Introduction

Shortly after invasion of the red blood cell (RBC), the malaria parasite Plasmodium falciparum initiates a profound remodeling of its host cell by delivering proteins to the RBC membrane1. As the parasite matures, the membrane properties of P. falciparum infected-RBCs (Pf-RBCs) are modified, with markedly decreased cell deformability and acquired cytoadherence properties2,3,4,5. Altogether, these modifications affect the dynamic behavior of Pf-RBCs altering blood flow and contributing to malaria pathophysiology6.

Pf-RBCs dynamic properties are critical at the ring-stage of the infection, since this is the only asexual intra-erythrocytic stage in the peripheral circulation. Both microcapillary circuits and splenic red pulp sinus slits dynamically challenge the mechanical properties of Pf-RBCs. Several single cell assays have documented a moderate decrease in deformability of ring-stage Pf-RBCs4,7. Furthermore, there occurs a decrease in membrane flickering of Pf-RBCs starting at the ring stage of the infection8. Its consequences for the dynamic microcirculatory behavior of ring-stage Pf-RBCs are unclear. How deformability measures of Pf-RBCs assayed in static conditions translate in alteration of the microcirculatory behavior remains an open question.

In the present study, we use different approaches to document the dynamic properties and the circulatory behavior of Pf-RBCs at the ring-stage of the infection and explore the contribution of the parasite-encoded Pf155/Ring-Infected Erythrocyte Surface Antigen (RESA), which has been identified as a major contributor to the reduced ring-stage Pf-RBCs deformability in static conditions9. RESA is located in the dense granules of the invasive merozoites10, released shortly after invasion into the parasitophorous vacuole11 and rapidly translocated to the internal face of the Pf-RBC membrane. RESA is then phosphorylated12 and remains associated with spectrin13 for the first 24 h of the asexual intra-erythrocytic development. The binding site of RESA to spectrin has been localized to repeat 16 of the β-chain14. This interaction appears to favor the tetrameric spectrin state, resulting in membrane mechanical stabilization and increased membrane thermal stability14. As a result, RESA protects Pf-RBCs from vesiculation damage induced at high temperature14,15,16. Our earlier work under static conditions revealed that the role played by RESA decreasing the deformability of Pf-RBCs at the ring stage of the infection was dramatically enhanced at febrile temperature9. Here we use wild-type resa1+, resa1− and resa1-rev (the revertant isogenic resa1-rev Pf-RBCs) genetically modified parasites9 to analyze the dynamic properties of ring-stage Pf-RBCs. The resa1-rev was constructed specifically to confirm targeted gene disruption. Since no differences in membrane stiffness between wild-type resa1+ and resa1-rev were observed9 either resa1+ wild type or resa1+ revertant are used in this study as a control condition. We analyze, by diffraction phase microscopy (DPM)17,18, the membrane dynamics of ring-stage Pf-RBCs at physiological normal and febrile temperature. To assess the physiological implication of decreased deformability of ring-stage Pf-RBCs in microcirculation, we used microfluidic technology with multiple constrictions19 that replicate in vitro the mechanical challenges imposed to Pf-RBCs in vivo. We document in real-time the ultimate impact of RESA protein on Pf-RBCs dynamic response, quantitatively comparing transit cell velocities of individual Pf-RBCs, expressing or not expressing RESA, as they are forced to traverse successive constrictions under controlled pressure gradients. Moreover we study the effects of febrile temperature and expression of RESA on the circulatory behavior of ring-stage Pf-RBCs through a microspheres microsphiltration system20, shown to mimic the mechanical challenge of Pf-RBCs by the human spleen by imitating the geometry of narrow and short inter-endothelial slits of the spleen sinuses.

Results

Decreased membrane dynamics in ring-stage Pf-RBCs are RESA-driven

We first addressed the effect of RESA on membrane dynamics of ring-stage Pf-RBCs using Diffraction Phase Microscopy (DPM). RBC samples were prepared under three different conditions: parasite-free RBCs, wild-type resa1+ and resa1-KO ring-stage Pf-RBCs (see Materials and Methods). In order to compare the dynamic membrane fluctuations of the different types of RBC samples, we calculated the root mean square (RMS) displacement of membrane fluctuation, , which covers the entire cell area for 2 s at 120 frames/s (Figure 1A). The RMS displacement of membrane fluctuations for parasite-free RBCs was 61 ± 4.0 nm. Membrane fluctuations decreased significantly to 55.0 ± 4.3 nm in wild-type resa1+ Pf-RBCs. However, membrane fluctuations in resa1-KO Pf-RBCs showed higher values than wild-type resa1+ Pf-RBCs (62.0 ± 9.0 nm). No significant difference between membrane fluctuations in resa1-KO Pf-RBCs and parasite-free RBCs was observed.

Figure 1
figure 1

Membrane dynamics of ring-stage Pf-RBCs at physiological body temperature.

(A) RMS displacements and (B) in-plane shear modulus values of parasite-free RBC, wild-type resa1+ and resa1-KO ring-stage Pf-RBCs. Open circles are experimental values and represent individual cell measurements. Significant differences are shown as * (p < 10−4) and * (p < 10−5) values.

Full size image

We then retrieved the in-plane shear modulus μ, which determines the ability of RBCs to deform. The results for the shear modulus are shown in Figure 1B. For parasite-free RBCs, μ = 6.2 ± 1.4 μN/m. A significant increase in μ was observed for wild-type resa1+ Pf-RBCs (15.0 ± 3.8 μN/m), indicating a decrease in cell deformability. Interestingly a significant decrease in μ is observed for resa1-KO Pf-RBCs (7.6 ± 2.6 μN/m). These results are consistent with literature values of membrane shear modulus obtained using independent experimental methods under static conditions that involved optical tweezers tests on parasite-free RBCs6,7, wild-type resa1+ and resa1-KO ring-stage Pf-RBCs9 and previous work on 3D7 ring-stage wild-type Pf-RBCs using DPM measurements8.

Temperature enhances membrane dynamics of ring-stage Pf-RBCs

Membrane fluctuation values derived from the DPM measurements susbtantially increased from normal physiological to febrile temperature in parasite-free RBCs (Figure 2A), while the in-plane shear modulus values decreased (Figure 2D). In contrast, membrane fluctuations of resa1+ Pf-RBCs were essentially similar at 37°C and 41°C (Figure 2B), but the in-plane shear modulus showed a significant decrease in cell deformability from body to febrile temperature, although the scatter range was large (Figure 2E). Resa1-KO Pf-RBCs DPM measurements showed that membrane fluctuations values were somewhat higher at 41°C than those observed at 37°C (Figure 2C), although the increase was modest compared to parasite-free RBCs. The resa1-KO Pf-RBCs in-plane shear modulus showed no significant difference between normal and febrile temperature as observed for parasite-free RBCs (Figure 2F). Results on membrane shear modulus measurements obtained previously using optical tweezers test on parasite-free RBCs, wild-type resa1+ and resa1-KO ring-stage Pf-RBCs9 are shown for comparison (closed circles Figure 2 D–F).

Figure 2
figure 2

Membrane fluctuations and in-plane shear modulus of ring-stage Pf-RBCs at body and febrile temperature.

RMS displacement histogram and in-plane shear modulus values (upper and lower panel, respectively) of (A), (D). parasite-free RBCs, (B), (E). resa1+ and (C), (F). resa1-KO ring-stage Pf-RBCs, measured at 37°C and 41°C. Open circles represent individual RBC measurements. Closed circles are shear modulus results obtained with optical tweezers and the same set of parasites9 shown for comparaison purposes. Significant differences are shown as * (p < 0.02) values.

Full size image

Microcirculation of ring-stage Pf-RBCs is influenced by RESA

The effect of RESA expression on the dynamic properties of ring-stage Pf-RBCs was measured using a microfluidic assay quantifying the transit cell velocity of Pf-RBCs moving through channels with multiple successive constrictions under controlled pressure and thereby simulating host microcirculation in vitro. Figure 3 shows in real time a representative example of the dynamic response of parasite-free RBCs, resa1+ and resa1-KO ring-stage Pf-RBCs when they were forced to pass through multiple constrictions at normal physiological (37°C) temperature. At a constant gradient pressure of 0.24 Pa μm−1 ring-stage Pf-RBCs were able to deform and squeeze through micro-sized constrictions of 3 µm. However, when compared to the transit cell velocity of parasite-free RBCs appearing in each field of view, the transit cell velocity of individual resa1+ Pf-RBCs (Figure 3A and supplementary material Video S1) was significantly slower than resa1-KO ring-stage Pf-RBCs (Figure 3B and supplementary material Video S2). As an illustrative example of all the cell velocity measurements performed, the spatial position as a function of time of individual resa1+ and resa1-KO Pf-RBCs compared to parasite-free RBCs (calculated from Figure 3A and B) are represented in Figure 3C. The average transit cell velocity of resa1+ Pf-RBCs (14.7 µm/s) was slower than parasite-free RBCs from the same culture (26.8 µm/s, p < 0.001) (Figure 4A). However, resa1-KO Pf-RBCs and bystander parasite-free RBCs exhibited similar transit cell velocities (24.5 and 25.9 µm/s, respectively, with p = 0.186) (Figure 4B).

Figure 3
figure 3

Differences in the dynamic response of ring-stage Pf-RBCs expressing or not RESA.

Real-time snap-shots transit cell velocity of a representative example of all the experiments performed of (A) resa1+ and (B) resa1-KO ring-stage Pf-RBCs, forced to traverse through 3 μm successive constrictions in micro-sized channels at a constant pressure gradient of 0.24 Pa μm−1. Pf-RBCs are labelled fluorescently using Thiazole orange. Yellow arrows indicate individual parasite-free RBC. (C). Illustrative example of all the transit cell velocity measurements performed. Each open circle represents the spatial position (μm) as a function of time (s) of individual resa1+ (red) and resa1-KO (green) Pf-RBCs compared to parasite-free RBCs (grey) calculated from Figure 3A and B.

Full size image
Figure 4
figure 4

Transit cell velocity of ring-stage Pf-RBCs at body temperature.

Transit cell velocities of (A) resa1+, (B) resa1-KO and (C) resa1-rev ring-stage Pf-RBCs compared to corresponding co-cultured parasite-free RBCs, passing through 3 μm constrictions in micro-sized channels. Measurements were performed at 37°C at a constant pressure gradient of 0.24 Pa μm−1. Each open circle represents an individual cell measurement. Significant differences are shown as * (p < 10−9) values.

Full size image

To further substantiate that the observed difference in transit cell velocity between resa1-KO Pf-RBCs and resa1+ Pf-RBCs was specifically due to the absence of RESA protein expression, we tested the dynamic response of the revertant isogenic resa1-rev Pf-RBCs9. The resa1-rev Pf-RBCs displayed markedly decreased average transit cell velocity (16.8 µm/s) in the same range as the original resa1+ parental line and substantially lower than bystander parasite-free RBCs (28.1 µm/s) (p < 0.001) (Figure 4C).

The effect of RESA on Pf-RBCs microcirculation in vitro was also investigated at febrile temperature. Indeed, the transit cell velocity of wild-type resa1+ Pf-RBCs was reduced (Figure 5A) when microfluidics experiments were performed at 41°C (10.2 µm/s compared to 14.7 µm/s at 37°C, p < 0.01). Resa1-rev Pf-RBCs behaved similarly, with an average transit cell velocity of 13.2 µm/s at 41°C and 16.8 µm/s at 37°C (p < 0.01) (Figure 5C). However resa1-KO Pf-RBCs, which had similar transit cell velocity values at both 37°C and 41°C (24.5 µm/s and 22.6 µm/s, respectively, p = 0.1863) displayed a similar trend as the parasite-free RBCs from 37°C to 41°C (26.8 µm/s and 23.30 µm/s p < 0.01) (Figure 5B).

Figure 5
figure 5

Transit cell velocity of ring-stage Pf-RBCs at body and febrile temperature.

Transit cell velocities of parasite-free RBCs (black), resa1+ (red), resa1-KO (blue) and resa1-rev (green) ring-stage Pf-RBCs, measured at 37°C and 41°C. Measurements were performed at a constant pressure gradient of 0.24 Pa μm−1. Each open circle represents an individual cell measurement. Significant differences are shown as * (p < 0.01) values.

Full size image

Temperature and RESA expression influence filterability of ring-stage Pf-RBC in microspheres

The effect of RESA on the circulatory behavior of ring-stage Pf-RBCs was assessed using microspheres, a microspheres filtration system simulating the narrow and short geometry of inter-endothelial slits of the spleen sinuses20. Highly synchronized cultures (ring stage at 15-18 h) pre-incubated for 3 h at 37°C or 40°C were loaded on the calibrated microspheres columns. Traversal of Pf-RBCs was quantified by calculating the retention rate in the microspheres. Retention of resa1-KO and resa1-rev Pf-RBCs did not differ significantly in the culture maintained at 37°C (Figure 6A). However, it was increased by 30-60% in resa1-rev Pf-RBCs pre-incubated for 3 h at 40°C and not in resa1-KO Pf-RBCs (Figure 6A). To rule out that increased retention of resa1-rev parasites was due to a more advanced parasite developmental stage, possibly more rapid development during the 3 h exposure at 40°C, the duration of the temperature shift was reduced to 2 h and synchronous parasites at different stage of development were used. Temperature-shifted 12–15 h old or 15–18 h old resa1-rev ring stage Pf-RBCs displayed a higher retention rate compared to the sibling culture maintained at 37°C. Such a temperature-induced reduced filterability was not observed for the resa1-KO Pf-RBCs cultures (Figure 6B). This indicated that RESA expression influenced microsphere filterability of fever-exposed ring-stages. At a later developmental stage, as anticipated, filterability was further reduced and was RESA-independent, as Pf-RBCs harbored mature, poorly deformable parasites.

Figure 6
figure 6

Incubation at febrile temperature increases retention rate by microspheres of RESA expressing Pf-RBCs (resa1-rev) and not of resa1-KO ring-stage Pf-RBCs.

(A) Synchronized cultures (15–18 h rings) of resa1-rev (red) and resa1-KO (blue) Pf-RBCs adjusted to 2% hematocrit and 10% parasitemia, were incubated at either at 37°C or 40°C for 3 h. Error bars represent the standard deviation. (B) Highly synchronized parasite cultures (12–15 h, 15–18 h, 18–21 h and 21–24 h after invasion) of resa1-rev (red) and resa1-KO (blue) Pf-RBCs adjusted to 2% hematocrit and 5% parasitemia, were incubated at 40°C for 2 h. Error bars represent the standard deviation. The p values for resa1-rev/resa1-KO are 0.07 for 12–15 h; 0.00066 for 15–18 h; 0.397 for 18–21 h and 0.641 for 21–24 h.

Full size image

Discussion

The results presented here demonstrate that expression of RESA protein in the ring stage of intra-erythrocytic development of P. falciparum has a significant effect on dynamic biophysical properties and microcirculatory response of Pf-RBCs. They also provide clear evidence that ring-stage Pf-RBCs markedly differ in their microcirculatory behavior from bystander parasite-free RBCs and from ring-stage Pf-RBCs devoid of RESA. Febrile temperature exacerbates the role of RESA in influencing the dynamic microcirculatory properties of Pf-RBCs.

Nanoscale fluctuations (commonly referred to as “flickering”) of the cell membrane at 37°C are markedly decreased by RESA expression in the ring stage. This is consistent with a reported reduction in thermally driven membrane flickering over the full range of (wild-type) infected stages of Pf-RBCs8. It had been speculated that parasite modifications to the host RBC cytoskeleton are responsible for changes in the Pf-RBCs membrane flickering profile. The membrane fluctuation values at 37°C on resa1-KO Pf-RBCs and parasite-free RBCs were comparable, indicating that the presence of RESA on host membrane, or downstream effects of its expression, are involved in the membrane flickering changes observed in young stages at normal body temperature. The in-plane shear modulus data retrieved from the measured membrane fluctuations values indicate increased rigidity for the Pf-RBCs expressing RESA but not for resa1-KO Pf-RBCs. These results are in agreement with optical tweezers measurements of quasi-static deformation9 and confirm that RESA protein modulates cell deformability of ring-stage Pf-RBCs.

We found febrile temperature to significantly influence the membrane dynamic properties of parasite-free RBCs present in the same culture as ring-stage Pf-RBCs. The membrane fluctuation of parasite-free RBCs increased from normal physiological (37°C) to febrile (41°C) temperature while the in-plane shear modulus decreased. This latter observation is in line with the reported decrease by ≈ 20% of the shear modulus of healthy RBC membrane when the temperature increased from 23 to 41°C21. This indicates an increase in the overall parasite-free RBCs deformability at febrile temperature, possibly reflecting structural changes of the RBC membrane phospholipid and/or spectrin network that alter its elastic properties8. One possible contributor is the transitional structural change in α- and β-spectrin molecules near 40°C22. Interestingly, different observations were made with ring-stage resa1+ Pf-RBCs. Since the in-plane shear modulus of RBCs in our results is calculated from the tangential component of displacement in membrane fluctuations23, both the measured axial membrane fluctuation and morphology of RBCs determined shear modulus value. No significant changes in membrane fluctuations were observed upon temperature shift in resa1+ Pf-RBCs while the in-plane shear modulus values increased, i.e. both parameters showed opposite trends compared to bystander parasite-free RBCs. Since RESA interaction with spectrin stabilizes spectrin’s tetrameric state in vitro14, it probably prevents the dissociation of spectrin tetramers at 41°C, resulting in overall decreased Pf-RBCs deformability and increased membrane mechanical stability14. This inference is also consistent with the earlier finding that RESA prevents membrane vesiculation at 50°C15,16. The RESA-dependent temperature enhancement of ring-stage Pf-RBCs stiffness observed here is consistent with previous static deformability measurements obtained using optical tweezers9. At 41°C, a modest increase in membrane fluctuations was recorded in resa1-KO Pf-RBCs compared to that observed at 37°C. This indicates that in addition to RESA expression, possibly other parasite-associated factors contribute to modulating ring-stage Pf-RBCs membrane fluctuations at 41°C. Possible parasite factors influencing membrane fluctuations changes at 41°C are additional parasite-encoded proteins and/or apical organelle-associated proteins discharged in the erythrocyte-membrane during invasion24,25,26. Alternative RESA-independent modifications are proteolysis of band3 and/or of erythrocyte cytoskeletal proteins needed to ensure invasion27,28,29. Although Pf-RBCs lacking RESA showed increased membrane fluctuations values at 41°C, the in-plane shear modulus remained essentially unaffected at 37°C and 41°C. The data confirm RESA as the main molecule involved in Pf-RBCs deformability changes observed at elevated temperature and support the role played by the specific RESA-spectrin interaction at 41°C.

Analysis of the dynamic properties of Pf-RBCs using a microfluidic device that replicates successive constriction obstacles encountered in the vascular microcirculation was done at physiological and febrile temperatures. Transit cell velocity of ring-stage resa1+ Pf-RBCs was markedly decreased compared to bystander parasite-free RBCs and was further reduced at febrile temperature. Although resa1-KO Pf-RBCs travelled slightly slower than the co-cultured parasite-free RBCs, this was not statistically significant and moreover transit cell velocity was independent of the temperature at which the experiments were performed. The similarly higher transit cell velocities of resa1-KO Pf-RBCs and parasite-free RBCs indicate that RESA is largely responsible for the reduced transit cell velocity of ring-stage Pf-RBCs. This data identifies for the first time RESA as a parasite factor is able to modulate the dynamic response of ring-stage Pf-RBCs during microcirculation.

In addition to vascular microcirculation, another anatomical site where RBC dynamic properties are stringently challenged is the red pulp of the human spleen where the RBCs must undergo extensive deformation as they cross the narrow inter-endothelial slits of the sinuses. In an ex vivo organ perfusion system, a significant proportion of ring-stage Pf-RBCs was retained in the red pulp of the human spleen, most probably due to their altered mechanical properties30. The bead microsphiltration system we used here mimics the mechanical sensing of RBCs by the human spleen, as it was shown to retain ring-stage Pf-RBCs at a rate similar to that observed in an isolated-perfused human spleen20. This system does not permit assessment of the behavior of bystander parasite-free RBCs, but allows us to monitor the filterability of Pf-RBCs and hence the impact of the presence of RESA and of a temperature shift on efficiency of traversal of the microsphere layer. Filterability of resa1-rev and resa1-KO Pf-RBCs was similar at 37°C, but a transient incubation at 40°C resulted in increased retention of the RESA expressing Pf-RBCs and not of the resa1-KO Pf-RBCs. Such a temperature-associated decreased filterability of RESA-expressing parasites was observed for late ring-stage Pf-RBCs and was not observed in mature stages, which no longer express RESA. The late ring stage parasites studied here were harvested before surface expression of the P. falciparum erythrocyte membrane protein 1 (PfEMP1)31. Furthermore no difference in retention with the bead system could be observed over the first 20 hours post invasion between K+ or K− parasites32 excluding a possible effect of knobs in bead retention. Thus, the microsphiltration results show that RESA expression clearly modulates the circulatory behavior of Pf-RBCs at febrile temperature.

Our data present the first detailed evidence for a major role of temperature and presence of RESA in modulating the dynamic microcirculation characteristics of ring-stage Pf-RBCs. This conclusion, stemming from the analysis of dynamic behavior of Pf-RBCs using complementary approaches that impose distinct geometric constraints on dynamic cell deformabilty, is consistent with observations of the static behavior of ring Pf-RBCs, using optical tweezers9 and biophysical measurements4,7. The consistency of data obtained on the static and dynamic behaviour on the one hand and biophysical measurements on the other is interesting as it cross-validates the various, complementary approaches. Moreover, data presented here as well as previous studies4,7,8,9 point to RESA as a major determinant of the temperature-enhanced deformability defect of ring-stage Pf-RBCs.

RESA may play a bifunctional role on influencing microcirculation of ring-stage Pf-RBCs during an in vivo infection. While its expression is essential for the young ring stage parasites to resist a transient exposure to febrile temperature and ensure their normal intra-erythrocytic cycle and parasite development9, upon exposure of febrile temperatures can possibly lead to a splenic retention and accelerated parasite clearance.

The present findings open new avenues for intervention to reduce parasite survival and pathogenesis in its human host.

Methods

Parasites and culture

The P. falciparum FUP/CB line (referred to as wild-type resa1+) and its derived resa1 gene knock-out P. falciparum clone (resa1-KO) and resa1-revertant clone (resa1-rev)9 were cultured in leukocyte-free human RBCs (Research Blood Components, Brighton, MA) under an atmosphere of 5% O2, 5% CO2 and 95% N2, at 5% haematocrit in RPMI culture medium 1640 (Gibco Life Technologies, Rockville, MD) supplemented with 25 mM HEPES (Sigma), 200 mM Hypoxanthine (Sigma, St. Louis, MO), 0.20% NaHCO3 (Sigma, St. Louis, MO) and 0.25% Albumax I (Gibco Life Technologies, Rockville, MD). Parasite cultures were routinely synchronized in ring stage by using Sorbitol lysis 2 h after merozoite invasion33. Prior to measurements the parasite cultures were enriched in late trophozoite stages at 96% purity using a Midi MACS LS magnetic column (Miltenyi Biotech, Auburn, CA) and diluted into RBCs obtained the same day, to harvest in the next cycle highly synchronous ring stages in fresh RBCs.

Diffraction phase microscopy

RBCs and Pf-RBCs samples were centrifuged at 125 g and diluted in Phosphate buffered saline (PBS), 1% Bovine Serum Albumin (BSA) (Sigma-Aldrich, St Louis, MO) to approximately 1×106 RBCs/mL. A 10 μL RBC suspension was introduced between two glass slides and the dynamic membrane fluctuations were measured using Diffraction Phase Microscopy (DPM)8. DPM is a highly stable and sensitive quantitative phase microscopy, which employs the laser interferometry in a common path geometry and thus provides full-field quantitative phase images of biological samples with unprecedented optical path length stability17,18. An Ar2+ laser (λ = 514 nm, Coherent Inc., Santa Clara, CA) was used as illumination source. An inverted microscope (IX71, Olympus American Inc., Center Valley, PA) was equipped with a 40x objective (0.65 NA), which facilitates a diffraction-limited transverse resolution of 400 nm. With the additional relay optics, the overall magnification of the system was 200×. EMCCD (Photonmax512B, Princeton Instruments, Trenton, NJ) was used to record interferograms. The instantaneous cell thickness map is obtained as h(x,y,t) = λ/(2πΔn)·φ(x,y,t), with φ the quantitative phase image measured by DPM. The refractive index contrast Δn between the RBC and the surrounding medium is mainly attributed to hemoglobin in RBC cytosol. The values for Δn were used from a previous study8. The DPM optical path length stability is 2.4 mrad, which corresponds to a membrane displacement of 3.3 nm in RBC membrane18.

By retrieving the optical path length shifts produced across the cell, the cell thickness profiles at a given time t, h(x,y,t), were obtained. The dynamic membrane fluctuation can then be calculated by subtracting the averaged cell shape from the instantaneous cell thickness map, , where the brackets represent the average over time.

The in-plane shear modulus of the RBC membrane, μ, can be retrieved from the measured membrane fluctuations8. We calculated μ by using Fourier-transformed Hamiltonian (strain energy) and equi-partition theorem23:

Where A is the diameter of RBC and a is the minimum spatial wavelength measured by DPM (400 nm). The tangential component of displacement in membrane fluctuations was decoupled from the measured axial membrane fluctuations and the normal diffraction of the membrane, which can be calculated from the mean cellular thickness map measured by DPM8.

The microscopic set up was equipped with a temperature controller (TC-202A, Warner Instruments, Hamden, CT), which uses a thermistor to set the temperature of the sample to within ±0.2°C. The well containing RBCs was placed in contact with the controller chamber and heat transfer and thermal equilibrium between the two systems were attained relatively fast, after 3–4 min. We measured the individual RBC response at different temperature points after a ~10 min incubation for each new temperature to ensure thermal equilibrium. DPM measurements were performed at 37°C and 41°C, on ring-stage parasites (14–18 h post-invasion).

Microfluidic channel fabrication

A microfluidic device was made of polydimethylsiloxane (PDMS) using photolithography and reactive-ion etching (RIE) techniques as described elsewhere19. The microfluidic device pattern was designed specifically for the capillary channels to test optimum RBC deformation. Briefly, the microfluidic device consisted of multiple parallel capillary channels with triangular pillars arrays, producing successive micro-sized constrictions of 3.0 μm between pillars, with a depth of 4.2 μm19. The inlet and the outlet reservoir have the same dimensions of 500 x 500 μm2.

Microfluidic measurements

RBCs and Pf-RBCs were centrifuged at 300 g and further diluted in PBS-1% BSA to a 0.5% hematocrit. Thiazole orange (0.05 μM) (Invitrogen, Carlsbad, CA) was added to the suspension and samples were incubated at room temperature for 20 min in the dark. Then, a 10 μL RBCs suspension was flowed through the microfluidic device previously coated with PBS-1%BSA, at a constant pressure gradient of 0.24 Pa μm−1 19. All experiments were imaged using an inverted fluorescent microscope (Olympus IX71, Center Valley, PA) equipped with a 60x objective and connected to a CCD camera (Hamamatsu Photonics, C4742-80-12AG, Japan). Images of individual RBC were acquired automatically using IPLab (Scanalytics, Rockville, MD) at 70 ms time interval using a green illumination (545 nm). Post-imaging analysis was done using imageJ software (National Institutes of Health).

The temperature was controlled using a heating chamber (Olympus, Center Valley, PA), which was preheated 30 min before the beginning of the experiment. Then a PBS-1%BSA coated microfluidic device was placed into the heating chamber 5 min before loading the RBCs suspension. A thermal meter was used to probe the exact temperature inside the heating chamber and 5 min were required for the temperature to be adjusted to a different value.

Transit cell velocities of individual RBCs and Pf-RBCs passing through microchannels were analyzed. Microfluidic measurements were performed at 37°C and 41°C, on ring stage-parasites (14–18 h post invasion).

Microsphiltration

Microspheres sphiltration was performed using methods previously established20. Briefly, 2 g of dry calibrated metal microspheres (96.50% tin, 3.00% silver and 0.50% copper; Industrie des Poudres Sphériques, France) with 2 different size distributions (5- to 15-μm-diameter and 15- to 25-μm-diameter) were mixed and then suspended in 8 mL of RPMI 1640/10% human serum. Nine hundred µL of the bead suspension was poured into an inverted 1000 µL anti-aerosol pipette tip (Neptune, BarrierTips) and allowed to settle, leading to the formation of a 5 mm-thick bead layer above the anti-aerosol filter. 600 µL of a 2% hematocrit RBC suspension containing less than 10% of potentially “retainable” RBC (to avoid bead saturation) was introduced upstream from the micro-bead layer. RBCs were flushed through the bead layer at a flow rate of 60 mL/h using an electric pump (Syramed µsp6000, ARCOMED AG, Regensdorf, Switzerland). The bead layer was then washed with 6 mL RPMI 1640/10% human serum. The downstream sample was retrieved. The percentage of ring-stage Pf-RBCs in upstream and downstream RBC samples was determined by examination of Giemsa-stained blood films. Experiments were performed on highly synchronized parasites (12–15 h, 15–18 h, 18–21 h and 21–24 h old rings) after incubation at 37°C or 40°C.

Statistical analysis

We calculated p values by two-tailed Mann–Whitney rank sum tests comparing the membrane fluctuations and shear moduli values between various test conditions. We used the student-T-test comparing the microspheres filtration results. All of the numbers following the ± sign in the text are standard deviations.