(Received for publication, October 28, 1996, and in revised form, November 11, 1996)
From the Department of Chemistry, University of
Wyoming, Laramie, Wyoming 82071-3838 and the
Department of
Physics, State University of New York at Stony Brook, Stony Brook, New
York 11794
In a process inhibited by the quinoline
antimalarial drugs, Plasmodia detoxify heme released during
the degradation of hemoglobin by aggregating it into malarial pigment,
an insoluble crystalline heme coordination polymer. Synchrotron x-ray
powder diffraction patterns for intact desiccated malarial trophozoites
and synthetic -hematin have been measured; both materials correspond
to a single crystalline triclinic lattice with unit cell parameters
a = 12.2176(4), b = 14.7184(5),
c = 8.0456(3) Å;
= 90.200(2),
= 96.806(3),
= 97.818(3) ° and Z = 2. These results unambiguously
demonstrate that hemozoin crystallites are identical to synthetic
-hematin.
Heme is a potent multifunction regulator whose biochemical levels
and distribution are precisely controlled on both intra- and
extracellular levels (1). Efficient regulation of heme is particularly
critical for intra-erythrocytic parasites such as plasmodia
which process large quantities of heme in the post-invasion digestion
of the erythrocyte's hemoglobin. Plasmodia, which lack heme
oxygenases, detoxify heme by sequestering it into an insoluble heme
aggregate termed malarial pigment or hemozoin (2). The quinoline-based
family of antimalarials interfere with this process by an as yet
unknown mechanism that has recently come under intense scrutiny as part
of the effort to combat the spread of chloroquine-resistant strains of
Plasmodium. A variety of spectroscopic and bioanalytical techniques indicate that hemozoin is similar to the synthetic aggregated heme phase -hematin, which is thought to form strands of
hemes linked by propionate oxygen-iron bonds as well as interstrand propionate hydrogen bonds, Fig. 1 (3, 4, 5).
Characterization of the carboxylate stretching bands for the propionic
acid side chains by IR and Raman spectroscopy provides the best
evidence for the presence of iron-oxygen bonds to the propionate side
chains (3, 4, 5). Unfortunately, crystallographic characterization of
these heme aggregates has been hampered by the phase heterogeneity of
many synthetic preparations as well as by the small size of the
synthetic and natural crystallites isolated from either trophozoites and infected hosts (6). High resolution powder diffraction has been
used extensively for the solution of many structural problems (7), and
it can solve problems posed by diffraction from microcrystalline
phases. In this communication we describe the characterization of
-hematin derived from both synthetic and natural sources and provide
new unambiguous evidence that the heme aggregate present in late stage
trophozoites is
-hematin.
Circa 3 × 109 chloroquine-susceptible
Plasmodium falciparum late trophozoites of the 3D7 clone
(NF54 strain) were synchronized with sorbitol and allowed to achieve a
92% level of parasitemia (8). The resulting intact trophozites were
isolated by rapid cooling in liquid nitrogen, and the frozen suspension
of cells was lyophilized by freeze-drying in vacuo at
10 °C. In general this freeze-drying process required 12-14 h and
resulted in a fine black free flowing powder. Uninfected erythrocytes
from the same source were treated in an identical manner to give a
bright red free flowing powder. In their respective experiments,
samples of lyophilized trophozoites and erythrocytes were anhydrously loaded onto the diffractometer goinometer which was shrouded by a
helium-filled glove bag during both the sample loading and diffraction pattern measurement procedures.
Synthetic -hematin was prepared from hemin
(Sigma) by treating it with noncoordinating bases in
anhydrous conditions, Equation 1 (3, 5). Briefly, the powder
diffraction patterns were measured on beamline X3B1 at the Brookhaven
National Laboratories National Synchrotron Light Source using an
angular step of 0.01° in 2
with count times of between 5 and 20 s/data point. Source x-rays were monochromated by flat Si(111)
crystals, and the diffracted beam was analyzed with a Ge(111) crystal.
The resulting x-ray flux was continuously monitored for intensity
throughout the data collection with an ion counting chamber placed in
the x-ray beam immediately before the sample. The wavelength was
calibrated by measuring the six most intense peaks of an aluminum oxide
standard, and the system resolution is 0.01° at full width
half-maximum (9). For all room temperature measurements a flat plate
(either brass or single crystal quartz) geometry was employed. For low temperature measurements a brass rectangular flat plate was used. For
the trophozoite measurements the lyophilized sample was packed into a
circular well 2 mm deep and 18 mm in diameter (volume = 0.51 milliliter). The data reduction was performed with the GUFI software
package, and the unit cell was determined by exhaustive trial and error
methods with ITO and TREOR (10). The resulting triclinic cells were
refined with FULLPROF (7). The final LeBail fits and lattice refinement
were performed with FULLPROF (11, 12).
Equation 1.
The diffraction patterns for trophozoites and uninfected
erythrocytes were measured from 4 to 30°, Fig. 2, traces a
and b, Table I. The difference
of traces a and b is shown in panel ii of Fig. 2 trace c. Both the desiccated erythrocytes and
trophozoites scatter x-rays in two broad bands between 7-10° and
13-20° in 2. The origin of these bands is most likely due to
scattering from the lipid bilayer membranes in the cells. What is
particularly important is that the trophozoites have clear Bragg
diffraction spikes from a crystalline species not found in the
erythrocytes; moreover, as shown in panel ii of Fig. 2,
trace d, these peaks are identical to the sharp diffraction
pattern obtained from synthetic samples of
-hematin. The hemozoin
formed within trophozoites thus crystallizes in the same unit cell as
-hematin. The similar intensities of the two patterns in Fig. 2,
traces a and d, suggest that the materials are
also crystallographically identical on the atomic level as well.
|
At room temperature high resolution powder diffraction patterns, full
width half-maximum < 0.05°, are obtained for -hematin, [Fe(protoporphyrin-IX)]n, Fig. 3. The
diffraction pattern for this coordination polymer corresponds to a pure
single phase which indexes to give a triclinic cell with a high figure
of merit, Table I. At lower temperatures, 50 K, the diffraction
profile(not shown) is also readily indexed to give a slightly smaller
unit cell. The crystallographic parameters for the two temperatures differ principally in contraction along the a and
b axes at low temperatures, suggesting that these correspond
to the directions of weakest interchain interactions.
The experimental density of -hematin, 1.45(1) g ml
1,
corresponds to the occupation of two heme molecules in the unit cell. For this crystal system there are only two possible space groups: either the two hemes are related by inversion symmetry in space group
P-1 or they are crystallographically inequivalent and the space group
is P1. In principle, Wilson statistics can be used to determine the
presence of an inversion center from a diffraction pattern, but the
method is notoriously inaccurate for powder data (13). In the present
case the mean of |E · E
1| is
0.892 for 266 data out to 28° in Fig. 3, a value ambiguously between
the limiting values of 0.968 and 0.736 for centro and noncentrosymmetric space groups. In single crystal x-ray diffraction experiments the space group is often confirmed by the complete and
successful solution to the crystal structure. In spite of the high
resolution present in the diffraction patterns in Figs. 2 and 3, we
have been unable to locate the hemes in the unit cells of
-hematin;
thus we are unable to unambiguously fit the topological model in Fig. 1
into the experimentally determined lattice with molecular modeling
techniques. We attribute this to the severe peak overlap present in the
single angular dimension at higher angles in 2
and the
concomitant lack of peak intensity data at higher resolution.
A variety of compositions have been proposed for hemozoin, and a great
deal of uncertainty surrounded early attempts to characterize this
seemingly intractable black solid (14, 15). Native isolated hemozoin is
a highly crystalline material and published transmission electron
micrograph images clearly indicate that there is either a regular
lattice or d spacing of spacing of 9 ± 2 Å (6). In spite of preliminary suggestions that the heme aggregate present in
hemozoin is -hematin (14), this hypothesis was discounted due to its
contamination with proteins from early isolation protocols (16). The
development of better isolation procedures, which often employ
proteases, has allowed for the separation of hemozoin consisting solely
of heme as determined by complete elemental analysis (4). Although the
hemozoin isolated by these more rigorous procedures is identical to
-hematin by IR spectroscopy, it has been suggested that the hemozoin
may have been modified during the isolation steps (17). The diffraction
results presented here provide compelling evidence against this
possibility and clearly indicate that the heme aggregate produced
in situ is crystalline and has the same lattice as
-hematin.
The central question relating to the structure of hemozoin concerns the mechanism of its biosynthesis and how this is inhibited by the quinoline antimalarials. The biochemical consequences of nonsequestered heme accumulation are profound and include membrane lysis (18) and protease inhibition (19, 20). It was recently recognized that one defense plasmodia have developed against these drugs is efficient drug excretion out of the digestive vacuole (21); however, there may be other adaptations in the chloroquine-resistant strains. Indeed, a host of mechanisms have been proposed for the drug action of the quinolines (15), with the main consensus being the location of their activity in the digestive vacuole. An important step in elucidating the drug action mechanism of these antimalarials is the recent description of the heme aggregating ability of an unusual class of histidine-rich proteins, HRP I-IV, from P. falciparum, which have been cloned and overexpressed (22). These proteins may be the heme polymerase suggested by Slater and Cerami (2) or they may instead initiate heme aggregation (22) which can then be propagated by either enzymatic (2, 23) or nonenzymatic (24, 25, 26, 27) heme-sequestering mechanisms. Regardless of these mechanistic details, it is clearly important to have a detailed structural understanding of the metabolic product of heme detoxification.
Several properties of hemozoin make it an ideal excretory product of heme detoxification. First, it is dense and insoluble under physiological conditions and thus represents an irreversible sink for heme released from hemoglobin degradation. Second, hemozoin is a coordination polymer of ferric(protoporphyrin-IX), and thus it will not contribute to the oxygen radical stress from the auto-oxidation that would result from a coordination polymer of ferrous(protoporphyrin-IX). Finally, formation of hemozoin corresponds to an important excretory system for iron, release of which would further exacerbate the oxidative stress of the trophozoites.
The methodology used in this report may be useful in the in
situ study of other biomineral phases. The study of these
materials is most often performed following exhaustive isolation and
work up procedures, steps that potentially may alter the crystalline phase. Biomineralization and biocrystallization are most often associated with intra or extracellular membranes and their integral scaffold proteins. While many biological processes are facilitated or promoted by biomineralization, examples of biocrystallization as a
mediator of toxicity or a regulated means of intracellular excretion
are rare. Probably the best example of this phenomenon is the
crystallization of -hematin by malarial parasites in the early
trophozoite stage. Finally, an important role for hemozoin released
into the vasculatur during merozoite release may be the suppression of
the host's immunological response by compromising macrophages that
phagocitize hemozoin particles (8, 28).
We thank Kathie Moch and Dr. Jeffrey A. Lyons of the Department of Immunology, Walter Reed Institute of Research, Washington, D. C. for the generous gift of the lyophilized late trophozoites.