J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141130
1. Introduction
The details of the antimalarial mode of actionof 4-aminoquinolines have been the subject ofintense scrutiny during the last 4 decades [1,2].Recent findings seem to elucidate the fundamen-tal elements of this mechanism [3–5]. Undoubt-edly, the action of this type of drugs is closelyrelated to the presence of free ferriprtoporphyrinIX (FP) that is generated inside the food vac-uole of the parasite during the digestion of in-gested host cell hemoglobin. In order to avoidthe demonstrable toxicity of FP to the parasite[6–8], it has hitherto been assumed that thisnoxious compound is converted to insoluble andharmless hemozoin [9]. Aminoquinolines accu-mulate inside the acidic food vacuole and areable to form complexes with FP and therebyinhibit the formation of hemozoin [10–16].However, we have recently demonstrated thateven in the absence of drug the polymerizationof FP into hemozoin is incomplete [17], imply-ing that the parasite must possess the means todetoxify this FP. We have suggested that FP isdegraded by glutathione as we have previouslydemonstrated for FP-loaded normal erythrocytes[18]. Thus, even if drug will totally inhibit FPpolymerization, this would add only 50% to theFP load on the detoxification, suggesting thatinhibition of FP polymerization per se is proba-bly insufficient to rationalize drug action. Wehave subsequently demonstrated thatchloroquine and amodiaquine inhibit the degra-dation of FP by glutathione, resulting in theaccumulation of FP in the membrane fraction ofdrug-treated infected erythrocytes [17]. This ac-cumulation of FP was shown to be directly pro-portional to the extent of parasite killing in adose- and time-dependent manner. These resultsare in agreement with the long-standing sugges-tion of Fitch and his colleagues [19] that FPserves as a receptor for chloroquine in the in-fected cell and that the binding of chloroquineto FP mediates cell killing [5], probably throughthe permeabilization of the parasite membraneto potassium [8,20]. We have recently estab-lished that while chloroquine, amodiaquine, py-ronaridine, halofantrine and some bis-quinolines
inhibit the degradation of membrane-associatedFP, quinine and mefloquine fail to do so.
In the present work we have probed the fateof FP in various strains of Plasmodium falci-parum in order to determine whether or nothandling of FP by these strains has any correla-tion to drug resistance. We have tested the ef-fect of drugs on hemozoin production and theaccumulation of FP in the membrane fraction inthese strains. These investigations were con-ducted in order to assess the universality of ourprevious observations that were obtained with asingle strain and with a limited assortment ofdrugs and in an attempt to get further insightsinto the mechanism of drug resistance.
2. Materials and methods
2.1. Materials
Fresh O+ or A+ blood and human O+ orA+ plasma were kindly donated by the ShaareiZedek or the Hadassah Hospitals. RPMI-1640was obtained from Biological Industries, Kib-butz Bet Haemek, Israel. N-2-hydroxyethylpiperazine-N %-2-ethanesulfonic acid (Hepes),Amodiaquine dihydrochloride, Quinine hy-drochloride and hemoglobin determination kit(525-A) were purchased from Sigma ChemicalCo. Chloroquine diphosphate was obtained fromServa and hemin from Porphyrin Products, Lo-gan, Utah. [3H]hypoxanthine (43 Ci mmol−1)was procured from Amersham. Mefloquine wasgenerously provided by A.F. Cowman. All otherchemicals were of the best available grade.
Parasite cultivation. The W2, D2, FCR3, G9and ITG chloroquine-resistant strains and thedrug-sensitive strains HB3, NF54 and D6 of P.falciparum were cultivated as previously de-scribed [21] in 10 mM glucose, 25 mM NaHCO3
enriched RPMI 1640 medium supplemented withHepes and 10% human heat-inactivated plasma.Cultures were synchronized by the sorbitol tech-nique [22] using the less toxic alanine. Parasiteswere fractionated to different stages using thePercoll-alanine gradient centrifugation protocol[23].
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141 131
2.2. Determination of parasite protein content
Infected cells were separated from synchronizedcultures as follows: cultures were centrifuged(12 000×g, 20 min at room temperature in a fixedangle SS34 rotor; Sorvall RC-5 centrifuge)through two layers of Percoll-wash medium sup-plemented with 6% (w/v) alanine): 60% Percoll(into which all infected cells migrate; ]90% para-sitemia) and 90% Percoll (which contains unin-fected cells). Cell numbers and parasitemia weredetermined and about 5×108 cells (as determinedin a hemacytometer) were lysed in 20 ml of0.003% (w/v) saponin in wash medium at roomtemperature. The free parasites were washed oncewith 40 ml buffered saline (pH 7.4) and the cellswere lysed in 1 ml of ice-cold distilled water for 10min. After centrifugation at 10 000×g for 5 minat 4°C in a microfuge, 50 ml of the supernatantwere taken for protein determination according toBradford, using bovine serum albumin as a stan-dard. The appropriate concentration of saponinneeded for the selective disruption of the host cellwithout infringement of parasite integrity, wasdetermined by parallel measurement of proteincontent and the activity of isocitrate dehydroge-nase in intact infected cells as compared to freeparasites [24]. Protein content/number of cells wascalculated from the standard curve.
2.3. Balance sheet of FP disposal in infectederythrocytes
Cultures of the various strains were synchro-nized and cultivated to the trophozoite. Infectedcells were separated from uninfected cells as de-scribed above. Both uninfected and infected cellswere used for the determination of hemoglobin,hemozoin and membrane-associated FP, as fol-lows: Cells of each layer were suspended in 20 mlwash medium supplemented with 6% (w/v) ala-nine, cells were counted in hemacytometer and theparasitemia was assessed by microscopic inspec-tion of Giemsa-stained thin blood smears. Cellswere spun down and the cell pellet was lysed inice-cold 5 mM phosphate buffer, pH 8.0 (5P8),followed by three cycles of freezing in liquid nitro-gen and thawing at room temperature. The dis-
ruption of cells and organelles was followed bystaining with rhodamine 123 and observation un-der the fluorescence microscope. Whereas hypo-tonic lysis left many food vacuoles intact, theirintegrity was totally destroyed after the freeze-thaw cycles. After centrifugation at 10 000×g for5 min, all subsequent steps were done at 4°C.Twenty ml of the lysate were added to 980 ml ofDrabkin reagent and the absorbance was read at540 nm and the hemoglobin content in mmol (1010
cells)−1 was calculated from a calibration curveconstructed with known amounts ofmethemoglobin. For the assay of hemozoin, awell-defined method [25] has been used with somemodifications. The pellet was washed three timeswith 5P8 at 10 000×g 5 min (it was ascertainedthat the supernatants of the washes did not con-tain any FP), dissolved in 1 ml of 2.5% (w/v) SDSand centrifuged again for 20 min (longer centrifu-gations did not increase the hemozoin content;data not shown). The absorbance of the superna-tant was read at 345 nm (at this wavelength thereis no contribution from the traces of hemoglobinthat could remain bound to the membrane) andthe amount of FP in mmol (1010 cells)−1 wascalculated from a calibration curve constructedwith freshly prepared FP dissolved in SDS. Wename this fraction of FP as membrane-associatedFP because after the disruption of the cells andthe intracellular organelles and the extensivewashing the pellet (prior to solubilization withSDS) it is assumed to consist essentially of mem-branes and nucleic acids. Given the high affinityof FP to membranes, it seems safe to suppose thatthe measured FP was associated with the mem-brane fraction in the intact infected cell. Theremaining pellet (hemozoin) was dissolved with100 ml of 0.2 M NaOH and 900 ml 2.5% SDS(37°C for 30 min with occasional mixing until novisible pellet could be observed), the absorbanceat 400 nm was determined and the amount ofhemozoin (1010 cells)−1 was calculated from acalibration curve of FP constructed in the samesolvents. The recovery of hemozoin was ascer-tained with synthetic b-hematin added at the be-ginning of the extraction procedure either touninfected or to infected cells and was found to beclose to 100% (data not shown). The percentage
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141132
of FP converted to hemozoin was calculated bydividing the hemozoin content by the differencebetween hemoglobin in uninfected cells andhemoglobin in infected cells ×100. From themany determinations thus performed, which usedvery different numbers of infected cells in thestarting material, we could not find any correla-tion between the number of cells and hence theinitial amounts of hemozoin, suggesting that thereis no need for a minimal amount of pigment foreffective and complete extraction.
2.4. Effect of drug treatment on hemozoinproduction and le6els of membrane-associated FP
Synchronized cultures were seeded at the ringstage and allowed to grow for another 20 h untilparasites reached the trophozoite stage. Para-sitemia and cell number were determined andcultures were cultured 9drug for periods of timethat are indicated in the Results section. One mlculture was washed twice in wash medium (cul-ture medium without plasma, 37°C) to remove thedrug and cells were seeded in 24 well cultureplates in full culture medium supplemented with 5mCi ml−1 of [3H]hypoxanthine. After 4 h of fur-ther cultivation, triplicate samples were trans-ferred into 96 well plate and parasite-associatedradioactivity was determined using the Filtermate/Matrix 96 Direct Beta counter. Inhibition of para-site growth was calculated compared to untreatedcontrols. The remaining cultures were used for thedetermination of hemozoin and membrane-associ-ated FP, as described above. The same parameterswere also determined in infected cells prior toincubation.
3. Results
3.1. Protein and hemozoin content during parasitede6elopment
Tightly synchronized cultures (two successivealanine treatments 30 h apart) were used for thedetermination of parasite cytosol protein andhemozoin contents at different time points alongthe development cycle of the parasite. Typical
results out of several repeats are shown in Fig. 1.They indicate that the time-course of hemozoinproduction and the increase in protein content inthe developing intraerythrocytic parasite is paral-lel to several known enzymatic activities [24,26–31]: they remain almost constant during the ringstage and increase abruptly during the trophozoitestage. There is some slight variation in proteincontent between the two strains tested, but thismay be due to conditions of the particular culture.The hemozoin content seems to be less variableand unlike what was previously assumed, it showsa slight increase during the ring stage and contin-ues to increase even at the schizont stage, suggest-ing that hemoglobin digestion occurs during thesestages.
3.2. Hemozoin content in different strains
For the comparison of hemozoin content indifferent parasite strains, it is imperative to assurethat all strains are exactly at the same develop-mental stage, otherwise erroneous results may be
Fig. 1. Production of parasite proteins and hemozoin duringthe development cycle. Representative results of several exper-iments with different strains are shown. Tightly synchronizedcultures of the HB3 (circles) and the W2 (squares) strains wereused for the determination of parasite cytosolic protein (lefty-axis; plain lines; filled symbols) and hemozoin (right y-axis;dotted lines; filled symbols) content as a function of time afterinvasion, as described in Section 2.
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141 133
Fig. 2. Relationship between hemozoin content and parasitecytosolic protein concentration in various strains of P. falci-parum. Infected cells were isolated from trophozoite stagecultures by the Percoll-alanine method and the contents ofcytosolic protein and hemozoin were determined as describedin Section 2. Linear regressions of hemozoin content vs.protein content were performed, yielding the following results(slope and regression coefficient, r): D2—0.181; 0.785.FCR3—0.191; 0.924. G9—0.195; 0.872. HB3—0.217; 0.988.ITG—0.209; 0.834. NF54—0.277; 0.963. W2—0.296; 0.998.The best fit for all data yielded a slope of 0.224; r=0.887.
tivity to chloroquine and hemozoin production.This conclusion holds even if the comparison isrestricted only to strains with the highest r-values,e.g. NF54, HB3, FCR3 and W2.
3.3. Con6ersion of FP to hemozoin at differentstages
We have shown previously that FP derivedfrom the digestion of host cell hemoglobin is notstoichiometrically converted into hemozoin [17].Here, we further probe this phenomenon by test-ing the efficacy of conversion of FP to hemozoin.To obtain the different stages for analysis, twodifferent protocols were used: the FCR3 strainwas tightly synchronized by two successive ala-nine treatments; and infected cells were isolatedfrom uninfected erythrocytes by the Percoll gradi-ent centrifugation technique. Alternatively, non-synchronized cultures were fractionated accordingto stage by the same technique. Results obtainedby the two protocols were joined together as theywere essentially identical. Infected and non-in-fected cells were analyzed for hemoglobin andhemozoin content as described in Materials andMethods. Results depicted in Fig. 3 show the plotof hemozoin content versus the amount ofhemoglobin digested [Hb(n–i)] (calculated by sub-tracting the hemoglobin content of infected cells([Hbi]) from that of uninfected cells ([Hbn]) ob-tained from the same culture), both expressed asFP equivalents. The linear correlation that wasfound between hemozoin production andhemoglobin digestion, suggests that the efficacy ofconversion of hemoglobin FP to hemozoin is con-stant throughout parasite development, as is therate of degradation of FP. Since the slope of theregression line is 0.431 (r=0.931; n=12), thismeans that only 43.1% of the FP generated duringhemoglobin digestion is converted into hemozoin.
In order to get a deeper insight into the processof FP production, we have assumed that FPwhich is produced during the digestion ofhemoglobin (obtained from the difference be-tween hemoglobin content of normal RBC andthat of infected cells [Hb(n–i)]) is either polymer-ized into hemozoin (hemozoin) or is degraded(D). All these parameters are essentially the inte-
obtained. Synchronized cultures were grown tothe trophozoite stage (as determined by micro-scopic inspection of Giemsa-stained thin bloodsmears), infected cells were separated by the Per-coll gradient technique and used for protein andhemozoin determination. Results from differentcultures of various strains are shown in Fig. 2,where the hemozoin content has been plottedagainst the protein content. Some strains, e.g. D2,ITG and G9, show a very high variability in thehemozoin to protein content relationship, whileothers, e.g. FCR3, W2, HB3 and NF-54 display amuch lower variability, as evidenced by the highregression coefficients. From the slopes of theregression lines (see legend to Fig. 2) it is evidentthat different strains generate different amountsof hemozoin, D2, FCR3 and G9 being the leastproductive, while NF-54 and W2 are the mostprolific (W2 produces 64% more pigment thanD2). Thus, there is no correlation between sensi-
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141134
grals of the rates of hemoglobin digestion, hemo-zoin production and FP degradation, respectively,up to the time of sampling. Thus, [Hb(n–i)]=[hemozoin]+D. Dividing this equation by[Hb(n–i)] and rearranging it, gives 1-[hemozoin]/[Hb(n–i)]=D/[Hb(n–i)]. The left side of thisequation has been plotted against 1/[Hb(n–i)] us-ing the data depicted in Fig. 3 and results areshown in Fig. 4. If the amount of FP degradationwere equal at all stages, we would have expected alinear correlation. This is clearly not met and theform of the curve indicates that proportionatelymore FP is degraded at the young ring stage thanat the latter stages of parasite development.
The conversion of FP to hemozoin was alsotested in different strains. Here, only the tropho-zoite stage was used for the analysis. Resultsportrayed in Fig. 5 show that despite the largevariation in the results (some of which may be dueto dissimilarity of stages in conjunction with the
Fig. 4. The extent of FP polymerization is stage dependent.The data of Fig. 3 were replotted as 1-[HZ]/[Hb(n–i)] vs.1/[Hb(n–i)], where [HZ] is the concentration of hemozoin and[Hb(n–i)] is the concentration of digested FP. The line con-necting the experimental points was drawn by eye. As 1-[HZ]/[Hb(n–i)] is also the relative amount of FP that has not beenpolymerized and hence, degraded (D), it is clear that Ddecreases with parasite maturation (going from right to left onthe x-axis).
rapid rate of hemozoin production during thetrophozoite stage; see Fig. 1), all data fit the linearrelationship (r=0.816) between % conversion andhemozoin content, here used as an indicator ofparasite stage. Hence, the less than stoichiometricconversion of FP to hemozoin can be regarded asa general phenomenon in P. falciparum.
Further inspection of the results of this analysisfor those strains where sufficient data are avail-able, indicates that slopes vary for differentstrains. A higher slope is indicative of more effi-cient conversion. Results of this analysis shown inTable 1, clearly show that they correlate with therelative productivity of hemozoin derived fromthe data of Fig. 2. Here again, no correlationbetween the efficacy of FP production and thesensitivity to chloroquine could be found.
3.4. Effect of drugs on accumulation ofmembrane-associated FP
We have shown previously that treatment oftrophozoite-infected cells with chloroquine oramodiaquine causes a time- and dose-dependentaccumulation of membrane-associated FP whoselevel correlates with the extent of parasite killing
Fig. 3. Relationship between hemozoin content and amount ofdigested hemoglobin. The FCR3 strains were grown to thetrophozoite stage. Infected cells were separated from unin-fected cells and both were used for the determination ofhemoglobin concentration and hemozoin content. The amountof hemoglobin that was digested (Hb(n–i)) was calculated bysubtracting the hemoglobin concentration of infected (Hb(i))cells from that of uninfected cells (Hb(n)) in the same culture.Depicted results were obtained from several individual cul-tures. The regression line yielded a slope of 0.48; r=0.931).
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141 135
[17]. We further show here that the effect ofchloroquine to increase the level of membrane-as-sociated FP is true for the ring and the schizontstages as well (Fig. 6). For the ring and thetrophozoite stages, the correlation between mem-brane-associated FP and parasite killing is main-tained. However, at the schizont stage, despite ahigher level of FP, parasite killing is less than atthe trophozoite stage. We interpret the refractori-ness of schizonts as follows: a vast increase inmembrane material occurs at this stage [32] andthus the concentration of FP may be actuallylower than in the younger stages and insufficientto perturb the integrity of the membranes (seeSection 4).
The correlation between the accumulation ofFP in the membrane fraction of trophozoite- in-
Table 1Relationship between efficacy of conversion of hemoglobin FPto hemozoin at the trophozoite stage and the productivity ofhemozoin in different strainsa
a Data from Fig. 2 were used to calculate the productivity ofhemozoin, that is the slope of hemozoin content to proteinconcentration. Data from Fig. 5 were used to calculate theefficacy of conversion of hemoglobin FP to hemozoin.
fected cells and parasite killing was also tested intwo additional strains, the chloroquine-sensitiveHB3 and the chloroquine-resistant W2. Resultsshown in Fig. 7 indicate that such correlationdoes exist for these strains as well and implies that
Fig. 5. The conversion of FP into hemozoin in different strainsat the trophozoite stage. The amounts of hemozoin and de-graded hemoglobin were determined in trophozoite stage-in-fected cells enriched by the Percoll-alanine method. The % ofconversion was calculated as [Hemozoin]/[Hb(n–i)]*100 andplotted against [Hemozoin], here used as a parameter forparasite maturation (the Percoll-alanine method isolates allinfected cells irrespective of their precise age at the trophozoitestage). The slopes for the different strains, indicating efficacyof conversion of hemoglobin FP to hemozoin and the correla-tion coefficients of linear regression (r), are: D2—35.2, 0.414;D6—10.2, 0.718; FCR3—7.5, 0.840; G9—5.1, 0.184; HB3—15.64, 0.979; ITG—10.1, 0.962; NF54—5.6, 0.990; W2—18.9,0.951. The slope for all data is 10.95, r=0.816.
Fig. 6. Effect of drug treatment on membrane-associated FP inthe three main parasite developmental stages, in FCR3.Tightly synchronized cultures harboring the desired stage wereincubated for 4 h under culture conditions with 10 mMchloroquine. Cells were processed for the determination ofmembrane-associated FP prior to incubation (t=0), after 4 hin absence of drug (t=4, −CQ) and in presence of drug(t=4, +CQ) and parasite viability was assessed, as describedin Section 2. Ring stage-white bars; Trophozoite stage-graybars; Schizont stage-black bars. Numbers above bars indicate% inhibition of parasite growth.
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141136
Fig. 7. Dose-dependent effect of chloroquine and amodiaquine on membrane-associated FP, hemozoin levels and parasite viabilityin the chloroquine-sensitive strain HB3 and the resistant strain W2. Data shown are representative results from at least threeidentical experiments. Cultures at the trophozoite stage were exposed for 4 h under culture conditions to the indicatedconcentrations of chloroquine and amodiaquine. Thereafter, the levels of hemozoin (left y-axis in FP equivalents; gray bars),membrane associated FP (left y-axis in heme equivalents×10; white bars) and parasite viability (right y-axis, % inhibition ofparasite growth; black bars), were determined as described in Section 2. Controls at the beginning (T=0) and the end (T=4) ofincubation with no drug, are also shown.
this occurrence may be universal for all strains.Also shown is the effect of drug treatment on thelevel of hemozoin. The inhibition of hemozoinformation is also dose-dependent and the reduc-tion in the levels of hemozoin more than accountfor the amount of FP that is associated with the
membrane fraction. Specific scrutiny of the dataindicate that the inhibition of hemozoin forma-tion and the increase in membrane-associated FPoccurs at lower drug concentrations in the HB3strains than in the resistant W2 strain, in agree-ment with their relative drug sensitivity.
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141 137
Fig. 8. Effect of drugs on the levels of hemozoin in differentstrains, as affected by various drugs. Results shown here andin Fig. 9 are representative of several identical experiments (atleast twice for each strain and up to six times for somestrains). Cultures at the trophozoite stage of various strainswere incubated for 4 h under culture conditions in the presenceof 10 mM of amodiaquine (AQ), chloroquine (CQ), quinine(Q) and mefloquine (MQ). Infected cells were then processedfor the determination of hemozoin levels and parasite viability.Controls at the beginning (T=0) and the end of incubation(T=4) were similarly analyzed. Inhibition of parasite growth(as % of untreated controls) are given in Table 2.
Fig. 9. Effect of drugs on the levels of membrane-associatedFP in different strains, as affected by various drugs. The samecultures used for the determination of hemozoin and parasiteviability depicted in Fig. 8, were used for the determination ofmembrane-associated FP.
continued in absence of drug to reach in somestrains twice the level from the beginning of theexperiment. This large increase is typical of thetrophozoite stage (see Fig. 1). However, in pres-ence of drugs, hemozoin production was almostcompletely inhibited. The extent of inhibition wasless pronounced in the highly resistant strains W2and ITG, as compared to the sensitive strain HB3and to the moderately resistant strain FCR3.
3.6. Effects of drug on membrane-associated FPand correlation with parasite killing
The same cells that served for hemozoin deter-mination, were also used for the determination ofmembrane-associated FP. Results depicted in Fig.9 indicate that there is a basal level of membrane-associated FP in infected cells and that this leveldoes not increase during further cultivation. Inuninfected cells from the same cultures no FPcould be detected in the membrane fraction. Inpresence of chloroquine this level increases sub-stantially and more so with amodiaquine. Mostimportantly, neither quinine nor mefloquine hadany effect on membrane-associated FP. In com-paring the different strains used for this analysis,we found that with amodiaquine the same in-crease in membrane FP caused a similar extent of
3.5. Effect of drugs on hemozoin production
Different strains were grown to the trophozoitestage and incubated with drugs at 10 mM for 4 h.They were then used for the determination ofhemozoin content as described in Section 2.Hemozoin content was also determined in controlinfected cells prior to drug treatment and after thesame period of incubation without drug. Resultsdepicted in Fig. 8 show that hemozoin production
Table 2Inhibition of parasite growth by various drugsa
AQ MQQCQStrain
82.3FCR3 74.8 71.347.477.8 70.172.5HB3 75.8
57.1 63.5W2 46.9 52.553.2 68.950.7D6 75.1
63.9ITG 68.484.853.1
a Results are related to data presented in Figs. 8 and 9. Theexperimental details are described in the legend to Fig. 8.
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141138
parasite killing (inhibition of hypoxanthine incor-poration). With chloroquine, the same increase inmembrane FP was more inhibitory for HB3, W2and D6 than for FCR3 and ITG (data notshown).
4. Discussion
Malaria-infected red blood cells are distin-guished from the cells of the infected host by theirhigh levels of non-hemoglobin FP, which is pro-duced during the digestion of the hemoglobin-richcytosol of the host erythrocyte [33]. The seques-tration of FP into the insoluble FP polymerhemozoin [34–36] has been thought heretofore toaccount for the efficient protection of the parasiteagainst the toxic effect of FP [6,7]. Hemozoinformation is assumed to be an effective process[9,11,12,16] and we show here that the pigmentaccumulates in infected cells parallel to parasitematuration (Figs. 1 and 2). Although the ringstage does not display a typical food vacuole(where FP is understood to be produced andhemozoin is seen to be accumulated), measurableamounts of hemozoin are produced during thefirst 20–24 h of parasite development. Indeed, ithas been proposed that the endocytic vesicles arethe site of hemozoin formation and the large foodvacuole is essentially the dumping site for thepigment, which appears only at the trophozoitestage when sufficient number of vesicles havecoalesced together [36]. Hemozoin levels werefound to correlate with hemoglobin digestion(Fig. 3). The efficacy of hemozoin productionseems to be stage-dependent: it is less effective atthe ring stage and becomes more adept as theparasite matures (Fig. 4). If the production ofhemozoin depends on the presence of histidine-rich protein(s) [16], whose own production isstage-dependent, such connection may explain thepresent observations. We also show that the accu-mulation of hemozoin during parasite develop-ment parallels the increase in protein content andthat this is true for seven different parasite strainstested here (Fig. 2). The slope of the linear regres-sion correlating hemozoin content to protein levelfor each parasite strain varies, suggesting that the
efficacy of hemozoin production may be differentfor each strain. This productiveness however, doesnot correlate with drug sensitivity.
Most importantly, we show here that substan-tial amounts of FP, on the average more than50% during the whole parasite cycle, escape poly-merization. To allow for normal parasite develop-ment, the FP must either be degraded in situ, orleave the food vacuole since it demonstrably in-hibits vacuolar proteases and thereby inhibits fur-ther digestion [37,38] and permeabilizesmembranes, consequently dissipating the protongradient needed for normal vacuolar activities[39]. FP can be degraded by H2O2 [40], which isproduced inside the food vacuole [41]. But theingested host cell cytosol contains catalase [42]which could reduce the peroxide to below thelevel necessary for FP degradation and FP itselfhas catalase and peroxidase activities, the latterdestroying FP itself. Although both activities areinhibited in vitro by antimalarials [43], their rolesin H2O2 disproportionation and FP degradationin the intact parasite remain to be shown. As FPdissolves into and translocate across, membranes[44–47], it could, once it reaches the cytosol of theparasite, be degraded by heme oxygenase, whoseactivity has been reported in P. berghei and P.knowlesi [48]. However, as P. falciparum does notdisplay such activity ([35,49] and our own unpub-lished observations), it can be surmised that FP isdecomposed by glutathione, as we have recentlydemonstrated [18]. Either vacuolar or cytosolicdegradation of FP can be the source of ironneeded for parasite anabolism.
How does the fate of FP in infected cells reflecton the antimalarial mode of action of 4-amino-quinolines? These drugs have been shown to formcomplexes with FP and thus inhibit the generationof hemozoin in vitro at the acid pH that prevailsinside the food vacuole [10–16], as well as inintact infected cells [25,50]. FP:drug complex for-mation does not prevent FP dissolving in phos-pholipid membranes [46,47] and possibly, theensuing translocation across it. In the presence ofdrugs, the levels of free FP are expected to in-crease, but its detoxification by glutathione isprobably not overwhelmed: at the trophozoitestage that is the most sensitive to chloroquine,
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141 139
unpolymerized FP would add at most 50% to theFP load. Increased levels of free FP have beenobserved in chloroquine- and amodiaquine-treated P. berghei-infected mouse cells, but notwhen cells were treated with mefloquine orquinine, although the levels of hemozoin werereduced with all drugs [50]. In the present investi-gation we show again that the accumulation ofFP in the membrane fraction of infected cellsincreases at all parasite stages (Fig. 6). The in-crease is dose-dependent in chloroquine-sensitive(HB3) and -resistant (W2) strains with bothchloroquine and amodiaquine and the levels ofFP correlate with parasite killing (Fig. 7), as wehave shown before for FCR3 [17]. The accumula-tion of FP is compatible with the reduction ofhemozoin production in the presence ofchloroquine and amodiaquine in all strains tested(Figs. 8 and 9). The inhibition of hemozoin pro-duction is compatible with the ability of the drugsto form complexes with FP and has been shownpreviously to occur in vitro and here in intactinfected cells treated with all drugs tested. Theaccumulation of FP in these cells is due to thedemonstrable ability of chloroquine and amodi-aquine to inhibit the degradation of membraneassociated FP by glutathione [17,51]. FP aloneand more so in the presence of chloroquine, per-meabilizes membranes to potassium [8] and inchloroquine- treated infected cells, all the potas-sium is lost and replaced by sodium in the para-site cytosol [20], thus accounting for the cytotoxiceffect of this drug. It has been recently shown thatthe saturable component of chloroquine uptakeinto infected cells is entirely responsible for theantimalarial action of the drug [5]. As the specificinhibition of hemoglobin digestion reduced thedrug’s binding capacity, it was concluded that FPis the receptor for chloroquine [52]. The presentresults suggest the amount of this receptor, i.e. thebinding capacity, increases with time of exposureof the parasite to the drug, as the drug inhibits thedegradation of FP leading to its accumulation.Until this prediction will be tested experimentally,one can attempt to compare the binding capacitywith the drug-dependent increase in FP levels.The maximal capacity of chloroquine binding issome 35 nmol (1010 infected cells)−1 seen at about
100 nM extracellular chloroquine, compared to of70–80 nmol (1010 infected cells)−1 in membrane-associated FP measured at the same drug concen-tration, fitting well the 1:2 stoichiometry of thechloroquine:FP complex.
Quinine and mefloquine also inhibit the pro-duction of hemozoin, but these drugs do not raisethe level of membrane-associated FP. This differ-ential effect agrees with the inability of quinineand mefloquine to inhibit the degradation ofmembrane-associated FP by glutathione, whileboth amodiaquine and chloroquine are very effi-cient inhibitors of this process [51]. The distincteffects of chloroquine and amodiaquine versusthose of quinine and mefloquine, suggest that themechanistic details of the antimalarial effect ofthe latter drugs are somewhat different from thefirst. They are also compatible with the fact thatmefloquine-sensitive parasite strains arechloroquine-sensitive and vice versa; that thereexists a degree of cross-resistance to chloroquineand amodiaquine, but not with quinine ormefloquine [53] and that verapamil which canenhance the sensitivity to chloroquine of somestrains [54], has no effect on the susceptibility tomefloquine [55].
It was rather surprising to find no correlationbetween the efficacy of hemozoin production orthe level of hemozoin itself in the various strainsand their sensitivity to chloroquine. One couldargue that a reduced efficacy would increase thelevel of free FP available for accumulation in thepresence of drug, which inhibits its degradationby glutathione. However, the accumulation of FPwould also depend on the concentration of glu-tathione, as the inhibition of FP degradation bychloroquine is competitive [51]. The final resolu-tion of this apparent inconsistency must await thedetermination of glutathione concentrations in thedifferent strains, which is currently under investi-gation in our laboratory.
Acknowledgements
This investigation received financial supportfrom the UNDP/World Bank/WHO Special Pro-gramme for Research and Training in Tropical
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141140
Diseases (TDR) and from the United States–Is-rael Binational Science Foundation.
References
[1] Ginsburg H, Krugliak M. Quinoline-containing anti-malarials—mode of action, drug resistance and its rever-sal—An update with unresolved puzzles. BiochemPharmacol 1992;43:63–70.
[2] Slater AF. Chloroquine—mechanism of drug action andresistance in Plasmodium falciparum. Pharmacol Ther1993;57:203–5.
[3] Foley M, Tilley L. Quinoline antimalarials: Mechanismsof action and resistance. Int J Parasitol 1997;27:231–40.
[4] O’Neill PM, Bray PG, Hawley SR, Ward SA, Park BK.4-aminoquinolines—Past, present and future: A chemicalperspective. Pharmacol Ther 1998;77:29–58.
[5] Bray PG, Mungthin M, Ridley RG, Ward SA. Access tohematin: The basis of chloroquine resistance. Mol Phar-macol 1998;54:170–9.
[7] Fitch CD, Chevli R, Banyal HS, Phillips G, Pfaller MA,Krogstad DJ. Lysis of Plasmodium falciparum by ferripro-toporphyrin IX and a chloroquine- ferriprotoporphyrinIX complex. Antimicrob Agents Chemother 1982;21:819–22.
[8] Chou AC, Fitch CD. Mechanism of hemolysis induced byferriprotoporphyrin IX. J Clin Invest 1981;68:672–7.
[9] Warhurst DC. Haemozoin and the mode of action ofblood schizontocides: More controversy. Parasitol Today1995;11:204–5.
[10] Dorn A, Vippagunta SR, Matile H, Jaquet C, Venner-strom JL, Ridley RG. An assessment of drug-haematinbinding as a mechanism for inhibition of haematin poly-merisation by quinoline antimalarials. Biochem Pharma-col 1998;55:727–36.
[11] Dorn A, Stoffel R, Matile H, Bubendorf A, Ridley RG.Malarial haemozoin b-haematin supports haem polymer-ization in the absence of protein. Nature 1995;374:269–71.
[12] Egan TJ, Ross DC, Adams PA. Quinoline anti-malarialdrugs inhibit spontaneous formation of beta-haematin(malaria pigment). FEBS Lett 1994;352:54–7.
[13] Blauer G, Akkawi M. Investigations of B- and b-hematin.J Inorg Biochem 1997;66:145–52.
[14] Slater AF, Cerami A. Inhibition by chloroquine of anovel haem polymerase enzyme activity in malariatrophozoites. Nature 1992;355:167–9.
[15] Chou AC, Fitch CD. Heme polymerase—modulation bychloroquine treatment of a rodent malaria. Life Sci1992;51:2073–8.
[16] Sullivan DJ, Gluzman IY, Goldberg DE. Plasmodiumhemozoin formation mediated by histidine-rich proteins.Science 1996;271:219–22.
[17] Ginsburg H, Famin O, Zhang J-M, Krugliak M. Inhibi-tion of glutathione-dependent degradation of heme bychloroquine and amodiaquine as a possible basis for theirantimalarial mode of action. Biochem Pharmacol1998;56:1305–13.
[18] Atamna H, Ginsburg H. Heme degradation in the pres-ence of glutathione—A proposed mechanism to accountfor the high levels of non-heme iron found in the mem-branes of hemoglobinopathic red blood cells. J Biol Chem1995;270:24876–83.
[20] Lee P, Ye Z, Van Dyke K, Kirk RG. X-ray microanalysisof Plasmodium falciparum and infected red blood cells:Effect of qinghaosu and chloroquine on potassium,sodium and phosphorus composition. Am J Trop MedHyg 1988;39:157–65.
[21] Trager W, Jensen JB. Human malaria parasites in contin-uous culture. Science 1976;193:673–5.
[22] Lambros CJ, Vanderberg JP. Synchronization of Plas-modium falciparum erythrocytic stages in culture. J Para-sitol 1979;65:418–20.
[23] Kutner S, Breuer WV, Ginsburg H, Aley SB, CabantchikZI. Characterization of permeation pathways in theplasma membrane of human erythrocytes infected withearly stages of Plasmodium falciparum : Association withparasite development. J Cell Physiol 1985;125:521–7.
[24] Vander Jagt DL, Hunsaker LA, Kibrige M, Campos NM.NADPH production by the malarial parasite Plasmodiumfalciparum. Blood 1989;74:471–4.
[25] Asawamahasakda W, Ittarat I, Chang CC, Mcelroy P,Meshnick SR. Effects of antimalarials and protease in-hibitors on plasmodial hemozoin production. MolBiochem Parasitol 1994;67:183–91.
[27] Deans JA, Thomas AW, Inge PM, Cohen S. Stage-spe-cific protein synthesis by asexual blood stage parasites ofPlasmodium falciparum. Mol Biochem Parasitol1983;8:45–51.
[28] Inselburg J, Zhang R-D. Study of dhydrofolate reductase-thymidylate synthase in Plasmodium falciparum. Am JTrop Med Hyg 1988;39:328–36.
[29] Rosenthal PJ, Kim K, McKerrow JH, Leech JH. Identifi-cation of three stage-specific proteinases of Plasmodiumfalciparum. J Exp Med 1987;166:816–21.
[30] Horrocks P, Jackson M, Cheesman S, White JH, KilbeyBJ. Stage specific expression of proliferating cell nuclearantigen and DNA polymerase delta from Plasmodiumfalciparum. Mol Biochem Parasitol 1996;79:177–82.
[31] Schmidt A, Schwarz RT, Gerold P. Plasmodium falci-parum : Asexual erythrocytic stages synthesize two struc-turally distinct free and protein-boundglycosylphosphatidylinositols in a maturation-dependentmanner. Exp Parasitol 1998;88:95–102.
J. Zhang et al. / Molecular and Biochemical Parasitology 99 (1999) 129–141 141
[32] Holz GG. Lipids and the malarial parasite. WHO Bull1977;55:237–48.
[33] Francis SE, Sullivan DJ, Goldberg DE. Hemoglobinmetabolism in the malaria parasite Plasmodium falci-parum. Annu Rev Microbiol 1997;51:97–123.
[34] Yamada K, Sherman IW. Plasmodium lophurae : composi-tion and properties of hemozoin, the malarial pigment.Exp Parasitol 1979;48:61–74.
[35] Slater AFG, Swiggard WJ, Orton BR, Flitter WD, Gold-berg DE, Cerami A, Henderson GB. An iron-carboxylatebond links the heme units of malaria pigment. Proc NatlAcad Sci USA 1991;88:325–9.
[36] Slomianny C. Three-dimensional reconstruction of thefeeding process of the malaria parasite. Blood Cells1990;16:369–78.
[37] Vander Jagt DL, Hunsaker LA, Campos NM. Character-ization of a hemoglobin-degrading, low molecular weightprotease from Plasmodium falciparum. Mol Biochem Par-asitol 1986;18:389–400.
[38] Gluzman IY, Francis SE, Oksman A, Smith CE, DuffinKL, Goldberg DE. Order and specificity of the Plasmod-ium falciparum hemoglobin degradation pathway. J ClinInvest 1994;93:1602–8.
[39] Schmitt TH, Frezzatti WA, Schreier S. Hemin-inducedlipid membrane disorder and increased permeability—Amolecular model for the mechanism of cell lysis. ArchBiochem Biophys 1993;30:96–103.
[40] Kremer M. The reaction of hemin with H2O2. EMBO J1989;185:651–8.
[41] Atamna H, Ginsburg H. Origin of reactive oxygen speciesin erythrocytes infected with Plasmodium falciparum. MolBiochem Parasitol 1993;61:231–41.
[42] Fairfield A, Abosch A, Ranz A, Eaton JW, Meshnick SR.Oxidant defense enzymes of Plasmodium falciparum. MolBiochem Parasitol 1988;30:77–82.
[43] Ribeiro MCD, Augusto O, Ferreira AMD. Influence ofquinoline-containing antimalarials in the catalase activityof ferriprotoporphyrin IX. J Inorg Biochem 1997;65:15–23.
[44] Cannon JB, Kuo F, Pasternack RF, Wong NM, Muller-Eberhard U. Kinetics of the interaction of hemin lipo-somes with heme binding proteins. Biochemistry1984;23:3715–21.
[45] Rose MY, Thompson RA, Light RW, Olson JS. Hemetransfer between phospholipid membranes and uptake byapohemoglobin. J Biol Chem 1985;260:6632–40.
[46] Ginsburg H, Demel R. The effect of protoporphyrin IXand chloroquine on phospholipid monolayer and thepossible implications to antimalarial activity. BiochimBiophys Acta 1983;732:316–9.
[47] Ginsburg H, Demel R. Interactions of hemin, antimalar-ial drugs and hemin- antimalarial complexes with phos-pholipid monolayers. Chem Phys Lipids 1984;35:331–47.
[48] Srivastava P, Puri SK, Dutta GP, Pandey VC. Studies onthe heme-oxygenase of a simian malarial parasite, Plas-modium knowlesi. Med Sci Res 1992;20:321–2.
[49] Goldberg DE, Slater AFG, Cerami A, Henderson GB.Hemoglobin degradation in the malaria parasite Plasmod-ium falciparum : An ordered process in a unique organelle.Proc Natl Acad Sci USA 1990;87:2931–5.
[50] Chou AC, Fitch CD. Control of heme polymerase bychloroquine and other quinoline derivatives. BiochemBiophys Res Commun 1993;195:422–7.
[51] Famin O, Krugliak M, Ginsburg H. Kinetics of inhibitionof glutathione-mediated degradation of ferriproto-prphyrin IX by antimalarial drugs. Biochem Pharmacol1999; in press.
[52] Bray PG, Janneh O, Raynes KJ, Mungthin M, GinsburgH, Ward SA. Cellular uptake of chloroquine is dependenton binding to ferriptoporphyrin IX and is independent ofNHE activity in Plasmodium falciparum. J Cell Biol 1999;in press.
[53] Foote SJ, Cowman AF. The mode of action and themechanism of resistance to antimalarial drugs. Acta Trop1994;56:157–71.
[54] Kyle DE, Oduola AM, Martin SK, Milhous WK. Plas-modium falciparum : modulation by calcium antagonists ofresistance to chloroquine, desethylchloroquine, quinineand quinidine in vitro. Trans R Soc Trop Med Hyg1990;84:474–8.
[55] Oduola AM, Omitowoju GO, Gerena L, Kyle DE, Mil-hous WK, Sowunmi A. Reversal of mefloquine resistancewith penfluridol in isolates of Plasmodium falciparumfrom South-West Nigeria. Trans R Soc Trop Med Hyg1993;87:81–3.
