The plasma membrane of microaerophilic protists: oxidative and nitrosative stress

The plasma membrane is the chemiosmotic barrier that provides the interface between the organism and its external environment. Across this phospholipid bilayer, a transmembrane electrochemical potential (negative inside) plays a pivotal role in the control of solute exchange (Mitchell, 1979; Harold, 1986; Smith, 1990). The most important monovalent ions implicated in gradients across this layer include Na⁺, K⁺, H⁺ and Cl and substrate anions. Divalent cations, especially Ca²⁺, are also important. Maintenance of this energy-requiring disequilibrium is by the action of specific electroneutral pumps (e.g. Na⁺/H⁺ symports), channels (e.g. those that facilitate K⁺ and Ca²⁺ transport) and electrogenic pumps (H⁺-translocating ATPases). It has also been suggested that NADH oxidase located on the outside of this membrane may play a key role (Morré et al., 1987). Little is known about the mechanisms by which the plasma-membrane potential is generated or regulated in anaerobic protists, apart from in the trophozoite of Giardia intestinalis, where it has been shown (Biagini et al., 2000) that K⁺, but not Na⁺, pumps acting alongside electrogenic H⁺ pumps are responsible for the normal maintenance of this membrane at 134 mV. This value was determined by the use of the oxonol probe bis(1,3-dibarbituric acid)-trimethine oxonol [DiBAC₄(3)], using a flow cytometric method, in which fluorophore uptake over a range of external dye concentrations was calibrated against fluorophore uptake (at zero potential) by dead organisms (Krasznai et al., 1995; Emri et al., 1998). Na⁺ does, however, play a role in pH regulation (Biagini et al., 2001b).

As the environmental interface for chemiosmotically driven processes, the plasma membrane senses and indicates physiochemical changes in the liquid phase of its immediate surroundings. Thus, perturbation of the plasma-membrane potential provides a sensitive and rapid indication of those stimuli likely to lead to functionally important physiological changes (Humphreys et al., 1994; Lloyd & Hayes, 1995; Lloyd et al., 2000). Where these effects are reversible, modification of membrane potential will be observable as a transient (often oscillatory) response. Irreversible effects elicit more profound disturbance (Scott & Rabito, 1988), often leading to cytotoxic cascades and even cell death by necrotic or apoptotic pathways (Ryu & Lloyd, 1995; Fernandes & Assreuy, 1997; Lloyd et al., 2003b).

In this study, we focus on the effects of reactive oxygen species (Lloyd et al., 2000), as well as of nitrosative stress (Lloyd et al., 2003a), as two of the most important sources of metabolic perturbation and cellular damage. These agents are implicated in significant early changes in the plasma membrane (Scott & Rabito, 1988), and thereby the onset of cellular ageing, senescence and cell death. In microaerophilic organisms, it would be expected that these sources of cellular stress might play major roles in cytotoxicity and that targets on the cytoplasmic membrane would be the primary sites of free-radical-mediated effects (Kayahara et al., 1998). Here, we demonstrate that fluorochrome-based measurements of plasma-membrane potential provide a sensitive and useful approach to the monitoring of cellular stress in these organisms as has previously been shown for a yeast (Dinsdale et al., 1995) and for bacteria (Mason et al., 1994; Lloyd et al., 2001).

Organisms and cultivation methods.
G. intestinalis (Portland1 strain) was grown as described previously on bile- and serum-containing medium (Keister, 1983). Trichomonas vaginalis and Tritrichomonas foetus KV1 were grown on Diamond's (Diamond, 1957) trypticase yeast extract/maltose medium with 10 % (v/v) heat-inactivated horse serum. Hexamita inflata, isolated by E. A. Meyer (while working with J. Kulda) from a freshwater lake, was grown axenically at 25 °C in medium containing 2 % (w/v) yeast extract, 0·5 % (w/v) maltose, 1 % L-cysteine, 10 mM potassium phosphate buffer and 10 % (w/v) heat-inactivated horse serum at pH 7·2. Mastigamoeba punctachora was isolated by C. Bernard from a freshwater pond near Sydney, Australia and grown as mixed cultures with bacteria and a boiled wheat grain at 22 °C. All organisms were grown in screw-capped tubes or culture flasks.

Organisms were grown to late-exponential phase before being harvested by centrifugation for 2 min at 650 g_av in an MSE bench centrifuge. Resuspension was in 150 mM NaCl, 5 mM K₂HPO₄ and 1·8 mM KH₂PO₄ PBS (pH 7·2), unless otherwise stated (e.g. where effects of K⁺ were studied, the buffer used was 140 mM N-methyl-D-glucamine chloride, 1 mM CaCl₂, 10 mM HEPES and 11 mM glucose).

Fluorescence microscopy.
Cells were examined using an Olympus BH2 triocular fluorescence microscope. Images were acquired on Fuji ISO400 (daylight) 38 mm film.

Confocal laser-scanning microscopy.
A Bio-Rad MRC confocal system attached to a research microscope (1024-Leica DMRB) was used with an argon/krypton air-cooled laser (emission at all lines, i.e. UV, 488 and 514 nm). Images were obtained with a x63 oil-immersion objective (Numerical Aperture 1·38). Section thickness was 5·5 µm. The 0·3 W laser was used at 10 % power to minimize photobleaching. Unless otherwise stated, organisms were washed and resuspended in 0·31 M mannitol before observation using FITC filters. Images were acquired on a Zip disk and printed using an Epson 750 colour printer.

Flow cytometry.
Cellular fluorescence (green emission, 530540 nm) was monitored by flow cytometry using a Mo-Flo cytometer (Cytomation PTY) with excitation at 488 nm from a water-cooled 200 mW argon-ion laser. In addition, forward light scatter and right-angle side scatter were measured and used for gating data collection. Typically, signals from 50 000 cells were acquired and analysed using CYCLOPS software (Cytomation PTY) for each sample. The flow cytometric histograms shown are representative of at least three independent experiments. For non-axenic cultures of M. punctachora, a selected population was analysed to exclude the contribution of signals from bacteria; cell sorting was used to validate the purity of the cohort selected.

Materials.
DiBAC₄(3) was from Molecular Probes (catalogue no. B-436); it was stored at 18 °C in the dark as a 100 µM ethanolic solution. Roussin's black salt, NH₄[Fe₄S₃(NO)₇], was synthesized in house. Gramicidin, NaNO₂, sodium nitroprusside, Na₂Fe(CN₅)NO and all other chemicals were from Sigma.

Fluorescence and confocal scanning microscopy
Fig. 1 shows the effects of oxidative and nitrosative stress on oxonol permeability of some microaerophilic protists. In G. intestinalis, DiBAC₄(3) is excluded from viable organisms (Fig. 1a), but exposure of washed organisms in air-saturated PBS for 18 h at 4 °C collapses the plasma-membrane potential and permits entry of the fluorophore (Fig. 1b). Viable Trichomonas vaginalis organisms (transmitted illumination, Fig. 1c) are seen with fluorescent haloes' (UV illumination, Fig. 1d); the dye accumulates in the periphery of the organism, but does not penetrate the plasma membrane, unless exposed to air for 3 h, or to nitrosative stress (Fig. 1f, 400 µM Roussin's black salt for 10 min). In Fig. 1(e) (confocal scanning microscopy, excitation with all laser lines), Tritrichomonas foetus hydrogenosomes show autofluorescence and, after nitrosation, the organisms become spherical and very swollen. M. punctachora (Fig. 1g, h) and H. inflata (Fig. 1i, j) also take up the oxonol fluorophore after exposure to air for 2 h. The aggregation of H. inflata in the presence of O₂ is a typical oxidative stress response in the organism (Biagini et al., 1997a).

(76K): Fig. 1. Micrographs of microaerophilic protists incubated with 1 µM DiBAC4(3). (a, b) G. intestinalis. Organisms were viewed under transmitted light (a) or UV light (b); fluorophore uptake was only evident after exposure to atmospheric O2 for 18 h at 4 °C. (cf) Tritrichomonas foetus. Under UV illumination organisms appear to have fluorescent haloes' due to fluorophore accumulation outside the plasma membrane (d). (e) Excitation at 488 nm in the confocal scanning microscope of a control incubate (no fluorophore) reveals autofluorescent organelles. (f) DiBAC4(3) fluorescence in 400 µM Roussin's black salt-treated Tritrichomonas foetus. (g, h) M. punctachora. Fluorescence micrographs of organisms incubated with DiBAC4(3) before and after exposure to air for 2 h. (i) H. inflata. Live organisms crowd together to minimize O2 toxicity (i) and after 2 h exposure at an air interface have become loaded with fluorophore (j). For details of fluorescence microscopy and confocal laser-scanning microscope see Methods.

Flow cytometry of DiBAC₄(3) permeability
Flow cytometry of G. intestinalis incubated for 50 min with 1 µM DiBAC₄(3) in PBS showed three subpopulations: after exposure to 100 µM O₂ for 6 h (Fig. 2a, b) intact, weakly fluorescent organisms accounted for 65 % of the total, 32 % were damaged and more highly loaded with the fluorophore, and 12 % were even more highly fluorescent and therefore probably non-viable. Microscopic examination of sorted populations confirmed that most of the organisms in the weakly fluorescent cohort were still highly motile, but that the organisms in the other two subpopulations were not. Heat treatment (60 °C for 3 min) resulted in virtually complete loss of viability (as indicated by loss of motility) and an increase in the most highly fluorescent population (to 77 % of the total); only about 1 % of weakly fluorescent (viable) organisms remained (Fig. 2c, d).

(30K): Fig. 2. Flow cytometry of G. intestinalis. Resolution of the population of DiBAC4(3)-treated organisms using gating of fluorescence intensity. (a, b) Analysis after 6 h exposure of organisms to 100 µM O2 in PBS, showing three subpopulations. (c, d) Heat-killed organisms. (a, c) Side scatter versus forward scatter plots, and (b, d) fluorescence intensity distributions. Details of the Mo-Flo cytometer employed are provided in Methods. SSC, side scatter; FSC, forward scatter; FL2, emission intensity at 580 nm.

Fig. 3 shows flow cytometric analyses of DiBAC₄(3)-treated Tritrichomonas foetus. Fluorescence emission of less than 10 arbitrary intensity units arose from very small particles in the medium, and autofluorescence (10100 units) (Fig. 3a) was likely to be from reduced nicotinamide nucleotides and oxidized flavins. Cellular fluorescence shows as a weak signal (Fig. 3b), until the plasma-membrane potential was collapsed (e.g. by exposure to a nitrosating agent). NaNO₂ (2·2 mM) was more effective than 7·5 mM sodium nitroprusside (Fig. 3d, e, respectively).

(11K): Fig. 3. Flow cytometry of Tritrichomonas foetus. (a) Control incubation, no fluorophore. (be) Organisms incubated with 1 µM DiBAC4(3) (b), together with NaNO2 at 90 µM (c) or 2·2 mM (d). In (e), the nitrosating agent was sodium nitroprusside (7·5 µM).

Although M. punctachora was grown in the presence of bacteria, the flow cytometric gating enabled measurements to be made selectively on the protist population (Fig. 4); cell sorting subsequently confirmed the identity of the gated cohort. The ATPase inhibitor N,N'-dicyclohexylcarbodiimide (DCCD) or the protonophore carbonyl cyanide m-chlorophenylhydrazone (CCCP) gave shifts of DiBAC₄(3) fluorescence to increasing intensities (Fig. 5). However, other V-type or P-type ATPase inhibitors, bafilomycin (40 µM) or vanadate (100 µM), respectively, were without effects. Similar effects to those with DCCD or CCCP were produced by the ionophore gramicidin (0·5 µg ml¹, not shown). Increasing [K⁺] from 0 to 40 mM increased the uptake of oxonol (and hence the fluorescence signal of the organisms) by partial depolarization of the plasma-membrane potential. Similar results were obtained in duplicate experiments.

(37K): Fig. 4. Flow cytometry of M. punctachora. Selective analysis of the largest organisms enables assessment of the effects of agents that perturb membrane potential. Organisms incubated with 1 µM DiBAC4(3) in the absence (a, b) or in the presence (c, d) of 100 µM DCCD.

(24K): Fig. 5. Effects of K+ on plasma-membrane potential in M. punctachora. KCl was added at varying concentrations to organisms suspended in N-methyl-D-glucamine chloride-containing solutions. Also shown are the effects of agents that perturb plasma-membrane function. The membrane potentials were estimated by the flow cytometric procedure at 0·5 µM DiBAC4(3). Results are means of two experiments.

Of the alternatives available for assessment and quantification of the transmembrane potential of the limiting lipoprotein membrane of an organism, microelectrode-based techniques (patch-clamping) are difficult for small organisms (<10 µm). Radioactive labels (charged permeant hydrophobic markers, e.g. tetraphenyl/phosphonium phosphate⁺; Kamo et al., 1979) have often been employed, but may be extruded by efflux pumps (Midgley, 1986). Thus, the use of charge-sensitive dyes that sense and indicate trans-electrochemical membrane potential presents many advantages. Both the slow voltage-sensitive cationic and anionic dyes have been used with bacteria and lower eukaryotes. One disadvantage of the cationic dyes is that they are often toxic (e.g. Rhodamine 123, the most popular first-generation dye employed for the measurement of bacterial or mitochondrial inner-membrane potential, inhibits respiration). Efflux has frequently been observed (Mason, 1994). Uptake into the mitochondria of intact organisms makes these dyes unsuited for measurements of the plasma-membrane potential of mitochondriate eukaryotes. Newer methods using a ratiometric method for a cyanine dye, which work well for some bacteria (Shapiro, 2000; Novo et al., 1999), are also not useful here. The advantages of the use of an anionic dye, such as DiBAC₄(3), include: (i) it shows limited uptake; (ii) it is not subject to efflux pump extrusion; (iii) it is itself non-toxic; (iv) it does not block ion channels; (v) it shows no aggregation; (vi) it does not contribute to the cellular signal in flow cytometric measurements; and (vii) where tested this method shows excellent correlation with patch-clamp measurements, e.g. for mammalian lymphocytes (Emri et al., 1998). However, binding of this dye to proteins does lead to altered fluorescence properties (enhancement of quantum efficiency). This may lead to a non-linear calibration and to overestimation of membrane-potential values, especially at higher dye concentrations. For this reason, when using the flow cytometric method, the most appropriate control treatment must be selected, so as to provide organisms with zero membrane potential and with unchanged volume. For this, the relative merits of aldehyde fixation (Krasznai et al., 1995; Emri et al., 1998) or heat-killed organisms (Lloyd et al., 2000; Biagini et al., 2000) have been outlined.

In this study, we have shown that the plasma-membrane potential provides a sensitive and convenient indicator of the vitality of several microaerophilic protist populations (Lloyd, 1993). Moreover, the use of flow cytometry enables resolution of heterogeneous populations, where damage affects only some of the organisms. It also enables experiments on non-axenic cultures as, for example, with M. punctachora as shown here. This is important to distinguish (a) damaged from intact organisms, and (b) the signals emitted from protists in non-axenic cultures. Assessment using fluorometric measurement of DiBAC₄(3) emission intensity has been used to determine injury by oxidative or nitrosative stress. Table 1 indicates their scavenging abilities as measured by the K_m O₂ values for O₂ consumption and the threshold values, above which O₂ becomes inhibitory to these oxygen-uptake systems. Previous work has shown that the plasma-membrane potential of G. intestinalis is maintained at 134±3 mV by a K⁺ diffusion pathway and an electrogenic H⁺ pump (Biagini et al., 2000); in this organism, sensitivity to reactive oxygen species (Lloyd et al., 2000, 2002b) and to nitrosative stress (Ryu & Lloyd, 1995; Lloyd et al., 2002a, 2003a) has been demonstrated. In M. punctochora, the mechanisms whereby the plasma-membrane potential is generated appear to be similar to those seen in G. intestinalis (G. Biagini & C. Bernard, unpublished data). Table 1 also compares the physiological characteristics of the microaerophilic species studied with respect to oxidative and nitrosative stress.

Table 1. Characteristics of some anaerobic protists

D. L. was visiting Professor in the University of New South Wales during some of this work, and thanks the Royal Society for a travel grant and K. Edwards for micrographs of M. punctachora.

Footnotes

†Present address: Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L35 QA, UK.