Equally contributing authors.
Haemoglobin S and C affect the motion of Maurer's clefts in Plasmodium falciparum-infected erythrocytes
Article first published online: 20 JAN 2013
© 2012 John Wiley & Sons Ltd
Volume 15, Issue 7, pages 1111–1126, July 2013
How to Cite
Kilian, N., Dittmer, M., Cyrklaff, M., Ouermi, D., Bisseye, C., Simpore, J., Frischknecht, F., Sanchez, C. P. and Lanzer, M. (2013), Haemoglobin S and C affect the motion of Maurer's clefts in Plasmodium falciparum-infected erythrocytes. Cellular Microbiology, 15: 1111–1126. doi: 10.1111/cmi.12102
- Issue published online: 13 JUN 2013
- Article first published online: 20 JAN 2013
- Accepted manuscript online: 26 DEC 2012 07:25AM EST
- Manuscript Accepted: 14 DEC 2012
- Manuscript Revised: 30 NOV 2012
- Manuscript Received: 14 SEP 2012
- Deutsche Forschungsgemeinschaft
- Chica and Heinz Schaller foundation
- Top of page
- Experimental procedures
- Supporting Information
The haemoglobinopathies S and C protect carriers from severe Plasmodium falciparum malaria. We have recently shown that haemoglobin S and C interfere with host-actin remodelling in parasitized erythrocytes and the generation of an actin network that seems to be required for vesicular protein trafficking from the Maurer's clefts (a parasite-derived intermediary protein secretory organelle) to the erythrocyte surface. Here we show that the actin network exerts skeletal functions by anchoring the Maurer's clefts within the erythrocyte cytoplasm. Using a customized tracking tool to investigate the motion of single Maurer's clefts, we found that a functional actin network restrains Brownian motion of this organelle. Maurer's clefts moved significantly faster in wild-type erythrocytes treated with the actin depolymerizing agent cytochalasin D and in erythrocytes containing the haemoglobin variants S and C. Our data support the model of an impaired actin network being an underpinning cause of cellular malfunctioning in parasitized erythrocytes containing haemoglobin S or C, and, possibly, for the protective role of these haemoglobin variants against severe malaria.
- Top of page
- Experimental procedures
- Supporting Information
The global burden of malaria has remained high – currently standing at approximately 216 million disease episodes and an estimated 0.655 million deaths, annually (World Health Organization, 2011). Of the five species that can cause malaria in humans Plasmodium falciparum is the most virulent and much of the malaria-related morbidity and mortality is due to infections by this protozoan parasite. The virulence of P. falciparum is associated with the intraerythrocytic life cycle of the parasite and has been attributed to the cytoadhesive properties of parasitized erythrocytes (Miller et al., 2002). During intraerythrocytic development, the parasite inserts immunovariant adhesins in knob-like protrusions in the erythrocyte plasma membrane (Flick and Chen, 2004). These adhesins mediate a broad range of adhesive interactions by binding to receptors on the surface of vascular endothelial cells, uninfected erythrocytes, platelets and monocytes (Flick and Chen, 2004). Due to their cytoadhesiveness, infected erythrocytes sequester in the deep vascular bed of inner organs, such as the brain, resulting in perturbed tissue perfusion, hypoxia and microvascular inflammation that can progress to life-threatening complications (Fairhurst et al., 2012).
Several polymorphisms have emerged in the human genome in response to the strong selective force exerted by P. falciparum on the human population, which protect carriers from severe malaria (Kwiatkowski, 2005). Prominent survival traits are the structural haemoglobinopathies S and C (Modiano et al., 2001; Aidoo et al., 2002; Taylor et al., 2012). Normal haemoglobin (HbAA) is a heterotetramer consisting of two α- and two β-globin chains. In haemoglobin S and C, the β-globin chain is altered at position 6 by a substitution of glutamic acid to valine and lysine respectively. Heterozygotes (HbAS and HbAC) have a significant survival advantage in P. falciparum infections while only suffering from mild to asymptomatic anaemia (Taylor et al., 2012). In contrast, individuals who are homozygote for HbS manifest the sickle cell disease and frequently die during early adolescence (Hebbel, 1991).
A recent study has suggested that HbS and HbC protect from severe malaria by interfering with parasite-induced host actin remodelling (Cyrklaff et al., 2011; 2012). Both HbS and HbC are unstable and react with oxygen (Bauminger et al., 1979; Chaves et al., 2008). Oxidized haemoglobins can impact on membrane cytoskeletal functions and actin dynamics (Jarolim et al., 1990; Hebbel, 1991; Tokumasu et al., 2005; Farah et al., 2011) and are thought to prevent the parasite from creating host-derived elongated actin filaments required for protein trafficking processes within the infected erythrocyte (Cyrklaff et al., 2011).
The actin filaments seem to serve as cables for vesicular protein trafficking to the erythrocyte surface. They connect the knobs with the Maurer's clefts (Cyrklaff et al., 2011; 2012), unilamellar membrane profiles within the cytoplasm of the infected erythrocyte that functions as an intermediary compartment for parasite-encoded proteins en route to the erythrocyte surface (Lanzer et al., 2006). Furthermore, the actin filaments seem to maintain the morphology of Maurer's clefts (Cyrklaff et al., 2011). Ineffective export of parasite-encoded proteins to the erythrocyte surface and aberrant anchoring of adhesins and components of the knobs to the parasite-induced actin network may explain the finding that parasitized erythrocytes containing HbS and HbC have fewer but unnaturally enlarged knobs, that they display diminished amounts of adhesins on their surface, that the adhesins that are displayed are aberrantly presented and that the ability to cytoadhere is reduced (Fairhurst et al., 2003; 2005; 2012; Cholera et al., 2008).
The host actin remodelling commences in the ring stage and progresses towards an elaborated actin network in the trophozoite stage (Cyrklaff et al., 2011; 2012). The completion of the actin network coincides with a drastic change in Maurer's clefts dynamics (Gruring et al., 2011). Maurer's clefts are highly motile in ring stages, whereas apparently immobile during the trophozoite stage (Gruring et al., 2011). The coincidence of the two events poses the question of whether the actin filaments tether the Maurer's clefts to the knobs and ultimately to the membrane skeleton of the erythrocyte, thereby arresting them. Since the actin network does not fully develop in parasitized erythrocytes containing HbS or HbC (Cyrklaff et al., 2011), this may affect the temporal changes in the motion of the Maurer's clefts during the intraerythrocytic developmental cycle. To address these questions and to better understand the function of the parasite-induced actin network we have investigated the motion of Maurer's clefts in infected erythrocytes containing different haemoglobin variants.
- Top of page
- Experimental procedures
- Supporting Information
Particle tracking tool to quantify Maurer's clefts motion
When trying to track individual Maurer's clefts within live P. falciparum-infected erythrocytes over time we encountered several problems that complicated a quantitative analysis. Frequently the infected erythrocyte changed its position and/or rotated during the measurement. To solve these problems, we wrote a script based on the software R that allowed us to semi-automatically track the motion of individual Maurer's clefts, follow their movement over time and compensate for translational and rotational movement of the infected cell. To validate the script we simulated a situation in which a green dot (imitating a Maurer's cleft) moved within a larger, also moving round object (imitating an erythrocyte) (Fig. 1A). The green dot moved by 35 pixels per frame over six continuous frames. The script correctly recognized the motion of the larger object, then compensated for this movement, and correctly calculated the total distance and the distance per frame covered by the green dot (Fig. 1A). We repeated the simulation but this time the larger object also rotated, in addition, to translational movement. The green dot did not move within the object (Fig. 1B). Again, the software correctly compensated for the movements of the larger object and correctly predicted that the green dot rested within the object.
We next applied the motion tracking tool to wild-type erythrocytes infected with a mutant of the P. falciparum clone 3D7 expressing a genomically integrated pfsbp1 gene juxtaposed at its 3′ terminus to the open reading frame of the green fluorescence protein (Saridaki et al., 2009). The gene product PfSBP1 is an established Maurer's clefts marker and the parasites expressing the PfSBP1/GFP fusion protein displayed fluorescently labelled Maurer's clefts (Blisnick et al., 2000; Saridaki et al., 2009). Figure 2A depicts a time series over 40 s. Note that the parasitized erythrocyte displayed both rotational and translational movement, which the tracking tool correctly recognized (Fig. 2B and C). The variance of the motion tracking procedure in quantifying the movement of a single Maurer's clefts was less than 0.3% as determined by having three different operators measure the same cell 10 times each and then calculating the mean and the standard error from these values (Fig. 2B).
Differential motion of Maurer's clefts in erythrocytes containing mutant haemoglobins
We next investigated the effect of haemoglobin S and C on Maurer's clefts movement. To this end, erythrocytes containing the haemoglobin variants HbAA, HbAC, HbAS, HbCC and HbSC were infected with the parasite clone expressing the PfSBP1/GFP fusion. The erythrocytes used for the infection study had mean disk diameters and mean corpuscular haemoglobin concentrations (MCHC) that were in the normal value ranges (Table 1). The mean corpuscular volumes (MCV) of the uninfected HbAA, HbAC and HbAS red blood cells were also in the normal range, whereas slightly lower MCV values were found for the HbCC and HbSC erythrocytes (Table 1), consistent with previous reports and indicative of the presence of hyperchromatic cells (Lew et al., 1995; Fairhurst and Casella, 2004).
|Erythrocyte variant||Disk diametera (μm)||MCVb (fl)||MCHCc (g dl−1)|
|HbAA||8.3 ± 1.2||85.1 ± 1.4||32.5 ± 1.2|
|HbAC||7.5 ± 1.3||85.7 ± 1.4||33.3 ± 1.3|
|HbAS||7.6 ± 1.3||82.4 ± 1.3||35.7 ± 1.4|
|HbCC||7.9 ± 1.6||77.5 ± 1.2||34.2 ± 1.3|
|HbSC||7.8 ± 1.3||76.7 ± 1.2||34.8 ± 1.3|
Parasites developed normally in the different erythrocyte variants under continuous in vitro culture conditions, although there was a statistically not significant trend of a slightly lower replication rate when parasites were grown in erythrocytes containing haemoglobin variants (Table 2, P > 0.05 compared with HbAA infected erythrocytes). The parasite maintained a knobby phenotype in the different erythrocyte variants, as shown by transmission electron microscopy of thin sections (Fig. S1A–C). Consistent with comparable replication rates, the area occupied by the parasites within the infected erythrocytes did not significantly differ between the different erythrocyte variants (P > 0.05 compared with HbAA infected erythrocytes) and increased from approximately 15 ± 3% in the ring stage (8–12 h post invasion) (Fig. 3A) to approximately 22 ± 3% in the trophozoite stage (20−26 h post invasion) (Fig. 3B). There were further no significant differences in the average number of Maurer's clefts per infected erythrocyte between the different erythrocyte variants (Fig. 3C and D). Consistent with previous reports, the number of detectable Maurer's clefts remained stable as the parasite developed from the ring to the trophozoite stage (Fig. 3C and D) (Gruring et al., 2011), suggesting that Maurer's clefts form early during parasite development. The average size of a Maurer's cleft, as expressed by the fluorescence coverage, varied from 100 to 175 pixels in rings and from 100 to 155 pixels in trophozoites, depending on the erythrocyte variant (Fig. 3E and F).
|Erythrocyte variant||Replication rate per 48 cycles||P-value|
|HbAA||5.6 ± 0.4 (24)||–|
|HbAC||4.6 ± 1.0 (18)||0.401|
|HbAS||4.0 ± 0.3 (18)||0.073|
|HbCC||4.0 ± 0.3 (23)||0.058|
|HbSC||5.1 ± 0.3 (24)||1.00|
To quantify the motion of Maurer's cleft in the different erythrocyte variants, we investigated, for each erythrocyte type, 10 or more independent cells and a total of at least 40 Maurer's clefts and determined the average speed at which the Maurer's clefts moved within the respective host cells. In each case, Maurer's clefts moved randomly within the host cell's cytoplasm (Fig. 4A), with an average speed ranging from 0.19 ± 0.02 μm s–1 to 0.14 ± 0.01 μm s–1 depending on the erythrocyte variant (Fig. 5A). The highest speed was observed in infected HbAA erythrocytes, the lowest in infected HbCC erythrocytes (Fig. 5A; P < 0.01). The other values were not significantly different from that in infected HbAA erythrocytes. Pre-incubating the cells with the actin depolymerizing agent cytochalasin D (10 μM) or the actin-stabilizing drug jasplakinolide (200 nM) for 10 min had no effect on Maurer's clefts movement during the ring stage, as exemplified by infected HbAA, HbCC and HbSC erythrocytes (Fig. 5B). This finding may suggest that motion of Maurer's clefts during the ring stage is an actin independent event.
During the trophozoite stage, the motion of Maurer's clefts slowed down (Figs 4B and 5C). However, movement did not completely cease, and there were significant differences in the speed of the Maurer's clefts between wild-type HbAA erythrocytes and erythrocytes containing haemoglobin mutants (P < 0.001; Fig. 5C). Maurer's clefts moved at an average speed of 0.034 ± 0.004 μm s–1 in HbAA infected erythrocytes (Fig. 5C). In comparison, Maurer's clefts moved significantly faster (twofold compared with HbAA red blood cells) in HbAC, HbAS, HbCC and HbSC erythrocytes, with a common average speed of 0.071 ± 0.004 (P < 0.001, Fig. 5C, see also Fig. S2 for a comparative analysis of the speed distributions).
Treating HbAA trophozoite-infected erythrocytes with cytochalasin D (10 μM) for 10 min significantly increased the movement of the Maurer's clefts to the speed observed for erythrocytes containing mutant haemoglobins (P < 0.001; Figs 4C and 5D). In comparison, cytochalasin D treatment did not affect the speed at which Maurer's clefts moved in HbSC erythrocytes. We further treated infected HbAA and HbSC erythrocytes with jasplakinolide (200 nM) but observed no effects on Maurer's clefts movement (Figs 4C and 5D).
Maurer's clefts move by Brownian motion
Since the motion of Maurer's clefts resembled a random walk (Fig. 4), we wondered whether it is due to Brownian motion. To approach this question we assessed the size and shape of Maurer's clefts. Maurer's clefts have been described as long slender or oblate membrane profiles depending on the viewing angle (Atkinson and Aikawa, 1990; Przyborski et al., 2003; Wickert et al., 2003; 2004; Lanzer et al., 2006; Hanssen et al., 2008; Cyrklaff et al., 2011). Investigation of published electron microscopic and cryo-electron tomographic images revealed that Maurer's clefts have length to width ratios of 10–20, with an average thickness of 30 ± 3 nm and an average length of 390 ± 40 nm (n = 27) in trophozoites (Wickert et al., 2003; 2004; Henrich et al., 2009; Cyrklaff et al., 2011). No such structures were found in ring stages. Instead there were large spherical membrane profiles (Fig. 6 and Fig. S1D–F), with an average diameter of 440 ± 40 nm (n = 21; radius of 220 ± 20 nm; average length to width ratio of 1.0–1.2) that are thought to be precursors of Maurer's clefts (Bannister et al., 2004; Cyrklaff et al., 2011). These spherical membrane profiles were surrounded by a dense protein coat (Fig. 6B and Fig. S1E) and they reacted with by an antisera specific to the Maurer's clefts marker PfSBP1 (Fig. 6C), which provides direct experimental support that they are indeed Maurer's clefts.
Entering the radius of 220 nm into the Stokes–Einstein equation revealed a diffusion constant of such a spherical particle of 6.9 × 10−15 m2 s–1, assuming a temperature of 37°C and a viscosity of 0.15 Pa·s for the erythrocyte cytoplasm (Williams and Morris, 1980). On the basis of the diffusion constant we calculated the total distance the particle would travel in 40 s and from these parameters the speed. One thousand repetitions of this simulation revealed that the particle would move at an average speed of 0.14 μm s–1 (Fig. S3). This value is in good agreement with the experimentally determined average speed of Maurer's clefts during ring stages (0.14–0.19 μm s–1 depending on the erythrocyte mutant; Fig. 5A).
To further validate the model of Maurer's clefts moving by Brownian motion, we recorded the movement of single Maurer's clefts in ring-stage infected wild-type erythrocytes over an extended period of 8.2 min (Fig. 7A). The Maurer's clefts covered almost half of the erythrocyte during that time, as seen in a total projection (Fig. 7B), exactly as one would predict for a Brownian particle (Fig. 7C).
We next simulated Maurer's clefts motion in trophozoites, using the Stokes–Einstein equation, but this time a Perrin factor of 1.7 was assumed to correct for the altered shape of Maurer's clefts in later stages. The Perrin factor was taken from the literature (Cantor and Schimmel, 1980) and is based on the ratio of the vertical versus the horizontal axis of Maurer's clefts and by assuming an oblate ellipsoid shape of Maurer's clefts, consistent with published electron microscopic images (Atkinson and Aikawa, 1990; Przyborski et al., 2003; Wickert et al., 2003; 2004; Lanzer et al., 2006; Hanssen et al., 2008; Cyrklaff et al., 2011). An average speed of 0.09 μm s–1 (diffusion constant of 3.2 × 10−15 m2 s–1) was calculated. This value is close to the experimental data, this being 0.07 ± 0.01 μm s–1, the speed at which Maurer's clefts moved in trophozoite-infected erythrocytes in the presence of cytochalasin D (Fig. 5C). As previously shown cytochalasin D destroys the actin network (Cyrklaff et al., 2011), thereby eliminating the contribution of the actin network in restraining Brownian motion of Maurer's clefts. These data are consistent with Brownian motion being the principal driver of the motion of Maurer's clefts.
Distinct patterns between parasitized erythrocytes containing haemoglobin variants
We repeatedly noted small, but not necessarily statistically significant, differences in several of the parameters measured. To explore the possibility that there, nevertheless, is an underlying but hidden structure in these data we conducted a principal component analysis. And indeed, patterns emerged for both ring- and trophozoite-infected erythrocytes (Fig. 8). In the case of the ring stages, parasitized erythrocytes that were heterozygous for HbA (HbAS and HbAC) clustered, indicating that they behaved very similarly with regard to the parameters examined in this study (Fig. 8A). Very different from infected HbAS and HbAC erythrocytes were the other three erythrocytes variants investigated, with infected HbSC erythrocytes being strongly inversely correlated to the HbAS and HbAC cluster. Also quite distinct were infected HbAA and HbCC erythrocytes, which formed independent groupings and which were strongly inversely correlated to each other (Fig. 8A). Note that the first two principal components sufficed to explain 78% of the variance contained in the data sets (Fig. 8A).
The principal component analysis on trophozoite-infected erythrocytes also revealed an underlying structure consisting of three distinct groupings (Fig. 8B). All erythrocyte variants containing haemoglobin C (HbAC, HbSC, HbCC) were very closely grouped together, mainly due to their similar parasite size and comparable fluorescence coverage and speed of their Maurer's clefts (Fig. 8B). It is further evident from the analysis that infected wild-type erythrocytes were strongly inversely correlated to all mutant erythrocytes, with the average speed of the Maurer's clefts being a major, but not the only, discriminator (Fig. 8B). The strongest inverse correlation existed between infected HbAA and HbAS infected erythrocytes. The principal component analysis represented the data sets quite well, with the first two principal components accounting for 97% of the information available for trophozoite stages (Fig. 8B). Hierarchical cluster analyses based on the co-ordinates of the data points in Fig. 8A and B confirmed the impression that infected HbAS and HbAC erythrocytes form a closely related group in ring stages (Fig. 9A). Similarly, the clustering of infected HbAC, HbSC and HbCC erythrocytes at the trophozoite stage was confirmed (Fig. 9B).
- Top of page
- Experimental procedures
- Supporting Information
Maurer's clefts are a novel type of organelle that, established by P. falciparum within the cytoplasm of its host erythrocyte, serve as an intermediary trafficking compartment for parasite-encoded proteins allotted to the erythrocyte plasma membrane (Lanzer et al., 2006). Maurer's clefts seem to originate from the parasitophorous vacuolar membrane and are present throughout the intraerythrocytic life cycle (Wickert et al., 2003; 2004; Hanssen et al., 2008; Henrich et al., 2009; Gruring et al., 2011). A recent study has investigated the movement of Maurer's clefts and found that Maurer's clefts are highly motile during the ring stages, but are apparently static during the trophozoite stage (Gruring et al., 2011). Our quantitative studies confirmed that Maurer's clefts move rapidly during ring stages (Fig. 5A). The movement resembled a random walk (Figs 4 and 7), and the measured apparent speed of 0.14–0.19 μm s–1 was in the range at which, according to the Stokes–Einstein equation, a spherical Brownian particle with a radius of 220 nm (the estimated average radius of Maurer's cleft precursors in rings) would move at a temperature of 37°C in a medium that has the viscosity of the erythrocyte's cytoplasm. Since the motion of Maurer's clefts was independent of actin (Fig. 5B) and since there is no evidence of microtubules being associated with these organelles, it is plausible that Maurer's clefts move by Brownian motion (for a discussion of the shape of Maurer's clefts see below).
We did not simulate the effect of variant-specific differences in the viscosity of the erythrocyte cytoplasm on Brownian motion of Maurer's clefts, although such differences are likely to exist. Both haemoglobin S and C can adopt a semi crystalline state under physiological conditions. As a consequence, infected HbS and HbC erythrocytes would have an increase cytoplasmic viscosity, which might explain the significantly reduced Brownian motion of Maurer's clefts observed in ring-infected HbCC erythrocytes (Fig. 5A).
The motion of Maurer's clefts slowed down fivefold to 0.034 μm s–1 (for HbAA erythrocytes) as the parasite matured to become a trophozoite; however, it did not stop (Figs 4 and 5C). The deceleration might be explained by a drastic change in the shape of Maurer's clefts and/or by a tethering mechanism that anchors Maurer's clefts to larger subcellular structures, such as the membrane skeleton of the host cell. Consistent with a tethering mechanism, a recent study has shown that long actin filaments connect the Maurer's clefts with the knobs and ultimately via KAHRP, the major component of knobs, to the erythrocyte membrane skeleton (Kilejian et al., 1991; Pei et al., 2005; Cyrklaff et al., 2011; 2012). Since this actin network is generated at approximately the same time as the motion of Maurer's clefts slows down one may propose a causative relationship between the two events. And indeed, treating the trophozoite-infected wild-type erythrocytes with the actin depolymerizing agent cytochalasin D significantly increased the speed at which Maurer's clefts moved (Fig. 5D; P < 0.001). However, the increase fell short of the speed observed during ring stages. Thus, while actin filaments slow down the motion of Maurer's clefts, their effect was insufficient to fully account for the drastic deceleration observed in trophozoites. That the contribution of the actin filaments in restraining the motion of Maurer's clefts is not larger may be due to the proposed dynamic nature of the actin network. It has been hypothesized that the actin filaments nucleate at the Maurer's clefts and depolymerize at the knobs, thereby creating a treadmilling mechanism that moves attached cargo vesicles towards the erythrocyte plasma membrane (Cyrklaff et al., 2012). It is also possible that filaments other than actin filaments hold the Maurer's clefts in place (Hanssen et al., 2008). However, if such tethers exist then they must be thinner and/or more fragile than the actin filaments and hence escaped detection by cryo-electron microscopy (Cyrklaff et al., 2011).
If not a tethering mechanism then how else can the change in Maurer's clefts motion be explained? Our electron microscopic and cryo-electron tomographic studies suggest that Maurer's clefts change their shape as they mature (Fig. 6). During the ring stages, Maurer's clefts appeared compact and spherical (Fig. 6) (Cyrklaff et al., 2011) – quite different from the published and generally accepted appearance of Maurer's clefts. The long slender or oblate membrane profiles (depending on the viewing angle) that characterize Maurer's clefts are only found in trophozoites (Atkinson and Aikawa, 1990; Przyborski et al., 2003; Wickert et al., 2003; 2004; Lanzer et al., 2006; Hanssen et al., 2008; Cyrklaff et al., 2011). Thus, it seems that Maurer's clefts become flattened and infolded as they mature. Maturation seems to be a rapid process since the motion of Maurer's clefts slows down within a short period of time (Gruring et al., 2011). This change in shape impacts on the diffusion constant and hence the speed at which Maurer's clefts move by Brownian motion. Assuming a Perrin factor of 1.7 to correct for the altered shape of Maurer's clefts, a diffusion constant of 3.2 × 10−15 m2 s–1 (corresponding to an average speed of 0.09 μm s–1) was calculated. This theoretical diffusion constant is comparable with the speed at which Maurer's clefts moved in the presence of cytochalasin D in trophozoite-infected wild-type erythrocytes (0.07 ± 0.01 μm s–1; Fig. 5D). The cytochalasin D value was chosen as a reference because cytochalasin D destroys the actin network (Cyrklaff et al., 2011), thus eliminating its contribution in restraining the motion of Maurer's clefts. On the other hand, treatment of infected erythrocytes with cytochalasin D might alter the morphology of Maurer's clefts (Cyrklaff et al., 2011), which, in turn, would affect the diffusion constant. Thus, the simulations conducted on Brownian motion of Maurer's clefts can only be considered first approximations. Irrespectively, our data suggest that Maurer's clefts move by Brownian motion and that the lower speed observed in trophozoites is partly due to an altered shape and partly due to the anchoring of the Maurer's clefts to the actin network. We did not consider possible parasite-induced temporal changes in the viscosity of erythrocyte cytoplasm.
The contribution of the actin network in restricting the motion of Maurer's clefts is further illustrated when comparing wild-type infected erythrocytes with infected erythrocytes containing haemoglobin variants (Fig. 5C). A previous study has shown that this actin network is only rudimentarily developed in parasitized erythrocytes containing HbSC and HbCC and does not connect the Maurer's clefts with the knobs (Cyrklaff et al., 2011). Consistent with an aberrant actin network, Maurer's clefts moved significantly faster in trophozoite infected erythrocytes containing haemoglobin variants as compared with wild-type erythrocytes (Fig. 5C). This was true for all mutant erythrocytes investigated, including HbAS, HbAC, HbCC and HbSC containing red blood cells. Importantly, the speed at which Maurer's clefts moved in these cases was comparable to the speed observed in trophozoite-infected wild-type erythrocytes treated with cytochalasin D (Fig. 5D).
It is unclear what might be the physiological implications of the motion of Maurer's clefts during ring stages, their sudden deceleration during the trophozoite stage, and the aberrant motion in host cells containing haemoglobin variants. One can only speculate that the temporal changes in the motion of Maurer's clefts relate to their function in protein trafficking and sorting and that an aberrant motion behaviour affects their role in delivering parasite-encoded proteins to the erythrocyte surface.
Differential movement of Maurer's clefts adds to the list of phenotypic alterations displayed by P. falciparum-infected erythrocytes containing the haemoglobin variants HbC and HbS. Other alterations includes: aberrant presentation of adhesin, diminished amounts of presented adhesins, enlarged but fewer knobs, reduced cytoadhesion, altered Maurer's clefts morphology and an aberrant host-derived actin network (Fairhurst et al., 2003; 2005; 2012; Cholera et al., 2008; Cyrklaff et al., 2011). Whether our findings can be extrapolated to other haemoglobin variations that protect carriers from severe malaria, such as α-thallassaemia and fetal haemoglobin F (Taylor et al., 2012), remains to be seen, but may be inferred from the common phenotypic alterations shared with infected HbS and HbC erythrocytes, which include abnormal knob morphology and density, aberrant PfEMP1 presentation and impaired cytoadhesion (Amaratunga et al., 2011; Fairhurst et al., 2012; Krause et al., 2012).
In addition to altered motion of Maurer's clefts, we further noted several other parameters, including red blood cell size, parasite size, number of Maurer's clefts and fluorescence coverage of Maurer's clefts, which taken individually were not necessarily significantly different from wild-type erythrocytes, but which when combined revealed patterns of common behaviour among erythrocytes containing HbS and HbC. Parasitized erythrocytes that behaved very similarly according to our principal component analyses were the infected heterozygous HbAS and HbAC erythrocytes during rings and the HbC containing infected erythrocytes (HbAC, HbSC and HbCC) during the trophozoite stage (Fig. 8). The principal component analyses further identified erythrocyte variants that were inversely correlated with these clusters, e.g. infected HbAA and HbAS erythrocytes behaved significantly different from the HbC cluster (Fig. 8). It seems that HbS and HbC, in addition to bringing about major phenotypic alterations, also cause subtle changes in the parasitized erythrocyte that combined might contribute to the protective effect against severe malaria. It remains to be seen whether all these major and minor alterations are a direct or indirect consequence of impaired host actin remodelling.
- Top of page
- Experimental procedures
- Supporting Information
The ethical review boards of Heidelberg University and the Biomolecular Research Center (CERBA/Labiogene) at the State University of Ouagadougou in Burkina Faso approved the study. Written informed consent was obtained from all blood donors.
Red blood cells
Haemoglobin genotypes were determined by cellulose acetate electrophoresis and by polymerase chain reaction (PCR) restriction fragment length polymorphism (RFLP) as described (Modiano et al., 2001). Erythrocytes were washed three times in cold AB-transfection medium (RPMI 1640 medium supplemented with 2 mM l-glutamine, 25 mM Hepes, 100 μM hypoxanthine, 20 μg ml–1 gentamicin). Cells were stored at 4°C before use. Erythrocytes were usually used for parasite culture within 7 days after donation, and the red blood cells present in the parasite culture were not older than 3 weeks. The MCV and the MCHC of different erythrocyte variants were determined using a haematology analyser (Sysmex).
Throughout the study we used a mutant of the P. falciparum strain 3D7 that expressed a genomically integrated PfSBP1–GFP fusion (Saridaki et al., 2009). The strain was continuously cultured using the appropriate red blood cells. Cells were grown at a haematocrit of 5.0% (50 μl of packed red blood cells and 1 ml of medium in a 12-well plate) and at a parasitaemia of not higher than 5% in AB-transfection medium. Cultures were maintained at 37°C under controlled atmospheric conditions of 5% O2, 3% CO2 and a humidity of 96%. The cultures were synchronized using sequential treatments with 5% d-sorbitol (Lambros and Vanderberg, 1979). Only parasite cultures synchronized within a 4 h window were used in this study.
Live cell imaging
Plasmodium falciparum-infected erythrocytes were allowed to settle on a glass slide in physiological Ringer's solution (122.5 mM NaCl, 5.4 mM KCl, 1.2 mM CaCl2, 0.8 mM MgCl2, 11.0 mM d-glucose, 1.0 mM NaH2PO4 and 25.0 mM Hepes, pH 7.4) for 10 min. The cells were then examined with an LSM510 confocal laser scanning microscope (Carl Zeiss). PfSBP1–GFP in living cells was excited at a wavelength of 488 nm with an argon laser (laser power 40%, transmission 3%, Plan-Apochromat 100×/1.4 oil DIC & C-Apochromat 63×/1.2 Water immersion), and the emission was captured using a 505- to 550-band pass filter. Time series of 40 s were recorded with an image (512 × 512 Pixels) taken every 2 s. Both pixel factor and the time between the frames were determined by the Carl Zeiss AIM Image Examiner Version 4.2. Where indicated cells were pre-incubated with cytochalasin D (10 μM) or jasplakinolide (200 nM) for 10 min prior to taking time series.
Preparation and imaging of P. falciparum-infected erythrocytes for electron microscopy and cryo-electron tomography have recently been described (Cyrklaff et al., 2011).
The script for motion tracking of individual Maurer's clefts is based on the statistical software R (R Development Core Team, 2011), the add-on package EBImage for image processing and analysis under R (Sklyar et al., 2011), the ImageMagick-librarys (http://www.imagemagick.org, Version 6.6.9 Q16) for reading various file formats, and the GTK framework (http://www.gtk.org, Version 2.16.6) for displaying images. The motion tracking script is available for free downloading at URL: http://www.klinikum.uni-heidelberg.de/Malaria-1-Lanzer.6573.0.html. The images were prepared as follows: the DIC and GFP images from the time series were prepared as JPEG image sequence with FIJI (http://fiji.sc/wiki/index.php/Fiji). The cell of interest was encircled with a black line (width 7) and a reference point (size at least 30 px) was placed on a recognizable spot close to the edge of the encircled erythrocyte. All other cells present in the images were covered to block out their fluorescence signals. The trafficking tool then determines the compensated distance travelled by the selected Maurer's clefts from one frame to the next. The individual distances are then added up to yield a cumulative forward distance. The speed at which the Maurer's cleft moved was subsequently calculated by dividing the cumulative forward distance by the time spent between the first and the last frame. Usually a time series consisted of 20 images taken at an interval of 2 s. To verify if an error is introduced by preparing the images for the motion tracking tool, three different operators measured the same fluorescent object 10 times each. The variance introduced was less than 0.3%, indicating that the operator-introduced error is insignificant if the image preparation is carried out carefully.
Principal component analysis and hierarchical cluster analysis were performed as described (Abdi and Williams, 2010). Briefly, principal component analysis was carried out on the mean values of all variables collected by the motion tracking program. The data set was centred and scaled prior to the analysis in order to give all variables an equal statistical weight. The scores of the first two principal components were subjected to hierarchical cluster analyses (Romesburg, 1989). The distance matrix was calculated using the Euclidian distance and the Ward's method was used for the fusion of clusters. For investigating statistical significance, the Kruskal–Wallis anova on ranks test followed by multiple Wilcoxon rank-sum tests with Bonferroni adjustment were used, as indicated in the respective figure legends.
- Top of page
- Experimental procedures
- Supporting Information
We thank the Deutsche Forschungsgemeinschaft (research focus ‘Host–Parasite Coevolution’) and the Chica and Heinz Schaller foundation for financial support. We thank the EMBL, COS and Bioquant (Heidelberg) for access to EM facilities. M.L. and F.F. are members of the Heidelberg Excellence Cluster ‘CellNetworks’ and the European Network of Excellence ‘EVIMalaR’. N.K. and M.D. are PhD students of the HBIGS Graduate School of Life Sciences at Heidelberg University. We thank David Modiano for introducing us to the field of haemoglobinopathies.
- Top of page
- Experimental procedures
- Supporting Information
- 2010) Principal component analysis. WIREs Comp Stat 2: 433–459. , and (
- 2002) Protective effects of the sickle cell gene against malaria morbidity and mortality. Lancet 359: 1311–1312. , , , , , , et al. (
- 2011) A role for fetal hemoglobin and maternal immune IgG in infant resistance to Plasmodium falciparum malaria. PLoS ONE 6: e14798. , , , , , , et al. (
- 1990) Ultrastructure of malaria-infected erythrocytes. Blood Cells 16: 351–368. , and (
- 2004) Three-dimensional ultrastructure of the ring stage of Plasmodium falciparum: evidence for export pathways. Microsc Microanal 10: 551–562. , , , , and (
- 1979) Quantitative studies of ferritinlike iron in erythrocytes of thalassemia, sickle-cell anemia, and hemoglobin Hammersmith with Mossbauer spectroscopy. Proc Natl Acad Sci USA 76: 939–943. , , , and (
- 2000) Pfsbp1, a Maurer's cleft Plasmodium falciparum protein, is associated with the erythrocyte skeleton. Mol Biochem Parasitol 111: 107–121. , , , , , , et al. (
- 1980) Biophysical Chemistry. San Francisco, CA: W. H. Freeman and Company. , and (
- 2008) Oxidative process in erythrocytes of individuals with hemoglobin S. Hematology 13: 187–192. , , and (
- 2008) Impaired cytoadherence of Plasmodium falciparum-infected erythrocytes containing sickle hemoglobin. Proc Natl Acad Sci USA 105: 991–996. , , , , , , et al. (
- 2011) Hemoglobins S and C interfere with actin remodeling in Plasmodium falciparum-infected erythrocytes. Science 334: 1283–1286. , , , , , , and (
- 2012) Host actin remodeling and protection from malaria by hemoglobinopathies. Trends Parasitol 28: 479–485. , , , and (
- 2004) Images in clinical medicine. Homozygous hemoglobin C disease. N Engl J Med 350: e24. , and (
- 2003) Aberrant development of Plasmodium falciparum in hemoglobin CC red cells: implications for the malaria protective effect of the homozygous state. Blood 101: 3309–3315. , , , , and (
- 2005) Abnormal display of PfEMP-1 on erythrocytes carrying haemoglobin C may protect against malaria. Nature 435: 1117–1121. , , , , , , et al. (
- 2012) Abnormal PfEMP1/knob display on Plasmodium falciparum-infected erythrocytes containing hemoglobin variants: fresh insights into malaria pathogenesis and protection. Microbes Infect 14: 851–862. , , and (
- 2011) Diverse protective roles of the actin cytoskeleton during oxidative stress. Cytoskeleton (Hoboken) 68: 340–354. , , , , and (
- 2004) Var genes, PfEMP1 and the human host. Mol Biochem Parasitol 134: 3–9. , and (
- 2011) Development and host cell modifications of Plasmodium falciparum blood stages in four dimensions. Nat Commun 2: 165. , , , , , and (
- 2008) Electron tomography of the Maurer's cleft organelles of Plasmodium falciparum-infected erythrocytes reveals novel structural features. Mol Microbiol 67: 703–718. , , , , , , and (
- 1991) Beyond hemoglobin polymerization: the red blood cell membrane and sickle disease pathophysiology. Blood 77: 214–237. (
- 2009) 3-D analysis of the Plasmodium falciparum Maurer's clefts using different electron tomographic approaches. Biotechnol J 4: 888–894. , , , and (
- 1990) Effect of hemoglobin oxidation products on the stability of red cell membrane skeletons and the associations of skeletal proteins: correlation with a release of hemin. Blood 76: 2125–2131. , , , and (
- 1991) Selective association of a fragment of the knob protein with spectrin, actin and the red cell membrane. Mol Biochem Parasitol 44: 175–181. , , , , and (
- 2012) Alpha-thalassemia impairs the cytoadherence of Plasmodium falciparum-infected erythrocytes. PLoS ONE 7: e37214. , , , , , , et al. (
- 2005) How malaria has affected the human genome and what human genetics can teach us about malaria. Am J Hum Genet 77: 171–192. (
- 1979) Synchronization of Plasmodium falciparum erythrocytic stages in culture. J Parasitol 65: 418–420. , and (
- 2006) Maurer's clefts: a novel multi-functional organelle in the cytoplasm of Plasmodium falciparum-infected erythrocytes. Int J Parasitol 36: 23–36. , , , , and (
- 1995) Generation of normal human red cell volume, hemoglobin content, and membrane area distributions by ‘birth’ or regulation? Blood 86: 334–341. , , , , and (
- 2002) The pathogenic basis of malaria. Nature 415: 673–679. , , , and (
- 2001) Haemoglobin C protects against clinical Plasmodium falciparum malaria. Nature 414: 305–308. , , , , , , et al. (
- 2005) Structural and functional studies of interaction between Plasmodium falciparum knob-associated histidine-rich protein (KAHRP) and erythrocyte spectrin. J Biol Chem 280: 31166–31171. , , , , , and (
- 2003) Maurer's clefts – a novel secretory organelle? Mol Biochem Parasitol 132: 17–26. , , , and (
- R Development Core Team (2011) R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing.
- 1989) Cluster Analysis for Researchers. Raleigh, NC: Lulu Press. (
- 2009) Export of PfSBP1 to the Plasmodium falciparum Maurer's clefts. Traffic 10: 137–152. , , , and (
- 2011) EBImage-Image processing toolbox for R 2011. 3.8.0 edn. , , , and (
- 2012) Haemoglobinopathies and the clinical epidemiology of malaria: a systematic review and meta-analysis. Lancet Infect Dis 12: 457–468. , , and (
- 2005) Band 3 modifications in Plasmodium falciparum-infected AA and CC erythrocytes assayed by autocorrelation analysis using quantum dots. J Cell Sci 118: 1091–1098. , , , , , , and (
- 2011) Clinical Hematology: Theory and Procedures, 5th edn. Philadelphia, PA: Lippincott Williams & Wilkins. (
- 2003) Evidence for trafficking of PfEMP1 to the surface of P. falciparum-infected erythrocytes via a complex membrane network. Eur J Cell Biol 82: 271–284. , , , , , and (
- 2004) Maurer's cleft organization in the cytoplasm of Plasmodium falciparum-infected erythrocytes: new insights from three-dimensional reconstruction of serial ultrathin sections. Eur J Cell Biol 83: 567–582. , , , and (
- 1980) The internal viscosity of the human erythrocyte may determine its lifespan in vivo. Scand J Haematol 24: 57–62. , and (
- World Health Organization (2011) World Malaria Report 2011. Geneva: WHO Press.
- Top of page
- Experimental procedures
- Supporting Information
Fig. S1. Morphological features presented by erythrocytes infected with the 3D7 mutant used in this study.
A–C. Transmission electron micrographs showing knobs at the surface of parasitized erythrocytes containing HbAA (A), HbCC (B) and HbSC (C). Sections of high-pressure frozen, freeze substituted cells are shown. Typical examples of knobs are highlighted by arrowheads. Knob density and morphology are known to be aberrant in parasitized erythrocytes containing haemoglobin S or C (Fairhurst et al., 2012), as also shown in these images, but were not further investigated in this study.
D–F. Gallery of representative young Maurer's clefts present in ring-infected erythrocytes. The preservation procedures were as follows: (D) cryo-preservation; (E) high-pressure freezing and freeze-substitution, and post-staining with uranyl acetate revealing a densely packed protein coat on the surface of the Maurer's clefts. (F) High-pressure freezing and freeze-substitution, and stained only with osmium tetroxide to reveal membranes. Occasionally, the two leaflets of the membrane bilayer can be seen both for Maurer's clefts (black arrowheads) and for the erythrocyte plasma membrane (white arrowheads).
Scale bars in (A) to (C), 1 μm; in (D) to (F), 100 nm.
Fig. S2. Frequency histogram showing the speed distributions of Maurer's clefts in trophozoite-infected erythrocytes containing HbAA (white bars) and HbAC (grey bars).
Fig. S3. Simulated Brownian motion of a spherical particle with the radius of a ring-stage Maurer's cleft. The following parameters were entered into the Stokes–Einstein equation: radius, 220 nm; temperature, 37°C; internal viscosity of an erythrocyte, 0.15 Pa·s (Williams and Morris, 1980).
A. Results of 1000 simulations. The speed at which the particle would move is shown.
B. Actual movement of the particle over 40 s.
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.