The response of Plasmodium falciparum to isoleucine withdrawal is dependent on the stage of progression through the intraerythrocytic cell cycle

The NF54 clone used for most experiments was produced by and obtained from the lab of David A. Fidock at Columbia University [15]. The Hsp101-DD conditional PTEX export 3D7 clone [16] was produced by and obtained from the lab of Dan Goldberg at Washington University.

Parasites were cultured using standard methods [17] in flasks gassed with a mixture of 90% N₂, 5% CO₂, 5% O₂ (Airgas, #Z03NI9022000033). Normal culture medium consisted of RPMI 1640 with l-glutamine and 25 mM HEPES (Corning, #10-041-CV), with 5.0 g/L Albumax II Lipid-Rich BSA (ThermoFisher Scientific, #11021029), 3.7 mM hypoxanthine, and 50 μg/mL gentamicin. The isoleucine deficient medium consisted of 10.3 g/L RPMI 1640 Ile-Dropout medium (US Biologicals, #R9014), supplemented with 2.0 g/L NaHCO3, 6.0 g/L HEPES, 5.0 g/L Albumax II, 3.7 mM hypoxanthine, and 50 μg/mL gentamicin. Trimethoprim was obtained from Sigma Aldrich (#T7883) and used at a final concentration of 10 μM.

For each sample to be analysed, 1 mL of parasite culture was pelleted and resuspended in 1 mL of PBS containing 4% w/v paraformaldehyde. Samples were rocked at 4 °C for 20–30 h of fixation. Samples were then pelleted and resuspended in PBS containing 0.1% v/v Triton-X 100 and rocked at room temperature for 1 h for permeabilization. This process was then repeated three additional times with PBS (without Triton-X 100) to remove as much haemoglobin as possible. Samples were then diluted approximately 100-fold into normal culture medium containing 1× SYBR Green I (ThermoFisher Scientific, #S7563) and analysed on a FACSCalibur (BD Biosciences) flow cytometer. The resulting data was further analysed using FLOWJO 10.0.7 analysis software. To determine the DNA content of infected red blood cells, an asynchronous culture was stained for DNA and analysed by flow cytometry to determine the median fluorescence intensity of the 1C parasite population. The higher DNA content values were called based on cell populations having an intensity peak at multiples of the median 1C fluorescence intensity obtained during the same flow cytometry session (e.g. a cell population with a median fluorescent intensity at twice the 1C intensity would be called as ‘2C’).

All statistical analyses were performed using R version 3.2.3 (2015-12-10) [18]. T-tests, linear and logistic regressions, and some figures were generated using R base functions. The drc package [19] was used for determining the 50% effect points of logistic models, and the ggplot2 package [20] was used for the production of several figures.

Babbitt et al. reported that when young rings are deprived of extracellular isoleucine the parasites enter a physiological state reminiscent of hibernation or aestivation in higher animals: a reversible state of low-metabolic activity and high resource conservation [12]. When resupplemented with isoleucine, parasites immediately resumed normal growth, displaying only modest losses in viability after 72 h of withdrawal. Over the course of 72 h of isoleucine deprivation, the parasite’s progression through the cell cycle was drastically retarded, advancing at 40% of the pace of control cultures and never progressing past haemozoin-containing, pre-S-phase trophozoites. The Babbitt et al. experiments were repeated to verify that the phenotype holds for the NF54 subclone of P. falciparum used in the authors’ laboratory.

As observed by Babbitt et al. [12], young rings (~ 6 h post-infection or hpi) washed thoroughly with PBS and transferred to culture medium lacking isoleucine remained at the starting parasitaemia after 72 h, whereas control parasites (standard RPMI 1640 with 382 μM Ile) completed a cell cycle and re-invaded red blood cells over the same time period (Fig. 2a). When the isoleucine-deprived parasites were resupplemented with complete medium for 72 h, they displayed growth similar to the control cultures that had never been deprived (Fig. 2a). Thus, the isoleucine response described by Babbitt et al. is robust and reproducible in the NF54 clone used in the authors’ laboratory.

Over the course of the experiments, parasites with greater than the 1C haploid DNA content were detected in cultures deprived of isoleucine for over 48 h, suggesting that, despite growth retardation, some parasites were able to commence DNA synthesis in the absence of extracellular isoleucine. To investigate this in more detail, isoleucine-dropout and isoleucine-normal cultures from late-stage, Percoll-isolated schizonts were initiated and tracked for their development over the ensuing 6 days.

Control parasites developed as expected, with schizonts appearing in Giemsa-stained smears roughly every 45 h, followed by an increase in parasitaemia with each generation (Fig. 2b, d). DNA content analysis by flow cytometry showed that the proportion of parasites containing 1C DNA peaked approximately every 45 h (Fig. 2c). Isoleucine-deprived parasites also progressed through the cell cycle, albeit at a much-reduced rate (Fig. 2b). Schizonts first became detectable at around 90 h of incubation, and the first rings of the second generation were observed at 100 h of incubation. DNA content tracking showed that 1C parasites peak at roughly 90 h, corresponding to about 50% reduction in the pace of the cell cycle (Fig. 2c). Thus, isoleucine-deprived parasites do not arrest prior to S-phase to assume a dormant state, but instead continuously progress through the cell cycle at a retarded pace. This continued growth during isoleucine withdrawal may account for why parasites that were deprived and then resupplemented, sometimes display higher growth than normal medium controls upon recovery (Fig. 2a). After 72 h of slow growth in media lacking isoleucine, the parasites would have reached S-phase. Upon return to normal media, the 72-h recovery period would allow these late stage parasites to egress and reinvade twice.

While control parasites increased in parasitaemia with each generation, the isoleucine-deprived parasites decreased in numbers at a modest, but steady rate over the course of a 6-day incubation (Fig. 2d). This suggests that, despite completing the cell cycle and re-invading at roughly 4-day intervals, the overall viability of the slow growing parasites was affected by exogenous isoleucine withdrawal. This agrees with the observation of Babbitt et al. of diminished recovery upon isoleucine resupplementation as a function of increased withdrawal time. The mean DNA content of parasites at schizont-enriched time points was substantially lower for isoleucine-deprived parasites compared to controls (Fig. 2e). This suggests that either isoleucine-deprivation produced schizonts with a lower average merozoite count, decreasing the growth potential of each cell, or that a majority of cells experienced catastrophic failure early in schizogony and produced no viable progeny. The low number of schizonts at these timepoints prevented experiments distinguishing these possibilities.

Next, the effect of the stage of cell cycle progression on the response to isoleucine withdrawal was investigated.

A culture of highly synchronous young rings was initiated in normal medium. At 3-h intervals, an aliquot of the culture was removed, washed repeatedly to remove any traces of isoleucine, and resuspended in medium lacking isoleucine. After 30 h or 60 h of incubation (for time points > 24 h or < 24 h respectively), the parasitaemia of the cultures was quantified by flow cytometry to determine the relative growth of each sample.

As expected from earlier experiments, young rings remained at roughly the starting parasitaemia over the course of the incubation. However, as the time of culture increased, a notable increase in the final parasitaemia became evident (Fig. 3). A logistic regression performed on the data confirmed that the likelihood of completing the cell cycle increased as the time of culture in the presence of isoleucine increased (p = 5.85 × 10⁻⁸², see Additional file 1: Table S1 for regression summary). Based on the logistic model fit to the data, at approximately 30 h post-invasion, 50% of parasites transferred from normal culture medium to isoleucine-lacking medium enter the growth-retarded ‘hibernatory’ state described by Babbitt et al., while the remaining 50% continue the cell cycle at the normal pace and enter the slow-growth state in the subsequent cycle. Beyond 30 h, an increasing proportion of parasites complete the cell cycle uninhibited by the absence of extracellular isoleucine. These data suggest that at around the 30 h into the cell cycle, molecular changes occur to render the parasite refractory to isoleucine withdrawal.

Inferring cellular events in the cell cycle of P. falciparum based on time estimates alone can be misleading, as the methods used for synchronization are imperfect, and the in vitro length of the intraerythrocytic cycle has been reported to range anywhere from 38 to 50 h for different parasite isolates and clones [12, 21]. Seeking a better understanding of the basis of refractoriness to isoleucine withdrawal, an attempt was made to link this response to well-established physiological milestones of the cell cycle.

The transition of the parasite from the ring form to the trophozoite form is marked by a dramatic change in the permeability of the host erythrocyte to a range of solutes. This new permeability pathway (NPP) is due to the activation of a parasite-derived or modified host channel in the erythrocyte membrane [22]. Isoleucine is one of the biologically relevant solutes affected by the NPP. While erythrocytes display a basal permeability to isoleucine via the erythrocyte L-system, activation of the NPP increases isoleucine uptake by fivefold [23]. In vitro, NPP activation renders parasites permeable to sorbitol, which leads to rapid lysis of infected erythrocytes with an active NPP (trophozoites and schizonts) but leaves impermeable cells (rings) unharmed. This property provided the basis for a commonly used method for ring-stage synchronization [24]. Given the role of the NPP in isoleucine uptake and the availability of a robust assay of NPP activation, the timing of the isoleucine refractory point relative to NPP activation was determined.

As before, synchronous parasites in normal medium culture were transferred to medium lacking isoleucine at 3-h intervals across the 45-h life cycle. At each time point, duplicate aliquots of culture were treated with 5% w/v sorbitol and resuspended in normal culture medium for a 72-h recovery period. The relative growth of the two treatments is shown in Fig. 4a. As with isoleucine withdrawal, a logistic regression fit to the sorbitol treated cultures shows a marked time-dependent response. Sorbitol-treated samples display decreased growth with increasing parasite age post-invasion, as would be expected from parasites becoming susceptible to sorbitol lysis as they reach the point of NPP activation (p = 8.73 × 10⁻⁹⁰). Notably, the logistic model fit to the data maps the 50% relative survival point for sorbitol treatment (23.3 ± 0.4 h) well in advance of the isoleucine refractory point (in this case, 32.2 ± 0.5 h), and finds that sorbitol sensitivity is only a weak predictor of the response to isoleucine removal (p = 0.0136) relative to time of isoleucine removal (p = 1.80 × 10⁻²⁸, Additional file 1: Table S1).

Another molecular milestone in the P. falciparum cell cycle is demarcated in late rings by export of certain parasite proteins across the parasitophorous vacuolar membrane into the host erythrocyte by the Plasmodium Translocon of Exported Proteins (PTEX) complex. Beck et al. [16], produced a transgenic parasite line in which an essential component of the PTEX complex, Hsp101, was modified such that its activity could be regulated by the chemical ligand trimethoprim (TMP). Removal of TMP from the culture medium leads to inhibition of Hsp101 function, blocking protein export by the PTEX complex. When TMP was removed from the culture at the ring stage, parasites arrested growth as morphological rings, but could resume growth without a loss in viability if the culture was resupplemented with TMP within 48 h. However, if TMP was removed from the culture during the mid-trophozoite phase, the parasites completed the cell cycle unperturbed and reinvaded new cells whereupon they arrested at the late ring stage [16]. This suggests that PTEX-dependent export defines another checkpoint-like event in the cell cycle of P. falciparum. This same parasite line was used to determine if the PTEX-dependent checkpoint-like event overlaps with the isoleucine refractory point.

Synchronous young rings were sampled at 3-h intervals across the cell cycle and transferred either to medium lacking isoleucine (but supplemented with TMP), or normal medium lacking TMP. The relative growth of each sample is shown in Fig. 4b. Logistic regression models fit to both treatments show a time-dependent binary response and that the 50% growth point for PTEX-dependent growth-arrest (17.8 ± 0.4 h) is well ahead of the isoleucine-dependent growth-retardation point (here 27.4 ± 0.4 h). Like sorbitol sensitivity, PTEX-dependent growth arrest is a weak predictor of isoleucine refractoriness (p = 0.00338) relative the point in the cell cycle at which extracellular isoleucine is removed (p = 5.59 × 10⁻⁴⁴, Additional file 1: Table S1). The 3D7 Hsp101-DD clone used for this assay displays a shorter cell cycle time than the NF54 clone, which likely contributes to the somewhat earlier timing the isoleucine response point. While it is difficult to project the PTEX-arrest point onto the sorbitol-NPP point calculated above with the NF54 clone, Beck et al. [16] demonstrated that PTEX-arrested parasites were resistant to sorbitol treatment, placing activation of the NPP after the PTEX-dependent arrest point.

Next, the mapping of the isoleucine refractoriness point relative to the transition into S-phase and schizogony was tested. Synchronous young rings were transferred from normal medium into isoleucine-depleted medium at 3-h intervals. At each time point, aliquots were fixed, permeabilized, and stained with SYBR GREEN I for DNA content analysis by flow cytometry (Fig. 4c). A logistic regression model fit to the proportion of parasites with > 1C DNA content overlaps with the proportion of parasites continuing normal growth, with 50% of the parasites transitioning from 1C DNA content to > 1C content (29.9 ± 0.5 h) at a point indistinguishable from the point at which 50% of the parasites become refractory to isoleucine withdrawal (29.4 ± 0.4 h). In fact, when a regression model was fit with both age post-invasion and proportion of > 1C parasites as factors, the DNA content alone predicts the response of the parasites to isoleucine (p = 5.38 × 10⁻¹² as compared to p = 0.577 for hour of Ile removal alone, Additional file 1: Table S1). From the same data, the timing of the transition from < 2C to > 2C DNA content relative to the isoleucine-dependent point was also tested. While temporally close, the transition from < 2C to > 2C (31.7 ± 0.5 h) occurs after cells have become refractory to the removal of isoleucine (Fig. 4d). The transition from below 2C to greater than 2C DNA content does not predict the response to isoleucine removal (p = 0.341, Additional file 1: Table S1).