Validation of Plasmodium vivax centromere and promoter activities using Plasmodium yoelii

Plasmodium vivax is the leading cause of malaria outside Africa and represents a significant health and economic burden on affected countries. A major obstacle for P. vivax eradication is the dormant hypnozoite liver stage that causes relapse infections and the limited antimalarial drugs that clear this stage. Advances in studying the hypnozoite and other unique biological aspects of this parasite are hampered by the lack of a continuous in vitro laboratory culture system and poor availability of molecular tools for genetic manipulation. In this study, we aim to develop molecular tools that can be used for genetic manipulation of P. vivax. A putative P. vivax centromere sequence (PvCEN) was cloned and episomal centromere based plasmids expressing a GFP marker were constructed. Centromere activity was evaluated using a rodent malaria parasite Plasmodium yoelii. A plasmid carrying PvCEN was stably maintained in asexual-stage parasites in the absence of drug pressure, and approximately 45% of the parasites retained the plasmid four weeks later. The same retention rate was observed in parasites possessing a native P. yoelii centromere (PyCEN)-based control plasmid. The segregation efficiency of the plasmid per nuclear division was > 99% in PvCEN parasites, compared to ~90% in a control parasite harboring a plasmid without a centromere. In addition, we observed a clear GFP signal in both oocysts and salivary gland sporozoites isolated from mosquitoes. In blood-stage parasites after liver stage development, GFP positivity in PvCEN parasites was comparable to control PyCEN parasites. Thus, PvCEN plasmids were maintained throughout the parasite life cycle. We also validated several P. vivax promoter activities and showed that hsp70 promoter (~1 kb) was active throughout the parasite life cycle. This is the first data for the functional characterization of a P. vivax centromere that can be used in future P. vivax biological research.

active throughout the parasite life cycle. This is the first data for the functional characterization of a P. vivax centromere that can be used in future P. vivax biological research.
falciparum [20]. Therefore, we decided to use a rodent malaria parasite Plasmodium yoelii as a platform, since many regulatory elements have been shown to be replaceable between rodent malaria parasites and old world monkey malaria parasites. In this study we evaluated P. vivax gene regulatory elements and centromere regions for activity in P. yoelii for potential suitability for genetic manipulation of P. vivax.

Generation of plasmid constructs containing a P. vivax centromere region
Putative P. vivax centromeres were predicted by gene synteny and whole genome sequence analysis for four putative centromeres identified in chromosomes 5,9,11, and 13 (Genbank numbers CM000446.1, CM000450.1, CM000452.1, and CM000454.1, respectively) [21][22][23]. We PCR-amplified and cloned these centromere regions (termed PvCEN5, PvCEN9, PvCEN 11, and PvCEN13, respectively) in E. coli. Plasmids containing the PvCEN5 and PvCEN9 fragments were unstable and were spontaneously deleted during the cloning procedure. Plasmids containing the PvCEN11 and PvCEN13 fragments were also unstable, but 10% of the transformed E. coli contained full-length fragments. After attempts to transfer these PvCEN fragments to a Gateway destination plasmid, we successfully generated a destination plasmid containing the PvCEN11 fragment (2211 bp). This fragment was then truncated in an attempt to make the PvCEN11-based plasmid more stable in E. coli; specifically, sequences 255-2142 (1888 bp) and 700-2234 (1535 bp) were selected to generate destination plasmids containing the fragments PvCEN11S2 or PvCEN11S3, respectively ( Fig 1A). These sequences contained a highly AT-rich fragment and a repetitive region, which are likely required for centromere function [23].
Since there is no available P. vivax long-term in vitro culture, we used P. yoelii to evaluate if the putative centromeres were functional in retaining episomal plasmids during blood-stage parasite propagation. As a positive control, the predicted P. yoelii centromere region of 1787 bp in chromosome 5 (DQ054838.1, PyCEN5) was used to make a destination plasmid ( Fig  1A). We then constructed three episomal centromere-based plasmids (PvCEN11S2-H86HD-TG, PvCEN11S3-H86HD-TG, and PyCEN5-H86HD-TG) expressing a drug selectable marker human dihydrofolate reductase (hDHFR) fused with a triple Ty1 tag sequence and a green fluorescence protein (GFP) marker driven by the P. falciparum hsp86 (pfhsp86) promoter ( Fig  1B). A negative control plasmid containing the same hDHFR-3Ty-GFP expression cassette without a centromere region was also prepared.
Transgenic lines were successfully generated with the 3 centromere-based plasmids using P. yoelii 17XL. However, transgenic parasites were not obtained with the negative control plasmid following multiple trials (two independent transfections and four times total). Thus, we decided to use the promoter region of P. vivax hsp86 (pvhsp86), an ortholog of Pfhsp86, in the syntenic region for the negative control plasmid without any centromere region (pNoCEN), and a transgenic line was successfully obtained (Fig 1B). The reason for the failure to obtain a transfectant with the Pfhsp86 promoter-based control plasmid is not clear, because the promoter activities in P. yoelii appear to be equivalent between the Pfhsp86 and Pvhsp86 promoters (described later). Of note is that the Pfhsp86 promoter region does not appear to play a role in plasmid segregation to daughter cells [24], thus replacing this promoter is considered not to affect the plasmid segregation efficiency. Under pyrimethamine (PYR) drug pressure, 3 transgenic parasites containing centromere-based plasmids grew similarly to P. yoelii wild type 17XL line parasites maintained without PYR (Fig 2A). However, pNoCEN parasites grew slower with PYR than the 17XL line without PYR (Fig 2A). Although it was not significant, pNoCEN parasites showed a lower maximum growth rate per day compared to the other lines (S1A Fig). The number of merozoites per schizont were equivalent among all parasite lines, indicating that this is not the determinant of the observed slow growth phenotype of pNoCEN parasites (S1B Fig). We interpret that the slower growth rate of pNoCEN parasites is likely due to a low segregation efficiency of the pNoCEN plasmid to daughter merozoites. Plasmids containing P. vivax or P. yoelii centromere regions were stably and similarly maintained in the rodent malaria parasite P. yoelii during asexual growth in vivo To evaluate the stability of the centromere-based plasmids during asexual growth, transgenic parasites were maintained without PYR pressure and the ratios of GFP-positive parasites were examined. Parasites were maintained in mice with PYR for one week to ensure that all parasites retained centromere plasmids, then infected red blood cells (iRBCs) were passaged to new mice and maintained without PYR for 28 days in naïve mice by syringe passage. When parasitemias reached 5-10%, blood was collected and 10 6 iRBCs were inoculated intraperitoneally into a new mouse. We found that parasites with control pNoCEN rapidly lost a GFP signal and only 2.5 ± 0.8% and 0% GFP-positive parasites were observed on day 10 and 14, respectively. In contrast, the ratio of GFP-positive parasites gradually declined in transgenic parasites with PvCEN11S2, PvCEN11S3, or PyCEN5 with the percentage of GFP-positive parasites averaging 44 ± 2, 50 ± 1, and 44 ± 3% on day 28, respectively ( Fig 2B). There was no significant difference in the percent infections among the three centromere-based plasmids ( Fig  2B). Because the average number of merozoites per schizont was about 11 (= 2 3.46 ), roughly 3.5 rounds of nuclear division was assumed to occur during one blood-stage schizogony of 18 hours [25]. Based on these assumptions, the segregation efficiency of the plasmid per nuclear division was calculated at each observation point. The segregation efficiency of parasites transfected with the PvCEN11S2 or PvCEN11S3 plasmids was always more than 99%, comparable with that of PyCEN5 parasites (> 99%), while the segregation efficiency of parasites transfected with a plasmid without a centromere, pNoCEN, was 88.3 ± 0.4% on day 5 and 92.2 ± 0.7% on day 10 (Table 1).
Next, we evaluated the plasmid copy number by qPCR before and 28 days after PYR removal. In the pNoCEN parasites, plasmid copy number was 14 ± 2 per parasite before PYR removal and 0.2 ± 0.1 per parasite on day 28 after PYR removal (~70-fold reduction, p < 0.01). The observation of a high copy number of pNoCEN is consistent with the concept that parasites with plasmid high copy numbers would be selected under drug pressure when the plasmid is not segregating efficiently. In the PvCEN11S2, PvCEN11S3, and PyCEN5 parasites, plasmid copy numbers were 4.7 ± 1.2, 6.5 ± 1.9, and 4.0 ± 1.8 per parasite before PYR removal and 4.4 ± 0.8, 2.1 ± 0.7, and 3.6 ± 1.9 per parasite on day 28 after PYR removal, respectively (Fig 3). These data indicate that the copy number of centromere-based plasmids is much lower than the copy number of pNoCEN with PYR (p < 0.01), but higher than that of pNoCEN The segregation efficiency of the plasmid per nuclear division was calculated (see Materials and methods) at each point of passage. Roughly 3.5 rounds of nuclear division occurring during one blood-stage schizogony of 18 hours was assumed [25], which resulted in the production of 11 daughter nuclei. Values are based on the parasitemia in 3 to 5 independent mice. Standard deviation is shown in parenthesis. Values were not calculated for pNoCEN after day 14, because no GFP-positive parasites were detected.
https://doi.org/10.1371/journal.pone.0226884.t001 without PYR (p < 0.01 for PvCEN11S2 and PyCEN5, but not significant for PvCEN11S3). Although the reduction in GFP positivity was similar among PvCEN11S2, PvCEN11S3, and PyCEN5 parasites, the plasmid copy numbers of PvCEN11S3 parasites were significantly reduced after 28 days cultivation without PYR (p < 0.01), albeit PvCEN11S2 and PyCEN5 parasites did not show significant reduction. The PvCEN11S2 445 bp sequence that is missing from PvCEN11S3 may have an important role in this observation. Iwanaga et al (2010) showed that approximately 90% of P. berghei parasites with a plasmid containing the complete P. berghei centromere region (pbCEN5) expressed GFP on day 21 after PYR removal [26], significantly higher than the~60% GFP positivity seen in this study in P. yoelii with PyCEN5 on day 18. This may be partly due to the difference in growth rates between the two parasite lines; P. yoelii 17XL line parasite invades a variety of RBCs and grows more rapidly than P. berghei which preferentially invades less abundant younger RBCs. The difference in the duration of one blood-stage schizogony may also contribute, which is 18 hours for P. yoelii [25] and 22-23 hours for P. berghei [27]. These are consistent with the observation that the segregation efficiency of the control pbGFPcon plasmid without a centromere region was 94% per nuclear division in their study, which was higher than the values for our control pNoCEN plasmid (88.3 or 92.2). Another factor may be that GFP was fused to hDHFR and expressed under one promoter in this study, whereas GFP and the drug selectable marker protein were expressed under different promoters for the P. berghei study. PvCEN11S2 and PyCEN5 parasites retained similar plasmid copy numbers per parasite before and on day 28 after PYR removal, albeit with a clear reduction in GFP signals, suggesting that the GFP signal decrease may be the result of transcriptional silencing in our plasmid expression cassettes. Nonetheless, our results indicate that during asexual blood growth the plasmid containing the PvCEN11S2 region was stably maintained equally well in parasites with low copy number and the P. yoelii centromere-based plasmid.

Transgenic parasites were able to retain plasmids with P. vivax centromere regions through the mosquito stages of the parasite life cycle
We examined if P. vivax centromere-based plasmids were stably maintained in parasites through the mosquito stages. On day 11 after mosquito feeding (day 14 after PYR removal), the median oocyst numbers in mosquito midguts were 39 and 52 (1st experiment) or 39 and 42 (2nd experiment) for PvCEN11S2 and PvCEN11S3, respectively. In contrast, those for PyCEN5 and pNoCEN transgenic parasites and P. yoelii wild type parasite were 52, 54, and 50, respectively (Figs 4 and 5A). These results indicate that all parasite lines yielded sufficient numbers of oocysts. Greater than 80% of oocysts showed GFP signals in all transgenic parasite lines with centromere-based plasmids, whereas pNoCEN parasites showed only 4% (Figs 4  and 5B). Salivary gland sporozoites collected on day 18 after mosquito feeding (day 21 after PYR removal) showed GFP signals only in PvCEN11S2, PvCEN11S3, and PyCEN5 transgenic parasite lines. No GFP signal was seen in pNoCEN parasites and P. yoelii wild type parasites (Fig 4). Due to a limitation of mosquito numbers, we did not evaluate the percentage of GFPpositive sporozoites. On day 3 or 4 after intravenous inoculation of pooled sporozoites (6,000~10,500) into naïve mice, blood-stage parasites emerged in all lines. On day 6 after mouse inoculation (day 30 after PYR removal), at least 40% of the blood-stage parasites showed GFP signals in PvCEN11S2, PvCEN11S3, and PyCEN5 transgenic parasite lines, whereas pNoCEN parasites and P. yoelii wild type parasite did not show a GFP signal (Figs 4 and Fig 5C). We found that PvCEN11S3 parasites retained GFP positivity as efficiently as PvCEN11S2 and PyCEN5 after mosquito and liver stage development, and the pattern of plasmid copy numbers with or without PYR pressure during the blood-stages were clearly different (Fig 3). It is formally possible that the lesser or absent GFP signal in pNoCEN parasites in the mosquito stages is not due to a loss of the plasmid, but a lack of transcription under the P. vivax HSP86 promoter in the pNoCEN plasmid; however, because HSP86 is shown to be highly transcribed in the P. vivax sporozoite stage [28] and pNoCEN blood-stage parasites did not show a GFP signal after passing through mosquitoes, we considered that the lack of GFP signal was due to a loss of the plasmid. Nonetheless, these results indicated that plasmids with PvCEN11S2 and PvCEN11S3 were maintained throughout the mosquito and liver stages of the parasite life cycle as efficiently as the PyCEN5 plasmid.

Evaluation of P. vivax promoters in blood and mosquito stages
To expand the repertoire of validated P. vivax promoters, we cloned from P. vivax gDNA the 5' untranslated regions (UTR) of hsp70, hsp86, and crt. For pvhsp70 5' UTR, 3 fragments with GFP signals from mosquito midgut oocysts, salivary gland sporozoites, and blood stages after liver stage development of transfected parasites. Forty mosquitoes were fed on day 3 after intravenous injection of 10 6 infected red blood cells into mice for each group (transgenic parasites with PvCEN11S2-H86HD-TG (PvCEN11S2), PvCEN11S3-H86HD-TG (PvCEN11S3), PyCEN5-H86HD-TG (PyCEN5), or pNoCEN and the parental parasite 17XL) and were dissected to count the numbers of midgut oocysts and to observe GFP signals on day 11 after feeding (day 14 after pyrimethamine removal). Whole salivary glands were examined to detect GFP signals on day 18 (day 21 after pyrimethamine removal). Ten to fifteen remaining mosquitoes were dissected to collect salivary gland sporozoites on day 21 after feeding (day 24 after pyrimethamine removal). Sporozoites were counted and observed for GFP signals. All possible collected sporozoites were injected intravenously into mice and parasites that appeared in the blood on day 6 after inoculation (day 30 after pyrimethamine removal) were observed for a GFP signal. Parasite nuclei were stained with Hoechst (Hx). DIC, differential interference contrast image. Bar indicates 10 μm.
https://doi.org/10.1371/journal.pone.0226884.g004 different lengths (approximately 1, 1.5, and 2 kb) were evaluated. The pfhsp86 promoter was used as a positive control. Two independent transgenic P. yoelii lines were generated for each plasmid and firefly luciferase activities were compared from purified schizont-stage parasites. Significantly higher luciferase activities were detected from all evaluated 5' UTR, with hsp70 promoter as the strongest, followed by the pvhsp86 and pvcrt promoters (Fig 6) by comparison with a negative control plasmid without a promoter.
An~1 kb fragment of the pvhsp70 5' UTR was selected for further evaluation because in other Plasmodium species this promoter has been frequently used to express proteins throughout the parasite life cycle including mosquito and liver stages. A piggyBac-based plasmid expressing NanoLuc 1 Luciferase-mCherry fusion protein under the pvhsp70 promoter (~1 kb) was generated and transfected to P. yoelii with a helper pHTH plasmid [29]. From two transfectant clones obtained (clones 5 and 15), bright mCherry-positive parasites were observed at blood, oocyst, and salivary gland sporozoite stages (Fig 7). Luciferase activities were detected before and after mosquito and liver stage development, indicating that the expression from the plasmid-inserted region was stable (Fig 8A). Luciferase activity was also detected in oocyst ( Fig 8B) and salivary gland sporozoite (Fig 8C) stages. These results indicate that the pvhsp70 promoter was active throughout the mosquito stages.

Conclusions
In this study, we cloned P. vivax centromere and promoter regions and confirmed their activities using P. yoelii. These experimentally validated components would serve as valuable tools for genetic manipulations to study P. vivax.
The piggyBac-based Multisite Gateway destination plasmid (pigHD_R43-3U) was generated from pXL-BacII-DHFR, by inserting a linker (StSmNhE5.ApaF and StSmNhE5.ApaR) into an ApaI site, a DNA fragment containing PbDT3U amplified from pHH1 [34] with primers PbDT3U.F3 and PbDT3U.R3 to a SmaI site, then a DNA fragment containing ccdB R43 amplified from pCHD43(II) with primers M13F and M13R into a StuI site. pigHD_vHSP70-1k-NLuc-mCherry was generated from pENT41-PvHSP70-5U-1k, pENT12-NLuc, pENT23-mCherry, and pigHD_R43-3U plasmids. Mice were infected with transgenic parasites with pigHD_vHSP70-1k-NLuc-mCherry or the parental parasite 17XL by intravenous injection of 10 6 infected red blood cells and maintained without pyrimethamine. Three days later thin blood smears were made (day 3) and mosquitoes were fed. Mosquitoes were dissected on day 11 after feeding (day 14 after pyrimethamine removal) to count the number of the midgut oocysts and to observe mCherry signals. Whole salivary glands were examined to detect mCherry signals on day 18 (day 21 after pyrimethamine removal). The remaining mosquitoes were dissected to collect salivary gland sporozoites on day 24, which were then injected intravenously into mice. Parasites appeared in the blood on day 6 after inoculation (day 30 after pyrimethamine removal) and were observed for mCherry signals. Parasite nuclei were stained with Hoechst (Hx). DIC, differential interference contrast image. Bar indicates 10 μm. https://doi.org/10.1371/journal.pone.0226884.g007

Malaria parasites, mice, and mosquitoes
The P. yoelii 17XL line was maintained using 6 to 8 week old female ICR mice (SLC Inc., Shizuoka, Japan) in Nagasaki University. Animal experiments conducted in this study were approved by the Animal Care and Use Committee of Nagasaki University (Permit number 1403031120-9) and mice were treated in accordance with the Act on Welfare and Management of Animals, Japan. Mice were kept in a temperature-controlled (20-26˚C) room under a 12-hour light/dark cycle and monitored at least twice a week. Availability of food and water was also monitored at least twice a week. Malaria parasites were intravenously inoculated to mice anesthetized with isoflurane. Euthanization of mice was done by deep anesthesia with 2-5% isoflurane followed by cervical dislocation.
Forty Anopheles stephensi mosquitoes for each group were fed on day 3 after intravenous injection of 10 6 parasite-infected RBCs into mice. Mosquitoes were dissected to observe midgut oocysts on day 11 after feeding (day 14 after PYR removal) and salivary gland sporozoites on day 18 (day 21 after PYR removal). To collect sporozoites for injection into mice, salivary glands were gently crushed in sterile phosphate buffered saline (PBS) using a plastic homogenizer and sporozoites were harvested from the supernatant after a brief low-speed centrifugation. Collected sporozoites were injected into mice intravenously.

Generation of transgenic parasites
P. yoelii schizont-enriched fractions were collected from mouse blood by differential centrifugation on a 50% Histodenz TM (Sigma-Aldrich, St. Louis, MO) solution and 20 μg of each plasmid was electroporated to 5 × 10 7 enriched schizonts using the Nucleofector™ 2b device (Lonza Japan) with Human T cell solution under program U-33 [35]. Transfected parasites were intravenously injected into 8-week old BALB/c female mice and transfectants were selected by oral administration of PYR and cloned by limiting dilution using mice. For pigHD_vHSP70-1k-NLuc-mCherry plasmid, 20 μg of pHTH plasmid was also transfected [29].

Live imaging of the parasites
Thin smears of parasite-infected blood, spots of purified sporozoites, and oocyst-attached mosquito midguts were prepared on glass slides, stained with Hoechst 33342, and fluorescent and differential interference contrast (DIC) images were captured using an AxioCam MRm CCD camera (Carl Zeiss, Germany) fixed to an Axio imager Z2 fluorescent microscope with a Plan-Apochromat 100 ×/1.4 oil immersion lens (Carl Zeiss) and Axiovision software (Carl Zeiss). Obtained images were processed using Zen lite 2011 software (Carl Zeiss) and Adobe Photoshop CS (Adobe Systems Inc., San José, CA).
The segregation efficiency of the plasmid per nuclear division was calculated using the following equation: segregation efficiency ¼ 100 � ðPgfp=100Þ 1=n , where P gfp is the percentage of GFP-positive parasites after PYR removal and n is the number of nuclear divisions. In this study, we assumed that 3.5 rounds of nuclear division (equivalent to the generation of 11 daughter parasites) occurred during one blood-stage schizogony of 18 hours.

Quantitative polymerase chain reaction (qPCR)
Genomic DNA was purified from the mixed stage parasites collected from mice using the QiaAmp DNA Mini Kit (QIAGEN, Germany). qPCR was conducted to quantitate the plasmids (targeting GFP gene with primers GFP.rtF1 and GFP.rtR1) and parasite genome (targeting met-tRNA ligase gene (PY17X_0519800) with primers PyMet-tRNA.rtF1 and PyMet-tRNA.rtR1) with Power SYBR™ Green PCR Master Mix (Thermo Fisher Scientific) using the 7500 Real-Time PCR system (Applied Biosystems, Foster City, CA). Copy number was estimated by using standard curves generated with serially diluted plasmids containing target PCR product sequences. Values for the GFP gene were normalized to those for the met-tRNA ligase gene and values from three independent experiments were assessed. A one-way analysis of variance (ANOVA) test and Tukey's post-hoc multiple comparisons tests were performed using PRISM (GraphPad Software, San Diego, CA) to analyze the difference between the plasmid copy number per parasite genome.

Luciferase assay
Purified schizont stage parasites were adjusted to 10 8 parasites/assay and firefly luciferase assays were performed using the ONE-Glo TM Luciferase system (Promega) according to the manufacturer's instructions. NanoLuc luciferase activity was measured by the Nano-Glo1 Luciferase Assay system (Promega) according to the manufacturer's instructions. Luciferase signals were immediately measured with a plate reader (ARVO MX 1420; PerkinElmer) for 15 sec. Plasmid copy numbers for each sample were estimated by targeting hDHFR with primers hDHFR.rtF and hDHFR.rtR as described above, and used to normalize the luciferase activities.
Supporting information S1 Table. Oligonucleotide primers used in this study.