A high-throughput automated RNA extraction method yields high-quality RNA
We developed a high-throughput platform to robustly screen mRNA levels and monitor gene expression profiles of Pf-infected RBCs (Pf-iRBCs), particularly to determine the immediate transcriptional responses of Pf sexual markers. A robot-automated technology, such as the Biomek i5 Automated Workstation used here, enables the use of very low volumes of Pf-iRBCs to process numerous samples simultaneously. Importantly, we established a platform with modular programmed methods for this robotic workstation, which consists of three sequential, ‘hands-free’ steps: (i) RNA extraction, (ii) cDNA synthesis and (iii) RT-qPCR plating. To ensure the suitability of the extracted RNA for the RT-qPCR assays, the yield and quality of the extracted RNA obtained using our robot-automated magnetic binding (rMB) method were determined. We compared our rMB method to two widely used RNA methods: the liquid–liquid solvent extraction method (LLE) [67] and the solid-phase extraction method (SPE), which employs silica columns [21].
RNA was extracted using each method (rMB, LLE and SPE), and UV spectrophotometry was used to assess the quantity and quality of the extracted RNA (Fig. 1a–d). We show that when using a minimal Pf-iRBC input (40 µl for trophozoite-stage parasites and 80 µl for ring-stage parasites), the yield of total RNA extracted from Pf-iRBCs using our rMB method is similar to the yield obtained using the SPE method (Fig. 1a, b). The LLE method seemed to yield a significantly higher RNA concentration than the two other methods. However, measurements of the A260/A280 purity ratio correlated to protein impurities in the extracted RNA [68], demonstrating that the rMB and SPE methods exhibit the highest RNA purity within the ideal range of 2.0–2.2 for pure RNA samples accepted in the field [68] (Fig. 1c, d). Conversely, the LLE method yielded significantly lower purity, indicative of high levels of protein contamination [68].
The A260/A230 ratio was found to be below the ideal value of 1.8 [68] when extracting RNA from ring or trophozoite Pf-iRBCs with all three methods (Additional file 4: Fig. S1). It is known that the A260/A230 ratio can be affected by numerous contaminants, such as guanidinium salts, ethanol, isopropanol and detergents, which are common reagents used in RNA extraction methods [68]. Furthermore, these spectrophotometric purity parameters have been demonstrated to be highly variable, and thus unreliable for samples with low RNA concentrations [68]. Of note, according to the current MIQE guidelines for RT-qPCR assays [64], reporting the A260/A230 ratio of RNA is not considered to be essential information. Nevertheless, we performed two additional assays (capillary electrophoresis and agarose gel electrophoresis) to assess the quality of the extracted RNA.
First, we assessed RNA integrity via agarose gel electrophoresis, and the results indicated non-degraded RNA samples [69, 70] (Additional file 5: Fig. S2). We further assessed RNA integrity by capillary electrophoresis. The RNA integrity number (RIN) was determined for a selected number of samples from each biological replicate (Fig. 1e, f). A RIN value of 8.5 is considered as the lower limit for high-quality RNA integrity [71, 72]. Here, we show that high RIN values were obtained for each of the three RNA extraction methods (Fig. 1e, f), indicating a high-quality integral RNA suitable for downstream RT-qPCR applications [71, 72]. Representative capillary electropherograms of samples extracted from each method are presented in Additional file 6: Fig. S3.
Collectively, these four different analyses demonstrate that the RNA extracted using our rMB method is comparable in yield, quality and integrity to traditionally accepted RNA extraction methods frequently used in gene expression studies in malaria [12, 21, 35].
The automated RT and RT-qPCR methods meet optimal requirements for gene expression analysis
Having established the reliability of our RNA extraction method, we could direct our automated platform towards the downstream profiling of gene expression in Pf-iRBCs. Importantly, this platform enables us to systematically monitor a large-scale number of samples (up to 96 simultaneously), with minimum manual intervention, a major advantage when compared to the standard Pf-iRBC RNA extraction methods currently employed in the field. Thus, we proceeded to generate automated protocols for two additional sequential steps: cDNA synthesis (reverse transcription, RT) and RT-qPCR. Since sexual commitment is an essential process in malaria transmission, we decided to apply our system to evaluate the expression of a selected panel of early gametocytogenesis-related genes (GRGs) (Fig. 2a). The RT-qPCR panel used for the analysis included a total of 10 parasite genes: two Pf reference genes, the protein kinase 4 pk4 and the ubiquitin-conjugating enzyme uce [21], the gametocytogenesis master regulator PfAP2-G [12], the sexual ring marker gexp05 [18], four early gametocyte markers, Pfg14.744, Pfg14.748, Pfs16 and Pfg27 [17, 73], the mature gametocyte marker Pfs25 [74] and the ring-stage marker sbp1 [75]—to serve as a non-gametocyte control.
To evaluate the technical repeatability of the RT-qPCR assay [64], we extracted RNA from either synchronized trophozoite-stage or ring-stage Pf-iRBCs. We then synthesized cDNA and performed RT-qPCR to estimate the basal expression levels of each of these genes on three independent biological replicates and, in each, three technical replicates. It is important to note that, although the predominant parasite forms in the cultures (rings or trophozoites) are asexual stages, a basal number of these asexual parasites continuously commit to sexual development and produce gametocytes. Thus, this existing low level of committed asexual and developing sexual parasites is responsible for the observed expression of gametocyte-specific transcripts [20, 76]. The Ct values of the amplification plots of the gene panel showed clear repeatability and reproducibility [64], as demonstrated by the standard deviation values of the replicates. Notably, these values were mostly smaller than one amplification cycle (Ct) between the three independent biological replicates for each Pf blood stage (Fig. 2a). Correspondingly, the standard curves applied for quantification of this gene panel showed optimal efficiencies (between 90–100%) and slopes for accurate quantification, as well as correlation coefficients (r2) of at least 0.99 (a representative example from one biological replicate is presented in Additional file 2: Table S2).
As a measure of reproducibility, the standard curves for each gene were plotted for three independent biological replicates of ring- and trophozoite-stage parasites (Additional file 7: Fig. S4) and the statistics applied to their means (mixed model ANOVA for batch effect). Both the efficiency and r2 of the average of the standard curves were also close to optimal values, and data dispersion was minimal (statistical parameters of the mean of the standard curves are presented in Additional file 3: Table S3), confirming the high level of reproducibility of the RT-qPCR standard curves. Analysis of the melting curve of the amplification product of each pair of primers in the gene panel showed one single discrete peak in the derivative of the fluorescence (Additional file 8: Fig. S5), demonstrating the absence of unspecific amplification products and primer specificity as per the accepted guidelines [64].
We further performed a No Reverse-Transcriptase (NRT) control assay , which enabled us to validate the absence of residual gDNA, post-RNA extraction (Additional file 9: Fig. S6). We extracted RNA using our rMB method from trophozoite-stage samples. During cDNA synthesis, reverse transcriptase was either added as per the protocol (RT-treated, yellow bars) or omitted (Non RT-treated, green bars). The cDNA synthetized from each sample was then used for RT-qPCR amplification as well as a no-template control (NTC, gray bars). As observed, the RT-treated samples presented Ct values well below the Ct threshold of 35, indicating positive amplification. However, the Non RT-treated samples and the no-template control presented no significant levels of amplification (Ct > 35 considered negligible [63]). These data indicate that there was no significant gDNA contamination in the DNAse-treated RNA samples, which could interfere with the RT-qPCR reaction cycles.
It is important to clarify that for several genes from our selected gene panel, the no-template control showed a weak amplification curve in some of the technical repeats with Ct values around cycle 35. This was likely due to the presence of primer dimers caused by multiple AT pairs in their sequence (presented in Additional file 1: Table S1), a common feature of malaria gene primers, as expected in cases of extremely high genomic AT content (more than 80% in coding sequences) [77]. This structural bias makes it technically challenging to generate highly sequence-specific primers that do not generate background fluorescence when using SYBR-green amplification. However, these low amplification curves were not observed consistently, even within the technical replicates of the no-template controls. Additionally, none of them appeared as considerable unspecific products in the melting curve analysis of the primers (Additional file 8: Fig. S5).
To determine potential contaminations originating from previous steps (i.e. RNA extraction, cDNA synthesis) that could inhibit the catalytic activity of the DNA polymerase during the RT-qPCR reaction, the SPUD assay is a reliable and quantitative quality control system [64]. Thus, a SPUD assay was carried out using cDNA prepared with our methods following RNA extraction from rings or trophozoite Pf-iRBCs. No statistically significant difference was observed between the mean Ct value of the SPUD amplicon alone or when the reaction is carried out in presence of spectator cDNA prepared from Pf-iRBCs (Fig. 2b), indicating absence of DNA polymerase inhibitors in the samples that could interfere with the RT-qPCR amplification. Phenol was used as a positive control for DNA polymerase inhibition [65], and no amplification of the SPUD template was observed in the presence of phenol (Fig. 2b).
Together, these data demonstrate the high robustness and reproducibility of the developed robotic methods and validate their effectiveness in RT-qPCR gene expression monitoring.
Treatment with DHA alters the PfAP2-G expression profile and affects Pf asexual stage composition
Once we ensured that our RT-qPCR platform met the accepted quality standards, we aimed to exploit this high-throughput system to monitor how GRGs modify their temporal expression pattern under a condition previously suggested to induce sexual conversion. The antimalarial drug dihydroartemisinin (DHA) has been previously reported to induce sexual conversion when applied at sublethal concentrations [35]; hence, DHA was selected as a reference treatment. Since trophozoites have been shown to be more responsive to gametocytogenesis induction [20, 35], we examined the effect of DHA on the expression profile of selected genes of our panel when applied for a short pulse (3 h) during the trophozoite stage at sublethal concentrations.
We thoroughly evaluated the temporal expression profile of Pf early GRG markers (PfAP2-G, gexp05, Pfg14.748, Pfs16, with sbp1 used as an internal control) over the course of 48-h post-treatment (hpt) at six different time points: 4, 8, 12, 25, 30 and 48 h (Fig. 3a). We observed a significant change in the expression pattern of PfAP2-G, the key gametocytogenesis regulator gene, at 25 hpt and 30 hpt when treated with the highest concentration of 10 nM DHA (Fig. 3b). As observed, PfAP2-G expression pattern peaked at 12 hpt for all samples (including the solvent control and markers treated with the low concentration of DHA) but rapidly decreased in the time points that followed, except for the samples treated with 10 nM DHA. Under this treatment, the parasites continued to express high levels of PfAP2-G even at 25 hpt, and the mRNA levels only started to decrease at 30 hpt and beyond (Fig. 3b). These results support a previous report demonstrating the induction of sexual commitment gene expression by DHA at sublethal concentrations [35].
Pretreatment of parasite cultures with choline, a known gametocytogenesis repressor [9, 33], has been used in previous studies to maintain parasites in a non-induced state and evaluate the effect of treatments instead of constant gametocytogenesis induction in the absence of choline [35]. In our study, treatment with choline to maintain parasites in non-inducing conditions was not required to observe changes in the pattern of PfAP2-G expression (Fig. 3b). These results suggest that our platform could provide an improved resolution, enabling detection of small changes in gene expression even in inducing conditions in absence of choline as well as a PfAP2-G induction in a different experimental setup than previous reports [35].
While an alteration in PfAP2-G expression was observed under the DHA treatment, we could not detect alterations in its target genes, Pfg14.744 and Pfs16, or in the sexual ring marker gexp05, over the course of 48 h (Fig. 3c–e). Thus, to determine whether pretreatment of the parasites with choline would increase our method’s ability to detect alterations in GRGs other than PfAP2-G by repressing the basal sexual commitment [19, 35], we cultured parasites in the presence of choline. Pf-iRBCs were pretreated with 2 mM choline for a minimum of two intraerythrocytic cycles; then, trophozoite-stage parasites were treated with a short (3 h) pulse of 10 nM DHA, and GRG expression was measured 24 hpt (Fig. 3g). Notably, we could detect a significant elevation in the PfAP2-G transcription level under DHA treatment. However, no significant changes were detected in the expression of PfAP2-G target genes, gexp05, Pfg14.744 and Pfg14.748 (Fig. 3g).
Importantly, DHA treatment resulted in a significant decrease in the expression level of sbp1, a highly specific ring-stage marker [6], when compared to the no drug control. This suggests either a delayed intraerythrocytic growth of the asexual forms in the DHA-treated samples or an increased number of parasites that converted into gametocytes (and consequently a reduced number of asexual parasites detectable by sbp1). This encouraged us to conduct a Pf growth assay combined with a blood stage composition analysis (rings vs. trophozoites vs. schizonts) under DHA treatment as well as a gametocyte culture assay to directly quantify gametocyte abundance by Giemsa smears after DHA treatment (Fig. 3h, Additional file 10: Fig. S7 and Additional file 11: Fig. S8). NF54 trophozoite-stage Pf-iRBCs were treated with DHA for 3 h, and FACS analyses combined with microscopy counting of Giemsa-stained smears were used to evaluate the effect of DHA on parasitic growth dynamics. As expected, exposure to sublethal concentrations of DHA (for 3 h) decreased parasite growth as early as 24 h post-treatment (upon the first RBC invasion). These results are in alignment with previously reported studies [35]. In our study, the detrimental effect of DHA was further observed during the second RBC invasion cycle (72 hpt). Moreover, stage composition analysis revealed delayed intraerythrocytic parasite development (Additional file 10: Fig. S7b), with a significantly smaller proportion of ring-stage parasites observed in the DHA-treated samples when compared to the non-treated control. These findings are also in accordance with previous published reports [78].
Furthermore, microscopic evaluation of gametocyte abundance after DHA treatment did not reveal a significantly increased proportion of sexual parasites (Additional file 11: Fig. S8), which supports the notion of the observed decrease in sbp1 expression being due to asexual growth delay and not to a significant contribution of increased commitment into the sexual stage. This result is in line with a previous report where DHA treatment did not significantly increase sexual commitment in inducing conditions (choline depletion) [35], suggesting that DHA treatment alone in inducing conditions (choline-depleted cultures) may initially alter the expression of PfAP2-G, but not necessarily produce an increased gametocyte abundance as opposed to choline supplementation, where an increase in sexual commitment after DHA treatment was previously reported [35].
Thus, the observed significant differences in asexual parasite stage composition may contribute, at least partially, to the detected differences in the expression dynamics of PfAP2-G (Fig. 3b). Indeed, under DHA treatment, PfAP2-G expression remained elevated for a longer period and peaked between 12 and 30 hpt (Fig. 3b), coinciding with a significantly lower percentage of ring-stage parasites (Additional file 10: Fig. S7b) but no net increase in gametocyte abundance (Additional file 11: Fig. S8). Since mature schizonts express PfAP2-G at high levels [21], our data suggest that changes in asexual stage composition caused by DHA can be responsible for the observed increase in PfAP2-G mRNA levels as detected by RT-qPCR; thus, an isolated change in PfAP2-G levels needs to be cautiously interpreted, in light of analysis of both asexual and sexual populations.
We next monitored the expression profiles of GRG under treatment with sub-lethal doses of chloroquine (CQ), another pertinent antimalarial drug. In this case, no significant alterations in sexual gene expression were observed (Additional file 12: Fig. S9), in line with previous findings [35, 40].
Overall, our findings emphasize the complexity in monitoring immediate responses in gametocyte markers (within a period of 48 h) and the importance of different experimental settings to evaluate gametocyte commitment (inducing or non-inducing conditions by choline supplementation/Kennedy pathway activation) and highlight the need to combine complementary approaches to interpret GRG expression changes.
Choline treatment represses PfAP2-G expression and affects Pf stage composition by increased growth
We showed that DHA treatment induced PfAP2-G expression as well as modified parasite growth and stage transition. Therefore, we decided to examine the expression profiles of Pf GRG using a gametocyte induction method that avoids stress on parasitic development. Choline is a key metabolite in the Kennedy pathway, a metabolic pathway active in malaria parasites and central to the formation of cellular membrane phospholipids during schizogony [34]. It has been previously reported that the presence of choline inhibits gametocyte formation [33] but does not inhibit parasitic growth [34]. Accordingly, we used our system to monitor the expression pattern of a selected number of GRG in choline-treated Pf cultures, following its depletion in the trophozoite stage. The NF54-gexp02-Tom fluorescent Pf line [19], previously used to evaluate both GRG expression by RT-qPCR and the percentage of gametocyte commitment rate (%GCR) [35] by flow cytometry, was used. This line expresses a fluorescent reporter under the control of the gexp02 promoter, a highly specific early gametocyte marker and a target of PfAP2-G [13, 20]. The expression of the fluorescent marker begins early on at the sexual ring stage [19].
Using our robotic platform, we generated a high-resolution expression profile for nine Pf genes over the course of approximately one intraerythrocytic cycle (~ 48 h) over nine time points: 0, 6, 12, 18, 24 30, 36, 42 and 48 h post-choline removal (Fig. 4a). Trophozoite-stage parasites routinely cultured with 2 mM choline supplementation for a minimum of 2 weeks were washed, and culture media were replaced with fresh media containing either no choline (− Choline) to induce gametocytogenesis [19, 35] or 2 mM choline (+ Choline), which served as a control to maintain gametocytogenesis repression. NF54-gexp02-Tom parasites were sampled every 6 h over the course of 48 h (Fig. 4a). RNA was extracted by means of the robotic apparatus, and gene expression levels were evaluated using RT-qPCR analysis. This is, to the best of our knowledge, the first high-throughput RT-qPCR study to robustly monitor the gene expression profiles of Pf sexual markers at multiple time points post-induction of choline depletion.
The Pf genes analyzed by this wide RT-qPCR screening included: the sexual master regulator PfAP2-G, the sexual ring marker gexp05, the early gametocyte markers Pfs16, Pfg14.744 and Pfg14.748, and two stage-specific genetic markers sbp1 [6] and rhoph2 [63], asexual rings and schizont-specific markers, respectively (Fig. 4b–f). As the NF54-gexp02-Tom line was used, we also decided to monitor the RNA expression patterns of the sexual ring (and early gametocyte) marker gexp02 [19, 79]. Thus, its RNA levels could be complemented by detection of the proportion of gexp02+ parasites as a proxy for sexually committed parasites using flow cytometry (Additional file 13: Fig. S10b).
This extensive screen was performed in parallel with sampling the parasite culture at each time point using Giemsa-stained smears to (i) measure parasitemia levels and (ii) analyze blood stage composition (rings vs. trophozoites vs. schizonts) (Additional file 13: Fig. S10a-c).
We found that PfAP2-G presented a “wave-like” transcriptional profile, with peak expression at ~ 12–18 hpt (Fig. 4d) in the presence or absence of choline. Depletion of choline (- Choline) led to statistically significant upregulation of PfAP2-G, but only for a short interval between 12–18 hpt. During this time, most of the Pf-iRBCs are predominantly in the late schizont and early ring stage, with relatively low levels of trophozoites (Additional file 13: Fig. S10c). Then, as the parasites grow and transition into the later ring and trophozoite stages, the difference in PfAP2-G expression levels decreases, but becomes significantly upregulated again under choline depletion at the last time points (between 42–48 hpt). Importantly, the stage composition analysis revealed no significant changes in the proportion of schizonts or other stages at 18 h post-treatment between the samples (Additional file 13: Fig. S10c). The specific schizont marker, rhoph2 [63], did not express differently between the treatments at any time points (Fig. 4c).
Overall, these data suggest that the changes observed in PfAP2-G expression were most probably due to ‘true’ alterations in transcriptional activity within the late schizonts/early rings at 12–18 hpt when choline is removed rather than due to differences in Pf stage composition. Choline depletion did not cause significant effects on parasite growth dynamics at these early time points.
In contrast to PfAP2-G, downstream GRGs were not affected by choline depletion as would have been expected (Fig. 4f). RNA expression levels of Pfg14.744, Pfg14.748 and Pfs16 were not affected by choline depletion over the course of 48 hpt (Fig. 4f). This could reflect the fact that the expression of these genes may be directly or indirectly induced by PfAP2-G or its downstream effectors later, after the initial 48 h of gametocytogenesis induction. These findings are consistent with previous data demonstrating that the expression of these early gametocyte genes usually occurs following the first 48 h of gametocytogenesis commitment [19]. Indeed, these markers were reported to be consistently expressed only once the committed parasites transition to stage I-II gametocytes [14, 17, 52] beyond the immediate 48 h of our analysis.
We identified a different transcriptional pattern for gexp02 and gexp05 (Fig. 4e), mostly considered as markers of the sexual ring stage and early gametocytes [16, 19, 79]. The expression levels of both genes peaked when most of the culture was in the ring-stage (24–36 hpt) as revealed by the stage composition analysis (Additional file 13: Fig. S10c) and by the temporal expression pattern of the ring-stage marker sbp1 (Fig. 4b). Interestingly, for both gexp02 and gexp05, transcript abundance was significantly higher in the presence of choline (Fig. 4e). This result is surprising as choline is expected to reduce sexual commitment levels [19, 33]. These data led us to suspect that choline treatment could also cause changes in parasite growth dynamics and thus stage composition, which could lead to changes in gene expression of these markers.
Evaluation of Pf growth in the presence of choline demonstrated a significant increase in parasitemia levels (Additional file 13: Fig. S10a). Moreover, stage composition analysis demonstrated a significantly increased proportion of ring-stage parasites at 30–36 hpt (Additional file 13: Fig. S10c) when most schizonts had already ruptured, when compared to the choline depleted samples. These findings were further supported by the detection of higher transcript abundance of the specific ring-stage marker sbp1 at 24 and 36 hpt, in the choline-supplemented parasites, compared to the control (Fig. 4b). Taken together, these observations indicate a higher abundance of rings in the choline-supplemented culture compared to cultures lacking choline, suggesting that choline significantly affects Pf proliferation.
To confirm these findings, we quantified the numbers of daughter merozoites per schizont-stage parasite in each condition. Remarkably, we observed an average of ~ 19 merozoites/schizont under choline supplementation vs. ~ 15 merozoites/schizont when choline was removed (Additional file 13: Fig. S10d-e). Our data determined a significantly smaller number of daughter cells per schizont in parasites grown post-choline depletion, in alignment with previous studies reporting increased asexual proliferation and merozoite productivity in choline supplementation or Kennedy pathway activation [33, 34, 39, 80].
Taken together, we conclude that the observed increase in gexp02 and gexp05 transcript levels in the choline-treated samples may be explained by the positive effect of choline on parasite proliferation and merozoite productivity. This could have significantly increased the proportion of all, asexual and sexual, ring-stage parasites, thus enriching the number of copies of both asexual and sexual ring-stage RNA species in choline-treated parasites.
Furthermore, our results demonstrate a ‘true’ upregulation in PfAP2-G by means of choline depletion, which can be clearly observed at 12–18 hpt, in the late schizont and early ring stages. Downstream target genes may not change immediately after PfAP2-G induction and their expression levels may be biased, affected by variations in asexual stage composition. These data emphasize a need for caution when interpreting immediate changes in GRG expression rather than relying exclusively on RT-qPCR assays.
The metabolites lactate and kynurenic acid do not immediately alter early GRG expression
We aimed to use our technology to explore the involvement of lactate and kynurenic acid in gametocytogenesis regulation. Both of these metabolites have been found to be abundant in severe malaria patients [45, 46, 58, 59, 81]. Indeed, kynurenic acid was found to accumulate in the cerebrospinal fluid of cerebral malaria patients [45, 46], and high levels of lactic acid have been detected in the circulation of patients suffering from severe malaria acidosis [58, 59, 81]. The latter was suggested as a potential gametocytogenesis regulator that does not affect parasite growth [43], although no direct PfAP2-G upregulation could be observed using RT-qPCR analysis. We therefore set out to evaluate whether lactate or kynurenic acid might alter immediate GRG expression and be involved in their regulation.
Based on our previous RT-qPCR results on GRG regulation using choline (Fig. 4), we directly tested the effect of lactate and kynurenic acid in trophozoites once the culture reached the late schizont stage (~ 16 hpt). NF54 wild-type Pf trophozoite-stage parasites were cultured in the presence or absence of choline (± Choline) and the two metabolites. The parasites were treated for 16 h with lactate or kynurenic acid at concentrations that mimicked the physiological levels estimated in severe malaria patients [45, 46, 58, 59]. NF54-gexp02-Tom parasites cultured in ± choline condition were used as a control, proven their ability respond to choline depletion during the previous screening (Fig. 4). The expression levels of the early gametocyte markers, PfAP2-G and gexp02, the ring stage marker sbp1 and the schizont stage marker rhoph2 were evaluated at the late schizont stage (~ 16 hpt) time point.
As shown in Fig. 5, indeed the NF54 parental line demonstrated choline-mediated PfAP2-G repression in the schizont stage (Fig. 5a, left), as was also observed for the NF54-gexp02-Tom line (Fig. 5a right). Nevertheless, at this time point neither of the two metabolites caused a significant change in PfAP2-G or gexp02 expression (Fig. 5a–b). Of note, an elevation in the expression level of rhoph2 under choline supplementation was observed under all treatments (Fig. 5d). rhoph2 is a known marker for rhoptry biogenesis, a main component of the apical complex of merozoites [82]; therefore, the increase in its transcript abundance is likely the result of a choline-mediated increase of daughter cell productivity during schizogony (Additional file 13: Fig. S10d).
These data suggest that treatment of trophozoites with kynurenic or lactic acids does not immediately affect Pf sexual commitment. Overall, this study lays the groundwork for future studies that employ robot-automated systems like the one developed here to evaluate the expression profiles of multiple GRG markers at different time points and under multiple conditions to advance our understanding of this essential process in malaria transmission.