In Situ Localization and Structural Analysis of the Malaria Pigment

Aug 25, 2007 - Heinrich-Buff-Ring 26-32, D-35392 Giessen, Germany, Institut für ... Raman microspectroscopy was applied for an in situ localization o...
0 downloads 0 Views 410KB Size
Download PDF
J. Phys. Chem. B 2007, 111, 11047-11056

11047

In Situ Localization and Structural Analysis of the Malaria Pigment Hemozoin Torsten Frosch,† Sasa Koncarevic,‡ Linda Zedler,† Michael Schmitt,† Karla Schenzel,§ Katja Becker,‡ and Ju1 rgen Popp*,†,| Institut fu¨r Physikalische Chemie, Friedrich-Schiller-UniVersita¨t Jena, Helmholtzweg 4, D-07743 Jena, Germany, Interdisziplina¨res Forschungszentrum, UniVersita¨t Giessen, Heinrich-Buff-Ring 26-32, D-35392 Giessen, Germany, Institut fu¨r Agrar-u Erna¨hrungswissenschaften, Martin-Luther-UniVersita¨t Halle-Wittenberg, Ludwig-Wucherer-Strasse 2, D-06108 Halle, Germany, and Institut fu¨r Photonische Technologien e.V., Albert-Einstein-Strasse 9, D-07745 Jena, Germany ReceiVed: March 5, 2007; In Final Form: June 22, 2007

Raman microspectroscopy was applied for an in situ localization of the malaria pigment hemozoin in Plasmodium falciparum-infected erythrocytes. The Raman spectra (λexc ) 633 nm) of hemozoin show very intense signals with a very good signal-to-noise ratio. These in situ Raman signals of hemozoin were compared to Raman spectra of extracted hemozoin, of the synthetic analogue β-hematin, and of hematin and hemin. β-Hematin was synthesized according to the acid-catalyzed dehydration of hematin and the anhydrous dehydrohalogenation of hemin which lead to good crystals with lengths of about 5-30 µm. The Raman spectra (λexc ) 1064 nm) of hemozoin and β-hematin show almost identical behaviors, while some low wavenumber modes might be used to distinguish between the morphology of differently synthesized β-hematin samples. The intensity pattern of the resonance Raman spectra (λexc ) 568 nm) of hemozoin and β-hematin differ significantly from those of hematin and hemin. The most striking difference is an additional band at 1655 cm-1 which was only observed in the spectra of hemozoin and β-hematin and cannot be seen in the spectra of hematin and hemin. Raman spectra of the β-hematin dimer were calculated ab initio (DFT) for the first time and used for an assignment of the experimentally derived Raman bands. The calculated atomic displacements provide valuable insight into the most important molecular vibrations of the hemozoin dimer. With help from these DFT calculations, it was possible to assign the Raman band at 1655 cm-1 to a mode located at the propionic acid side chain, which links the hemozoin dimers to each other. The Raman band at 1568 cm-1, which has been shown to be influenced by an attachment of the antimalarial drug chloroquine in an earlier study, could be assigned to a CdC stretching mode spread across one of the porphyrin rings and is therefore expected to be influenced by a π-π-stacking to the drug.

Introduction Malaria is still one of the most important infectious diseases. According to the WHO1 about 40% of the world’s population is at risk, and up to three million people die every year from malaria.2 This is especially true for sub-Saharan African countries, where every 30 s a child younger than 5 years is killed by malaria, and has tremendous negative impact on their economical development.1,3 The malaria parasite Plasmodium falciparum has an extremely complicated life cycle in its vector, the Anopheles mosquito, and its human host. In the human host, the red blood cell serves as one habitat for the parasite where an asexual life cycle takes place. During this intraerythrocytic cycle, the parasite ingests vast amounts of host hemoglobin. It is transported to the parasitic food vacuole where it is degraded in order to use its amino acids for de noVo protein biosynthesis, maintain osmotic balance, or simply to make room for itself.4-6 Free, toxic heme, ferriprotoporphyrin IX (FP), is generated during the degradation of host hemoglobin by the parasite and * Author to whom correspondence should be addressed. † Friedrich-Schiller-Universita ¨ t Jena. ‡ Universita ¨ t Giessen. § Martin-Luther-Universita ¨ t Halle-Wittenberg. | Institut fu ¨ r Photonische Technologien e.V.

is detoxified by sequestration to an inert, crystalline substance, hemozoin, the brownish malarial pigment in the parasitic food vacuole.7 Chloroquine has been used as the leading antimalarial for some decades, because of its outstanding properties and has caused one of the most important health advances ever achieved by a drug against an infectious disease.8,9 Nowadays resistance against chloroquine is spreading around the globe. However, the molecular mode of action of the drug is still not well understood. The most widespread hypothesis is that the quinoline class of antimalarials is believed to act by interfering with the detoxification process of the hemoglobin digestion byproducts in the red blood cell state of the plasmodium’s asexual life cycle.10-16 NMR studies of the in Vitro interactions between chloroquine and different forms of hematin suggested that π-πinteractions between the quinoline ring system of the drug and the porphyrin system of the target structures are very likely to exist.17-20 Resonance Raman (RR) spectroscopy has emerged as an extremely capable tool to investigate biological systems. The method is characterized by high sensitivity and selectivity and can be applied to study biological samples with low laser power and almost no further preparation.21,22 In combination with a conventional light microscope, RR spectroscopy allows for a study of living cells with a spatial resolution of about the size

10.1021/jp071788b CCC: $37.00 © 2007 American Chemical Society Published on Web 08/25/2007

11048 J. Phys. Chem. B, Vol. 111, No. 37, 2007 of the excitation wavelength.23 The Raman microscopy technique has been shown to be very useful in investigating Plasmodia and locating the malaria pigment hemozoin.24-28 Resonance Raman spectroscopy allows for a selective enhancement of single chromophoric groups in extended biological environments by tuning the excitation wavelength into the electronic absorption bands. The drug-target-interaction can be monitored via resonance Raman spectroscopy by studying either the drug’s influence onto sensitive Raman modes of the biological target29,30 or vice versa. Thus, the application of deep UV Raman excitation wavelengths leads to a selective resonance enhancement of antimalaria active agents.31-34 Recently we have demonstrated that UV Raman microscopy is able to localize quinine in situ in cinchona bark.32 Furthermore we studied and located the promising, new active agent, the naphthylisoquinoline dioncophylline A, in different parts of the tropical liana Triphyophyllum peltatum by means of UV resonance Raman microscopy.33 UV resonance Raman spectroscopy has even been demonstrated to be sensitive to the influence of the water environment32 and the state of protonation of chloroquine34 (which is crucial for an accumulation of chloroquine inside the food vacuole of Plasmodium falciparum35-37) and even to distinguish between the diastereomers quinine and quinidine.32 Upcoming Raman experiments monitoring possible changes of sensitive Raman modes of these active agents due to the presence of the hematin target molecules will help to understand their molecular mode of action and thus to tailor new, more effective drugs. The second approach is to use visible Raman excitation wavelengths to selectively enhance the hematin target structure modes by exciting the well known absorption bands of these molecules in the visible range. First, visible RR studies concerning the interaction of chloroquine with hematin showed that covalent interactions can be excluded, in agreement with the NMR findings,19,20 because no wavenumber shifts larger than 2 cm-1 have been detected. However, a significant change in the depolarization ratio of a hematin Raman band at 1569 cm-1 has been detected upon chloroquine docking in the hydrous environment.29 In a landmark, high-resolution X-ray study,38 the structure of the malaria pigment hemozoin has been elucidated to consist of dimers, where the hematin molecules are bonded by reciprocal iron-carboxylate bonds. The dimers are linked by hydrogen bonds and built-up triclinic crystals. From this point of view, the quinoline class of antimalarials is expected to selectively block the small active growing face of the hemozoin crystal and act almost as tailor-made crystal growth inhibitors.39,40 The presented work contributes a detailed Raman and resonance Raman spectroscopic investigation of the malaria pigment hemozoin in combination with first DFT calculations of this complex system. The obtained results are an indispensable prerequisite to interpret and understand Raman spectroscopic drug-target interaction studies which are currently underway in our laboratory. Materials and Methods Cultivation of P. falciparum and Sample Preparation. Blood stages of P. falciparum strain Dd2 (chloroquine-resistant) were permanently maintained in culture using a modified protocol of Trager and Jensen.41,42 RPMI medium 1640 supplemented with NaHCO3 and HEPES, pH 7.4, 20 µg/mL gentamicin sulfate, 2 mM glutamine, 200 µM hypoxanthine, 0.2% Albumax II, and 4.1% of human serum was used for cultivation. Washed human erythrocytes of blood group A+ were added to

Frosch et al. a hematocrit of 5%. Parasites were maintained at a parasitaemia of 1-10% in an atmosphere of 90% N2/5% O2/5% CO2 at 37 °C and synchronized to the ring stage by the sorbitol method.43 Additionally, trophozoite-infected erythrocytes were enriched by Gelafundin separation.44 Blood smears were prepared on conventional glass slides and fixed in methanol. Extraction of Hemozoin. Parasites were isolated by lysing red blood cells for 10 min at 37 °C in 20 volumes of saponin containing buffer (7 mM K2HPO4, 1 mM NaH2PO4, 11 mM NaHCO3, 58 mM KCl, 56 mM NaCl, 1 mM MgCl2, 14 mM glucose, 0.02 mM saponin, pH 7.5). The pellet obtained by centrifugation (1500g, 5 min, 4 °C) was washed three times, and subsequently the P. falciparum cells were disrupted by three cycles of freezing in liquid N2 and thawing. The supernatant after centrifugation at 13.000g for 30 min was discarded. The obtained brownish pellet was lyophilized, and hemozoin was extracted according to.45 Briefly, the dry pellet was subsequently extracted with chloroform-methanol (2:1 v/v) and chloroformmethanol-water (10:10:3 v/v). After drying, the residue was resuspended in 100 mM Tris-HCl, 1 mM CaCl2, pH 7.5, and digested with Pronase to remove proteins and protein-linked GPIs. The insoluble residue was extracted with 50 mM sodium phosphate, pH 7.2, 4 M guanidine hydrochloride, 0.5% Triton X-100, and stirred overnight at 4 °C to remove nucleic acids. The residue (insoluble pigment, hemozoin) was recovered by centrifugation, washed three times with water and once with 80% 1-propanol, and dried. Synthesis of β-Hematin. β-Hematin was synthesized according to the acid-catalyzed dehydration of hematin.46,47 In a first method46 a solution of 0.109 g of hemin in 25 mL of 0.1 M NaOH was stirred for 30 min under reflux. Propionic acid was added until the pH reached 4.1. During the addition of the first drops of the acid, the precipitation of a dark solid could be observed. The suspension was then stirred for 18 h at 70 °C under reflux. After being cooled at ambient temperature, the reaction mixture was filtered through a nylon filter (Millipore, 22 µm) and the solid was washed, alternating with methanol and water and finally with 0.1 M NaHCO3 solution. The procedure was finished by washing with water until the filtrate became neutral. The black solid was dried over phosphorus pentoxide and will be referred to as β-hema_p in the following text. In the second acid-catalyzed method to prepare of β-hematin with acetate,47 82.3 mg of hemin was dissolved in 16 mL of 0.1 M NaOH solution, and 1.6 mL of 1 M HCl was added to neutralize the NaOH solution. The precipitation of a dark solid could be observed. After addition of 10 mL of 12.8 M acetate buffer, the reaction mixture was stored in the dark without stirring for 6 days. The filtration through a nylon filter (Millipore, 22 µm) resulted in a black solid (called β-hema_a in the following text), which was washed with 100 mL of methanol and water before drying over phophorous pentoxide. Following the anhydrous dehydrohalogenation of hemin48,49 all steps were carried out under a dry argon atmosphere with dried reactants. A 0.151 g amount of hemin was first dissolved in 3 mL of 2,6-lutidine for 15 min. A mixture of 50 mL of dry DMSO and 50 mL of dry methanol was added without stirring. The dark suspension was stored in the dark for 4 weeks under argon. After the allotted time, the mixture was released from the inert atmosphere and filtered through a nylon filter (Millipore, 22 µm). The precipitate was washed with methanol until the filtrate became colorless and dried over phosphorus pentoxide. This anhydrous synthesized sample of β-hematin is more

Malaria Pigment Hemozoin

J. Phys. Chem. B, Vol. 111, No. 37, 2007 11049

Figure 1. In situ Raman mapping of a P. falciparum-infected erythrocyte. Raman point spectra with an excitation wavelength of 532 nm have been taken every 0.5 µm two dimensionally across the infected erythrocyte shown in the microscope picture (C). The spatial resolved Raman spectra show significant different features. Spectra of hemozoin (A) and hemoglobin (B) are located where the microscopic inspection shows inclusions or no inclusions in the erythrocyte (C). These spectra are significantly different in the relative intensity of the Raman band at 1372 cm-1. The spatial distribution of the band at 1372 cm-1 is imaged with a linear intensity scale (D) and corresponds to the microscopic inspection. The in situ-located Raman spectra of hemozoin (A) and hemoglobin (B) agree well with the Raman spectra of extracted hemozoin (E) and hemoglobin (F).

crystalline (crystal lengths of about 5-30 µm) compared to the acid-catalyzed samples and is called β-hema_b in the following text. Spectroscopy. The nonresonant Raman spectra of the samples were recorded with a Bruker FT Raman spectrometer (RFS 100/S) at the macroscopic mode with a spectral resolution of 2 cm-1. A Nd:YAG laser operating at its fundamental wavelength of 1064 nm with an estimated laser power of approximately 100 mW at the samples was used as the excitation source. The Raman scattered light was collected by means of a liquid nitrogen-cooled Ge-detector. The resonance Raman spectra were acquired with a Raman setup (HR LabRam, Horiba/Jobin-Yvon) equipped with an Olympus IX71 microscope, a video camera, and liquid nitrogencooled CCD detector. An Olympus 20×/0.5 objective focused

the laser light on the sample slides. As an excitation wavelength, the 568 nm line of a krypton ion laser (Coherent Innova 300C) was used. The micro-Raman spectra were acquired with a Raman setup (HR LabRam, Horiba/Jobin-Yvon) equipped with an Olympus IX70 microscope, a video camera, and an air-cooled CCD detector operating at 220 K. A Leica 100×/0.9 objective focused the laser light on the sample slides. As excitation wavelengths, the 532 nm line of a frequency doubled Nd:YAG laser (Coherent Compass) and the 633 nm of a HeNe laser were used. For the x/y-scans, the samples were moved relative to the fixed laser spot with the help of a motorized stage. Validation of the wavenumber axis was performed between the measurements using the Raman signals from TiO2 (anatase).

11050 J. Phys. Chem. B, Vol. 111, No. 37, 2007

Frosch et al.

Figure 2. In situ Raman mapping of a P. falciparum-infected erythrocyte. Raman point spectra with an excitation wavelength of 633 nm have been taken every 0.5 µm across the infected erythrocyte shown in the microscope picture (D). The spatial resolved Raman spectra show significant different features. While very strong Raman signals of hemozoin (A) are located at regions, where the microscopic picture shows inclusions, very weak Raman signals could be detected outside these inclusions (B). The intensity scale of graph B is stretched 5 times for better illustration. The very intense Raman band at 755 cm-1 has been used to image the hemozoin distribution (C) with a linear intensity scale.

BP86/6-31G* and BP86/TZVP are not shown. The Raman intensities were calculated from the Raman scattering activities.56 The Raman spectra with finite bandwidth were simulated by convoluting the theoretical stick spectra with a Gauss-Lorentz weighted profile. Results and Discussion

Figure 3. Raman spectra of extracted malaria pigment hemozoin with excitation wavelengths 1064, 830, 633, 568, and 532 nm.

Density Functional Theory Calculation. DFT calculations were performed with Gaussian 0350 applying the exchange correlation functional BP8651,52 since this functional is known to provide reliable estimates of experimental wavenumbers without the necessity of applying scaling factors to the calculated harmonic vibrational wavenumbers (error compensation).53 SDD,54 6-31G*,50 and TZVP55 basis sets have been tested. However, the results with BP86/SDD were in better agreement with the experimental findings, and therefore the results with

Raman Mapping of Plasmodium Trophozoites. Within this work, Raman point scanning studies have been performed on Plasmodium falciparum-infected erythrocytes. Raman images have been constructed out of the array of Raman spectra. The experimental procedure is somewhat different from imaging studies, which have been done by Wood et. al26 with 780 nm laser excitation. In the case of filter imaging,26,28 an optical bandpass filter is placed into the beam line and an image of small spectral bandwidth is taken across a whole cell. This technique allows for fast imaging but with limited spatial and spectral resolution.57 However, the main advantage of the point scanning technique, applied in this contribution, is the availability of the whole Raman spectrum at every spatially resolved point of the infected erythrocyte. This allows for construction of images of any significant band and is a necessity for tracing slight influences on previously unknown Raman bands due to drug attachment. In Figure 1, a microscopic picture of a P. falciparum-infected erythrocytes is shown (1C). This erythrocyte was investigated by applying a resonance Raman excitation wavelength of 532 nm. The erythrocyte (fixed onto a glass slide) was moved relative to the laser beam with a step size of 0.5 µm by means of a motorized microscope stage. A Raman spectrum was taken at every point. The step size of 0.5 µm corresponds to the wavelength of the laser excitation (532 nm) and is approximately the diffraction limit of the spatial resolution. The spatially resolved Raman spectra differ significantly across the erythrocyte sample. There are some regions, with very intense Raman signals as depicted in Figure 1A and regions with less intense signals (see Figure 1B). The signals in Figure 1A correspond well with the Raman spectrum of extracted hemozoin as shown in Figure 1E while the signals in Figure 1B correspond to the

Malaria Pigment Hemozoin

J. Phys. Chem. B, Vol. 111, No. 37, 2007 11051

Figure 4. Raman spectrum (excitation wavelength 1064 nm) of hemozoin (D) compared with Raman spectra of synthesized β-hematin (A-C). The intensity scale is stretched by a factor of 5 from 100 to 1100 cm-1 for better illustration. β-Hematin was synthesized according to the acidcatalyzed dehydration of hematin (A: propionic acid; B: acetate) and the anhydrous dehydrohalogenation of hemin (C). Crystals with lengths of about 5-30 µm were derived as indicated in the inset (inset C).

Raman spectrum of hemoglobin (see Figure 1F). The spatially resolved Raman signals of hemozoin and hemoglobin of the P. falciparum-infected erythrocytes differ significantly in its intensity patterns. This can be best seen at the mode at 1372 cm-1 which is much more intense in the Raman spectrum of hemozoin as compared to hemoglobin. This Raman band has therefore been used to construct a Raman image of the hemozoin distribution across the P. falciparum-infected erythrocyte, as shown in Figure 1D. The derived spatial distribution of hemozoin with a resolution of 0.5 µm (Figure 1D) corresponds very well to the microscopic image (Figure 1C). Different resonance Raman excitation wavelengths (λexc) have been tested to choose the most appropriate λexc for hemozoin imaging experiments. The results obtained for an excitation wavelength of 633 nm are shown in Figure 2. Like for the experiment with λexc ) 532 nm, a P. falciparum-infected erythrocyte (see microscope picture Figure 2D) has been scanned with a step size of 0.5 µm. While very intense Raman signals of hemozoin (Figure 2A) can be detected at distinct regions of the erythrocyte sample, only very weak background signals can be seen outside the hemozoin inclusions (Figure 2B). The very intense Raman band at 755 cm-1 has been used to construct an image of the hemozoin distribution across the P. falciparum-infected erythrocyte (Figure 2C). Because of the very high signal-to-noise ratio of the in situ Raman signals of hemozoin compared to the signals outside the hemozoin inclusions (pigment vacuole of P. falciparum), the Raman excitation wavelength of 633 nm is ideally suited for imaging hemozoin. However, in order to interpret these complex Raman spectra, it is a necessity to investigate the malaria pigment hemozoin both experimentally and theoretically carefully to understand the molecular vibrations in detail. Raman Spectroscopic Investigation of the Malaria Pigment Hemozoin. Hemozoin was extracted from parasites45 and β-hematin, the synthetic analogue (see ref 38 and references

cited therein) of hemozoin, was synthesized according to the dehydration of hematin46,47 and the anhydrous dehydrohalogenation of hemin.48,49 All samples were inspected by FT-IR spectroscopy prior to the Raman studies and the characteristic IR absorption bands at 1211 cm-1, 1644 cm-1, and 1712 cm-1 58 were detected (results are not shown). The extracted malaria pigment hemozoin was investigated with a variety of resonant and nonresonant Raman excitation wavelengths. The results with λexc ) {1064, 830, 633, 568, 532} nm are shown in Figure 3 and are in good agreement with the spectra from Wood et al.,27 taken from synthetic β-hematin. The Raman spectra of hemozoin (Figure 3) show a very complex enhancement pattern, depending on the applied Raman excitation wavelength. For example, the intensity of the important mode at 1372 cm-1, known as π density marker band, is strongly influenced by the excitation wavelength. This band is enhanced for an excitation wavelength of 532 nm and is even more dominant in the resonance Raman spectrum excited with 568 nm, while it exhibits a much weaker intensity in the 633 nm Raman spectrum. It is completely dominant in the Raman spectrum recorded for an excitation wavelength for 830 nm and is again weaker in the nonresonant Raman spectrum (λexc ) 1064 nm). These interesting Raman spectra of hemozoin are compared to spectra of β-hematin samples as well as spectra of hematin and hemin in more detail in the following text. The nonresonant Raman spectra (λexc ) 1064 nm) of the two acid-catalyzed samples of β-hematin (β-hema_p and β-hema_a) are shown in Figures 4A and 4B together with the spectra of β-hema_b (the sample synthesized according to the anhydrous dehydrohalogenation of hemin) (Figure 4C) and the extracted hemozoin (Figure 4D). The intensity scale in Figure 4 is stretched by a factor of 5 from 100 to 1100 cm-1 for better illustration. The nonresonant spectra of the four samples in Figure 4 are almost identical, proving the assumption that

11052 J. Phys. Chem. B, Vol. 111, No. 37, 2007

Frosch et al.

TABLE 1: Wavenumbers of Experimentally Derived Nonresonant Raman Spectra (λexc ) 1064 nm) of Hemozoin, β-Hema_b, β-Hema_a, and β-Hema_p According to Figure 4. Band Intensities are Weighted as Weak (w), Medium (m), Strong (s), Very Srong (vs), and Shoulder (sh) β-hema_b

hemozoin 107 118 123 132 139 147 153 201 219 229 264 282 306 315 321 343 368 587 632 678 698 711 725 735 755 773 796 821 975 1000 1023 1032 1047 1055 1077 1091 1123 1144 1170 1176 1189 1203 1219 1235 1270 1279 1292 1307 1327 1339 1372 1409 1430 1447 1460 1465 1475 1491 1530 1550 1568 1585 1622 1654

m w w w w w w w w w m w w w w w m w w m w w w w s w w w w m w w w w m m s w w w w sh s sh w w w m w m vs w m w w w w w vs vs vs vs vs sh

107 118 124 130 140 148 155 199 219 231 264 282 305 316 323 344 369 588 629 678 698 712 726 736 755 774 797 821 973 1000 1023 1032 1048 1058 1077 1091 1122 1146 1168 1174 1186 1203 1219 1233 1273 1279 1293 1307 1327 1340 1370 1408 1430 1448 1459 1465 1472 1491 1531 1550 1567 1584 1621 1656

β-hema_a m w w w w w w w w w m w w w w w m w w m w w w w s w w w w m w w w w m m s w w w w sh s sh w w w m w m vs w m w w w w w vs vs vs vs vs w

β-hematin is the synthetic analogue of hemozoin. Figure 4 illustrates that the most intense Raman bands can be seen at 1622 cm-1, 1585 cm-1, 1568 cm-1, 1550 cm-1, 1372 cm-1, and 755 cm-1. A small shoulder is observed at approximately 1654 cm-1. Characteristic Raman bands are also observed in

β-hema_p

107 117 124 131 140 147 154 200 218 230 264

m w w w w w w w w w m

303

w w w w m w w m w w w w s w w w w m w w w w m m s w w w w sh s sh w w w m w m vs w m w w w w w vs vs vs vs vs w

323 344 367 588 630 677 700 712 723 738 755 775 796 821 978 999 1025 1048 1056 1079 1089 1123 1148 1168 1173 1186 (1203) 1220 1237 1275 1281 1288 1307 1324 1343 1370 1410 1430 1449 1458 1466 1469 1490 1528 1550 1567 1585 1622 1656

107 117 124 132 139 148 154 201 222 231 262 280 306 323 344 368 590 (636) 677 700 715 723 737 755 772 795 819 972 998 (1030) 1031 1046 1056 1077 1088 1122 1147 1168 1187 (1203) 1219 (1237) 1276 1282 1306 1322 1339 1370 1409 1430 1446 1461 1467 1477 1491 1527 1549 1567 1584 1622 1656

m w w w w w w w w w m w w w w w m w w m w w w w s w w w w m w w w w m m s w w w w sh s sh w w w m w m vs w m w w w w w vs vs vs vs vs w

the low wavenumber region at 368 cm-1 and 264 cm-1. All Raman signals in Figure 4 are summarized in Table 1.While the β-hematin samples synthesized according to the acidcatalyzed dehydration of hematin show an amorphous, agglutinate morphology, the samples synthesized according to the

Malaria Pigment Hemozoin

J. Phys. Chem. B, Vol. 111, No. 37, 2007 11053

Figure 5. Raman spectra of hemozoin (A), β-hematin, synthesized according to the anhydrous dehydrohalogenation of hemin, (B), hematin (C), and hemin (D) recorded with excitation wavelength 568 nm. While the Raman spectra of hemozoin (A) and β-hematin (B) are identical, they differ significantly from those of hematin (C) and hemin (D). A new Raman band at 1655 cm-1 can be seen in the spectra of hemozoin (A) and β-hematin (B).

anhydrous dehydrohalogenation of hemin show a crystalline morphology with crystal lengths of 5-30 µm, represented in the inset of Figure 4C. A comparison of the Raman pattern of the bands between 400 cm-1 and 250 cm-1 shows some differences. It is suggested that these differences are due to different sample morphology. This assignment is strongly supported by DFT calculations, as discussed in the following subsection. While further analysis is necessary, the result would be very important for a routine control as well as for scientists working in other disciplines. For a further examination of the Raman spectra of hemozoin and β-hematin in comparison to other hematins, like hematin (hematine porcine) and hemin, different resonance Raman excitation wavelengths were applied. Figure 5 compares the resonance Raman spectra of hemozoin (Figure 5A) and β-hema_b (Figure 5B) excited with λexc ) 568 nm with the spectra of hematin (Figure 5C) and hemin (Figure 5D). While the spectra of hemozoin (Figure 5A) and β-hema_b (Figure 5B) are almost identical, significant differences in the Raman intensity pattern can be seen in comparison to hematin (Figure 5C) as well as to hemin (Figure 5D). The spectrum of hemozoin (Figure 5A) is dominated by a very strong Raman peak at 1373 cm-1 while the Raman spectrum of hemin (Figure 5D) shows the most intense band at 1622 cm-1. A different behavior can also be seen when comparing the resonance Raman spectrum of hematin (Figure 5C) and hemozoin (Figure 5A). The hematin spectrum shows almost equal intensity of the Raman peaks at 1622 cm-1, 1565 cm-1, and 1372 cm-1. Also differences in the low wavenumber region (below 400 cm-1) are observed between the different hematins (see Figure 5). These very intense Raman bands below 400 cm-1 in the spectrum of hemin (Figure 5D) support the suggestion derived from Figure 4 that this wavenumber region might be useful for the characterization of the morphology of the samples, because the hemin sample is very crystalline (Teichmann crystals). In addition the low

wavenumber modes in the spectrum of hematin (Figure 5C) are weak, and the hematin sample is not expected to be very crystalline. Furthermore, a Raman band at 1655 cm-1 can be seen in the spectra of hemozoin (Figure 5A) and β-hematin (Figure 5B), which is absent in the spectra of hematin (Figure 5C) and hemin (Figure 5D). Ab Initio (DFT) Calculation of the Hemozoin Central Unit and Mode Assignment. As mentioned above when discussing the Raman mapping experiments on Plasmodium falciparuminfected erythrocytes, a detailed analysis of the Raman spectra, i.e., mode analysis, is absolutely necessary. So far, most of the porphyrin Raman spectra were successfully interpreted according to the very important mode assignments of Abe59 as well as Spiro and co-workers21,60-62 following a D4h symmetry of planar porphyrin systems (and considering perturbations to this approximation due to side groups; e.g. refs 60, 61). However, many new, not understood features are observed in the Raman spectra of the three-dimensional, crystalline structure of hemozoin (Figure 3). One possible explanation might be an aggregation enhancement of the Raman signals due to the intermolecular excitonic interaction between the carboxylate and hydrogenbonded porphyrin units.63 Otherwise, DFT calculations of antimalarials have been very helpful to assign even complex resonance Raman spectra and give valuable insight into the molecular vibrations by analyzing the atomic displacements.31-34 Therefore, the geometry of the hemozoin dimer (which has been shown to have a single Fe(III) high spin environment64), being the elementary cell of hemozoin, was optimized starting from a X-ray structure derived from Pagola et al.38 Subsequently the Raman spectrum was calculated (DFT: BP86/SDD). The experimentally derived nonresonant Raman spectrum of hemozoin and the DFT calculation of the Raman spectrum of the hemozoin dimer are in good agreement to each other (see Figure S1, Supporting Information). Some of the most prominent

11054 J. Phys. Chem. B, Vol. 111, No. 37, 2007

Frosch et al.

Figure 6. Atomic displacements of calculated vibrational modes of hemozoin assigned to the modes at 1654 cm-1 (A), 1622 cm-1 (B), 1568 cm-1 (C), 1372 cm-1 (D). These modes are discussed in detail in the text.

Raman modes are displayed in Figure 6 and are explained in the following. The Raman peak at 1655 cm-1 in the experimental resonance Raman spectrum (λexc ) 568 nm) of hemozoin in Figure 5A (only seen as a small shoulder in the nonresonant spectrum with λexc ) 1064 nm in Figure 4A), and which is totally absent in the resonance Raman spectra of hematin and hemin, is assigned to two normal modes located at the two free propionic side chains of the hemozoin dimer. This band is very weak in the nonresonant Raman spectrum (Figures S1A and S1B), but two normal modes are observed in the calculated stick spectrum at 1660 cm-1. The atomic displacement of one of the two similar molecular vibrations is shown in Figure 6A. The normal mode is assigned to a strong CdO stretching vibration, a (C)OH bending vibration, and a CH2 wagging vibration. Obviously the propionic acid side chains of the hemozoin dimer are important for the formation of the hydrogen linkage and thus for the hemozoin crystals. A resonantly enhanced Raman signal of this mode is therefore a very good explanation of the experimental finding that this mode can only be seen in the hemozoin (see Figure 4). The very strong Raman signal at 1622 cm-1 which is present in the Raman spectra of hemozoin and β-hematin as well as in hematin and hemin can be assigned to localized normal modes at 1622 cm-1. One of these two modes is shown in Figure 6B and reflects a very strong CdC stretching mode of the two vinyl

side chains of one of the two porphyrins (while the second mode is identical and just located at the other porphyrin system). Also a CH2 scissoring (and (C)CH scissoring) is part of this combined vibration. Another normal mode at 1624 cm-1 is a combined stretching vibration in the vinyl group and in the pyrrole part. The band at 1622 cm-1 in the nonresonant Raman spectrum of hemozoin (Figure 4D) is therefore the envelope of these normal modes. Since the vinyl groups are known to have an important impact onto the Raman spectra of porphyrins60,61 and since they are present in hemozoin and β-hematin as well as hematin and hemin, this assignment convincingly describes the experimental findings. The Raman band located at 1568 cm-1 has been shown to be sensitive to interactions of chloroquine and hematin in a hydrous environment29 via a dramatic change in the depolarization ratio pointing toward ordering effects as expected for a stacking of the drug to the porphyrin ring. This Raman peak can be assigned to two normal modes at 1568 cm-1. One of these two modes is shown in Figure 6C and is located at one of the two porphyrins (while the second mode is identical just located at the other porphyrin ring system). This mode is a combined CdC stretching vibration of the porphyrin ring where also the iron-carboxylate propionic side chain is involved. Furthermore CH bendings, CH2 twistings, and CH3 rockings are part of this normal mode. The assignment of this mode gives strong evidence that the recently observed

Malaria Pigment Hemozoin

J. Phys. Chem. B, Vol. 111, No. 37, 2007 11055

change in the depolarization ratio of this mode due to the presence of chloroquine29 is because of a π-π stacking of chloroquine and hematin which is in agreement with NMR studies.19,20 The Raman peak at 1372 cm-1 is strongly influenced by the excitation wavelength as discussed in Figure 3. Obviously, this very complex behavior can only be understood by taking into account the extended, three-dimensional structure of the hemozoin crystal. The Raman peak at 1372 cm-1 is assigned to a normal mode at 1373 cm-1 which is displayed in Figure 6D. The mode consists of strong asymmetric CdC stretching modes, mainly located at the four pyrrole rings marked with the numbers 1-4. The stretching is also extended into the linked propionic acid side chains. Furthermore CH3 scissorings, CH2 scissoringwaggings, and CH bendings are part of this normal mode. The localization of the normal mode at regions where the two porphyrins are stacked to each other provides strong evidence that this Raman band is very much influenced by the extended hemozoin structure. The atomic displacement picture therefore helps to understand and interpret this very complex behavior of the Raman/resonance Raman spectra of hemozoin in terms of the complex three-dimensional structure of hemozoin. Two of the low wavenumber Raman bands at 368 cm-1 and 264 cm-1 are shown in Figures S2E and S2F, respectively. Both modes are very complex consisting of combined out-of-plane C-C stretchings as well as CH, CH2, and CH3 bendings. These modes are very likely to be sensitive to the crystalline morphology of the samples as suggested by the experimental findings displayed in Figure 4. These out-of-plane modes might also be influenced by drug-hemozoin interactions. However, these modes are not very much enhanced in the resonance Raman spectra.

atomic displacement pictures provide valuable insight into the structure of the malaria pigment hemozoin. This is a prerequisite for understanding the Raman spectra and the molecular vibrations of hemozoin and is absolutely necessary for interpreting upcoming Raman monitoring experiments of drug-hemozoin interactions. The investigation of in situ drug-hemozoin experiments in Plasmodium falciparum-infected erythrocytes as well as in Vitro experiments with synthesized β-hematin crystals is currently underway in our laboratories. The experimental findings will be assisted by the assignments and atomic displacement pictures derived from the calculations of the hemozoin structures presented within this paper.

Conclusion and Outlook

(1) http://rbm.who.int/wmr2005. (2) Snow, R. W.; Guerra, C. A.; Noor, A. M.; Myint, Hla Y.; Hay, Simon I. Nature 2005, 434 (7030), 214. (3) Sachs, J.; Malaney, P. Nature 2002, 415 (6872), 680. (4) Allen, R. J.; Kirk, K. Trends Parasitol. 2004, 20 (1), 7; discussion 10-1. (5) Krugliak, M., Zhang, J.; Ginsburg, H. Mol. Biochem. Parasitol. 2002, 119 (2), 249. (6) Lew, V. L., Macdonald, L.; Ginsburg, H.; Krugliak, M.; Tiffert, T. Blood Cells Mol Dis. 2004, 32 (3), 353. (7) Fitch, C. D. Life Sci. 2004, 74, 1957. (8) Hastings, I. M.; Bray, P. G.; Ward, S. A. Science 2002, 298, 74. (9) Wellems, T. E. Science 2002, 298, 124. (10) Ridley, R. G. Nature 2002, 415, 686. (11) Foley, M.; Tilley, L. Pharmacol. Ther. 1998, 79, 55. (12) Francis, S. E.; Sullivan, D. J., Jr.; Goldberg, D. E. Annu. ReV. Microbiol. 1997, 51, 97. (13) Ursos, L. M. B.; Roepe, P. D. Med. Res. ReV., 2002, 22, 465. (14) Wiesner, J.; Ortmann, R.; Jomaa, H.; Schlitzer, M. Angew. Chem. 2003, 115, 5432. (15) Sullivan, D. J., Jr.; Matile, H.; Ridley, R. G.; Goldberg, D. E. J. Biol. Chem. 1998, 273, 31103. (16) Sullivan, D. J., Jr.,; Gluzman, I. Y.; Russell, D. G.; Goldberg, D. E. Proc. Nat. Acacd. Sci. U.S.A., 1996, 93, 11865. (17) Leed, A.; DuBay, K.; Ursos, L. M. B.; Sears, D.; de Dios, A. C.; Roepe, P. D. Biochemistry 2002, 41, 10245. (18) de Dios, V.; Tycko, V.; Ursos, V.; Roepe, P. D. J. Phys. Chem. A 2003, 107, 5821. (19) Constantinidis, I.; Satterlee, J. D. J. Am. Chem. Soc. 1988, 110, 927. (20) Constantinidis, I.; Satterlee, J. D. J. Am. Chem. Soc. 1988, 110, 4391. (21) Spiro, T. G., Ed. Biological Applications of Raman Spectroscopy; Wiley & Sons: New York, 1988; Vol. 1-3 and refs cited therein. (22) McHale, J. L. In Handbook of Vibrational Spectroscopy; Chalmers, J. M., Griffiths P. R., Eds.; 2002; Vol. 1, and refs cited therein. (23) Puppels, G. J.; De Mul, F. F. M.; Otto, C.; Greve, J.; Robert-Nicoud, M.; Arndt-Jovin, D. J.; Jovin, T. M. Nature 1990, 347, 301. (24) Ong, C. W.; Shen, Z. X.; Ang, K. K. H.; Kara, U. A. K.; Tang, S. H. Appl. Spectrosc. 1999, 53 (9), 1097.

This study shows the high potential of Raman mapping experiments of Plasmodium falciparum-infected erythrocytes. It was possible to localize hemozoin with a spatial resolution of 0.5 µm with help of Raman microspectroscopy. The Raman excitation wavelength 633 nm has been demonstrated to provide very intense Raman signals of hemozoin against only a very weak background. While important questions in malaria research are still not resolved, the demonstrated strengths of resonance Raman microscopy for a precise and sensitive localization of the malaria pigment will contribute understanding, e.g., the exact mechanism of hemozoin formation. This knowledge will be of utmost importance for scientists working in other disciplines (e.g., biochemistry and medicine). The malaria pigment hemozoin has been compared to synthesized samples of β-hematin of different morphology. The Raman signals of hemozoin and crystalline β-hematin have been shown to be identical, while the amorphous samples might be distinguished with the help of some low wavenumber bands. The resonance Raman spectra of hemozoin and β-hematin differ from those of hematin and hemin in their intensity pattern and by the appearance of a Raman band at 1655 cm-1. The Raman spectrum of the hemozoin dimer has been calculated ab initio for the first time, following the structure given by Pagola et al.38 With help from this calculation, it was possible to assign the prominent Raman modes of hemozoin. For an interpretation of the complex resonance Raman spectra of hemozoin, the extended, threedimensional structure of hemozoin has to be taken into account. The derived assignment has also been shown to be useful for an interpretation of drug-hematin interactions.29 Altogether the experimentally derived Raman spectra as well as the calculated

Acknowledgment. The authors gratefully acknowledge support from the Fonds der Chemischen Industrie and the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 630 “Recognition, Preparation, and Functional Analysis of Agents Against Infectious Diseases”, project C1 and grant BE1540/4-4 to KB). The authors wish to acknowledge that computations were run at ZIH of TU Dresden and for the technical assistance of Elisabeth Fischer in cell culture. Supporting Information Available: Figure S1: Comparison of the DFT calculation of the Raman spectrum of the hemozoin dimer and the experimental nonresonant Raman spectrum (λexc ) 1064 nm) of hemozoin. Figure S2: Atomic displacements of calculated vibrational modes of hemozoin assigned to the modes at 368 cm-1 (S2E) and 264 cm-1 (S2F). This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes

11056 J. Phys. Chem. B, Vol. 111, No. 37, 2007 (25) Ong, C. W.; Shen, Z. X.; Ang, K. K. H.; Kara, U. A. K.; Tang, S. H. Appl. Spectrosc. 2002, 56 (9), 1126. (26) Wood, B. R.; Langford, S. J.; Cooke, B. M.; Glenister, F. K.; Lim, J.; McNaughton, D. FEBS Lett. 2003, 554 (3), 247. (27) Wood, B. R.; Langford, S. J.; Cooke, B. M.; Lim, J.; Glenister, F. K.; Duriska, M:; Unthank, J. K.; McNaughton, D. J. Am. Chem. Soc. 2004, 126 (30), 9233. (28) Wood, B. R.; McNaughton, D. Expert ReV. Proteomics 2006, 3 (5), 525-544. (29) Frosch, T.; Ku¨stner, B.; Schlu¨cker, S.; Szeghalmi, A.; Schmitt, M.; Kiefer, W.; Popp, J. J. Raman Spectrosc. 2004, 35, 819. (30) Frosch, T.; Meyer, T.; Schmitt, M.; Popp, J. Anal. Chem. 2007, published on Web. (31) Frosch, T.; Schmitt, M.; Popp, J. Anal. Bioanal. Chem. 2007, 387, 1749. (32) Frosch, T.; Schmitt, M.; Popp, J. J. Phys. Chem. B 2007, 111 (16), 4171-4177. (33) Frosch, T.; Schmitt, M.; Noll, T.; Bringmann, G.; Schenzel, K.; Popp, J. Anal. Chem. 2007, 79 (3), 986. (34) Frosch, T.; Schmitt, M.; Bringmann, G.; Kiefer, W.; Popp, J. J. Phys. Chem. B 2007, 111 (7), 1815. (35) Olivaro, P. L.; Goldberg, D. E. Parasitology Today, 1995, 11, 294; Hawley, S. R.; Bray, P. G.; O’Neill, P. M.; Park, B. K.; Ward, S. A. Biochem. Pharmacol. 1996, 52, 723. (36) Yayon, A.; Cabantchik, Z. I.; Ginsburg, H. EMBO J. 1984, 3, 2695. (37) Ginsburg, H.; Geary, T. G. Biochem. Pharmacol. 1987, 36, 1567. (38) Pagola, S.; Stephens, P. W.; Bohle, D. S.; Kosar, A. D.; Madsen, S. K. Nature 2000, 404, 307. (39) Buller, R.; Peterson, M., L.; Almarsson, O ¨ .; Leisirowitz, L. Cryst. Growth Des. 2002, 2, 553. (40) Solomonov, I.; Osipova, M; Feldman, Y.; Baehtz, C.; Kjaer, K.; Robinson, I. K.; Webster, G. T.; McNaughton, D.; Wood, B.; Weissbuch, I.; Leiserowitz, L. J. Am. Chem. Soc. 2007, 129 (9), 2615. (41) Trager, W.; Jensen, J. B. Science 1976, 193, 673. (42) Cranmer, S. L.; Magowan, C.; Liang, J.; Coppel, R. L.; Cooke, B. M. Trans. R. Soc. Trop. Med. Hyg. 1997, 91, 363. (43) Lambros, C., Vanderberg, J. P. J. Parasitol. 1979, 65, 418-420. (44) Pasvol, G.; Wilson, R. J.; Smalley, M. E.; Brown, J. Ann. Trop. Med. Parasitol. 1978, 72, 87. (45) Jaramillo, M.; Gowda, D. C.; Radzioch, D.; Olivier, M. J. Immunol. 2003, 171, 4243. (46) Bohle, D. S., Conklin, B. J., Cox, D., Madson, S. K., Paulson, S., Stephens, P. W., Yee, G. T. ACS Symp. Ser. 1994, 572, 497, and refs. therein. (47) Egan, T. J.; Mavuso, W. W.; Ncokazi, K. K. Biochemistry 2001, 40, 204.

Frosch et al. (48) Bohle, D. S.; Helms, J. B. Biochem. Biophys. Res. Commun. 1993, 193, 504. (49) Bohle, D. S.; Kosar, A. D.; Stephens, P. W. Acta Crystallogr. D 2002, 58, 1752. (50) Gaussian 03, Revision C.02, Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.; Gaussian, Inc., Wallingford CT, 2004. (51) Becke, A. D. Phys. ReV. A 1988, 38, 3098. (52) Perdew, J. P.; Yue, W. Phys. ReV. B 1986, 33, 8800. (53) Neugebauer, J.; Hess, B. A. J. Chem. Phys. 2003, 118, 7215. (54) Dunning, T. H., Jr.; Hay, P. J. Modern Theoretical Chemistry; Plenum: New York, 1976, and refs cited therein. (55) Schaefer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829. (56) Long, D. A. The Raman Effect; Wiley: New York, 2002. (57) Chalmers, J. M.; Griffiths P. R., Eds. Handbook of Vibrational Spectroscopy; 2002. (58) Slater, A. F. G.; Swiggard, W. J.; Orton, B. R.; Flitter, W. D.; Goldberg, D. E.; Cerami, A.; Henderson, G. B. Proc. Nat. Acad. Sci. U.S.A. 1991, 88, 325. (59) Abe, M.; Kitagawa, T.; Kyogoku, Y. J. Chem. Phys. 1978, 69, 4526. (60) Choi, S.; Spiro, T. G.; Langry, K. C.; Smith, K. M. J. Am. Chem. Soc. 1982, 104, 4337. (61) Choi, S.; Spiro, T. G.; Langry, K. C.; Smith, K. M.; Budd, D. L.; La Mar, G. N. J. Am. Chem. Soc. 1982, 104, 4345. (62) Hu, S.; Smith, K. M.; Spiro, T. G. J. Am. Chem. Soc. 1996, 118, 12638. (63) Akins, D. L.; O ¨ zcelic, S.; Zhu, H. R.; Guo, C. J. Phys. Chem. 1997, 101, 3251. (64) Sienkiewicz, A.; Krystek, J.; Vileno, B.; Chatain, G.; Kosar, A. J.; Bohle, D. S.; Forro, L. J. Am. Chem. Soc. 2006, 128, 4534.