl. clone. Insertion sites of piggyBac transposon were shown blow. U upper band of cl.1, L lower band of cl.1, Chr. chromosome. g Growth assay of two pacPB integrated clones. Female ICR mice were infected intravenously with 1000 infected P. berghei erythrocytes. Parasitaemia of infected mice (Wild: n = 5, cl.1: n = 4, cl.2: n = 3) was monitored daily by examination of blood smears. cl. clone, Wild wild type. h Growth inhibition of pXL/hdhfr-pac-egfp [8]-transfected parasites (pbhsp70-pac), pXL/hdhfr-pbef-1α-pac-egfp-transfected parasites (pbef-1α-pac) obtained by pyrimethamine selection using hdhfr marker and pXL/ pbef-1α-pacPB-egfp-transfected parasites (pbef-1α-pacPB) was evaluated by determining schizont development within a range of puromycin concentrations. i Real-time quantitative RT-PCR analysis of marker gene expression in pbhsp70-pac, pbef-1α-pac, and pbef-1α-pacPB. The value was normalized to the level of pbhsp70 transcripts in each sample. **p < 0.01 (Dunnett's multiple comparisons tests) established (Fig. 3a) [15]. To determine its suitability as an in vitro selection method, it was used to select pXL/pbef-1α-pacPB-egfp-transfected parasites. After transfection, puromycin selection was performed by both the original method used in the previous study [8] and the two cellcycle culture method. The egfp-expressing mutant ratio was examined before and after each selection for both methods. The ratio was 0.02 ± 0.03, 17.96 ± 13.51, 94.60 ± 8.66 and 94.59 ± 8.20% (mean ± SD) (Fig. 3b, c). The target mutant ratio after one selection using the two cell-cycle culture method was equal to that of the original method after a second selection. Flow cytometry analysis also confirmed  that more than 94% of the parasites expressed eGFP after the two cell-cycle selection procedure (Fig. 3d). Thus, an in vitro puromycin selection method using a two cell-cycle culture method was successfully developed in P. berghei.

Discussion
In this study, an improved in vitro drug selection method for generating transgenic parasites in P. berghei was described. Securing a variety of promoters and simplification of the selection procedure were attempted. In previous studies, the pbhsp70 promoter was solely used to drive marker gene expression. In the present study, the pbef-1α promoter, most commonly used for marker gene expression, was successfully applied for in vitro selection by using a codon-optimized marker sequence. The pbef-1α promoter is bidirectional and its size (~ 600 bp) is smaller than pbhsp70 (~ 1400 bp) [12]. This reduced size is an advantage, because large constructs often result in low recombination efficacy at the gene manipulation level as well as instability in Escherichia coli caused by the ATrich Plasmodium genome [16]. The pbhsp70 promoter is known to promote constitutive and strong expression of the gene throughout the life cycle [10]. Flexibility in the selection of a promoter is desirable to minimize any

A T GGC CA A GC C T T T GT C T CA AGA AGA A T CC AC C CT CA T T G A A AGA GCA A C A T GGC A A A A C C A T T A T C A CA AGA AGA A T CA AC A T T A A T AG A A AGA GCA A C GGC T A C A A T C AA C A GCA T CC C CA T CT C T GA AGA CT A C A GC GT CGC C A GCG AGC A A CA A T A AA T T C A A T A C C A A T A T C AGA AGA T T A T T C A GT T GC A T C AG
CAGCT  In the bsdPB sequence, modified nucleotides are highlighted. b Schematic diagram of the piggyBac transposon vector containing bsdPB and egfp expression cassettes (pXL/pbef-1α-bsdPB-egfp). ITR inverted terminal repeat. c The eGFP-positive parasite ratio after each blasticidin selection. The ratio was analysed using fluorescence microscopy. Each bar represents the mean ± SD of three independent experiments. **p < 0.01, n.s.: not significant (paired t-tests). d Fluorescence images of parasites after each selection. Parasites were stained with Hoechst 33342. The scale bar represents 10 µm. e Flow cytometry analysis of the typical transfection line of eGFP positive mutant ratio after each blasticidin selection. Numbers above the bracketed lines indicate the percentage of parasites with eGFP expression unexpected effects on the parasite phenotype after transfection. The use of the pbef-1α promoter as an in vitro selection method was, therefore, examined. In a pacpuromycin system, the mutant parasite with pac driven by the pbef-1α promoter could not be obtained. There were indications that the expression level of Pac under the pbef-1α promoter was insufficient for puromycin selection. The GC-richness (around 70%) of pac might affect the expression level of this marker gene in the ATrich parasite genome [17]. To overcome this problem, codon usage within the pac sequence was optimized to that employed in the parasite genome. It has been reported that codon optimization of antibiotic genes results in the improvement of the drug selection efficacy in other organisms [18] as well as increased translation efficacy [19]. The pbef-1α promoter was successfully used with the codon-optimized pacPB for puromycin selection, suggesting that codon optimization increased the expression level of the marker gene. In fact, gene expression analysis by real-time RT PCR and IC 50 value analysis showed a significantly increased expression level of the marker gene. As this codon optimization strategy could also be used in a blasticidin selection system, it has been suggested that a selection marker with GC-richness should be applied after sequence optimization for codon usage in malaria parasites [20][21][22]. The eGFP-positive parasite ratio after each puromycin selection performed by original method and two cell-cycle culture method. The ratio was analysed using fluorescence microscopy. Each bar represents the mean ± SD of four independent experiments. **p < 0.01, n.s.: not significant (paired t-tests). c Fluorescence images of parasites after each selection and two cell-cycle culture selection. Parasites were stained with Hoechst 33342. The scale bar represents 10 µm. d Flow cytometry analysis of the typical transfection line of eGFP positive mutant ratio after each puromycin selection and two cell-cycle culture method. Numbers above the bracketed lines indicate the percentage of parasites with eGFP expression In this study, an in vitro selection method using a two cell-cycle culture method was also successfully developed. In previous studies, established selection procedures based on in vitro culture and in vivo parasite recovery had to be applied twice [8,9] as P. berghei merozoites cannot egress from the schizont in vitro without mechanical shear stress [3]. To simplify this procedure, a two cell-cycle culture method was applied as an in vitro selection system. The problem was addressed by applying a filtration protocol previously used for the purification of Plasmodium yoelii merozoites [15]. An in vitro selection procedure based on two cell-cycle culture with drug was successfully established. The desired mutant parasites were enriched to > 90% by a single in vitro selection. Overall selection efficacy was equivalent to that of the method used in the original in vitro selection [8]; however, the two cell-cycle culture approach shortened the total period required for generating mutant parasites. It took ~ 11 days to enrich the desired mutant parasite to > 90% using the two cell-cycle method, compared to ~ 14 days in the original method [8]. Therefore, the two cell-cycle method also permits rapid isolation of many mutant parasites. In addition, the two cell-cycle culture method requires only a single infected mouse for the procedure (Fig. 3a). The method's high selection efficacy should allow us to isolate a clone using fewer than five mice, the number previously used for in vitro selection [8]. This method thus also has benefits for animal welfare.

Conclusion
The pbef-1α promoter was successfully applied in an in vitro selection system. This resulted in an improved potential promoter variety for the in vitro selection system. Furthermore, the in vitro selection procedure could be simplified using the two cell-cycle culture method. This approach might minimize the time and animal usage often required in other in vitro selection systems. These improvements provide us a more flexible and userfriendly platform to perform genetic manipulation of P. berghei. Furthermore, the selection systems developed in this study could enable us to develop a genetically modified parasite-library, although further experiments are needed.