Microarray analyses of stage-specific gene expression of Plasmodium falciparum require purification of RNAs from highly synchronized cultures. To date, no reliable method to control the quality of synchronization of P. falciparum cultures is available.
A double-staining method using hydroethidine and thiazole orange for nucleic acid staining was carried out to compare by flow cytometric analysis the nucleic acid labeling of synchronized P. falciparum in cultures at different time points of the 48-h intraerythrocytic cycle.
With this method, we determined the quality of culture synchronization in schizont and ring stages. Nucleic acid analysis, based on thiazole orange fluorescence, clearly showed that low levels of schizonts in ring cultures results in a high contamination of ring nucleic acids by schizonts. Conversely, nucleic acids from trophozoite or schizont cultures containing ring stages did not present a significant contamination by ring nucleic acids.
Complete Plasmodium falciparum genome sequencing was recently achieved (1–4). This offers the possibility of genome-wide analysis of stage-specific gene expression in different situations, including drug pressure, immune pressure, and oxidative stress. This requires the purification of RNAs from highly synchronized cultures to compare the expression profiles of well-defined developmental stages under standardized conditions. Consequently, the control of synchronization quality is a critical step of the process.
For several years, flow cytometry offered the possibility to study the cell cycle of P. falciparum by using nucleic acid staining. Dertinger et al. (5) and Torous et al. (6) showed that Plasmodium nucleic acid staining is reproducible. They used asynchronous rodent malaria parasites as a positive control in flow cytometric analysis of micronuclear formation induced by clastogenic agents. Propidium iodide (7, 8), acridine orange (9–11), Hoechst 33258 (12, 13), and Hoechst 33342 (14, 15) were commonly used as DNA dyes. Propidium iodide and acridine orange required fixation and permeabilization of parasites before use. Hoechst is a vital dye and was used by Janse et al. (16) to demonstrate that the stage of the parasite could be determined from the intensity of fluorescence. However, it requires an ultraviolet laser for cell analysis. Recently, hydroethidine (HE), a new vital nucleic acid dye, was introduced to study the malaria parasite by flow cytometry with analyzers equipped with a 488-nm laser (17). From these studies, it was clear that the fluorescence intensities correlated with the ring and schizont stages of the parasite. With regard to the nucleic acid staining by thiazole orange (TO) (18, 19), we adapted it for monitoring parasite growth in vitro (20). We found that it is possible to perform double staining of nucleic acids of P. falciparum parasites with Hoechst and TO without pretreament of infected red blood cells, thus respecting their integrity and allowing immunologic detection of parasite surface antigens (20). However, this method required two lasers for analysis.
In the present study, we analyzed parasite populations stained with HE instead of Hoechst by using a flow cytometer equipped with a single 488-nm laser. We combined HE staining, which allows the visualization of all parasitized erythrocytes, regardless of the parasite stage, with TO staining, which provides a wide scale of fluorescence labeling. Accordingly, we investigated the relative distribution of parasite subpopulations at different parasite intraerythrocytic stages by using synchronized P. falciparum cultures, fluorescence double staining, and flow cytometric analysis to control the parasite synchronicity, a critical step before transcriptome microarray analysis. This method permits not only visualization of a parasite population with high accuracy but also following the parasite's development during its life cycle.
MATERIAL AND METHODS
Parasite and Culture Conditions
We used the chloroquine-sensitive strain 3D7. Parasites were grown according to the method of Trager and Jensen (21), in RPMI-1640 medium with glutamine (Life Technologies, Bethesda, MD) containing 0.2% sodium bicarbonate, 25 mM HEPES, 0.2% glucose, and 10% human AB+ serum in the presence of O+ erythrocytes at 37°C in a humidified atmosphere consisting of 5% CO2, 5% O2, and 90% N2. Cultures were performed routinely in duplicate culture flasks.
For different cell cycle stage examinations, P. falciparum cultures were synchronized by a double sorbitol treatment conducted 26 h apart, as described elsewhere (22).
Intraerythrocytic Parasite Staining With HE
The method of van der Heyde et al. (17) was adapted to enumerate erythrocytes containing viable malaria parasites. Cultures of infected erythrocytes were incubated with the vital dye HE, which is converted by metabolically active cells to ethidium. The interaction of ethidium and nucleic acids of parasites results in a red fluorescence emission that allows a differentiation between infected erythrocytes, at whatever stage, and uninfected erythrocytes when using flow cytometry.
For HE staining, a stock solution of HE (10 mg/ml in dimethyl sulfoxide; Polysciences, Warrington, PA) was prepared and stored at −20°C. HE was added directly to the culture medium at a final concentration of 50 μg/ml. The culture was incubated for 20 min at 37°C, in the dark. After two washes in phosphate buffered saline, TO staining was performed.
Intraerythrocytic Parasite Staining With TO
TO was selected as the best dye for RNA reticulocyte analysis (18) because of it strong enhancement of green fluorescence upon binding to RNA. In fact, DNA and RNA are stained by TO.
Intraerythrocytic parasite staining was performed as described previously (20). After staining by HE, 5 × 106 parasitized red blood cells were pelleted by centrifugation and resuspended in 1 ml of a solution of TO (Retic-Count, Becton-Dickinson, San Jose, CA) and kept in the dark for 30 min at room temperature before flow cytometric analysis.
Flow Cytometric Analysis
Flow cytometric data acquisition and analysis were done on a FACSCalibur (Becton-Dickinson). The detectors of forward scatter, side scatter, FL1 (for detection of TO), and FL2 (for detection of HE) were set in logarithmic mode. Appropriate electronic color compensations were adjusted between TO green fluorescence and HE red fluorescence. Infected and uninfected erythrocytes were gated on the basis of their forward scatter and side scatter signals. List mode data from 50,000 cells were stored and processed with CellQuest software.
For analysis of parasitized red blood cells, HE and TO fluorescence intensities were expressed as geometric means values of events occurring in the different parts of the quadrant (fluorescence arbitrary units, FAU). Raw FAU values were divided by the fluorescence intensity of the non-parasitized red blood cells (corresponding to the basal autofluorescence of the non-parasitized red blood cells) to give the processed values, HEp and TOp. The total fluorescences, HEt and TOt, corresponding to a quadrant part was obtained by multiplying HEp and TOp, respectively, by the number of parasitized red blood cells observed in the corresponding area.
Asynchronous Culture Analysis
Figure 1A displays a typical flow cytometric dot plot of an asynchronous culture of the P. falciparum 3D7 strain doubly stained with HE and TO: in the lower left panel (LL), the uninfected red blood cells are doubly negative; in the lower right panel (LR), the red blood cells parasitized with rings and young trophozoites (R + YT) are intensively labeled by HE, but their TO labeling does not differ from that of the uninfected red blood cells. In contrast, the red blood cells infected by trophozoites and schizonts (T + S) are strongly stained by both dyes, as shown in the upper right panel (UR).
Table 1 shows the raw data of each part of the quadrant of Figure 1 and the values obtained after processing. Raw HE and TO fluorescence values were expressed as geometric mean values of events occurring in the different parts of the quadrant (FAU). Processed values, HEp and TOp, were obtained by dividing the raw FAU value by the fluorescence intensity of the non-parasitized red blood cells. This step was used to overcome inter-experiment variations of the fluorescences measurements. Total fluorescence values, HEt and TOt, were obtained by multiplying the HEp and TOp values, respectively, by the number of parasitized red blood cells observed in the corresponding part of the quadrant.
Table 1. Processed Data of the Raw Values collected in Different Parts of the Quadrant of Figure 1*
Differential parasitemia (%)
% Total fluorescence
HE, hydroethidine; LL, lower left corner in Figure 1; LR, lower right corner in Figure 1; PRBC, parasitized red blood cell; R + YT, rings and young trophozoites; RBC, red blood cell; T + S, trophozoites and schizonts; TO, thiazole orange; UR, upper right corner in Figure 1.
LR (R + YT)
UR (T + S)
The processed results presented in Table 1 show that the parasitized red blood cell population was selectively discriminated by HE staining and consisted of 70% rings and young trophozoites (R + YT) and 30% trophozoites and schizonts (T + S), and this result was confirmed by microscopic examination of Giemsa-stained slides (data not shown). From these results, it appeared that the major parasite population, R + YT (70% of the parasites), contributed 33% of the total HE fluorescence and 12% of total TO staining only. These results are in accordance with respective nucleic acid contents of young and old stages of erythrocytic forms of P. falciparum (23–25).
Synchronous Culture Analysis
Next we examined synchronized parasites cultures to analyze the populations in terms of TOt labeling. Figures 2A1 and 2A2 show two independent experiments of HE/TO double staining of synchronous cultures sampled just at the end of re-invasion, when the percentages of R + YT were 99.2% (Fig. 2A1) and 94.7% (Fig. 2A2). In each culture TOt fluorescence from R + YT accounted for 75.3% and 46.5% of the total TOt labeling, respectively. The presence of only few schizonts in these preparations resulted in a high proportion of TOt labeling from schizonts in the cultures (24.6% and 53.4% for Figl 2A1 and 2A2, respectively, for only 0.8% and 5% contaminating schizonts, respectively).
Figures 2B1 and B2 show results obtained with two independent synchronous cultures sampled 24 h after re-invasion and present a large difference in synchronicity according to the distribution of parasites in the UR and LR of the quadrant. Despite a large fraction of R + YT in the first sample (45%), the trophozoites TOt label accounted for 92% of the total labeling. In the second sample, where R + YT represented 39% of total parasites gated, trophozoites TOt labeling contributed to 97% of the total fluorescence. In contrast to TOt labeling of synchronized R + YT, TOt labeling of trophozoites was far less sensitive to contamination by younger parasite stages. Further analysis of 16 samples from seven different synchronous cultures shown in Figure 3 confirmed that a low contamination of ring stages by trophozoites and schizonts is correlated to a dramatic increase of the contaminant-stage contribution to the fluorescence measured. With more than 5% of mature stages in a ring preparation, the cross contamination by mature-stage nucleic acids (>40%) may be difficult to avoid and may compromise specific ring-stage transcriptome analysis.
The present data show that estimation of the degree of synchronization, particularly for rings stages, is feasible based on the distribution of parasites in the quadrants. By using a logarithmic scale for data acquisition, we could cover the large range of fluorescence labeling observed from young to old parasite stages. Although this double-staining technique did not allow the determination of the actual number of parasite per infected red blood cell, it did not influence dramatically the distribution of the different parasite populations identified. Concerning gametocytes, this approach did not allow their differentiation, but analysis of Giemsa-stained slides throughout the different experiments showed that they represented fewer than 1% of the total parasite population.
Despite the fact that HE and TO stain both DNA and RNA, we observed that the contributions of fluorescence by R + YT and T + S, respectively, to the total fluorescence were in accordance with previous studies on nucleic acid synthesis throughout the P. falciparum erythrocytic cycle (23–25).
For P. falciparum transcriptome studies, synchronous cultures are required to avoid cross contaminations between different stages. Our approach showed that a very low level of contamination of the ring population by schizonts results in a significant contamination of ring nucleic acids by schizont nucleic acids. The results of experiments searching for specific ring-stage gene expression may be severely biased in such conditions. The reverse is not true because ring contamination of trophozoite and schizont preparations resulted in only minor contamination by ring nucleic acids. Transcriptome analysis of trophozoite and schizont stages may be more representative. Because the different parasite stages can be purified by using centrifugation on Percoll gradient as seen by flow cytometric analysis (20) or metrizamide, the results obtained can be expected to be more accurate by combining sorbitol synchronization and Percoll centrifugation.
In conclusion, double labeling of P. falciparum-synchronized cultures with HE and TO is a powerful and rapid method to control their degree of synchronization when using a single 488-nm laser flow cytometer. We demonstrated the low nucleic acid content in the ring stage when compared with the high nucleic acid content of schizont-stage parasites. The results in this study call for caution in interpreting ring transcriptome profiles because of the risk of contamination by nucleic acids of other stages. This should be taken into account in any study of P. falciparum transcriptome during its erythrocytic cycle.
We thank Gaetane Woerly for critical reading and Peter David for helpful comments.