INTRODUCTION
Gamma-aminobutyric acid (GABA) is the major neurotransmitter responsible for synaptic inhibition in the mammalian central nervous system where it mediates a rapid inhibitory transmission by activating ionotropic receptors (GABAA and GABAC) coupled to chloride channels (Barnard et al. Reference Barnard, Skolnick, Olsen, Mohler, Sieghart, Biggio, Braestrup, Bateson and Langer1998). In invertebrates, GABA has been reported to be present in the nervous system of several phyla (Walker et al. Reference Walker, Brooks and Holden-Dye1996). Contrasting with the existence of numerous studies in nematodes (Johnson and Stretton, Reference Johnson and Stretton1987; Martin, Reference Martin1987; McIntire et al. Reference McIntire, Jorgensen, Kaplan and Horvitz1993; Walker et al. Reference Walker, Brooks and Holden-Dye1996), the physiological actions of GABA in platyhelminths remains relatively poorly explored, with limited reports of GABA-immunoreactivity in the central and peripheral nervous system of several species, including the cestode Moniezia expansa, planarian Girardia tigrina and trematode Fasciola hepatica (Eriksson and Panula, Reference Eriksson and Panula1994; Eriksson et al. Reference Eriksson, Maule, Halton, Panula and Shaw1995) and depression of the electrically evoked activity in the submuscular ventral longitudinal nerve cords of Notoplana acticola (Keenan et al. Reference Keenan, Koopowitz and Bernardo1979). In schistosomes, the trematode responsible for human schistosomiasis, there was no direct evidence for a GABA transmission pathway until our recent description of GABA immunoreactivity (GABA-IR) in nerve cells and fibres of the cerebral ganglia and longitudinal nerve cords and the nerve plexuses ramifying throughout the parenchyma of male adult Schistosoma mansoni. In addition, strong GABA-IR was also found associated with the oral and ventral suckers as well as in testes, indicating a role for GABA in fixation to the host vascular wall and spermatogenesis (Mendonça-Silva et al. Reference Mendonça-Silva, Gardino, Kubrusly, de Mello and Noël2004). Furthermore, measurement of a glutamate decarboxylase (GAD) activity and Western blot analysis with antibodies raised against mammalian GAD showed that S. mansoni was able to synthesize GABA from glutamate (Mendonça-Silva et al. Reference Mendonça-Silva, Gardino, Kubrusly, de Mello and Noël2004). Thus, GABA is to be considered a putative neurotransmitter in S. mansoni, although not all classical criteria that define a neurotransmitter have been met hitherto, such as a detectable influence of GABA on any neurophysiological end-point and the identification of a specific receptor. Such receptors could be GABAA-like ionotropic receptors since this is the type of GABA receptor that is the most widely studied in invertebrates (Walker et al. Reference Walker, Brooks and Holden-Dye1996) and have been well characterized in somatic smooth muscles of A. suum (Holden-Dye et al. Reference Holden-Dye, Krogsgaard-Larsen, Nielsen and Walker1989). Furthermore, our recent data showing that picrotoxin, a classical blocker of the GABAA channel, is able to modify the motor activity of adult S. mansoni support this idea (Mendonça-Silva et al. Reference Mendonça-Silva, Gardino, Kubrusly, de Mello and Noël2004).
The objective of the present work was to search for, and characterize, GABAA receptors in adult S. mansoni using [3H]-flunitrazepam, a classical tool for labelling the allosteric benzodiazepine site present on the (mammalian) GABAA receptor/channel complex (central benzodiazepine receptor). This strategy was also stimulated by the existence of one study describing the characteristics of a benzodiazepine binding site in S. mansoni (Bennett, Reference Bennett1980) and the necessity to revise this relatively old work in the light of the recent knowledge on benzodiazepine binding sites. In fact, enormous advances have been made in the molecular biology and pharmacological analysis of central benzodiazepine receptor subtypes (Barnard et al. Reference Barnard, Skolnick, Olsen, Mohler, Sieghart, Biggio, Braestrup, Bateson and Langer1998; Möhler et al. Reference Möhler, Crestani and Rudolph2001). Also of importance is the recent discovery of peripheral benzodiazepine receptors that constitute a totally distinct molecular entity (Gavish et al. Reference Gavish, Bachman, Shoukrun, Katz, Veenman, Weisinger and Weizman1999), mainly present in the internal and external mitochondrial membranes where it is part of the mitochondrial permeability transition pore (McEnery et al. Reference McEnery, Snowman, Trifiletti and Snyder1992; Papadopoulos et al. Reference Papadopoulos, Boujrad, Ikonomovic, Ferrara and Vidic1994). Finally, our work characterizes a new putative target for the development of anti-parasitic agents related to the neuromuscular system of this flatworm.
MATERIALS AND METHODS
Preparation of subcellular fractions of S. mansoni
The investigation was in accord with the Institutional Ethical Committee for animal care (Federal University of Rio de Janeiro, Brazil). Adult male S. mansoni (BH strain) were obtained from mice, sacrificed by cervical dislocation, that have been infected 45 days earlier with approximately 150 male cercariae each, as previously described (Cunha et al. Reference Cunha, Mever-Fernandes and Noël1992). About 1500 male worms were then homogenized in a Dounce homogenizer at 4°C in 0·25 m sucrose solutions (5 mmTris (hydroxymethyl) aminomethane, Tris-HCl, pH 7·4) using 3 sequences of 10 passes of the pestle. The homogenate was centrifuged to obtain 4 pellets (P1, P2, P3, P4) sedimenting respectively at 300 gav (5 min); 1000 gav (10 min); 8000 gav (10 min) and 100 000 gav (1 h). These fractions have been previously characterized by electronic microscopy as heterogeneous (P1), nuclear (P2), mitochondrial (P3) and microsomal (P4) (Cunha et al. Reference Cunha, De Souza and Noël1988). The pellets were resuspended in buffered sucrose solution and stored at −70°C until use. Protein concentration was determined by the method of Bradford (Reference Bradford1976), using bovine serum albumin as standard.
Preparation of rat brain synaptosomes
Brains, without cerebellum and brain stem, were obtained from adult male Wistar rats sacrificed by decapitation. Briefly, tissues were homogenized in a Potter apparatus with a motor-driven Teflon pestle at 4°C in 15 volumes of ice-cold 0·32 m buffered sucrose (pH 7·4) per gram of organ. After centrifuging at 1000 gmax for 10 min, the supernatant was centrifuged at 48 000 gav for 20 min to obtain the crude synaptosomes that were resuspended in buffered Krebs solution (1·4 ml/g tissue) and stored at −80°C until use. Protein concentration was determined by the method of Lowry et al. (Reference Lowry, Rosebrough, Farr and Randall1951) using bovine serum albumin as standard.
Preparation of rat kidney mitochondrial fraction
Kidneys from adult male Wistar rats were homogenized in a Potter apparatus with a motor-driven Teflon pestle at 4°C in 10 volumes of ice-cold 50 mm Tris-HCl (pH 7·4) per gram of organ. After centrifuging at 500 gav for 10 min, the supernatant was centrifuged at 15 000 gav for 15 min. After resuspension of the pellet in 3 volumes of 50 mm Tris-HCl (pH 7·4), the material was centrifuged at 50 000 gav for 20 min to obtain the mitochondrial fraction that was resuspended in the buffer (3 ml/g tissue) and stored at −80°C until use.
[ 3H]flunitrazepam binding assay
P1-P4 fractions of S. mansoni (100–200 μg protein) or rat synaptosomes (200 μg protein) were incubated at 4°C for 90 min in a buffered Krebs solution containing 0·2 nm or 1 nm [3H]flunitrazepam (71–85 Ci/mmol, New England Nuclear Life Science Products, USA). After incubation, samples were rapidly diluted with 3 ml of ice-cold Krebs buffer and immediately filtered on glass fibre filters (GMF 3, Filtrak, Germany) under vacuum. Filters were then washed once more with 3 ml of buffer, dried, immersed in a scintillation mixture [1,4-bis-[2-(5-phenyloxazolyl)]-benzene, POPOP, (0·1 g/l) and 2,5-diphenyloxazole, PPO, (4·0 g/l) in toluene], and the radioactivity retained in the filters was counted with a Packard Tri-Carb 1600 TR liquid scintillation analyser. Saturation experiments were performed by the addition of increasing concentrations of unlabelled flunitrazepam (0·1 nm to 5·0 nm) to 0·2 nm [3H]flunitrazepam. This protocol, classically referred to as a competition experiment, has already been used with the same radioligand (Lopes et al. Reference Lopes, Caruso, Castro, Costa, da Silva and Noël2004) and is of particular interest for binding studies with S. mansoni since it allows a great economy of radioligand and protein (Mendonça-Silva et al. Reference Mendonça-Silva, Pessôa and Noël2002). Non-specific binding was estimated in the presence of 5 μm unlabelled flunitrazepam.
[ 3H]PK11195 binding assay
P1-P4 fractions of S. mansoni (100–200 μg protein) or rat kidney mitochondrial fraction (100 μg protein) were incubated at 4°C for 60 min in 50 mm Tris-HCl buffer (pH 7·4) containing 0·5 nm [3H]PK11195 (83·5 Ci/mmol, New England Nuclear Life Science Products, USA), a selective antagonist of the peripheral binding receptor that directly interacts with its ‘isoquinoline-binding protein’ subunit. After incubation, samples were rapidly diluted with 4 ml of ice-cold buffer and immediately filtered on glass fibre filters (GMF 3, Filtrak, Germany) under vacuum. Filters were then washed twice more with 4 ml of buffer and treated as described above. Non-specific binding was estimated in the presence of 10 μm unlabelled PK11195.
Statistics
Statistical comparisons of recoveries of flunitrazepam or PK11195 binding in the different subcellular fractions were determined by using ANOVA followed by Bonferroni test for multiple comparisons. Binding data from saturation experiments were presented as a classical Scatchard plot. The model parameters (B max and K d, for saturation assays; IC50, for competition assays) were estimated using a computerized non-linear regression analysis of the untransformed data (Prism 4.0, GraphPad Software Inc.), assuming a single population of binding sites. The time-course of flunitrazepam binding was analysed using the model of one phase exponential association.
Drugs
Bicuculline, DMCM, GABA, flunitrazepam, pentobarbital, picrotoxin, PK11195 and Ro5-4864 were obtained from Sigma-Aldrich Chemicals (Milwaukee, WI, USA). Clonazepam, diazepam, midazolam and zolpidem were kindly supplied by Roche Indústrias Químicas (Brazil). Flumazenil was from Cristália Produtos farmacêuticos. DMCM, flunitrazepam, PK11195, Ro5-4864, clonazepam, diazepam and midazolam were dissolved in 100% ethanol whereas flumazenil and picrotoxin were dissolved in 100% dimethyl sulfoxide (DMSO) to obtain stock solutions (usually 30 mm) and further diluted in water or Krebs solution. Pentobarbital was dissolved in water (5 mm). At the final concentrations used, ethanol and DMSO had no effect in our assays.
RESULTS
[ 3H]flunitrazepam binding
Specific binding sites for 1 nm [3H]flunitrazepam were detected in all subcellular fractions (P1-P4), but a higher specific binding was obtained in the mitochondrial (P3) fraction (Table 1). Therefore this fraction was used to further characterize the binding of [3H]flunitrazepam to S. mansoni. As [3H]flunitrazepam binding to its allosteric site present on the GABAA receptor complex of mammalian brain is sensitive to temperature (Maguire et al. Reference Maguire, Villar, Davies and Loew1992), we compared the binding of 1 nm [3H]flunitrazepam to the P3 fraction of S. mansoni at 4°C, 25°C and 37°C (Table 2). The binding was significantly higher at the lowest temperature, in good accordance with the increase of affinity observed in rat brain when lowering the temperature (Maguire et al. Reference Maguire, Villar, Davies and Loew1992). These data also support the use of this temperature (4°C) throughout our study in S. mansoni, as is generally the case for studies in mammals. Figure 1 shows the time-course of 0·2 nm [3H]flunitrazepam binding to P3 in our standard conditions (4°C). [3H]flunitrazepam binding is very rapid (half-life=5 min) reaching a plateau after about 30 min and remaining stable during the 90-min incubation period. After having determined these optimal experimental conditions, we studied the concentration-dependence of flunitrazepam binding to S. mansoni by incubating fraction P3 at 4°C during 90 min, in the presence of 0·2–5 nm of the ligand (see Materials and Methods section for details). As shown in Fig. 2 for a typical experiment, the Scatchard plot was linear, indicating the labelling of a single homogeneous population of receptors. These experiments were repeated twice in order to obtain good estimations of the parameters K d and B max, by averaging the values calculated independently in 3 different preparations (Table 3). Flunitrazepam binds in the nanomolar range of concentration, i.e with a relatively high affinity, to a very large number of sites. When compared to rat brain synaptosomes (Table 3), a preparation enriched in GABAA receptors, the density of flunitrazepam binding sites is about 4 times higher in the mitochondrial fraction of S. mansoni whereas the affinity is about 15 times lower. In order to further characterize the flunitrazepam binding sites present in S. mansoni, we firstly tested the effect of allosteric modulators of the mammalian benzodiazepine binding site present on the GABAA receptor complex (central benzodiazepine receptor). Neither GABA (1 mm) nor pentobarbital (100 μm) altered the binding of [3H]flunitrazepam to P3, in contrast to observations in mammalian brain preparations washed from endogenous GABA (Lopes et al. Reference Lopes, Caruso, Castro, Costa, da Silva and Noël2004). Bicuculline (30 μm) and picrotoxin (100 μm), which modulate the binding of benzodiazepine to mammalian brain preparations in the presence of GABA (Lopes et al. Reference Lopes, Caruso, Castro, Costa, da Silva and Noël2004), were also without effect in our preparation from schistosomes. As the pharmacological modulation of a binding site is a powerful tool for its classification, we performed whole competition curves for [3H]flunitrazepam binding to the P3 fraction of S. mansoni using different putative ligands of the benzodiazepine receptor and compared their IC50 values. Figure 3 shows competition curves for 4 classical ligands of the mammalian central benzodiazepine receptors. The IC50 values for zolpidem, flunitrazepam and diazepam were not very different from the values observed in rat brain, being about 2–58 times higher (Table 4). On the other hand, the IC50 for clonazepam (4940 nm) was surprisingly very high, about 5000 times higher than in rat brain. The fact that clonazepam is particular in that it is very selective for the central benzodiazepine receptors in mammals (it has a very low affinity for the peripheral benzodiazepine receptors (Hirsch et al. Reference Hirsch, Beyer, Malkowitz, Loullis and Blume1988)), together with the lack of effect of GABA and other classical modulators of the central benzodiazepine receptor led us to investigate if the [3H]flunitrazepam binding sites detected in S. mansoni could correspond to peripheral-like binding sites instead of central-like benzodiazepine binding sites. Firstly, we tested flumazenil, another benzodiazepine which is very selective for the mammalian central benzodiazepine receptor; in this case, no effect was observed even using a concentration as high as 10 000 nm (Table 4). A second approach was to assay 2 ligands very selective for the mammalian peripheral benzodiazepine receptor, namely PK11195 and Ro5-4864 (Maguire et al. Reference Maguire, Villar, Davies and Loew1992). PK11195 exhibited an IC50 value (4710 nm) similar to the one measured in rat brain for competition to the central benzodiazepine receptor (19 400 nm, Table 4). On the other hand, the IC50 value of Ro5-4864 was much lower in S. mansoni (110 nm) than in rat brain (33 100 nm) being even similar to the IC50 measured for binding to the mammalian peripheral benzodiazepine receptor present in rat kidney (25 nm, Table 5). This result is compatible with the idea that [3H]flunitrazepam could be labelling a peripheral-like benzodiazepine receptor in S. mansoni. On the other hand, the IC50 value of PK11195 should be too high for such peripheral sites, based on comparison with values for rat kidney (Hirsch et al. Reference Hirsch, Beyer, Malkowitz, Loullis and Blume1988 and Table 5), indicating that PK11195 is here competing with [3H]flunitrazepam for a schistosomal central benzodiazepine receptor. In an attempt to directly challenge this hypothesis, we decided to search for PK11195 binding sites with higher affinity (hence, presumably of the peripheral type), using [3H]PK11195 as a radioligand.
Time-course of [3H]flunitrazepam binding to the Schistosoma mansoni P3 fraction. Approximately 100 μg protein were incubated at 4°C in a modified Krebs solution containing 1 nm [3H]flunitrazepam. The theoretical curve was obtained by non-linear regression analysis using the model of one phase exponential association. Each point represents the mean±s.e.m. obtained from experiments performed in triplicate with 3 different preparations.
Typical Scatchard plot for [3H]flunitrazepam binding to the P3 fraction of Schistosoma mansoni. The curve was drawn using the parameters fitted by non-linear regression analysis with the model of a single class of independent binding sites. Each point represents the mean of triplicate determinations in a typical experiment. Non-specific binding accounted for 10–30% of total [3H]flunitrazepam binding.
Competition curves for 0·2 nm [3H]flunitrazepam binding to P3 fraction of Schistosoma mansoni. Flunitrazepam (●), zolpidem (▾), diazepam (■) and clonazepam (▲). The curves were drawn using the parameters fitted by non-linear regression analysis using the model of a single class of independent binding sites. Each point represents the mean±s.e.m. obtained from 2–4 experiments performed in quadruplicate.
Subcellular distribution of 1 nm [3H]flunitrazepam binding in subcellular fractions of Schistosoma mansoni
† % Recovery=100×binding (specific binding×protein content) in each individual fraction divided by the sum of the binding of the 4 fractions. Values represent means±s.e.m. from 3 different preparations, each experiment being performed in triplicate.
* P<0·05, ANOVA followed by Bonferroni test (P3vs P1, P2 and P4 fractions).
+ P<0·05, ANOVA followed by Bonferroni test (P4vs P1 and P2).
Effect of temperature on [3H]flunitrazepam binding to the P3 fraction of Schistosoma mansoni
(Approx. 100 μg protein were incubated for 90 min at different temperatures in a modified Krebs solution containing 1 nm [3H]flunitrazepam. Values represent means±s.e.m. from 3 different preparations, each experiment being performed in triplicate.)
* P<0·05, ANOVA followed by Bonferroni test (4°C vs 25°C and 37°C).
Equilibrium constants for the specific binding of [3H]flunitrazepam
(Values of parameters represent means±s.e.m. of 3 individual experiments performed in triplicate.)
Inhibition of [3H] flunitrazepam binding in rat brain synaptosomes and in P3 fraction of Schistosoma mansoni by different ligands of benzodiazepine receptors
(IC50 values were calculated by non-linear regression analysis of competition curves like those shown in Fig. 3. Drugs were incubated along with 0·2 nm [3H] flunitrazepam at 4°C for 90 min in a buffered Krebs solution in the presence of 100–200 μg protein of rat synaptosomes or P3 fraction of S. mansoni. #No inhibition was obtained even at 10 μm flumazenil in S. mansoni, indicating that the IC50 should be much higher than this concentration, or that flumazenil has no affinity at all for these binding sites. In mammals, clonazepam and flumazenil are selective ligands of the central benzodiazepine receptors whereas PK11195 and Ro5-4864 are selective for the peripheral benzodiazepine receptors (see Table 5).)
Affinities (IC50) of various benzodiazepine ligands for central and peripheral benzodiazepine receptors of the rat
(IC50 values were calculated by non-linear regression analysis of competition curves for [3H] flunitrazepam binding in rat brain synaptosomes (central benzodiazepine receptor) or for [3H]PK11195 binding in rat kidneys (peripheral benzodiazepine receptor). The ratio (PK/FLU) is a measure of the selectivity between these 2 benzodiazepine receptors: values higher than 1 indicates a selectivity for central benzodiazepine receptors and values lower than 1 indicate a selectivity for peripheral benzodiazepine receptors.)
a Less than 40% inhibition was obtained even at 30 μm flumazenil, indicating that the IC50 should be higher than this concentration.
Subcellular distribution of 0·5 nm [3H]PK11195 binding in subcellular fractions of Schistosoma mansoni
(Values represent means±s.e.m. from 3 different preparations, each experiment being performed in triplicate.)
† % Recovery=100×binding (specific binding×protein content) in each individual fraction divided by the sum of the binding of the 4 fractions.
* P<0·05, ANOVA followed by Bonferroni test (P1vs P2 and P3 fractions).
[ 3H]PK11195 binding
Specific binding sites for 0·5 nm [3H]PK11195 were detected in all subcellular fractions (P1-P4) with a similar value of specific binding (Table 6). Here too, the mitochondrial (P3) fraction was chosen for further characterization of these binding sites. The best way to study the pharmacological modulation of a binding site is to compare the IC50 values calculated from whole competition curves performed with different putative ligands. Unfortunately, such a protocol was not possible here due to the technical limitation imposed by the very low specific binding (only about 20–50% of total binding at 0·5 nm [3H]PK11195, in the absence of competitor). This low level of specific binding also explains why we did not obtain sufficiently reliable saturation curves in order to calculate K d and B max in P3 (5 assays) but indicates that the number of [3H]PK11195 binding sites in P3 fraction is lower than of [3H]flunitrazepam binding sites, the first indication that the two ligands label different sites. Note that the scarcity of material (3·17±0·26 mg protein of P3 per 1000 worms, n=15) was an additive constraint throughout the present work. As a surrogate, we tested only 1–3 concentrations of each putative ligand, near their IC50 values that we previously estimated in a screening assay. Figure 4 shows that we were able to obtain good estimates of IC50 for all drugs tested with the exception of flumazenil, which produced only 10% inhibition at 10 μm. As a whole, the IC50 values for [3H]PK11195 binding in our schistosomal fraction were very similar to the values for the mammalian peripheral benzodiazepine receptor, as measured in the rat kidney preparation (compare Table 5 with Table 7). Only flunitrazepam and Ro5-4864 exhibited higher IC50 values in schistosomes than in rat (17 and 28 times higher, respectively). The fact that we were able to obtain good estimates of IC50 values for both [3H]flunitrazepam and [3H]PK11195 in schistosome and rat preparations, in the same experimental conditions, allowed us to compare the selectivity of these drugs for the 2 benzodiazepine binding sites, in both species. Comparison of the ratio of PK11195/flunitrazepam IC50 values (Tables 5 and 7) indicated that no drug was very selective in S. mansoni, since only PK11195 exhibited a noticeable selectivity (55 times) for the ‘peripheral’ binding site (ratio=0·018). Note that this drug had a 1140 times selectivity for the peripheral benzodiazepine receptor in the rat (ratio=0·00088) just as R05-4864 (ratio=0·00076). On the other hand, clonazepam that was 11 500 times more selective for the central benzodiazepine receptor in the rat (ratio=11500) was nearly equipotent for competing at both [3H]flunitrazepam and [3H]PK11195 binding sites, in schistosomes (ratio=6). With respect to IC50 values for the mammalian peripheral benzodiazepine receptors (Table 5), we performed whole competition curves for 0·5 nm [3H]PK11195 binding in our rat kidney mitochondrial fraction. Note that this preparation contains a relatively high density of peripheral benzodiazepine receptors, as calculated from saturation curves performed either with [3H]PK11195 (B max=6·6 pmol . mg protein−1, n=1) or with [3H]flunitrazepam (B max=8·26±1·04 pmol . mg protein−1, n=3), allowing good technical conditions (specific binding about 80% at 0·5 nm [3H]PK11195).
Pharmacological modulation of 0·5 nm [3H]PPK11195 binding to P3 fraction of Schistosoma mansoni. Each point represents the mean±s.e.m. obtained from 1–4 experiments performed in quadruplicate. Non-specific binding accounted for 48–82% of total [3H]flunitrazepam binding. FLU=flunitrazepam.
Affinities (IC50) of various benzodiazepine ligands for ‘central’ and ‘peripheral’ benzodiazepine receptors of Schistosoma mansoni
(IC50 values were calculated by non-linear regression analysis of competition curves for [3H]flunitrazepam (‘central’-like benzodiazepine receptor) or roughly estimated from partial inhibition assays (see Fig. 4) for [3H]PK11195 binding (‘peripheral’-like benzodiazepine receptor) in the P3 fraction of S. mansoni. The ratio (PK/FLU) is a measure of the selectivity between these two benzodiazepine receptors: values higher than 1 indicates a selectivity for central benzodiazepine receptors and values lower than 1 indicate a selectivity for peripheral benzodiazepine receptors.)
a No inhibition was obtained at 10 μm flumazenil, indicating that the IC50 should be much higher than this concentration.
b Only 10% inhibition was observed at 1 μm flumazenil, indicating that the IC50 should be much higher than this concentration.
DISCUSSION
Our recent evidence for the presence of a GABAergic neurotransmission in the trematode S. mansoni (Mendonça-Silva et al. Reference Mendonça-Silva, Gardino, Kubrusly, de Mello and Noël2004) led us to investigate the presence of ionotropic GABAA receptors using [3H]-flunitrazepam, a classical tool for labelling the allosteric benzodiazepine site (central benzodiazepine receptor) present on the mammalian GABAA receptor/channel complex. The specific binding of [3H]flunitrazepam was higher in the P3 fraction of adult male S. mansoni, a fraction previously characterized as a mitochondrial fraction (Cunha et al. Reference Cunha, De Souza and Noël1988), not coinciding with the binding sites for kainic acid, which labels another putative amino acid neurotransmitter receptor (glutamate), that have been previously detected only in the P1 fraction (Mendonça-Silva et al. Reference Mendonça-Silva, Pessôa and Noël2002). This was the first reason to imagine that the benzodiazepine binding sites present in this worm could be different from the classical central benzodiazepine receptors of mammals and even to suspect flunitrazepam labelling of peripheral benzodiazepine binding sites, initially discovered in mitochondria (Gavish et al. Reference Gavish, Bachman, Shoukrun, Katz, Veenman, Weisinger and Weizman1999; McEnery et al. Reference McEnery, Snowman, Trifiletti and Snyder1992). This hypothesis had to be seriously considered since flunitrazepam is known to be poorly selective for the central benzodiazepine receptors, labelling the peripheral benzodiazepine receptors present in rat kidney mitochondria with a IC50 of about 30 nm i.e. a value very similar to that measured in our P3 fraction of S. mansoni (41 nm). Note that IC50 values reported in the literature for flunitrazepam binding to peripheral benzodiazepine receptors vary largely, from 20 to 500 nm (Hirsch et al. Reference Hirsch, Beyer, Malkowitz, Loullis and Blume1988; Lueddens and Skolnick, Reference Lueddens and Skolnick1987), reinforcing the importance of comparing drugs and species in experiments performed under the very same conditions, as we did here. Another strong argument apparently supporting flunitrazepam labelling of peripheral benzodiazepine receptors was the pharmacological profile of [3H]flunitrazepam binding that was very different from that expected for the ‘classical’ (mammalian) central binding site present on the GABAA receptor/channel complex and much more similar to that observed for the peripheral benzodiazepine receptor (compare first column of Table 7 to first and second columns of Table 5). On the other hand, zolpidem and, more significantly, PK11195 IC50 values did not fit with this hypothesis. The IC50 value for PK11195 (4710 nm) was more similar to the value for the rat central benzodiazepine receptor (19400 nm) than for the rat peripheral benzodiazepine receptor (17 nm). Another argument favouring the initial hypothesis that [3H]flunitrazepam was effectively labelling a central-like benzodiazepine receptor, and not a peripheral-like receptor, was the temperature-dependence of the binding, with higher affinity at lower temperature. Finally, the direct evidence that [3H]PK11195 labelled a different population of benzodiazepine binding sites, with a somewhat different subcellular distribution, (apparently) lower density, high affinity for PK11195 and a pharmacology very similar to the rat peripheral benzodiazepine receptor supports the existence of 2 distinct populations of benzodiazepine binding sites in adult S. mansoni. We will now discuss the characteristics of these 2 populations, in the light of our knowledge on mammalian and invertebrate benzodiazepine receptors.
The first population of schistosomal benzodiazepine binding sites should correspond to a central-like benzodiazepine receptor, albeit significantly different in various aspects to the mammalian site present on the GABAA receptor/channel complex, as discussed below. (1) Preferential localization in the mitochondrial fraction. (2) Lack of effect of GABA and other classical modulators of the central benzodiazepine receptor like pentobarbital, bicuculline and picrotoxin. (3) Singular pharmacological modulation, that does not fit well with a mammalian central (nor a peripheral) benzodiazepine receptor, due to a very high affinity for Ro5-4864 and a very low affinity for clonazepam. These differences of affinities could be explained by putative structural peculiarities of the worm receptors as a function of the large phylogenetic distance between mammals and platyhelminths. Note that platyhelminths are the first metazoan group to possess a centralized nervous system, being considered as the link between lower and higher invertebrates (Day and Maule, Reference Day and Maule1999). Indeed, numerous authors reported differences between mammalian and invertebrate GABAA receptors (Walker et al. Reference Walker, Brooks and Holden-Dye1996). As an example, the pharmacological profile of the benzodiazepine sites present in GABA-activated chloride channels in insects is more similar to mammalian peripheral than central benzodiazepine receptors (Sattelle et al. Reference Sattelle, Lummis, Wong and Rauh1991). In the somatic musculature of the nematode Ascaris suum the GABA receptors are insensitive to the antagonists bicuculline and picrotoxin and to allosteric modulators including the benzodiazepines and barbiturates (Holden-Dye et al. Reference Holden-Dye, Krogsgaard-Larsen, Nielsen and Walker1989). At the molecular level, such differences could be due to a different arrangement of the subunits constitutive of the pentameric receptor complex, since the benzodiazepine binding site is localized at the interface of the α and γ subunits and their affinities depend on the composition of these 2 subunits (Smith and Olsen, Reference Smith and Olsen1995). The combination of a α1, α2, α3, or α5 subunit with a γ2 or γ3 subunit, for example, is necessary for a high affinity (McKernan et al. Reference McKernan, Wafford, Quirk, Hadingham, Harley, Ragan and Whiting1995; Barnard et al. Reference Barnard, Skolnick, Olsen, Mohler, Sieghart, Biggio, Braestrup, Bateson and Langer1998) whereas receptors formed with α4 or α6 are virtually insensitive to most benzodiazepines. The peculiar pharmacology of the central-like benzodiazepine receptor of S. mansoni could also be partially explained by the presence of an ε subunit as suggested by analysis of the S. mansoni transcriptome (Verjovski-Almeida et al. Reference Verjovski-Almeida, DeMarco, Martins, Guimarães, Ojopi, Paquola, Piazza, Nishiyama, Kitajima, Adamson, Ashton, Bonaldo, Coulson, Dillon, Farias, Gregorio, Ho, Leite, Malaquias, Marques, Miyasato, Nascimento, Ohlweiler, Reis, Ribeiro, Sá, Stukart, Soares, Gargioni, Kawano, Rodrigues, Madeira, Wilson, Menck, Setubal, Leite and Dias-Neto2003), since this subunit can substitute the γ2 in the pentameric complex being responsible for an atypical binding of the benzodiazepines, as reported in the mammalian locus coeruleus (Moragues et al. Reference Moragues, Ciofi, Tramu and Garret2002) for example. Note that we looked in the genomic (http://www.genedb.org/genedb/smansoni/) and EST (Schistosoma mansoni EST Genome Project website at: http://cancer.lbi.ic.unicamp.br/schisto6/) data sets and found 2 pieces of indirect evidence for a central-type benzodiazepine receptor: an EST compatible with the above-cited ε subunit of the pentameric GABAA receptor complex (SmAE 709016.1) as well as an EST (SmAE 604360.1) and a coding sequence (Smp_073790) compatible with a GABAA receptor associated protein.
The sites labelled with [3H]flunitrazepam in our P3 fraction of S. mansoni are probably similar to those detected by Bennett in his pioneering, and unique, work on benzodiazepine binding sites in schistosomes, using [3H]clonazepam and [14C]3-methylclonazepam as probes (Bennett, Reference Bennett1980). Although this author worked in a non-equilibrium condition (3 min incubation) and at a higher temperature (37°C), that decreases the affinity of flunitrazepam to its binding sites, the IC50 reported for clonazepam and diazepam (4 μm and 6 μm, respectively) are roughly similar to our values.
The second population of schistosomal benzodiazepine binding sites should correspond to a peripheral-like benzodiazepine receptor, with pharmacology very similar to the mammalian correspondent. (1) The IC50 values for [3H]PK11195 binding in our schistosomal fraction are very similar to the values for the mammalian peripheral benzodiazepine receptor, as measured in the rat kidney preparation. (2) The sequence of potency, a classical pharmacological tool for receptor classification, is nearly identical to that characterizing the mammalian peripheral benzodiazepine receptor (PK11195=Ro5-4864=Flunitrazepam>Diazepam≈Zolpidem≫Clonazepam>Flumazenil) but not the mammalian central benzodiazepine receptor (Clonazepam=Flumazenil=Flunitrazepam>Diazepam>Zolpidem≫ PK11195=R05-4864). Note that a gene encoding a protein with homology to the human peripheral benzodiazepine receptor has been cloned supporting our hypothesis (Santos et al. Reference Santos, Johnston, Azevedo, Ridgers, Martinez, Marotta, Santos, Fonseca, Ortega, Rabelo, Saber, Ahmed, Romeih, Franco, Rollinson and Pena1999). With respect to the localization and functional role of such receptors, we can just speculate that they could be localized in the mitochondria or even in the tegument, based on their main recovery (55%) in the P1 fraction, a heterogeneous fraction containing pieces of tegument. Note that there is evidence for the presence of mammalian peripheral benzodiazepine receptors in the plasma membrane of different cells (Olson et al. Reference Olson, Ciliax, Mancini and Young1988; Woods and Williams, Reference Woods and Williams1996).
As a conclusion, this work not only contributes to the basic knowledge of S. mansoni biology but also should enable investigations on the likelihood that these worm benzodiazepine receptors are involved in the proposed GABA neurotransmission and are the molecular targets for the schistosomicidal effect of clonazepam and 3-methylclonazepam (Stohler, Reference Stohler, Siegenhaler and Luthy1978). We are now investigating this relevant question for the putative development of new anti-parasitic agents related to the neuromuscular system of this flatworm (Geary et al. Reference Geary, Klein, Vanover, Bowman and Thompson1992).
The authors wish to thank Dr Lygia dos Reis Corrêa (Instituto Oswaldo Cruz, Rio de Janeiro) for kindly providing the infected mice. This work was supported by CNPq (Nr. 420015/05-1 (IM-INOFAR) and Nr. 472638/2004-1) and FAPERJ (Nr. E-26/153·239/2004) grants; F. Noël is recipient of a CNPq research fellowship.
