Cytochemical techniques and energy-filtering transmission electron microscopy applied to the study of parasitic protozoa.

The study of parasitic protozoa plays a major role in cell biology, biochemistry and molecular biology. Numerous cytochemical techniques have been developed in order to unequivocally identify the nature of subcellular compartments. Enzyme and immuno-cytochemistry allow the detection of, respectively, enzymatic activity products and antigens in particular sites within the cell. Energy-filtering transmission electron microscopy permits the detection of specific elements within such compartments. These approaches are particularly useful for studies employing antimicrobial agents where cellular compartments may be destroyed or remarkably altered and thus hardly identified by standard methods of observation. In this regard cytochemical and spectroscopic techniques provide valuable data allowing the determination of the mechanisms of action of such compounds.

Introduction

Parasitic protozoa comprise valuable tools in cell biology, providing excellent and often unique experimental models (e.g. 1, 2). The use of microbial pathogens to approach basic questions in cell biology is considered a new science termed 'cellular microbiology' (3, 4). The first living cells observed under a microscope were most probably parasitic protozoa (Giardia lamblia, Opalina, Nyctotherus and coccid parasites of rabbits) as well as bacteria collected by Antoni van Leeuwenhoek. Knowledge about cell functioning and structure relied largely on the development of microscopic approaches and the protozoology of parasites plays a pivotal role in the advancement not only of parasitology, but also of cell biology, biochemistry and molecular biology.

Ultrastructural And Cytochemical Study Of Parasitic Protozoa

The study of parasitic protozoa using light and transmission electron microscopy has often been faced with difficulties in subcellular compartment identification. Organelles with similar structural features but rather distinct compositions were frequently subsumed under the same designation and vice versa. Therefore the identification of a compartment by routine TEM may be considered purely arbitrary (5). We have previously studied the effects of the putrescine analogue 2,4-diamino butanone on the trichomonad parasite Tritrichomonas foetus. The drug remarkably reduced protozoan proliferation in vitro and led to profound alterations in the redox organelles termed hydrogenosomes. These organelles are important in the chemotherapy of trichomoniasis since the drugs of choice, the 5-nitroimidazoles such as metronidazole (flagyl®) are activated by reduction within these compartments. The TEM of the drug-treated parasites strongly suggested that the organelles were drastically altered and destroyed after prolonged incubations. We were only able to confirm the hydrogenosomal destruction by the putrescine analogue by the use of immunogold cytochemistry employing a monoclonal antibody to a marker enzyme i.e. β-succinyl-CoA synthetase (6).

Sterol biosynthesis inhibitors (SBI) are potent antifungal compounds that comprise important candidates for the chemotherapy of Chagas disease and leishmaniasis (7-10). We have previously noticed that SBI such as ketoconazole and terbinafine induced uncontrolled autophagy in Leishmania amazonensis parasites (8). The cytochemical detection of acid phosphatase was employed to confirm the autophagic nature of multivesicular bodies observed in SBI-treated parasites. More recently we used cytochemistry and electron energy-loss spectroscopy (EELS) to further address the SBI effects on Leishmania (11). In this report, we demonstrate that cytochemistry and energy-loss spectroscopy are valuable tools in protozoology.

Energy-Filtering Transmission Electron Microscopy

Energy-filtering transmission electron microscopy has become a powerful tool to study the element distribution in cellular compartments. This technique allows the calculation of element distribution images in a sample with high spatial resolution and sensitivity (12). The underlying principle of this technique is the fact that beam electrons in a transmission electron microscope lose characteristic or element-specific energies during inelastic collisions with the sample atoms. The electron energy-loss spectrum shows the features of this interaction with the specimen.

The excitation events of inner-shell electrons of a certain atom that compose the specimen appear in the spectrum as an increase in electron counts in the core-loss ionization edge followed by a progressive decrease with the increasing energy loss. Every ionization edge is superimposed on a spectral background due to other energy-loss processes. To obtain an image that represents the characteristic energy-loss electrons, and consequently reflects the elemental distribution in the specimen, the background contribution must be subtracted. Various methods have been developed to obtain the element distribution in a specimen. In the spectrum-imaging and image-EELS techniques, a three-dimensional dataset of the electron energy-loss spectrum is collected for each pixel of the analysed field. In spectra imaging in a scanning transmission electron microscope, a small area of the sample is analysed and a parallel EELS spectrum is acquired for each point. Alternatively, the element distribution can be calculated from a small number of images acquired at defined energy losses. The images are obtained in a fixed beam transmission electron microscope that contains either an in-column or a post-column filter (13). The filter separates electrons according to their energy losses after interaction with the specimens and a slit is used to define an energy window that will contribute to the recorded image. The image is recorded with a camera at energy losses related to the element to be mapped and sorted in the computer' memory for future calculations. To obtain an elemental map with energy-filtered images it is necessary to remove the non-element-specific background from the specific signal of the ionization edge.

Thickness of the sample is an important parameter to be considered in electron energy-loss spectroscopy of sections of biological specimens. This is mainly because, when an accelerated beam electron passes through a specimen, it is possible that it is scattered twice or even several times. The probability of plural or multiple scattering increases in proportion to the specimen thickness and rising atomic number or density. All scattering combinations can occur (elastic plus elastic, elastic plus inelastic or multiple inelastic) and multiple-scattered electrons lose energy with each inelastic impact. The information content of the element-characteristic single scatter electrons is greatly reduced when multiple scatter increases. Multiple scattered electrons provide no information value for energy loss analysis, especially for elemental mapping. The more multiple scatter a sample may present, the less reliable the analysis for the element-characteristic single scatter. This leads to a considerable reduction of the signal-to-noise ratio and a decrease in detectability. Thus, the section thickness of a specimen for acquisition of an electron energy-loss spectrum or calculation of a elemental map should be optimised for best results.

Different protocols are used to estimate the background and identify the element signal to calculate the elemental map from the element-enhanced image. The most common are the two- and three-window methods. In the two-window method, a background image is a pre-edge image that is subtracted (difference method) or divided (jump-ratio method) from the element-enhanced image. The two-window method gives elemental maps with high signal to noise ratios because it uses only two images instead of three as in the case of the three-window-power-law method. The three-window-power-law method is the most conventional method for background subtraction and calculation of elemental maps. In this case, the background is estimated by a power-law function I = A . E-r, where I is the electron count, E is the energy loss and A and r are the two independent fitting parameters (13). To calculate the elemental map with this method, two images below and one image above the ionization edge are acquired. Background correction is done by extrapolating the background image according to the power function model. Both A and r vary within the samples as a result of variations in the thickness and composition and a separate estimation of the background is required at each pixel. To do this, extrapolated background is subtracted pixel by pixel from the element enhanced edge (the image above the ionization edge). The net result of this subtraction is considered the elemental map for that element. Variations of the two- and three-window methods were developed (12).

Materials and Methods

Cytochemical Identification Of Leishmania Subcellular Compartments

In order to identify the parasite organelles, we performed the detection of acid phosphatase (14). This technique is based on the cerium chloride-mediated capture of liberated phosphate groups in the presence of a phosphatase substrate such as glycerophosphate or p-nitrophenylphosphate (p-NPP). The use of p-NPP poses the advantage of having as a product p-nitrophenol, which is easily detected spectrophotometrically allowing a parallel quantitative approach. Nevertheless the use of p-NPP is associated with its non-enzymatic hydrolysis in the presence of cerium chloride.

The cells were fixed in 1% glutaraldehyde (type I, Sigma, St. Louis) in 0.1M sodium cacodylate buffer, pH 7.2, for 10 min. at 4°C. Afterwards they were washed twice in the same buffer and once in Tris-HCl buffer, pH 5.0. The cells were then incubated for 60 min. in Tris-HCl buffer, pH 5.0 with 1mM sodium-β-glycerophosphate, 5% (w/v) sucrose and 2mM CeCl3. After that, the cells were washed in the same buffer and fixed in 2.5% glutaraldehyde in cacodylate buffer, pH 7.2, for 60 min at room temperature and processed for TEM.

Endocytic compartments can be detected by the use of tracers such as peroxidase and gold-labelled proteins that can be readily identified by electrondense reaction product precipitates and characteristic particles, respectively. We employed horseradish peroxidase (HRP) as a fluid-phase endocytic tracer. Living cells were incubated with HRP for 60-120 min. and then with 500 μg/ml 3:3'-diaminobenzidine (DAB), and 0.01 % (w/v) hydrogen peroxide (H2O2) as substrates for more 60 min. at room temperature. Afterwards the cells were processed for TEM. Endocytic compartments can be also traced by incubation of living cells with cationized ferritin, which binds to anionogenic sites on the cell surface. Living Leishmania amazonensis promastigotes were incubated with 10 mg/ml cationized ferritin for 60-120 min. and then washed, fixed in 2.5% glutaraldehyde and processed for TEM.

Subcelullar compartments can also be studied by the use of gold-conjugated lectins. We incubated thin sections of lowicryl K4M-embedded parasites with several gold-labelled lectins (15). For this procedure, the cells were fixed in 1% glutaraldehyde, 4% formaldehyde (freshly prepared) in cacodylate buffer, pH 7.2, for 60 min. at 4°C, dehydrated in methanol and embedded in lowicryl K4M by the progressive lowering of temperature method (16).

Gold-labelled transferrin and cystatin C were also employed as endocytic tracers. Living cells were incubated with the gold-conjugated proteins for 60 min. and then processed for TEM.

Elemental Mapping By Energy-Filtered Transmission Electron Microscopy

In order to detect the elements present in subcellular structures of the parasites, we prepared cells for electron microscopy. The most prevalent elements in the spectra, which we decided to map, were calcium, oxygen and phosphorus. Cells were fixed in 2.5 % glutaraldehyde in cacodylate buffer 0.1 M, pH 7.2, for 2 h at room temperature, washed in the same buffer, post-fixed in a buffered 1% OsO4 solution for 20 minutes, washed in buffer, dehydrated through an acetone series and then embedded and polymerized in PolyBed 812 resin. Unstained 30-50 nm-thick sections were analysed with a Zeiss CEM902 transmission electron microscope equipped with an IBAS (Kontron Elektronik GmbH) image analysis system. For elemental mapping, the two or three window method was used to map phosphorus (post-edge: 150 eV; pre-edge: 100 and 110 eV; energy-selecting window of 20 eV), calcium (post-edge: 360 eV; pre-edge: 330 eV; energy-selecting window of 20 eV and O (post-edge: 545 eV; pre-edge: 515 and 490 eV; energy-selecting window of 20 eV). The objective aperture was 90 μm (17 mrad approximately) and the accelerating voltage was 80 kV. Elemental maps were directly photographed from the host computer monitor. For electron energy-loss spectroscopy, the microscope was operated in spectrum mode with an objective aperture of 30 μm (5.8 mrad approximately) and accelerating voltage of 80 kV. The energy-selecting window was of approximately 2 eV. Electron counts were measured at 2 eV intervals by a digital multimeter connected to a photomultiplier attached to the microscope. The output readings were fed into the computer memory and processed by the software developed by Kontron Elektronik GmbH.

Results and Discussion

Ultrastructural Cytochemistry Of Leishmania

Cytochemical approaches have allowed the bona fide identification of subcellular compartments within whole cells. Cell organelles such as nucleus and mitochondria are readily identified under TEM. Nevertheless, other less conspicuous or drug-modified compartments cannot be unequivocally recognised simply by structural features (6).

Different protozoa species display compartments termed acidocalcisomes, which are acidic and store calcium (17). Nevertheless this definition is mainly based on functional assays and methodological artefacts have hampered the identification of these organelles in TEM.

Since we have previously noticed (8) that Leishmania parasites treated with the antifungal compounds presented structural alterations that included increased numbers of organelles resembling the acidocalcisomes (Figs. 1a and b), we decided to further approach this phenomenon. We employed cytochemical methods to characterize the SBI-treated parasites. Although these compartments are acidic and accumulate calcium, experiments carried out using other protozoa indicate that they do not take part in the endocytic pathway (18, 19). Nevertheless different endocytic tracers and long incubation periods should be tested. The observation of membranous, material could be indicative of a presumably autophagic nature to this compartment and the cytochemical detection of acid phosphatase (Fig. 1c) activity would corroborate this inference. Nevertheless these compartments present high pyrophosphatase activity (these enzymes may also be detected by cerium-based cytochemical procedures) and considerable amounts of inorganic phosphates (20). Thus the acid phosphatase detection within this compartment may be misleading unless suitable controls are carried out.

Fig. 1

A) TEM of control Leishmania amazonensis promastigotes displaying the normal organelles, including some acidocalcisomes (arrowheads) N - nucleus, K - kinetoplast.; B) Terbinafine-treated promastigote showing numerous and large acidocalcisomes (Arrows).; C) Acid phosphatase detection in L. amazonensis acidocalcisome (arrowheads) Note the presence of membrane unit lining the organelle core (arrows). D) HRP activity detection in acidocalcisome (arrowheads); Scale bars indicate 1 μm (A-B) and 0.25 μm (C-D); (reproduced from (11) with publisher permission).

Interestingly under such conditions it was possible to observe a membrane unit lining the organelle core (Fig. 1c).

The use of endocytic tracers makes possible the identification of compartments of the endosomal/lysosomal/autophagic pathway in different cell types. The use of peroxidase and DAB in ultrastructural cytochemistry was introduced by Graham and Karnovsky (21, 22) to trace endocytosis but it was also widely employed for the demonstration of catalase and cytochrome oxidase as well as for detection of endogenous peroxidase activities. Heme-derivatives can also mediate the oxidative polymerization of DAB. This would not be a problem here since trypanosomatid parasites do not produce heme groups, which therefore must be incorporated from the medium. Thus if ingested these groups would also be found in endocytic compartments. The oxidation leads to DAB polymerization via nitrogen atoms and the resulting polymer is highly osmiophilic and insoluble, therefore allowing easy identification of electrondense precipitates of the enzymatic reaction product with minimal diffusion artefacts.

The incubation of SBI-treated and control parasites with HRP as an endocytic tracer resulted in the acidocalcisome labelling (Fig. 1d).

Living Leishmania parasites incubated with cationized ferritin presented labelled acidocalcisomes (Fig. 2a). Similarly, we have previously observed (15) the Leishmania acidocalcisome staining (Fig. 2b) by the gold-labelled lectin from Limax flavus (specific for neuraminic acid).

Fig. 2

Living Leishmania parasites were incubated with the endocytic tracers: A) cationized ferritin; C) transferrin-Au and D) cystatin C-Au, for 60-120 min or fixed and processed for post-embedding cytochemistry (see methodology section) and the Limax flavus agglutinin-binding sites (B) were detected using the gold conjugated lectin. In the conditions tested here labelling was observed in acidocalcisome-like compartments. Scale bars indicate 0.1 μm (A), 0.2 μm (B) and 0.25 μm (C-D); (reproduced from (11) and (15) with publishers permission).

Gold-labelled proteins can also be used successfully as endocytic tracers (23). They pose the remarkable advantage of allowing the use of different proteins conjugated with different particle diameters but it must be kept in mind that the gold conjugation, particularly with particles larger than 5nm, may alter the endocytic pathway of a protein such as transferrin (24). In such cases, the protein may be followed by immunocytochemistry (e.g. 25, 26). Interestingly, immunocytochemical detection of one or more proteins may be performed simultaneously with the cytochemical detection of an enzyme (e.g. 27).

Here we employed gold-conjugated transferrin and cystatin C. In both cases acidocalcisome-like organelles were labelled (Figs. 2c and d), indicating a possible endosomal/autophagic nature. Unfortunately, we presently cannot exclude the possibility of mistaking other compartments with structural similarity.

Application Of The Eftem To Microbiological And Chemotherapeutical Studies In Leishmania

In our previous report (11), electron spectroscopic imaging was used to map phosphorous, oxygen and calcium in thin sections of polybed-embedded parasites. Inelastically scattered electrons with element-specific energy losses were used to determine the distribution of several elements in Leishmania promastigotes before and after cultivation of the SBI ketoconazole and terbinafine. Both the electron energy-loss spectra and elemental maps of the acidocalcisomes revealed the presence of phosphorous, oxygen and calcium.

Elemental mapping of acidocalcisomes showed a homogenous distribution of phosphorous (Fig. 3b), calcium (Fig. 3d) and oxygen (Fig. 3c) all over the organelle dense core (Fig. 3a). Electron energy-loss representative spectra confirmed the presence of these elements by demonstrating calcium L2,3 (346 eV), oxygen K (532 eV) and nitrogen (402 eV) edges (Fig. 4a). The L2,3 edge for phosphorous (132 eV; Fig. 4b) and occasionally zinc (not shown) were also detected in this organelle. We cannot exclude the possibility that other elements were washed out during sample processing and/or sectioning. The detection of P and O confirm the organelles studied were indeed the acidocalcisomes and phosphate groups could be found within these compartments due to the polyphosphate and/or pyrophosphate breakdown (20) or to the degradation of material derived from the endocytic/autophagic pathway (11).

Fig. 3

Elemental mapping (b-d) of an L. amazonensis acidocalcisome shown in (a). The phosphorus (b) and oxygen (c) maps were calculated by the three-window methods from two energies below and one above the energy edge for each element. The calcium map (d) was calculated by the two-window method using one energy below and one above the energy edge of that element. Scale bar indicates 0.5 μm (reproduced from (11) with publisher permission).

Electron energy loss spectra of an acidocalciosome showing (a) the L2,3 edge of the element calcium (346 eV) and K edges of the elements nitrogen (402 eV) and oxygen (532 eV) and (b) the L2,3 edge of the element phosphorus (132 eV). Insets are the derivatives of the spectra shown, highlighting the edges of the elements. (reproduced from (11) with publisher permission).